The Ties That Bind and the Bonds That Break by maqboolshahin


									The Ties That Bind and the Bonds That Break: Social Capital and Attitudes Toward Political Violence
Terrence Chapman* Emory University

Paper Prepared for the Annual Meeting of the Southern Political Science Association New Orleans, LA, January 8-10, 2004

*I would like to thank Frank Boyd, Ben Chapman, Ilaria Ossella-Durbal, Ozlem Elgun, Eric Reinhardt, and Greg Shaw for helpful comments and suggestions. Special thanks go to Kathleen Montgomery for helping guide and shape this project. I am responsible for any errors or inaccuracies. Comments welcome.

1 Introduction What determines support for political violence? If violence is inherently costly and carries with it material and normative sanctioning, why should individuals ever believe it is justified, particularly if other alternative, less costly forms of action are available? Answers to these questions have tended to focus on rational and psychological incentives. Rationalist accounts, however, fail to fully specify the conditions under which the expected benefits of violence outweigh the costs, and psychological theories fail to explain the widespread variation in levels of violence, given the general ubiquity of frustration and suffering. This paper argues that social capital theory offers a fresh perspective on why individuals might support political violence, given the difficulties in organizing violence and the associated costs. Despite modernization, increased standards of living, and long established democratic practices, political violence is a pervasive problem for many developed countries. Countries as diverse as Great Britain, the United States, and Japan have each experienced occurrences of political violence in the last decade. For the developing and democratizing world, on the other hand, discovering the sources of political violence and discovering mechanisms for managing it are even more immediate problems. Understanding the dynamics of support for nationalist violence, for instance, has helped shed light on the horrors that accompanied the disintegration of the former Yugoslavia. Understanding support for political violence can further help explain the sources of evolving group conflict in the Afghanistan and post-war Iraq. Rationalist theories explain violence as a function of the expected costs and benefits of collective violence. Psychological theories, on the other hand, point to alienation and frustration as the sources of political violence. Rationalist scholars have generally viewed violence as an extension of politics by any other means. Individuals will engage in violent collective action if their expected gains outweigh the costs of acting. Psychological theories, such as relative deprivation, suggest individuals’ perceptions of their position, relative to expectations, leads to frustration and aggression. Empirical studies have not found

2 strong supporting evidence of relative deprivation, but socio-psychological approaches that emphasize the importance of individual and group identity remain of interest to many scholars. Scholars have increasingly been interested in how types of social capital and civic engagement contribute to better democracy and help subdue violence. Social capital, according to Robert Putnam’s definition (2000), refers to civic networks and related norms of trust and reciprocity. Civic networks that are inclusive of broad segments of society may foster norms of generalized reciprocity and trust, in turn nurturing peaceful forms of democracy and overall more efficient social interaction. Interethnic associations are also thought to play a role in dispelling rumors driving group violence. Conversely, networks that are insular and exclusive may reinforce existing prejudices and close off members to ideas and innovations from wider society, fostering in-group versus out-group enmity, stagnation of group development and, at worst, inter-group violence. This paper investigates how varieties of civic networks influence individuals’ attitudes towards political violence. Social capital theory has implications for rational actor theories of collective violence and social-psychological theories, as well as for broader expectations about what types of societies are most likely to experience political violence. If social capital fosters norms of trust and reciprocity, together with building networks and opportunities for the articulation of grievances, then it can lead to the promotion of low-cost alternatives to violence and to norms rejecting violence. Notably, if violence is just an extreme form of political expression, social networks may lead to more “unconventional participation,” such as signing petitions and engaging in protests, while diminishing support for outright violence.1 Social capital that cuts across traditional lines of group identification can also reduce insularity and ingroup out-group distrust. On the other hand, if social capital takes the form of insular, exclusive and isolated organizations, inter-group distrust is likely to be heightened while low-cost alternatives to violence will diminish.


Dalton, for instance, argues that so-called unconventional forms of political action are increasingly becoming mainstream (2002).

3 The paper proceeds as follows: the first section synthesizes theories of violence, political participation, and civic networks. The second section will discuss the research design and data. The third section will present results of cross-national analysis of the relationship between social capital and attitudes towards political violence, using data from the World Values Survey (Inglehart et al., 2000). The final section will provide concluding remarks and discuss implications for the role of networks in creating, maintaining, or frustrating peaceful democracy. Violence, Participation, and Civic Networks Theories of Political Violence Why might individuals choose to participate in political violence? Scholarship addressing this question has a long pedigree, but the literature has converged on two general classes of explanation, rational actor theories and psychological theories.2 Variations exist within these traditions, and both have influenced contemporary theories of group conflict. Rationalist accounts, sometimes referred to as “political” theories of violence, generally posit that individuals will engage in violence given a strong enough incentive and the opportunity. Charles Tilly, arguing against earlier theories which suggested that violence was accounted for by moral breakdown or isolated social deviance, posited that the proximate cause of political violence was the imposition of suffering (Tilly, 1963). Yet whereas suffering is pervasive, violence is episodic and not all groups who suffer commit violent acts. Under some conditions those who suffer will organize collectively to commit violence, whereas under many other conditions individuals either suffer without organizing or address their grievances through some other means besides violence. Thus, the relevant question is when will incentives for collective violence outweigh the costs? According to Tilly’s view, violence is purposive action aimed at fulfilling claims made on other groups and on the state (1974 a, b). The outcome of collective violence is the provision of collective goods for group members. Groups are thought to


See Rule, 1988, for a history of theories of civil violence.

4 continuously weigh the costs and benefits of actions that could potentially obtain such goods; if the costs of violence are outweighed by the expected benefits, collective violence ensues (Tilly, 1978). Other research in this tradition has focused on the amount of resources groups are able to bring to bear when engaging in collective action (McCarthy and Zald, 1973, 1987; Oberschall, 1973, Jenkins, 1983). Established organizational resources make it easier for groups to organize and execute collective action, thus lowering overall costs and enabling action. Notably, resource mobilization literature also assumes that discontent is pervasive, but variability exists in the ability of groups to act to rectify their situation. Not only will some groups be able to organize and act while others will not, but the character of action may vary with group characteristics. Oberschall (1973), for instance, suggests that collective action is especially likely to take a violent form when networks of insurgents are organized but insulated from broader society. Recent research has focused on the opportunities and incentives for collective action, arguing specifically that political violence is more likely where repression is not too severe (Schock, 1996) and where terrain and the political climate make insurgency easier to carry out (Fearon and Laitin, 2003). Political violence is often strategically used by extremists to induce breakdowns of peace talks (Kydd and Walter, 2002). Furthermore, violence that results in success breeds more violence, as it becomes obvious that the tactic is effective (Crenshaw, 1990; Pape, 2003). Such arguments are supported by strong empirical evidence and fit in the rationalist tradition that the decision to commit violent acts is a function of relative costs and expected benefits. In contrast, psychological theories posit that violence is a function of individual and group perceptions. An early exposition of this position was Gurr’s theory of relative deprivation (1970). Gurr argued that violence is an outgrowth of frustration occurring when “value expectations” exceed “value capabilities.” According to this logic, episodes of political violence are likely to be preceded by a period of rising expectations. When it becomes apparent that these expectations will not be met, groups undertake collective action out of frustration and in an effort to bring expectations and realized goals into

5 closer accordance. Early empirical tests of Gurr’s theory were suggestive, but mediating variables hypothesized to affect the process by which deprivation translates into actual violence, such as regime legitimacy, explained much more variation in violence than measures of deprivation (Gurr and Duvall, 1973). Political-rationalist accounts offer one explanation for this finding, namely, that whereas deprivation is relatively widespread, the ability to channel frustration into action is highly variable. Gurr later modified his theoretical framework, arguing that group identity and group cohesion interacts with active grievances and the potential for mobilization to generate violence (Gurr, 1993). Muller (1979) enhanced the understanding of psychological sources of political violence, incorporating social-psychological variables into his explanation. These variables, such as belief in the efficacy of violence, alienation from the political system, and exposure to social norms that encourage aggressive action consistently outperformed deprivation variables in analyses. Muller concluded that such factors, together with calculations of expected benefits, account for much of the variation in patterns of political violence. Gurr’s later analysis and Muller’s work reflect the evolving recognition of interaction between psychological and strategic motivations for violence. Other approaches take a social-psychological perspective, focusing on the role of group pressures and socialization. Case study evidence documents how members of extremist social movements become socialized into the movement and grow to accept violence as a legitimate means of affecting political change. Della Porta (1995) documents how members of the Red Brigades were recruited and insulated from broader society in order to prepare them for carrying out violent acts against the state, and a similar process was used in recruitment to the Weathermen in America during the 1960s and 1970s (Sprinzak, 1990).3 Interestingly, many of the extremists in these groups were educated and middle class (See also Whitaker, 2001), which is consistent with evidence suggesting that suicide bombers do not necessarily fit a certain “profile” of uneducated and underprivileged individuals (Spinzak, 2000). Rather, social psychological pressures may trump pre-established individual characteristics, such as social class. Jerald

Also see Moss, 1989; Ferracuti, 1990 for discussion of psychology of left-wing violence in Italy.

6 Post (1990) documents the process by which groups exert pressures to conform and commit acts of violence, and the influential psychologist Albert Bandura has explicated how terrorists undergo training in “moral disengagement,” by which recruits are transformed and taught to dehumanize the intended enemy (Bandura, 1990). It is important to note that during such processes individuals are typically surrounded by individuals with similar perspectives and goals and are frequently intentionally insulated from broader segments of society. Rationalist and psychological theories of violence are not necessarily mutually exclusive. It is possible, indeed probable, that a lack of low-cost alternatives to violence and the opportunity for violence interact with socialization and moral disengagement in facilitating the occurrence of violence. Moreover, the resources individuals have at hand to realize their political goals are intimately related to factors that influence socialization and the development of one’s perspective of others. The very networks that enable individuals to carry out plots to commit violence simultaneously act as socializing agents, shaping individual worldviews and ideas of inter-group relations.4 The next section discusses how varieties of civic networks and norms of interpersonal trust may influence the resources and motivations for political violence. Social Capital and Theories of Violence Social capital theory is concerned with how linkages and networks facilitate individual and group action (Coleman, 1988). It also relates to the process by which individuals become socialized and internalize certain norms that can facilitate social interaction (Portes, 1998). Social capital itself is a notoriously diffuse concept, and many alternative definitions have been proposed.5 Putnam, for instance, defines social capital as “the social networks and norms of reciprocity and trustworthiness that arise from them” (2000: 19). For the purposes of this paper, Putnam’s definition, which considers social networks and norms to be the central elements of social capital, is adopted. Putnam has argued that social capital

Social identity theory (Tajfel, 1978) suggests that individuals are motivated by an innate desire for self-esteem when they develop identity and perceptions of themselves versus others. See Mercer (1995), for a discussion of identity formation under the security dilemma. 5 See Narayan and Cassidy for discussion of alternative definitions and measures.

7 explains patterns in government performance in Italy (1993), and others have argued that societies rich in social capital produce more efficient economic outcomes (Narayan and Pritchett, 1999; Narayan and Cassidy, 1999; Grootaert, 1999; Fukuyama, 1995). Notably, such accounts bear close resemblance to theories of economic interaction that is “embedded” in a complex web of social structures (Polyani, 1944; Granovetter, 1985; Evans, 1995; 1996). Scholars have directly argued that social capital creates opportunities for citizens and groups to channel demands, suggesting that declining social capital partially accounts for alienation from government and declining levels of confidence in institutions (Putnam and Pharr, 2000).6 Social capital theorists have also increasingly begun to distinguish between types of social capital. For instance, Putnam’s study of declining civicness in America discusses the concepts of “bridging” and “bonding” social capital (2000). Bridging social capital refers to types of networks that link diverse segments of society, fostering generalized (as opposed to specific) trust and reciprocity. Bonding social capital, on the other hand, facilitates action, but involves linkages and networks that reinforce existing group affiliations and increase solidarity. While neither type is necessarily “good” or “bad,” the presence of bonding social capital in the absence of bridging social capital may facilitate distrust and misunderstanding of other groups. Others have noticed the potential of insular, tight-knit social capital to lead to negative outcomes, such as cronyism and corruption (Evans, 1989; Mauro, 1995). Varshney’s recent analysis of the causes of communal conflict in India delineates between types of networks, finding that cities with interethnic community associations were much less prone to outbursts of violence than communities lacking such organizations, in large part because interethnic groups were effective at dispelling rumors and spirals of conflict generated by isolated incidents (2002). Oliver and Wong (2003) find that “close proximity to out-groups corresponds with less racial antagonism” and that “negative perceptions of out-groups are higher for those who live in neighborhoods with more of their own racial
Policy makers at the World Bank have also begun to recognize the potential of social capital to facilitate development and economic performance, as evidenced by the World Bank “Social Capital Initiative.” See

8 group” (580), findings which are consistent with theoretical speculation about the effects of bridging and bonding social capital. Social capital theory holds direct implications for both rationalist and psychological accounts of political violence. On the one hand, if social capital lowers the costs of collective action, then the presence of social capital should make it easier for groups to mobilize to address collective grievances. However, the socialization effects of social capital, reflecting social-psychological theories of identity formation and group pressure, intersect with the instrumental-facilitative role of social capital to mediate the type of collective action that should occur. Bridging, cross-cutting, or inclusive social capital should facilitate collective action overall, but because it also eases interethnic tension and promotes generalized trust and reciprocity this action should not take violent forms. On the other hand, bonding social capital should ease action, but because it is, by definition, insular, it may also promote distrust and tension. Moreover, although it will facilitate organizational capacity, its insularity may eliminate access to channels with broader society which can serve as low-cost alternatives to violence. Under such circumstances, action is more likely to take violent forms. This logic leads to the hypotheses that higher levels of bridging social capital, in the absence of bonding social capital, should lead individuals and groups to choose non-violent forms of political expression. On the other hand, high levels of bonding social capital in isolation should lead to high levels of distrust, and general acceptance of political violence. When both forms of social capital are high individuals are not expected to support political violence. This is because ethnic pride and strong kinship ties alone should not make violence a less costly alternative or necessarily foster inter-group animosity. In other words, strong in-group ties will not have the same socialization effects when they exist along side broader social networks. Rather, it is the presence of strong insular ties in the absence of broader engagement that should foster distrust and facilitate violence. Low levels of both types of social capital will generate atomized individuals who, while perhaps frustrated and unsatisfied, will have low levels of attachment to groups and little resources to at hand to facilitate collective action. These individuals may

9 favor violence, but are unlikely to take part in concerted acts of political violence. These implications are developed and tested below. Research Design and Data An ideal test of the core hypotheses about the impact of bridging, or inclusive, and bonding, or exclusive, forms of social capital on political violence would involve measuring data on the occurrence of political violence and various indicators of social capital at the aggregate level of state, region, or locality and comparing geographic units, controlling for confounding factors. Several practical problems exist with this approach, however. These are discussed in turn and a strategy for exploring the relationship between social capital and acceptance of political violence is outlined below. Measuring Political Violence and Levels of Analysis One initial difficulty for any study of political violence is defining the dependent variable. Because studies often have different theoretical foci, and because some grey area exists between examples of nonconventional participation, such as protests, demonstrations, and strikes (Dalton, 2002) and actual occurrence of violence, different operational definitions have been used. For instance, Gurr (1970) defines political violence as “collective attacks within a political community against a political regime, and Tilly (1978) defines violence in general as “any observable interaction in the course of which persons or objects are seized or physically damaged in spite of resistance.” Jenkins and Schock (1992) refer to “domestic political conflict” as “noninstitutionalized coercive or threatening interactions between citizens and states.” In light of the present theoretical concerns, political violence will be defined as any event in which damage or seizure of persons or properties occurs, with political motivations. The theoretical foundation for this definition is in agreement with rationalists’ general position that violence is simply an extreme form of a continuum of political participation, ranging from generally-accepted forms, such as voting and contacting elected officials, to extreme forms, such as bombing, assassinations, and rioting. Alternative forms of participation, such as signing petitions or participating in demonstrations and strikes are arrayed in between these extremes, reflecting the relative costs of such acts. It is important to note that

10 according to this definition violence may be directed against the state or against other groups within the state. The operational definition implies an obvious choice for a measure of the dependent variable: realizations of political violence, such as bombings, assassinations and violent strikes and protests. However, reliable data on these events exists either for a limited sample of countries, a limited sample of phenomena, or for a limited time period, depending on the data source.7 More importantly, given the theoretical focus on individual and group explanations of conflict, it is difficult to collect reliable data on political violence at the level of individual and group involvement. A second problem arises at the level of measurement for the core explanatory variables. Setting aside for now the problems of operationalizing a concept as abstract as “social capital” (which will be discussed below), there are simply not adequate measures of the indicators that have been used by Putnam and others at the level of country or region for a reasonable amount of countries and time periods to make analysis feasible. For these reasons, analysis is restricted to individual level data collected in multiple countries at two time points, employing Inglehart’s World Values Survey. This survey includes items that are theoretically proximate to what others have used to operationalize social capital. This level of analysis is also theoretically appropriate, given the interest in the individual determinants of support for political violence. While the World Values Survey does not measure involvement in political violence directly, there is a question about individual belief in the legitimacy of political violence. The exact wording of the question is, “Here is another statement: how strongly do you agree or disagree with it? Using violence to pursue political goals is never justified.” Respondents are asked if the agree, disagree or strongly agree or disagree. This is an imperfect measure, but it captures individual attitudes towards the use of political violence.

For instance, the Banks data has only been extended through the present but its reliability has been questioned. Other data sources, such as the PRIO conflict data detail instances of low and intermediate civil conflict, but do not specify type of conflict.


11 To further test whether types of social capital facilitate other forms of political participation and whether the relationship between social capital and violence is qualitatively different than that between social capital and other non-conventional forms of participation, models are estimated with individual willingness to sign petitions, join boycotts, join lawful demonstrations, join unofficial strikes, or occupy a building in protest as dependent variables. Responses to these questions are “have done” (1), “might do” (2), and “would never do” (3). These questions allow tests of whether support for violence is differs substantially from support for other forms of unconventional participation that vary in terms of legitimacy and risk. Analyses are restricted to the 1990-93 and 1995-97 waves of the survey to ensure that the same countries are included in both waves and to focus analysis on a reasonably short interval of time to assume that political culture is relatively constant within countries. The survey includes data from 66 countries; a list of countries included in the final analysis (after eliminating cases with missing data) is included in the appendix. Measuring Social Capital Social capital has been used to refer to many things and is therefore difficult to definitively operationalize, let alone measure. After all, how does one define and measure norms, trust, reciprocity, and solidarity? Fortunately, literature on social capital has developed enough to provide examples to follow in this area. Putnam (1993) uses voluntary association membership, referenda turnout, and local newspaper readership as proxy members of social capital. Knack and Keefer (1997) use the World Values Survey to measure social capital, employing items on generalized trust and civic norms as measures of social capital. Likewise, Narayan and Cassidy (2001), in surveying various social capital measurement instruments, indicate that trust, membership in voluntary associations, family and friend connections, tolerance and reciprocity are key measures of social capital. The World Values Survey contains indicators of several of the concepts mentioned by Narayan and Cassidy. Given the interest in distinguishing between forms of social capital, indicators of bridging, or inclusive forms of social capital, are first selected. First, responses to the following question are used as a

12 measure of generalized trust, “Generally speaking, would you say that most people can be trusted or that you can't be too careful in dealing with people?” Respondents answer one for yes and zero for no. Generalized trust is thought to reflect bridging social capital much more than bonding because it specifically refers to trust without personal reference. Whereas specific trust would indicate trust in certain types of individuals, general trust is indifferent to social and personal characteristics, which is a norm that should arise from connections among diverse individuals. Another indicator of this type of social capital is involvement in inclusive civic organizations, or organizations that should have little or no social or personal prerequisites for membership. Responses to items asking about membership in sports, recreation, arts, music, education, environmental, or charitable organizations are collapsed into an index of organization membership. Responses to individual items were coded three if the respondent was an active member, two if the respondent was an inactive member, and one if the respondent did not belong. The responses comprise a simple additive index, varying from zero to one, reflecting organizational involvement. Other items that indicate various attitudes towards political and social involvement exist in the survey, but analyses are restricted to these measures because they are closest to the theoretical focus on social networks and norms of trust and because they have precedent in the literature. Informal ties are no doubt also a large component of bridging social capital; unfortunately, only indirect information on such ties is available in the format of the survey. Measuring bonding social capital is more difficult since scholars are only recently beginning to distinguish between types of social capital and recognize the potential harmful impact of exclusive networks. Such networks are also arguably more likely to be informal than bridging networks, which often have formal realizations in voluntary organizations. This poses difficulties for measurement, since “bonding” types of networks may be more unobservable than other networks. Several proxy measures of this concept are employed. First, respondents were asked to identify groups of people who they would not like to have as their neighbor. Individuals are coded one if they select the group in question, and zero otherwise. Individuals who mentioned immigrants, neighbors of a different race, or neighbors of a

13 different religion are collapsed into one additive index, reflecting distrust of individuals who are unlike the respondent. Several variables are also created based on responses to questions about immigrants and minorities. The first of these is coded one if individuals chose the most restrictive option, “prohibit people coming here from other countries,” to the question of what the government should do about immigrant workers. The second item is coded one if individuals picked immigrants or another minority group as the group the least liked from a list of possible social groups. These items are meant to reflect a general distrust of outsiders, which is theoretically indicative of strong insular ties. Another measure of level of insularity is the size of geographic area one most identifies with. Responses to this question range from one for town or locality to five for world as a whole. Respondents that identify first as residents of their town above region, state, nation, or large entities are thought to have a more inclusive conception of belonging, indicative of the tight-knit insularity of bonding social capital. Finally, to identify pride in ethnic grouping individuals are coded one if they answered that “above all, I am a black [ethnic group] American [nationality]” and zero if they answered “I am an American [nationality] first and a member of some ethnic group second.” Again, these are proximate measures of strength of homogenous group ties, and the presence of strong levels of bonding ties by no means precludes high levels of bridging ties. Table one summarizes core social capital variables; complete survey wordings are listed in the Appendix. [table one about here] Control Variables It is important to control for individual demographic characteristics and general satisfaction with life and government when examining individual support for political violence. Age, education, income, ideology, gender, life satisfaction, confidence in the national government, and town size are included as control variables. First, if support for violence is a function of the expected costs and benefits to violence, then costs should increase and benefits decrease with education, age, and income. More educated individuals are more likely to be members of middle and upper classes, also indicated by income, and are

14 thus more likely to have a stake in a strong stable political system. Older individuals are also more likely to have solidified positions in society or have a stake in seeing political stability to ensure social welfare, thus raising the costs for political violence. Gender is also included to control for differences between men and women in support for political violence. Gender is coded one for males and zero for females. Scholars of political participation have noted that there has been a generally shift towards unconventional forms of participation on the part of educated, middle class individuals, suggesting that these demographic characteristics should increase the likelihood of participating in collective action short of violence (Dalton, 2002; Inglehart, 1997) Ideology is included as a control for several reasons. First, those on the political right in industrial democracies are likely to be capitalists with a vested stake in the system. Moreover, Muller (1979) has argued that, in the psychological tradition, leftist ideology is correlated with political alienation and belief in the efficacy of violence. However, one might also posit that political extremists on either end of the political spectrum are more likely support political violence. After all, some of the most heinous examples of political violence in the past century were committed by right-wing extremists. Therefore, a quadratic term for ideology is included in order to test for non-monotonicity. Ideology is measured on a scale from one to 10, with one being extreme left and 10 being extreme right. Life satisfaction is controlled for in order to account for variance in support for political violence due to general frustration with one’s life position. Life satisfaction is measured on a scale from one to 10, with one indicating dissatisfied and 10 indicating satisfied. Similarly, we control for confidence in the national government in order to control for general frustration in the current government. This variable varies from “a great deal [of confidence]” (1) to “none at all” (4). Finally, town size is included because in many cases political violence is an urban phenomenon. However, theory does not suggest a strong directional hypothesis with regards to town size because urban residents who are exposed to more

15 violence may have a negative reaction to it or conversely may be more accustomed to the idea that violence is a form of participation.8 Finally, several country-level control variables are used to identify whether the hypothesized relationships hold across diverse countries. GDP per capita is included to control for level of development within countries, which is likely to be correlated with political violence (Powell, 1982), and regional dummy variables are included to control for variation by region. Analysis Ordered probit analysis is appropriate in this context because the dependent variable is categorical Robust standard errors are clustered by country to control for correlation in responses within countries (Jones and Steenbergen, 1997) and probability weights are included to correct for differences in sampling design within countries. 9 Results for models with attitudes towards violence and other participation forms as the dependent variables are presented in Table two below. [table 2 about here] Evidence is mixed. The index of civic involvement and trust are expected to have negative impacts on individuals’ support for political violence, but only the civic index has the correct sign and is statistically significant. Its substantive significance is suggestive, but not strong, given the scale of the threshold parameters reported at the bottom of the table. Holding all else equal, moving from 0 to 1 on the civic index (from its minimum to maximum) moves an individual .347 units towards the “agree” category with the statement that “the use of violence for political goals is never justified.” Depending on an individual’s scoring on other independent variables, this change could be enough to move the individual from one category to another. This magnitude change might also have no impact on an individual


Greater political violence in urban areas may also reduce social capital in such areas, leading to more atomized individuals. Lower levels of social capital may exist in urban areas at the outset, given that urban dwellers are less reliant on recurrent social interaction or kinship ties for everyday living. Implications of potential endogeneity and directions for dealing with these issues in future research are briefly discussed below. 9 Several model specifications were tried, including fixed-effects, to control for individual country characteristics. Some country dummies were significant, but substantive inferences did not change.

16 response, however, if other characteristics placed an individual in an intermediate range of support. Thus, civic engagement is likely to decrease one’s acceptance of political violence, but only marginally. The evidence for indicators of insular ties is similarly mixed, but suggestive. Distrust of neighbors of a different race, religion, or country has a positive impact on individual likelihood of rejecting the statement that political violence is never justified, as does strong ethnic identification and identifying minorities or immigrants as “least liked” groups. However, geographic identification has no effect, and holding anti-immigration attitudes actually decreases the likelihood of that one will reject political violence. Overall, three of the five hypothesized indicators of insular ties increase individual propensity to support political violence, but their substantive significance is not overwhelming. These results are nonetheless suggestive that insular ties make individuals more likely to support political violence. Most other variables perform as expected. There is evidence of a curvilinear relationship with regards to ideology, with more extreme individuals more likely to reject the notion that violence is never justified. Age, income, and education decrease the likelihood that individuals disagree with the statement, but age is the only variable to be statistically significant. Several regional dummy variables are significant, with Europe as the reference category. All else equal, respondents in North America and Africa are more likely than those in Europe to accept that political violence is justified, while respondents in Asia are less likely to belief in the legitimacy of violence. Country GDP per capita is negatively associated with attitudes in favor of political violence, as expected. The model results for other forms of participation are also substantively interesting. First, it is important to note that responses to questions regarding other forms of participation are coded as one for “have done,” two for “might do,” and three for “would never do.” Given the hypotheses about inclusive forms of social capital facilitating collective action, civic involvement and trust should have negative coefficients, or increase the likelihood of engaging in collective political action, while indicators of insular forms of social capital should have positive signs, or decrease the likelihood of engaging in such nonviolent forms of participation. The results provide some support for these hypotheses. The index of civic

17 involvement has a positive and significant impact on the likelihood of having engaged in forms of collective political action ranging from signing a petition to occupying a building in protest, but generalized trust displays no significant effect. Notably, indicators of insular ties generally decrease the likelihood that an individual has engaged in these forms of non-violent expression. With the exceptions of the effect of distrust of neighbors of different race, religion, or nationality on the probability of having joined a boycott and the effect of holding strong anti-immigration attitudes on the probability of joining a demonstration, indicators of exclusive forms of social capital decrease the propensity to engage in group forms of political expression. The control variables generally perform as expected in these models, as well. Consistent with notions of a “new citizen politics” (Dalton, 2002) and Inglehart’s “postmaterialism” thesis (Inglehart, 1997), education, income, and country-level GDP per capita increase individuals’ propensity to engage in collective forms of political expression and protest politics. This evidence is also supportive of resource mobilization theories of collective action (McCarthy and Zald, 1987), which suggest that mobilization is partially a function of collective resources and ability. The most interesting aspect of the empirical results is the difference between the model with attitudes towards political violence as the dependent variable and the models with various forms of protest politics as the dependent variables. The evidence is consistent with the above argument that the impact of social capital on political violence is qualitatively different than its impact on other forms of political expression. Strong inclusive civic ties are likely to decrease support for outright violence, while enabling other forms of unconventional participation, including the relatively extreme action of occupying a building in protest. On the other hand, strong exclusive social attitudes and norms increase individual support for political violence and decrease the probability of joining in non-violent forms of protest politics. This is supportive of the general hypotheses that bridging social capital facilitates norms of peaceful articulation of grievances while providing resources and channels for such articulation, while bonding social capital can foment distrust and limit access to broader channels of political expression.

18 Discussion and Conclusion This paper argues that social capital theory can serve as an extension to existing theories of political violence. Varieties of social capital that strengthen ties between diverse segments of society are likely to foster norms of interpersonal trust, provide outlets for dissatisfaction with the political system, and provide means of coordinating and acting to address collective concerns. On the other hand, varieties of social capital that reinforce existing group loyalties, particularly in the absence of more inclusive civic ties, can create inter-group distrust and limit individual participation in broader collective action. The cross-national empirical analysis largely supports these arguments and suggests that the relationship between social capital and other forms of non-conventional participation is different than that between social capital and political violence. This analysis is consistent with other recent findings that inter-ethnic ties facilitate peaceful conflict resolution, whereas the absence of such ties can reinforce animosity. Likewise, the arguments and empirical results in this paper are also consistent with evidence that those who participate in political violence can come from a variety of socioeconomic and educational backgrounds.10 What distinguishes those who condemn violence from those who are active participants may very well be the relative strength of inclusive to exclusive forms of engagement with one’s community and society. There are several reasons to regard the empirical results of this paper as only suggestive, however. First, individual attitudes towards political violence are not equivalent with individual participation in violence. Violence is extremely costly for individuals and individual willingness to participate in political violence is likely to be much lower than individual support for the idea of political violence. Second, although a number of important alternative explanations were controlled for, including life satisfaction, confidence in the government, and level of development, individual support for political violence is very likely influenced by recent experience with political violence. Future research should control for the frequency and severity of past violence in individual countries. Finally, and most importantly, the

Recall that the education and income were not statistically significant predictors of attitudes towards political violence.

19 presence of violence in society is likely to be damaging to social capital itself. Widespread societal violence, such as experienced in the Israeli-Palestinian conflict, may depress civic engagement and foster distrust and inter-group hatred. Therefore, not controlling for previous episodes of political violence may lead to spurious relationships between social capital and attitudes toward political violence, when the root cause might be the destruction of inclusive civic engagement by recurring violence. Future research should examine this possibility with more sophisticated econometric techniques such as endogeneity modeling. Despite potential problems with parsing out the various causal relationships between social capital and political violence, the results of this paper are an initial step towards developing a more nuanced understanding of rational and psychological motivations for violence. Political violence is extremely costly to contemporary societies and the relationship between civic institutions and violence is only made more salient by ongoing efforts to rebuild war-torn countries in the face of extremism. For established democracies, episodes of violence are more isolated, but not less destructive, as the bombing of the Oklahoma City federal building in 1995 demonstrated. Given the implications of technology and social change on civic involvement and the changing nature of security threats, as evidenced by extremist violence, more research on the social roots of political violence is warranted. The general results presented in this paper can be regarded as either pessimistic or optimistic. On the one hand, some evidence suggests that social capital is deeply rooted in societal traditions and is relatively unchangeable (Putnam, 1993). On the other hand, many have argued that social capital can be promoted throughout the world and the World Bank has increasingly turned to social capital creation and facilitation as a cornerstone of its development efforts. From this latter perspective, understanding the relationship between social capital, violence, and participation may bode well for civic initiatives aimed at strengthening democracy and peaceful conflict resolution. The high cost of violence alone is reason for continued examination in this area.


21 Tables Table1: Social Capital Indicators Inclusive Social Capital: Civic Networks: Membership in Arts, Education, Music, Sports, Recreation, Environmental, and Charitable Organizations Generalized Trust: Generally people in society can be trusted. Insular Social Capital: Distrust of Neighbors: Dislike of neighbors who are immigrants, a different race, or a different religion Least-Liked members of society: Dislike of immigrants or members of a different race Subjective ethnic Identification: Identify as member of an ethnic group above all else Geographic identification: Identify with town or locality above state or country Anti-Immigration: Animosity towards Immigrants; favor exclusion

Table 2: Ordered Probit Model Results
Violence Trust Petition Boycott Demonstration Strike .070 .100 .150†† .111 .028 (.193) (.082) (.057) (.072) (.181) Civic Index -.347** -1.837** -1.853** -1.717** -.675** (.022) (.290) (.028) (.056) (.044) Neighbor .120* -.030†† .104** .046** .045* Distrust (.010) (.028) (.022) (.052) (.067) Anti-.065†† .017** .152** -.027†† .122** Immigration (.013) (.005) (.027) (.009) (.052) Least-Liked .074** .133** .091** .161** .014 (.014) (.027) (.014) (.034) (.030) Subjective .200** .340** .425** .268** .203** Ethnic ID (.030) (.030) (.008) (.007) (.020) Geo-group ID .004 .042* .046 .086†† .008 (.011) (.020) (.041) (.027) (.015) Ideology .037** -.083 .135 .098 .120** (.012) (.059) (.090) (.115) (.005) Ideology2 -.005** .011** -.010 -.006 -.005* (.001) (.005) (.007) (.007) (.003) Gender .071 .071 -.170** -.216** -.163** (.053) (.017) (.006) (.026) (.019) Age -0.005** -.001 .008** .011** .005* (.0003) (.001) (.002) (.002) (.001) Education -0.007 -.090** -.058** -.056 -.049** (0.008) (.018) (.010) (.036) (.013) Income -.011 .001 .002 .0009 .001† (.020) (.002) (.003) (.0006) (.0004) Confidence in -0.038†† -.144** -.139** -.030 -.078 Nat. Govt. (0.013) (.048) (.030) (.023) (.051) Life -.015** .001 .002 .026* -.009** Satisfaction (0.002) (.013) (.021) (.013) (.003) Townsize 0.012† .030* -.076** .005 -.007** (0.005) (.007) (.002) (.031) (.0255) GDP per -0.00001** -.00004** -.00004** -.000010* -.000005** capita (.000004) (.000006) (.000007) (.000005) (.000006) N. America .2878†† -.195 .298 .578†† -.212 (.101) (.155) (.173) (.118) (.150) S. America -.016 -.405† .473†† .400†† .106 (0.262) (.176) (.166) (.132) (.083) Asia -.141†† -.480†† .250†† -.348†† -.931†† (.031) (.057) (.048) (.106) (.071) Africa 0.195†† .482†† .033 .528†† .021 (.024) (.041) (.057) (.058) (.039) Tau_1 -.390†† -2.131†† -2.49†† -.678†† -1.888†† (0.039) (.199) (.298) (.105) (.144) Tau_2 .510†† -1.021†† -1.09†† .561† -.560†† (.047) (.111) (.320) (.303) (.094) Tau_3 1.187†† ----(.153) N 9522 10555 9926 10086 9983 9913 Robust std. errors in parentheses. *1-tailed z-test significant at .05 level. **1-tailed z-test significant at the .01 level. †2-tailed z-test significant at .05 level. ††2-tailed z-test significant at .01 level. β0 set equal to 0 for identification. Occupy Building -.103 (.086) -.508** (.121) .053* (.023) .124 (.105) .134* (.065) -.031† (.014) -.016 (.016) .224** (.018) -.011** (.001) -.093 (.140) .012** (.001) -.032** (.009) -.005** (.001) .042†† (.007) -.025** (.005) .010 (.016) .000002 (.00001) -.721† (.320) -.488†† (.062) -.589†† (.064) -.895†† (.044) -2.199†† (.099) -.712†† (.164) --

23 Appendix Countries and number of respondents included in final analysis: Spain (635) U.S.A. (545) Mexico (480) Sweden (198) Brazil (684) Nigeria (783) Chile (498) Belarus (505) India (452) Lithuania (458) Latvia (631) Estonia (590) Ukraine (590) Russia (569) Peru (163) Venezuela (582) Armenia (2) Azerbaijan (682) Croatia (369) Bosnia Hercegovina (497) Descriptive Statistics Variable Obs violence 9520 petition 9691 boycott 9618 demons 9728 strike 9733 occbuild 9913 trust 9913 civicindex1 9913 nei_distrust 9913 antimm1 9913 least_liked1 9913 ethnicid 9913 geo_group1 9913 ideol 9913 ideol2 9913 gender 9913 age 9913 educ 9913 income 9335 conf_natgov 9913 lifsatis 9913 townsize 9913 gdppc 9913 Mean 1.776366 2.131359 2.572156 2.277138 2.723518 2.839806 .7663674 .3855824 .457581 .4133965 .2408958 .7993544 3.630485 5.553011 36.19923 .4897609 40.68556 5.649955 4.568613 2.654898 6.032583 5.1149 5209.096 Std. Dev. .9645945 .7810831 .6081167 .7387716 .5328296 .4069628 .4231625 .1018328 .8408706 .4924676 .4276488 .4005036 1.330536 2.315997 27.40088 .4999204 16.12438 2.29598 2.491077 .9371064 2.616196 2.52106 7538.783

24 Survey Wordings: V27. Generally speaking, would you say that most people can be trusted or that be too careful in dealing with people? 1 Most people can be trusted 2 Can't be too careful [TRANSLATION: ="have to be very careful"] Now I am going to read off a list of voluntary organizations; for each one, could you tell me whether you are an active member, an inactive member or not a member of that type of organization? Active(1) ; Inactive(2); Don’t Belong (3) V29 Sport or recreation organization V30 Art, music or educational organization V33 Environmental organization V34 Professional association V35 Charitable organization 1 2 3 On this list are various groups of people. Could you please sort out any that you would not like to have as neighbors? V52 People of a different race V56 Muslims* V57 Immigrants/foreign workers * In Western Europe; in other countries, you may substitute "Jews," "Christians," "Gypses" or some other small but salient minority group. V65. All things considered, how satisfied are you with your life as a whole these days? Please use this card to help with your answer. (answers range from 1(dissatisfied) to 10(satisfied)) Now I'd like you to look at this card. I'm going to read out some different forms of political action that people can take, and I'd like you to tell me, for each one, whether you have actually done any of these things, whether you might do it or would never, under any circumstances, do it. (Have Done, Might Do, Would Never Do) V118 Signing a petition V119 Joining in boycotts V120 Attending lawful demonstrations V121 Joining unofficial strikes V122 Occupying buildings or factories V123. In political matters, people talk of "the left" and "the right." How would you place your views on this scale, generally speaking? (Range from 1 to 10, 1 is Left and 10 is Right) V134. How about people from other countries coming here to work. Which one of the following do you think the government should do? 1. Let anyone come who wants to? 2. Let people come as long as there are jobs available? 3. Place strict limits on the number of foreigners who can come here? 4. Prohibit people coming here from other countries?

25 I am going to name a number of organizations. For each one, could you tell me how much confidence you have in them: is it a great deal of confidence, quite a lot of confidence, not very much confidence or none at all? V142 The government in [YOUR CAPITAL] Here's one more statement. How strongly do you agree or disagree with it? V164. "Using violence to pursue political goals is never justified." (Strongly Agree, Agree, Disagree, Strongly Disagree.) V167. I'd like to ask you about some groups that some people feel are threatening to the social and political order in this society. Would you please select from the following list the one group or organization that you like least? 1. Jews* 4. Immigrants *[if necessary, use functional equivalent for these items] V203. To which of these geographical groups would you say you belong first of all? Locality or town where you live 1 State or region of country where you live 2 [The U.S.] as a whole* 3 [North America]* 4 The world as a whole 5 V208. Which of the following best describes you? 1 [A] Above all, I am an Hispanic American 2 [B] Above all, I am a Black American 3 [C] Above all, I am a white American 4 [D] Above all, I am an Asian American 5 [E] I am an American first and a member of some ethnic group second [modify the ethnic groups in this question to fit your own society] V217. What is the highest educational level that you have attained? (use functional equivalent of the following, in given society) 1. No formal education 2. Incomplete primary school 3. Complete primary school 4. Incomplete secondary school: technical/vocational 5. Complete secondary school: technical/vocational 6. Incomplete secondary: university-preparatory 7. Complete secondary: university-preparatory 8. Some university-level education, without 9. University-level education, with degree V232. Size of town: 1 Under 2,000 2 2,000 - 5,000 3 5 - 10,000 4 10 - 20,000 5 20 - 50,000

26 6 50 - 100,000 7 100 - 500,000 8 500,000 and more

27 Bibliography Bandura, Albert. 1990 (1998). Mechanisms of Moral Disengagement. In Origins of Terrorism: Psychologies, Ideologies, and States of Mind. Baltimore: Johns Hopkins University Press. Coleman, James. 1988. Social Capital in the Creation of Human Capital.” American Journal of Sociology 94(Supplement): S95-S120. Crenshaw, Martha. 1990 (1998). The Logic of Terrorism: Terrorist Behavior as a Product of Strategic Choice. In Origins of Terrorism: Pyschologies, Ideologies, and States of Mind. Baltimore: Johns Hopkins University Press. Dalton, Russel J. 2002. Citizen Politics: Public Opinion and Political Parties in Advanced Industrial Democracies. 3rd ed. New York: Chatham House. Della Porta, Donatella. 2000. Social Movements, Political Violence, and the State: a Comparative Analysis of Italy and Germany. Cambridge: Cambridge University Press. Evans, Peter. 1995. Embedded Autonomy. Princeton, NJ: Princeton University Press. ______. 1996. Government Action, Social Capital, and Development: Reviewing the Evidence on Synergy. World Development 24(6): 1119-1132. Fearon, James D. and David D. Laitin. 2000. Violence and the Social Construction of Ethnic Identity. International Organization 54(4):845-877. ______. 2003. Ethnicity, Insurgency, and Civil War. American Political Science Review 97(1): 75-90. Ferracuti, Franco. 1990 (1998). Ideology and Repentance: Terrorism in Italy. In Origins of Terrorism: Pyschologies, Ideologies, and States of Mind. Baltimore: Johns Hopkins University Press. Fukuyama, Francis. 1995. Trust: The Social Values and the Creation of Prosperity. New York: Free Press. Gurr, Ted Robert. 1970. Why Men Rebel. Princeton: Princeton University Press. ______. 1993. Minorities at Risk: A Global View of Ethnopolitical Conflict. Washington, D.C. United States Institute for Peace. Gurr, Ted Robert and Raymond Duvall. 1973. Civil Conflict in the 1960s: A Reciprocal Theoretical System with Parameter Estimates. Comparative Political Studies 6: 135-71. Granovetter, Marc. 1985. Economic action and Social Structure: the Problem of Embeddedness. American Journal of Sociology 91: 481-510. Grootaert, Christiaan. 1999. Social Capital, Household Welfare and Poverty in Indonesia. Local Level Institutions Working Paper No. 6. Washington, DC: The World Bank. Inglehart, Ronald. 1997. Modernization and Postmodernization: Cultural, Economic, and Political Change in 43 Societies. Princeton: Princeton University Press.

28 Inglehart, Ronald et al. 2000. World Values Surveys and European Values Surveys, 1995-1997. [Computer file]. ICPSR version. Ann Arbor, MI: Institute for Social Research [producer], 2000. Ann Arbor, MI: InterUniversity Consortium for Political and Social Research [distributor], 2000. Jenkins, Craig J. 1983. Resource Mobilization Theory and the Study of Social Movements. In Annual Review of Sociology Vol. 9. Jones, Bradford S. and Marco R. Steenbergen. 1997. Modeling Multilevel Data Structures. Paper prepared for the 14th annual meeting of the Political Methodology Society, Columbus, OH, July 25, 1997. Knack, Stephen and Philip Keefer. 1997. Does Social Capital Have an Economic Payoff? A Cross-Country Investigation. Quarterly Journal of Economics 112(4): 1251-88. Kydd, Andrew and Barbara F. Walter. 2002. Sabotaging the Peace: The Politics of Extremists Violence. International Organization 56(2): 263-296. McCarthy, John D. and Mayer N. Zald. 1973. The Trends of Social Movements in America. Morristown, PA: General Learning Press. ______. 1987. Resource Mobilization and Social Movements: A Partial Theory. In Social Movements in Organizational Society, Mayer N. Zald and John D. McCarthy eds. New York: Transaction Books. Mercer, Jonathan. 1995. Anarchy and Identity. International Organization 49: 229-52. Moss, David. 1989. The Politics of Left-Wing Violence in Italy, 1969-85. New York: St. Martin’s Press. Muller, Edward N. and Erich Weede. 1990. Cross-National Variation in Political Violence: A Rational Action Approach. Journal of Conflict Resolution 34(4): 624-651. Narayan, Deepa and Michael F. Cassidy. 2001. A Dimensional Approach to Measuring Social Capital: Development and Validation of a Social Capital Inventory. Current Sociology 49(2): 59-102. Oberschall, Anthony. 1973. Social Conflict and Social Movements. Englewood Cliffs, NJ: Prentice-Hall. Oliver, J Eric and Janelle Wong. 2003. Intergroup Prejudice in Multiethnic Settings. American Journal of Political Science 47(4): 567-582. Pape, Robert. 2003. The Strategic Logic of Suicide Terrorism. American Political Science Review 97(3): 343362. Polanyi, Karl. 2001 (1944). The Great Transformation: The Political and Economic Origins of Our Time. Boston, MA: Beacon Press. Portes, Alejandro. 1998. Social Capital: Its Origins and Applications in Contemporary Sociology. Annual Review of Sociology 24:1-24.

29 Post, Jerrold M. 1990 (1998). Terrorist Psycho-Logic: Terrorist Behavior as a Product of Psychological Forces. In Origins of Terrorism: Pyschologies, Ideologies, and States of Mind. Baltimore: Johns Hopkins University Press. Powell, G. Bingham. 1982. Contemporary Democracies: Participation, Stability, and Violence. Cambridge: Harvard University Press. Putnam, 1993. Making Democracy Work: Civic Traditions in Modern Italy. Princeton: Princeton University Press. ______. 2000. Bowling Alone: The Collapse and Revival of American Community. New York: Simon & Schuster. Putnam, Robert, Susan Pharr, and Russel J. Dalton. 2000. Introduction: What’s Troubling the Trilateral Democracies? In Disaffected Democracies: What’s Troubling the Trilateral Countries? Princeton: Princeton University Pres. Rule, James B. 1988. Theories of Civil Violence. Berkeley: University of California Press. Schock, Kurt. 1996. A Conjunctural Model of Political Conflict: The Impact of Political Opportunities on the Relationship Between Economic Inequality and Violent Political Conflict. Journal of Conflict Resolution 40(1): 98-133. Sprinzak, Ehud. 1990 (1998). The Psychological Formation of Extreme Left Terrorism in a Democracy: the Case of the Weathermen. In Origins of Terrorism: Pyschologies, Ideologies, and States of Mind. Baltimore: Johns Hopkins University Press. ______. 2000. Rational Fanatics. Foreign Policy 120: 66-73. Tajfel, Henri ed. 1978. Differentiation Between Social Groups: Studies in the Social Psychology of Intergroup Relations. London: Academic Press. Tilly, Charles. 1963. Queries on Social Change and Political Upheaval in France. Princeton: Center for International Studies _______. 1974a. Do Communities Act? Sociological Inquiry 43:209-40 _______. 1974b. Town and Country in Revolution. In Peasant Rebellion and Communist Revolution in Asia, ed. John Wilson Lewis. Stanford: Stanford University Press. _______. 1978. From Mobilization to Revolution. Reading, MA: Addison-Wesley. Varshney, Ashutosh. 2002. Ethnic Conflict and Civic Life: Hindus and Muslims in India. New Haven: Yale University Press. Whitaker, David J. ed. 2001. The Terrorism Reader. New York: Routledge.

To top