Does Democracy Enhance or Reduce Lethal Violence? Examining the Role of the Rule of Law

Abstract

This study builds on prior cross-national criminological literature by using disaggregated measures of democracy, notably rule of law, to examine the influence key components of democracy have on homicide rates. To assess this relationship, the current study uses two measures of rule of law: (a) a measurement of an independent judiciary; (b) a measurement of “Law and Order Tradition” from the International Country Risk Guide (ICRG). Findings indicate that the measures of rule of law have a significant negative effect on homicide rates.

Keywords

democracy comparative criminology international homicide trends rule of law transitional democracy homicide

Introduction

Following the third wave of democratization in the last quarter of the 20th century (Huntington, 1993), social science researchers have devoted increasing attention to the association between democracy and a wide range of social and political phenomena. Often evidence suggests democracy decreases undesirable social outcomes (e.g., infant mortality; see Gerring, Thacker, & Alfaro, 2012) or increases prosocial outcomes (e.g., economic development; see Doucouliagos & Ulubaşoğlu, 2008). Yet, research on the relationship between democratization and homicide generally displays a weak and inconsistent association, even though extant sociological theory often predicts a negative association between democratization within a country and levels of crime and violence (see LaFree & Tseloni, 2006).

One possible explanation of the weak relationship between democracy and lethal violence is that measures used in prior research fail to capture the core components of democracy that influence violence. Indeed, Nivette (2011) recently noted that although “democracy indices proved rather weak influences on homicide overall, this may be attributable to the lack of empirical and theoretical development of specific mechanisms involved” (p. 123). Likewise, Stamatel (2008) contends that theorizing about specific components of democracy and political order that may influence criminal activity is needed to advance research. Following these recommendations, the current study focuses on a key subcomponent of democracy by examining on the impact of the “rule of law” on violent behavior.

Rule of law is considered a major subcomponent in operationalization of democracy in which “all actions of citizens, the state, and government are equally subject to legal scrutiny, independent of position, status, and power” (Karstedt & LaFree, 2006, pp. 11-12). In the current study, we contend that rule of law is a key construct of democracy that can impact violence of a society, as it serves as a state’s ability to embody and enforce law in a manner citizens perceive as just, and therefore should reduce violence by diminishing antisocial views and desires of revenge among citizens, as well as through the transfer of power to resolve conflicts from citizens to the state (Black, 1983; Herrman, Thöni, & Gächter, 2008).

To assess this relationship, we present two measures of rule of law—(a) a measurement of an independent judiciary from the Political Constraint Index (POLCON) data set (see Henisz, 2002); (b) a measurement of “Law and Order Tradition” from the International Country Risk Guide (ICRG)—and use country-level homicide rates from the World Health Organization (WHO) as a measure of homicide. Although our main focus is on explaining the association between rule of law and homicide rates, we aim to create a more complete test of rule of law, as well as competing theoretical perspectives by including two other measures of violence as alternative outcomes: country-level suicide rates and country-level fatalities from terrorist attacks. In the following sections, we outline competing theoretical perspectives regarding the relationship of democratization and lethal violence. We then describe our sample, measures, and empirical strategy and conclude with a discussion of the implication of our findings for future research.

Conceptual Framework

Prior cross-national criminological research typically draws upon two theoretical perspectives in explaining the relationship between democracy and violence: (a) Durkheim’s modernization perspective and (b) Elias’s civilizing process. In this section, we briefly describe these theoretical positions and formulate corresponding hypotheses. Next, we expand upon this literature by disaggregating measures of democracy and offering “rule of law” as an alternative theoretical avenue that can explain differences in lethal violence cross-nationally.

Modernization Theory

Modernization theory, as developed within and after the work of Durkheim suggests that crime and violence result when values and norms breakdown and the established system is disrupted by the emergence of a new social structure. The modernization perspective postulates a curvilinear relationship between democracy and violent crime. During a period of social transformation, increased anomie derived from readjustment of social institutions will initially produce greater levels of violence. However, after the new social order is firmly established, violent behavior should decrease as the newly reconstructed norms are stabilized and the anomie of social transition fades (Durkheim, 1892/1947, 1897/1950). Although Durkheim did not explicitly address democratic transitions, the perspective is often applied to the study of democratization (LaFree & Tseloni, 2006) as readjustments of the political structure often generate instability and periods of anomie (Zhao & Cao, 2010).

Much empirical work has relied on the use of cross-sectional data to test the modernization process in a cross-national perspective. Evidence from these initial studies provide weak support for the theory, leading to questions about the base predictions of modernization theory, as much work finds either no relationship or a negative relationship between measures of economic or political development and lethal violence (Bennett, 1991; Messner & Rosenfeld, 1997). Although most of these studies do not grapple with democracy as a central concept, the limited numbers of states in these samples tend to be democracies of some form, thus providing indirect evidence about the relationship between democracy and crime.

However, later research identified key issues with early test of modernization theory. Notably, the cross-sectional nature of prior work does not observe the process overtime, which is central to the theory as such data snapshots do not tell us where on a potential curve the society lies. LaFree and Drass (2002) have noted these limitations of the cross-sectional designs and argued that modernization is an inherently longitudinal concept. In their longitudinal study of 34 countries from 1956 to 1998, LaFree and Drass (2002) found support for modernization theory’s prediction that industrializing nations are more likely than industrialized nations to experience rapid increases in violent crime as a result of a breakdown in the social order. LaFree and Tseloni (2006) found mixed support for modernization theory’s predictions regarding transitional democracies and cross-national homicide rates in their longitudinal study of 44 countries, finding “that countries moving from autocratic to transitional democracies experienced a significant increase in homicide rates.” Yet, contrary to expectations, after completing a full transition, democracy itself did not reduce homicide, as “homicide rates of full democracies were not on average significantly different from countries with autocratic governments” (p. 43).

In sum, modernization theory predicts a curvilinear relationship between democracy and violence. Specifically, violence should increase during the time when a society’s political structure is in transition, and then as the process stabilizes violence should plateau and then decrease as the new social order is consolidated. Accordingly, autocracies will have lower rates of violence than transitional states and once a country fully transitions to democracy, levels of violence should be lower in democracies than both autocratic and transitional societies. Furthermore, while most cross-national criminological research has focused on the effects of modernization on homicide (LaFree & Tseloni, 2006), the current study aims to present a fuller test of the theory by testing the propositions of modernization theory in regard to multiple forms of violence, including terrorist fatalities and suicide rates. Specifically, we suggest that a similar curvilinear relationship should operate in relation to terrorist violence and modernization within a society. Finally, following Durkheim’s (1897/1950) work on the impact of social forces on suicide, we suggest that suicide rates should follow the same curvilinear pattern as well. Therefore, we propose the following hypotheses:

Hypothesis 1: Countries in transition to democracy will experience higher levels of homicide, suicide, and terrorism fatalities than autocracies and democracies.

Hypothesis 2: Countries that have fully transitioned to a democracy will have significantly lower levels of homicide, suicide, and terrorism fatalities in comparison to autocracies.

Civilizing Processes

The work of Norbert Elias provides an oppositional view of the potential relationship between democratization and violent crime (see LaFree & Drass, 2002; LaFree & Tseloni, 2006). The civilizing process anticipates that as societies modernize, violence will decline through two main processes. First, as states develop, simultaneously more and more power over people’s lives is absorbed by the state via its monopolization of law and violence. Accordingly, while in the past interpersonal disputes would be decided privately, the civilizing process transforms the decision of a person to use violence as a means to take revenge into a matter of public interest that is enforced by the laws of the state. Second, the civilizing process influences cultural attitudes held by individuals, which impacts decision making processes via an insertion of anti-violence norms into a cultural framework. Therefore, as the state develops, citizens also undergo psychological changes in response to the burgeoning social system that requires complex interpersonal interactions, self-restraint over aggressive impulses, and higher levels of self-control (Whitt, 2010). This multi-loci restraint produces less violence; especially once attitudes are internalized by citizens and not just embodied in external social structures (Elias, 2000).

In support of Elias’s work, those using historical data have noted the long-term decline in rates of violence in Western countries from approximately the 13th century into the 20th century (Eisner, 2001; Gurr, 1981). For instance, Eisner (2001) finds support for the civilizing process noting that emerging institutional arrangements, such as expansion of schools, rise of religious reform movements, and the organization of work in manufacturing were instrumental in the substantial declines in homicide rates from the Middle Ages into the 20th century in Europe. However, other research focusing on homicide trends after World War II contradicts Elias’s predictions. LaFree and Drass (2002) find substantial increases in homicide rates among 34 industrialized countries from 1958 to 1998. Similarly, Eisner (2001) finds increases in homicide rates among European democracies after 1950. Contrary to the argument from a civilization perspective that violent crime will decline as autocratic regimes transition to democracies, LaFree and Tseloni (2006) did not find homicide rates were significantly lower in full democracies relative to autocracies in a sample of 44 countries from 1950 to 2000.

Other work following Elias’s claims has sought to more closely assess the effect of the civilizing process on violence in democratic societies. Karstedt’s (2006) exploratory work on democratic regimes and violence finds that values typically associated with democracy are negatively associated with homicide rates, concluding “the more countries adopt individualistic and egalitarian values, the lower their levels of crime; though egalitarian values do not achieve this without egalitarian structures” (see also Karstedt, 2015). Karstedt’s caveat is an important one as theories within the Anomie tradition highlight the criminogenic effect of individualistic values but moderate this through egalitarian structures (Messner & Rosenfeld, 1997).

Although most criminological research to date using the civilizing process has focused on homicide (Eisner, 2001; LaFree & Drass, 2002; LaFree & Tseloni, 2006), the civilizing process does not anticipate that all forms of violence will dissipate. Rather, as societies modernize and promote virtues such as self-control, individualism, and self-responsibility, there is simultaneously an internalization of blame and aggressive impulses (Whitt, 2010). Indeed, van Krieken (1989) suggests that the civilizing process is often linked to a transition from a culture based on shame to one based on guilt. This is an important distinction as societies that are more sensitive to shame and disrespect of one’s honor are more likely to resolve their disputes through interpersonal violence (Nisbett & Cohen, 1996), yet societies more prone to guilt are often more likely to internalize such feelings and turn violence inward on one’s self, suggesting that the civilizing process should result in a higher rate of suicide (Whitt, 2010).

To conclude, the civilizing process predicts a negative relationship between democracy and homicide rates, driven by attitudinal changes but strengthened when coupled with democratic social intuitions and structures. Similarly, as societies civilize, it is likely that other forms of interpersonal violence, such as terrorism should decrease as well. Yet, as attributions toward violence turn inward, the civilizing process is expected to be associated with an increase in suicide. Based on this discussion, we derive the following hypothesis:

Hypothesis 3: Autocratic countries will have the highest rates of homicide and terrorism fatalities, followed by transitional states, with fully democratic countries having the lowest rates.

Hypothesis 4: Fully democratic countries will have the highest rates of suicide, followed by transitional states, with autocratic countries having the lowest rates.

Rule of Law

Although the previously discussed theories illustrate how a general change from autocratic to democratic institutions might influence lethal violence, they do not discuss the influence of specific institutions within these frameworks. Rule of law is one particular social institution associated with democracy that can influence violent crime. At the broadest level, rule of law is a system of political rules and rights “applied impartially to all citizens” (Weingast, 1997, p. 1). For the law to be impartial, the state must be sufficiently strong to punish transgressions, but also checked so that no one portion of society is above it. An independent judiciary is one observable institution that can assure the state can achieve these conditions. As Tyler (2006) argues people are more apt to comply with laws that are perceived as just (see also LaFree, 1998). Accordingly, citizens evaluate the fairness of these laws by analyzing the procedures police and the courts use to exercise authority (Tyler & Huo, 2002). When the judiciary is politically independent, it should be more likely to be viewed as having just procedures. Although more stable and democratic societies should have a strong institutional and normative tradition emphasizing rule of law, there is variance across democracies, as well as theoretical reasons that rule of law, on its own, should have a strongly negative relationship with levels of violence.

First, a society with an independent judiciary should have a populace that perceives a greater amount of justice in the legal system. Legitimacy, of any institution, should act as an internal control (see LaFree, 1998). Moreover, the idea that functioning courts grounded in the concept of rule of law are more likely to punish individuals in an appropriate manner, and, thus, reduce the attractiveness of a given criminal opportunity is a long-standing concept. For example, in summarizing the reasons for the effectiveness of the HOPE Probation Program, Judge Steven Alm specifically highlights fairness in punishment as a key element in maximizing deterrence, stating,

When the system isn’t consistent and predictable, when people are punished randomly, they think, My probation officer doesn’t like me, or, Someone’s prejudiced against me, rather than seeing that everyone who breaks a rule is treated equally, in precisely the same way. (Rosen, 2010)

Second, recent extensions of social disorganization theory including collective efficacy and research on legal cynicism demonstrate mechanisms whereby stronger rule of law should curtail violent behavior. Briefly, collective efficacy suggests that reductions in violence are attributable to informal social control and social cohesion among residents of a given community (Sampson, Raudenbush, & Earls, 1997). Although traditional social disorganization theory contends that strong community ties are instrumental to facilitating community organization, collective efficacy underscores a key distinction by noting that social ties must be activated to achieve collective goals such as the control of crime. Indeed, a long line of research on collective efficacy suggests that the capacity of a locale to come together and activate mechanisms of social control, such as collectively bringing to action the criminal justice system, is instrumental in controlling crime and other problem behaviors (Sampson et al., 1997). Moreover, recent research links sociolegal institutions and collective efficacy suggesting that respect and strong ties among law enforcement organizations and the community may be influential in promoting collective efficacy. For instance, Sampson (2002) notes that social institutions that are viewed as legitimate and supported by a strong government are key in generating collective efficacy. Kochel (2012) suggests that by providing quality services and fostering strong relationships with citizens, police can promote greater collective efficacy within communities.

Strengthening rule of law should also reduce legal cynicism among citizens by providing a legitimate outlet to settle grievances rather than seeking revenge. Black (1983) notes that most homicides are not predatory actions but rather a retaliatory response to actions that the perpetrator themselves view as offensive or deviant. Interpersonal violence then is often the end result of grievances that occur in situations where the law is unavailable or incapable of resolving disputes, leading individuals to take law into their own hands and use violence as a method of social control (Black, 1983). In support of this perspective, a burgeoning literature expanding social disorganization models suggest legal cynicism is a core feature that mediates the influence of structural features of neighborhoods on crime and violence (see Kirk & Papachristos, 2011). Specifically, areas with high legal cynicism are prone to greater levels of interpersonal violence as views of the law as illegitimate, unresponsive, and ill-equipped to ensure public safety results in a loosening of the social control aspect of the law. Furthermore, if individuals do not see the criminal justice system as a legitimate outlet for resolving disputes, they may seek alternative measures of dispute resolution. For instance, Kirk and Papachristos (2011) find cynicism toward legal institutions increases violence at the neighborhood level, citing that it “constrains choice if individuals presume that the law is unavailable or unresponsive to their needs, thus pushing individuals to engage in their own brand of social control” (p. 1228).

Empirically, measures of rule of law have shown support in reducing the capacity for violent behavior across a variety of countries. Herrman et al. (2008) finds that rule of law predicts the degree to which citizens favorably view the use of antisocial revenge. In explaining the strong association between rule of law in a society and the antisocial views of citizens, Herrman et al. (2008) explain, “If the rule of law is strong, people trust the law enforcement institutions, which are perceived as being effective, fair, impartial, and bound by the law. Revenge is shunned. If the rule of law is weak, the opposite holds” (p. 1365). At the cross-national level, Karstedt (2015) finds a strong negative bivariate relationship between rule of law and homicide rates in a cross-sectional study of 98 countries. Similarly, Nivette’s (2016) examination of conflict resolution in Latin America suggests that subjective indicators of institutional illegitimacy are robust predictors of support of violent vigilantism.

Moreover, the effects of rule of law in combating violence may be widespread, curtailing not only homicide but terrorist violence as well. For instance, Choi (2010) suggests that rule of law protects the rights of both citizens, as well as foreigners by providing a means to resolve disputes in nonviolent ways. Thus, having strong rule of law is a means in which a society can create an environment where it is unnecessary for individuals to rely on terrorist violence to resolve grievances. In support of this proposition, Choi (2010) finds that maintaining strong rule of law reduces the likelihood of terrorist events, and rule of law within democratic institutions is a key defense against terrorism.

The above discussion highlights that rule of law is a core social institution within democratic societies that impacts rates of violence. Although rule of law may influence violent behavior through a variety of mechanisms, the anticipated effect remains the same. Rule of law is expected to independently reduce homicide rates even when combined with other improvements in democracy. Moreover, as rule of law enhances and formal institutions gain legitimacy and autonomy, there may be a spillover effect, where the standards of behavior drawn from sociolegal institutions are applied to other social situations, irrespective of their applicability to this new context (see Mullins & Young, 2012). Specifically, as the state encompasses more control over interpersonal violence among citizens and homicide declines, this pacification effect may spillover into other areas of interpersonal violence such as terrorism (Choi, 2010) and intrapersonal violence such as suicide. This discussion leads to our final hypotheses:

Hypothesis 5: Countries with stronger rule of law will have lower levels of homicide, suicide, and terrorism fatalities than countries with weaker rule of law

Research Design

Dependent Variables

Homicide rate

The primary dependent variable of interest is the country-level homicide victimization rates per 100,000 residents obtained from the WHO for years 1960 to 2004. Homicide victimizations are defined in accordance with the International Classification of Diseases (ICD) as “the death, or the circumstances of the accident or violence which produced the fatal injury.” Although the WHO homicide data faces limitations such as a lack of data on underdeveloped countries, especially in the regions of Africa and Asia, WHO homicide data is often regarded by researchers as the most valid and reliable source of cross-national homicide data (LaFree & Tseloni, 2006; Messner & Rosenfeld, 1997). Table 1 presents the descriptive statistics for each estimation sample used to model homicide rates. Across the three samples, the mean homicide rate is consistent, ranging from 4.88 to 4.34 homicides per 100,000 residents cross-nationally.¹

Table 1.

Descriptive Statistics for Independent and Dependent Variables (Homicide Rates).

Variables	Homicide rates						Suicide rates
	Model 1		Models 2		Models 3		Model 4		Model 5		Model 6
	M	SD	M	SD	M	SD	M	SD	M	SD	M	SD
Homicide rate	4.88	8.16	4.82	8.20	4.34	8.28	—	—	—	—	—	—
Suicide rate	—	—	—	—	—	—	12.10	8.00	12.00	7.97	11.67	7.70
Transitional	0.26	0.44	0.25	0.43	0.29	0.45	0.29	0.45	0.28	0.45	0.35	0.48
Full democracy	0.46	0.50	0.47	0.50	0.56	0.50	0.49	0.50	0.50	0.50	0.51	0.50
Independent judiciary	—	—	0.59	0.49	—	—	—	—	0.64	0.48	—	—
Total population	2.54	4.25	2.50	4.21	2.99	4.93	2.77	4.51	2.72	4.46	3.14	5.01
Young male population	12.17	1.55	12.19	1.56	11.93	1.73	12.18	1.61	12.19	1.62	12.14	1.76
Infant mortality	28.79	28.73	28.97	28.98	12.88	11.21	22.70	21.15	22.77	21.34	13.98	12.05
Law and order	—	—	—	—	4.82	1.27	—	—	—	—	4.62	1.38
Sample size	1,922		1,880		736		1,794		1,759		833
Number of countries	81		80		63		73		71		65
Years	1960-2004		1960-2004		1980-2004		1960-2004		1960-2004		1980-2004

Suicide rate

The suicide rate is the number of suicides in a country-year per 100,000 residents obtained from the WHO. Suicides are classified by the WHO using ICD and are defined as an act of deliberately killing oneself. During the time period between 1960 and 2004, the ICD underwent three revisions (starting with ICD-8 and ending with ICD-10). Suicide is coded in line with the following ICD codes: E950-E959 in ICD-8 and ICD-9, and X60-X84 in ICD-10 (see Liu, 2009). Table 1 demonstrates that the suicide rate is fairly consistent across our three estimation samples ranging from 12.10 to 11.67 per 100,000 residents.

Terrorism fatalities

A count of fatalities resulting from terrorism events in a given country-year are collected from the Global Terrorism Database (GTD), an open source database on worldwide terrorist attacks and one of the most comprehensive sources of terrorism incident databases available (see LaFree & Dugan, 2007). The current study uses data on terrorism fatalities from 1970 to 2004. Table 2 shows that terrorism fatalities are relatively consistent across each estimation sample ranging from 32.14 to 42.81 terrorism fatalities on average.²

Table 2.

Descriptive Statistics for Independent and Dependent Variables (Terrorism Fatalities).

Variables	Model 7		Model 8		Model 9
Variables	M	SD	M	SD	M	SD
Terrorism fatalities	32.14	163.81	32.65	165.47	42.81	188.28
Transitional	0.33	0.47	0.32	0.47	0.41	0.49
Full democracy	0.39	0.49	0.40	0.49	0.42	0.49
Independent judiciary	—	—	0.54	0.50	—	—
Total population	3.41	10.43	3.37	10.50	4.32	13.55
Young male population	12.62	1.71	12.64	1.72	12.55	1.90
Infant mortality	26.41	23.69	26.55	23.89	18.71	15.68
Law and order	—	—	—	—	4.29	1.49
Sample size	1,948		1,908		1,077
Number of countries	90		90		81
Years	1970-2004		1970-2004		1980-2004

Independent Variables

Democracy scale

To test the Hypotheses 1 to 4, we created the measures of democracy from the Polity database, similar to those outlined by LaFree and Tseloni (2006). The Polity database measures are based on an 11-point scale measuring annual rates of democracy and autocracy for all countries dating back to 1800. Although other scales are available, such as Freedom House, prior research notes that these measures are highly correlated with the Polity data and that the Polity database is beneficial because it provides a larger and more diverse set of countries for analysis (see LaFree & Tseloni, 2006, p. 46).³

Measurement of democracy is coded via a system of points allocated based on four characteristics of a countries national government: (a) the competiveness of political participation, (b) the openness of executive recruitment, (c) the competiveness of executive recruitment, and (d) constraints on the chief executive (LaFree & Tseloni, 2006; Marshall & Jaggers, 2004). This produces in a 21-point scale (−10 to +10). Following the procedure outlined by LaFree and Tseloni (2006), we create a series of variables that collapse the democracy scale into three categories: full democracy (10 points on the Polity scale), transitional (1-9 points), and autocracy (−10 to 0 points).⁴ Our analysis includes dichotomous indicators for full democracies and transitional states, and uses autocracies as the reference category.

Independent judiciary

A measure of an independent judiciary is collected from the POLCON data set (Henisz, 2000, 2002). The POLCON data set defines an independent judiciary via the joint existence of a Polity score on the executive constraint dimension and the ICRG score on Law and Order. Specifically, Henisz (2000, 2002) applies the average (1985-1990) value of Law and Order Tradition from the ICRG to compute a dichotomous measure variable if the Law and Order Tradition is defined as having “a relatively strong law and order tradition.” However, as law and order may be the result of either a well-functioning legal system or a repressive political system, only those country-years in which a Polity score of political constraints was coded as having at least “slight to moderate limitations to executive authority” were coded as an independent judiciary. Using this coding scheme, judicial independence is coded as a binary measure where 1 equals a country that has a judiciary that is politically independent and 0 is a politically dependent judiciary. This indicator is available from 1960 through 2004 and has been used widely in prior research (see Henisz, 2000; Quinn, 2004).

Law and Order Tradition

The measure of “Law and Order Tradition” comes from the ICRG. According to the ICRG’s codebook, this variable “reflects the degree to which the citizens of a country are willing to accept the established institutions to make and implement laws and adjudicate disputes.” Higher scores indicate “sound political institutions, a strong court system, and provisions for an orderly succession of power.” Lower scores indicate “a tradition of depending on physical force or illegal means to settle claims.”⁵ This measure is based on subjective assessments by subject area experts and ranges from 0 to 6. This measure is available from 1980 to 2004, and in contrast to the independent judiciary measure, there are far fewer countries that ICRG codes.

Control Variables

To control for other cross-national predictors of crime that could also explain rule of law, we control for indicators of population demographics and economic development.⁶

Infant mortality

The infant mortality rate is used as a proxy for poverty (see Pridemore, 2008). Data on infant mortality come from World Bank’s World Development Indicators (WDI) database and are defined as the rate of infant deaths before reaching 1 year of age per 1,000 live births in a given country-year. Prior research often finds infant mortality rate to be an important and significant predictor of homicide rates (Nivette, 2011; Pridemore, 2008).⁷

Total population

Population (in ten thousands) is measured as the total population in a country in a year and is collected from the World Bank’s WDI database. Although population density measures may be better at predicting homicides, this variable is not available for our temporal and spatial domain. Therefore, we include the total population as a means to control for components of the population structure (see Young & Dugan, 2011).

Young males

The proportion of the population, that is male and aged 15 to 29 years, is collected from the World Bank’s WDI database. The demographics of the young male population are included based on the assumption in prior cross-national criminological research that young people (especially young males) commit a disproportionate amount of violent crime (see Bennett, 1991; LaFree & Tseloni, 2006).

Estimation

We estimate a fixed-effects (FE) linear regression model with AR1 Prais–Winsten transformation for models estimating homicide rates and suicide rates. FE models are preferable to alternative estimators such as Random Effects (RE) as the FE model focuses on within-unit change and provides less restrictive assumptions relative to alternative models. Specifically, RE models assume that time-varying independent variables (i.e., rule of law) are uncorrelated with unmeasured country-level time invariant factors, whereas the FE model does not make this restrictive assumption. Rather the FE model controls for the influence of time invariant predictors, whereas RE do not. Moreover, as panel heterogeneity and contemporaneous correlation may bias standard errors, we use Beck and Katz (1995) panel-corrected standard errors, which have been shown using Monte Carlo simulation to be robust to these problems.⁸

For models estimating terrorism fatalities, our outcome is a count variable. Most quantitative studies of terrorism use either Poisson or negative binomial regression (NBRM). However, Poisson models are based on the restrictive assumption that the mean is the same as the variance. Following previous research examining democracy and transnational terrorism and we use NBRM (see Young & Dugan, 2011) and cluster standard errors by country to account for within panel correlations that would artificially deflate standard errors. Finally, Wald tests are performed for all models to examine whether there is a statistically significant difference between the full democracy and transitional democracy coefficients.

Results

Table 3 presents the results from the FE regression models testing the hypotheses developed above. The first hypothesis is derived from modernization theory and suggests that transitional states will have higher levels of homicide, suicide, and terrorism fatalities than both autocracies and democracies. In support of this hypothesis, Models 1 through 3 in Table 3 demonstrate that transitional states have higher rates of homicide than autocracies while controlling for other factors. However, this hypothesis is only partially supported, as Wald tests reveal, there is no significant difference between transitional states and full democracies. Next, Models 4 through 6 in Table 3 demonstrate there is no significant difference in the suicide rate between transitional states and autocracies. Again a Wald test reveals no significant difference between transitional states and full democracies as well. Finally, Models 7 through 9 indicate that transitional states consistently have higher levels of terrorism fatalities than autocracies and a Wald test reveals no significant difference between transitional states and full democracies. Thus, hypothesis 1 receives only partial support.

Table 3.

Fixed-Effects Regression Results of Lethal Violence Outcomes.

Variables	Homicide rate			Suicide rate			Terrorism fatalities
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
Transitional	3.07***	3.62***	3.15***	0.14	0.25	−0.59	2.11***	2.57***	2.34***
Transitional	(0.54)	(0.56)	(0.60)	(0.27)	(0.28)	(0.32)	(0.40)	(0.39)	(0.42)
Full democracy	2.75*	3.62**	4.10***	0.32	0.45	−0.17	1.69*	2.70***	2.97***
Full democracy	(1.21)	(1.22)	(1.07)	(0.52)	(0.52)	(0.57)	(0.69)	(0.68)	(0.49)
Independent judiciary	—	−2.60***	—	—	−0.61**	—	—	−0.53	—
Independent judiciary	—	(0.64)	—	—	(0.24)	—	—	(0.43)	—
Total population (10K)	0.35	0.58	1.80**	1.11***	0.68*	1.94***	0.08*	0.09*	0.12**
Total population (10K)	0.46	(0.49)	(0.70)	(0.30)	(0.31)	(0.53)	(0.04)	(0.04)	(0.05)
Young male population	0.42**	0.44**	0.07	0.29**	0.28**	0.10	0.46**	0.54**	0.52**
Young male population	(0.15)	(0.15)	(0.21)	(0.09)	(0.08)	(0.11)	(0.16)	(0.16)	(0.15)
Infant mortality	−0.04	−0.04	0.63***	0.06*	0.05*	0.36***	0.03*	0.03**	−0.01
Infant mortality	(0.03)	(0.03)	(0.13)	(0.03)	(0.03)	(0.08)	(0.01)	(0.01)	(0.01)
Law and order (ICRG)	—	—	−0.47**	—	—	0.06	—	—	−0.79***
Law and order (ICRG)	—	—	(0.19)	—	—	(0.09)	—	—	(0.12)
Constant	−2.17***	−1.87***	−9.47***	4.36***	6.19***	−0.44	−5.38**	−6.86**	3.26
Constant	(0.45)	(0.46)	(0.49)	(0.15)	(0.15)	(0.24)	(2.09)	(2.06)	(2.09)
Wald test coefficient	0.09	0.00	1.18	0.17	0.22	0.84	0.58	0.06	2.92
Observations	1,922	1,880	736	1,794	1,759	833	1,948	1,908	1,077
Within R²	.027	.037	.124	.023	.019	.035	.027	.030	.049
Number of countries	81	80	63	73	71	65	90	90	81
Years estimated	1960-2004	1960-2004	1980-2004	1960-2004	1960-2004	1980-2004	1970-2004	1970-2004	1980-2004

Note. Standard errors in parentheses. ICRG = International Country Risk Guide.

p < .05, two-tailed. **p < .01, two-tailed. ***p < .001, two-tailed.

Our second hypothesis draws upon modernization theory and proposes that full democracies would have significantly lower levels of homicide, suicide, and terrorism fatalities in comparison to autocracies. In contrast to this hypothesis, Models 1 through 3 in Table 3 suggest that democracies consistently have higher rates of homicide than autocracies. Regarding suicide, Models 4 through 6 show no significant difference between full democracies and autocracies. Finally, terrorism fatalities are consistently higher in democracies than in autocracies in Models 7 through 9. Therefore, the results fail to provide support for the second hypothesis.

Next we examine the third hypothesis testing the civilization perspective, which proposes both transitional states and fully democratic countries will have lower levels of homicide and terrorism fatalities compared to autocratic countries. In contrast to the predictions of the civilization perspective, the results in models 1 through 3 suggest that both full democracies and transitional states have higher rates of homicide than autocracies. Moreover, Models 7 through 9 demonstrate that both full democracies and transitional states consistently have higher levels of terrorism fatalities than autocratic states. However, the results from the Wald test show no significant difference between transitional states and full democracies.

Our fourth hypothesis examines the predictions of the civilizing process that suggests that fully democratic states would have the highest rates of suicide, followed by transitional states, with autocratic countries having the lowest rates. The findings from Models 4 through 6 show no significant difference in suicide rates between full democracies and autocratic states, as well as transitional societies and autocracies. In addition, a Wald test comparing democracies and transitional states finds no significant difference in suicide rates. Thus, we fail to find support for Hypotheses 3 and 4.

The fifth hypothesis turns the focus to our rule of law measures and proposes that countries with stronger rule of law will have lower levels of homicide, suicide, and terrorism fatalities. Model 2 adds a separate measure for independent judiciary, which is negatively associated with homicide rates. The presence of an independent judiciary, on average, is expected to decrease the homicide rate by 2.3 murders per hundred thousand. As the mean rate in our sample is 4.82, in substantive terms, the presence of an independent judiciary is expected to decrease the mean homicide rate by about 48%, while holding other variables including democracy constant. Model 3 uses the ICRG measure of Law and Order instead of judicial independence. The results are similar with Law and Order having a negative and statistically significant effect on the homicide rate with no large differences between either models providing support for Hypothesis 5.

Model 5 demonstrates that an independent judiciary has a statistically significant and negative association to suicide rates. Model 6 uses the ICRG measure of Law and Order and finds that this rule of law measure is not related to suicide rates. Overall, these results suggest that while an independent judiciary carries a strong negative association to suicide rates, the ICRG measure of Law and Order is unrelated to suicide in both models, providing mixed support for hypothesis 5.

Model 8 reports the results of an independent judiciary on terrorism fatalities. The findings suggest that an independent judiciary has a negative but non-statistically significant relationship to homicide rates, while controlling for our primary measures of transitional states and democracy. Model 9 indicates the ICRG measure of Law and Order has a negative and statistically significant association with terrorism fatalities. Thus, the findings provide contradictory evidence, suggesting that while independent judiciary is unrelated to terrorism fatalities, the ICRG measure of Law and Order reduces terrorism-related fatalities in support of Hypothesis 5.

Discussion

In this article, we have examined multiple variant predictions of the relationship between levels of lethal violence, a society’s level of democracy, and sociolegal values and democratic institutions representative of the rule of law. In our models testing the predictions of modernization theory’s on homicide rates, we verify previous results and find partial support for our first hypothesis finding a positive and linear relationship between homicide rates and societies in the process of democratization (see LaFree & Tseloni, 2006). However, in bringing in suicide rates and terrorism fatalities to offer a fuller test of modernization theories predictions, the effects of transitional societies and democracy became inconsistent. In contrast to our second hypothesis, the effect of democracy itself proved inconsistent with the magnitude and statistical significance of democracy varying across models for all three outcomes measuring lethal violence.

A civilization process, however, also predicts a negative relationship between these variables in relation to homicide rates and terrorism fatalities. As we find transitional regimes can have higher incidences of homicide and terrorism, and full democracies have a positive and significant association with homicide and terrorism, these findings are not supportive of our third hypothesis. Similarly, the civilizing process also predicts that as societies increase in self-restraint and internalize norms against the use of interpersonal violence, an emergence in individualistic norms will also lead to increased rates of suicide. However, the results of the current study find that democratization has no association with suicide rates, failing to support our fourth hypothesis, and lending further evidence against the propositions of the civilizing process.

In support of our fifth hypothesis, our disaggregation of democracy highlights the specific power of the judiciary in reducing homicide. Other factors held constant, societies that possess an independent judiciary have lower homicide rates. These values evolve at the same time as those that discourage violence as a resolution to interpersonal disputes and other attitudes, which encourage self-restraint. These sets of beliefs serve to reduce criminal motivations through increasingly defining such actions and actors as problematic and peaceable interactions as more desirable. Thus, from this standpoint, there is nothing essential in democracy itself that reduces crime. Rather democracy arises and stabilized at the same time as other attitude shifts, which we now think of as modernity. In addition, important here is the fact that if courts are perceived as just and fair by the public (more likely if they are independent), the perceived legitimacy of the courts and law is heightened. This increases the probability of rule internalization by the citizenry (or at least rule abidance).

Following LaFree’s (1998) Institutional Legitimacy theory, an independent judiciary is part and parcel of building a network of institutions that citizens perceive of as just and rightful in the state’s exercise of authority and power (see also Tyler, 2006). As rule of law increases, conformity to the law is more likely to be internalized, reducing lethal violence not as a function of fear of punishment but as a reflection of more coherent social organization and shifts in attitudes about law and the courts. This operates separately from deterrence as fear of punishment and from normative transitions that deemphasize using violence as an acceptable form of conflict resolution.

We do use two separate measures of rule of law here, and they do explain different amounts of variation in homicide at the state level. Although both exhibit strong and statistically significant relationships with homicide, the Henisz (2002) independent judiciary measure produced stronger effects in the regression models of homicide rates. This highlights the unique power of the courts in controlling crime, yet without specific measures of attitudes toward institutions and the law, it is not possible to say exactly what this effect may be. It could be a deterrent effect, a legitimacy effect (LaFree, 1998), a reflection of procedural justice (Tyler, 2006; Tyler & Huo, 2002), or something else entirely.

Yet our measure of independent judiciary was found to be unrelated to terrorism fatalities. This leads to the question of how far an independent judiciary goes in curtailing violence. Although previous research finds that rule of law is effective in reducing terrorism (Choi, 2010), the current study finds mixed support for this proposition using different measurements of the rule of law. Although the measure of independent judiciary, which had strong effects in reducing homicide, is found to have no statistically significant association with terrorism fatalities, the ICRG measure of Law and Order Tradition is found to significantly reduce deaths related to terrorism within societies. Overall, these results suggest that rule of law may play an important role in reducing terrorism fatalities; however, future research must continue to examine whether different measures of rule of law have a consistent effect on terrorism and what explains diverging results across these measures. Finally, an independent judiciary was found to be effective in reducing suicide rates as well, suggesting that the emergence of strong legal institutions may serve “as a very powerful agent of socialization” that can reduce multiple forms of violent behavior (Mullins & Young, 2012, p. 33). Again, however, the measures of rule of law proved somewhat inconsistent as the ICRG measure of Law and Order tradition had no effect on suicide rates.

Still, there are a number of limitations in the current study, which future research can expand upon. First, as with prior work, we cannot separate a change in real behavior from an artifact in the data collection process. There is good reason to believe that reporting/recording of lethal violence will improve during a democratic transition. Second, fatalities resulting from homicide, as well as suicide and terrorist attacks are potentially dependent upon a country’s level of development as well as trust and legitimacy of a country’s government. For instance, highly developed countries will have better equipped hospitals, well-trained doctors, and faster emergency response time, all of which should reduce the likelihood that a given act of violence results in a fatality relative to a less developed country. Similarly, as legitimacy in the government increases, citizens may be more likely to call the police or other emergency services after a violent event has taken place, rather than attempting to respond on their own to the medical needs of the injured. Accordingly, rule of law within a country may also be related to the probability that an act of violence results in death rather than an injury. Although we have used measures on homicide, suicide, and terrorism fatalities widely considered among the highest quality data available, our data do not allow us to empirically investigate this potential limitation in greater detail. Accordingly, we encourage future research to examine the association between development or institutional legitimacy within a country and the adequacy of medical responses to violent events.

Another limitation is in regard to missing data in the current study. Although the current study includes more countries and a broader spatial coverage than previous longitudinal research on democratization and crime (LaFree & Tseloni, 2006), in our sample, like most cross-national criminological research missing data substantially reduces the sample size. Although there is unfortunately limited theoretical basis for knowing exactly how non-reporting of information might influence the findings of this study, it is plausible that the inclusion of data from certain countries that were not available would have changed the results.

Finally, it may be suggested that an independent judiciary on its own does not lead to reductions in lethal violence, but rather both the formation of independent judiciary and reductions in lethal violence are consequences of improvements of democratic development. The current study sought to isolate the effect of an independent judiciary by using a variety of statistical approaches and controlling for multiple alternative measures democracy, as well as demographic and economic covariates. Although we remain confident in the potential for an independent judiciary and the “rule of law” more broadly to reduce violent behavior, we encourage future research to continue to investigate this question and explore innovative approaches to identify the effect of rule of law on lethal violence.

In sum, the current study aims to contribute to the literature by displaying the benefits of disaggregation of measures of democracy, as a means to further our understanding of the relationship between political structure and violence (Nivette, 2011; Stamatel, 2008). In doing so, the current study adds to the existing literature through specifying the importance of investigating measures of the “rule of law” as our study shows that on their own, our two measures of rule of law can reduce lethal violence, especially homicide. Although in some sense an obvious finding, it does emphasize the fact that the development of sociolegal values and institutions is central in the process of democratization and that court development specifically may be one area of particular importance.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Alexander Testa is a PhD candidate in the Department of Criminology and Criminal Justice at the University of Maryland, College Park. His research interests include the consequences of criminal justice contact across the life course, criminal justice decision-making, and cross-national causes and consequences of violence.

Joseph K. Young is an Associate professor and chair of the department of Justice, Law, and Criminology at American University. His research addresses the causes and consequences of public violence. His recent publications appear in Criminology and Public Policy, Terrorism and Political Violence and Journal of Politics.

Christopher Mullins, PhD (UMSL—2004) is a Professor of Criminology and Criminal Justice at Southern Illinois University. His research focuses on the structural and cultural aspects of violence.

References

Beck

Katz

J. N.

(1995). What to do (and not to do) with Time-Series Cross-Section Data. American Political Science Review, 89, 634-647.

Bennett

R. R.

(1991). Routine activities: A cross-national assessment of a criminological perspective. Social Forces, 70, 147-163.

Black

(1983). Crime as social control. American Sociological Review, 48, 34-45.

Bollen

K. A.

Paxton

(2000). Subjective measures of liberal democracy. Comparative Political Studies, 33, 58-86.

Choi

(2010). Fighting terrorism through the rule of law? Journal of Conflict Resolution, 54, 940-966.

Doucouliagos

Ulubaşoğlu

M. A.

(2008). Democracy and economic growth: A meta-analysis. American Journal of Political Science, 52, 61-83.

Durkheim

(1947). The division of labor in society ( Simpson

, Trans.). New York, NY: Free Press. (Original work published 1892)

Durkheim

(1950). Suicide: A study in sociology ( Simpson

, Trans). New York, NY: Free Press. (Original work published 1897)

Eisner

(2001). Modernization, self-control and lethal violence: The long-term dynamics of European homicide rates in theoretical perspective. British Journal of Criminology, 41, 618-638.

10.

Elias

(2000). The civilizing process: Sociogenetic and psychogenetic investigations (Rev. ed.). Oxford, UK: Wiley-Blackwell.

11.

Gerring

Thacker

S. C.

Alfaro

(2012). Democracy and human development. The Journal of Politics, 74, 1-17.

12.

Giannone

(2010). Political and ideological aspects in the measurement of democracy: The Freedom House case. Democratization, 17, 68-97.

13.

Gurr

T. R.

(1981). Historical trends in violent crime: A critical review of the evidence. Crime and Justice, 3, 295-353.

14.

Henisz

W. J.

(2000). The institutional environment for economic growth. Economics & Politics, 12, 1-31.

15.

Henisz

W. J.

(2002). The Political Constraint Index (POLCON) dataset. Retrieved from http://www.management.wharton.upenn.edu/henisz/POLCON/ContactInfo.html

16.

Herrman

Thöni

Gächter

(2008). Antisocial punishment across societies. Science, 319, 1362-1367.

17.

Huntington

S. P.

(1993). The third wave: Democratization in the late twentieth century. Norman: University of Oklahoma Press.

18.

Karstedt

(2006). Democracy, values and violence: Paradoxes, tensions, and comparative advantages to liberal inclusion. The ANNALS of the American Academy of Political and Social Science, 605, 50-81.

19.

Karstedt

(2015). Does democracy matter? Comparative perspectives on violence and democratic institutions. European Journal of Criminology, 12, 457-481.

20.

Karstedt

LaFree

(2006). Democracy, crime and justice. The ANNALS of the American Academy of Political and Social Science, 605, 6-23.

21.

Kirk

D. S.

Papachristos

A. V.

(2011). Cultural mechanisms and the persistence of neighborhood violence. American Journal of Sociology, 116, 1190-1233.

22.

Kochel

T. R.

(2012). Can police legitimacy promote collective efficacy? Justice Quarterly, 29, 384-419.

23.

LaFree

(1998). Losing legitimacy: Street crime and the decline of social institutions in America. Boulder, CO: Westview Press.

24.

LaFree

Drass

K. A.

(2002). Counting crime booms among nations: Evidence for homicide victimization rates, 1956 to 1998. Criminology, 40, 769-800.

25.

LaFree

Dugan

(2007). Introducing the global terrorism database. Terrorism and Political Violence, 19, 181-204.

26.

LaFree

Tseloni

(2006). Democracy and crime: A multilevel analysis of homicide trends in forty-four countries, 1950-2000. The ANNALS of the American Academy of Political and Social Science, 605, 25-49.

27.

Liu

(2009). Suicide rates in the world: 1950-2004. Suicide and Life-Threatening Behavior, 39, 204-213.

28.

Marshall

M. G.

Jaggers

(2004). Polity IV project: Political regime characteristics and transitions, 1800-2004. College Park: Center for International Development and Conflict Management, University of Maryland.

29.

Messner

S. F.

Rosenfeld

(1997). Political restraint of the market and levels of criminal homicide: A cross-national application of institutional anomie theory. Social Forces, 75, 1393-1416.

30.

Mullins

C. W.

Young

J. K.

(2012). Cultures of violence and acts of terror: Applying a legitimation–habituation model to terrorism. Crime & Delinquency, 58, 28-56.

31.

Nisbett

R. E.

Cohen

(1996). Culture of honor: The psychology of violence in the South. Boulder, CO: Westview Press.

32.

Nivette

A. E.

(2011). Cross-national predictors of crime: A meta-analysis. Homicide Studies, 15, 103-131.

33.

Nivette

A. E.

(2016). Institutional ineffectiveness, illegitimacy, and public support for vigilantism in Latin America. Criminology, 54, 142-175.

34.

Pridemore

W. A.

(2008). A methodological addition to the cross-national empirical literature on social structure and homicide: A first test of the poverty-homicide thesis. Criminology, 46, 133-154.

35.

Quinn

K. M.

(2004). Bayesian factor analysis for mixed ordinal and continuous responses. Political Analysis, 12, 338-353.

36.

Rosen

(2010, January 10). Prisoners of parole. The New York Times. Retrieved from http://www.nytimes.com/2010/01/10/magazine/10prisons-t.html?pagewanted=all&;_r=0

37.

Sampson

R. J.

(2002). Transcending tradition: New directions in community research, Chicago style. Criminology, 40, 213-230.

38.

Sampson

R. J.

Raudenbush

S. W.

Earls

(1997). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277, 918-924.

39.

Scoble

H. M.

Wiseberg

L. S.

(1981). Problems of comparative research on human rights. In Nanda

V. P.

Scarritt

J. R.

Shepherd

G. W.

(Eds.), Global human rights: Public policies, comparative measures, and NGO strategies (pp. 147-171). Boulder, CO: Westview Press.

40.

Stamatel

J. P.

(2008). Using mortality data to refine our understanding of homicide patterns in select postcommunist countries. Homicide Studies, 12, 117-135.

41.

Steiner

N. D.

(2014). Testing for a political bias in freedom house democracy scores: Are US friendly states judged to be more democratic? Journal of Comparative Policy Analysis, 18, 329-394.

42.

Tyler

T. R.

(2006). Why people obey the law. Princeton, NJ: Princeton University Press.

43.

Tyler

T. R.

Huo

(2002). Trust in the law: Encouraging public cooperation with the police and courts. New York, NY: Russell Sage Foundation.

44.

van Krieken

. (1989). Violence, self-discipline and modernity: Beyond the “civilizing process.” The Sociological Review, 37, 193-218.

45.

Weingast

B. R.

(1997). The political foundations of democracy and the rule of law. American Political Science Review, 91, 245-263.

46.

Whitt

H. P.

(2010). The civilizing process and its discontents: Suicide and crimes against persons in France, 1825-18301. American Journal of Sociology, 116, 130-186.

47.

Young

J. K.

Dugan

(2011). Veto players and terror. Journal of Peace Research, 48, 19-33.

48.

Zhao

Cao

(2010). Social change and anomie: A cross-national study. Social Forces, 88, 1209-1229.