! NON-STATE SECURITY, STATE LEGITIMACY AND POLITICAL PARTICIPATION IN SOUTH AFRICA By Danielle Carter A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Political Science-Doctor of Philosophy 2013 ! ABSTRACT NON-STATE SECURITY, STATE LEGITIMACY AND POLITICAL PARTICIPATION IN SOUTH AFRICA By Danielle Carter This dissertation explores the political consequences of non-state security provision in the context of South Africa. In the developed and developing world alike, many individuals rely on various forms of non-state security including commercial security companies, voluntary associations, and faith and community-based organizations to meet their security needs. This reliance is especially pronounced in developing countries where the state’s capacity to provide security is often weak. Though reliance on non-state security is quite widespread in South Africa (and the developing world more generally), little is known about the political consequences of this reliance. This dissertation therefore seeks to probe the attitudinal and behavioral consequences of non-state security reliance. I begin by testing the effect of non-state security reliance on individuals’ perceptions of state legitimacy. Using original survey data collected in Cape Town, Durban and Johannesburg, South Africa, I find that individual reliance on non-state security may strengthen or undercut perceptions of state legitimacy, depending on how individuals view the state’s normative role in service delivery. Where individuals feel the state should be responsible for the direct, day-to-day production of security, individual reliance on non-state security decreases perceptions of a legitimate state. But where individuals feel that the state should play more of a facilitative role in security provision, individual reliance on non-state security increases perceptions of a legitimate state. I then test the effect of non-state security reliance on political participation using Afrobarometer survey data. I examine five types of ! political participation including joining, collective action, contacting, protesting and voting. I argue that non-state security reliance should increase most forms of political participation (except protest), but only when the state is viewed as legitimate. I find that those who rely on non-state security and see the state as legitimate, are less likely to vote, but more likely to engage in non-electoral forms of participation such as protest and collective action. These results suggest that non-state security provision is key to shaping individual political attitudes and behavior in South Africa and other sub-Saharan African countries. ! Copyright by DANIELLE CARTER 2013 For Guy ! v! ACKNOWLEDGMENTS This Dissertation would not have been possible without the assistance of many people. I would like to thank the chair of my dissertation committee, Michael Bratton, for his guidance over the years. Thank you for being a superb model of scholarly success. I am grateful to the other members of my committee for their commitment to helping me further this dissertation. Thank you Jeffrey Conroy-Krutz, Richard Hula, and Mahesh Nalla. I would like to offer a special thanks to Adrienne Lebas who continued to mentor me and provide intellectual support, even though she was no longer at MSU. I extend a huge thanks to Seoyoun Choi, Helen Lee, and Nicholas Kerr for tirelessly listening to my ideas and providing feedback. I would also like to thank Funmbi Elemo for useful comments on my dissertation and years of friendship. I would not have been able to carry out my field research in South Africa without the generous assistance of many individuals and institutions in that country. I am especially grateful to Clifford Shearing and Bob Mattes at the University of Cape Town, and the Centre for the Study of Civil Society at the University of KwaZulu-Natal in Durban. I am also eternally grateful to the many South African men and women who participated in my survey and elite interviews, and thus made the writing of this dissertation possible. Finally, I am deeply grateful to all my family and friends who provided so much encouragement and support over the years. I would especially like to thank my mom, Felicia Wilson, my late father, Daniel Carter, Jr., my late paternal grandmother, Joan Pate, and my maternal grandmother, Edith Wilson, for always believing that I’d accomplish great things. I’d like to thank my nieces, Kyia, Na’jai, and Nyla, for bringing me so much joy over the years and my stepmother, Synthia Carter, for her support. Last but not least, I’d like to thank my partner, Jeff Kushner, for helping me, in so many ways, to make it to the finish line. ! vi! TABLE OF CONTENTS LIST OF TABLES…………………………………………………………………………….......x LIST OF FIGURES………………………………………………………………………….......xii CHAPTER 1: INTRODUCTION…………………………………………………….……….......1 INTRODUCTION OF RESEARCH QUESTION……………………………….……….2 WHY STUDY THE POLITICAL CONSEQUENCES OF NON-STATE PROVISION?...................................................................................5 THEORIZING THE RELATIONSHIP BETWEEN NON-STATE SECURITY AND PERCEPTIONS OF STATE LEGITIMACY……………………….……...8 Functions of the State as an Arranger………………………………... 13 DISSERTATION OUTLINE…………………………………..………………………...17 CHAPTER 2: LEGITIMACY, THE STATE AND SECURITY………………………………..20 INTRODUCTION…………………………………………………………………..……21 UNDERSTANDING LEGITIMACY…………………………………………..……….21 Conceptualizing Legitimacy……………………………………………………. 21 Dimensions of Legitimacy………………………………………………………23 Instrumental Legitimacy…………………………………………………24 Procedural Legitimacy…………………………………………………...25 Symbolic Legitimacy…………………………………………………….26 Objects of Legitimacy…………………………………………………………..27 THE STATE……………………………………………………..………………………29 Conceptualizing the State……………………………………………………….29 Functions of the State…………..……………………………………………....30 The Importance of Security Provision…………………………………..30 The State in Africa……………………………………………………………….33 TOWARD NON-STATE SECURITY PROVISION…..………………………………...36 EMPIRICAL WORKS ON LEGITIMACY…………………………..…………………39 CHAPTER 3: CRIME AND SECURITY IN SOUTH AFRICA……………………………......45 INTRODUCTION………………………………………………………………….……46 CRIME AND SECURITY UNDER APARTHEID………………………………….….46 The Politicization of Crime………………………………………………………46 The Distribution of Policing Services……………………………………………48 Non-State Security under Apartheid……………………………………………..49 Market-Based Security…………………………………………………...49 Societal-Based Security………………………………………………….51 CRIME AND SECURITY SINCE THE 1994 TRANSITION……………………….…53 Crime Trends in South Africa: 1994-2012………………………………………56 Popular Perceptions of Crime……………………………………………………62 Rate of Crime/Personal Security…………………………………………64 Victimization……………………………………………………………..70 ! vii! Evaluations of the Police and Courts……………………………………73 STATE RESPONSES TO CRIME IN POST-APARTHEID SOUTH AFRICA….….…77 NON-STATE SECURITY IN POST-APARTHEID SOUTH AFRICA………………...81 Market-Based Security…………………………………………………………...81 Societal-Based Security………………………………………………………….87 The Good, the Bad, and the Ugly of Non-State Security………………………..89 CHAPTER 4: RESEARCH DESIGN AND METHODOLOGY……..…….…………………...93 INTRODUCTION……………………………………………….……...……………….94 ORIGINAL SURVEY RESEARCH IN URBAN SOUTH AFRICA.…...……………...94 Why Design an Original Survey?..........................................................................94 Sampling Procedures and Study Participants……………………………………96 Description of the Survey Instrument…………………………………………..103 Data Collection Procedures and Survey Administration……………………….106 Recruitment/Training of Fieldworkers…………………………………107 Pilot Survey with Fieldworkers………………………………………...108 Administration of the Survey……………………………………….109 Quality Control…………………………………………………………111 Data Capturing………………………………………………………….111 QUALITATIVE INTERVIEWS WITH ELITES IN SOUTH AFRICA..….………….112 AFROBAROMETER SURVEY DATA……………………………….………………114 VICTIMS OF CRIME SURVEY DATA………………...….…………………………115 OFFICIAL STATISTICS…………..……………………….………………………….118 CHAPTER 5: EXPLAINING THE EFFECT OF NON-STATE SECURITY ON PERCEPTIONS OF STATE LEGITIMACY.......................................................................119 INTRODUCTION……………….……….………………………………………….....120 DESCRIBING NON-STATE SECURITY AND STATE LEGITIMACY…………….121 HYPOTHESES………….…………….………………………………………………..134 Controls…………………………………………………………………………139 EXPLAINING STATE LEGITIMACY…...…………………………………………...141 Testing the Effect of Non-State Security Reliance on Perceptions of State Legitimacy Using Afrobarometer Data……………………….154 CONCLUSION……………….……….………………………………………………..160 CHAPTER 6: EXPLAINING THE RELATIONSHIP BETWEEN SECURITY AND POLITICAL PARTICIPATION………......................................................................................163 INTRODUCTION………….…………………………………………………………..164 EXISTING EXPLANATIONS OF POLITICAL PARTICIPATION…...……………..166 Extant Literature on the Link Between Security and Participation…………….170 THEORIZING THE RELATIONSHIP BETWEEN SECURITY AND PARTICIPATION………………………..………………...172 Non-State Security and Political Participation…………………………172 Victimization and Political Participation……………………………….175 HYPOTHESES……………………….………………………………………………...177 Joining and Collective Action Hypotheses……………………………………..177 ! viii! Contacting and Voting Hypotheses…………………………………………….178 Protesting Hypotheses………………………………………………………….179 DATA AND INDICATORS………………..…………………………………………..180 Political Participation…………………………………………………………...180 Determinants of Political Participation…………………………………………183 EXPLAINING THE SECURITY-PARTICIPATION LINK…….………………….…187 Non-State Security, State Legitimacy and Political Participation……………...187 The Effect of Insecurity and Victimization on Participation…………………...189 Other Important Variables……………………………………………………...199 CONCLUSION…………………………….…………………………………………...201 CHAPTER 7: CONCLUSION…………………………………………………………………203 INTRODUCTION………………………………..…………………………………….204 SUMMARY OF KEY FINDINGS……………………………………………………..204 GENERALIZABILITY OF THE FINDINGS…………………………...……………..207 The Effect of Non-State Security on Perceptions of State Legitimacy in Africa…………………………………………...208 Non-State Security, State Legitimacy and Political Participation in Africa…………………………………………………..210 The Effect of Insecurity and Victimization on Participation in Africa…………215 Summarizing the Relationship between Security and Political Participation in Africa………………………………………………….217 CONTRIBUTIONS OF THE RESEARCH………………...………………………….219 The Political Consequences of Non-State Security Provision………………….219 The Political Consequences of Victimization and Personal Insecurity………...221 LIMITATIONS OF THE RESEARCH………………………………..……………….221 FUTURE DIRECTIONS……………………………………………………………….223 APPENDICES…..…………………………………….………………………………………..224 APPENDIX A: LIST OF SURVEY NEIGHBORHOODS BY CITY………...……….225 APPENDIX B: MASS SURVEY QUESTIONNAIRE ………………..………………227 APPENDIX C: MASS SURVEY CONSENT FORM…………………………………253 APPENDIX D: ELITE INTERVIEW CONSENT FORM………………....…………..257 APPENDIX E: QUESTION WORDING AND CODES FOR TABLES 5.5-5.6……………………………………………………….....261 APPENDIX F: QUESTION WORDING AND CODES FOR TABLES 5.7-7.1…………………………………………………..……...266 APPENDIX G: QUESTION WORDING AND CODES FOR TABLES 6.1-6.2, AND 7.2-7.3……………………………………….….270 REFERENCES…………………………………………………………………………………276 ! ix! LIST OF TABLES Table 3.1: Perceptions of Violent Crime in South Africa by Province, 2011……...……………67 Table 3.2: Perceptions of Personal Security by Province, 2011…………………...…………….69 Table 3.3: Victimization in South Africa by Race, 2011…………………………...……………71 Table 4.1: Proposed Sample Frame for Cape Town, Durban and Johannesburg, South Africa…………………………………...………………………………………..99 Table 4.2: Sample Design for Cape Town, South Africa………………………….…………...100 Table 4.3: Sample Design for Johannesburg, South Africa………….…………………………100 Table 4.4: Sample Design for Durban, South Africa…….…………………………………….101 Table 4.5: Total Sample Frame…………….…………………………………………………...102 Table 4.6: Elite Interviews in Cape Town, Durban and Johannesburg, South Africa.…………112 Table 5.1: Government Legitimacy by Race, 2010-2012………………………………………126 Table 5.2: Perceived ANC Control of State by City, 2010-2012…………………………...….127 Table 5.3: Type of Non-state Security Reliance by Race, 2010-2012………………………….133 Table 5.4: African Reliance on Non-state Security, 2010-2012………………………………..134 Table 5.5: Pooled Logistic Regression Output on Perceptions of State Legitimacy, South Africa ………………………………………………………………………......141 Table 5.6: Logistic Regression Output On Perceptions Of State Legitimacy, By City………...147 Table 5.7: Afrobarometer Logistic Regression Output On Perceptions Of State Legitimacy, South Africa…………………………………………………………………….155 Table 5.8: Afrobarometer Logistic Regression Output on Perceptions Of State Legitimacy, South Africa (Fixed Effects Model)……………………………………………158 Table 6.1: Regression Analyses of Joining and Collective Action……………….…………….191 Table 6.2: Regression Analyses of Contacting, Protesting and Voting……….………………..196 ! x! Table 7.1: Pooled Logistic Regression Output on Perceptions of State Legitimacy, 11 African Countries……………………….………………………………………209 Table 7.2: Regression Analyses of Joining, Collective action and Contacting, 11 African Countries……………………………….……………………………………….212 Table 7.3: Regression Analyses of Protesting and Voting, 11 African Countries….………..…214 Table 8.1: List of Survey Neighborhoods by City……………………………………………...225 Table 8.2: Question Wording and Codes for Tables 5.5-5.6………….………………………..261 Table 8.3: Question Wording and Codes for Tables 5.7 and 7.1……………………………….266 Table 8.4: Question Wording and Codes for Tables 6.1-6.2, and 7.2-7.3……………………...270 ! xi! LIST OF FIGURES Figure 1.1: Conventional View of The Relationship Between Capacity and Legitimacy…….…10 Figure 1.2: An Alternate Path from Capacity to Legitimacy………….…………………………11 Figure 1.3: The Affect of Non-state Security on State Legitimacy in Low Capacity States.……12 Figure 3.1: Rates of Murder/Attempted Murder per 100,000 of the Population, 1994-2012…....57 Figure 3.2: Rates of Assault per 100,000 of the Population, 1994-2012………….……………..58 Figure 3.3: Rates of Rape per 100,000 of the Population, 1994-2012……….…………………..59 Figure 3.4: Rates of Robbery per 100,000 of the Population, 1994-2012……….………………60 Figure 3.5: Rates of Motor Vehicle Theft per 100,000 of the Population, 1994-2012….……….61 Figure 3.6: Rates of Commercial and Residential Burglary per 100,000 of the Population, 1994-2012………………………………………………………………………..62 Figure 3.7: Rates of Commercial and Residential Burglary per 100,000 of the Population, 2003-2012………………………………………………………………………..62 Figure 3.8: Perceptions of Crime in South Africa, 2012………………………………………...64 Figure 3.9: Perceptions of Crime in South Africa, 2003-2012…………………………………..65 Figure 3.10: Perceptions of Violent Crime in South Africa by Race, 2011……………………..66 Figure 3.11: Perceptions of Property Crime in South Africa by Race, 2011…………………….66 Figure 3.12: Perceptions of Personal Security, 1998-2012………………………………………68 Figure 3.13: Perceptions of Personal Security by Race, 2011…………………………………...69 Figure 3.14: Victimization in South Africa, 1998-2012…………………………………………70 Figure 3.15: Primary Reasons for Not Reporting Crime to the Police, 2011……………………72 Figure 3.16: Evaluations of Police Performance, 2003-2012……………………………………73 Figure 3.17: Evaluations of Courts, 2007-2012………………………………………………… 74 Figure 3.18: Evaluations of Courts by Province, 2011…………………………………………..74 ! xii! Figure 3.19: Evaluations of Police Performance by Province, 2011…………………………….75 Figure 3.20: Evaluations of Police Performance by Race, 2011………………………………...76 Figure 3.21: Evaluations of Courts by Race, 2011………………………………………………76 Figure 3.22: Growth of the Commercial Security Industry by Number of Security Guards, 2001-2011………………………………………………………………………..82 Figure 3.23: Growth of the Commercial Security Industry by Number of Security Businesses, 2001-2011………………………….………………………………..83 Figure 5.1: Legitimacy of State Institutions, 2010-2012……………………………………….122 Figure 5.2: Legitimacy of the State in South Africa, 2010-2012……………………………….123 Figure 5.3: Legitimacy of the State in South Africa by Race, 2010-2012……………………...124 Figure 5.4: Legitimacy of the State in South Africa by City, 2010-2012………………………126 Figure 5.5: Security Reliance Crime Prevention, 2010-2012…………………………………..128 Figure 5.6: Security Reliance Post Victimization, 2010-2012………………………………....128 Figure 5.7: Primary Responsibility for Public Good Provision, 2010-2012……………………129 Figure 5.8: Primary Responsibility for Public Good Provision in Southern Africa, 20082009……………………………………………………………………………..130 Figure 5.9: Type of Non-state Security Reliance, 2010-2012……………………………...…..132 Figure 5.10: Crime and Security as the Most Important Problem, 2008-2009…………………135 Figure 5.11: Conditional Effect Of Non-state Security On Perceived State Legitimacy………144 Figure 5.12: Satisfaction with Non-state Security in South Africa, 2010-2012………………..150 Figure 6.1: The Effect of Non-State Security and State Legitimacy on Participation………….175 ! xiii! CHAPTER 1 INTRODUCTION ! 1! INTRODUCTION OF RESEARCH QUESTION In July 2012, a member of the Women’s Development Forum in South Africa emphasized the need to revive street committees. She said “We need street or court committees like in the old days to keep the gangsters out of our area. If you see them with guns walking around you start clamping down. If we have street committees then the gangs can no longer just come and shoot 1 as they like. We need to get every person who's at home to check up on what is happening ”. In July 2009, South African Police Minister Nathi Mthethwa suggests that street committees and community police forums need to be strengthened. He also mentions that unemployed youth could assist in combating street robberies, which comprise a substantial amount of crime in 2 South Africa . In 2008, speaking of the days of Apartheid, anti-apartheid activist Lungi Sisulu remarks that "We had a death rate in Soweto alone of 45 people per weekend. Now we talk of a death rate of 42 a month in the whole of Gauteng. I am not condoning murders, robberies or attacks in any sense: nobody can defend crime at any place, anytime. But you have to balance things when you say crime is bad.” He goes on to speak to the necessity of street committees, saying “That was why we formed committees to defend ourselves against this sort of crime. We did something about it in the townships then and it needs to be the same today. Everybody needs to be involved in trying to solve crime; you cannot have a policeman standing at every gate. People need to get !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 1 Cape Argus. “ ‘Let Army Sort Them Out; Frustrated Pastor Leads Demand for State of 2 Agence France Presse. “South Africa May Re-open Specialized Crime Units: Police Minister”. 1July2009. ! 2! organised, know their neighbours, keep in touch. We formed street committees and they 3 helped” . In 2007, the then minister of Safety and Security for South Africa, Minister Charles Nqakula, said that more resources would be provided to help communities fight crime. He said “The partnership between the communities and the police with respect to social crime prevention, especially serious and violent attacks, is going to be strengthened. One of the ways of doing that 4 is going to be the revamping of the Community Police Forums” . Later in 2007, the Minister credited local communities for helping to reduce murder and other violent crimes in the Western 5 Cape . In April 2006, a group of residents from neighborhoods in Durban, South Africa launched an organization called Chatsworth and District Against Crime (Cadac) to deal with issues of crime. The chairman, Dr. Paul Lutchman, said that the reason they launched this organization is because "The police have failed to ensure our safety. It is time we recruited our own troops -- the residents of Chatsworth -- to restore law and order…We will not stand for it any longer. We 6 want to send a strong message to the criminals that we will no longer be soft targets.” !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 3 The Sunday Independent. “He (Lungi Sisulu) Gave Up His Childhood To See Aparthied finally banished”. 10August2008. 4 BuaNews. “South Africa; Mobilization of People Against Crime to Shape Community Police Forums”. 16February 2007. 5 Africa News. South Africa; Substance Abuse Crimes Rise in Cape”. 7December2007. 6 ! Sunday Times. “South Africa; ‘People’s Army to Fight Crime”. 7May2006. 3! These stories provide a snapshot of the many calls by citizens and state officials alike for greater involvement of non-state actors in the provision of security in South Africa. A range of individuals from within the state and society support non-state security provision and this support has persisted over time. These stories demonstrate that many ordinary citizens and state officials view security in South Africa as an arena in which non-state actors should participate. Non-state security structures such as street committees, neighborhood watches and commercial security firms are seen as essential forces in combating crime and violence in South African society. Yet for all the many calls for greater non-state provision of security; little is known about the consequences of the non-state provision of this key good. This dissertation seeks to understand the political consequences of non-state security provision. Specifically, it explores the impact that individual reliance on non-state security has on citizens’ political attitudes and behavior. This question arose as I began to think about the very basic reason why states exist. The social contract theorists’ response to this question would be that states exist, by and large, to provide security. These theorists, as well as empirical political scientists who focus on the importance of public good provision, would argue that the provision of security not only provides the raison d’être of states, but that the social contract between state and society is predicated upon how well the state is able to deliver this good. The key question raised here is one of quality and how the state’s ability to provide adequate protections for individuals shapes their attitudes toward the state. Essentially, the emphasis is a state-centered one that focuses on whether and how well the state is able to provide this good. Yet, increasingly, citizens in both the developed and developing world extensively rely on non-state providers to meet their security needs. This raises a new and important question, one that focuses on the ! 4! consequences of who provides. In a world where non-state actors from commercial firms to vigilante groups participate heavily in the provision of security, the question of who provides is as essentially political as the question of the quality of state-provided services. This dissertation therefore probes the behavioral and attitudinal consequences of non-state security provision. I define non-state security as the assurance of personal physical safety and safety of property that is provided by private actors or actors outside of the public sector. I examine two types of nonstate security; market-based and societal-based. Market-based security is security that is provided in exchange for a fee. This form of security is therefore provided by commercial security companies. But since there are a plethora of non-state actors involved in the provision of security in South Africa, it is important to explore the full range of non-state security providers. Therefore, I also investigate the political consequences of societal-based security. Unlike marketbased security, the provision of societal-based security is not contingent on financial exchange. Instead, societal-based security is largely voluntary. In this dissertation, I account for several types of societal security actors that individuals may turn to, including relatives, community police forums, street committees and neighborhood watch groups. WHY STUDY THE POLITICAL CONSEQUENCES OF NON-STATE PROVISION? Political scientists have long studied the role that state provided goods and services play in boosting citizens’ support for the regime, government and individual political leaders. The conventional wisdom, perhaps best illustrated by the economic voting literature, is that citizens will punish political leaders when times are bad and reward them when times are good. Thus, we ! 5! have solid theoretical propositions and empirical support for the positive way that citizens respond to political elites when they can, in fact, provide the goods. But what about when those providing the goods are not state agents, but non-state actors? Scholars and donor agencies alike have come to recognize the important role that non-state actors play in the provision of basic goods and services. Particularly in the developing world, non-state actors such as corporations, non-profit organizations, community organizations and faith-based organizations play a key role in ensuring that important goods and services are delivered to the citizenry. Yet, there are very few empirical investigations of the political consequences of non-state provision. Important exceptions include the work by Sacks (2012) which explores how donor and NGO provision shapes legitimacy and the work by Cammett and MacLean (forthcoming) that investigates the implication of non-state provision for state capacity, equity of access and experiences of citizenship. This dissertation makes an empirical contribution by analyzing the political consequences of non-state security provision. Rather than focusing on non-state education or healthcare (as the few existing works on non-state provision do), I argue that security provides a logical starting point for illuminating the political importance of non-state provision. First, scholars often argue that security is not only an important public good, but a chief good that should primarily be provided by the state. More so than any other good, the provision of security is seen as a key, defining characteristic of the state, and, importantly, one that connects citizens to the state. This suggests that attitudes toward the state will suffer when this good is not publicly provided. Yet, it is an empirical question as to whether a lack of state-provided security will dampen state-society ! 6! relations as suggested by extant literature. Therefore, while scholars have emphasized the centrality of state-provided security, this dissertation will explore the importance of this criterion for ordinary citizens. In particular, I test whether publicly provided security is, in fact, necessary for favorable popular assessments of statehood, or whether non-state provision may generate a similar outcome. In this sense, the dissertation allows us to explore the role that non-state actors may play in linking citizens back to the state. South Africa presents an opportune place within which to study the political consequences of non-state security provision. First, there is a long history of non-state security reliance in this country, both at the level of state and society. Apartheid was a system of white minority rule that restricted the civil and political rights of the majority population. Under this system, the police were militarized, focusing less on crime detection and prevention and more on enforcing segregationist policies and quelling the political opposition. However, the public police force was understaffed and therefore the state often turned to the commercial security industry for assistance with these tasks. Non-state security mechanisms were also developed at the societal level. Africans, denied the protection of the state and often the direct targets of state violence during apartheid, developed a culture of “self-reliance” and created non-state security mechanisms such as street committees within their communities (Dixon et al. 2003; Emmett and Butchart 2000). These organizations provided protection against violence that was unleashed both by state agents and members of liberation movements who sought to secure a loyal base of supporters for the anti-apartheid ! 7! struggle (Baker 2002). Therefore, state and society heavily relied upon non-state security during the apartheid era. However, extensive reliance on non-state security is not a thing of the past. Even though many believed that individuals would rely primarily on the state for their security needs with the transition to a democratic regime in 1994, this has not been the case. Even under the new order, South Africans continue to rely on various non-state actors from commercial security companies to community police forums to meet their security needs. Part of the reason for the persistence of individual reliance on non-state security is the continued salience of crime and security under democracy. Of 20 sub-Saharan countries included in the 2008 Afrobarometer sample, South Africa is the country that has the highest proportion of individuals who feel that crime and security is the most important problem facing their nation. The salience of crime and security coupled with the widespread belief that the police are not capable or willing to adequately address crime, leads to continued reliance on non-state security. The persistence of this reliance makes South Africa an ideal place for studying the political consequences of non-state security. THEORIZING THE RELATIONSHIP BETWEEN NON-STATE SECURITY AND PERCEPTIONS OF STATE LEGITIMACY The conventional wisdom suggests that when political goods such as security are not adequately provided by the state, then citizens’ attitudes toward the state may become more negative. In fact, existing literature suggests that the lack of widespread security and a rule of law can lead to state weakness and eventually failure (Rotberg 2003; Wood and Dupont 2006). But increasingly ! 8! when the state fails to provide a good to citizens’ satisfaction, they may turn to non-state sources, be it the market or society, for the provision of this good. The question then becomes whether individual reliance on non-state security has an impact on citizens’ attitudes toward the state. In particular, I seek to understand whether non-state security reliance strengthens or undercuts citizens’ views of the legitimacy of the state. Drawing on Lipset’s classic definition of legitimacy, I define legitimacy as the capacity of the state to “engender and maintain the belief that existing political institutions are the most appropriate or proper ones for the society” (Lipset 1959, 86). In order to illuminate the effect of non-state security reliance on perceptions of state legitimacy, I distinguish between two roles that the state may play in service delivery; that of producer or arranger. This distinction is borrowed from Savas (2000) where he explains that “the service producer directly performs the work or delivers the service to the consumer [while]…the service arranger assigns the producer to the consumer, or vice versa, or selects the producer who will serve the consumer” (Savas 2000, 64-65). Importantly, Savas suggests that collective goods may be produced and arranged by either the public or private sector (Savas 2000). 7 While many citizens continue to see the provision of security as a government duty , the reality is that producers of that good are increasingly non-state agents. Thus, the important question for the purposes of this study is whether the production of security by non-state actors (state as arranger) makes for a more or less legitimate state in the eyes of ordinary individuals. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 7 When respondents from my 2010-2012 survey in Cape Town, Johannesburg and Durban were asked whose responsibility it is to keep people safe, 86% of respondents chose “central government” or “local government” as their response ! 9! Conventional works that stress the importance of state-provided public goods, have conceived of the state’s responsibility in one way, that is the state as producer. Hence, under these conditions, when the state fails to adequately produce security (as we see in many countries where state capacity is weak), the legitimacy of the state suffers. This outcome is depicted in Figure 1.1. Conceiving of the state’s role in such a narrow way means that under conditions of weak state capacity, state illegitimacy is the only logical outcome, as the ability of the state to be perceived as legitimate by the citizenry is intimately tied to its ability to adequately produce security (and other public goods). Figure 1.1: Conventional View of The Relationship Between Capacity and Legitimacy (State as Producer) Low State Security Capacity (Perceived) Low State Legitimacy (Perceived) High State Security Capacity (Perceived) High State Legitimacy (Perceived) I suggest that the assertion that the legitimacy of the state will necessarily suffer when the state fails to adequately produce security may be overstated. Therefore, my theory proposes a third way, one in which the legitimacy of the state may remain afloat, even under conditions of weak state security capacity (Figure 1.2). However, I argue that this third way is only possible when we account for the role of non-state actors in security provision (Figure 1.3). ! 10! Figure 1.2: An Alternate Path from Capacity to Legitimacy Low State Security Capacity (Perceived) Low State Legitimacy (Perceived) Reliance on NonState Security High State Security Capacity (Perceived) High State Legitimacy (Perceived) The theory put forth here suggests that under conditions of weak state capacity, citizens may turn to non-state actors for this provision of security. As depicted in Figure1.3, individual reliance on non-state security may strengthen or undercut perceptions of state legitimacy, depending on how individuals see the proper role of the state in service delivery (producer vs. arranger). From this perspective, responsibility for service provision may be subdivided into two categories. In the first instance, the state’s burden of responsibility is heavier. Here, the state is directly involved in the day to day production of security via an effective public police force. This is the state as producer. In the second instance, the state is still responsible for security, but acts primarily in a facilitative role, creating the space, regulations and policies that stipulate whether and how nonstate actors may participate in the governance of security. This is the state as arranger. In the first instance, when individuals rely on non-state security, but see the state’s proper role as one of a producer, perceptions of state legitimacy suffer. This may be the case when citizens see the responsibility for the production of security as largely resting with the state. For citizens who participate in political life and generate the revenue that states use to govern (payment of ! 11! taxes), they may expect basic goods and services to be directly produced by the state. Thus, when citizens rely on alternate sources to fulfill what are viewed as state responsibilities, citizens’ confidence and belief in the state’s right to rule may suffer. I posit that the gap between citizens’ security demands and the state’s security supply may contribute to views of an illegitimate state. When this good is produced by non-state sources, citizens’ needs are met, but citizens may come to view market or societal structures as competing sources of authority in which they vest their trust and loyalty. In this instance, state production of security provides the raison d’être for states from the perspective of ordinary citizens. Thus, when citizens rely on non-state sources for the production of this fundamental good, the legitimacy of the state itself may be thrown into question. Figure 1.3: The Affect of Non-state Security on State Legitimacy in Low Capacity States Low State Security Capacity State as Producer Low State Legitimacy (Perceived) State as Arranger High State Legitimacy (Perceived) Reliance on Non-State Security In the second instance, when individuals rely on non-state security and see the state as an arranger of this good, the state is able to bank legitimacy dividends. In this scenario, citizens may recognize that the state provides the constitutional framework within which non-state security actors are allowed to operate. Without the freedom of association and the legal right to bear arms ! 12! and secure contracts, non-state groups would not be able to provide security. Moreover, citizens may recognize that the state regulates non-state actors in their capacity as security providers. Therefore, when the state is seen as an arranger of security, it may receive credit from their citizens for the moral, technical and administrative support it provides for these actors. This may particularly be the case if these non-state security actors are viewed as effective in their provision of security. The “third way” that I introduce suggests that non-state actors may play a key role in attenuating the link between weak state capacity and state legitimacy. Figure 1.3 accounts for the legitimacy outcomes that we should witness depending on whether individuals view the state as an arranger or producer of security. Functions of the State as an Arranger The role of state as producer is the one that is probably best understood by scholars and citizens alike, since this is how we have traditionally thought of the state’s role in society. In fact, we have often conflated state provision of public goods with state production of them, assuming that if states are adequately providing education, security, and healthcare, for example, then we should correspondingly observe a plethora of public schools, police forces and clinics. However, states may also ensure that the citizenry is provided for without directly producing the goods themselves. Because the state as arranger is a relatively new idea, it is worth explaining what functions the state might perform in this role. When the state is acting as an arranger of security, it might do one of three things. First, the state may create space for non-state actors to participate in the provision of security. This is largely a ! 13! constitutional act, one in which the state recognizes which entities are legally allowed to provide security. In South Africa, there are constitutional provisions that allow for two main types of non-state security actors. These are the commercial security industry and individual citizens who are envisioned as participating in crime prevention through structured Community Police Forums (See chapter 3). Second, state officials may issue rhetorical or symbolic endorsements of non-state security actors. Beginning with his presidential campaign in 2008, current President Jacob Zuma made several references to the need to revive street committees, local community-based structures (initially developed under apartheid) that organize around issues of crime and violence at the street level. He has made comments such as “We reiterate the call for ANC branches to provide support to law enforcement agencies in the fight against crime, including establishing and 8 strengthening street committees” . Calls for greater community involvement in the fight against crime have been echoed by many public and state officials including provincial premiers, the national police commissioner Nathi Mthethwa and other members of the South African Police 9 Service . In fact, some police officials directly attribute reductions in crime to the work of 10 ordinary citizens organized through street committees . In these verbal endorsements, state !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 8 Africa News . “NGC 2010-Political Report of the President of the ANC Jacob Zuma”. September 22, 2010. 9 Sunday Times. “It’s time for a New Approach”. September 12, 2010. Agence France Presse. “South Africa May Re-Open Specialized Crime Units: Police Minister”. July, 1 2009. Daily News. “All Help is Welcome to Fight Crime: SAP”. November 12, 20009. 10 Cape Argus. “Mitchells Plain ‘Not Saturated With Drugs’ Despite Having Highest Number of Arrests”. September 25, 2009. ! 14! officials often call on ordinary citizens to get involved in crime fighting, primarily by participating in crime prevention activities. However, there is also some evidence that state officials often endorse non-state security actors in non-verbal ways. In particular, it has been suggested that the police will often allow civilian members of community police forums and street committees to harshly “deal” with alleged criminals. A member of the Social Justice Coalition, a security focused NGO based in the Khayelitsha township of Cape Town, spoke to this tacit endorsement of violence on behalf of the police. He suggested that at times, the police fail to intervene and even allow the community to 11 kill people . During my own field research in South Africa, a chairperson of a community police forum admitted that they would use violence to discipline alleged criminals. Members of the community police forum, in turn, are provided with extra protection from police in case 12 alleged criminals decide to retaliate . Therefore, the police often tacitly endorse violence as a legitimate response to crime by allowing community members to essentially participate in law enforcement duties by way of their role in sentencing and meting out punishment to alleged offenders. Finally, the state may act in a regulatory capacity, deciding who may have license to provide security and empowering, monitoring and sanctioning these actors according to a set of statedevised standards. In South Africa, the key body responsible for this task is the Private Security Industry Regulatory Authority. Its executive council is appointed by the minister of police, and is !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 11 12 ! Los Angeles Times. “The horror of vigilante justice”. September 24, 2012. CPF chairperson. Personal interview. 30December2010.! 15! responsible for monitoring and evaluating the activities of private security providers within the commercial security industry. As part of this task, the PSIRA has the authority to approve and revoke licenses for commercial security providers. State regulation of societal-based security is much more difficult to achieve. Because societal security structures are largely staffed by ordinary citizens and found throughout many local communities, the costs of effectively monitoring these structures would be very high. Thus, many of these structures function autonomously of the state. However, even in the area of societal-based security, the South African state has attempted to gain leverage and control. The clearest way in which it has attempted to do this is by encouraging citizens who care about crime and security issues to join Community Police Forums (CPFs). Because community police forums are officially housed at police stations, and because at least one police officer is supposed to sit on each forum, the state has attempted to use the CPF as a way of regulating the behavior of ordinary citizens and setting the tone and direction of their participation in crime fighting. The purpose of the above discussion has been to provide tangible examples of the activities the state may undertake in its capacity as an arranger of security. However, it is worth noting here that the theoretical thrust of this dissertation is not concerned with whether the state actually acts as a producer or arranger, but whether citizens believe it should act in one of these two different capacities. ! 16! DISSERTATION OUTLINE Chapter 1 has presented an overview of the main research questions raised throughout the dissertation and my theoretical approach for explaining how non-state security shapes perceptions of state legitimacy. Chapter 2 reviews relevant literature on the key concepts that this dissertation is concerned with; legitimacy, the state and security. The chapter begins by discussing the conceptualization of legitimacy. It then goes on to discuss various dimensions of the concept and the objects to which citizens may attach legitimacy. This provides a nice segue into a discussion of the state, including the state in Africa. Finally, this chapter reviews literature pertaining to security and the proposed relationship between state security provision and state legitimacy. It outlines some of the theoretical arguments in favor of state security provision as well as examines some of the empirical works that explore the factors that shape individual perceptions of state legitimacy. Chapter 3 is a contextual chapter that delves into the nature of crime and security in South Africa. It begins by examining crime and security issues in this country under apartheid, including a thorough discussion of the historical role that non-state actors have played in security provision. The chapter then discusses contemporary crime and security trends and challenges using official police crime statistics, policy documents and public opinion data. It examines state responses to crime as well as the range of non-state actors that are involved in security provision in post-apartheid South Africa. In essence, this chapter helps to answer why South Africa provides an optimal setting for exploring the political consequences of non-state security. ! 17! Chapter 4 lays out the research design and methodology devised for this dissertation. It describes at length an original mass survey that I carried out in Cape Town, Johannesburg, and Durban South Africa. Specifically, I discuss the design and implementation of the survey, including the development of fieldwork protocols and methods for the recruitment, training and supervision of fieldworkers. The chapter also discusses the elite interviews and other key sources of data I employ throughout the dissertation such as the Afrobarometer and Victims of Crime surveys. Chapter 5 is the first empirical chapter. It tests the effect of non-state security reliance on perceptions of state legitimacy using data from my original survey. The first part of the chapter is descriptive, showing, for example, the distribution of legitimacy attitudes and the percentage of individuals that rely on each type of security (i.e. public police, community police forums, commercial security companies). I also introduce a discussion about responsibility, exploring whom South Africans see as responsible for the provision of security and other public goods. I then move on to discuss and explain the results of regression analyses, highlighting what the results tell us about the effect of non-state security on state legitimacy in South Africa. I find that the identity of security providers does matter for individual perceptions of state legitimacy, but that other factors, such as being a victim of violent crime, hold explanatory power as well. I end the chapter by conducting a robustness check of my model using Afrobarometer data. Given that my survey was limited to urban areas of South Africa and that I lacked some variables that would allow me to successfully control for some competing explanations, I decided to run a similar model using nationally representative survey data collected from October-November 2011 in South Africa. ! 18! Chapter 6 is the second results chapter. While chapter 5 is concerned with understanding citizens’ attitudes toward the state, chapter 6 attempts to move beyond attitudes to assess political behavior. Given the importance of victimization in chapter 5, I expand the explanatory scope of chapter 6 to investigate the behavioral consequences of several security-oriented factors. Most importantly, I explore the impact of non-state security and victimization on political participation, but I also examine the role of personal insecurity and evaluations of government performance on political order. I investigate the impact of these variables on electoral and non-electoral forms of political participation, including joining, collective action, contacting, protesting and voting. I find that, who individuals turn to for their security needs is a key determinant of political participation, as is insecurity and victimization. Importantly, the impact of victimization varies depending on the type of victimization under consideration (property vs. contact) and the type of political participation. Analyses for this chapter rely on the latest round of Afrobarometer data collected in South Africa from October-November 2011. Chapter seven concludes the dissertation. This chapter briefly re-states the main findings and thoroughly discusses their implications. It speaks to the generalizability of these findings and tests how well my theory holds in other African countries using survey data from Afrobarometer. Finally, this chapter discusses the limitations of this study and suggests directions for future research. ! 19! CHAPTER 2 LEGITIMACY, THE STATE AND SECURITY ! 20! INTRODUCTION This chapter reviews literature on legitimacy and the state. I first discuss how legitimacy is conceptualized and the various dimensions of the concept. Next, I explore the objects to which citizens may attach legitimacy. I then focus on the key object of legitimacy that this study is concerned with, namely, the state. I introduce the conceptualization of the state that is employed throughout this dissertation and move on to highlight theoretical literature that speaks to the link between state security provision and state legitimacy. Finally, I end by exploring the existing empirical works on legitimacy. UNDERSTANDING LEGITIMACY Conceptualizing Legitimacy The classic definition of political legitimacy, and the one that will be employed throughout this dissertation, comes from Seymour Martin Lipset’s article “Some Social Requisites of Democracy: Economic Development and Political Legitimacy” written in 1959 (Lipset 1959). Lipset defines legitimacy as “the capacity of the political system to engender and maintain the belief that existing political institutions are the most appropriate or proper ones for the society” (Lipset 1959, 86). There are two important points to note regarding this definition. The first is that the political system must be able to do something; that is it must be capable of convincing society that it, in its current form, is superior to all other alternatives. At a later point I will examine the various component parts of the “political system” and explore the ways in which ! 21! they may go about convincing society of their appropriateness. The second important point has to do with the question of who needs convincing of the appropriateness of current political institutions. These individuals may consist of powerful elites within society who could threaten the political order if unsatisfied or, ordinary citizens. Two years before Lipset’s influential article was published, David Easton published an equally important work dealing with the concept of political support(Easton 1957). Easton defines legitimacy as “the conviction that it is right and proper …to accept and obey the authorities and to abide by the requirements of the regime” (1975, 451). Like Lipset, Easton sees legitimacy as involving the idea that the political system is viewed as the most proper. He emphasizes the element of obedience, suggesting that those who see the system as legitimate will usually feel an obligation to obey authorities and to accept their decisions as binding(Easton 1975). The primary concept of interest for Easton is that of political support. While the concept of support is not exactly the same as legitimacy, it is related. Easton (1957) defines support as “energy in the form of actions or orientations promoting and resisting a political system”. If, for example, citizens’ actions are aimed at resisting a political system, this may be indicative of the fact that they do not see the existing institutions as the most appropriate or proper. However, if citizens’ actions promote the existing institutions, it may be safe to say that they see them as legitimate and thus are likely to be in favor of retaining them. I will address the different types of support that Easton examines in the next section, but for now suffice it to say that both concepts, legitimacy and support, bring the attitudes and behaviors of ordinary citizens to the forefront and stress the importance of favorable popular orientations toward the political system. ! 22! Since Easton and Lipset, other authors have advanced similar conceptualizations of the term legitimacy. For example, Booth and Seligson (2009) in their work on political legitimacy in Latin America, use legitimacy and support interchangeably, arguing that legitimacy consists of citizen support for government. Jackman (1993) suggests that “a regime is legitimate to the extent that it can induce a measure of compliance from most people without resort to the use of physical force” (98-99). This definition focuses more on the behavioral dimension of legitimacy, but is consistent with other studies which have argued if people see political institutions as legitimate and hold favorable attitudes toward these institutions, the result will be voluntary compliance (Gibson and Caldeira 2003; Levi et al. 2009) In short, most definitions of legitimacy emphasize the importance of citizens who hold supportive attitudes and perhaps even engage in behavior that upholds the current political institutions. Therefore, while acknowledging that elite support of the political system is important, I focus my attention primarily on the attitudes of ordinary citizens, and examine the factors that contribute to popular perceptions of legitimacy. But how exactly are perceptions of legitimacy generated among the masses? That is the subject of the next section. Dimensions of Legitimacy Now that we understand what political legitimacy means, it is important to explore the ways in which legitimacy is developed. There are three primary ways through which legitimacy is generated, through the provision of basic goods and services (instrumental legitimacy) through adherence to/respect for rules and procedures (procedural legitimacy) or through affective means ! 23! (symbolic legitimacy). We will first discuss the instrumental sources of legitimacy before turning to the other types. Instrumental Legitimacy Lipset (1959) makes a critical distinction between effectiveness and legitimacy. He sees the performance of a political system as mainly an attribute of effectiveness, arguing that a political system is deemed as effective or ineffective, based, in large part, on its performance on economic development. Lipset separates legitimacy and effectiveness because he is interested in understanding political stability and, in turn, how effectiveness can help to sustain a political system while legitimacy is being built. Conversely, Lipset encourages readers to ponder how legitimacy can help to keep a democratic system afloat when its effectiveness is down. It would therefore seem that for Lipset, effectiveness is inherently instrumental, while legitimacy is more intrinsic. However, it is possible that individuals may come to see the state as more or less legitimate depending on how well it is able to meet their basic instrumental needs. Lipset allows for this possibility when he states that “prolonged effectiveness which lasts over a number of generations may give legitimacy to a political system; in the modern world, such effectiveness mainly means economic development” (Lipset 1959, 91). Therefore, Lipset acknowledges that the delivery of economic goods (in this case economic development) can be a crucial determinant of legitimacy. ! 24! With the heavy emphasis on the provision of economic development, the question remains open 13 as to whether and how the delivery of political goods shapes perceptions of legitimacy. Easton also acknowledges the performance-based dimension of political support, which he refers to as “specific support”. For Easton, members of society come to support political authorities when these authorities are able to generate outputs that meet the needs and demands of citizens (Easton 1957). The outputs that Easton speaks of here are what he refers to as “policy decisions” and essentially are the goods and services that governments supply for the citizenry. As acknowledged earlier, Easton is speaking of political support, but it easy to see how political authorities may also gain legitimacy through the effective provision of basic goods. Thus, the performance-based dimension of legitimacy stems from how well citizens are supplied with goods and services. Procedural Legitimacy The importance of basic provision to the development of political legitimacy has been noted. But not all legitimacy is conditional on material exchange. Several scholars have distinguished between legitimacy or supportive attitudes that come from the provision of material goods to ones that are derived from the way in which rules and procedures are upheld within a given society (Bratton and Mattes 2001; Diamond and Morlino 2004). Those who focus on procedurally based sources of legitimacy follow the path of Weber, emphasizing rational-legal notions of legitimacy and the importance of rule-based behavior. These works suggest that the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 13 Pennock (1966) describes political goods as the set of collective goals that “makes the polity valuable to man”. His list of core political goods includes security, justice, liberty and welfare ! 25! development of rules that are perceived to be fair and the fair and equal application of such rules will lead to legitimating attitudes among the citizenry (Tyler 2003; Levi et al. 2009). Many works that examine procedurally based support of political institutions have focused on democracy and elections. For example, Jackman (1993) notes that “regimes are often judged by their basis in “popular consent”…so that the refusal of any government to employ the electoral test is itself taken as evidence against its legitimacy” (p 97). Many studies therefore use citizens’ perceptions of election quality to gauge individual support for the regime. But more recent work has focused on particular institutions of the state, exploring, for example, the important role that perceived procedural fairness has on legitimating courts and police forces (Gibson and Caldeira 2003; Mondak 1993; Tyler and Fagan 2010). Procedural legitimacy is seen to be a particularly important way to garner the trust and obedience of citizens, even if and when they are unhappy with political outcomes. However, as Saward (1992) notes, a policy might be regarded as legitimate because due process requirements are satisfied…and yet the consequences flowing from the implementation might be morally problematic or disastrous in some way” (3536). Saward’s insight highlights the importance of accounting for both procedural and instrumental forms of legitimacy (Saward 1992). Symbolic Legitimacy The final dimension of legitimacy focuses more on the affective basis through which individuals come to identify with and lend credence to the political system. The affective dimension, whereby citizens have deeply held attachments to political institutions, can be developed through ! 26! processes of socialization and/or through experience with rituals and symbols that reinforce the “rightness” or “appropriateness” of those institutions. These deeply held attachments or affective bases of legitimacy are what Easton (1975) refers to as diffuse support, support that is not contingent on what political authorities do. Instead, this dimension of legitimacy is more about what the political institutions represent for individuals (Easton 1957). This is what Crawford Young in his most recent work refers to as the “state as idea”, an idea that has been “imprinted in the minds of its subjects and agents as an array of images, norms, and expectations” (Young 2012, 42). When considering this dimension, it becomes possible to see how individuals may continue to see that state as legitimate, even when they are dissatisfied with the provision of material goods and unhappy with how political processes are conducted. From this perspective, individuals see the state as legitimate because they have been taught to see it as such(Young 2012); perceptions of state legitimacy remain positive more so out of habit than as a result ongoing critical evaluation (Jackman 1993). Objects of Legitimacy When Lipset published his classic work on legitimacy and economic development in 1959, he was focused on the political system broadly speaking. However when speaking of legitimacy, it is important to disaggregate the “political system” and specify which component part is under examination. Key scholars have suggested that we may distinguish between the political community, the political regime and political authorities (Easton 1957, 1975; (Norris 1999). ! 27! When individuals attach legitimacy to the political community, they see themselves as an integral part of a collective. Individuals see their interests as compatible with the interests of the broader community or nation and are therefore willing to sacrifice and contribute to the good of that broader community. According to Norris, “the boundaries of a political community can be defined more narrowly in terms of a local or regional community, or a community defined by political cleavages based on ethnic, class, or religious identities” ( 1999, 10-11). The political community (and the legitimacy thereof) refers to how connected one feels to other citizens. Easton (1957, 1975) suggests that regime legitimacy, by contrast, refers to the notion that the rules of the game are seen as right or appropriate. Norris (1999) later makes even more finegrained distinctions of the regime. Her classification suggests that citizens may not only see the regime as distinct from the other political entities such as the political community, but that citizens also make distinctions within the regime itself. She argues that citizens in fact distinguish between the principles of the regime (i.e. democratic principles), the performance of the regime as a whole, and the performance of particular regime institutions. Finally, the legitimacy of political authorities depends on the view that those who occupy office are ruling in a manner that is acceptable to those they rule. According to Easton, political authorities may include “all public officials from chief executives, legislators, judges and administrators down to local city clerks and policemen, as well as the institutions, such as legislatures or courts, of which they are a part” (1975, 438). Both Easton’s and subsequently Norris’ discussion of political authorities includes institutions that may be considered to belong to the realm of the state. However, neither of these authors specifically distinguishes the state as ! 28! an entity to which citizens may attach legitimacy. Therefore, I will spend the next section discussing the state, the object with which this dissertation is concerned. THE STATE Conceptualizing the State One of the most frequently employed definitions of the state is that of Max Weber(Weber 1984). Weber defines the state as “a human community that successfully claims the monopoly of the legitimate use of force within a given territory” (Weber 1984, 33). For Weber, there must be certain attributes present before we can say that a state exists. These attributes include a bounded territory, a population that resides within that territory, and perhaps most importantly for Weber, a monopoly over the means of coercion. Moreover, all modern states are run by an executive who is ultimately responsible for commanding institutions of coercion such as the police and armed forces. Weber sees the legitimate use of force as a key, defining attribute of the state. From his perspective, a monopoly over the use of force is crucial to state survival, as one of the state’s main goals is to maintain dominance in society and secure obedience from the citizenry. Importantly, others may gain the right to use force, but only when the state so allows. Thus, the state is always sovereign in all areas, but especially in deciding who may use force and how. ! 29! Functions of the State Many scholars have followed in the footsteps of Weber in defining the state in terms of its means (i.e. monopoly over violence). But in addition to defining what constitutes the core coercive functions of a state, it may also be helpful to speak to the additional key functions that most modern states perform. States are responsible for providing a whole range of goods and services. These include the building of domestic infrastructure including roads, schools, and clinics, the collection of revenue and the maintenance of standing armies, to name just a few. Most of these 14 goods are known as public goods , since they are difficult for individuals to produce on their own (Olson 1971; Ostrom 1990). Thus, the state, in most instances, assumes responsibility for the production of these goods, not least of which is security. The Importance of Security Provision Section I, with its introduction of performance legitimacy, suggests that legitimacy perceptions stem, in part, from the ability of the state to provide a range of substantive goods. Political theorists and empirical political scientists alike have suggested that the provision of security is chief among these goods and, indeed, comprise the basic raison d’être of states. (Hobbes 1998; Locke 1966; Rotberg 2003). Social contract theorists have helped to link the importance of security provision to state formation by imagining a hypothetical situation called !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 14 Public goods are goods that are non-subtractable and non-excludable. Non-subtractable simply means that the availability of goods will not decrease when used by one person or set of persons. The term non-excludable means that it is difficult or impossible to restrict access to and use of the good, either because of the inherent characteristic of the good or because the costs of doing so would be too high. Because these goods are difficult to provide through the market and are seen to benefit large sections of the population, these goods are often provided by the state. ! 30! the “state of nature”. In this pre-governmental state, laws of nature prevail; however, there exists no entity to ensure that men will not violate these natural laws, and in turn, each other. In an effort to insure themselves against the uncertainties of life, individuals form a government, and thereby ensure the provision of a collective good that they could not provide through their own individual efforts. As Locke aptly states, government is seen as the most efficient means by which to “preserve men in this world from the fraud and violence of one another” (Locke 1966, x). Social contract theorists agree that this hypothetical state would lead rational individuals to the logical conclusion that a state is in fact necessary; so necessary, that people are willing to give up rights and a measure of freedom in order to achieve a government that is powerful enough to ensure security for all. The state’s primary obligation in this contract is to develop an extensive security apparatus to protect its borders and citizens from internal and external threats. Based on this view, the state-society relationship is contingent on how well the state can provide physical security. Its failure to do so may render the citizenry’s allegiance to it void. The suggestion is that where the state fails to protect citizens, the legitimacy with which it has been vested may be undermined. Weber’s classic definition of the state further emphasizes the role of security provision. Weber’s focus on the state’s means of coercion largely emphasizes elites’ interest in using the state to maintain dominance. From this perspective, state officials spend extensive amounts of time, energy, and money, building strong armies that can protect their borders and, in turn, preserve their sovereignty. The focus here has therefore primarily been on understanding how elites’ monopoly of force and coercion allows them to retain power. However, the emphasis on the state’s monopoly over force can serve a second key, albeit less discussed, purpose. Specifically, ! 31! the state’s efforts to achieve and maintain a monopoly over the use of force may be interpreted as attempts to credibly commit to providing security. If, as suggested above, states were formed in large part to provide security for citizens, elites’ survival in office should be enhanced when they can signal to members of society that they are able to provide this good. One such way of doing so is by maintaining a monopoly over force. Beyond political philosophy, political scientists have emphasized the importance of security to the viability of the state. (Bratton and Chang 2006; Carothers 1998; Huntington 1968; Linz and Stepan 1996; Rose and Shin 1999; Rotberg 2003). Rotberg (2003), for instance, has gone as far as to say that states become weak and ultimately fail when they cannot adequately provide this chief political good. Moreover, some scholars have emphasized the foundational role that 15 security and political order play in enabling the effective delivery of other goods and services and further political development (Huntington 1968; Weingast 1997). Others, while focusing on factors that are crucial to democracy and democratic consolidation, have suggested that a strong rule of law, and the provision of political order, is key to successful democratization and consolidation of democracy (Carothers 1998; Linz and Stepan 1996; Rose and Shin 1999). Finally, Bratton and Chang (2006) make a key contribution on the importance of security from the vantage point of ordinary citizens, showing that of several key state capacities, law enforcement capacity is paramount to citizen’s understanding of effective statehood. Above I have highlighted some of the key responsibilities of the state. This discussion has centered on the important role of the state in the provision of public goods, most notably the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 15 Huntington (1968) describes the absence of political order or disorder as evidenced by “ violence, coups, insurrections and other forms of instability” (vii). ! 32! provision of security. Bearing these functions in mind, I next turn to explore the state in the context that this dissertation focuses on, the state in Africa. The State in Africa Most of the literature on the African state focuses to varying degrees on the issue of state building and more frequently, on the issue of state weakness and failure. One of the foremost theorists on the state in Africa, Jeffrey Herbst, suggests that “the central problem confronting leaders of almost all African states is how to broadcast power over sparsely settled lands” (Herbst 2000, 3). In the Weberian tradition, Herbst suggests that in order to be viable, states must control the populations within their borders and gain the loyalty of their citizens (Herbst 2000). He argues that this has been challenging in the African context because of the high costs associated with expanding the domestic infrastructure, the nature of national boundaries and the design of state systems. Thus, in African states, central authorities have often failed to extend their authority to the outermost regions, specifically rural ones. In these instances, competing sources of authority take root and we may find citizens vesting their loyalty in traditional and non-state sources of authority, for it is these sources that most directly meet their needs and, in turn, command their obedience. Jackson and Rosberg (1982) too focus on the fragility of the African state, suggesting that statehood in Africa has largely been juridical, with the international community preserving African statehood on legal grounds in spite of their lack of capacity in critical areas of governance (Jackson and Rosberg 1982). International involvement in African statehood is ! 33! echoed in Pierre Englebert’s study of state legitimacy in Africa (Englebert 2002). Englebert develops a theory to explain variation in levels of state developmental capacities across Africa. He argues that variation in state legitimacy accounts for variation in the developmental capacities of African states. Englebert (1982) defines legitimacy as “a structural variable that is determined by history: a state is deemed legitimate when it has evolved endogenously to local social relations of power and authority or when, having originally been imported, it is then absorbed by such pre-existing endogenous institutions” (72) . Englebert suggests that Africa’s state weakness is a product of its history and that state weakness is largely a function of its imported nature and form, one that does not resonate with African citizens. Studies of the state in Africa thus suggest that African state survival has largely been divorced from its performance and that the state in Africa has been perpetuated through international actors rather than through domestic ones. In other words, states in Africa remain weak, in part, because they remain disconnected from their citizenries. This idea is captured in Hyden’s metaphor of the state as a “balloon suspended in mid air”, whereby he suggests that African states fail to insert themselves into the day to day productive activities of a society (Hydén 2006). One of the important by-products of the weak state in Africa is that non-state actors frequently fill the gap in state service delivery. Donor agencies in particular have recognized the steady if 16 not increasing importance of “non-state actors or providers ” in the provision of goods and services in the developing world, including Africa (Batley and Mcloughlin 2009; Moran and !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 16 Moran and Batley (2004) define non-state providers as all providers existing outside the public sector, whether they operate for profit or for philanthropic purposes. ! 34! Batley 2004; Pavanello and Darcy 2008; Rose 2007; Sacks 2012). For example, Batley and Mcloughlin (2009) have suggested that “in Nigeria and Malawi, Christian medical missions provide around 60 percent and 37 percent of healthcare services respectively, and in addition there is a myriad of for-profit providers (15). Thus, non-state actors are key to ensuring that the basic needs of ordinary citizens in Africa and the developing world more generally are met. Donors recognize that there is great variation in the quality of services offered by non-state actors, but that in most instances, non-state provision improves access to goods and services, especially for vulnerable populations such as the poor. However, what is interesting is that most donors see non-state provision as a temporary solution, rather than an enduring service-delivery arrangement (Pavanello and Darcy 2008). In essence, non-state provision is seen as a way to fill the gap in service delivery until the state is strong enough to take primary responsibility for this function. Donors have pointed to the legitimacy problems that may arise when non-state actors become chiefly responsible for public good provision. I suggest, however, that whether non-state provision undermines legitimacy is a question to be empirically explored, as I do in Chapter 5. Because non-state provision is so prevalent in African life, it is worth investigating the political consequences of this phenomenon. Moreover, given the extent to which citizens have relied on non-state actors for key goods and services, it is not a given that this service delivery arrangement will pass away, even when the state becomes stronger. The possibility of non-state provision persisting as a permanent route of African service delivery lends even more importance to the study of this question. ! 35! Most donor studies of non-state provision have focused on sectors such as education and health, goods that can more easily be classified as private goods. I explore the effect of non-state provision via the lens of security, a good that has been seen as a chief public good. Since 17 arguments have been made that non-state provision of public goods can undercut legitimacy , the security sector is a suitable place to start to test this proposition and determine whether and how non-state actors and their role in public good provision may strengthen or undercut individual perceptions of state legitimacy. As a result, this dissertation will contribute, in part, to understanding how non-state security provision might attenuate the link between weak state capacity and state legitimacy. TOWARD NON-STATE SECURITY PROVISION Most social contract theorists and empirical researchers of public good provision focus narrowly on security that is provided by the state and the various ways in which state-provided security benefits the state. In essence, these works imply that the legitimacy of the state will be determined by whether, and how well, it can provide security. However, these works fail to take account of an empirical reality in which a range of actors from the private realm participate. This is particularly true in Africa where the state’s capacity to provide key public goods, including security, is weak. Therefore, for this dissertation, I am interested in non-state security and the impact the non-state provision of security has on popular perceptions of state legitimacy. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 17 See Sachs 2012 “Can Donors and Non-State Actors Undermine Citizens’ Legitimating Beliefs?”! ! 36! Drawing from criminologists Johnston and Shearing (2003) I define security as “personal, physical safety, as well as the safety of … [individuals’] belongings from damage or depredation” (p.1). This notion of security is strictly concerned with the protection of an individual’s person and property. It does not deal with other basic human security needs such as food, water and sanitation. Non-state security then, is simply the assurance of personal physical safety and safety of property that is provided by private actors or actors outside the public sector. When I speak of non-state actors, I include entities as diverse as corporations, voluntary associations, and even families. All of the non-state actors that I examine in this dissertation play some role in security provision. The specific types of non-state actors that I explore will be discussed more fully in the following chapter. Numerous studies have documented the extensive role that actors other than the state now play in the provision of security (Baker 2008; Bayley and Shearing 1996; Johnston and Shearing 2003; Kempa et al. 1999; Shearing 1992; Wood et al. 2006). According to these works, non-state security arose in response to many factors, including but not limited to rising crime rates and fear of crime, the emergence of private property, and the perception that police are unable to manage crime on their own. Whatever the cause, there is widespread reliance in the developed and developing world on a range of non-state security structures including NGOs, commercial security companies, and individual volunteers from local communities. In effect, policing and security provision is no longer the strict prerogative of the state; instead, policing has become pluralized to the extent that in many countries, private security guards outnumber the public police force (Kempa et al. 1999; Bayley and Shearing 1996). ! 37! Within the field of criminology, a theory of “nodal governance” has been developed to account for the various and diverse range of actors involved in security provision. Burris et al. (2005) explain that these “nodes” may take a variety of forms, “from legislatures and government agencies through neighborhood associations and other NGOs to firms and gangs” (12). The theory of nodal governance, suggests that non-state actors not only participate in the provision of security, but they also help to set the tone and direction of security policy (Burris et al. 2005). From this perspective, non-state actors are seen as extraordinarily powerful players in the security game. Theorists of nodal governance still see that state as an important player in the governance of security, but only one among many. Nodal governance theory dovetails nicely with literature from political science on the “regulatory state” in that they are both able to account for the role of non-state actors in delivering security. Where they differ is in the relative power that these theories accord to the state versus non-state actors. In the regulatory-state account, it is assumed that the state is still largely responsible for commanding and controlling; that the state is primarily setting the tone and direction of security governance, even if government officials allow other actors to participate in the provision of this good. However, nodal governance does not assume that the state is always or even often in control of non-state entities involved in delivering security. In this sense, nodal governance takes us a bit further than the regulatory state approach by allowing for an examination of not only formal and legal non-state actors, but also informal and illegal ones that operate beyond the long arm of the law. This literature therefore allows for the possibility that non-state actors may come to participate in the provision of security by being actively encouraged to do so by the state, or by more autonomous means. The purpose of this dissertation is to account for the various types ! 38! of actors that deliver security to citizens, and determine how their role in security provision impacts citizens’ attitudes toward the state and shapes their political behavior. EMPIRICAL WORKS ON LEGITIMACY The theoretical basis for the relationship between state security provision and state legitimacy is well established within existing political science literature. But what about empirical examinations of this relationship? Surprisingly, there has been virtually no empirical work that investigates the relationship between security provision and state legitimacy. Of the empirical studies that posit state legitimacy as the object of explanation, scholars have examined the effect of rights, governance and welfare gains (Gilley 2006); institutional trust(Fernandez and Kuenzi 2008; Levi et al. 2009; Peltier 2007); and procedural fairness (Levi et al. 2009; Tyler and Fagan 2010) . Gilley (2006) conducts the most comprehensive study of state legitimacy to date, exploring the determinants of state legitimacy across seventy-two countries. His sample includes countries spanning Western and Eastern Europe, Asia, Latin America, Africa and the Middle East. Using survey and expert data, Gilley finds that good governance, democratic rights, and welfare gains 18 most strongly contribute to state legitimacy (Gilley 2006). Gilley notes that his findings are !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 18 Gilley measures state legitimacy using nine indicators from the World Values Survey, GlobalBarometer regional surveys, the World Handbook of Political and Social Indicators IV, International Institute for Democracy and Electoral Assistance and the International Monetary Fund. These include evaluations of state respect for individual human rights, confidence in police, confidence in civil service, satisfaction with democratic development, evaluation of current political system, satisfaction with operation of democracy, use of violence in civil protest, voter turnout and quasi-voluntary taxes. ! 39! most important because they demonstrate that “politics and politically mediated social and economic outcomes seem to matter most to legitimacy” (2006, 58). Gilley’s work makes several important empirical contributions, and he is to be applauded for his attempt to highlight the universal determinants of legitimacy. But Gilley’s methodology is questionable, as he relies on a system of ranked bivariate correlations from which to formulate his theoretical and empirical model. Moreover, Gilley’s measure of governance includes indicators on the rule of law and the control of corruption. These two variables are certainly important to any consideration of legitimacy. However, most studies on governance see the provision of security and public order as key to an assessment of governance, and that consideration is missing from this analysis. More recently, other scholars have investigated the sources of state legitimacy within specific regional contexts. Here I restrict my focus to the empirical works that speak to sub-Saharan Africa (Bratton and Mattes 2001; Fernandez and Kuenzi 2008; Levi et al. 2009; Peltier 2007). Peltier and Levi et al. explore the impact that institutional trust has on spurring legitimating attitudes. Peltier (2007) examines the sources of legitimacy in sub-Saharan Africa using survey data on 18 African countries. His model tests the effect of social structure, ethnicity, institutional performance, performance evaluations and trust on attitudes toward the legitimacy of the state. Peltier finds that “the only significant indicator across each country was an individual’s trust in institutions” (2007, 107). Peltier’s work is a significant contribution to our understanding of state legitimacy across sub-Saharan Africa. But some rightfully note issues with examining the ! 40! relationship between trust and legitimacy, issues including endogeneity and the fact that trust itself is often used as a proxy for legitimacy (Booth and Seligson 2009). Levi et al. (2009) corroborate Peltier’s findings on the significance of trust in explaining legitimacy, but they go one step further by introducing the importance of procedural justice. Their conceptual model posits that three factors feed into the perceived trustworthiness of government: government performance, leadership motivations and administrative competence. They hypothesize that those who view the government as trustworthy and those who feel the government exercises authority through fair procedures will be more likely to view the government as legitimate. Using survey data on 18 Africans countries, the authors find support for their argument that trust in government and procedural fairness shape legitimacy perceptions. While I would agree that government performance, leadership motivations and administrative competence may have an impact upon trust perceptions, the Afrobarometer survey provides a more direct way of measuring trust, namely by providing questions that ask individuals whether or not they trust the police and court officials. In my estimation, government performance and trust should be considered as separate determinants of legitimacy, and tested as such to determine which one better accounts for individuals’ willingness to defer to state authority. Not only would keeping these variables separate allow us to see the distinct impact of these predictors, but the measures I propose specifically ask people about their levels of trust in the agents who work within the state institutions that comprise Levi’s et al object of explanation. Thus, a more fine-grained and relevant measure of trust is in order. ! 41! Notwithstanding the issues with indicators of trust, however, Levi et al. make an important contribution by looking at how procedural fairness affects support for the state. They employ survey questions that ask respondents whether they feel people are treated equally under the law and whether they feel members of their ethnic group are treated fairly. They find that those who perceive the government to enact procedures in a fair and impartial manner are much more likely to voluntarily defer to state authority. This finding corroborates findings by Tyler and Fagan (2010) and Gibson (2008) that emphasize the primacy of procedural justice in spurring citizen support of state institutions and policies. Many other scholars have also spoken to how the state’s provision of economic and political goods helps to legitimate states (Schaar 1981; Schatzberg 2001). Both Schaar (1981) and Schatzberg (2001) examine how the distribution of economic goods helps to build instrumental legitimacy. Schaar (1981) makes the case that modern states largely gain instrumental legitimacy because of their role as “provider and guarantor of increase” (25). Taking a cultural approach to legitimacy, Schatzberg (2001) notes that “when political fathers care for, nurture and provide wealth for their children [i.e. citizens], their political legitimacy is enhanced” (24). Bratton et al. (2002) is one of the few empirical works to confirm the effect of economic goods provision on state legitimacy(Bratton et al. 2002). They find that find that “Malians grant legitimacy to the state to the extent that its agencies prove themselves capable of solving basic economic problems” (Bratton et al. 2002, 230). More to the issue of political goods, the literature suggests that there should be a positive link between the provision of political goods and individual views on the legitimacy of the state. ! 42! Existing theoretical work pegs the rule of law as an important determinant of state legitimacy and a foundational feature of strong states (Weingast 1997). Since Hobbes, political philosophers have asserted the importance of state security provision in legitimating the social contract between the state and society. In fact, some have gone as far as to suggest that the lack of widespread security can lead to state weakness and eventually failure (Rotberg 2003; Wood and Dupont 2006). Speaking to this issue more recently and in the context of South Africa, Marks and Goldsmith suggest that “an important source of state legitimacy is its capacity to protect its citizens from unprovoked violence and depredation; a state that will or cannot make this a core responsibility has little claim on the allegiances of the people living under it” (2006, 157). Marks and Goldsmith thus support a Weberian notion of the state and see peoples' continued support of it as contingent on its ability to provide basic protections. However, most of these propositions regarding the relationship between political good provision (i.e. security, rule of law) and state legitimacy have not been tested empirically. One exception is the recent work by Fernandez and Kuenzi (2010). Working in the African context, they investigate the relationship between crime-both perceptions of and experience with itinstitutional trust and support for democracy in Africa. While the authors do not focus on state legitimacy specifically, they do find that perceptions of crime and actual victimization both negatively affect support for democracy in Africa. This finding is interesting as it empirically shows how legitimacy (of the regime) may be undercut when individuals’ personal security is jeopardized. ! 43! In sum, extant theoretical literature ties the emergence and legitimation of states to individuals’ innate desire for protection. These works imply that state legitimacy will critically hinge on whether and how well the state is able to provide safety and security for citizens. Yet empirical explorations of this proposed relationship are lacking. Moreover, the prevailing normative aspiration of political theorists and donors seems to be in favor of security coming directly from the state. Yet, the empirical reality is that increasingly, non-state actors participate in the provision of this good. In many instances, states either directly involve other actors in the provision of security, or, at the very least, provide a context within which other actors may participate in the provision of security. Therefore, the question of the political consequences of non-state provision remains open. I explore this question empirically in Chapter 5, investigating how individuals’ reliance on non-state security shapes their perceptions of state legitimacy. ! 44! CHAPTER 3 CRIME AND SECURITY IN SOUTH AFRICA ! 45! INTRODUCTION This chapter describes the nature of the crime and security atmosphere in South Africa. It begins with a discussion of crime and security under apartheid and examines the variety of non-state security structures that were active during this tumultuous political time. After reviewing historical patterns of crime in South Africa, the chapter then reviews issues of crime and security since the transition to a democratic political regime in 1994, assessing trends and challenges using official statistics and public opinion data. Finally, the chapter explores how the state has responded to crime in the post-apartheid context, and identifies the non-state entities that participate in the provision of security in contemporary South Africa. CRIME AND SECURITY UNDER APARTHEID There are several important factors to note about crime and security under apartheid. The first refers to the acts that were considered criminal, including politically motivated behavior. The second important point concerns the distribution of security services, and the third, the extensive role that non-state actors came to play in policing. This section addresses each of these components in turn. The Politicization of Crime It is impossible to understand the nature of crime under apartheid without understanding its explicit relationship to politics (Bayart et al. 1999; Kynoch 2005; Shaw 1995, 2002). How the ! 46! apartheid state defined a crime often had little to do with the objective of maintaining a safe and secure atmosphere for all. Instead, criminality was largely politicized, and those who engaged in political acts were often labeled as terrorists. One way in which the state was able to criminalize the political, was by using the legal system to buttress its political goal of maintaining white minority rule (Schönteich and Louw 2011). Several scholars have shown how draconian laws were used to criminalize everyday activity and how opposition political parties that were seen as a threat to the ruling National Party, most notably the African National Congress (ANC), were banned. Thus, any individual who expressed discontent with or opposition to the prevailing political order was, in effect, guilty of breaking the law. In this context, the police were much less focused on conventional policing duties such as crime prevention and investigation. On the contrary, the job of the police was to specialize in political repression. Therefore, the police in South Africa were militarized and engaged in a war with the vast majority of the South African population, who were seen as the enemy. In this climate, the police’s efforts and resources were largely dedicated to monitoring the African population as a means to trying to prevent political uprisings, and quell them where they did appear. The police were also known to participate in “death squads,” which tortured and assassinated those who worked with the ANC and other opposition parties (Pauw 1991). Thus, some would argue that the state itself was criminalized, using torture and other intimidation techniques to discourage any political behavior that would contribute to the rise of a viable opposition. The politicization of crime and violence was not, however, solely the preserve of the state. The armed wing of the African National Congress (Umkhonto we Sizwe), and to a lesser extent the United Democratic Front (a coalition of politicized civic organizations) often used violence, both ! 47! directly against state agents and as a way of achieving allegiance among the masses. In fact, 19 political activists sometimes used the violent technique known as necklacing against public officials and ordinary citizens who were suspected of cooperating with the state (Buur and Jensen 2004). Violence was therefore used by the opposition as a means of securing a base of supporters in the fight against apartheid and discouraging any potential collaboration with the state. More generally, the ANC encouraged widespread mass disregard for the law in their campaign to make the state “ungovernable”. The main tactic was to withhold revenue from urban local governments (“the townships”) by engaging in tax and rent boycotts. Where these tactics failed, the UDF, acting as an agent of the ANC, was not averse to intimidating township residents and, at times, even assassinating state officials (Neocosmos 1998). The Distribution of Policing Services Aside from political violence, there were regular forms of everyday criminal activity under apartheid. As Mayekiso (1996) notes, there is bound to be criminal activity where high levels of unemployment and alienation exist(Mayekiso 1996). So, ordinary forms of crime became ways for people to subsist in poverty stricken townships. As with most other services under apartheid, the provision of policing services was highly skewed along racial lines. According to the 1998 White Paper on Safety and Security, “in 1994, 74% of the country's police stations were situated in White suburbs or business districts” ("White Paper on Safety and Security" 1998). This suggests that, on the rare occasions that the police did engage in normal duties of crime control and prevention, they did so primarily in White residential and business areas. The principal !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 19 Necklacing refers to the practice of setting an oil-filled tire about someone’s neck and then setting the tire aflame ! 48! mission of the South African Police force [SAP] was to protect and insulate White South Africans from crime, violence and political insurrection. Crime in African townships was rarely, if ever, investigated and punished (Reconciliation 2009). The fact that crime was interpreted narrowly in political terms and that police services were concentrated in minority neighborhoods, created space for the involvement of various non-state actors in the provision of security. The next section will explore the various actors that became involved in non-state policing under the apartheid regime. Non-State Security under Apartheid Market-Based Security Under apartheid, the commercial security industry thrived. In large part, the demand for commercial security was driven by the state. Since minority rule involved controlling and restricting key freedoms of the majority population, massive amounts of manpower were needed to ensure this end. The state often turned to security companies to supplement its understaffed public police force (Brogden and Shearing 1993). In this political climate, security companies were given extensive policing powers to the extent that some argued that they constituted a parallel police force. This included the powers to arrest and well as full rights of search and seizure (Irish 1999). ! 49! That the commercial security industry took on the political tone of the apartheid regime and became militaristic in its approach is by now, undeniable. Singh (2008) notes that commercial security training manuals from the 1980s made constant reference to the “the enemy”, evidence that the commercial security industry had become a key ally of the state in its efforts to repress political dissent. Not only did the state extend contracts and many policing powers to the commercial security industry, but, according to Singh, it was also able to gain indirect and direct control over the industry and ensure the industry’s allegiance to its political agenda with key pieces of legislation. Indirectly, the state politicized commercial security and the expansion of the commercial security industry with the National Key Points Act of 1980 and later an amendment to the Prevention of Illegal Squatting Act in 1988. With the former piece of legislation, once a site was designated a Key Point, or of relevance to national security interests, the state could then mandate owners of the property to hire commercial security. In this way, the commercial security industry largely became the guardian of White property interests. The amendment to the Prevention of Illegal Squatting Act also allowed for the state’s indirect control of the industry, namely by shifting the onus to citizens for evicting squatters from their property. To assist in this task, those who owned property often turned to the commercial security industry. The Security Officers Act (SOA) of 1980 allowed for more direct state control of the commercial security industry. As Singh (2008) notes, “the SOA granted the Minister of Law and Order and the South African police significant authority in the control of private security, thus providing at least the potential for the identification of industry objectives with state security interests” (45). The SOA was the first piece of legislation developed to officially regulate the ! 50! commercial security industry. In so doing, it established a regulatory board, appointments of which were made by the Minister of Law and Order. Thus the informal linkages that existed between the commercial security industry and the South African Police began to be formalized through the Security Officers Act, with the industry brought into an increasingly tight relationship with the state. Societal-Based Security Aside from the commercial security industry, non-state security initiatives were also to be found more generally within society under apartheid. Most studies agree that these community based structures reached their apex during the height of the struggle against apartheid in the 1980s (Adler and Steinberg 2000; Bundy et al. 2000; Mayekiso 1996; Schärf and Nina 2001). Scharf and Nina (2001) explore the rise of non-state initiatives like self-defence units, anti-crime committees, and people’s courts from 1984 onwards and note that the development of these structures coincided with the ANC’s goal of making the townships ungovernable by the state. In their view, the exclusion of the state from these locales, led to an increasingly important role for informal institutions in the provision of security, justice and order. As Buur and Jensen (2004) note, more often than not, the emergence of vigilante formations is premised on a deep seated mistrust of the police or perceived lack of initiative by police in providing basic human and economic security. These were the case in townships under apartheid. Mistrust of the police created space for the development of non-state township structures such as self-defence units (SDUs), street committees and people’s courts. The civics ! 51! movement, which consisted of a range of locally based organizations led by African notables such as clergymen and clerks, was responsible for developing these community-based structures (Adler and Steinberg 2000; Seekings 2000). The street committees and self-defence units served several purposes in their communities. On the one hand, they were responsible for the provision of basic goods and social services. In this sense, they played a key social welfare role. On the other hand, they were largely responsible for the provision of security and order. Because the state police, the South African Police (SAP) did not control crime in townships, street committees took on the responsibility of policing African communities and protecting them from crime. Moreover, because township residents were often the direct targets of state-led violence, these structures were also responsible for protecting their communities from official coercion. Importantly, those involved in the various organizations that were part of the civics movement often came into direct, sometimes violent, conflict with the state. In particular, members of the civics often came into conflict with Black councilors from the Black Local Authorities (BLA) who were seen as puppets of the apartheid regime. In opposition to this form of local governance, the civics boycotted local elections, burned government buildings, encouraged residents to halt the payment of rental and service charges, and sometimes injured and killed councilors (Adler and Steinberg 2000; Buur 2010). Thus, under apartheid there existed market and societal-based forms of non-state security, with the market–based version tightly aligned with the state and societal forms often standing in stark opposition to this political entity. ! 52! CRIME AND SECURITY SINCE THE 1994 TRANSITION Apartheid, a system of racial segregation and White minority rule instituted in 1948, began to unravel in 1990 as domestic and international pressure led President F.W. de Klerk to unban the opposition African National Congress (ANC) party and release its leader, Nelson Mandela from prison. It had become clear that political transition was underway and the country adopted an interim constitution and began to strike down many apartheid-era laws. Four years later in 1994, the first multi-party elections were held in South Africa. The ANC won with a landslide victory and Nelson Mandela was elected as the country’s president. Even though the country successfully transitioned to democracy in 1994 and managed to get rid of many draconian apartheid-era laws, the politicized nature of crime under apartheid would hold two very important implications for the post-apartheid security atmosphere. The first implication is that the post-apartheid state inherited weak policing capacity. Because the police were primarily concerned with the political priority of keeping the apartheid regime in power, they lacked traditional policing skills that focused on crime prevention, detection and investigation (Baker 2008). The second implication is that the new state had to contend with deeply entrenched levels of mistrust between the police and ordinary citizens. Thus, after the transition there was a heavy state focus on police transformation (Marks 2005; Shaw 2002). The concept of police transformation had two primary objectives. The first was for the police to learn basic policing skills needed to police citizens in a civil way as opposed to the authoritarian, militaristic style of policing carried out under apartheid. The second objective was for the police ! 53! to transform in a way that would enable them to earn the trust and respect of the people (Marks 2005). To this end, the name of the police force was changed from the South African Police (SAP) to the South African Police Service (SAPS) to signal to members of society that the police 20 had a new identity, one that was severed from the politics of the past . In other words, the police force, as a key institution of the new democratic order, recognized that it not only needed to build capacity, but also legitimacy. In an effort to build legitimacy, the police force not only attempted to become more service-oriented, but it also made diversification of the upper ranks of 21 the police force a key priority . In fact, South Africa adopted a national police force, rather than local forces, in part, to ensure that policing services would be distributed in a more fair and 22 equitable manner under the new order . A problematic police force was not, however, the only difficulty for creating a atmosphere conducive to liberal democracy and the rule of law in South Africa. Another issue is that !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 20 In an interview with the provincial head for Business Against Crime, a major non-profit that deals with crime in South Africa, he mentioned that the change to the new police force after the 1994 transition could be witnessed by the change in uniform. He mentioned that the police had “traded in their fatigues for their baby blues” referring to the blue uniforms that police now wear. For him, this change was not a good one as he mentioned that police under the new order had become soft and that people had “lost respect” for them. 21 Data on the exact composition of the police force are no longer made publicly available. However, SAPS does post the names and pictures of high-ranking commissioners and other police officials on their website. From these profiles, it is clear that the upper ranks of the police force has been transformed from predominately White to predominately African. There is also a greater degree of gender diversity with the current national police commissioner being a woman. 22 Under apartheid, the ruling National Party divided the country into 11 different states known as “homelands”. These homelands were developed so that each ethnic group could have its own territory. Under this system, Africans were not citizens of South Africa. They were citizens of their respective homelands. Needless to say, the provision of public goods were substandard in the homelands. Under this system, there was a separate policing agency for each homeland. The interim constitution abolished the homeland system and the new constitution signed into law in 1996 established a single, national police force for all of South Africa. ! ! 54! ordinary citizens, even if for understandable reasons, had become accustomed to rebelling against the law during the struggle for liberation (Gibson and Gouws 1997). Moreover, many South Africans developed a culture of self-reliance under the old regime, and were used to relying on non-state, community-based structures to solve disputes. Ordinary citizens thus had to learn the value and necessity of working through state structures and procedures for dispute settlement and the attainment of justice. For many citizens, the idea that one could and should turn to the state to regulate private affairs was new and would take time to embrace. These historical realities thus helped to set the tone of the security atmosphere in post-apartheid South Africa. The newly elected government inherited an environment where there was widespread criminality, weak policing capacity and high levels of mistrust in the police. All of these factors collectively created a context that made security a primary policy issue with which to be reckoned in the post-apartheid state. With the transition to democracy, many South Africans were hopeful that the challenges of the past, including crime and insecurity, would successfully be dealt with by a popularly elected government. However, levels of violence remain extraordinarily high in South Africa, although this violence is now more criminal than political in nature (Harris 2003). South Africans have 23 consistently ranked crime and security as one of their top five concerns over time . Moreover, crime remains a salient political issue because it throws into question the ability of elected officials to effectively govern. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 23 Afrobarometer, the largest cross-national public opinion survey in Africa, asks citizens what they perceive to be the most important problem facing their nation. South Africans have consistently rated crime and security as one of the tops concerns, often coming only after the issue of unemployment. ! 55! Whether crime has increased in South Africa since the transition is difficult to know, especially since crime statistics from that time included not only criminal, but political offenses (Schönteich and Louw 2011). What is clear is that the state continues to struggle to provide a sense of peace, security and order for its citizens. A 2007 report compiled by the Center for the Study of Violence and Reconciliation using data from the United Nations Office on Drugs and Crime (UNODC), showed that South Africa had the second highest murder rate of all countries 24 included in the sample . In fact, South Africa has often been referred to as the crime capital of the world(Altbeker 2005). This reputation, whether justly warranted or not, has implications for investments in the economic realm, and political behavior and attitudes in the political sphere. The first part of this section will review trends in crime based on official statistics from the South African Police Service (SAPS). The second component of this section will focus on popular evaluations of crime to see how ordinary citizens perceive the crime situation in South Africa. Crime Trends in South Africa: 1994-2012 While South Africa has often been pointed to as a country with extraordinarily high levels of crime, the overarching trend is that crime has been steadily and sharply decreasing since the 25 transition to democracy in 1994 . The South African Police Service considers 7 types of crimes as “contact crimes”. These are considered some of the most serious crimes as they bring physical !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 24 Columbia had the highest murder rate at the time. Others countries in the sample were (in descending order of murder rate) Venezuela, Brazil, Ecuador, Swaziland, Mongolia, Suriname, Lithuania, Latvia, Zimbabwe, Belarus, Kyrgystan, Ukraine, Uganda, Estonia, Moldova, Sri Lanka, Costa Rica, Georgia, Uruguay, Peru, and the United States. 25 Data on crime trends come from annual reports released by the South African Police Service (SAPS). See chapter 4 for more details on these data. ! 56! harm or death to victims. Contact crimes include murder, attempted murder, sexual offenses, common assault, assault with grievous bodily harm, aggravated robbery and common robbery. In the case of murder, the number of murders per 100,000 of the population decreased from 67 in the 1994/1995 reporting period to 31 in the 2011/2012 reporting period, a decrease of 54%. Attempted murder has also sharply declined. Attempted murder reached its peak in 2002/2003, but has since been on the decline. Murder and attempted murder, two of the most serious contact crimes have been cut in half since the transition to multiparty rule. While the improvement in these two forms of crime is to be commended, we should, however, note that the murder rate per capita still remains quite high in South Africa. Over 15,000 people were murdered in South Africa in 2011/2012. This is the same number of individuals that were murdered in the United States in 2011 even though the U.S. is 6 times as large as South Africa (Uniform Crime Reporting Program 2011). Therefore, while great improvements have been made, there is still substantial work to be done in these areas. Figure 3.1: Rates of Murder/Attempted Murder per 100,000 of the Population, 1994-2012 “For interpretation of the references to color in this and all other figures, the reader is referred to the electronic version of this dissertation.” 26 Source: South African Police Service (SAPS) !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 26 The following crime statistics presented in figures 1-7 have been compiled from various South African Police Service (SAPS) annual reports. See saps.gov.za ! 57! Turning next to instances of assault, we see that this type of crime is very high in South Africa. In 1994/1995 there were over 500 reported cases of both common assault and assault with grievous bodily harm per 100,000 of the population. Assault with grievous bodily harm peaked in 2000/2001 and common assault reached its peak two years later, but both have steadily been on the decline since. Both of these crimes have decreased by one-third in the past seventeen years. Figure 3.2: Rates of Assault per 100,000 of the Population, 1994-2012 Source: South African Police Service (SAPS) Next, we examine instances of rape in South Africa. In 1994/1995, there were 116 rapes per 100,000 individuals. By 2011/2012, this number had decreased slightly to 95, a decrease of 18 percent. It is difficult to know exactly how well official statistics represent instances of rape in South Africa, as this type of crime is grossly underreported. Moreover, there have been significant changes to the definition of rape over time. Before 2007, the definition of rape used by the South African Police Service was “unlawful and intentional sexual intercourse with a female without her consent”. Therefore, males who had been the victim of rape were not included in rape statistics prior to 2007. The problems with rape statistics notwithstanding, the overall picture shown here is that rape is slightly on the decline. ! 58! Figure 3.3: Rates of Rape per 100,000 of the Population, 1994-2012 Source: South African Police Service (SAPS) Finally we come to robbery, the category of contact crime that has experienced the least amount of improvement over time. First, it is worth noting that aggravated robbery is much more prevalent than common robbery. A great number of robberies that are committed are done so with a weapon, contributing to the large amount of victims that are injured or killed in the commission of a robbery. As of 2011/2012, a total of 200 aggravated robberies per 100,000 of the population were reported. Only slightly more aggravated robberies were reported in 1994/1995, a total of 220 to be exact. The occurrence of aggravated robbery has decreased over time, but only by 9% since 1994/1995. Common robbery, by contrast, actually increased during this period from 85 per 100, 000 of the population to 105 per 100,000 of the population. It may 27 therefore come as no surprise that robbery is one of the crimes that citizens fear the most , given the difficulty of policing this kind of crime and reducing its occurrence. Aggravated robbery is most common in the province of Gauteng (where the crime-ridden city of !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 27 Results from the 2012 Victims of Crime Survey show that 50% of respondents fear home robbery the most and 40% fear street robbery the most. The fear or robbery comes only second to the fear of burglary, with 57% of respondents reporting that they fear this crime the most. Note that individuals were allowed to choose more than one crime that they feared the most. ! 59! Johannesburg is located), with 35% of all aggravated robberies occurring there in 2011/2012. This province also has the highest rate of robbery per 100,000 of the population of all nine provinces (312/100,000). Figure 3.4: Rates of Robbery per 100,000 of the Population, 1994-2012 Source: South African Police Service (SAPS) Next, we turn to property crime. Here I examine the trend for theft of motor vehicles and burglary over time. As with contact crime, theft of motor vehicles has been drastically reduced over time. While motor vehicle theft reached 274 per 100,000 of the population in 1994/1995, by 2011/2012 this figure had been cut down to 117, a 57% decrease. This may be owed in large part to the increasing adoption of sophisticated technology. For instance, the crime-focused nonprofit organization called Business against Crime (BAC) was instrumental in bringing about the use of MicroDot technology that allows for the tracking and identification of stolen cars. At the more extreme end of the spectrum, it was legal at one point to purchase flame throwers for cars! In the event of a carjacking, the driver could flip a switch to literally release fire from the side of his/her car to ward off carjackers ("Flamethrower Now an Option on South African Cars" 1998). The flamethrower as well as other, much less extreme and dangerous technological developments such as MicroDot seem to have gone far in reducing motor vehicle theft. ! 60! Figure 3.5: Rates of Motor Vehicle Theft per 100,000 of the Population, 1994-2012 Source: South African Police Service (SAPS) Likewise burglary (residential and commercial combined) has decreased substantially, from 826 per 100,000 of the population in 1994/1995 to 624 as of 2011/2012. This represents a percent decrease of 26%. However, it seems that most of the decrease in burglary has largely been driven by a decrease in residential burglary. Prior to 2003, data made available on burglary were only the combined figure for both commercial and residential burglaries. From 2003 forward, however, separate burglary figures were made available for commercial and residential burglaries. This is helpful because first we are able to see that residential burglary is much more prevalent than commercial burglary. While individuals and businesses alike heavily rely on alarm systems and armed guards for protection, it seems that businesses are still better positioned to insulate themselves from crime. Second, while residential burglary decreased by 25% between 2003/2004 and 2011/2012, levels of commercial burglary has remained virtually the same. When we break these figures down by province, we see that a greater percentage of all burglaries occur in Gauteng, but that a higher number of burglaries per 100,000 of the population occur in the ! 61! Western Cape. Much of this is probably driven by the higher levels of property crime committed in the wealthy areas of Cape Town. Figure 3.6: Rates of Commercial and Residential Burglary per 100,000 of the Population, 1994-2012 Source: South African Police Service (SAPS) Figure 3.7: Rates of Commercial and Residential Burglary per 100,000 of the Population, 2003-2012 Source: South African Police Service (SAPS) Popular Perceptions of Crime When assessing the crime situation in any country, it is important to examine public perceptions of crime. Official reports on crime and crime statistics may have an impact on individuals’ perceptions of crime, but viewpoints on crime are rarely simply a function of actual levels of ! 62! crime. In fact, there are often wide gaps between the actual prevalence of crime and individual perceptions of the pervasiveness and severity of crime. On the one hand, many crimes that occur are never reported to the police; on the other, sensational press reporting may inflate popular perceptions of the prevalence of crime. Thus, reviewing micro-level data on individual perceptions of crime may complement or provide an alternative view to state-produced statistics. 28 I rely on Victims of Crime Survey (VOCS) data to depict citizens’ perceptions of crime in South Africa. To date, five national VOCS have been conducted in South Africa. The first was carried out in 1998 by Statistics South Africa. The Institute for Security Studies (ISS) was responsible for conducting the 2003 and 2007 versions of the VOCS. In 2011 and 2012, Statistics South Africa once again resumed responsibility for administering the VOCS and they will continue to do so on an annual basis from this point forward. . The VOCS is a nationally representative household-based survey conducted with South African residents that are 16 years of age or older. It measures citizen’s perceptions of and reported experiences with crime. It also gauges their attitudes toward key criminal justice institutions such as the police and courts. The sample size was 31,007 dwelling units for the 2012 VOCS, and 29,754 dwelling units in 2011. The institute for security studies carried out a total of 4,500 interviews in 2007 and 4,860 interviews in 2003. The total sample size in 1998 was 4,000. Although the last VOCS was conducted in 2012, I was not able to obtain access to this raw data. In fact, I was only able to gain access to the raw survey data collected in 2011. Therefore, I have !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 28 See chapter four for a full description of these data. ! 63! 29 primarily relied on reports and presentations on the data for my over-time figures .The crosstabulated data that I depict below will be from the survey that was administered in 2011. Rate of Crime/Personal Security To begin, we can explore whether citizens’ views on the crime rate are consistent with the trend captured in official crime statistics. In 2012 when respondents were asked whether they thought crime had increased, decreased or remained the same in the last three years, one-third of respondents reported that they felt crime had “increased”, while almost thirty percent felt that it had “remained the same”. As of 2012, a clear majority of respondents felt that crime had actually not improved or become worse over time. This stands in stark contrast to official statistics, which show a dramatic decline in most categories of crime. Figure 3.8: Perceptions of Crime in South Africa, 2012 Question: How do you think the level of property/violent crime in your area has changed in the last three years? Note: The 2012 survey asked about property crime and violent crime separately. The figures shown are the average of these two indicators for each category. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 29 Many times these reports did not include full descriptive information such as the N for the particular indicator that was being reported. ! ! 64! However, the percentage of individuals who report that crime has “increased” has dropped over time. In 2003, a majority of individuals (53%) reported that crime was one the rise. Nine years later, this percentage has dropped by twenty percentage points. Therefore, even though one-third of respondents felt that the crime situation was getting worse in South Africa as of 2012, more people now than at any other time that this survey was run feel that crime is improving. Figure 3.9: Perceptions of Crime in South Africa, 2003-2012 Question (2011-2012): How do you think the level of property/violent crime in your area has changed in the last three years? Question (2003-2007): Do you think that crime has increased, decreased, or stayed the same in your area in the last four years? Note: Figures shown are the percentage of individuals who say that crime has “increased”. Note: The 2011-2012 surveys asked about property crime and violent crime separately. The figures shown are the average of these two indicators for each category. Please note that what the VOCS refers to as “violent” crime is referred to as “contact” crime in other parts of the dissertation. When examining perceptions of crime by race, we see that Asians and Coloureds are slightly more likely than Africans and Whites to feel as if violent crime has increased, and that they are significantly more likely than Africans to feel the property crime is on the rise. In the next section we will explore whether the gap between Africans and minorities in their views on property crime may have to do with minorities’ greater experiences of this type of crime. ! 65! Figure 3.10: Perceptions of Violent Crime in South Africa by Race, 2011 Question: How do you think the level of violent crime in your area has changed in the last three years? Note: Figures shown are the percentage of individuals who say that crime has “increased”. Figure 3.11: Perceptions of Property Crime in South Africa by Race, 2011 Question: How do you think the level of property crime in your area has changed in the last three years? Note: Figures shown are the percentage of individuals who say that crime has “increased”. ! 66! When investigating perceptions of crime by province, it is somewhat surprising that more citizens in Gauteng than in any other province feel as if violent crime is on the decline (51%). There is a similar trend for property crime. During informal conversations with Durban residents, many people said that crime was being displaced from Johannesburg to Durban. There may be an element of truth to this, especially since a majority of Johannesburg respondents feel that crime has decreased. These perceptions may be influenced, in part, by the leadership of the Lieutenant General Petros, the fairly new police commissioner appointed to Gauteng in early 2011. It is widely perceived that Petros has been highly effective in cleaning up crime in Gauteng. One interviewee credited him for pushing important initiatives such as the one that requires that a 30 corruption hotline number be painted on police cars . Table 3.1: Perceptions of Violent Crime in South Africa by Province, 2011 W. E. N. Free KwaZulu N. Gau- MpumaCape Cape Cape State -Natal West teng langa Limpopo Increased 37% 38% 34% 42% 31% 36% 22% 29% 38% Decreased 29% 36% 26% 34% 45% 41% 51% 49% 33% Same 34% 25% 40% 23% 24% 23% 25% 21% 28% Unspecified % % % % % % 1% 2% 1% Question: How do you think the level of violent crime in your area has changed in the last three years? Next, I explored citizens’ sense of personal security by using indicators that asked how safe they feel walking around their area during the day and at night. As of 2012, almost 90% of respondents reported that they felt safe walking around during the day. However, after dark this figures drops substantially to only slightly more than one-third of individuals who feel safe walking around at night (37%). As of 2012, not only do one-third of people feel that crime is getting worse over time, but most people do not feel secure after dark. I witnessed this sense of !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 30 Chairperson of Community Police Forums for Gauteng. Personal interview. ! 67! insecurity during my fieldwork in South Africa, as people wrestled to secure a seat on the minibus taxi 31 after work to make sure they arrived to their homes in the township before dark. Figure 3.12: Perceptions of Personal Security, 1998-2012 Question(2011-2012): If you had to walk, how safe would you feel walking alone in your area during the day/when it is dark? Question(1998-2007): How safe do you feel walking alone in your area during the day/after dark? Note: Figures shown are the percentage of individuals who say that they feel “very safe” or “fairly safe” !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 31 The minibus taxi is a popular mode of transportation in South Africa. Each taxi carries approximately sixteen passengers. The cost of transportation is relatively inexpensive (generally no more than 5 Rand (less than one U.S. Dollar) and therefore comprises the mode of transportation most often used by poor people. ! ! 68! Figure 3.13: Perceptions of Personal Security by Race, 2011 Question: If you had to walk, how safe would you feel walking alone in your area when it is dark? Note: Figures shown are the percentage of individuals who say that they feel “very safe” or “fairly safe” While substantial percentages of all South Africans do not feel secure outdoors after dark, this sense of personal insecurity is most pronounced for Asians. A little over one-third of Whites, Africans and Coloureds feel safe outdoors at night, but for Asians this figure drops to roughly one-quarter. Moreover, in roughly all provinces, at least half of individuals feel unsafe leaving the house at night. Feelings of insecurity are most pronounced in Free State province where a full 80% of respondents report that they feel “very unsafe” or “a bit unsafe” walking alone at night. Table 3.2: Perceptions of Personal Security by Province, 2011 W. E. N. Free KwaZulu N. Gau- Mpumal LimCape Cape Cape State -Natal West teng -anga popo Very/fairly safe 37% 33% 52% 18% 46% 34% 33% 28% 48% A bit/very unsafe 62% 67% 48% 81% 53% 65% 65% 67% 51% Unspecified 1% 0% 1% 1% 1% 1% 2% 4% 1% Question: If you had to walk, how safe would you feel walking alone in your area when it is dark? ! 69! Victimization Of all the crimes explored here, individuals most frequently report being a victim of burglary, 32 followed by robbery . These findings are consistent with official statistics that show burglary as the most prevalent crime in South Africa. In 2003, 8% of respondents reported experiencing a burglary. As of 2012 this figure has dropped to 5%, but burglary is still, by far, the most commonly reported form of victimization. It may come as no surprise then that a majority of respondents (57%) say that burglary is the most feared crime followed by robbery (50%) 33 (Victims of Crime Survey 2012). Data from figure 14 are also in line with the official statistics in that they generally depict a downward trend in experiences of crime. Figure 3.14: Victimization in South Africa, 1998-2012 Base Question: Having asked in general about your perceptions of crime, I would like to ask you about your experiences of crime over the past five years, and in particular, within the past twelve months. I am going to read out a list of crimes, and I would like you to tell me if you or any member of your household have been a victim of any of these crimes in the past five years, and then in the past twelve months: Question: Housebreaking/burglary (when someone was at home)? !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 32 Because of the various ways in which robbery has been defined over time, VOCS reports only release the robbery figures for 2011 and 2012, since these are the only two years for which the data are comparable. 33 Respondents were allowed to give more than one response to the question of which crimes were feared the most.! ! 70! Figure 3.14 (cont’d) 34 Question: Robbery (excluding home robbery and car/truck hijackings) ? Question: Assault? Question: Theft out of motor vehicle? Question: Theft of car? Question: Sexual offence (including rape)? Question: Murder? Note: Figures shown are the percentage of individuals who report that they have experienced these crimes within the past12 months If we explore victimization rates by race, the picture we come away with is that Whites and 35 Asians are more likely to be victims of crime, at least crimes that are property-related . This may help to explain why, in part, Asians are more likely to report that property crime is increasing. If we were to examine violent crimes in more detail, however, a different picture 36 might arise as to who bears the brunt of victimization . But only one form of violent crime is captured here, and we see that less than one percent of all racial groups report experiencing the murder of a relative. Table 3.3: Victimization in South Africa by Race, 2011 Afric an Coloured Asian White Burglary 11% 11% 13% 18% Car Theft 1% 3% 8% 10% Theft from Car 2% 6% 10% 11% !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 34 For robbery, assault and murder, the base question reads as follows: “Having asked about household crime, I would like to ask you about your personal experiences (underline in the original question) of crime over the past five years, and in particular, within the past twelve months. I am going to read out a list of crimes, and I would like you to tell me if you have been a victim of any of these crimes in the past five years, and then in the past twelve months:” 35 Because of data restrictions I was only able explore the racial breakdown of the four crimes shown in table 3. Indicators of other forms of victimization were not made available to me in the dataset I was given. ! 36 The conventional view is that Africans are more likely to be victims of contact, violent crime. ! 71! Table 3.3 (cont’d) Murder .6% .3% .0% .3% Note: Figures shown are the percentage of individuals who report that they have experienced these crimes within the past 5 years As with official statistics, it is also possible that individuals may underreport experiences of victimization when participating in surveys. While it is impossible to know for sure whether individuals underrport in surveys, it seems clear that individuals underreport crime to the police. Respondents for the 2011 South Africa Afrobarometer survey were asked what the main reason is for citizen’s failure to report crimes. A plurality of respondents (34%) feel that people fail to report crimes because of police apathy. This suggests that the police have quite a bit of work to do in convincing citizens that they are actually concerned about their safety and security. Figure 3.15: Primary Reasons for Not Reporting Crime to the Police, 2011 Question: Some people say that many crimes are never reported to the police. Based on your experience, what do you think is the main reason that many people do not report crimes like thefts or attacks to the police when they occur?. Source: South Africa Afrobarometer Round 5 ! 72! Evaluations of the Police and Courts It is clear from the figure 15 that many South Africans feel that the police do not care about their security needs. But what other views do South Africans hold of those officials responsible for providing security and justice? As of 2012, a clear majority (62%) of individuals feel that the police are performing well in their area. In fact, evaluations of police performance have become more positive over time, rising from 52% in 2003 to over 60% in 2012. A strong majority of individuals are also satisfied with the performance of the courts (64% as of 2012). Figure 3.16: Evaluations of Police Performance, 2003-2012 Question(2011-2012): Are you satisfied with the police in your area? Question(2003-2007): How do you think the police are doing in your area? Note: Figures shown are the percentage of individuals who report that the police are doing a “good job” in 2003/2007 and those who report “yes” in 2011/2012 ! 73! Figure 3.17: Evaluations of Courts, 2007-2012 Question: Are you satisfied with the way courts generally deal with perpetrators of crime? Note: Figures shown are the percentage of individuals who respond “yes” to this question Regionally, we see that individuals from the Western Cape are the most dissatisfied with the way that courts deal with criminals. In every other province, at least 60% of individuals are satisfied with the performance of the courts. Figure 3.18: Evaluations of Courts by Province, 2011 Question: Are you satisfied with the way courts generally deal with perpetrators of crime? Note: Figures shown are the percentage of individuals who respond “yes” to this question ! 74! Dissatisfaction with the police is most pronounced in Mpumalanga, but it is noteworthy that in all provinces a majority of citizens are satisfied with the way the police are performing as of 2011. Figure 3.19: Evaluations of Police Performance by Province, 2011 Question: Are you satisfied with the police in your area? Note: Figures shown are the percentage of individuals who report “yes” Finally, when we examine perceptions of the police and courts by race, we see that Whites are the most satisfied with the police, but the least satisfied with the courts. For every racial group, however, a strong majority of individuals report being satisfied with the police. This is also the case for the courts with the exception of Whites. ! 75! Figure 3.20: Evaluations of Police Performance by Race, 2011 Question: Are you satisfied with the police in your area? Note: Figures shown are the percentage of individuals who report “yes” Figure 3.21: Evaluations of Courts by Race, 2011 Question: Are you satisfied with the way courts generally deal with perpetrators of crime? Note: Figures shown are the percentage of individuals who respond “yes” to this question ! 76! STATE RESPONSES TO CRIME IN POST-APARTHEID SOUTH AFRICA One of the first policy documents on crime introduced by the Post-Apartheid state was the National Crime Prevention Strategy (NCPS) of 1996. The strategy included a four-pillar approach to crime prevention. The first focused on revamping the criminal justice system to make it more efficient and, in turn, a deterrent to crime. Second, the state attempted to use the environment to its advantage, focusing on the ways in which smart environmental design could decrease opportunities for crime. Third, there were “public values and education” initiatives which attempted to change the way that people thought about and responded to crime. Fourth, the state saw the effective targeting of trans-national crime as a key component of crime prevention. Throughout the NCPS, the state heavily focused on mediating the social causes of crime. This task required key partnerships with other government departments such as health, education and welfare. The thinking was that many people who engaged in crime were themselves victims, in that they had been denied access to education and other basic goods and services under apartheid (Singh 2008). Therefore, with education and increased access to opportunities, there would be less of a need for individuals to engage in criminal behavior. The approach was a multi-agency one that encouraged coordination across a range of governmental departments in order to address crime. However partnerships were not just restricted to governmental ones. The NCPS also focused on building strong partnerships with the commercial security industry and civil society. According ! 77! to the NCPS, “there are many important partners in the fight against crime. These include, among others, organizations of civil society, particularly business and community organizations, citizens who volunteer for service as Police Reservists, as well as the commercial security industry, which performs a useful role. The role of such players is, in principle, one of partnership with the State”("National Crime Prevention Strategy" 1996). It is further noted that the role of the commercial security industry was so crucial in the fight against crime that its role and duties would be further elaborated on in legislation designed specifically for that industry. Throughout the NCPS, there is a call for greater participation by civil society and ordinary citizens in crime prevention. At one point in the document, it says that “to effectively reduce crime, it is necessary to… facilitate real community participation”("National Crime Prevention Strategy" 1996). This document thus sets the tone for the involvement of citizens and organizations in the prevention of crime in the post-apartheid state. Crime, in effect, became not 37 just the business of police, but everyone’s business . In this sense, the state accommodated and incorporated citizens into the fold of policing (Buur and Jensen 2004). Moreover, the state attempted to create a clear-cut divide in security functions, making non-state actors heavily responsible for crime prevention and the state more responsible for law enforcement. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 37 This sentiment was echoed as recently October 2012 when Commissioner Lamoer for the Western Cape said “Safety is not only a police problem, it needs efforts from all of us. Cooperation must be strengthened and we don't need to be vigilantes to clear the streets." October 19, 2012. Cape Argus (South Africa) “New station for hot spot; Second cop shop for Nyanga the W Cape's murder capital”. ! ! 78! In 1998, the Department of Safety and Security released the next major policy document to address issues of crime in South Africa. This document was the White Paper on Safety and Security. In the Foreword by the then Minister of Safety and Security, he notes that government’s initial agenda on crime was largely concerned with reforming the police and mobilizing people’s participation in safety and security. Further, the minister noted that while the “NCPS continue to frame the development of policy within the department…the emphasis has now shifted towards improved service delivery. This means that the Department’s approach continues to be underpinned by the philosophy of community policing. These have at their heart the principle that a partnership between the police and communities is essential to effective service delivery.” ("White Paper on Safety and Security" 1998) The above quote shows that the emphasis on citizen involvement in policing remained a key approach to addressing crime and violence in South Africa. But the initial state emphasis on crime prevention soon gave way to a much greater focus on tough enforcement (Dixon 2004, Shaw 2002). To be fair, the White Paper states that it advocates a “dual approach to crime” by focusing on tough enforcement and crime prevention. However, in terms of spending and budgeting priorities, it seems that the state has mostly supported the enforcement side of the equation. One obvious priority has been on increasing the visibility of the police. In 2007/2008, for example, the South African Police Service (SAPS) budget was 38, 371 million rand. By 2011/2012, it had increased to 60, 658 million rand, an increase larger than that of other criminal justice institution such as the courts and prisons combined. Not only has the police budget grown, but the number of personnel has also grown substantially. The number of police officers grew from 87, 643 in 2002/2003 to 128, 542 in 2009/2010. In just a seven-year period, the size of the police force grew by 32 percent. ! 79! The increase in the number of officers corresponds to the priorities of SAPS, which has been on visible policing. SAPS has five sectors or programs on which they spend; administration, visible policing, detective services, crime and intelligence and protection and security services which is responsible for providing protection for government officials. As of 2010/2011, a plurality of the total budget (43%) was spent on visible policing (Newman 2011). Even when exploring budgetary changes, we find that substantial growth in the visible policing budget has occurred over time, with the total percent increase from 2003/2004 to 2011/2012 coming second only to the percent increase for administration. Visible policing allows cops to quickly respond to crime when it occurs, however, one could also argue that having cops visibly on the beat also contributes to crime prevention. Unfortunately, funding for detective services, which could potentially help to build stronger cases for prosecution, has grown much slower. This is perhaps why even though the police have been more successful in making arrests, conviction rates have not increased much over time and conviction rates for serious crimes stand at less than 5 percent (Altbeker 2007). The shift to a greater focus on enforcement is, perhaps, understandable in a country that has come under fire for its widespread crime. Amid rising levels of crime and increasing fear of crime, enforcement approaches provide something that can be measured and sold to the public to convince them that the police are making progress. Whereas the efficacy of crime prevention is difficult to measure, enforcement activities (i.e. number of arrests) can be measured and can provide some indicator of police performance. This is particularly important in a context where the police force is under constant pressure to demonstrate that it is improving the security landscape. The official statistics shown above suggest that the police have been making progress ! 80! with crime on the decline in virtually every area (with the exception of robbery). Why then, does the use of non-state security remain so popular in post-apartheid South Africa. I turn to that subject in the next section. NON-STATE SECURITY IN POST-APARTHEID SOUTH AFRICA As discussed above, the new state-with its weak security capacity, its desire to legitimate itself, and pressure to reduce crime rates- incorporated citizens into the fold of policing. This created space for members of society to participate in ensuring their own safety and security. In addition, with South Africa shifting to a free market economy, there were continued opportunities for the operation of the commercial security industry. As Scharf and Nina note, “The irony is that the liberal state was supposed to reduce the need for non-state forms of ordering, but the inability of the transforming state to rise to the level and scope of service delivery has had the opposite effect” (2001, p.6). This section will discuss the role that each of these types of non-state security structures is playing under the new order. Market-Based Security Cross-national comparisons of commercial security suggest that the growth of the sector may be the result of three factors; the withdrawal of the state from some of its functions, the growth of mass private property, and the inability of the police to protect citizens from crime (Shaw 1995). In South Africa, the steady growth of the commercial security industry may be a result of all of ! 81! these factors, but chief among these is the perception that police lack sufficient capacity to protect citizens from crime. Today, the commercial security industry is a key actor in controlling crime and protecting private as well as public spaces (Brogden and Shearing 1993). One way of measuring the persistent salience of crime and security issues in South Africa is by looking at the continued growth of the commercial security industry. Figure 3.22 depicts the total number of registered active security guards in the country from 2001 when these statistics were first recorded. It shows that the number of registered active guards grew from a total of 194, 525 in 2001 to over 400,000 as of 2011. This represents a percent increase of 111 percent in 10 years! By comparison, SAPS grew from a total of 87, 643 officers in 02/03 to 156,745 by 2012, a percent increase of 79 percent. Figure 3.22: Growth of the Commercial Security Industry by Number of Security Guards, 2001-2011 Similarly, the number of security businesses in operation have increased substantially over time. Figure 3.23 shows that there were 5,491 registered active security businesses in South Africa in 2001, compared to over eight thousand in 2011, a 61 percent increase. So it is not simply that individual companies are getting larger over time, but that more competitors are entering the marketplace every year. Of the various types of commercial security companies that exist, those ! 82! that provide guarding or patrolling services-whether commercial or residential-are by far the 38 most popular with 6,324 such companies in operation as of 2011 . While many people rely on technology and alarm systems as a first line of protection, it is clear that businesses and individuals alike are very dependent on guards (many of them armed) to ensure their physical safety and the safety of their property. The steady increase in commercial security guards and commercial security companies over time shows that demand for commercial security has continued to grow over the years. This demand may largely stem from increasing levels of fear of crime, but also an increase in the number of individuals with the means to purchase marketbased security. Of the 8,828 security companies in existence in 2011, the vast majority could be found in three provinces. These are Gauteng (3,202), KwaZulu-Natal (1,502) and the Western Cape (956), all provinces with large urban centers. Figure 3.23: Growth of the Commercial Security Industry by Number of Security Businesses, 2001-2011 It is impossible to speak of the commercial security industry in South Africa without mentioning its relationship to the state, mainly via the formalized institution of the Private Security Industry !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 38 Other types of commercial security companies include, for example, those that provide bodyguarding, cash-in-transit services, security consultancy, entertainment/venue control, and car watch, just to name a few. In many cases, a business will provide several of these services and therefore may be counted more than once. ! 83! Regulatory Authority (PSIRA). This body was created in 2001 in order to conduct oversight of the commercial security industry. According to the Private Security Industry Regulation Act of 2001, key functions of the authority include: ! ! ! ! ! Granting registration to as well as suspending the registration of security service providers Developing and maintaining standards related to security provision Ensuring quality training of security service providers, primarily through accreditation of security training institutions Processing complaints concerning private security providers and Protecting private security guards from exploitation The body is governed by a council consisting of a chairperson, a vice-chairperson and three additional councilors, all of whom are appointed by the Minister of Police. The council members serve for a term of three years with the option to be reappointed for up to two additional terms. One of the key ways the PSIRA exercises control over the commercial security industry is through inspections. In 2011, a total of 6,611 inspections of commercial security providers were carried out in South Africa. Of these inspections, the vast majority were carried out in three provinces, Gauteng (2,138), KwaZulu-Natal (2,144) and the Western Cape (717). Moreover, the vast majority of these inspections were carried out on security businesses that provided guarding or patrolling services (85% in Gauteng, 84% in KwaZulu-Natal and 68% in the Western Cape). There are four different type of inspections carried out by the authority. First, the authority conducts routine inspections to ensure that service providers are in line with training and accreditation requirements. First time inspections are thorough inspections conducted for newly registered businesses. Next, the authority carries out infrastructure inspections. Finally, the authority is responsible for employing triggered inspections. Of these various types of ! 84! inspections, priority is given to infrastructure and triggered inspections. Triggered inspections are important because these are spontaneous inspections that are carried out in response to complaints about security service providers. Therefore, these inspections provide a mechanism for corporations and individuals who rely on commercial security to come into contact with the state and voice their dissatisfaction with the industry. In 2011, 577 triggered inspections were carried out in Gauteng, compared to 595 in KwaZulu-Natal and 90 in the Western Cape. Over time, the state has called on the commercial security industry to be a key partner in the fight against crime and the commercial security industry has, in turn, presented itself as a major ally of the state. For the most part, the relationship between the state and the commercial security industry has been characterized by cooperation. However, there have certainly been and continue to be times of conflict between these two entities. One of the biggest sources of conflict stems from the criminality embedded in the commercial security industry itself. In an interview with a high-ranking police official in Pretoria, he mentioned that the commercial security industry presents lots of problems for the police in South Africa, citing the fact that you often find “fly39 by-night” companies that are operating illegally and that employ illegal immigrants . Perhaps even more disturbing is the sentiment that many security guards are directly involved in crime, for example, helping to set up burglaries of the very homes they are supposed to 40 protect . Part of the reason for the inspections carried out by the PSIRA is to detect unregistered !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 39 South African Police Service (SAPS) Head Office Brigadier. Personal interview. 15March2011. 40 Member of Johannesburg Mayoral Committee on Public Safety. Personal interview. 22March 2011.! ! 85! security business and businesses in violation of other PSIRA regulations (i.e. unregistered or improperly registered firearms). Where these are detected, criminal cases are opened with the South African Police Service. As of March 2010, 648 cases were pending with the SAPS, cases that inspectors of the authority had filed. Of the 648 pending cases, 257 had been opened in the last year. In addition to criminal cases being filed, the Authority often refuses to register security providers or revokes their registration if they are found guilty of criminal offenses. Between 2010 and 2011 alone, the authority refused 11,810 individual security service provider applications due to the applicants being guilty of disqualifiable offences(Authority 2010/2011). Another 168 individual security service provider registrations were withdrawn after conviction of a criminal offense. In sum, the commercial security industry is a key player in the security atmosphere in South Africa. Its substantial growth over time shows that businesses and citizens alike (and at times even the state) see it as an necessary actor in protecting their person and property. But commercial security can be extremely costly, limiting access to the well-to-do. Therefore this service is not one to which the poor are likely to turn. To account for the full range of non-state security actors in post-apartheid South Africa, we must also examine those voluntary, community-based forms of non-state security, what I refer to throughout this dissertation as societal-based security. ! 86! Societal-Based Security Earlier in the chapter I noted that societal forms of non-state security largely arose in African communities under the banner of the “civics” before the transition to democracy. Seekings (2000) notes that most civics died off very shortly after the transition. This was, in part, because the mobilization role that they played in the struggle for liberation was no longer necessary. Yet even though the civics as a broad overarching organizational structure may have ceased to exist after the transition, some structures that operated under its umbrella (such as street committees) did not. In fact, some scholars have argued that there has been a concerted effort on the part of the state to integrate these non-state structures into state structures, via for example, organizations such as Community Police Forums (Buur 2005) Section 18 of the South African Police Service Act of 1995 stipulates a key role for community to play in security provision through the creation of Community Police Forums (CPFs). According to this act, the purpose of CPFs is to formalize a relationship between community and police, improve transparency and accountability of the police, and involve community in problem identification and problem solving regarding issues of crime. In the immediate aftermath of the 1994 transition to democracy, community policing was seen as way of improving public perceptions of police, ensuring that the police became a “people’s police”. In practice, however, some forums have often exacerbated tensions and mistrust between the police and community (Brogden 2004; Shaw 2002; Baker 2008). During my fieldwork in South Africa I was often told by police that community members are problematic because they want to act as if they are the police. Sometimes police even suggested that they questioned citizens’ motives for ! 87! joining the CPF, as it had been suspected that in some cases criminals themselves were joining in an effort to gain access to information that would be of help to them in their criminal pursuits. From the community side, I frequently heard citizens complain about the police not wanting to listen and fully take their concerns and suggestions into consideration. CPFs are established at the police station level and officially there should be one CPF for every police station. In practice, however, the establishment and activity of CPFs varies widely across regions and even across individual police stations. The CPF is supposed to consist of both police officers (usually, at least, the station commander) and community members, with a community member responsible for holding the position of chairperson. Each CPF has a chairperson, vicechairperson and however many additional members as determined by the executive committee of the CPF. Each CPF is governed by a constitution that is supposed to be drafted by members of the executive committee within 30 days of its establishment. In many instances, if the police station covers a wide geographical area, Community Police Sub-Forums will be established in an attempt to deal with crime specific to smaller areas under the police station jurisdiction. Of all the societal forms of non-state security explored here, CPFs are by far the most institutionalized. They embody the most formal relationship between citizens and agents of the state such as the police. As a result, they are perhaps seen as the most legitimate form of societalbased security. However, this is not to say that the state is always in control of or perfectly monitors the behavior of CPFs. In an interview with a CPF chairperson, I was told that the CPF often patrolled the neighborhood with no police, although police are supposed to be a part of all neighborhood patrols. Moreover, although citizens involved with the CPF are only supposed to ! 88! act to help prevent crime, they often react to crime and alleged criminals, taking on more of an enforcement role. In speaking to this same CPF chairperson I was told that his CPF is effective 41 because, off the record, “criminals will get a beating” . So CPFs do not necessarily act lawfully, even though they are most closely associated with those charged with upholding the law. What’s more is that in addition to community members of CPFs sometimes using violent and unlawful means for dealing with criminals, there is some evidence that the police are often aware of this 42 behavior and tacitly endorse it . The Good, the Bad and the Ugly of Non-State Security Until now, we have focused almost exclusively on how security is provided either by the state or through non-state actors that the state authorizes to provide security services. However, this only tells one side of the story of how state actors become involved in the provision of security. In fact, it is a top down story that assumes that the state is completely in control of the security atmosphere. In reality, there are many organizations and individuals that participate in security provision that remain beyond the reach of the state. Some more autonomous organizations include those like street committees and neighborhood watches. Street committees originally arose under apartheid. These were local community structures that largely protected African communities from state violence, but also dealt with criminal elements within their own communities since the policed rarely patrolled African areas. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 41 42 CPF chairperson. Personal interview. 30December2010. See Steinberg Thin Blue: The Unwritten Rules of Policing South Africa for a vivid account of how police are aware of, yet turn a blind eye to, community violence against alleged criminals.! ! 89! While many other locally-based structures under the umbrella of the “civics” died off after apartheid, the street committee has persisted as a viable way of dealing with crime at the local level (Kempa et al. 1999). Neighborhood watches are also locally-based groups of citizens who volunteer their time in an effort to protect their neighborhoods. As the name suggests, street committees are concerned with particular streets or blocks within the community, whereas neighborhoods watches are interested in security issues relevant to a larger community. It is not uncommon, therefore, for those members of various street committees to band together to form a neighborhood watch. Street committees and neighborhood watches often function within the ambit of the law, though their relationship to the state is not institutionalized like community police forums. Therefore, it is conceivable that street committees and neighborhood watches may use illegal and undemocratic means to achieve their ends, especially since it is more difficult for the state to monitor their activity. In fact, in the early 1990s, street committees developed a particularly strong anti-crime focus and would use violence to bring criminals into compliance. Under the new order, the challenge is to shape these committees in ways that are consistent with democracy. To assist with this task, the new democratic state employed NGOs to train street committees in practices that were consistent with the new focus on human rights and the rule of law (Singh 2008). On the other end of the non-state security spectrum, we have organizations such as gangs and vigilante groups (Jensen 2008). These groups do not have the permission of the state to operate, nor do they seek it. The most well-known vigilante group in South Africa is People Against ! 90! Gangsterism and Drugs (PAGAD) which organizes against gangsters and drug dealers in the Cape Flats area of Cape Town. Members of the organization often protest in front of drug dealers’ homes, and often severely beat and or kill those accused of gangsterism and drug dealing (Dixon and Johns 2001; Gottschalk 2011). Therefore, recognizing the plural nature of security provision in South Africa means acknowledging that security may be provided by entities and individuals that cooperate and coordinate with the state, or by entities or individuals that function autonomously of the state, and that either one of these may employ legal or extralegal means or some mix of the two for achieving their ends. Recognizing the various types of actors involved in the provision of security in this context begins to highlight the potential achievements and pitfalls of non-state security. Non-state security, in its many forms, can provide services to those who are either not completely satisfied with or unable to gain adequate access to state security services. However, non-state provision of security can also raise several problems that should be of interest to any democratic state. The first is that it may exacerbate inequality, as only the wealthy and elite members of society can purchase and thus gain access to commercial security services (Bayley and Shearing 1996). Therefore, the poor will be left to rely solely on what is considered an under-equipped, ill-trained, and apathetic police force, especially in poor areas where stocks of social capital are not sufficient enough to allow for the production of societal non-state security. The expectation is not that non-state security will be a panacea for underperforming state security. As Baker (2008) notes, the state has never provided security on a equal footing, so we should not expect non-state security to achieve this great feat. However, it is worth ! 91! highlighting the potential for non-state security to exacerbate inequality, and not only with regard to security, since the poor also become cut off from potential middle class networks that could help to improve their lives when the middle class and rich “fortress” themselves behind gated communities and armed security guards (Lemanski 2004). The second issue arises with the possibility of “mob justice” and vigilantism when communities are allowed to participate in security provision (Kempa et al. 1999). In this case, the concern is not with the uneven supply of security, but rather with instances of extralegal measures that could perhaps unjustly target individuals and lead to harsh and violent punishment by the community. Extreme forms of punishment administered by community were prevalent under apartheid in the form of “necklacing” and were even witnessed recently in spates of violence against foreigners (Lubbe 2008; Mosselson 2010). Societal participation in security provision therefore opens the door to societal administration of what they perceive to be proper forms of punishment and justice, whether legal or not. Putting aside problems with both commercial and societal forms of non-state security, what is most important to note is that citizens rely quite extensively on both forms to meet their security needs. It seems that non-state security, at least for the foreseeable future, will continue to play a critical role in the governance of security in South Africa. The question then becomes what are the political consequences of this provision. I empirically explore this question in chapters 5 and 6. ! 92! CHAPTER 4 RESEARCH DESIGN AND METHODOLOGY ! 93! INTRODUCTION This dissertation makes use of several data sources in order to describe the crime and security atmosphere in South Africa and to explain the political consequences of non-state security reliance and other security factors. I employ official police statistics and existing public opinion survey data in several chapters. In this chapter I will describe these various data sources, but I first begin with a discussion of the public opinion survey that I developed and administered during my 2010-2012 fieldwork in South Africa. Since I was responsible for the design and implementation of this mass survey, and because I rely extensively on this survey for my analysis in chapter 5, I will discuss this survey at length before moving to a discussion of other data sources that were employed throughout the dissertation. ORIGINAL SURVEY RESEARCH IN URBAN SOUTH AFRICA Why Design an Original Survey? Given that a plethora of survey data exists in South Africa, it is worth discussing why I designed and carried out an original survey in this context. This survey was primarily designed to provide data to test the effect of individual reliance on non-state security on perceived state legitimacy. Afrobarometer, the largest cross-national public opinion survey in sub-Saharan Africa, contains a battery of questions that assess individual attitudes toward state institutions and actors. At the time that I was developing my research project, however, it did not contain questions that would ! 94! 43 allow me to capture who South Africans turn to for their security needs . Moreover, the argument that I put forth about the conditional affect of non-state security on state legitimacy (depending on whether individuals see the state as a producer or arranger-see chapter 1) could not be tested with Afrobarometer data. The survey simply contained no indicator that would allow me to ascertain what citizens believe the state’s proper role should be in security provision. Therefore, while the Afrobarometer survey contained many questions that would allow me to measure the dependent variable of interest (state legitimacy) it contained no measures of my primary independent variables of interest (non-state security and state arranger). Conversely, the Victims of Crime Survey, which is a nationally representative survey, contained a range of indicators on crime and security (including who individuals rely on for security), but lacked sufficient indicators of legitimacy that would allow me to test the relationship between individual reliance on non-state security and their perceptions of state legitimacy. Both of these surveys, then, while providing a rich source of public opinion data, were not sufficient for an analysis of the relationship I was most interested in explaining. Therefore, a supplementary survey with ordinary citizens was in order. The survey that I designed would allow me to include various indicators of crime and security, as well as indicators of legitimacy. Moreover, developing this survey provided an opportunity to develop a new and important indicator that gauges citizens’ normative views of the proper role of the state in service (in this case security) delivery. Therefore, the new survey data would allow me to fully test the conditional effect of !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 43 When I conducted my fieldwork in South Africa in 2010, Afrobarometer survey data did not include a question that gauged who individuals turned to for their security needs. However, Afrobarometer added a security question in the survey that was administered in OctoberNovember 2011. This question asked individuals “if you were a victim of crime in this country, who, if anyone, would you go to first for assistance”? !! ! 95! non-state security on perceptions of legitimacy when individuals see the state as a producer or arranger of security. Sampling Procedures and Study Participants When conducting survey research, researchers must first identify the relevant population for their study. Once the target population has been identified, researchers may then compile a sampling frame (if possible), and select the appropriate sample. For those who are interested in generalizing from the sample to the population, it is necessary to employ some form of probability sampling. For this research, the relevant population would be all adult South Africans. Since I am interested in understanding citizens’ attitudes toward the legitimacy of the state in South Africa, all adult South Africans would need to have an equal probability of being included in the sample to achieve the greatest generalizability. However, limited resources made a survey that spanned all nine provinces in South Africa unfeasible. Therefore, I conducted this survey on a smaller scale, relying on a multi-stage sampling process. First, I chose to explore the relationship between non-state security and perceptions of state legitimacy in the country of South Africa for several reasons. South Africa has been known to contend with extraordinarily high levels of crime and violence (See chapter 3). Therefore, issues of crime and security are particularly salient in this context and have remained so over time. The salience of crime and security issues coupled with perceptions that the police are unable to adequately address these issues has created a large role for non-state security providers to play in this country. Thus, South Africa is an opportune context within which to study how individuals’ ! 96! reliance on non-state security shapes their perceptions of state legitimacy. The selection of the country is therefore a purposive one. I also used a purposive sampling technique to choose three major cities- Cape Town, Durban and Johannesburg- in which to administer the survey. Due to the limited resources mentioned above, I chose to restrict the survey to the three largest urban centers in the country. Since crime is heavily concentrated in urban areas, it seemed likely that these would be the places where I would most likely find a significant number of individuals who rely on non-state security. I would expect those who live in urban areas to be more likely to need and have the means to employ commercial security firms. But, I also expected urban areas to have the greatest concentration of individuals who rely on societal-based non-state security, given the historical prevalence of these structures in places like urban townships. After choosing the cities where I would carry out the survey, I then created a sampling frame of 44 neighborhoods using South African census data collected in 2001 . I employed a disproportionate stratified random sampling procedure to select neighborhoods based on race and 45 income levels . Because the distribution of political and economic goods has historically been uneven across racial groups in South Africa, and because of the racial hierarchical order in !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 44 When I began my fieldwork in 2010, the 2001 census data was the most recent census data available. 45 Because census data had been collected nine years prior to the start of my fieldwork, before beginning data collection I wanted to ensure that these neighborhoods has not drastically changed in terms of their socio-economic makeup. Therefore, I consulted with fieldworkers who were familiar with local neighborhoods in each city to determine if the classification of these neighborhoods held face validity. In virtually every instance they confirmed that the classification of neighborhoods made sense, suggesting that the composition of South African neighborhoods is not changing much over time. ! 97! political and social life, it was important to stratify the sample along racial lines. It was also useful to stratify the sample by income, since reliance on non-state security (especially marketbased security) may be a function of growing socio-economic affluence. I oversampled Whites, Asians and Coloureds because their views on the legitimacy of the state were likely to differ starkly from the majority African population and I wanted to ensure that their perceptions were adequately represented in the survey. Moreover, by including Cape Town and Durban in the sample, I was also able to capture the responses of Coloureds and Asians respectively, where each group is concentrated. Oversampling minorities afforded the opportunity to make meaningful statistical inferences for each sub-group. My original plan was to stratify the sample by race and income in order to yield six strata consisting of low and high-income neighborhoods for each major racial group (see Table 1). According to the sample frame, I planned to randomly select 6 neighborhoods within each of the six strata, yielding a total of 36 neighborhoods in each city. Next, I would have randomly selected 4 households within each neighborhood and the individual to be interviewed within the household would have been selected randomly, but alternated by gender. Thus, the total sample size for ordinary citizens was to be 144 respondents in Cape Town, 144 respondents in Durban and 144 respondents in Johannesburg, for a total sample size of 432. ! 98! Table 4.1: Proposed Sample Frame for Cape Town, Durban and Johannesburg, South Africa Race Income Africans Whites Low 6 n, 24r High 6 n, 24r n=neighborhood(s), r=respondent(s) 6 n, 24r 6 n, 24r Coloureds in Cape Town (Asians in Durban) 6 n, 24r 6 n, 24r N=144/city Total N=432 In reality, however, there were often little to no White neighborhoods that qualified as lowincome and in many instances there were no African or Coloured neighborhoods that classified as high-income. Therefore, the sample frame had to be adjusted accordingly. The actual sample frame employed in each city is depicted in Tables 2-4. In all three cities, neighborhoods where at least 55% of residents were African, White, Coloured or Asian were classified as African, White, Coloured, or Asian neighborhoods, respectively. Mixed neighborhoods (Cape Town only) consisted of neighborhoods where no one racial group comprised a majority (50+1). 46 In Cape Town, neighborhoods where at least 55% of residents earned R0-6400 /month were classified as low-income neighborhoods, whereas neighborhoods where at least 55% of residents earned R6401 or above/month were classified as high-income neighborhoods. Given the lack of high-income African and Coloured neighborhoods, I increased the number of respondents in some of the larger low-income African and Coloured neighborhoods, and also included some of the few mixed neighborhoods that exist in the sample. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 46 As of March 2013, R6400 was equal to about $700.00. ! 99! Table 4.2: Sample Design for Cape Town, South Africa Race Income Africans Whites Coloured Low 5 n, 36r 2 n, 8r 6 n, 36r High 7 n, 28r n=neighborhood(s), r=respondent(s) Mixed 6n, 24r N=132 In Johannesburg and Durban, there were both high and low-income neighborhoods for each racial group. In these cities, I classified neighborhoods where at least 20% of residents earned at least R6401 or more per month as high-income neighborhoods. Given that there were so few neighborhoods where a majority of residents earned over R6401 per month (regardless of race), I lowered the income bar in order to include a range of neighborhoods in the Johannesburg and Durban samples. With this cut point, I was able to incorporate both high and low-income neighborhoods for each racial group. Table 4.3: Sample Design for Johannesburg, South Africa Race Income Africans Whites Coloureds Low 5n, 24r 6n, 24r 3n, 12r High 6n, 24r 6n, 24r 3n, 12r n=neighborhood(s), r=respondent(s) Asians 3n, 12r 3n, 12r N=144 ! 100! Table 4.4: Sample Design for Durban, South Africa Race Income Africans Low 4n, 30r High 2n, 12r n=neighborhood(s), r=respondent(s) Whites 11n, 48r Asians 6 n, 24r 5n, 24r N=138 Stratifying the sample by race and income yielded 5 strata in Cape Town and Durban and 8 strata in Johannesburg (see Tables 4.2-4.4). I then randomly selected neighborhoods within each 47 strata . Randomly selected neighborhoods were achieved by placing the names of all neighborhoods that met the criteria for a particular stratum into a bag. I then withdrew the amount of names that corresponded to number of neighborhoods needed for each stratum. These 48 names were the neighborhoods in which the interviews would be conducted . 49 Next, I randomly selected households within each neighborhood . Starting at a random start point, field workers were instructed to select houses based on a 10-house walk pattern (counting houses on the right and left (see the “Data collection” section below for more details). Once a household was selected, an individual in the household that was eighteen years of age or older was selected as the survey respondent. To ensure random selection of the respondent, names of all members of the household (including those who were not home at the time of the interview) were recorded next to a numbered list in the questionnaire. The fieldworker would then use !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 47 In general, 6 neighborhoods were selected per strata, though some strata have more or less neighborhoods. 48 A list of all neighborhoods included in the sample can be found in Appendix A. 49 In general, 4 households/respondents were selected per neighborhood, but more interviews were conducted in especially large neighborhoods, meaning that some neighborhoods had more than 4 households/respondents. ! ! 101! playing cards with numbers that corresponded to the numbered list of household members. The individual with whom the fieldworker was speaking would then blindly choose a card, and the individuals’ whose name corresponded to the name on the card would become the survey respondent. Interviews were alternated by gender. Given the smaller scope of my survey, the sample was not proportionate to population size. However, the sampling method that I employed ensured that a diverse range of neighborhoods in all cities were included in the sample. The total sample size is 132 respondents in Cape Town, 138 respondents in Durban and 144 respondents in Johannesburg for a total sample size of 414 (see Table 4.5). Below I have listed the various stages I employed in the multi-stage sampling process. o Stage 1: country selection (purposive) o Stage 2: city selection (purposive) o Stage 3: stratified neighborhoods by race and income levels o Stage 4: neighborhood selection within strata (random) o Stage 5: household selection within neighborhood (random) o Stage 6: individual selection within household (random with gender quota) Table 4.5: Total Sample Frame Africans 130 Whites 144 Cape Town 132 By Race Coloureds 63 By City Johannesburg 144 Asians 73 Other 4 Durban 138 N=414 ! 102! Description of the Survey Instrument The survey that I carried out between October 2010 and July 2012 was conducted in English 50 and administered through face-to-face interviews. A total of 414 interviews were conducted. Samples of this size yield a margin of error of +/- 5%. The questionnaire was comprised of nine sections. The full questionnaire can be found in Appendix B. The first section gathered data on the respondents including their age, primary language of use and contact information. The second section included questions designed to tap individuals’ attitudes toward the state, government and regime. For the purposes of this research project, I was most concerned with measuring citizens’ attitudes toward the state, specifically their perceptions of state legitimacy. To measure state legitimacy at the individual level, I 51 adopted indicators from the Afrobarometer survey . These three indicators, which I used to create an index of state legitimacy, are as follows: For each of the following statements, please tell me whether you disagree or agree: ! The courts have the right to make decisions that people always have to abide by. ! The police always have the right to make people obey the law. ! The tax department always has the right to make people pay taxes. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 50 Given that all surveys were administered in urban areas where most people speak English well, the use of English as the survey language did not pose problems. 51 Afrobarometer is the largest cross-national public opinion survey in Africa. Greater details on this survey will be given in a later section of this chapter.!! ! 103! These measures of legitimacy have been used by other scholars in the discipline and are consistent with my conceptualization of legitimacy, which focuses on individuals’ perceptions that institutions of the state are the right and proper ones for society. A subsection of section two (section 2B) included a battery of questions to determine citizens’ normative views on the role the state should play in service delivery. In particular, these questions asked who, in particular should be responsible for the provision of a range of goods and services like education, health care and security. The main variable of interest here was one in which I asked citizen’s if they believed the state should directly provide security or act as more of a facilitator in security provision. The specific question is as follows: “Which of the following statements is closest to your view? Choose statement 1 or statement 2. Statement 1: The state should take the main responsibility for protecting citizens from crime. Statement 2: The state should mainly provide support to private groups such as the commercial security industry or neighborhood watches so that they may protect citizens from crime.” This variable becomes key to my analysis in chapter 5 where I show that the effect of non-state security on perceptions of state legitimacy depends upon how citizens respond to this question. Section three of the survey focuses on citizens’ perceptions of crime and personal insecurity. It also ascertains their views on the police and courts. Specifically, the questions in section three ask citizens to evaluate the performance of the police in a number of areas. Section four builds upon these questions by asking respondents to indicate the extent to which they trust these state agents and believe they are involved in corruption. ! 104! Section five comprises another key section of the instrument as this is the section that asks individuals how they meet their security needs. In particular, this section asks who individuals rely on for security, both for crime prevention and after a crime has occurred. Depending upon who individuals report relying on for security, more detailed questions are asked such as who pays for the security (in the case of commercial security), and whether the respondent helped to form the group that they rely on (in the case of societal security). If respondents rely on non-state security, they are also asked to evaluate the performance of the non-state security entity that provides their security. The main object of interest in section five is measured by an indicator of individual reliance on non-state security. This is the key explanatory variable in the chapter 5 analysis. Recall that I define non-state security as the assurance of safety of person and property that is produced by private actors. To measure whether individuals rely on non-state sources for protection, I utilize a question that asks: “Please tell me who you would call on if you or someone in your family had been a victim of a crime?” The response categories consist of “police”, “some other group”, “no one” and “don’t know”. Section 6 of the survey attempts to gauge the extent to which individuals see non-state security organizations as legitimate. Here, I ask citizens whether they believe non-state security organizations should have the right to patrol neighborhoods, arrest criminals and punish criminals. In addition to knowing whether individuals view institutions of the state as legitimate, I thought it equally interesting to know whether citizens believe non-state actors should have the right to fulfill what have conventionally been thought of as state functions. ! 105! Section seven is a small section that contains questions on social capital. These questions are primarily concerned with perceived levels of social trust in the respondent’s area of residence. Section eight collects more information on the respondent. It asks about their level of education, employment and marital status, religious affiliation and occupation. This section also attempts to gauge the level of impoverishment that each respondent has faced in the last year, as well as the respondent’s race and gender. I asked the final set of questions last due to their sensitive nature. Section nine asks respondents about their personal experiences of victimization. In particular, respondents are asked whether they ever experienced any contact or property crimes. Contact crimes consist of murder, robbery, assault or rape. The battery of property crime questions include burglary, motor vehicle theft and theft of a range of other goods including bicycles and items from cars such as car radios. Data Collection Procedures and Survey Administration The data for this project were collected from November 2010-June 2012. The Cape Town interviews were carried out from November 2010-February 2011. About one-third of the Durban interviews were completed in 2011. In particular, interviews with Africans were completed in 52 April 2011, while the remaining interviews were completed from June-July of 2012 . Finally, interviews in Johannesburg were carried out between August and September of 2011. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 52 Interviews in Durban were delayed due to difficulties recruiting fieldworkers from Asian and White communities. ! 106! Recruitment/Training of Fieldworkers The recruitment of fieldworkers was greatly facilitated by institutional affiliations that I had within South Africa. During my fieldwork in Cape Town, I had a formal affiliation with the Democracy in Africa Research Unit (DARU) of the University of Cape Town. DARU was a part of a larger research center housed at the University of Cape Town, the Centre for Social Science Research (CSSR). In addition, I was greatly assisted by the Centre of Criminology under the direction of Professor Clifford Shearing. With the help of these two centers, I was able to recruit fieldworkers who had extensive experience conducting interviews in Cape Town. In Durban, I recruited fieldworkers based on my contacts from the Fulbright-Hayes Zulu Group Project Abroad (GPA) program that I participated in 2008. The fieldworkers that I hired in Durban assisted with the implementation of the GPA program in 2008. This consisted of assisting students in the program with learning to speak, read and write the Zulu language at the advanced level. In survey research, researchers always attempt to recruit fieldworkers that are representative of the communities in which they conduct interviews. In Cape Town and Durban I was successful in recruiting Coloured (Cape Town only) and African fieldworkers. I failed to, however, recruit White and Asian (Durban only) fieldworkers. Therefore, I worked with a local South African firm to conduct interviews with Whites in Cape Town and with Asians and Whites in Durban. Due to the lack of institutional contacts in Johannesburg, I also relied on this firm to conduct all interviews in Johannesburg. The name of this local firm is Dibanaa Field and Research Services, ! 107! a firm that often works with graduate students and other individuals and corporations that are interested in carrying out public opinion surveys on a smaller scale. The founder of this company is an employee for Citizens Surveys, a large firm that specializes in conducting mass public opinion surveys. Their clients include the Afrobarometer survey, among others. In order to ensure that fieldworkers understood the aim of the survey and their specific responsibilities and duties as fieldworkers, briefing sessions were held with all fieldworkers (including Dibanaa field supervisors) prior to the start of fieldwork. During the briefing session I gave an overview of the project, discussed the importance of informed consent and reviewed, in detail, sampling procedures. I also conducted mock interviews to ensure that fieldworkers were clear on how to administer the survey. Briefing sessions for the Johannesburg and the June-July 53 2012 portion of the Durban data collection were conducted via Skype . Pilot Survey with Fieldworkers Many of the key questions utilized in this survey have been adopted from other large public opinion surveys such as Afrobarometer. In many cases, these questions have been asked across several rounds of surveys and shown to be reliable and valid measures. However, I developed several measures (i.e. state arranger) myself and thus was unsure as to how well the measures would be understood by ordinary citizens. To minimize problems in the field with the new questions, I conducted a pilot of the survey with field workers in Cape Town. Because the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 53 My fieldwork period ran from October 2010-April 2011. Therefore, I was out of the country at the time these interviews were underway.!!! ! 108! fieldworkers themselves were South African citizens, I felt it was appropriate to test the questionnaire with them. Administration of the Survey The survey employed a pre-coded questionnaire with 65 items which was administered through face-to-face interviews, each of which lasted approximately 30-45 minutes. I thoroughly 54 explained the consent form to all fieldworkers and emphasized that a consent form signed by the respondent must accompany each completed questionnaire. Also, all fieldworkers were given letters with my contact information to explain the work they were doing in local communities, if needed. In most instances, field workers reported that they faced no constraints in administering 55 the surveys . I was responsible for selecting all neighborhoods where surveys were to be carried out. Once fieldworkers were ready to begin data collection, I provided them with the names of the neighborhoods where they were to conduct interviews and designated a specific start point for each neighborhood. The instructions to fieldworkers for the selection of households were as follows: ! The project director (Danielle Carter) will choose a random start point within the community where interviews are to be conducted. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 54 55 See Appendix C for the full consent form that was administered with each survey. There were are few respondents in Johannesburg that refused to sign consent forms. Otherwise, field workers encountered no other problems when conducting interviews.! ! 109! ! Beginning at the start point, field workers should then walk in the opposite direction of th each other until they reach the 10 house, counting houses on the right and left. The 10th house will be the house where the fieldworker asks for the interview. ! If no one is at home, fieldworker may substitute with the next household. If the interview is refused, fieldworkers must use an interval of 10 to select a substitute household. If the interview is refused, the fieldworker must record why on the table on the first page of the questionnaire. Once fieldworkers found a home where someone was home and did not refuse the interview, the instructions to fieldworkers for the selection of respondents were as follows: ! Within the household, fieldworkers are responsible for selecting an individual AT RANDOM. This individual becomes the respondent. Because the study needs to include an equal number of men and women, fieldworkers should use the table on page two to help ensure that they are alternating the respondents by gender. ! The fieldworker should ask for the names of all the males/females (dependent on which is needed for the interview; i.e. if the interview must be with a female, only record female names) who live in the household. In addition, the fieldworker is responsible for making sure that they are only recording the names of people who are at least 18 years of age and citizens of South Africa. The fieldworker must list names of all males/females who are at least 18 and citizens of South Africa, even those who are not presently home. ! Take out deck of cards that correspond to the number of males/females (i.e. if there are 3 women, take out the ace=equal to number 1, number 2 and number 3 cards). Place the ! 110! cards face down and have the person choose one. Let them know that the person who corresponds to the number chosen will be the person interviewed. ! If that person is not home and will return that day, select a time to return for the interview. If that person is not home and will not return that day, substitute with the household to the right and repeat the respondent selection procedure. ! FIELDWORKERS MAY ONLY SUBSTITUTE HOUSEHOLDS, NOT RESPONDENTS. Quality Control In order to ensure that interviews were appropriately conducted, I conducted household backchecks of approximately 50% of the interviews. All back-checks were conducted via phone. During the back-check, I asked to speak with the name of the respondent recorded on the original questionnaire. I then asked if the person recalled completing an interview with (interviewer’s name). Once the respondent confirmed that they had, I also asked them to confirm two unique identifiers (usually the name of the neighborhood in which the respondent lived and their age). Data Capturing To assist with data capturing, I hired an assistant to input the data from all questionnaires into an Excel spreadsheet. The assistant manually input data from Johannesburg, Cape Town and data on the African respondents in Durban. Given the late date of data collection, a field supervisor from Dibanaa captured the data on Asian and White respondents from Durban when they completed data collection. As a second check, I then checked the codes for each question from ! 111! the physical questionnaire against the codes in the excel file to ensure they were inputted correctly. I then created a dataset from the excel file using SPSS statistical software. QUALITATIVE INTERVIEWS WITH ELITES IN SOUTH AFRICA To gain a fuller understanding of the nature of crime and security issues in South Africa, I conducted 21 qualitative interviews with elected officials, members of the South African Police 56 Service (SAPS) and civil society organizations working on crime and security issues . With regard to elected officials, I attempted to interview individuals who were directly involved with committees that deal with crime and security in South Africa. Therefore, the Members of Parliament that I interviewed were all members of the Select Committee on Security and Constitutional Development. This committee is responsible for oversight of the Department of Correctional Services and the Department of Police, to name just a few. Therefore, these officials were deeply aware of contemporary crime challenges facing South Africa, and public policy responses to them. As with national level officials, I also conducted interviews with local government officials who were specifically responsible for security issues in their respective cities. In Cape Town and Johannesburg, I interviewed two members of the Mayoral committee. These two individuals were the members of the committee who were directly responsible for public safety. In addition, !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 56 See Appendix D for the consent form used with elites. ! 112! 57 in Durban I interviewed the head of the Independent Complaints Directorate and the head of Safer Cities, a local unit aimed at combating crime in Durban through a multi-agency approach. Table 4.6: Elite Interviews in Cape Town, Durban and Johannesburg, South Africa Type of Interview Number of Interviews City Members of Parliament 4 Cape Town Local Government Officials 4 Cape Town, Johannesburg, Durban Police Officials 4 Johannesburg/Pretoria, Cape Town, Durban Civil Society Groups Business Against Crime Community Police Forum Officials Institute for Security Studies Center for the Study of Violence and Reconciliation Commercial Security Industry 2 3 1 1 Cape Town, Durban Cape Town, Johannesburg, Durban Pretoria Cape Town 2 Johannesburg, Durban Total Elite Interviews=21 Two of the police officials that I interviewed were key to building partnerships between the community and police. One official was the head of community policing for the province of KwaZulu-Natal and the other was the commander of Partnership Policing for the Western Cape. These interviews proved valuable as they provided the opportunity to hear from police officials about their perceptions of, for example, community police forums and the commercial security industry. These interviews provided valuable insights on how the police view the involvement of citizens and private corporations in security provision. The other two police officials were less directly involved in partnership work, but still offered helpful information. One of these officials was an officer with the Johannesburg Metro Police Department, and therefore lent helpful insights into the functions of the metropolitan department. The other was a high-ranking official !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 57 This body is responsible for investigating complaints of police misconduct.!! ! 113! at the head office with many years of experience in SAPS and broad knowledge of crime and security challenges in South Africa. Finally, the perspective of police officials was balanced by my meetings with leaders from civil society who gave their perspectives on security issues and their views on the police. The interviews with community police forum officials were particularly helpful in understanding citizens’ involvement in security. These interviews also provided a useful means for contrasting how police officials understand community involvement with how citizens themselves perceive their own involvement in security matters. Individuals from the Center for the Study of Violence and Reconciliation and the Institute for Security Studies were helpful in outlining crime and security challenges as well as providing access to useful resources on these issues. Finally, interviews with representatives from Business against Crime and the commercial security industry proved useful for understanding, in greater depth, the role they play and the challenges they face in dealing with crime. Insights gleaned from qualitative interviews with elites were used, primarily, to frame my interpretations of the statistical results in chapters 5 and 6. AFROBAROMETER SURVEY DATA 58 Afrobarometer is the largest cross-national public opinion survey in sub-Saharan Africa. It is a nationally representative survey that measures citizens’ attitudes toward democracy, markets and !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 58 For more information see www.afrobarometer.org ! 114! civil society. Afrobarometer has been in existence for over ten years. The last completed round of surveys, administered in 2008, included twenty sub-Saharan African countries. The round that is currently underway includes 35 African countries. To date, six rounds of surveys have been carried out in South Africa with the last being conducted from October-November 2011. The South African Afrobarometer survey consists of 2400 South African citizens who are 18 years of age or older. Samples of this size yield a margin of error of +/- 2%. I make extensive use of the latest round of South Africa Afrobarometer data in chapters 5 and 6, while also relying on Afrobarometer data more generally in several other areas of the dissertation. VICTIMS OF CRIME SURVEY DATA The Victims of Crime Survey (VOCS) is a nationally representative survey conducted with residents of South Africa that are 16 years of age or older. It captures individual and household experiences with crime, perceptions of crime and security and perceptions of state institutions such as the police force and courts. Five rounds of victimization studies have been carried out in South Africa. The first was carried out in 1998 by Statistics South Africa. The Institute for Security Studies (ISS) was responsible for conducting the 2003 and 2007 versions of the VOCS. In 2011 and 2012, Statistics South Africa once again resumed responsibility for administering the VOCS and they will continue to do so on an annual basis from this point forward. In 1998, the sampling procedures were as follows: ! A probability sample of 800 enumerator areas was drawn from a sampling frame of 86 000 enumerator areas as demarcated in the 1996 population census. ! 115! ! The sample was explicitly stratified by the nine provinces and implicitly stratified by the 42 police areas in the country. ! From each of the 800 enumerator areas, a systematic sample of five households was selected for interviewing. ! One respondent aged 16 years and above was selected from each of the five households. In 2003, the sampling procedures were as follows: Multi-stage cluster sampling was utilized, with Enumerator Areas (EAs) from the 2001 Census selected at the first stage of the sampling, households within the EAs at the second stage, and individuals within the household at the third stage. Based on the total number of households in South Africa (identified by the 2001 Census as 11,205,705), a total of 80,787 EAs were allocated. Ten interviews were conducted within each EA. In total, a sample of 4,860 respondents was realized. The same sampling procedure was utilized in 2007 with a representative sample of EAs, households, and individuals within households being selected at random. In 2011 and 2012, the same sampling procedure was followed. Both surveys used a Master Sample (MS) originally designed for the Quarterly Labour Force Survey (QLFS) as a sampling frame. The MS is based on information collected during the 2001 Population Census conducted by Statistics South Africa (SSA). The MS has been developed as a general-purpose household survey frame that can be used by all household-based surveys irrespective of the sample size requirement of the survey. The VOCS 2011 and 2012 use an MS of primary sampling units (PSUs) which comprise census enumeration areas (EAs) that are drawn from across the country. ! 116! The sample for the VOCS 2011 and 2012 used a stratified two-stage design with probabilityproportional-to-size (PPS) sampling of PSUs in the first stage, and sampling of dwelling units (DUs) with systematic sampling in the second stage. The sample was designed to be representative at provincial level. A self-weighting design at the provincial level was used and MS stratification was divided into two levels. Primary stratification was defined by metropolitan and non-metropolitan geographic area type. During secondary stratification, the Census 2001 data were summarized at PSU level. The following variables were used for secondary stratification: household size, education, occupancy status, gender, industry and income. A Probability Proportional to Size (PPS) systematic sample of PSUs was drawn in each stratum, with the measure of size being the number of households in the PSU. The Master Sample consists of 3 039 PSUs. In each selected PSU, a systematic sample of dwelling units was drawn. The number of DUs selected per PSU varies from PSU to PSU and depends on the Inverse Sampling Ratios (ISR) of each PSU and the number of dwelling units in that PSU. The total sample size in 1998 was 4,000. The institute for security studies carried out a total of 4,860 interviews in 2003 and 4,500 interviews in 2007. The sample size was 29,754 dwelling units in 2011 and 31,007 dwelling units in 2012 . I rely on VOCS data in chapter 3 of the dissertation. This chapters outlines the crime and security context. I use VOCS survey data to tap individuals’ perceptions of crime, their reported experiences of victimization and their attitudes toward state officials such as police and court officials. ! 117! OFFICIAL STATISTICS In addition to reporting citizens’ perceptions of crime in chapter 3, I also report on official statistics, including official police statistics and statistics from the Private Security Industry Regulatory Authority (PSIRA). Official crime statistics are published annually through the South 59 African Police Service (SAPS) . I use these statistics to give a thorough description of the security climate in South Africa. In particular, I examine the prevalence of different types of crime, and crime trends over time. 60 Likewise, data are readily available on the commercial security industry in South Africa . These data allow me to explore the extent to which the private security industry permeates South African society. In chapter 3, I chart how demand for private security varies across space and time, the types of commercial security services people most often rely on, and, importantly, the growth of the private security industry. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 59 60 ! For more information see saps.gov.za For more information go to psira.co.za! 118! CHAPTER 5 EXPLAINING THE EFFECT OF NON-STATE SECURITY ON PERCEPTIONS OF STATE LEGITIMACY ! 119! INTRODUCTION This chapter empirically tests the relationship between citizens’ reliance on non-state security and their perceptions of state legitimacy in South Africa. It draws on original survey data collected in Cape Town, Johannesburg and Durban, South Africa between November 2010 and 61 July 2012. The sample consists of 414 ordinary South African citizens eighteen years of age or older. Samples of this size yield a margin of error of +/- 5%. South Africa is considered a deeply divided society, one in which race, and increasingly class, substantially shapes individuals’ attitudes and experiences. In order to ensure that the attitudes of all relevant groups were adequately represented within the sample, I oversampled minorities and stratified the sample by 62 race and income, creating a total of 8 strata . Unless otherwise specified, all the data reported in this chapter will be from the data collected in South Africa during the 2010-2012 time period. This chapter primarily uses this data to explain whether and how individuals’ reliance on nonstate security shapes their perceptions of a legitimate state. However, before attempting to explain this relationship, I will first provide a descriptive account of popular perceptions of state legitimacy and non-state security. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 61 The total number of respondents was 132 in Cape Town, 138 in Durban and 144 in Johannesburg. 62 These strata consisted of four strata of low-income groups; low-income Africans, Whites, Asians and Coloureds and four strata of high-income groups; high-income Africans, Whites, Asians and Coloureds. For details on income cut-points and sampling, please see the methodology chapter (chapter 4). ! 120! DESCRIBING NON-STATE SECURITY AND STATE LEGITIMACY In chapter two, I make use of Lipset’s classic conceptualization of legitimacy as “the capacity of the political system to engender and maintain the belief that existing political institutions are the most appropriate or proper ones for the society” (Lipset 1959, 86). I focus on the state as the key object of legitimacy. Because I rely on a Weberian notion of the state, I am primarily interested in citizens’ perceptions of the legitimacy of those institutions that are responsible for exercising coercion. Therefore, I assess citizens’ perceptions of three state institutions that are chiefly responsible for this task; the police force, the courts and the tax agency. The specific questions asked of survey respondents are as follows: “For each of the following statements, please tell me whether you agree or disagree: ! The police always have the right to make people obey the law.” ! The courts have the right to make decisions that people always have to abide by.” ! The tax department always has the right to make people pay taxes.” In general, perceptions of these state institutions are quite positive, with over three-quarters of respondents saying that the police, courts and tax agency have a legitimate right to rule. When accounting for the five percent margin of error, we see that an almost identical percentage of 63 respondents rank each state institution as legitimate . (Figure 5.1). !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 63 I also examine the distribution of legitimacy perceptions using South Africa Afrobarometer data from 2011. Afrobarometer employs a nationally representative sample of 2400 South African citizens and asks the same questions explored above to tap legitimacy perceptions. I find that a slightly lower percent of individuals’ view all three institutions in the nationally representative sample as legitimate (i.e. the percentage of those who say they “agree” or ! 121! Figure 5.1: Legitimacy of State Institutions, 2010-2012 Question: The police always have the right to make people obey the law. Question: The courts have the right to make decisions that people always have to abide by. Question: The tax department always has the right to make people pay taxes. Note: Percentages shown are the respondents who “agree” or “strongly agree” with the above statements N=413/N=411/N=410 While it is helpful to know how individuals perceive the legitimacy of individual state institutions, this dissertation is more concerned with the legitimacy of the state as a whole. Therefore, I develop a composite measure of state legitimacy using the three variables discussed above. I create an index of state legitimacy by using the mean score of these three items (values 0-4). I then use this index to create a binary variable, setting to one those whose mean score on the index is 3 or higher and setting to 0 those mean score on the index is less than 3. I create a binary variable from the index primarily to be used for logistic regression analysis in the next section. Since the distribution of the state legitimacy data are highly skewed, an analysis that makes use of a dichotomous dependent variable is most appropriate. Moreover, the binary !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! “strongly agree” with these statements is 74, 79 and 72 percent for the police, courts, and tax agency, respectively), but that these figures still come very close to three-quarters of respondents who view these institutions as legitimate in my sample. ! ! 122! variable used here is a strong composite measure of state legitimacy because it, in effect, codes those who “strongly disagree”, “disagree” or “neither agree nor disagree” with any of the three items as 0 or “not legitimate”. This allows me to compare those who see all three institutions of the state as legitimate to those who do not. Using the binary variable, we see that the state is perceived to be highly legitimate in South Africa (Figure 5.2). Almost 70% of individuals agree that the state should have a recognized right to rule. However, it is worth noting that almost a third of South African citizens do not see the state as legitimate. Figure 5.2: Legitimacy of the State in South Africa, 2010-2012 Question: The police always have the right to make people obey the law. Question: The courts have the right to make decisions that people always have to abide by. Question: The tax department always has the right to make people pay taxes. N=413 Given the deeply divided nature of South Africa, I next explore differences in legitimacy perceptions across race and region. Figure three shows the percentage of individuals who see the state as legitimate by race. Immediately, we see that Whites are substantially more likely than Africans or Coloureds to see the state as legitimate. Under apartheid, these two groups were the ! 123! most drastically disenfranchised and they were the two groups that bore the brunt of repressive 64 state policies . From this perspective, their significantly lower levels of perceived state legitimacy may come as no surprise. Figure 5.3: Legitimacy of the State in South Africa by Race, 2010-2012 Question: The police always have the right to make people obey the law. Question: The courts have the right to make decisions that people always have to abide by. Question: The tax department always has the right to make people pay taxes. Note: Percentages shown are the respondents who see the state as legitimate (state legitimacy=1) To illustrate further the potential role of history in shaping these groups’ perceptions of the state, let’s consider the fact that Africans, who now comprise the majority in power, are the least likely to see the tax agency as legitimate. Seventy-two percent of Africans reported that the tax agency has the right to collect taxes, compared with 83% of Whites who feel the same way. Lower levels of perceived tax agency legitimacy may be a reflection of current economic realities, namely, higher levels of unemployment among Africans, and their general lack of contact and experience with the tax agency. However, it may also reflect a historical feature of South African politics, one in which Africans employed tax and rate boycotts as part of the struggle for !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 64 It is worth noting that Asians were also victims of racist policies under apartheid, though to a lesser extent than these other two groups. ! 124! liberation. It is possible that through participation in economic strikes aimed at crippling the state, African citizens have developed deeply entrenched negative attitudes toward the state body that is responsible for collecting revenue. When exploring racial attitudes toward the state, it may also help to examine perceptions of the government. In addition to measures of state legitimacy, I also included a measure of government legitimacy in my survey to see whether perceptions of the state and perceptions of the government converge. In particular, the question asks: “For each of the following four statements, please tell me whether you agree or disagree: An elected government always has the right to make laws and policies that the country must adopt”. Values range from 0 (strongly disagree) to 4 (strongly agree). When I investigate citizens’ perceptions of government in South Africa, I find a story that is somewhat reversed from the story on state legitimacy. Taking a look at Table 5.1, we find that 64 percent of Africans believe that the government is legitimate, compared to 56 percent of Whites who report the same. Therefore, Africans are less likely than Whites to see institutions of the state as legitimate, but more likely than Whites to see the government as legitimate. Given the historical relations between Africans and state institutions like the police, this gap makes sense. Conversely, it is intuitive that Whites would be less supportive of an ANC government that they feel is not representative of or responsive to their interests. Perhaps what is most interesting about this finding, however, is that it provides evidence that citizens in South Africa are, in fact, distinguishing between the state and the government, and evaluating them on their own terms. ! 125! Table 5.1: Government Legitimacy by Race, 2010-2012 African White Coloured Asian Strongly disagree/ disagree 15% 29% 29% 15% Neither agree nor disagree 22% 15% 11% 10% Strongly agree/agree 64% 56% 60% 75% Total 100% 100% 100% 100% Question: Please tell me whether you agree or disagree. An elected government always has the right to make laws and policies that the country must adopt. Examining regional differences, we find that Cape Town residents hold the harshest critique of the state, while residents of Durban and Johannesburg have roughly equal percentages of individuals who see the state as legitimate (Figure 5.4). The steep differences between residents of Cape Town and residents of other cities may have to do with who exercises power in this city. Cape Town is one of the few places in South Africa controlled by the opposition party the Democratic Alliance (DA). If citizens view institutions of the state as being under the control of the ruling party (ANC), then those who support the opposition in Cape Town may be less likely to support the state. Figure 5.4: Legitimacy of the State in South Africa by City, 2010-2012 Question: The police always have the right to make people obey the law. Question: The courts have the right to make decisions that people always have to abide by. Question: The tax department always has the right to make people pay taxes. Note: Percentages shown are the respondents who see the state as legitimate (state legitimacy=1) ! 126! To see if this is the case, I ran a cross-tabulation between ANC control and “city” (Table 5.2). The former variable asks the following question: “Which of the following statements is closest to your view? Choose Statement 1 or Statement 2. 1: The ANC directly controls state institutions like the police force and courts. 2: The police force and courts act independently of the ANC.” The purpose of this variable is to gauge the extent to which citizens see the state as distinct from and independent of the ruling party. According to Table 5.2, 69% of Cape Town residents feel that the ruling party does in fact control institutions of the state. Given that the majority of individuals in Cape Town see state institutions as extensions of the ruling party, we would expect this to suppress their perceptions of state legitimacy. Conversely, in a place like Durban where 65 there are more ANC supporters , we would expect tighter ruling party-state relations to boost perceptions of state legitimacy, which it seems to do. Table 5.2: Perceived ANC Control of State by City, 2010-2012 Cape Town Durban Johannesburg ANC doesn’t control state 31% 22% 57% ANC controls state 69% 100% 78% 100% 43% 100% Question: Which of the following statements is closest to your view? Choose Statement 1 or Statement 2. 1: The ANC directly controls state institutions like the police force and courts. 2: The police force and courts act independently of the ANC. Turning to the sources of security provision, we investigate variation in public versus non-state security reliance. This main object of interest is measured by asking respondents the following question: “Please tell me who you would call on if you or someone in your family had been a !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 65 In the last national election of 2009, the ANC won 68% of the vote in Durban, compared to 33% of the vote in Cape Town. ! 127! victim of crime?” The response categories consist of “police only”, “police and some other group”, “other groups only”, or “none of these.” Figure 5.5 shows that a slim majority of respondents (51%) relies solely on the police for security. This figure captures the percentage of individuals who turn to the police for help after they or someone in their family had been a victim of crime. However, there is a similar trend for individuals who turn to the police for crime prevention. For example, when asked who they would turn to keep themselves or someone in their family from becoming a victim of crime, a plurality (46%) of respondents says that they would rely solely on the police. In both instances, however, we see that approximately one-third of respondents rely on the police and non-state groups for security, suggesting a hybrid approach to protection from crime and violence. Figure 5.5: Security Reliance Crime Prevention, 2010-2012 Question: Please tell me who you would call on if you or someone in your family had been a victim of crime? N=410 Figure 5.6: Security Reliance PostVictimization, 2010-2012 Question: Please tell me who you would rely on to keep you and your family from becoming victims of crime? N=408 The fact that so many people rely on the police as their primary source of security should come as no surprise. When asked who they thought should have the primary responsibility for keeping ! 128! 66 people safe, 86% of respondents said that government should primarily be responsible for this task. However, what is interesting is that less respondents see the state as being chiefly responsible for this good than for other basic services like education and health care (Figure 5.7). In the context of urban South Africa, citizens seem to be more open to creating space for nonstate actors to participate in the provision of security as opposed to the provision of other goods. In fact, this trend holds in much of Southern Africa (Figure 5.8). Figure 5.7: Primary Responsibility for Public Good Provision, 2010-2012 Question: Who do you think should have primary responsibility for managing each of the following tasks: keeping people safe? Question: Who do you think should have primary responsibility for managing each of the following tasks: managing schools? Question: Who do you think should have primary responsibility for managing each of the following tasks: managing health clinics? N=408/408/409 In 2008, the Afrobarometer survey asked citizens who they believed should have primary responsibility for a range of public goods. Figure 8 shows the percentage of respondents who said that “central government” or “local government” should have primary responsibility for law and order and maintaining schools and clinics. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 66 Government was split into two categories, central government and local government. The 86% of respondents who are reported as saying that government is responsible for safety represent the combined local and central government categories. Other response categories included “traditional leaders”, “community members”, “voluntary associations or NGOs” and “ the private sector, i.e. ADT (for safety question only)”. ! 129! Figure 5.8: Primary Responsibility for Public Good Provision in Southern Africa, 2008-2009 Question: Who do you think should have primary responsibility for managing each of the following tasks: maintaining law and order? Question: Who do you think should have primary responsibility for managing each of the following tasks: managing schools? Question: Who do you think should have primary responsibility for managing each of the following tasks: managing health clinics? Note: Figures shown are the percentage of individuals who report that “central government” or “local government” is primarily responsible for the above tasks N=2400/1200/1200/1200/1200/1500 Source: South Africa Afrobarometer Round 5 The figure shows that in practically every southern African country, more people feel that government should primarily be responsible for healthcare and education than for the provision of law and order. In these countries, more than 10% of respondents feel as if traditional leaders or community members should be chiefly responsible for undertaking law and order functions. The percentage of individuals who feel this way range from 12% in Zambia to a substantial 33% in Lesotho. In Malawi, Mozambique and Madagascar, roughly equal percentages of individuals feel that the state should be responsible for law and order and the provision of health clinics. Moreover, fewer individuals in these countries feel that education should be publicly provided ! 130! when compared to law and order and health clinics (for example, 74% and 76% of Malawians respectively feel that government should be responsible for law and order and health clinics. Only 67% feel that schools are the primary responsibility of government. This is probably a result of the long history of Christian education in this country). Given that individuals are quite supportive of non-state actors providing security, to which types of non-state security are they most likely to turn? My survey shows that reliance on market-based security dominates (Figure 5.9). Of those who rely on non-state security, a majority of 58% report relying on a commercial security company for security services. In what I refer to as societal-based forms of non-state security, 19% of respondents report turning to a community police forum (CPF), while 8% report turning to family and 7% to a neighborhood watch scheme (NWS). Of the various forms of non-state security, respondents are least likely to report relying on street committees(5%). Thus, in South Africa, individuals from all walks of life rely on a range of actors for security from the conventional police force to the market and society. However, in most instances the choice of a security provider is not a competition between public versus non-state sectors, as a full one-third of respondents rely on both public police and non-state sources jointly to meet their security needs (Figures 5.5 and 5.6). ! 131! 67 Figure 5.9: Type of Non-state Security Reliance , 2010-2012 Question: In addition to or besides the police, which of these groups or organizations would you be most likely to turn to for help after the crime? N=178 But who is most likely to rely on what forms of non-state security? The data (Table 5.3) show that there is a racial gap in terms of non-state security reliance, with Whites and Asians being more likely to rely on commercial non-state security and Africans and Coloureds more likely to rely on societal non-state security. In fact, Whites make up 56% of those who rely on commercial non-state security, although they are only about 10% of the total population. Africans, on the other hand, comprise a majority of respondents that rely on societal forms of non-state security like neighborhood watch schemes, street committees and community police forums. These figures suggest that individuals’ reliance on non-state security will depend, to a great extent, on their access to financial resources. Only individuals who have achieved a certain degree of wealth can afford to hire a commercial security company to protect their person and property. In most instances, only Whites and Asians earn a high enough income to afford these services. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 67 The survey also asked individuals about their reliance on gangs, vigilante groups and traditional leaders, however no respondents reported turning to these sources for security, even if in reality some people do. ! 132! Table 5.3: Type of Non-state Security Reliance by Race, 2010-2012 African White Coloured Asian Total Commercial Security 14% 56% 4% 26% 100% Neighborhood Watch 50% 8% 42% 0% 100% Street Committee 100% 0% 0% 0% 100% Community Police Forum 88% 3% 9% 0% 100% Question: In addition to or besides the police, which of these groups or organizations would you be most likely to turn to for help after the crime? N=178 However, commercial security companies are not without some support from African consumers (Table 5.4). In fact, among Africans who rely on non-state security, 24% rely on commercial security, coming only second to reliance on community police forums. Most of this demand for commercial security by Africans is driven by Africans respondents from Johannesburg where the African business class is concentrated. ! 133! Table 5.4: African Reliance on Non-state Security, 2010-2012 Commercial Security 24% Neighborhood Watch 10% Street Committee 15% Community Police Forum 48% Family 2% Other 2% Total 100% Question: In addition to or besides the police, which of these groups or organizations would you be most likely to turn to for help after the crime? N=62 HYPOTHESES South Africa has consistently performed well across a number of social and economic indicators since the transition to democracy in 1994. However, the country has continued to struggle with high levels of crime and public insecurity (see chapter 3). Public opinion data show that these issues have persisted as a top concern for individuals over time. The Afrobarometer survey asks citizens “In your opinion, what are the most important problems facing this country that government should address?”A cross-national examination of responses to this question shows that issues of crime and security were most salient for citizens of Madagascar and South Africa between 2008 and 2009 (Figure 5.10). Moreover, a majority of respondents (50%) from my 2010-2012 survey feels that crime is increasing and feel unsafe walking around at night (71%). ! 134! Figure 5.10: Crime and Security as the Most Important Problem, 2008-2009 Question: In your opinion, what are the most important problems facing this country that government should address? Source: South Africa Afrobarometer Round 4 I posit a security-driven explanation to citizens’ perceptions of state legitimacy. When the state’s capacity to provide security is low, citizens often turn to non-state sources for security. The question then becomes whether citizen perceptions of the state’s legitimacy are strengthened or weakened when they rely on non-state sources for this good. Recall that in chapter 1, I lay out a theory which makes a distinction between two types of roles that the state may play in service delivery, that of producer or arranger. This terminology is borrowed from Savas’ work on privatization, where he explains that the “service producer directly performs the work or delivers the service to the consumer [while]…the service arranger assigns the producer to the consumer, or vice versa, or selects the producer who will serve the consumer” (Savas 2000, 64-65). I argue that the state may act as a producer or arranger of security. Moreover, I suggest that individual reliance on non-state security may strengthen or undercut their perceptions of state legitimacy, depending on whether they see the state as a producer or arranger of this good. ! 135! When individuals view the state as a producer of security, they believe that the responsibility for the production of security largely rests with the state. For citizens who participate in political life and generate the revenue that states use to govern (payment of taxes), they may expect basic goods and services to be directly produced by the state. Thus, when citizens rely on alternate sources to fulfill what are viewed as state responsibilities, citizens’ confidence and belief in the state’s right to rule may suffer. I posit that the gap between citizens’ security demands and the state’s security supply may contribute to views of an illegitimate state. When this good is produced by non-state sources, citizens’ needs are met, but citizens may come to view marketbased or societal structures as competing sources of authority in which they vest their trust and loyalty. In this instance, state production of security provides the raison d’être for states from the perspective of ordinary citizens. Thus, when citizens rely on non-state sources for the production of this fundamental political good, the legitimacy of the state itself may be thrown into question. HYP. 1: Individuals who rely on non-state security and see the state as a producer will be less likely than those who rely on state security to see the state as legitimate In a competing hypothesis, I suggest that individual reliance on non-state security may improve perceptions of the state’s legitimacy. The reasons for this are twofold. In the first instance, the state provides a constitutional framework within which non-state security actors are allowed to operate. Without the freedom of association and the legal right to bear arms and secure contracts, non-state groups would not be able to coordinate and operate to provide security. Thus, it may be the case that citizens recognize that the state provides the broader legal context within which non-state groups may operate and the provision of this open environment is enough for citizens ! 136! to assign legitimacy credits to the state. In other words, the state may be seen as a sort of arranger, while the producers of security are non-state actors. In the second instance, governments may receive credit from their citizens just by way of supporting non-state actors that produce this good. State leaders who openly support non-state forms of security may receive credit for the moral, technical and financial support they provide for these actors. This may particularly be the case if these actors are viewed as effective in their provision of security. HYP. 2: Individuals who rely on non-state security and see the state as an arranger will be more likely than those who rely on state security to see the state as legitimate I measure non-state security using the following indicator “Please tell me who you would call on if you or someone in your family had been a victim of crime?” The response categories are “police only”, “police and some other group”, “other groups only” and “none of these”. Those who reply “police only” are coded as 0, whereas those who reply “other groups only” are coded as 1. For those who reply “police and some other group” I code them as 0 or 1 depending on how they respond to a subsequent question. For those who reply “police and some other group”, they are then asked: “If you were a victim of a crime, who would you FIRST call for help: the police or some other group?” For those who say they would call “the police” first, I code them as 0, and for those who say they would call “some other group” first, I code them as 1. Those who respond 68 “none of these ” are coded as missing. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 68 Two percent of the sample fall into this category ! 137! In addition to the impact of non-state security, I also examine the effect of other security-related variables. If one of the states’ key responsibilities is to provide security, then one could argue that the state will be judged based on how well it protects citizens from predation. Becoming a victim of crime should therefore lead people to view the state as less legitimate. Fernandez and Kuenzi (2010) have shown that experiences of crime negatively impact citizens’ attitudes toward the regime. This research also expects to find a similar impact on popular perceptions of the state. HYP. 3: Individuals who have past experiences as a victim of crime will be less likely to view the state as legitimate In any country, the police force is central to the maintenance of security and order. This institution more so than any other is seen as responsible for crime prevention and detection and security provision. Tyler (2004) shows how legitimacy is enhanced when individuals’ perceive the police to act in a fair manner. Similarly, I would expect that those who evaluate the police negatively will be more likely to see the state as illegitimate. HYP. 4: Individuals who evaluate the police negatively will be less likely to view the state as legitimate ! 138! Controls In addition to security-based variables, I also control for a number of competing explanations of state legitimacy. Several studies have pointed to the importance of institutional trust in legitimating the state (Levi et al 2009; Peltier 2007). The argument is that where individuals feel they can place confidence in institutions of the state, they will be more likely to view the state in favorable terms. We should therefore expect views of state legitimacy to be dampened where levels of trust in institutions are low. I measure institutional trust with a two-item construct comprised of the following two questions: “How much would you say you trust: 1) The police? and 2) Courts?” Values range from 0 (not at all) to 3 (a lot). For the rule of law to be credible, laws must be known to individuals and they must apply not only to ordinary citizens, but also to those who hold the reins of power (Carothers 1998; Diamond and Morlino 2004). The rule of law gauges the extent to which leaders are likely to be held accountable for wrongdoing, but also the likelihood that members of the general public who engage in criminal activity will be detected and punished. Where the rule of law is upheld in a fair and impartial manner, I would expect this to have a positive effect on perceptions of legitimacy. I measure the rule of law using an index of the following three indicators: “And with regards to enforcement of the law, In your opinion, how often, in this country: 1) Are people treated unequally under the law?, 2) Do officials who commit crimes go unpunished?, and 3) Do ordinary people who break the law go unpunished?” Values range from 0 (always) to 3 (never). ! 139! Other scholars have looked at the relationship between the regime and the state, and have argued that regime characteristics influence perceptions of legitimacy. In particular, these studies have explored the legitimating power of democracy and how a democratic regime helps to legitimate the state (Bratton and Chang 2006; Sil and Chen 2004). The expectation would be that those who show greater support for democracy would be more likely to view the state as legitimate. I measure support for democracy in the following way: “Which of these three statements is closest to your own opinion? Statement 1: Democracy is preferable to any other kind of government. Statement 2: In some circumstances, a non-democratic government can be preferable. Statement 3: For someone like me, it doesn’t matter what kind of government we have. Those who choose statement 1 are coded as 1 and all others are coded as 0. Finally, some scholars have examined levels of political participation as a key indicator of legitimacy. The thinking here is that those who participate in political life are more likely than those who do not to see the state as properly constituted. I measure political participation narrowly, examining only individuals’ electoral participation. I employ the following indicator: “With regard to the most recent, 2009 national elections, which statement is true for you?” I code as 1 those who report that they “voted” and code as 0 all those who did not vote. I also control for age, gender, lived poverty, education, employment status, city, and race. ! 140! EXPLAINING STATE LEGITIMACY To explain the impact that reliance on non-state security has on perceptions of state legitimacy, I employ a logistic regression analysis. As explained above, the dependent variable, state legitimacy, relies on three indicators that tap citizens’ attitudes toward the police, courts and tax agency. I then create a binary variable from the composite index. The index of perceived state legitimacy employed here has been used by other scholars studying this topic and has been found to be a compelling way of measuring citizens’ attitudes toward the legitimacy of the state (Levi et al. 2009). Table 5.5 presents the pooled estimation results of perceived state legitimacy. When examining the interaction between non-state security and state arranger, we find that the coefficient on the interaction term is positive and statistically significant. This suggests that individuals who rely on non-state security and see the state as an arranger will be more likely to perceive the state as legitimate. This finding supports hypothesis 2. Table 5.5: Pooled Logistic Regression Output on Perceptions of State Legitimacy, South Africa Perceptions of State Legitimacy! Independent Variables! Non-state security -.487 (.320) -.625* (.379) 1.550** (.637) -2.382*** (.513) State Arranger Non-state Security*State Arranger Contact Victimization ! 141! Table&5.5&(cont’d) Police Performance Evaluations .346* (.189) .0000 (.187) -.341 (.219) -.210 (.309) .162 (.303) .013 (.011) -.218 (.257) .150 (.243) Institutional Trust Rule of Law Support for Democracy Political Participation Age Male Lived Poverty Black (reference) White .524 (.356) .887** (.422) .113 (.403) Coloured Asian Employed .181 (.285) .082 (.096) Education Johannesburg (reference) Cape Town -.836** (.337) .396 (.385) .287 (.765) Durban Constant N=358 Standard errors in parentheses Statistical significance denoted by ***=p≤ .01,**= p≤ .05, *=p≤.10 R2: .143 -2 Log likelihood: 383.002 ! 142! To further illustrate the conditional effect of non-state security on perceptions of state legitimacy, I plot the effect of non-state security across the range of state arranger values (Figure 5.11). The graph shows that the relationship between non-state security and perceived state legitimacy is negative when state arranger equals zero (state as producer). The coefficient on non-state security when state arranger equals zero is -.487, but it is not statistically significant (p value=.129).However, there is a positive and statistically significant effect of non-state security on perceived state legitimacy when state arranger equals one (state as arranger). The coefficient on non-state security when state arranger equals one is 1.06 and it is statistically significant (p value=.056). These are the results when all covariates at set at their mean. When examining the predicted probabilities, we find that for those who rely on non-state security, the probability that one sees the state as legitimate increases from 66% to 83% when an individual moves from seeing the state as a producer to seeing the state as an arranger. This is an 17 percentage point increase, which means that individuals who rely on non-state security and see the state as an arranger have an 26% greater probability of seeing the state as legitimate when all covariates are held at their mean. ! 143! Figure 5.11: Conditional Effect Of Non-state Security On Perceived State Legitimacy The next variable, contact victimization, also has a substantial affect on how legitimate citizens perceive the state to be. This result suggests that when people become victims of contact crimes 69 like rape, assault, robbery and murder , their perceptions of state legitimacy plummet. Thus, it !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 69 The contact crime questions ask respondents the following: Rape: “People sometimes grab, touch or assault others for sexual reasons in a really offensive way, or sometimes they even rape others. This can happen either at home or elsewhere, for example in a pub or shebeen, in the street, at school, on public transport, in cinemas, on the beach or at your workplace. The person doing this could be someone you don’t know, or it could be a relative, friend or family member. Has anyone ever done this to you?” Assault: “Have you personally ever been attacked, physically beaten or threatened by someone in a way that really frightened you, either at home or elsewhere, for example in a pub or a shebeen, in the street, at school, on public transport, in cinemas, on the beach or at your workplace?” Robbery: “Has anyone ever taken something from you by using force or threatening you, or did anyone try to do so?” Murder: “Has anyone who lived in your present household been deliberately killed or murdered?” ! 144! 70 seems to be the case that when people’s person is violated in particular , that their perceptions of the state will be tainted. Next we see that perceptions of the police also factor into citizens’ assessments of the state. The positive and statistically significant coefficient on the “police performance evaluations” variable tells us that when people feel the police do a good job preventing crime, apprehending criminals, and making people feel safe, they are more likely to feel that the state has a legitimate right to rule. Thus, while the non-state provision of security may contribute to legitimacy perceptions, the performance of the state itself still matters. Finally, of the variables included to test for the effect of social structure, only two are significant. In particular, we find that Coloureds are significantly more likely than Blacks to see the state as legitimate. In fact, the coefficients on the “White” and “Asian” variables are also positive (though not significant) suggesting that all minorities are more likely than the majority African population to see the state as legitimate. The fact that all three minority groups are more likely to see the state as legitimate than Africans may reveal something about their better standing with the state, historically. To be sure, Coloureds and Asians also suffered state repression under apartheid. However, it was Africans that faced the worst treatment during this time. Therefore, their more negative attitudes toward the state may be a reflection of their historical treatment by these entities. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 70 I also explored the effect of property victimization (i.e. burglary, theft) on perceptions of state legitimacy. The coefficient for this variable was not statistically significant. ! 145! Finally, we also see that residents of Cape Town are significantly less likely than residents of Johannesburg to see the state as legitimate. I will now turn to results that are disaggregated by city in an attempt to account for this regional variation. Table 5.6 displays the logistic regression results by city. It is worth noting that when 71 disaggregating the data by city the sample size becomes small . Therefore, some of the variables that were significant in the pooled model may have lost their significance here because of the relatively limited statistical power afforded by a small sample. I will therefore focus on the signs of the coefficients. To begin with Cape Town, we see that neither non-state security nor the interaction term between non-state security and state arranger reaches statistical significance. However, the sign on the interaction term is positive, which is consistent with the direction proposed by my theory. In Durban, by contrast, the coefficient on the non-state security variable is negative (as expected), and reaches statistical significance. This means that individuals who rely on non-state security (and see the state as a producer) are less likely than those who rely on the state security to see the state as legitimate. Durban also differs in that the sign on the coefficient for the interaction term is negative (though not significant). This suggests that individuals who rely on non-state security and see the state as an arranger are less likely to see the state as legitimate. This finding counters my expectation about how viewing the state as an arranger mitigates the relationship between non-state security !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 71 The total number of respondents was 132 in Cape Town, 138 in Durban and 144 in Johannesburg.! ! 146! and state legitimacy. In the case of Durban, individuals seem to punish the state when they must rely on non-state security, no matter what they believe the state’s proper role in security provision should be. In Durban, whenever citizens must turn to the private realm for security, their perceptions of state legitimacy are undermined. However, we should take caution when interpreting the interaction action term for Durban, as only a very small number of individuals see the state as an arranger in Durban, and even less see the state as an arranger and rely on nonstate security. The very small cell count for individuals that meet all the criteria for the interaction here may therefore make this finding unreliable. Finally, in Johannesburg, we see that individuals who rely on non-state security and see the state as an arranger are more likely to see the state as legitimate. The coefficient on the non-state security and state arranger interaction term is positive and statistically significant. This suggests that in Johannesburg, people are much more open to the state as an arranger of security, and credit the state with legitimacy dividends when it plays this role. Table 5.6: Logistic Regression Output On Perceptions Of State Legitimacy, By City Independent Variables Non-state security State Arranger Non-state Security*State Arranger Contact Victimization ! Cape Town Durban Johannesburg -.458 (.624) -.563 (.694) 1.628 (1.331) -2.465*** (.768) -2.237** (1.066) -17.331 (28411.725) .202 (.707) -1.204* (.710) 2.320** (1.189) -2.887*** (1.094) -2.835*** (1.119) -3.300*** (1.012) 147! Table 5.6 (cont’d) Police Performance Evaluations .788** (.408) .009 (.422) -.026 (.360) Support for Democracy -.024 (.363) -1.095*** (.436) -1.028* (.555) -.075 (.453) .141 (.463) .531 (.973) .016 (.353) .217 (.436) .201 (.542) Political Participation -.497 (.609) .209 (.846) 1.132** (.544) .062*** (.023) .443 (.478) -.226 (.477) .007 (.028) -1.032** (.627) -.452 (.509) -.050** (.021) .068 (.484) .975* (.551) .736 (.727) 1.126 (.748) - -.551 (.991) - 2.154*** (.845) 1.487* (.811) .936 (.719) .493 (.537) .142 (.206) .081 (1.432) 124 .203 119.374 Institutional Trust Rule of Law Age Male Lived Poverty Black (reference) White Coloured Asian Employed .016 (.894) -.310 (.813) .312 (.214) 1.930 (1.996) 112 .302 83.371 .220 (.515) Education -.275 (.197) Constant .471 (1.553) N= 122 R2: .286 -2 Log likelihood: 120.566 Standard errors in parentheses Statistical significance denoted by ***=p≤ .01,**= p≤ .05, *=p≤.10 How do we explain the differences across cities found above? Why do citizens punish the state for being an arranger in Durban but not in Johannesburg or Cape Town? In Johannesburg, forty- ! 148! 72 three percent of respondents are open to the idea of state as an arranger of security and hence to the inclusion of non-state actors in the production of security. In Cape Town, 20 percent of respondents see the state as an arranger. But in Durban, this figure drops to just 10 percent. Why then, are citizens so much less accepting of non-state involvement in security provision in Durban, and why do these citizens sanction the state when the state is seen to have arranged this non-state provision? Perhaps this unexpected relationship can be explained by the extent to which people see nonstate security as efficacious. In Cape Town and Johannesburg, we see that over 85 percent of respondents who rely on non-state security, feel that they are doing well at patrolling neighborhoods and arresting and punishing criminals (Figure 5.12). In Durban, by contrast, these numbers are practically halved. A little over half of the respondents who rely on non-state security in Durban feel that they do a good job patrolling neighborhoods, but only one-third are satisfied with the way that they deal with criminals. Therefore, we may have a situation in which citizens only accord the state legitimacy when it is perceived to be leveraging effective security services from non-state security providers. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 72 Question: Which of the following statements is closest to your view? Choose Statement 1 or Statement 2. Statement 1: The state should take the main responsibility for protecting citizens from crime.Statement 2: The state should mainly provide support to private groups such as the commercial security industry or neighborhood watches so that they may protect citizens from crime. I code those who “agree” or “strongly agree” with statement 1 as 0, or those who see the state as a producer. I code those who “agree” or “strongly agree” with statement 2 as 1, or those who see the state as an arranger. ! ! 149! Figure 5.12: Satisfaction with Non-state Security in South Africa, 2010-2012 Question: How satisfied are you with the performance of [name of private organization] in: Patrolling and securing your neighborhood? Question: How satisfied are you with the performance of [name of private organization] in: Arresting criminals? Question: How satisfied are you with the performance of [name of private organization] in: Punishing criminals? Shifting to the other security-oriented variables, we find that citizens sanction the state when they become victims of violent predation in all three cities. The effect of the “contact victimization” variable is positive and statistically significant across all three cities. Most studies 73 look at the effect of perceptions of crime in shaping peoples’ attitudes toward the state . However these results show that people who report actually being a victim of crime are less likely to see the state in a positive light. Whether or not individuals actually experience contact crime has a significant impact on how much legitimacy they accord to the state. Finally, of the security-related variables, evaluations of police performance only reach statistical significance in Cape Town. As in the pooled results, here we see that people who feel that the police are performing well, are more likely to believe that the state should have the right to rule. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 73 “Perceptions of crime” variable was tested but does not reach statistical significance in this model ! 150! Turning next to test the effect of competing hypotheses, I examine the effect of institutional trust, the rule of law, support for democracy, and political participation on perceptions of state legitimacy. In Cape Town, the coefficients on support for democracy and the rule of law are negative and statistically significant. The direction of these signs are surprising, as they suggest that individuals who support democracy and those who believe the rule of law is frequently upheld, are less likely to see the state as legitimate. These results are counterintuitive, especially given the extensive literature that alludes to the legitimating affects of a strong rule of law and democracy. In Durban, none of the competing hypotheses reach statistical significance, but in Johannesburg one does. In particular, those who reported voting in the last national election were more likely to see the state as legitimate. The positive relationship here is in line with my expectations and confirms previous studies that have argued that individuals who participate in the political system will be more likely to see it as legitimate. Finally, we examine the impact of social structure on perceptions of state legitimacy. We find that age matters in the context of Cape Town and Johannesburg, but that its impact pulls in different directions in the two cities. In Cape Town, older individuals are more likely to see the state as legitimate, but in Johannesburg the opposite is true. In the case of older individuals seeing the state as legitimate, this could perhaps be attributed to their coming of age during a time when people were taught to acquiesce to the state. However, for older people in Johannesburg, it is not exactly clear why they would be less likely to see the state as legitimate. ! 151! Lastly, we come to gender and race. In Durban, males are less likely than females to see the state as legitimate. This is the only city in which the coefficient on gender is statistically significant. With regard to race, Whites and Coloureds are more likely than Africans to hold positive attitudes toward the state in the city of Johannesburg. Historical reasons for the racial divide in legitimacy perceptions have been discussed in earlier sections of this chapter. However some of this gap, particularly with regard to Coloureds, may be driven by the fact that stocks of police 74 legitimacy in particular, have been growing among Coloureds over time . Moreover, Coloureds (63%), more so than any other group included in the sample, rely solely on the police for security provision. Greater levels of state legitimacy may stem from greater levels of contact with representatives of the state, at least where the police is concerned. Interestingly, the coefficient for Coloured is only significant in Johannesburg. It is somewhat surprising that this coefficient doesn’t gain statistical significance in the context of Cape Town where Coloureds constitute a plurality of the population. What about the effect of individual reliance on different types of non-state security? Does reliance on market versus societal forms of non-state security make a difference for individual perceptions of state legitimacy? To test this possibility, I ran a difference of means test using a dummy variable for commercial non-state security. The purpose was to examine whether those who rely on commercial security are substantively different in their legitimacy perceptions from those that rely on societal forms of non-state security such as street committees, neighborhood !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 74 Data from Afrobarometer show that among Coloureds, those who “agree” or “strongly agree” that the police always have the right to make people obey the law has increased from 57% in 2002 to 71% as of 2008 ! 152! watch schemes and community police forums. The difference of means was significant at the .10 level, showing that individuals who rely on commercial security are more likely than those who rely on societal security to see the state as legitimate. This relationship perhaps signals something important about the legacy of state-society relations in South Africa and how different segments of society had different relations with the state. Under apartheid, the commercial security industry worked very closely with the state, often supplementing its understaffed public police force. In fact, the commercial security industry was given almost parallel policing powers under apartheid and some would argue that the distinction between the two became increasingly blurred during that time (Brogden and Shearing 1993). Because this formal and institutionalized relationship between the state and the commercial security industry persists, individuals might more readily credit the state with leveraging security resources from the security industry. In contrast to the commercial security industry, some societal-based security structures such as street committees developed in direct opposition to the state. Because these organizations were directly involved in the struggle for liberation, their relationship to the state was inherently hostile. This legacy, along with the fact that these societal-based groups are relatively more autonomous from the state in present-day, may mean that individuals who rely on this form of security, are less likely than those who rely on the market-based form to credit the state with legitimacy dividends. ! 153! Testing the Effect of Non-State Security Reliance on Perceptions of State Legitimacy Using Afrobarometer Data The results discussed above lend credence to my theory on the effect of security provision on state legitimacy. But how valid and reliable are these results? The data collected during the 20102012 period constitutes a large-N survey with 414 respondents. However, one may want to know how well these findings hold when using a larger sample that is more representative of the country under study. One may also be curious as to how the “non-state security” variable performs when other variables that tap more substantive dimensions of legitimacy are included in the model. The data collected during the 2010-2012 time periods incorporates variables that test for alternative explanations, but most of the variables represent procedural determinants of legitimacy. Therefore, to check the robustness of these results, I triangulate on the same research question and run a similar model using Afrobarometer data. Afrobarometer is the largest, cross-national public opinion survey being conducted in subSaharan Africa. Afrobarometer has been conducting surveys in Africa for over ten years, with the most recent round of surveys being carried out in 35 sub-Saharan African countries. The survey measures citizens’ attitudes toward democracy, civil society and markets in Africa. The most recent round of Afrobarometer data were collected in South Africa from October to November 2011. This data contains a nationally representative sample of 2400 adult South African citizens. Samples of this size yield a margin of error of +/- 2 % and thus provide a sample large enough to test the robustness of the results discussed above. ! 154! The results from the logistic regression model using Afrobarometer data validate support for security explanations of state legitimacy (Table 5.7). As in the pooled results shown above, the coefficient on the non-state security variable is negative. But here it is also statistically significant, suggesting that those who rely on non-state security are less likely than those who rely on public police to see the state as legitimate. Unfortunately, the Afrobarometer survey does not have a question that allows me to gauge whether individuals see the state as a “producer” or “arranger”. Thus, I am not able to test the conditional hypotheses that I propose about the interaction between the non-state security and state arranger variables and their impact on perceptions of state legitimacy. However, the effect of non-state security holds up. Moreover, the effect of victimization also proves to be significant in the Afrobarometer model. In particular, those who report experiencing a contact crime are much less likely to look favorably upon the state. Table 5.7: Afrobarometer Logistic Regression Output On Perceptions Of State Legitimacy, South Africa Independent Variables Non-state security *Contact Victimization *Economic Performance Evaluations Institutional Trust Transgression of Rule of Law *Satisfaction with Democracy Political Participation Age ! Perceptions of State Legitimacy -.354*** (.102) -.246*** (.093) -.121* (.075) .238*** (.064) .219*** (.049) .212*** (.055) -.003 (.115) .007** (.004) 155! Table 5.7 (cont’d) Male -.202** (.093) .060 (.062) Lived Poverty Black (reference) White -.098 (.171) Coloured -.067 (.160) Asian .088 (.277) Employed -.226** (.101) Education .008 (.035) *The question wording for these indicators differ from the question wording for indicators in Table 1 N=2400 Standard errors in parentheses Statistical significance denoted by ***=p≤ .01,**= p≤ .05, *=p≤.10 Constant: -.513* (.297) R2: .041 -2 Log likelihood: 2688.562 Moving to a test of alternate hypotheses, the Afrobarometer model differs slightly with the original pooled model. Here, institutional trust proves to be significant. The positive sign on this coefficient suggests that as institutional trust improves, individuals are more likely to see the state as legitimate. This finding is in line with previous studies that have stipulated the importance of trust in institutions in spurring legitimacy attitudes. Like the original model, the measure for the rule of law is also statistically significant and positive, suggesting that as the rule of law is transgressed, individuals are more likely to see the state as legitimate. Satisfaction with ! 156! 75 democracy also proves to be significant in this model . Importantly, “economic performance 76 evaluations” , which was included to tap into the more economic bases of legitimacy, is not significant. Since my original model primarily measured the supply of political goods, I thought it important to account for the affect of substantive economic goods in this model. This variable, however, fails to reach statistical significance. This means that, at least in current day South 77 Africa, the legitimacy of the state largely hinges on the perceived supply of political goods . Finally, of the social structure variables, age, gender and employment status prove to be significant. Specifically, older people, women and the unemployed are more likely to see the state as legitimate. We may expect older individuals and women to be more likely to defer to authority and thus more likely to see the state as legitimate, yet it is not exactly clear why the unemployed would see the state as more legitimate. After all, we might expect that those who are unemployed would have more grievances against the state. But perhaps this has something to do with the oft-proposed taxation-representation link, and suggests that taxpayers (the employed) perhaps see institutions of the state such as the tax agency as illegitimate because it takes their !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 75 The original model asks about support for and NOT satisfaction with democracy. Support for democracy was tested here, but was not statistically significant. 76 This variable is measured using the following four questions: “How well or badly would you say the current government is handling the following matters, or haven’t you heard enough to say: Creating jobs? Keeping prices down? Narrowing gaps between rich and poor? Ensuring everyone has enough to eat?’ I used the mean of the four items to create an index of economic performance (values 0-3), high numbers being more positive assessments of government performance. 77 In another paper that I wrote using Afrobarometer data, I compared the effect of political goods and economic goods on perceptions of state legitimacy. There I also found that political goods carried more weight than economic ones. ! 157! revenue, but in turn does not satisfactorily provide goods and services or represent the interests of this particular constituency. One may, however, wonder whether these results hold when accounting for higher level units of analysis. Therefore, I rerun the model displayed in Table 5.7 with provincial fixed effects. I find that the results are similar with regard to the sign and statistical significance of most 78 coefficients . Moreover, individuals in practically every province are more likely than those who reside in the Western Cape to see the state as legitimate. Recall in an earlier section I show that a strong majority of individuals (69%) who reside in the city of Cape Town see the state as under the control of the ANC government (Table 5.2). Because Cape Town (and the Western Cape province more generally) is ruled by the opposition Democratic Alliance party, I suggest that the lower percentage of individuals who see the state as legitimate in this city may be tied to the fact that so many individuals believe that institutions of the state are controlled by the ruling party, and that those who do withdraw legitimacy dividends from the state as a result. It seems as if this explanation may also carry weight in explaining variation in perceptions of legitimacy across provinces. TABLE 5.8: Afrobarometer Logistic Regression Output on Perceptions Of State Legitimacy, South Africa (Fixed Effects Model) Independent Variables Non-state security *Contact Victimization Perceptions of State Legitimacy -.337*** (.108) -.251*** (.096) !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 78 A few exceptions include age, which becomes significant at the .01 rather than .05 level. Education loses some of its statistical power, becoming significant at the .10 rather than the .05 level. Finally, the sign on the “White” and “Coloured” coefficients becomes positive, though these variables remain statistically insignificant.!!! ! 158! Table 5.8 (cont’d) *Economic Performance Evaluations -.133* (.079) .241*** (.067) .211*** (.052) Institutional Trust Rule of Law *Satisfaction with Democracy .269*** (.058) -.001 (.004) .010*** (.004) -.207** (.096) .059 (.065) Political Participation Age Male Lived Poverty Black (reference) White .075 (.179) .280 (.182) .080 (.285) -.180* (.104) .018 (.037) Coloured Asian Employed Education Western Cape (reference) Eastern Cape 1.422*** (.203) .817*** (.248) .985*** (.176) .963*** (.190) 1.669*** (.240) .758*** (.237) -.212 (.242) Free State Gauteng KwaZulu-Natal Limpopo Mpumalanga Northwest ! 159! Table 5.8 (cont’d) Northern Cape 1.040*** (.351) *The question wording for these indicators differ from the question wording for indicators in Table 1 N=2400 Standard errors in parentheses Statistical significance denoted by ***=p≤ .01,**= p≤ .05, *=p≤.10 Table 5.8 (cont’d) Constant: -1.664*** (.351) R2: .092 -2 Log likelihood: 2573.776 The results from the Afrobarometer model therefore validate the importance of security provision in shaping perceptions of state legitimacy. However, these results also suggest that procedural dynamics of legitimacy are important too, particularly those elements that focus on 79 trust and satisfaction with democracy. As my city level results have pointed out , the effect of these variables may change depending on region. The Afrobarometer data used in this analysis contain a representative sample the draws respondents from the entire country, whereas my sample is restricted to three cities. The larger geographical coverage of Afrobarometer may help to explain the disparity in results. CONCLUSION This research has tapped into a key dimension of the micro-foundations of statehood. While much of the literature on the state has asserted the provision of security as a key characteristic of !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 79 For example, the “support for democracy” and “lived poverty” variables are only significant in Johannesburg. ! 160! statehood, this research has tested the importance of this criterion for ordinary citizens. In particular, it has shown that ordinary citizens see the provision of security as a key criterion of legitimate statehood. Moreover, the results presented here move beyond conventional works that have simply looked at the importance of state security provision in legitimating the state-society relationship. Interestingly, these results highlight the importance of non-state actors in shaping individual attitudes toward the state. My research shows that individual reliance on non-state security can boost perceptions of state legitimacy, but only when individuals see the state as an arranger of that good. The positive direction of this relationship holds true in Cape Town and Johannesburg. However in Durban, this relationship is negative (though not statistically significant). This means that, contrary to my expectations, those who rely on non-state security and see the state as an arranger are less likely to see the state as legitimate in Durban. Essentially, citizens reward the state for acting as an arranger in Johannesburg, but condemn the state for playing this role in Durban. I attribute this difference to the lower levels of satisfaction with non-state security in Durban as compared to other cities and to the fact that the more perilous security situation in Johannesburg (at least in the case of contact crime) causes citizens to laud supplementary security provision. At least in Johannesburg, it seems as if citizens still expect the state to participate actively in the production of security; but they do not expect the state to be omnipotent. Instead, citizens in this context will improvise with state limitations by supplementing state security provision with non-state security services. ! 161! In conclusion, this research demonstrates that attitudes toward the state depend not just on whether key goods like security are provided, but also by whom they are provided. It shows that actors from the non-state sphere have an important role to play in shaping the relationship between state and society, and that their services may strengthen or undercut perceptions of state legitimacy depending on how people view the proper role of the state in service provision. ! 162! CHAPTER 6 EXPLAINING THE RELATIONSHIP BETWEEN SECURITY AND POLITICAL PARTICIPATION ! 163! INTRODUCTION The previous chapter examined how individual’s reliance on non-state security shaped their political attitudes. Specifically, it found that those who rely on non-state security may be more or less likely to see the state as legitimate, depending on what they perceive the proper role of the state to be in service delivery. That chapter therefore established the key role that non-state actors may play in shaping popular perceptions of state legitimacy. This chapter shifts to a focus on political behavior and examines the ways in which security factors influence individuals’ political participation. I explore two key security variables and their affect on participation. First, I investigate the effect of non-state security reliance on political participation. Does an individuals’ choice of security provider (non-state versus state) shape whether and how they participate in the political sphere? In essence, this question asks whether the non-state provision of a key public good (security) will cause individuals to withdraw from or engage more fully with other citizens and the state. The second security factor is victimization. This chapter will therefore explore how becoming a victim of crime shapes individuals’ willingness to be a part of political associations and processes. Following Bratton et al. (2005), I examine five types of political participation: joining, collective action, contacting, protesting, and voting(Bratton et al. 2005). I want to know not only whether non-state security and victimization have an impact on political participation, but also on what types. Do, for example, non-state security and victimization only matter for electoral participation, or do they also hold consequences for how individuals participate between elections? ! 164! While considering other security factors in addition to non-state security and victimization, I focus most heavily on these two considerations as key explanatory factors of participation. I emphasize the affect of non-state security because very little is known about the political consequences of non-state security provision in South Africa, even though non-state security usage is widespread in this country (See chapter 3). Studies by political scientists and donor organizations have probed the political consequences of the non-state provision of health, education, and water, but there is virtually no work that examines how the non-state provision of security shapes political behavior (Batley and Mcloughlin 2009; Moran and Batley 2004; Rose 2007; Sacks 2012). Scholars have often implied that the extent and quality of state provided goods (including security) could impact citizens’ political behavior and attitudes (Levi et al. 2009) . But it is a completely open question as to how the non-state provision of goods would shape individuals’ political activity. Likewise, victimization becomes important in the context of South Africa because of the widespread nature of crime and violence found here (See chapter 3). But again, most studies do not examine the affect of reported victimization, but rather focus on how perceptions of crime shape individuals’ attitudes and behavior (Lemanski 2004; Louw 1997; Louw). Therefore, it seems both timely and relevant to investigate the political consequences of reported victimization in South Africa. Chapter 5 showed that individuals who report being a victim of a contact crime are less likely to see the state as legitimate. This chapter will explore whether this factor also has an impact on political behavior. ! 165! EXISTING EXPLANATIONS OF POLITICAL PARTICIPATION The foundational work on political participation originated primarily in the context of the United States and focuses heavily on the importance of social structure (Leighley 1995; Nie et al. 1969; Rosenstone and Hansen 1993; Verba and Nie 1972; Schlozman and Brady 1995). These works, beginning with that of Verba and Nie, stress the importance of socio-economic status as a key determinant of political participation. Underpinning the socio-economic approach, is the idea that resources are key to individuals’ political participation, and that the more resources individuals have (in the form of time, skills and finances), the more likely they will be to engage politically. The most powerful and consistent finding with regard to socio-economic status has been that individuals with higher levels of education are more likely to participate in public life (Almond and Verba). Moreover, not only does the level of education matter for participation, but also the type of education that one receives. Recent scholars have shown that informal modes of education, as well as education that focuses on citizenship and civic engagement, have a positive effect on political participation(Kuenzi 2006; Print 2007). In addition to education, scholars have focused on other facets social structure in order to account for variation in political participation, including age, gender and ethnicity, among others. The general consensus in both the developing and developed world is that older people are more likely than younger individuals to engage in politics (Henn et al. 2002; Print 2007; Plutzer 2002; Resnick and Casale 2011). While there is agreement that younger people are less likely to ! 166! participate, the reasons that scholars give for younger peoples’ lower levels of political participation are varied. Some suggest that younger people are simply less interested in formal, conventional modes of political participation such as voting (Henn et al. 2002) or lack the political knowledge and other important resources necessary to effectively participate (Print 2007). With regard to the effect of ethnicity and gender on political participation; there is a huge literature that could not be fully explored here. Studies of gender have many different foci, but several point to the greater obstacles that women face in participating in politics and gaining meaningful representation (Burns et al. 2001; Hirschmann 1991; Iwanaga 2008; Lowndes 2004; Matland 2005). Similar arguments have been advanced to explain lower levels of participation among ethnic and racial minorities in the United States , though some note that members of these groups may be more likely than dominant groups to participate under certain circumstances; i.e. when there is a politicized sense of group consciousness (Miller et al. 1981; Uhlaner et al. 1989). Perhaps nowhere is the effect of ethnicity on political participation more deeply felt than in developing countries. In Africa in particular, ethnicity is a key driving force behind politics, shaping whether and how citizens participate in political life(Bratton et al. 2012; Bratton and Kimenyi 2008; Dunning and Harrison 2010; Ferree 2006; Kramon 2009; Norris and Mattes 2003). But while ethnicity plays a major role in Africans’ participatory considerations, it is worth noting that these citizens also rely on performance evaluations to inform their electoral decisions. Recognizing that gender and ethnicity may sometimes prohibit equal participation , there has been lots of discussion on electoral engineering and how the design of institutions may ! 167! encourage greater participation and inclusion in the political arena (Lijphart and Waisman 1996; Norris 2004; Reilly 2001). Many scholars have proposed that the structure of the electoral systems matters, with proportional representation systems providing greater points of political access for those from disadvantaged groups. Aside from social structure, other factors have been known to influence citizens’ political participation including partisanship, political knowledge and social capital. Partisan ties have been shown to have a particularly strong effect on voting behavior in developing and developed countries alike (Bartels 2000; Bratton et al. 2012; Campbell et al. 1986; Cho and Bratton 2006; Norris and Mattes 2003). In both contexts, individuals who express a closeness to or affinity for a political party are more likely to vote for that party on election day. However, attachments to a political party are not enough to determine citizens’ participation. In a democratic society, we assume that access to information is vital if citizens are going to make informed participatory choices. Research therefore shows that those who are politically knowledgeable and who gain access to information via media exposure are more likely to engage in political participation (Howe 2006; McLeod et al. 1999). The extent to which citizens feel they can trust other citizens and political elites and the extent to which citizens are embedded in networks of reciprocity also influences citizens’ political participation. Strong networks that emphasize reciprocity and political and social trust form the foundation of what we know as social capital (Putnam 2001; Putnam et al. 1994). Social capital scholars have shown the vital importance of trust and associational membership in spurring electoral and non-electoral forms of participation (Kaase 1999; Paige 1971; Platt 2008; Portney ! 168! and Berry 1997; Morales and Giugni 2011; Wagle 2006). For example, scholars have shown that where trust is low, individuals are more likely to participate in contentious forms of political behavior such as rioting (Kaase 1999; Paige 1971), while high levels of trust and associational membership have a positive relationship to more traditional forms of political participation such as voting (Platt 2008; Wagle 2006). Finally, one of the largest bodies of literatures explaining political participation is the literature on economic voting (Anderson 2000; Lewis-Beck and Paldam 2000; Powell Jr and Whitten 1993). The whole thrust of the economic voting literature is that citizens will punish incumbents during bad economic times, but reward them during good ones (Anderson 2000; Lewis-Beck and Paldam 2000; Powell Jr and Whitten 1993). This work has shown that socio-tropic economic considerations are generally more powerful determinants of voting behavior than egocentric (pocketbook) ones. In sum, the participation literature speaks to the many individual level factors that influence participation, but largely fails to account for individual’s micro-level security experiences. This omission is ironic, given that the emergence of the modern state is predicated on the promise of ensuring security for all. Moreover, there is a vast literature on how individuals’ evaluations’ of the state’s performance on the delivery of economic goods shapes political participation. Yet there has been scant attention to the behavioral consequences of the provision of political goods like security. A few exceptions, however, are in order. ! 169! Extant Literature on the Link Between Security and Participation The vast majority of literature explaining political participation fails to centrally examine the importance of security factors. However, there are a few works that examine this connection. For example, some scholars have examined the effect of victimization and public insecurity on political participation(Blattman 2009; Pérez 2003). Blattman, in particular, shows that excombatants who were conscripted via abduction in Uganda, were much more likely to be engaged politically after the conflict ended. Essentially, those who were abducted were much more likely to vote and to engage in community leadership, suggesting that victimization encourages political participation. In addition, others have shown how fear and feelings of public insecurity can suppress political behavior (Pérez 2003; Salamon and Van Evera 1973). Salamon and Van Evera (1973) demonstrate how fear of physical and economic reprisal suppressed Black political participation in the South during the 1960’s, while Perez working in the context of El Salvador and Guatemala has shown how public insecurity increases support for authoritarianism (and, by extension, decreases support for democracy and democratic practices). Thus, there has been some work which has looked at the impact of security related factors on participation, but not much. The work by Blattman in particular is especially instructive, but it focuses on experiences of violent political conflict whereas my interests lie in understanding the political consequences of everyday forms of crime and violence. His work therefore begs the question of how victimization during more ordinary political times would shape political behavior. The work that has addressed this question most directly and thoroughly is the research by Bateson (2012). Bateson (2012) examines the effect of crime victimization on political ! 170! participation in countries spanning five continents; Europe, Asia, Latin America, Africa and North America. She finds that there is a worldwide association between crime victimization and political participation(Bateson 2012). In particular, Bateson argues that being a victim of crime 80 makes individuals more likely to participate in all forms of politics included in her study . Importantly, Bateson argues that all crime victimization, whether serious or petty, violent or nonviolent, increases all forms of political participation. Bateson hints at several reasons why victimization might cause individuals to be more politically active, including post-traumatic 81 growth theory, instrumental reasons, and emotional and expressive factors . Bateson’s work presents the first attempt, to my knowledge, to systematically explore the effect of everyday violence on political behavior, and she does so in a way that is far-reaching and compelling. Therefore, I would like to build on Bateson, delving more deeply into how we might account for the relationship between victimization and participation, but also generating a security-driven model of participation that accounts for other security concerns, most notably, the impact of non-state security provision. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 80 Forms of political participation explored by Bateson include including discussing politics, interest in politics, and participation in protests and community meetings. 81 Post-traumatic growth theory: argues that personal growth and development result from traumatic experiences. This theory usually only applies to major life crises or experiences with very serious crimes that cause trauma. Instrumental reasons: The idea here is that citizens may organize to bring pressure to act on their particular cases. Emotional/expressive factors: Focuses on how becoming activists helps in healing process (emotional), and with recasting identity, i.e. people transition from being victims to victors (expressive). ! ! 171! THEORIZING THE RELATIONSHIP BETWEEN SECURITY AND PARTICIPATION Non-State Security and Political Participation Given that little is known about the behavioral consequences of non-state security provision, I develop a theory to account for how individual reliance on non-state security might shape political participation. MacLean (2011) provides a useful starting point for the development of this theory(MacLean 2011). MacLean (2011) seeks to understand how the retrenchment of the state in Africa shapes political behavior. In particular, she explores the effect of the frequency of individual reliance on public education and health services on citizens’ political participation. MacLean examines electoral and non-electoral forms of participation, including voter registration and turnout, political contacting and joining. Using Afrobarometer survey data from 18 countries, she finds that individuals who report having more frequent experience with public schools and clinics are more likely to participate in both electoral and non-electoral political activity. According to MacLean, the mechanism linking citizens’ experience with public services and their greater levels of participation is the quality of experience citizens have when making use of these services. Specifically, MacLean argues that “the experience of the declining quality of publicly provided health and education services…mobilized greater citizen participation and engagement” (MacLean 2011, 1256). While MacLean’s article does not directly examine security issues or issues of non-state provision, her theory about state retrenchment could easily be applied to both. If, in fact, individuals who rely on public services are more likely to participate in politics, then, by ! 172! extension, those who rely on non-state services should be less likely to participate. For MacLean, the driving factor behind citizen’s mobilization is their discontent with the quality of public services, which gives them an incentive to lobby elected officials for improved public services. Similarly, many citizens who primarily rely on the non-state actors for their basic needs, do so, in part, because of their dissatisfaction with public provision. The more economically well-off citizens who have the means to purchase commercial services simply remove themselves from dependency on the state rather than lobbying government to improve the quality of public service delivery. As such, their limited contact with the public realm, and the relative convenience and higher quality services received from non-state actors should make these citizens less likely to engage in politics. This is so because their reliance on non-state actors means that government performance essentially has no bearing on their lives or the quality of services that they receive. Therefore, I apply MacLean’s argument about the implications of state retrenchment in Africa to the case of non-state security, expecting that those who rely on non-state sources of security will be less likely to involve themselves in politics. While I expect individual reliance on non-state security its own to suppress political participation, I expect perceptions of state legitimacy to mediate the relationship between nonstate security reliance and political participation. Several scholars have shown that perceived state legitimacy has important behavioral consequences such as inducing obedience, compliance and cooperation with the state and government (Gibson and Caldeira 2003; Levi et al. 2009; Mondak 1993; Tyler and Fagan 2010). More directly, there has been a demonstrated link between legitimacy and various forms of political participation in Costa Rica (Booth and ! 173! Seligson 2005). Therefore, I would expect the positive effect of non-state security reliance on participation to only be realized when we account for individual views of a legitimate state. But why would the positive effect of non-state security reliance only be realized when accounting for state legitimacy? Recall in chapter 5, that I show that individual reliance on nonstate security has a positive effect on perceptions of state legitimacy when individuals view the state as an arranger of security. In that same vein, I also expect non-state security to impact individuals’ decision to participate in political life. Essentially, I argue that state legitimacy should mediate the relationship between non-state security and participation because individuals who rely on non-state services should only feel compelled to engage with the state when they feel that it still possesses stocks of legitimacy. Put differently, if individuals’ reliance on nonstate security does not undercut their legitimacy perceptions, then citizens may still find it worthwhile to engage with the state, even if it is not directly producing security for them via an effective public police force. In fact, those who rely on non-state security and continue to see the state as legitimate may do so, in part, because they see the state as effectively leveraging nonstate security services (see chapter 5). While I generally expect those who rely on non-state security and see the state as legitimate to be more likely to participate, I maintain that these individuals may be less likely to engage in contentious forms of political behavior like protest. ! 174! Figure 6.1: The Effect of Non-State Security and State Legitimacy on Political Participation Reliance on Non-State Security State Legitimacy High Political Participation Low Political Participation Victimization and Political Participation When it comes to the relationship between victimization and participation, I generally expect that victimization will make individuals more likely to engage in political acts. However, I argue that individuals’ decisions to join associations and participate in collective action will be contingent on trust. Participation in associational life is about citizens working together through organizations to form and address their collective interests. Because this form of participation (along with collective action) requires extensive interaction with fellow citizens, issues of trust become important and perhaps key to understanding the relationship between victimization and participation. For example, crime victimization has been shown to suppress trust (Delhey and Newton 2003). Therefore, we may expect citizens who have been victims of crime to be less trustful of others, and, in turn, less likely to join in local organizations and engage in collective action with them. The extremely important work by Putnam (1994) has shown the significance of trust and other forms of social capital for democracy. In this same vein, I suggest that trust becomes an essential element for developing and sustaining a key feature of modern ! 175! democracies; political participation. As Putnam notes, “the greater level of trust in a community, the greater the likelihood of cooperation.” (1994, 171). In a deeply divided country like South Africa, I would argue that it is not only important to explore whether trust exists, but also the locus of that trust. I suggest that understanding this factor will be important for whether victimization will lead people to be more or less likely to join and engage in collective action. Putnam hints at the importance of the locus of trust in his discussion of networks and whether they are bridging or bonding. Importantly, he notes that segregated networks (bonding) foster cooperation within the group. Because of the racially segregated nature of life in South Africa, it would seem that most organizations would be of the bonding type. However, I suggest that intra-group and intra-community trust is not a given, and that this type of trust must be present for individuals to join local associations and engage in collective action. To illustrate the importance of the locus of trust, let’s examine a public opinion question about criminals. When asked where they believe criminals reside, many White South Africans report that criminals reside outside their communities, whereas Africans largely report that criminals 82 reside within their communities . So those who feel as if criminals come from within their communities (and therefore may have less trust in neighbors) may be less likely to join in local organizations and engage in collective action. Conversely, those who believe that crime comes !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 82 According to the 2007 Victims of Crime Survey, 46% of African respondents feel that property crime is committed by people living in their area, compared to 16% percent of Whites who report the same. The majority of Whites feel that property crime is committed by people living outside their area (53%). Similarly, the majority of Africans feel that violent crime is committed by people living in their area (50%), while a majority of Whites feel that violent crime is committed by people living outside their area (57%). ! 176! from outside their communities (and are therefore more trusting of their neighbors) might be more likely to band together in organizations to keep criminals out. Therefore, I expect social trust to mediate the relationship between victimization and joining/collective action. HYPOTHESES Joining and Collective Action Hypotheses Given my theories about non-state security provision and victimization, I develop the following conditional hypotheses about individual’s willingness to join and participate in collective action: Hyp. 1: Individuals who rely on non-state security and see the state as legitimate will be more likely to join a voluntary association Hyp. 2: Individuals who rely on non-state security and see the state as legitimate will be more likely to engage in collective action Hyp. 3: Individuals who report being a victim of crime (contact and property) and trust their neighbors will be more likely to join a voluntary association Hyp. 4: Individuals who report being a victim of crime (contact and property) and trust their neighbors will be more likely to engage in collective action ! 177! Contacting and Voting Hypotheses As opposed to joining and collective action, which requires contact between citizens, contacting and voting explicitly speak to citizen-elite linkages. Extant literature states that during bad times, citizens are more likely to contact their elected leaders and more likely to punish incumbents at the polls (Hero 1986; Lewis-Beck and Paldam 2000; Sharp 1982). Therefore, individuals who have fallen prey to victimization may be more likely to engage politically. This suggests an instrumental motivation for these types of political participation, one of the potential motivators that Bateson points to in her work. I also expect that those who rely on non-state security and see the state as legitimate will be more likely to contact and vote. In terms of the relationship between security and contacting and voting, I expect the following: Hyp.5: Individuals who rely on non-state security and see the state as legitimate will be more likely to contact Hyp. 6: Individuals who rely on non-state security and see the state as legitimate will be more likely to vote Hyp. 7: Individuals who report being a victim of crime (contact and property) will be more likely to contact ! 178! Hyp. 8: Individuals who report being a victim of crime (contact and property) will be more likely to vote Protesting Hypotheses Finally, we come to protesting, the most contentious of all forms of political behavior discussed thus far. Following a similar logic from contacting and voting, my expectations are that: Hyp. 9: Individuals who rely on non-state security and see the state as legitimate will be less likely to protest Hyp. 10: Individuals who report being a victim of crime (contact and property) will be more likely to protest In sum, my general expectations are that individual reliance on non-state security will increase individual political participation when individuals view the state as legitimate. I also generally expect crime victimization to increase political participation, though, again, with joining and collective action, I expect this relationship to be conditional on trust. ! 179! DATA AND INDICATORS The following analyses rely on data from Round 5 Afrobarometer 83 surveys carried out from October-November 2011 in South Africa. Afrobarometer is the largest, cross-national public opinion survey in Africa. The survey measures citizens’ attitudes toward social, economic and political issues in their country. All surveys are based on random samples that are nationally representative of the adult population 18 years old or older. The South Africa survey consisted of a sample of 2400 such citizens. Samples of this size yield a margin of error of +/- 2%. In describing the measures employed in this study, I first begin with the dependent variables of interest. Political Participation One of the ways that citizens participate as active members of society is through joining organizations. In Africa, joining organizations has been an important way of exercising political voice, with political activity often being organized through key organizations such as voluntary associations, trade unions and churches. Because this dissertation is chiefly concerned with issues of security, ideally I would examine the extent to which individuals join societal-based security organizations. In the absence of such a measure, I employ the closest proxy that asks individuals the following: “Now I am going to read out a list of groups that people join or attend. For each one, could you tell me whether you are an official leader, an active member, an inactive member, or not a member: Some other voluntary association or community group? I dichotomize !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 83 www.afrobarometer.org ! ! 180! this variable, coding as 1 those who say that they are “an official leader” or “an active member” and coding as 0 those who report that they are “an inactive member” or “not a member”. Next we explore the extent to which citizens connect with one another by engaging in collective action. Here I combine two questions that tap the extent to which citizens fellowship with each other over matters that are important to them as a community. The questions are: “Here is a list of actions that people sometimes take as citizens. For each of these, please tell me whether you, personally, have done any of these things during the past year. If not, would you do this if you had the chance: 1) Attended a community meeting? and 2) Got together with others to raise an issue? Values range from 0 (“no, would never do this”) to 4” (yes, often). I created a construct labeled “collective action” by taking the mean of these two variables. The form of participation referred to as contacting, investigates the extent to which citizens engage with public leaders. I utilize two separate measures of contacting; one to gauge the extent to which individuals contact local government councilors, and the other to examine the extent to which individuals contact members of parliament. The questions are as follows: “During the past year, how often have you contacted any of the following persons about some important problem or to give them your views: 1) A local government councilor, and 2) A Member of Parliament?”3) Values on these questions range from 0-3, with high scores representing greater frequency of contact. I analyze these forms of contacting separately as citizens may contact officials at different levels of government depending on the issue at hand. Thus, by analyzing these measures of contacting separately, we will have a better sense of not only whether crime and security issues cause individuals to contact, but also whom. ! 181! Of the forms of participation examined in this chapter, protesting perhaps represents the most contentious form. I created an additive index to gauge the extent to which individuals go on strike, participate in demonstrations or protests, or even engage in political violence. The specific questions used to create this index are as follows: “Here is a list of actions that people sometimes take as citizens. For each of these, please tell me whether you, personally, have done any of these things during the past year. If not, would you do this if you had the chance: 1) Refused to pay for services provided by government like water, electricity or property rates?, 2) Refused to pay a tax or fee to government?, 3) Attended a demonstration or protest march?, 4) Gone on strike in order to demand a higher salary or better working conditions?, and 5) Used force or violence for a political? Values range from 0-4, with four indicating that individuals have undertaken these actions “often”. While it may be clear to see how questions three through five tap into more contentious forms of political behavior, questions one and two may deserve further explanation. In addition to protests and the use of political violence, South Africans often used tax, rent and rate boycotts as a form of protest under apartheid. In fact, non-payment of fees was a strategic form of political participation under apartheid, one whose intention was to cripple the economic strength of the apartheid state. Thus, in South Africa, there is a long legacy of withholding payments of fees in order to make a political statement. Therefore, I examine these forms of participation along with strikes, protests and political violence. My theoretical reasons for exploring these five activities together is empirically supported (one unrotated factor explains 61% of the variance (eigenvalue=3.166), alpha=.850). ! 182! Finally, I examine electoral behavior. In particular, I employ a question that asks respondents the following: “With regard to the most recent national election in 2009, which statement is true for you?” Response categories consist of “You were too young to vote”, “You were not registered to vote”, “You voted in the elections”, “You decided not to vote”, “You could not find the polling station”, “You were prevented from voting”, “You did not have time to vote”, “You did not vote because you could not find your name in the voters’ register”, and “Did not vote for some other reason”. I code as 1 those who reported voting in the last election and all others as 0. Determinants of Political Participation The main explanatory variables of interest are non-state security and victimization. I measure individual reliance on non-state security by using the following question: If you were a victim of crime in this country, who, if anyone, would you go to first for assistance?” Response categories include “the police”, “a security service or security company that you pay for”, “a traditional leader or traditional court”, “a street committee or local security organization”, “a powerful local person or local gang”, “you would personally take revenge”, “you would join with others to take revenge”, “your own family or friends”, and “the family of the perpetrator”. I code as 0 those who report that they would go to the police and all others as 1. Therefore, my measure of nonstate security is a broad one that includes both market and societal forms (See chapter 3). Because I argue that state legitimacy may mediate the relationship between non-state security and political participation, I also include a measure of state legitimacy. To stay consistent with ! 183! the coding of this variable from the last chapter, I first develop a composite measure (index using the mean of three variables) of state legitimacy from the following three questions: “Please tell me whether you agree or disagree: 1) The police always have the right to make people obey the law, 2) The courts have the right to make decisions that people always have to abide by, and 3) The tax department always has the right to make people pay taxes. Original values ranged from 0 (“strongly disagree”) to 4 (“strongly agree”). Because I am simply interested in those who see that state as legitimate versus those who do not, I then create a binary variable with values <3=0 84 and values equal to or >3=1 . I measure two types of victimization; contact and property victimization. My measure of contact victimization asks: “During the past year, have you or anyone in your family: Been physically attacked?” My measure of property victimization asks” During the past year, have you or anyone in your family: Had something stolen from your house?” The contact victimization variable is essentially gauging whether respondents or someone in their families have been assaulted in the past year, whereas the property victimization measure assesses whether respondents or someone in their families have been victims of theft. I dichotomize these variables, coding as 0 those who respond “no” to those questions and coding as 1 those who respond “yes, once”, “yes, twice”, and “yes, three or more times”. Because I argue that social trust may mediate the relationship between security and collective action and joining, I also include a measure of social trust that asks: “How much do you trust each of the following types of people: Your neighbors?” I also create a binary variable here, coding as 0 those who say they trust their neighbors “not at all” or “just a little” and coding as 1 those who say they trust their neighbors “somewhat” or “a lot”. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 84 I also ran the regression analyses with a continuous measure of state legitimacy and the results were similar. ! 184! In addition to the above security factors, I control for two additional security considerations; personal insecurity and evaluations of government performance on the provision of political order. These are important considerations as they may directly impact an individuals’ decision to participate, particularly between elections. For example, individuals may be less likely to attend community meetings and join associations that meet after work when they feel it is not safe to go outdoors after dark. Perez (2003) has shown that greater levels of insecurity decrease support for democracy. Similarly, I would expect higher levels of insecurity to have a negative effect on political participation. The personal security indicator asks: “Over the past year, how often, if ever, have you or anyone in your family: Felt unsafe walking in your neighbourhood?” Values range from 0 (never) to 4 (always). For evaluations of political order, I utilize a question that asks citizens how well they believe government is reducing crime. Values range from 0 (very badly) to 3 (very well). In addition to the security-related variables, I also control for other factors that are known to influence political participation, including economic performance evaluations, personal efficacy, partisanship, political sophistication, media exposure and social structure. Scholars have long shown that the provision of economic goods (i.e. a strong economy) have a substantial influence on voting. Therefore, we should expect economic evaluations to have an impact on whether and how individuals participate. To measure performance on economic goods, I create an index from four items that asks individuals how well the government is doing with inflation, providing jobs, providing food, and narrowing the gap between the rich and poor. Values range from 0-3, with higher numbers representing better performance on these issues. ! 185! Next, I create a measure of political efficacy. Existing research has shown that individuals will be more likely to participate in politics when they feel elected officials are responsive to their demands (Finkel 1985). Therefore, I include a construct of political efficacy using the mean of the following two indicators: How much of the time do you think the following try their best to listen to what people like you have to say: 1) Members of Parliament, and 2) Local government councilors?” Values range from 0 (never) to 3 (always). I also control for partisanship, which has been shown to be one of the most important predictors of voting behavior (Bartels 2000; Campbell et al. 1986; Bratton et al. 2012) I code those who identify with the ruling party as 1 and those who identify with an opposition party as 0. The extent to which citizens participate in political life has also been shown to be influenced by their levels of political sophistication and access to information (Kenski and Stroud 2006; McLeod et al. 1999).Two different measures of political sophistication are explored here. The first asks: “When you get together with your friends or family, would you say you discuss political matters: Frequently, occasionally, or never?” Values range from 0 (“never”) to 2 (“frequently”). The second measure asks: “How interested would you say you are in public affairs?” (0=not at all interested; 3=very interested). I measure media exposure by creating an additive index from the following four indicators: “How often do you get news from the following sources: 1) Radio, 2) Television, 3) Newspapers, and 4)Internet? Values range from 0 (never) to 4 (everyday). ! 186! Finally, I control for the affect of social structure by including variables for age, gender, race, 85 urban/rural residence, level of poverty, employment, and education . EXPLAINING THE SECURITY-PARTICIPATION LINK Non-State Security, State Legitimacy and Political Participation In an attempt to illuminate the relationship between security factors and participation, I first examine the relationship between the non-state security and state legitimacy interaction term and various forms of political participation. First, the interaction term fails to reach statistical significance in all but one of the participation models explored (voting). Though this variable fails to reach statistical significance in most instances, it generally runs in the hypothesized direction. That is, there is a positive relationship between seeing the state the interaction term and joining, engaging in collective action, contacting, and even protesting (though I expected a negative relationship here). This suggests that individuals who rely on non-state security and see the state as legitimate will be more encouraged to connect with fellow citizens and elected officials between elections. Only in the case of voting does the interaction tem reach statistical significance. Moreover, the relationship between the interaction variable and voting is unexpectedly negative. Interestingly, !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 85 See Appendix G for full question wording and coding. ! 187! those who see the state as legitimate and rely on state security 86 are more likely to vote. But those who see the state as legitimate and rely on non-state security are less likely to vote. At least in the specific case of electoral participation, this suggests that individuals will withdraw from the state when they must turn to sources other than the state for their security needs, even if they continue to see the state as legitimate. In other words, legitimacy attitudes on their own are not enough to keep state-society relations afloat in the electoral arena. In order for citizens’ to engage in electoral politics, they must feel not only that the state is legitimate, but also that this legitimate state is providing the goods. Aside from the interaction term, we see that the state legitimacy variable has a negative and statistically significant impact on collective action and protest behavior. This suggests that when individuals rely on state security and see the state as legitimate, they are less likely to engage in contentious forms of political activity. Seemingly, the role that citizens’ perceive the state to play in security provision matters not only for political attitudes, but also for political behavior. The results indicate that the direction of the relationship between state legitimacy and participation largely rests on whether individuals’ see the state or some other entity as meeting their security needs. Specifically, security services provided by a legitimate state seems to strengthen individuals’ engagement in electoral politics, but undercut their involvement in contentious forms of political participation (collective action and protest) between elections. However, when citizens see the state as legitimate, but rely on non-state security, this seems to encourage collective action and protest (as evidenced by the positive sign on the state legitimacy/non-state !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 86 The coefficient on the state legitimacy variable represents the coefficient for individuals who see the state as legitimate, when other variables are held constant. Therefore, this is the coefficient for individuals who see the state as legitimate when non-state security =0 (i.e. when individuals primarily rely on state security ( public police) for their security needs). ! 188! security interaction terms in tables 6.1-6.2). It may be that citizens who rely primarily rely on non-state security and still see the state as legitimate are more prone to take to the streets and organize with other citizens to petition for greater state involvement in the provision of security. The Effect of Insecurity and Victimization on Participation The above results have shown that who individuals turn to for security and their views of state legitimacy have an impact on whether and how they participate. Now we turn to a discussion of the relationship between other security considerations and political participation. First, the coefficient on personal insecurity is positive and statistically significant in all cases of political participation except for collective action (positive but not statistically significant) and voting (negative and not statistically significant). This means that individuals who frequently feel insecure are more likely to join associations, contact elected officials, and protest. Feelings of personal insecurity therefore have an important role to play in shaping individuals’ political behavior. It encourages them to bond with other citizens’ in associational forums, but it also encourages them to connect with public officials. Next, we examine the political impact of victimization. To begin with joining, recall that I expect the relationship between this form of political participation and victimization to be conditional on social trust. We find that the coefficient on the interaction term for contact victimization and trust is statistically significant. Contrary to my expectations, however, the sign on the interaction term is negative. This suggests that those who have been a victim of contact crime and trust their neighbors are less likely to join a voluntary association. Interestingly, contact victimization on ! 189! its own makes individuals more likely to join a voluntary association, but when coupled with social trust the effect on joining is negative. Perhaps this has something to do with the relationship between perpetrators and victims. For example, in South Africa-as well as in most places-many assaults happen between people who know each other. Perhaps then, individuals who have good relations with other members of the community are less likely to risk tarnishing those relations or their reputation by focusing on what may be perceived as an unseemly and private affair between family members or friends. This explanation would especially ring true in cases of domestic violence, since a widespread sentiment is that this issue should be dealt with in 87 the privacy of one’s own home . In the case of collective action, none of the victimization variables reach statistical significance. While the signs on the coefficients indicate a positive relationship between victimization and collective action, none of these variables are statistically significant. Victimization therefore does not seem to shape whether individuals engage in collective action with their fellow citizens. However, it may be that the relationship between trust and victimization only matters for individuals’ participation in more structured organizations (i.e. voluntary associations), but that this relationship matters less for activities like community meetings where an individuals’ involvement may be sporadic and not contingent on formal membership status. To account for this possibility, I reran the collective action model excluding the interactions between victimization and trust. When I do this, I find that the coefficient on property victimization is positive and statistically significant. This means that individuals who have had something stolen !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 87 In the 2007 VOCS respondents were asked if they would report a case of domestic violence to the police if they were aware of such a case. Almost 20% of individuals said “no”. When asked why they would not report the crime, the main reason that people gave was “it’s none of my business” (51%). ! 190! from them in the last year were more likely to attend community meetings or join with others to raise issues. With regard to collective action then, property victimization mobilizes collective action, but the same is not true for contact victimization which does not reach statistical significance. 88 Table 6.1: Regression Analyses of Joining and Collective Action , South Africa Independent Variables Joining Collective Action Collective Action (no victimization/trust interactions) Non-state Security Contact Victimization*Social Trust .031 (.314) -.309 (.217) .356 (.394) .802** (.369) .028 (.302) .302 (.217) -1.124** (.532) .033 (.107) -.138* (.074) .135 (.134) .005 (.138) .127 (.100) .123* (.071) -.017 (.187) .037 (.107) -.135* (.073) .132 (.134) .047 (.095) .173*** (.068) .144*** (.059) - Property Victimization*Social Trust .175 (.390) .083 (.134) - Personal Insecurity .147** (.067) .026 (.023) .025 (.023) Political Order -.044 (.108) .003 (.037) .003 (.037) Economic Goods -.136 (.162) -.119** (.055) -.119** (.055) State Legitimacy Non-state Security*State Legitimacy Contact Victimization Property Victimization Social Trust !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 88 This model includes provincial fixed effects, though the coefficients for the provincial dummy variables are not shown. ! 191! Table 6.1 (cont’d) Political Efficacy ANC Partisan Discuss Politics Interest in Public Affairs Media Exposure Age Male Lived Poverty White Coloured Asian Urban Employed Education Constant .140*** (.041) -.047 (.288) .512*** (.146) .168* (.098) .393*** (.119) .018*** (.006) .003 (.174) .027 (.118) -.559 (.438) -.242 (.386) -1.115 (.829) -.505** (.229) .149 (.187) .004 (.069) -4.426*** (.673) .033** (.016) -.075 (.098) .257*** (.050) .108*** (.033) .134*** (.039) .010*** (.002) .006 (.059) .172*** (.040) -1.035*** (.146) -.438*** (.125) -.462** (.235) -.274*** (.067) -.050 (.065) .007 (.024) 1.163*** (.211) N=2400 R square: .117 .253 -2 Log likelihood 948.346 standard errors in parentheses statistical significance denoted by ***=p≤ .01,**= p≤ .05, *=p≤.10 .034** (.016) -.073 (.098) .257*** (.049) .107*** (.033) .133*** (.039) .010*** (.002) .006 (.058) .172*** (.040) -1.034*** (.145) -.436*** (.124) -.459** (.235) -.275*** (.076) -.049 (.065) .008 (.024) 1.149*** (.210) .253 Examining local forms of contacting first, we see that, as with collective action, being a victim of property crime makes individuals more likely to contact a local government councilor. Individuals may feel especially comfortable bringing issues of crime and security to leaders that ! 192! are close to their communities. But is this also true for parliamentarians, government officials with whom citizens may feel less familiar? With regard to contacting members of parliament, both contact victimization and property victimization are positive and statistically significant predictors of contacting, but only when individuals feel they can trust other people from their community. Interestingly, property victimization on its own suppresses the likelihood that one will contact a parliamentarian, but when individuals who have been victims of property crime feel they can trust their neighbors, they are more likely to contact. Contact victimization on its own also has a negative relationship with contacting MPs, though the coefficient is not statistically significant. Thus, the impact of victimization on contacting national government officials is conditional on social trust. The contacting findings raise interesting ideas about how victimization shapes citizen engagement with officials across levels of government. The theoretical discussion above only points to the need for a victimization/trust interaction with regard to joining and collective action, but some works have shown that strong community ties may also have an impact on other forms of participation such as contacting (Davidson and Cotte 2006). In fact, this logic is implied in the literature on social capital which points to how strong horizontal relations between citizens may contribute to strong vertical relations between citizens and elites (Putnam 2001; Putnam et al. 1994). This seems to be the case with citizens who contact MPs. There are a couple of reasons why the effect of victimization on contacting MPs may only be realized when accounting for trust. First, attracting the attention of a national government official is more costly than securing that of a local government official. National officials have larger ! 193! constituencies and thus competition for their attention is steeper. Therefore, individuals who seek the time and attention of MPs may only be willing to pay the costs of taking issues to higher governmental heights when they feel they have the backing of their local community. When social trust is present in a community, citizens feel that they have stronger ties within the community, and a strong network such as this may be needed to motivate victims of crime to approach national level officials. In other words, individuals who are embedded in strong local communities may be more likely to receive the encouragement, motivation and support necessary to make claims on national level public officials. Secondly, when individuals feel close to other members in the community, they may also be more likely to view their victimization as not just an individual problem, but a collective one. Hirlinger (2003) distinguishes between two modes of contacting which he refers to as particularized or general referent contacting. The latter is contacting that individuals do on behalf of a larger community problem. It is not based on narrow, individualized problems like particularized contacting. Hirlinger argues that different explanatory models may be needed depending on the type of contacting that individuals are undertaking. But I argue that it may be possible that who individuals decide to contact may be linked to whether they are dealing with more particularized or general issues. Having a vested interest in community needs may therefore drive individuals to try to push issues of crime and security higher up on the national agenda. This may especially be the case in South Africa where those who are ultimately 89 responsible for dealing with crime operate at the national level . !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 89 South Africa has a national police force. While there are some metropolitan police departments, they are restricted to dealing with minor issues like traffic control. They are not authorized, for example, to conduct criminal investigations. ! 194! Next, we explore the effect of victimization on protest behavior. The contact victimization variable is positive and statistically significant. So, those who report being a victim of a contact crime are not only more likely to engage with elected officials (see results for contacting MPs), but they are also more likely to take to the streets. This suggests a hybrid approach to mobilization on issues of crime and security between elections; on the one hand, individuals who have fallen prey to violent criminals will engage directly with elected officials, on the other hand, they will attempt to express their grievances through more contentious forms of political behavior including the use of violence, demonstrations and withholding of fees and payments. Property victimization, while also exhibiting a positive sign, is not a statistically significant predictor of protest behavior. Finally, we come to electoral participation. Here, the coefficient on contact victimization is negative and statistically significant, meaning that those who have been victims of contact crime are less likely to vote. It seems then that violent crime in particular has a way of suppressing voter turnout and keeping individuals from voicing their political preferences on election day. These results show that criminal and everyday forms of violence have an important impact on citizens’ participation in electoral politics. However, why would contact victimization make people more likely to contact their elected leaders, but less likely to vote? If contact victimization makes individuals more likely to engage with public officials (in this case MPs), shouldn’t it also make them more likely to engage in elections that decide who their representatives will be? One would think that victims of violent crime would not only be more motivated to contact elected officials, but also to go to the polls to punish incumbents for not doing a good job preventing ! 195! crime. I argue that the nature of party competition within South Africa may be the reason why victimization suppresses voter turnout. In a country that has a dominant party (the ANC in South Africa), victims of crime may feel that the likelihood of punishing leaders (i.e. ejecting them from office) is slim to none. Therefore, those who have been victims of violent crime may feel that their time is best spent working with the leaders they have, (as in the case of contacting MPs) instead of wasting energy and time participating in an electoral arena that is not likely to result in changed leadership. 90 Table 6.2: Regression Analyses of Contacting, Protesting and Voting , South Africa Independent Variables Contacting Contacting Protesting Voting LGC MP Non-state Security State Legitimacy Non-state Security*State Legitimacy Contact Victimization Property Victimization Social Trust Contact Victimization*Social Trust Property Victimization*Social Trust Personal Insecurity Political Order .055 (.088) -.029 (.060) .032 (.110) .042 (.078) .129** (.056) -.001 (.049) -.076* (.046) -.031 (.032) .067 (.058) -.055 (.060) -.090** (.043) -.050 (.031) -.088 (.061) -.130*** (.042) .117 (.077) .200*** (.055) .022 (.039) -.040 (.034) .315 (.282) .451** (.203) -.795** (.362) -.713*** (.230) -.092 (.185) .290* (.167) - .152* (.081) .114** (.058) - - - - .060*** (.019) .039*** (.010) .063*** (.013) -.026 (.063) -.035 (.030) .010 (.016) .030 (.021) -.021 (.103) - !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 90 This model includes provincial fixed effects, though the coefficients for the provincial dummy variables are not shown.! ! 196! Table 6.2 (cont’d) Economic Goods Political Efficacy (LGC indicator for LGC model and MP indicator for MP model) ANC Partisan Discuss Politics Interest in Public Affairs Media Exposure Age Male Lived Poverty White Coloured Asian Urban Employed Education Constant -.058 (.045) .166*** (.029) .008 (.024) .130*** (.015) -.026 (.032) .040*** (.009) .008 (.154) .033 (.046) -.149* (.081) .173*** (.041) .049* (.027) .098*** (.032) .010*** (.002) -.007 (.048) .093*** (.033) -.749*** (.120) -.251*** (.102) -.395** (.194) -.011 (.063) .039 (.053) .021 (.020) -.060 (.042) .075*** (.021) -.016 (.014) .054*** (.017) .000 (.001) -.010 (.025) .040** (.017) -.221*** (.063) -.057 (.054) -.213** (.102) .100*** (.033) .041 (.028) .008 (.024) .017 (.056) .045 (.029) .053*** (.019) .061*** (.022) -.001 (.001) .022 (.034) .070*** (.023) -.192** (.084) -.060 (.072) -.056 (.136) .005 (.044) -.008 (.037) -.003 (.014) -.278 (.284) .079 (.136) -.068 (.089) .195* (.105) .049*** (.007) -.235 (.164) -.011 (.111) -.711* (.420) -.342 (.349) .776 (1.026) .109 (.206) .180 (.184) -.009 (.070) -.375** (.173) -.353*** (.092) .205 (.121) -1.292** (.624) N=2400 R square: .183 .162 .146 -2 Log likelihood standard errors in parentheses statistical significance denoted by ***=p≤ .01,**= p≤ .05, *=p≤.10 ! 197! .094 1020.406 In sum, victimization suppresses joining and voting, but encourages collective action, contacting and protesting. Unlike Bateson (2012), I do not find that all forms of victimization have the same affect on all forms of participation. Contact victimization proves to be the driving factor behind most forms of political participation (joining, contacting MPs, protesting and voting). Property victimization only matters in two cases; with regard to contacting local government officials and collective action. Therefore, it seems that individuals are most likely to rely on local government and loose networks of collective action when victimization is property-related. Yet when individuals’ have had their person violated, this affects their willingness to contact national level officials, vote for these officials, protest, and embed themselves in dense and tightly connected associational networks. In particular, those who have been the victim of assault, are less likely to join and vote, but more likely to contact MPs and voice their grievances through protest. This suggests that contact victimization suppresses the likelihood of citizens connecting to each other in meaningful ways (i.e. through associational membership), but it also suppresses the likelihood of citizens engaging in electoral forms of participation, a key way in which citizens express their political preferences and thereby connect with the state. However, all is not lost, as citizens who have been victims of contact crime are also more likely to contact national level elected leaders and engage in protest. This shows that victims of violent crime are willing to use more than one approach to make their voices heard, one formal and institutionalized (contacting), the other informal and uninstitutionalized (protest). ! 198! Other Important Variables In all of the models explored here except one, political efficacy has a positive and statistically significant effect on participation. Specifically, individuals who feel more efficacious are more likely to join, engage in collective action, contact, and protest. Political efficacy does not reach statistical significance in the case of voting (though the sign is still positive). Moreover, media exposure has a positive and statistically significant effect in all models of political participation. This means the more often people gain access to news via the radio, television or internet, the more likely they are to get involved in politics. The findings on media exposure and political efficacy are in line with many studies which have noted the importance of access to information and elite responsiveness in shaping individuals’ political behavior. Turning to a discussion of control variables for specific models, I begin with joining. Aside from variables already discussed, joining seems to largely be determined by structural factors. In particular, older individuals and those who reside in rural communities are more likely to join voluntary associations and community groups. In addition, those who discuss politics frequently and show an interest in public affairs are also more likely to join. A similar explanatory pattern emphasizing structural concerns is also true for collective action. As with joining, collective action is an activity in which older individuals and rural residents are more likely to participate. In addition, Whites, Coloureds and Asians are all less likely than Africans to engage in collective action, while those who are more impoverished are more likely. But in addition to social structure, other factors also matter for collective action. In particular, those who feel they can trust their neighbors are more likely to engage in collective action, as are those who frequently ! 199! discuss politics and show an interest in public affairs. Finally, economic performance evaluations matter, showing that individuals who feel the government is doing well providing economic goods, are less likely to attend meetings and join with other citizens to raise issues. When it comes to contacting both local and national government officials, social structure is also important. With regard to contacting at the local level, Whites, Coloureds and Asians are all less likely than Africans to contact local government councilors, while older folk, the heavily impoverished and those interested in politics are more likely to contact. Partisanship also matters here, with those who are ANC partisans being less likely to contact a local government councilor. At the parliamentarian level, as with local government councilors, minorities (Asians and Whites, though not Coloureds in this case) are less likely to contact. But the heavily impoverished, urbanites and those who frequently discuss politics are more likely to contact. In the case of protest behavior, only three additional variables reach statistical significance. In particular, Whites are less likely than Africans to engage in this type of behavior, while the impoverished and those with an interest in public affairs are more likely. Finally, in terms of voting, we find that Whites are less likely than Africans to vote, while older people are more likely to do so. In addition, those who trust their neighbors are also more likely to engage in electoral politics. ! 200! CONCLUSION This chapter has shown that security factors are important determinants of political participation. It finds that those who see the state as legitimate and rely on non-state security are less likely to participate in electoral politics (voting), but that those who see the state as legitimate and rely on state security are more likely to engage in this form of political participation. This suggests that individuals’ electoral participation is strengthened when they feel that the state is legitimate and that they can rely on it to provide the most fundamental of political goods, security. The provision of security by a legitimate state also seems to matter for collective action and protest, with those who both see the state as legitimate and rely on state security being less likely to engage in these political actions. Conversely, seeing the state as legitimate and relying on nonstate security seems to boost collective action and protest. These results suggest that seeing the state as an arranger of non-state security may not only improve perceptions of state legitimacy (See chapter 5), but that it may also boost non-electoral forms of participation. In addition, victimization and feelings of personal insecurity are powerful predictors of political participation. In most of the models of participation explored here, personal insecurity has a positive and statistically significant effect, meaning that the more frequently individuals feel insecure, the more politically engaged they become. Only in the cases of collective action and voting does this variable not reach statistical significance. Finally, this chapter shows that, contrary to claims made in the literature (Bateson 2012), the effect of victimization on participation varies both according to the type of victimization and the ! 201! type of participation. Property victimization only proves to be a significant predictor of participation in the cases of collective action and contacting local government officials. In both of these instances, the effect is positive, meaning that property victimization mobilized participation. When it comes to contact victimization (like assault, rape, murder and robbery), however, not only does this have an impact on more forms of participation, but the direction of the relationship varies according to the type of participation under consideration. In particular, contact victimization suppresses joining (when there is social trust) and voting, but increases protesting and contacting (of MPs). These findings suggest that victims of property crime turn to local government in their time of need, but that victims of violent crime feel the need to take their grievances to higher political heights. Perhaps the sentiment is that property crime represents less serious forms of crime that can be dealt with by local leaders, whereas violent crime, as a more serious form of crime needs to be addressed by those who potentially have more resources to bring to bear on the issue. While many studies have examined how fear of crime shapes attitudes and behavior, these results show that reported experiences of actual victimization go a long way in shaping political behavior. ! 202! CHAPTER 7 CONCLUSION ! 203! INTRODUCTION This dissertation has sought to understand how individuals’ everyday experiences of crime and security shape their political attitudes and behavior. It began by acknowledging two important considerations that are unique to the security atmosphere in developing countries. The first relates to how widespread individual reliance on non-state security is in developing contexts. While private citizens all over the world rely on a range of non-state actors to meet a variety of basic needs, this reliance is particularly pronounced in developing countries where the state’s capacity to provide is often weak. Secondly, this dissertation acknowledged the widespread nature of crime and victimization in South Africa, but in many developing countries more broadly. These realities make a study that attempts to explain the political consequences of nonstate security and other security factors more broadly, timely and relevant. In what follows, I will summarize the main findings of the dissertation, speaking to how well security factors help to account for variation in perceived legitimacy and individual political participation in South Africa. But I also attempt to move beyond South Africa to explore how generalizable these findings are to other sub-Saharan African countries. I end by discussing the contributions and limitations of the research, and suggesting directions for future study. SUMMARY OF KEY FINDINGS The core of this dissertation empirically explored the relationship between non-state security reliance and perceptions of state legitimacy in South Africa. I found that individual reliance on ! 204! non-state security undercuts perceptions of state legitimacy when individuals view the state as a producer of security. However, perceptions of state legitimacy improve for those who rely on non-state security and see the state as an arranger. Chapter 5 therefore highlights how the provision of goods by non-state actors may not necessarily undermine the state-society relationship, and indeed, may even strengthen it; at least in the context of urban South Africa. In addition, the dissertation moved beyond political attitudes to explore the effect that non-state security has on political behavior in South Africa. In chapter 6, I proposed that the effect of nonstate security on political participation should be conditional on perceived state legitimacy. In particular, I argued that individuals who relied on non-state security and saw the state as legitimate, should be more likely to participate, except in the case of protest where I would expect them to be less likely to participate. I found that those who relied on non-state security and saw the state as legitimate were more likely to join, engage in collective action, contact and even protest, but less likely to vote. In short, relying on non-state security and seeing the state as legitimate strengthened non-electoral forms of participation, but undercut electoral ones. These findings, along with the findings from chapter 5, show that the non-state provision of goods may strengthen or undercut political attitudes and behavior, depending on how individuals’ see the proper role of the state in service delivery. Moreover, chapter 6 also investigated the impact of other security variables on participation, examining how different types of victimization and feelings of personal insecurity shape political behavior. In sum, I found that feelings of personal insecurity generally mobilize citizens to act, except in the cases of collective action and voting. Victimization also proves to be ! 205! an important determinant of political participation, though the effect of victimization varies depending upon the type of victimization and the form of political participation under consideration. In South Africa, property victimization only proves to be a significant predictor of participation in the cases of collective action and contacting local government officials. In both of these instances, the effect is positive, meaning that property victimization mobilized participation. When it comes to contact victimization however, not only does this form of victimization have an impact on more forms of participation, but the direction of the relationship varies according to the type of participation. In particular, contact victimization suppresses joining (when there is social trust) and voting, but increases protesting and contacting (of MPs). These findings suggest that victims of property crime turn to local government in their time of need, but that victims of violent crime feel the need to take their grievances to higher political heights. Perhaps the sentiment is that property crime represents less serious forms of crime that can be dealt with by local leaders, whereas violent crime, as a more serious form of crime, needs to be addressed by those who potentially have more resources to bring to bear on the issue. These results show the varied political approaches that individuals take in response to different forms of victimization. They suggest that in South Africa, when individuals experience less serious, property-crime, they will rely more on local level officials and loose networks of collective action in response to suffering the loss of material goods. However, when individuals have their person violated, they will be more likely to respond to the shock of victimization by joining in protests and contacting national level officials. I reiterate that the fact that victimization ! 206! suppresses voting may be a function of South Africa’s dominant party system. Individuals who have suffered from violent crime therefore make rational calculations about how to exert their energy and influence in the aftermath of victimization. I suggest that in the context of a dominant party state, individuals may find that working with parliamentarians in office to bring about change may be a surer bet than attempting to punish them at the polls. GENERALIZABILITY OF THE FINDINGS The importance of citizens’ concerns with the delivery of security is demonstrated in the context of South Africa, but just how generalizable are these findings to other contexts in Africa and elsewhere? Given that South Africa is often considered to be “exceptional”, do findings on the relationship between non-state security and perceived state legitimacy hold in other African contexts? Do the relationships that I find between security factors and participation cross country lines? In an attempt to gauge the extent to which the results of this study travel to other African countries, I run the state legitimacy and political participation models with a cross-national 91 dataset of 11 African countries . These data were collected by Afrobarometer between 2011 and 2012. The sample is comprised of 20,414 African citizens. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 91 Countries included in the sample are Benin, Botswana, Cape Verde, Ghana, Kenya, Liberia, Malawi, Mauritius, Tanzania, Uganda, and Zimbabwe. This round of surveys includes 35 African countries, however, the full merged dataset with all 35 countries was not available at the time I was writing this chapter. ! 207! The Effect of Non-State Security on Perceptions of State Legitimacy in Africa I first examine the relationship between non-state security and perceptions of state legitimacy. Table 7.1 shows the pooled logistic regression output of perceived state legitimacy across 11 African countries. The results show that the relationship between non-state security and perceived state legitimacy is negative and statistically significant, meaning that individuals who rely on non-state security are less likely than those who rely on state security to see the state as legitimate. This is the expected direction of the relationship between non-state security and state legitimacy when we do not take into account the conditional effect of how individuals view the state’s role in security provision. Unfortunately, this most recent round Afrobarometer data does not contain an indicator that allows us to gauge whether individuals believe the state should be an arranger or producer of security. Therefore, I am not able to test this conditional effect. However, the importance of who provides comes through as the non-state security variable maintains its significance, even when accounting for other competing hypotheses. While I am not able to directly test the conditional affect of non-state security on state legitimacy when individuals see the state as an arranger, I do believe that this general argument is generalizable to other African, developing and developed contexts. Given that citizens in all these contexts should hold opinions about the proper role of the state in service delivery, I would expect my theory to travel. Moreover, citizens in the developed and developing world alike rely on non-state actors for the provision of key goods. Even though I conducted my survey in urban areas where issues of crime are likely to be more salient, the question is just as applicable to rural communities. In fact, Afrobarometer data show that across the 11 countries included in the ! 208! sample, 41% of rural respondents report going to the police for help after a crime, but that over one-quarter (26%) of rural respondents would go to a traditional leader first. With substantial numbers of individuals in rural communities who feel as if traditional leaders should be responsible for maintaining law and order, my theory would seem to have some applicability in these areas as well. Table 7.1: Pooled Logistic Regression Output on Perceptions of State Legitimacy, 11 92 African Countries Independent Variables Perceptions of State Legitimacy Non-state security -.276*** (.049) Contact Victimization -.047 (.039) Economic Performance Evaluations .031 (.036) Institutional Trust .348*** (.026) Transgression of Rule of Law -.034 (.023) Satisfaction with Democracy .179*** (.025) Political Participation -.054 (.056) Age .002 (.002) Male -.003 (.044) Lived Poverty -.001 (.028) -.083* (.046) .001 (.013) Employed Education N=20,414 Standard errors in parentheses !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 92 This model includes provincial fixed effects, though the coefficients for country dummy variables are not shown.! ! 209! Table 7.1 (cont’d) Statistical significance denoted by ***=p≤ .01,**= p≤ .05, *=p≤.10 Constant: -.232 R square:.076 -2 Log likelihood: 12527.918 In addition to non-state security, other factors also prove to be important determinants of perceived state legitimacy. Here, as in the South Africa model, being a victim of contact crime suppresses attitudes of state legitimacy. However, the coefficient on contact victimization while remaining negative, does not reach statistical significance in the cross-national analysis. This may be due to lower levels of crime found in other African countries. Moreover, individuals who trust institutions of the state and are satisfied with democracy are more likely to see the state as legitimate, while the employed are less likely to see the state as such. In the South Africa model, institutional trust is also positive (though not statistically significant), while support for 93 democracy is negative (though not statistically significant) . Finally, in the South Africa model, the sign on the employed coefficient is positive, though not statistically significant. While employment makes individuals more likely to see the state as legitimate in South Africa, it suppresses legitimacy attitudes in the eleven countries included in this sample. Non-State Security, State Legitimacy and Political Participation in Africa Next, we test the non-state security-state legitimacy-participation link with cross-national data. In the case of joining, the interaction term between non-state security and state legitimacy is !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 93 The difference in signs in the two models may be due to the different measures employed. In South Africa I measured support for democracy, while measuring satisfaction with democracy in the cross-national analysis. Support for democracy was not statistically significant in the crossnational regression. ! 210! positive (as in South Africa). Though here, the coefficient is also statistically significant whereas it was not in South Africa. This suggests that individuals who rely on non-state security and see the state as a legitimate are more likely to join in associational life. Also like South Africa, those who rely on non-state security and see the state as a legitimate are more likely to engage in collective action (though the coefficient here, as in the South Africa case, is not statistically significant). The same effect is witnessed for contacting local government officials where the relationship between the interaction term is positive, but not statistically significant (again, as in South Africa). Therefore, the findings from South Africa with respect to the positive influence of the interaction term on joining, engaging in collective action and contacting local government councilors, are generalizable to other African countries. The participation picture starts to diverge, however, when we explore contacting members of parliament and protesting. In the cross-national analysis, individuals who rely on non-state security and see the state as legitimate are less likely to contact parliamentarians (coefficient negative, but not statistically significant) and less likely to engage in protest behavior (coefficient negative and statistically significant). These results differ from the South Africa results and suggest that individuals in other sub-Saharan African countries withdraw from contacting national level officials and protesting when they rely on non-state security and see the state as legitimate. Moreover, if we examine the coefficient on the state legitimacy variable, we see that individuals who see the state as legitimate and rely on state security are also less likely to contact parliamentarians (the coefficient is negative but not statistically significant). This suggests that whenever the Africans included in this sample see the state as legitimate, they are less likely to engage with national level officials, regardless of the entity from which they get their security. ! 211! This supports the common view that individuals are less likely to make political noise when the state is perceived to be doing well. The same pattern holds true for protest behavior, with individuals who see the state as legitimate being less likely to participate in this political activity. Finally, when we explore the relationship between the non-state security/state legitimacy interaction term and voting, we find the same relationship that we witnessed in South Africa. Specifically, when individuals rely on non-state security and see the state as a legitimate, there is a negative and statistically significant effect on voting. This demonstrates, as in South Africa, that individuals’ voting behavior is suppressed when they rely on non-state security , but heightened when they rely on state security (as evidenced by the positive coefficient on the state legitimacy variable). Table 7.2: Regression Analyses of Joining, Collective action and Contacting, 11 African 94 Countries Independent Joining Collective Contacting (LGCs) Contacting Variables action (MPs) Non-state Security .078 (.063) -.164*** (.053) .040 (.027) -.011 (.022) .033 (.026) -.027 (.020) .014 (.019) -.013 (.015) Non-state Security*State Legitimacy .138* (.077) .012 (.033) .012 (.031) -.020 (.022) Contact Victimization .258*** (.059) .106*** (.026) .064*** (.024) .093*** (.018) State Legitimacy !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 94 These models include provincial fixed effects, though the coefficients for country dummy variables are not shown. ! 212! Table 7.2 (cont’d) Property Victimization Social Trust Personal Insecurity Political Order Economic Goods Political Efficacy (LGC indicator for LGC model and MP indicator for MP model) Discuss Politics Interest in Public Affairs Media Exposure Age Male Lived Poverty Urban Employed Education Constant N= 20, 414 R square: -2 Log likelihood ! .260*** (.043) .066*** (.019) .092*** (.018) .050*** (.013) .077* (.040) -.029* (.018) .058*** (.017) -.001 (.008) .039*** (.015) -.007 (.007) .009 (.012) 4.091 (.005) .012 (.021) .212*** (.031) .170*** (.024) .021** (.009) -.010 (.013) .118*** (.010) .005 (.009) -.032*** (.013) .273*** (.009) -.011* (.006) .004 (.009) .198*** (.007) .160*** (.031) .209*** (.021) .161*** (.013) .089*** (.009) .123*** (.013) .050*** (.008) .088*** (.009) .029*** (.006) .076*** (.024) .016*** (.001) .007 (.038) .180*** (.024) -.313*** (.045) .112*** (.039) .078*** (.013) .010 (.010) .010*** (.001) .139*** (.016) .064*** (.010) -.291*** (.019) .022 (.017) .025*** (.005) .021** (.009) .009*** (.001) .129*** (.015) .052*** (.010) -.139*** (.017) .048*** (.016) .015*** (.005) .028*** (.007) .005*** (.000) .042*** (.011) .029*** (.007) -.090*** (.013) .038*** (.011) .017*** (.004) -3.680*** (.140) 1.002*** (.051) -.243*** (.043) -.249*** (.032) .104 17937.11 .290 .154 213! .103 Table 7.2 (cont’d) standard errors in parentheses statistical significance denoted by ***=p≤ .01,**= p≤ .05, *=p≤.10 Table 7.3: Regression Analyses of Protesting and Voting, 11 African 95 Countries Independent Variables Non-state Security State Legitimacy Non-state security*State Legitimacy Contact Victimization Property Victimization Social Trust Personal Insecurity Political Order Economic Goods Political Efficacy Discuss Politics Interest in Public Affairs Protesting Voting .095*** (.015) -.104*** (.012) -.063*** (.018) .104 (.071) .076 (.056) -.144* (.085) .135*** (.014) -.050 (.066) .034*** (.010) -.020** (.009) .011*** (.004) .027 (.047) .018 (.043) -.075*** (.019) -.018*** (.005) .057*** (.007) .045*** (.006) .024 (.024) -.145*** (.034) -.097*** (.027) .068*** (.007) .007 (.005) .213*** (.035) .053** (.023) !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 95 These models include provincial fixed effects, though the coefficients for country dummy variables are not shown. ! 214! Table 7.3 (cont’d) Media Exposure Age Male Lived Poverty Urban Employed Education Constant .050*** (.006) -.002*** (.000) .032*** (.009) .034*** (.006) -.004 (.010) .003 (.009) -.007** (.003) .053** (.027) .003 (.026) .064*** (.002) -.028 (.041) .060** (.027) -.234*** (.048) .412*** (.043) .018 (.014) 1.033*** (.155) N= 20, 414 R square: .117 .135 -2 Log likelihood 15509.221 standard errors in parentheses statistical significance denoted by ***=p≤ .01,**= p≤ .05, *=p≤.10 The Effect of Insecurity and Victimization on Participation in Africa In most cases, feelings of personal insecurity decreases individuals’ willingness to engage in politics in Africa. Unlike in South Africa, other Africans who frequently feel insecure are less likely to join, engage in collective action and contact local government councilors (insecurity is negative in all these cases, but only statistically significant in the case of joining). Perhaps South Africans are more likely than other Africans to be moved to engage with other citizens and enlist the help of local officials because of the widespread nature of insecurity and the greater salience of crime and security issues in this country. Like in South Africa, feelings of personal insecurity encourage people to protest, but discourage them from voting. ! 215! With regards to victimization, there is a positive and statistically significant relationship between both forms of victimization (contacting and property) and all forms of political participation except one. In particular, those who have been victims of property and contact crime are more likely to join, engage in collective action, contact elected officials, and protest. Only in the case of voting do contact and property victimization not reach statistical significance. Moreover, the coefficient on contact victimization is negative, suggesting that those who have been victims of violent crime are less likely to vote, as in South Africa. Unlike in South Africa, the effect of property victimization is not simply limited to contacting local government councilors and engaging in collective action. In the cross-national model, the positive impact of property victimization extends to all modes of political participation (though this variable is not significant in the case of voting). As in South Africa, being a victim of a contact crime encourages most citizens to become politically active, except, as already acknowledged, in the case of voting. In sum, feelings of personal insecurity generally decrease individuals’ willingness to participate in politics, while experiences of victimization increase the likelihood that they will do so. It seems as if experiences of crime, whether violent or propertyrelated, mobilize citizens across Africa to connect, between elections, to other citizens and elected officials. ! 216! Summarizing the Relationship Between Security and Political Participation in Africa The results from the cross-national analysis and the South Africa model suggest that generally, individuals who rely on non-state security and see the state as legitimate are more likely to participate. But why would this move individuals to political action? The reasons may be political or economic. First, for individuals who rely on non-state security, the added layer of physical security may give them the confidence needed to participate in political life. This may become particularly important when we consider political activities such as attending meetings that often happen in the evening (times when citizens feel most unsafe). Therefore, an extra layer of physical protection may be just the resource needed to get citizens connecting with one another through meetings and associations. From an economic perspective, citizens who rely on non-state security are essentially taxed twice for the provision of security. One the one hand, they pay for non-state security (in the case of market-based security) from a commercial security firm. On the other hand, as citizens, they pay the government for state security services (via taxation) from which they may well no longer benefit. Therefore, these citizens may be mobilized to engage in political life and petition for better state provided security, in the hopes that stronger state capacity to produce security may mean that non-state security will no longer be necessary. In effect, it would mean only a single tax payment for them as citizens. Therefore, citizens who have a vested economic interest in state-building may be more likely to engage politically. ! 217! While reliance on non-state security and seeing the state as legitimate generally encourages participation, there are two exceptions in the cross-national model. Specifically, it seems to be the case that whenever individuals’ view the state as legitimate-regardless of who provides their security-they are less likely to contact parliamentarians and protest. Therefore, perceived state legitimacy is key to decreasing Africans’ protest behavior and contacting (MPs). However, it is not exactly clear why the relationship between legitimacy and contacting hinges on how one’s security needs are met in the case of contacting local councilors, but not contacting parliamentarians. Perhaps the fact that state legitimacy makes individuals less likely to contact national level officials supports the notion that individuals are more likely to contact when times are bad, and less when times are good. One more relationship regarding the non-state security-legitimately-participation link is worth discussing here. In South Africa, as well as cross-nationally, those who rely on non-state security and see the state as legitimate are less likely to vote. In both cases, this suggests that individuals will only be moved to participate in electoral politics, when they feel the state is actively involved in the day to day production of security. In all, reliance on non-state security and seeing the state as legitimate seems to boost non-electoral forms of political participation (with the exception of protesting and contacting parliamentarians at the cross-national level), but undercut electoral ones. Finally, to summarize the political affect of personal insecurity and victimization, we generally find that feelings of personal insecurity encourage participation in South Africa, but discourage participation in other parts of Africa. Moreover, while victimization depresses some forms of ! 218! participation (joining and voting) and strengthens others (collective action, contacting, and protesting) in South Africa, its impact largely runs in one direction in the cross-national analysis (increasing participation). These findings collectively show the importance of security considerations for political attitudes and behavior across Africa. They show that the choice of security provider and how legitimate one perceives the state to be, affects individuals’ willingness to participate in political life. Moreover, other security factors, such as victimization and feelings of personal insecurity, also prove to be important determinants of Africans’ political behavior. CONTRIBUTIONS OF THE RESEARCH This research makes several key contributions to the existing literature. First, it empirically explores the political impact of security considerations and their political importance from the vantage point of ordinary citizens. While scholars have often asserted that the provision of security is of paramount importance for citizens, this research shows, empirically, that security factors do actually shape popular political attitudes and behavior. In particular, I explore the impact of three key security variables; non-state security, victimization, and personal insecurity. The Political Consequences of Non-State Security Provision State-provided security has often been asserted as key to legitimating the relationship between state and society. By focusing on the relationship between non-state security and perceptions of ! 219! state legitimacy, this dissertation has explored whether this is in fact the only route to legitimacy from a micro-perspective. I’ve found that the state is not doomed (as some would suggest) when they are not the primary producers of security. For example, the donor community has often stressed the potential political pitfalls of non-state provision of goods. But by showing that reliance on non-state security may strengthen perceptions of legitimacy, (at least in the context of urban South Africa), my research shows the role that non-state actors may play in connecting citizens back to the state in a healthy way. In short, non-state provision can be good for the state. The fact that citizens may see the states’ role as one of a facilitator and that the state may bank legitimacy dividends when they do, means that the state is not necessarily doomed when its capacity to produce security is weak. It does mean, however, that the state needs some capacity; if not in the direct production of goods, then at least the capacity to regulate and somehow leverage goods from non-state actors. Therefore, citizens may not expect the state to be omnipotent, but they do expect it to do be capable of performing some tasks, such as effective regulation. Finally, while the emergence of non-state actors may be linked to weak state production capacity; the persistence of these actors over time may indicate a renewed role of the state from the vantage point of citizens (Singh 2008) Donors speak of non-state provision as if it’s a temporary service delivery feature. But in societies that have a long history of relying on nonstate actors, non-state service delivery may be an enduring feature of the service delivery environment, even when the state does build greater production capacity over time. ! 220! From a policy standpoint, this research highlights the importance of public-private partnerships. When the state acts as an arranger, this implies a partnership with non-state actors. Since individuals who see the state as an effective arranger of goods often credit the state with legitimacy and are more likely to participate politically, public-private partnerships may prove to be goods ways of building the state, and even democracy, from the bottom up. The Political Consequences of Victimization and Personal Insecurity Finally, this research shows that victimization and personal insecurity are key political variables that shape individual attitudes toward the state (victimization only) and individuals’ willingness to participate in political life. Contact victimization, or reported experiences of violent crime, undercut individuals’ perceptions of state legitimacy and boost participation in most forms of political behavior; whereas property victimization increases some forms of participation in South Africa and most forms of participation cross-nationally. This research therefore shows that security variables that tap into people’s everyday experiences of crime and violence should be central to studies of political attitudes and behavior. LIMITATIONS OF THE RESEARCH While this research has made several important contributions to the literature, there are, as with any research, certain shortcomings. First, I focus, to a great extent, on popular perceptions of state legitimacy. However, it is not exactly clear just how individuals conceive of the state. A number of questions remain to be answered here, not least of which is what types of institutions ! 221! most readily come to mind when citizens in Africa think of “the state”. In this study, I have focused on state institutions that are responsible for providing security and exercising coercion. I focus on these institutions because I want to know how African’s come to form their assessments of state institutions that are responsible for providing what has often been called a chief public good-security. Yet, it would be worthwhile to understand what institutions are top of the mind for ordinary Africans’ when they think of “the state”. Second, while I argue that citizen’s perceptions of state legitimacy will hinge on how they see the state’s role in service delivery, it is still not exactly clear how individuals come to see the state as rightfully playing the role of arranger or producer. Moreover, do citizens feel that the state should be a producer or arranger for all goods alike, or do they have different preferences for the role that the state should play in service delivery depending upon the type of good under consideration? Therefore, much more could be known about how individuals’ arrive at their normative views of what role the state should play in service delivery, and variation in these views across policy sectors. Finally, as it relates to victimization, there remains much more to be known about how, exactly, victimization shapes attitudes toward the state and political participation. While Bateson (2012) proposed several theories as to why victimization might increase political participation, much more work needs to be done to determine which of these proposed mechanisms carry the greatest explanatory weight. Therefore, the question remains as to whether crime victims are driven to political action primarily due to instrumental or expressive reasons. ! 222! FUTURE DIRECTIONS In future work, I would like to build upon the research that I’ve begun here. In particular, I’d like to investigate, more broadly, how the non-state provision of a range of goods and services shapes a variety of political outcomes. While I feel that exploring the political consequences of nonstate security has been illuminating, it is, in fact, a starting point. Moving forward, I would like to investigate how various non-state forms of provision affect attitudes toward the state, but also attitudes toward democracy across Africa. Moreover, I would like to move beyond an analysis of state legitimacy to explore the legitimacy of non-state actors. In places like Africa where individual reliance on non-state security is so widespread, do non-state actors build their own independent stocks of legitimacy, or is their legitimacy simply derived from a perceived connection with the state? In fact, it’s a fascinating question as to who legitimates whom. In other words, when partnerships between the state and non-state actors arise, is it the state that legitimates the non-state, the non-state that legitimates the state, or is legitimation mutually reinforcing? These are questions to which I hope to attend in the near future. ! 223! APPENDICES ! 224! APPENDIX A: LIST OF SURVEY NEIGHBORHOODS BY CITY Table 8.1: List Of Survey Neighborhoods By City Cape Town No. of Durban Neighborhoods Intervie Neighborhoods ws Gugulethu 8 Phoenix (Clayfield) ! Du noon 4 Langa No. of Johannesburg Interviews Neighborhoods 4 4 4 Chatsworth (Bayview) Avoca Doornbach Khayelitsha Lansdowne 4 16 4 Bonela Lotus Park Springfield 4 4 4 Schaap Kraal Maitland Garden Village/Maitland Scottsdene Bellville Mitchells Plain Somerset West 4 4 Isipingo Hills Reservoir Hills 4 8 4 4 16 4 Genazzano La Mercy Tongaat Beach Hilary 4 4 4 4 Marina Da Gama Gordon’s Bay 4 4 Essenwood Musgrave 2 4 Kirstenhof Signal Hill/Lion’s Head Constantia 4 4 Bothas Hill Pine Town 4 Milnerton 4 Camps Bay Claremont Sybrand Park Westlake Mowbray 4 Soweto Freedom Park Hillbrow Kapok Informal Soweto-Mmesi Park Bramley View Crystal Gardens 4 4 4 4 4 4 4 Kew Lyndhurst Wits University EdenvaleEdenvale Hospital Glenkay Klipriviersberg Estate Martindale Rossmore New Germany (Central) Umhlanga (Umhlanga Rocks) 2 South Hills 4 4 RandburgJukskei Park 4 4 Morningside 4 4 4 4 4 4 Oceanview Hillcrest (Central) Kloof (Central) Avoca Hills 4 4 8 6 SandtonFourways Berario Fairway Kensington Rembrandt Park 225! 4 Alexandra-East Bank Diepsloot No. of Intervi ews 4 8 4 4 4 4 4 4 4 4 4 4 4 4 Table 8.1 (cont’d) Ysterplaat 4 Avoca 6 Eldorado Park Ext 8 4 Wingfield 4 Cato Manor 8 4 Zonnebloem 4 Chesterville Clermont Umkhumbane Glenmore/Glenwood 8 7 7 4 Ennerdale Ext. 6 Newclare Bosmont Eldorado Estate Ennerdale Ext. 10 Homestead Park Lenasia Ext. 2 Mayfair Lenasia South Lenasia South Ext. 1 Mayfair West Total Johannesburg Interviews=144 4 4 4 4 Total Cape Town Interviews=132 ! Total Durban Interviews=138 226! 4 4 4 4 4 4 APPENDIX B: MASS SURVEY QUESTIONNAIRE NON-STATE SECURITY AND STATE LEGITIMACY IN CONTEMPORARY SOUTH AFRICA Mass Questionnaire Danielle Carter Doctoral Candidate Michigan State University Respondent Number Leave Blank Fieldworker Number Fill in from List of Codes Province Circle Correct Code Western Cape KwaZulu-Natal Gauteng 500 501 502 City Circle Correct Code Cape Town Durban Johannesburg 600 601 602 Field Number Suburb: Write in Field Name Household Selection Procedure It is critical at this stage that you select a random (this means any) household. Your starting point in each community has been randomly chosen by the field supervisor. Starting at this th point, field workers will walk in opposite directions of each other until you reach the 10 house, th counting houses on both the right and the left. The 10 house will be the house where you ask for the interview. If you are covering an entire community by yourself, choose a direction from th the start point and walk until you reach the 10 house, counting houses on both the right and left. Once the first interview has been completed, return to the start point and then walk in the th opposite direction until you reach the 10 house, counting houses on both the right and the left. Repeat this pattern until all interviews in this area have been completed. If a visit is unsuccessful, use the table below to record your progress until you make a successful visit. Circle a code number for unsuccessful visits only. ! 227! NOVISIT Reasons for Unsuccessful Visits NOVISIT_1 NOVISIT_ NOVISIT NOVISIT_4 NOVISIT_5 2 _3 Household Household Household Household Household 1 2 3 4 5 Refused to be interviewed 01 01 01 01 01 Person selected was never at home after at least two visits 02 02 02 02 02 03 03 03 03 03 Not a citizen/Spoke only a foreign language 04 04 04 04 04 Deaf/Did not speak a survey language 05 05 05 05 05 Did not fit gender quota 06 06 06 06 06 No adults in household 07 07 07 07 07 Other (specify)____________ ___________________ ____ 08 08 08 08 08 Not Applicable 97 97 97 97 97 Household/Premises empty for the survey period after at least two visits If no one is at home (i.e., premises empty), substitute with the very next household. If the interview is refused, use an interval of 10 to select a substitute household, counting houses on both the right and the left. When you find a household with someone home, please introduce yourself using the following script and this script only. Please do not deviate from this script. ! 228! Interviewer: Hello. My name is ____________________. I am working on behalf of a dissertation study that is being conducted at Michigan State University in the United States. I do not represent any government or political party. The purpose of this study is simply to understand individual views of crime, security, and the state in South Africa. I would like to interview someone from your household on these issues. All information that I am provided with during the interview will be kept confidential. Moreover, the interview may be stopped at any point and the individual who is being interviewed may refuse to answer any question they do not want to answer. Would you be willing to help me select an adult from your household for the interview? Note: The person must give his or her informed consent by answering positively. If participation is refused, walk away from the household and record this in the above table on “Reasons for Unsuccessful Visits.” Substitute the household using an interval of 10 households. If consent is secured, have respondent sign the consent form and then proceed to Respondent Selection. Respondent Selection Procedure Within the household, you must select a random (this means any) individual. This individual becomes the interview Respondent. In addition, the study needs to include an equal number of female and male respondents. Please use the table below to help you to alternate between genders. Circle the correct code below. Note that “First Interview” should ONLY be used for your very first interview on the first day of fieldwork, NOT your first interview every day. First Interview Male Female 0 PREVINT. Previous interview was with a: 1 1 THISINT. This interview must be with a: 2 2 Interviewer: Please tell me how many adults older than 18 years living in this household. That is, all household members (18+ years) who live in this household for more than 15 days per month. Record total number of adults who live in household Interviewer: Please tell me the names of all males / females [select correct gender] who presently live in this household. I only want the names of males / females [select correct gender] who are citizens of South Africa and who are 18 years and older. If this interview must be with a female, list only women’s names. If this interview is with a male, list only men’s names. List all eligible household members of this gender who are 18 years or older, even those not presently at home but who will return to the house at any time that day. Include only citizens of South Africa. ! 229! Women's Names Men's Names 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 Please record the total number of adult women / men [select correct gender, from above table] who are citizens of South Africa in the household, i.e., how many names did you write in either the left of the right column above. Enter a two-digit number. ADULT_CT. Number of women / men in the household (adult citizens) Take out your deck of numbered cards. Present them face-down so that the numbers cannot be seen. Ask the person who is selecting respondents to pick any card, by saying: Interviewer: Please choose a card. The person who corresponds to the number chosen will be the person interviewed. [REMEMBER to circle the code number of the person selected on the table above.] Interviewer: The person I need to speak to is [insert name] _______________________________. Is this person presently at home? ! If yes: May I please interview this person now? If no: Will this person return here at any time today? If no: Thank you very much. I will select another household. Substitute with the next household to the right and repeat the respondent selection procedure. (NOTE: YOU CAN ONLY SUBSTITUTE HOUSEHOLDS NOT INDIVIDUALS.) If yes: Please tell this person that I will return for an interview at [insert convenient time]. If this respondent is not present when you return, replace this household with the next household to the right. 230! If the selected respondent is not the same person that you first met, repeat Introduction: Interviewer: Hello. My name is ____________________. I am working on behalf of a dissertation study that is being conducted at Michigan State University in the United States. I do not represent any government or political party. The purpose of this study is simply to understand individual views of crime, security, and the state in South Africa. Any answers you provide will be confidential and anonymous. Your responses will be combined with the responses of approximately four hundred other South Africans so that it will be impossible to pick you out from other respondents. This interview will take approximately 45 minutes. You will not be penalized if you refuse to participate and if you do choose to participate, you may refuse to answer any question you do not want to answer. Do you wish to proceed? ! ! ! ! If no: o Interviewer: Thank you very much. I will select another household. If yes, ask the following question: o Interviewer: I am only allowed to conduct this interview with citizens of South Africa. Are you a citizen of South Africa, with an ID Book, or birth certificate issued by the Department of Home Affairs? Please remember, there is no penalty for not participating, and your answer will be kept confidential. If yes, proceed with the interview If no: Interviewer: Thank you very much. I will select another household. Substitute with the next household to the right and repeat the respondent selection procedure. (NOTE: YOU CAN ONLY SUBSTITUTE HOUSEHOLDS NOT INDIVIDUALS.) VISITS. Circle number How many visits were made to the household where the interview actually took place? DATEINTR. Date of interview [Interviewer: Enter day, month, and year] Day STRTIME. Time interview started [Interviewer: Enter hour and minute, use 24 hr. clock] 1 Month Year Hour Min If a respondent firmly refuses to answer any question, write “refused” in the answer space and continue to the next question. Also, if the respondent does not know an answer to a question write “don’t know”. NEVER READ THE “REFUSED” OR “DON’T KNOW” RESPONSES. ! 231! 2 BEGIN INTERVIEW NOTE TO INTERVIEWER: DO NOT READ SECTION HEADINGS SECTION 1: RESPONDENT INFORMATION DO NOT READ Interviewer: I would like to begin by asking you some information about yourself. 1. How old are you? Enter three digit number. Don't Know = 999 If respondent is aged less than 18, stop interview and use cards to randomly draw another respondent in the same household Which South African language do you usually speak at home? (DO NOT READ RESPONSES) English 001 Swazi 010 Portuguese 002 Venda 011 Afrikaans 003 Zulu 012 Ndebele 004 Other (specify) 995 ________________________ Xhosa 005 Refused to answer 998 Pedi/Spedi/North 006 Don’t know 999 Sotho Sesotho/Sotho/South 007 Sotho Setswana/Tswana 008 Shangaan 009 2. Is there a number where I can contact you to follow up if need be? __________________________________________________ Record number in space provided SECTION 2: ATTITUDES TOWARD LEGITIMACY OF THE STATE (LEGAL DIMENSION)/REGIME AND GOVERNMENT DO NOT READ Interviewer: Next I’d like to ask you some questions about your attitudes toward the state. There are no right or wrong answers, we are simply interested in your opinions. 3. For each of the following four statements, please tell me whether you agree or disagree: Strongly Disagree Neither agree Agree Strongly agree disagree nor disagree ! 232! A. The police always have the right to make people obey the law. B. The courts have the right to make decisions that people always have to abide by. C. The tax department always has the right to make people pay taxes. D. An elected government always has the right to make laws and policies that the country must adopt. 4. 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 Which of these three statements is closest to your own opinion? Statement 1: Democracy is preferable to any other kind of government. Statement 2: In some circumstances, a non-democratic government can be preferable. Statement 3: For someone like me, it doesn’t matter what kind of government we have. Statement 1: Democracy preferable 1 ! Statement 2: Sometimes non-democratic preferable 2 Statement 3: Doesn’t matter 3 Don’t know 9 233! 5. There are many ways to govern a country. Would you disapprove or approve of the following alternatives: Strongly Disap Neither Approv Strongly disapprove prove approve nor e approve disapprove A. Only one political party is allowed to stand 1 2 3 4 5 for election and hold office? B. The army comes in to govern the 1 2 3 4 5 country? C. Elections and the National Assembly are abolished so that 1 2 3 4 5 the president can decide everything? 6. And with regards to enforcement of the law, In your opinion, how often, in this country: Never Rarely Often Always A. Are people treated unequally under the law? 0 1 2 3 B. Do officials who commit crimes go unpunished? 0 1 2 3 C. Do ordinary people who break the law go unpunished? 0 1 2 3 SECTION 2A: ATTITUDES TOWARD LEGITIMACY OF THE STATE (ACTS OF CONSENT) DO NOT READ Interviewer: Let’s talk a little about your own personal political participation. 7. With regard to the most recent, 2009 national elections, which statement is true for you? You were not registered or you were too young to vote You voted in the elections ! 234! 0 1 You decided not to vote You could not find the polling station You did not have time to vote You were prevented from voting Did not vote for some other reason Don’t know/Can't remember 8. 2 3 4 5 6 9 If a presidential election were held tomorrow, which party’s candidate would you vote for? DO NOT READ RESPONSES UNLESS RESPONDENT CANNOT ANSWER African Christian Democratic Party (ACDP) African Muslim Party African National Congress (ANC) Afrikaner Unity Movement / Afrikaner Eenheidsbeweging Alliance of Free Democrats Azanian People’s Organization (AZAPO) Democratic Alliance – Democratic Party / Demokratiese Alliansie – Demokratiese Party (DA-DP) Federal Democrats Freedom Front / Vryheidsfront Independent Democrats (ID) Inkatha Freedom Party / Inkatha Vryheidsparty (IFP) Minority Front National Democratic Convention (NADECO) New National Party / Nuwe Nasionale Party (NNP) Pan Africanist Congress (PAC) United Christian Democratic Party (UCDP) United Democratic Movement (UDM) United Independent Front Christian Party COPE Others Would not vote Refused to answer Don’t know 9. 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Here is a list of actions that people sometimes take as citizens. For each of these, please tell me whether you, personally, have done any of these things during the past year. If not, would you do this if you had the chance? No, would never ! 1 2 3 4 5 6 235! No, but would do this Yes, once or Yes, several times Yes, often do this Attended a community meeting B. Got together with others to raise an issue C. Attended a demonstration or protest march D. Used force or violence for a political cause A. if I had the chance twice 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 SECTION 2B: STATE AS PRODUCER VS. ARRANGER OF SECURITY (and other goods)/INDEPENDENCE OF POLICE FORCE AND COURTS DO NOT READ 10. Which of the following statements is closest to your view? Choose Statement 1 or Statement 2. 1: The state should take the main responsibility for protecting citizens from crime. 2: The state should mainly provide support to private groups such as the commercial security industry or neighborhood watches so that they may protect citizens from crime. 1=Agree very strongly with Statement 1 Agree with Statement 2 Agree with Statement 1 Agree very strongly with Statement 2 1 2 3 4 Agree with neither (DO NOT READ) 5 Don’t know (DO NOT READ) 9 11. Who do you think should have primary responsibility for managing each of the following tasks: Central Gov’t ! Local Gov’t Traditional leaders 236! Commu nity member s Voluntary association s or Nongovernmen tal organizatio ns (NGO) The private sector (i.e. ADT) Only read this response for question 11a and 11g A. Keeping people safe? B. Keeping the community clean? C. Managing schools? D. Managing health clinics? E. Collecting income taxes? F. Solving local disputes? G. Maintaining law and order? 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 INDEPENDENCE OF STATE INSTITUTIONS DO NOT READ 12. Which of the following statements is closest to your view? Choose Statement 1 or Statement 2. 1: The ANC directly controls state institutions like the police force and courts. 1=Agree very strongly with Statement 1 2: The police force and courts act independently of the ANC. Agree with Statement Agree with Statement Agree very strongly 1 2 with Statement 2 4 1 2 Agree with neither (DO NOT READ) Don’t know (DO NOT READ) ! 3 5 9 237! SECTION 3: PERCEPTIONS OF CRIME AND SECURITY/EVALUATIONS OF POLICE AND COURTS DO NOT READ Interviewer: Ok, now we will talk for a while about your perceptions of crime and security in South Africa and your thoughts on the police and courts. 13. Do you think that crime in your neighborhood is increasing, decreasing, or remaining the same? Decreasing 1 Very safe 14. a. How safe do you feel walking alone in your neighborhood during the day? 14. b. How safe do you feel walking alone in your neighborhood after dark? 15. Remaining the same 2 1 Fairly safe 2 A bit unsafe 3 Increasing Don’t know 3 9 Very unsafe Don’t know 4 9 1 2 3 4 9 How well do you think the police are doing at the following activities: Very badly Fairly Fairly well Very well badly A. Preventing 1 2 3 4 crime? B. Catching 1 2 3 4 criminals? C. Making people 1 2 3 4 feel safe? D. Prosecuting 1 2 3 4 Criminals? Don’t know 9 9 9 9 SECTION 4: TRUST AND CORRUPTION IN POLICE AND COURTS DO NOT READ 16. How much would you say you trust: Not at all Just a little The police? Courts of law? ! Somewhat A lot Don’t know/Haven’t head enough 0 1 2 3 9 0 1 2 3 9 238! Would you say that you trust the following more, less or about the same amount as you did under apartheid: About the More Less Don’t know same The 1 2 3 9 police? The 1 2 3 9 courts? 17. How many of the following do you think are involved in corruption: Some of None Most of them All of them them Police? 0 1 2 3 Court 0 1 2 3 officials? 18. Don’t know 9 9 SECTION 5: RELIANCE ON NON-STATE SECURITY DO NOT READ Interviewer: The next set of questions will address some steps that you may take to meet your security needs. First I’d like to focus on what you would do AFTER a crime happened to you. (Post Victimization)-DO NOT READ Please tell me who you would call on if you or someone in your family had been a victim of crime? If respondent answered that they would only call the police, SKIP questions 20 and 22-25. 19. 20. If response= 2:If you were a victim of a crime, who would you FIRST call for help: the police or some other group? Why would you call this group or organization first? 22. If response= 2 or 3: Police only Police and some other group Other groups only None of these Don’ t know 4 9 1 2 3 Police Some other group Don’t know 1 2 9 21. ! Verbatim: Commercial security industry (i.e. ADT) 239! 1 In addition to or besides the police, which of these groups or organizations would you be most likely to turn to for help after the crime? Allow respondent to choose ONLY one response If response =2-9 skip 23 A-D. 23. If response=1: A. Who pays for the commercial security company that protects your family and property? B. How much of your income would you say you spend on hiring a security company? C. And how long have you relied on this company or a company like it for protection? ! Neighborhood watch scheme 2 Street Committee 3 Gang Vigilante group Community Police Forum 4 5 6 Family members Traditional leader Some other group 7 8 9 You or some other member of your household 1 Your employer 2 Another family member or friend 3 Other_______________________________________ 4 1-20% 21-40% 41-60% 61-80% 1 2 3 4 In the last year Last five years Last ten years Last fifteen years Since before the transition, pre-1994 1 2 3 4 5 D. And how much of the Rarely time would you say that you call on this Often company when you Always have a problem? 1 24. If responses=2-6: A. Did you help to form this group? Yes 1 No 2 B. Please tell me your current level of Not a member Inactive member 0 1 2 3 240! involvement with this group C. And how long have your relied on this group or a group like it for protection? D. And how much of the time would you say that you call on this group when you have a problem? Active member Official leader 2 3 In the last year Last five years Last ten years Last fifteen years Since before the transition, pre-1994 Rarely 1 2 3 4 5 1 Often 2 Always 3 25. How satisfied are you with the performance of [write in name of private organization that respondent is most likely to rely on___________________________________] in: Not at all satisfied A. Patrolling and securing your neighborhood? B. Arresting criminals? C. Punishing criminals? Not very satisfied Very satisfied 1 2 3 4 1 2 3 4 1 2 3 4 26. Is there any other type of group that you would turn to for help that was not listed above? Please do not give me specific names of groups. I am just interested in other types of groups that you may turn to for help that I did not list. ! Fairly satisfied 241! (Crime Prevention) )-DO NOT READ Interviewer: The next set of questions are the same as the last ones. But instead of thinking about what you would do AFTER a crime happened to you, I want you to think about what you would do to try to keep the crime from occurring in the first place. 27. Please tell me who you rely on to keep you and your family from becoming victims of crime? If respondent answered that they would only call the police, SKIP questions 28 and 30-33. 28. If response =2: Who are you most likely to rely on to keep you and your family from becoming victims of crime: the police or some other group? 29. Why would you be more likely to call on this group or organization? 30. If response= 2 or 3: In addition to or besides the police, which of these groups or organizations would you be most likely to turn to keep you and your family from becoming victims of crime? Allow respondent to choose ONLY one response Police only Police and some other group None of these Don’ t know 4 9 1 2 3 Police Some other group Don’t know 1 2 9 Verbatim: Commercial security industry (i.e. ADT) 1 Neighborhood watch scheme 2 Street Committee 3 Gang Vigilante group Community Police Forum 4 5 6 Family members Traditional leader Some other group 7 8 9 If response =2-9 skip 31 A-D (DO NOT READ) ! Other groups only 242! ! 31. If response=1: A. Who pays for the commercial security company that protects your family and property? You or some other member of your household 1 Your employer 2 Another family member or friend 3 B. If you or some other member of your household pays for this service, how much of your income would you say you spend on hiring a security company? C. And how long have your relied on this company or a company like it for protection? 1-20% 21-40% 41-60% 61-80% 1 2 3 4 In the last year Last five years Last ten years Last fifteen years Since before the transition, pre-1994 1 2 3 4 5 Other_________________________________________ 4 D. And how much of the Rarely time would you say that you call on this Often company when you Always have a problem? 1 32. If responses=2-6: A. Did you help to form this group? Yes 1 No 2 B. Please tell me your current level of involvement with this group Not a member Inactive member Active member Official leader 0 1 2 3 C. And how long have your relied on this group or a group like it for protection? In the last year Last five years Last ten years Last fifteen years 1 2 3 4 2 3 243! D. And how much of the time would you say that you call on this group when you have a problem? Since before the transition, pre-1994 Rarely 5 1 Often 2 Always 3 33. How satisfied are you with the performance of [write in name of private organization that respondent is most likely to rely on] in: Not at all satisfied A. Patrolling and securing your neighborhood? B. Arresting criminals? C. Punishing criminals? Not very satisfied Very satisfied 1 2 3 4 1 2 3 4 1 2 3 4 34. Is there any other type of group that you would turn to for help that was not listed above? Please do not give me specific names of groups. I am just interested in other types of groups that you may turn to for help that I did not list. 34a. I just have one more question for this section. Remember there is no right or wrong answer. Do you or anyone else in your household own a gun or firearm? ! Fairly satisfied 244! SECTION 6: LEGITIMACY OF NON-STATE SECURITY ORGANIZATIONS DO NOT READ Interviewer: I would like to ask you a few more questions about how you think we should deal with crime. 35. Do you agree or disagree that private organizations should have the right to patrol and secure local neighborhoods? Strongl y disagre e Disagree Neither agree nor disagree Agree Strongl y agree If respondent disagrees, SKIP the following question. 36. If response=4 or 5: Which organizations do you think should have the right to patrol and secure local neighborhoods? Allow respondent to choose more than one response 37. Do you agree or disagree that private organizations should have the right to arrest criminals? 1 2 3 4 Commercial security industry (i.e. ADT) Neighborhood watch scheme Street Committee Gang Vigilante group Community Police Forum Family members Traditional leader Some other group Strongl Disagre Neither Agree y e agree disagree nor disagree 5 1 2 3 4 5 6 7 8 9 Strongl y agree If respondent disagrees, skip the following question. 1 38. If response=4 or 5: which organizations do you think should have the right to arrest criminals? Allow respondent to choose more than one response 39. Do you agree or disagree that private organizations should have the right to punish those who break the law? If respondent disagrees, skip the ! 2 3 4 Commercial security industry (i.e. ADT) Neighborhood watch scheme Street Committee Gang Vigilante group Community Police Forum Family members Traditional leader Some other group Strongl Disagre Neither Agree y e agree disagree nor disagree 245! 5 1 2 3 4 5 6 7 8 9 Strongl y agree following question. 1 2 3 4 Commercial security industry (i.e. ADT) Neighborhood watch scheme Street Committee Gang Vigilante group Community Police Forum Family members Traditional leader Some other group 40. If response=4 or 5: which organizations do you think should have the right to punish criminals? Allow respondent to choose more than one response 5 1 2 3 4 5 6 7 8 9 SECTION 7: SOCIAL CAPITAL DO NOT READ Interviewer: Now I’d like to talk about your views on the area in which you live and your level of trust in other people. 41.How much do you trust each of the following: A. Your relatives? B. Other people in your neighborhood? Not at all 0 Just a little 1 Somewhat 2 A lot 3 0 1 2 3 42. Which of the following statements is closest to your view? Choose Statement 1 or Statement 2. 1: I live in a neighborhood where people 2: I live in a neighborhood where people do mostly help each other. not want to get involved. Agree very 2=Agree with 3=Agree with 4=Agree very strongly with Statement 1 Statement 2 strongly with Statement 1 Statement 2 2 3 1 4 Agree with neither (Do Not Read) 5 Don’t know (Do Not Read) 9 SECTION 8: RESPONDENT INFORMATION (CONT’D) DO NOT READ Interviewer: I just have a few more questions about you. 43. What is the highest level of education you have completed? ! No formal schooling Informal schooling only (including Koranic schooling), Some primary schooling Primary school completed 246! 0 1 2 3 Some secondary school/ high school Secondary school completed/high school completed Post-secondary qualifications, other than university e.g. a diploma or degree from polytechnic or college Some university University completed Post-graduate 44. Over the past year, how often, if ever, have you or anyone in your family gone without: A. Enough food to eat? B. Enough clean water for home use? C. Medicine or medical treatment? D. Enough fuel to cook your food? E. A cash income? 45. Do you have a job that pays a cash income? Is it full-time or part-time? Never Just once or twice 7 8 9 Many times Alwa ys 1 1 2 2 3 3 4 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 Yes, full time Yes, part time No 2 3 1 2 3 4 5 R0-1600 R1601-6400 R6401-25600 25601-102400 102401 or more 47. Are you married? Yes 1 48. How many children under the age of 18 live in your household? ! 6 0 0 1 46. Approximately how much money would you say you that you, personally, earn in one month? Several times 4 5 247! No 2 49. What is your occupation? 50. What is your religion, if any? _________________________________________ Fill in response Christian only 1 Jehovah’s Witness 17 Roman Catholic Orthodox Coptic, Anglican Lutheran Methodist 2 18 Presbyterian 8 Baptist 9 Quaker/Friends Mennonite 10 11 Dutch Reformed Evangelical Pentecostal Independent Zionist Christian Church 12 Seventh Day Adventist Mormon Muslim only Sunni only Ismaeli Mouridiya Brotherhood Tijaniya Brotherhood Qadiriya Brotherhood Shia only Traditional/ethnic religion, Hindu 13 14 15 16 Bahai Agnostic Athiest Other 29 30 31 32 Do Not Read Responses, Fill in Code for the Answer Given 3 4 5 6 7 51. How important is religion in your life? 24 25 26 27 28 Not at all important Not very important Somewhat important Very important 1 2 3 4 Asian other 52. Respondent’s Black/African White/European race (Interviewer ! 19 20 21 22 23 248! Coloured to record) 1 53. Respondent’s gender (Interviewer to record) 2 3 4 Male Female 1 2 5 SECTION 9: VICTIMIZATION DO NOT READ Interviewer: The final section of the interview will talk about your personal experiences with crime. Remember that you do not have to answer any questions that you are not comfortable answering. 54. Has anyone who lived in your present household been deliberately killed or murdered? 55. Has anyone ever taken something from you by using force or threatening you, or did anyone try to do so? 56. Have you personally ever been attacked, physically beaten or threatened by someone in a way that really frightened you, either at home or elsewhere, for example in a pub or a shebeen, in the street, at school, on public transport, in cinemas, on the beach or at your workplace? 57. Has anybody ever attempted to steal or actually stolen a car, van, truck or bakkie by force, when you or other members of your present household were inside or near the vehicle? No, have not experienced this 1 Yes, have experienced the this No, have not experienced this 1 Yes, have experienced this No, have not experienced this Yes, have experienced this 1 2 No, have not experienced this Yes, have experienced this 1 2 No, have Yes, have experienced this 58. People sometimes grab, touch or assault ! 249! 2 2 others for sexual reasons in a really offensive way, or sometimes they even rape others. This can happen either at home or elsewhere, for example in a pub or shebeen, in the street, at school, on public transport, in cinemas, on the beach or at your workplace. The person doing this could be someone you don’t know, or it could be a relative, friend or family member. Has anyone ever done this to you? 59. Has anyone ever broken into your dwelling/s without permission and stolen or tried to steal something? 60. Have you or any other member of your present household ever had any of their cars, vans, trucks or bakkies stolen when nobody was in the vehicle? 61. Have you or any other member of your present household ever had any of their motorcycles, motor scooters or similar vehicles stolen? 62. Have you or any other member of your present household ever had any of their bicycles stolen? 63. Have you or any member of your present household ever been a victim of theft of a car radio, or something else which was left in your car, or theft of a part of the car, such as a car mirror or wheel? ! not experienced this 1 2 No, have not experienced this 1 Yes, have experienced this No, have not experienced this 1 Yes, have experienced this No, have not experienced this 1 Yes, have experienced this No, have not experienced this 1 Yes, have experienced this No, have not experienced this Yes, have experienced this 1 2 250! 2 2 2 2 64. Have you or any other member of your present household ever had any of their animals stolen? 65. In some areas or countries, there is a problem of corruption among government or public officials. Has any government official, for instance a customs official, police officer or inspector ever asked you or wanted you to pay a bribe for his/her service? No, have not experienced this 1 Yes, have experienced this No, have not experienced the crime Yes, have experienced the crime 1 2 2 Closing Script Interviewer: Thank you very much for participating in this survey. Your answers greatly contributed to this research. If you have any further questions, please feel free to contact the project investigator at the numbers/e-mail address that you have been provided. ! 251! INTERVIEWER TO RECORD THE FOLLOWING AFTER THE INTERVIEW ENDS ENDTIME. Time interview started [Interviewer: Enter hour and minute, use 24 hr. clock] Hour Minute Interview Address:____________________________________________________ Was the house enclosed by a security gate? Was the house in a gated community? Was the house protected by a watchdog? Did the house have a commercial security sign posted (i.e. Chubbs/ADT)? Was there a police station near the house? Were there police visible in the community at the time of the interview? No Yes 1 2 No Yes 1 2 No Yes 1 2 No Yes 1 2 No Yes 1 2 No Yes 1 2 SUPERVISOR ONLY Household backchecked? ! No Yes 1 2 252! APPENDIX C: MASS SURVEY CONSENT FORM Research Participant Information and Consent Form You are being asked to participate in a research project. Researchers are required to provide a consent form to inform you about the study, to convey that participation is voluntary, to explain risks and benefits of participation, and to empower you to make an informed decision. You should feel free to ask the researchers any questions you may have. Study Title: Private Security and State Legitimacy in Contemporary South Africa Principal Investigator: Michael Bratton, Ph.D, University Distinguished Professor Department and Institution: Department of Political Science, Michigan State University Address and Contact Information: 323 South Kedzie Hall, Michigan State University East Lansing, MI 48824, Phone: 517-353-3377 Email: mbratton@msu.edu Co-principal Investigator: Danielle Carter, Doctoral Candidate Department and Institution: Department of Political Science, Michigan State University Address and Contact Information: 232 South Kedzie Hall, Michigan State University East Lansing, MI 48824, Phone: (443) 983-9963 Email: carte165@msu.edu PURPOSE OF RESEARCH: You are invited to participate in a research study that examines individual views of crime, security and the state in South Africa. This study is being conducted by Dr. Michael Bratton, University Distinguished Professor in the Department of Political Science at Michigan State University and Danielle Carter, doctoral candidate in the Department of Political Science at Michigan State University. The study is research towards a doctoral dissertation. ! 253! The purpose of the study is to investigate how individual reliance on private security shapes attitudes toward the legitimacy of the state. From this study, the researcher hopes to understand both the political causes and consequences of individual reliance on private security in South Africa. You have been selected as a possible participant in this study by chance. Four hundred thirty two South African citizens have been asked to participate in this study. Your answers will therefore be combined with the responses of hundreds of other individuals to generate an overall understanding of these issues. WHAT YOU WILL DO: At an agreed date, time and location, we will meet and I will ask you a number of questions during a face-to-face interview that will last approximately 30 minutes. The questions relate to your views on crime, security and the South African state. At the completion of the research study, I will provide you with an electronic copy of the key findings if you so desire. POTENTIAL BENEFITS: You may not benefit directly from your participation in the study. However, your participation in the study may help to shape crime policy in ways that are consistent with the values and interests of people in your community who grapple with issues of crime and insecurity in their everyday lives. POTENTIAL RISKS: For most participants the potential risks of participating in this study are very minimal. A series of steps will be taken to ensure the complete confidentiality of your responses, and these steps will significantly reduce the already minimal risks of participating in the study. PRIVACY AND CONFIDENTIALITY: It is important to note that your confidentiality will be protected to the maximum extent allowable by law. To ensure your confidentiality and the confidentiality of the data collected through the study, the following procedures will be followed. First, to guarantee your privacy, all interviews will be conducted in an area of your house or community that is away from other people. All conversations that occur before and after the interview will be kept strictly confidential. During the interview, your responses will be recorded on a questionnaire sheet. To ensure the confidentiality of your responses, a de-identification process will be used to confidentially link you with your responses. In all notes and data your identification code will be used in place of your name. Moreover, the electronic key that links your identification code with your name will be stored in a password protected document for at least 3 years and then ! 254! destroyed. All completed questionnaires will be securely stored in a locked file cabinet at the secondary investigator’s office during the data collection period. The principal investigator and Institutional Review Board (IRB) staff at Michigan State University are the only parties that will be allowed to see the data at this stage of the process. Raw study data will be kept for at least three years after the project closes in a locked file cabinet at Michigan State University. Thereafter, the electronic key and all questionnaires will be destroyed. The results of this study may be published or presented at professional meetings, but the identity of all research participants will remain confidential. Moreover, data repositories such as ICPSR at the University of Michigan, Ann Arbor may have access to this data, but the electronic database will only include the respondent’s assigned ID number. YOUR RIGHT TO PARTICIPATE, SAY NO, OR WITHDRAW: Your participation in this study is voluntary. You may choose not to participate, and you may withdraw at any time during the questionnaire without any penalty to you. In addition, you may choose not to answer any question with which you are not comfortable. COSTS AND COMPENSATION FOR BEING IN THE STUDY: There are no costs involved in participating in this research. You will not receive money or any other form of compensation for participating in this study. CONTACT INFORMATION FOR QUESTIONS AND CONCERNS Thank you for your time and contribution. If you have concerns or questions about this study, such as scientific issues, how to do any part of it, or to report an injury (i.e. physical, psychological, social, financial, or otherwise), please contact the principle researcher, Michael Bratton, by phone 517-353-3377, fax 517-432-1091, email mbratton@msu.edu or regular mail 323 South Kedzie Hall, Michigan State University East Lansing, MI 48824. Or, you may contact the secondary researcher, Danielle Carter, by phone 443-983-9963, fax (517) 432-1091, email carte165@msu.edu or regular mail 232 South Kedzie Hall, Michigan State University, East Lansing MI, 48823. If you have questions or concerns about your role and rights as a research participant, would like to obtain information or offer input, or would like to register a complaint about this study, you may contact, anonymously if you wish, the Michigan State University’s Human Research Protection Program by phone 517-355-2180, fax 517-432-4503, e-mail irb@msu.edu or regular mail at 207 Olds Hall, MSU, East Lansing, MI 48824. DOCUMENTATION OF INFORMED CONSENT. Your signature below means that you voluntarily agree to participate in this research study. ! 255! ________________________________________ Signature Date ________________________________________ You will be given a copy of this form to keep. ! _____________________________ 256! APPENDIX D: ELITE INTERVIEW CONSENT FORM Research Participant Information and Consent Form You are being asked to participate in a research project. Researchers are required to provide a consent form to inform you about the study, to convey that participation is voluntary, to explain risks and benefits of participation, and to empower you to make an informed decision. You should feel free to ask the researchers any questions you may have. Study Title: Private Security and State Legitimacy in Contemporary South Africa Principal Investigator: Michael Bratton, Ph.D, University Distinguished Professor Department and Institution: Department of Political Science, Michigan State University Address and Contact Information: 323 South Kedzie Hall, Michigan State University East Lansing, MI 48824, Phone: 517-353-3377 Email: mbratton@msu.edu Co-principal Investigator: Danielle Carter, Doctoral Candidate Department and Institution: Department of Political Science, Michigan State University Address and Contact Information: 229 South Kedzie Hall, Michigan State University East Lansing, MI 48824, Phone: (443) 983-9963 Email: carte165@msu.edu PURPOSE OF RESEARCH: You are invited to participate in a research study that examines individual views of crime, security and the state in South Africa. This study is being conducted by Dr. Michael Bratton, University Distinguished Professor in the Department of Political Science at Michigan State University and Danielle Carter, doctoral candidate in the Department of Political Science at Michigan State University. The study is research towards a doctoral dissertation. ! 257! The purpose of the study is to investigate how individual reliance on private security shapes attitudes toward the legitimacy of the state. From this study, the researcher hopes to understand the political consequences of private security provision (including commercial security companies and civil society groups) in South Africa. Data for this study will be collected from two primary sources. At the mass level, I will conduct a survey with a total of four hundred thirty two South African citizens. At the elite level, this research will involve interviews with elected officials (Members of Parliament and Local Government Councilors), police officials, and members of commercial security companies and other civil society groups that organize around security issues. You have been selected as a possible participant in this study because you are a member of one of these key security stakeholder groups and I believe that you will bring important qualitative insights to bear on this research. WHAT YOU WILL DO: At an agreed date, time and location, we will meet and I will ask you a number of questions during a face-to-face interview that will last approximately 45 minutes. The questions relate to your views on crime and security in contemporary South Africa and how these issues are being addressed at the state and societal levels. At the completion of the research study, I will provide you with an electronic copy of the key findings if you so desire. POTENTIAL BENEFITS: You may not benefit directly from participating in this study. However, the overall findings may be of help to elected officials and civil society groups who are interested in understanding citizen perceptions of the proper role of state and society in security provision. POTENTIAL RISKS: The potential risks of participating in this study are very minimal. A series of steps will be taken to ensure the complete confidentiality of your responses, and these steps will significantly reduce the already minimal risks of participating in the study. PRIVACY AND CONFIDENTIALITY: It is important to note that your confidentiality will be protected to the maximum extent allowable by law. To ensure your confidentiality and the confidentiality of the data collected through the study, the following procedures will be followed. First, to guarantee your privacy, all interviews will be conducted in your office. All conversations that occur before and after the interview will be kept strictly confidential. During the interview, your responses will be recorded by hand. To ensure the confidentiality of your responses, a de-identification process will be used to ! 258! confidentially link you with your responses. In all reports and papers, your identification code will be used in place of your name. Moreover, the electronic key that links your identification code with your name will be stored in a password protected document for at least 3 years and then destroyed. All hand-written notes will be securely stored in a locked file cabinet at the secondary investigator’s office during the data collection period. The principal investigator and Institutional Review Board (IRB) staff at Michigan State University are the only parties that will be allowed to see the data at this stage of the process. Raw study data will be kept for at least three years after the project closes in a locked file cabinet at Michigan State University. Thereafter, the electronic key and all hard copies of notes will be destroyed. The results of this study may be published or presented at professional meetings, but the identity of all research participants will remain confidential. YOUR RIGHT TO PARTICIPATE, SAY NO, OR WITHDRAW: Your participation in this study is voluntary. You may choose not to participate, and you may withdraw at any time during the interview without any penalty to you. In addition, you may choose not to answer any question with which you are not comfortable. COSTS AND COMPENSATION FOR BEING IN THE STUDY: There are no costs involved in participating in this research. You will not receive money or any other form of compensation for participating in this study. CONTACT INFORMATION FOR QUESTIONS AND CONCERNS Thank you for your time and contribution. If you have concerns or questions about this study, such as scientific issues, how to do any part of it, or to report an injury (i.e. physical, psychological, social, financial, or otherwise), please contact the principle researcher, Michael Bratton, by phone 517-353-3377, fax 517-432-1091, email mbratton@msu.edu or regular mail 323 South Kedzie Hall, Michigan State University East Lansing, MI 48824. Or, you may contact the secondary researcher, Danielle Carter, by phone 443-983-9963, fax (517) 432-1091, email carte165@msu.edu or regular mail 232 South Kedzie Hall, Michigan State University, East Lansing MI, 48823. If you have questions or concerns about your role and rights as a research participant, would like to obtain information or offer input, or would like to register a complaint about this study, you may contact, anonymously if you wish, the Michigan State University’s Human Research Protection Program by phone 517-355-2180, fax 517-432-4503, e-mail irb@msu.edu or regular mail at 207 Olds Hall, MSU, East Lansing, MI 48824. DOCUMENTATION OF INFORMED CONSENT. ! 259! Your signature below means that you voluntarily agree to participate in this research study. ________________________________________ Signature Date ________________________________________ You will be given a copy of this form to keep. ! _____________________________ 260! APPENDIX E: QUESTION WORDING AND CODES FOR TABLES 5.5-5.6 Table 8.2: Question Wording and Codes for Tables 5.5-5.6 Variable Name Question Text Codes State Legitimacy, 3 Item Index, One unrotated factor explains 62% of the variance (eigenvalue=1.867), alpha=.692 Police Legitimacy Court Legitimacy Please tell me whether you agree or disagree: The police always have the right to make people obey the law? The courts have the right to make decisions that people always have to abide by? The tax department always has the right to make people pay taxes? Tax Legitimacy Non-State Security State Arranger Please tell me who you would call on if you or someone in your family had been a victim of crime? !!See!www.afrobarometer.org! ! 261! 96 AB Survey Original values: 0=Strongly disagree 1=Disagree 2=Neither agree nor disagree 3=Strongly agree 4=Agree Created an index using the mean of 3 items then, used the index to created a binary variable with values<3=0, values equal to or >3=1 Police only=0 “Police and some other group” =0 if call police first; =1 if call other groups first “Other groups only”=1 Which of the following statements is closest Agree with/agree to your view? Choose Statement 1 or very strongly with Statement 2. 1: The state should take the 1=0 main responsibility for protecting citizens Agree with/agree from crime. very strongly with 2: The state should mainly provide support to 2=1 private groups such as the commercial security industry or neighborhood watches so that they may protect citizens from crime. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 96 Source New questio n New questio n Table 8.2 (cont’d) Variable Name Question Text Codes Source Contact Victimization, 4 Item Index, One unrotated factor explains Victims of Crime 47% of the variance (eigenvalue=1.870), alpha=.619 Survey Murder Has anyone who lived in your No, have present household been not deliberately killed or murdered? experienced this: 0 Yes, have experienced Robbery Has anyone ever taken this: 1 something from you by using force or threatening you, or did anyone try to do so? Assault Have you personally ever been attacked, physically beaten or threatened by someone in a way that really frightened you, either at home or elsewhere, for example in a pub or a shebeen, in the street, at school, on public transport, in cinemas, on the beach or at your workplace? Rape People sometimes grab, touch or assault others for sexual reasons in a really offensive way, or sometimes they even rape others. This can happen either at home or elsewhere, for example in a pub or shebeen, in the street, at school, on public transport, in cinemas, on the beach or at your workplace. The person doing this could be someone you don’t know, or it could be a relative, friend or family member. Has anyone ever done this to you? Police Performance Evaluations, 3 Item Index, One unrotated factor explains 88% of the variance (eigenvalue=2.638), alpha=.931 ! 262! Adapted from Victims of Crime Survey Table 8.2 (cont’d) Variable Name Prevent Crime Catch Criminals Question Text Codes How well do you think the police Very badly: 0 are doing at the following activities: Fairly badly: 1 preventing crime? Fairly well: 2 Very well: 3 Source How well do you think the police are doing at the following activities: catching criminals? How well do you think the police are doing at the following activities: making people feel safe? Make Safe Institutional Trust, 2 Item Construct, One unrotated factor explains 78% of the variance (eigenvalue=1.553), alpha=.710 Trust Police How much would you say you Not at all: 0 trust: the police? Just a little: 1 Somewhat: 2 How much would you say you A lot: 3 Trust Courts trust: courts? Rule of Law, 3 Item Index, One unrotated factor explains 47% of the variance (eigenvalue=1.398), alpha=.425 Unequal And with regards to enforcement of Never: 3 Treatment the law, In your opinion, how often, Rarely: 2 in this country: are people treated Often: 1 unequally under the law? Always: 0 Officials Unpunished And with regards to enforcement of the law, In your opinion, how often, in this country: do officials who commit crimes go unpunished? Ordinary People Unpunished And with regards to enforcement of the law, In your opinion, how often, in this country: do ordinary people who break the law go unpunished? ! 263! AB!Survey AB!Survey Table 8.2 (cont’d) Variable Name Support for Democracy Question Text Codes Which of these three statements is closest to your own opinion? Democracy preferable: 1 Statement 1: Democracy is preferable to any other Sometimes kind of government. nonStatement 2: In some circumstances, a nondemocratic democratic government can be preferable. preferable: Statement 3: For someone like me, it doesn’t matter 0 what kind of government we have. AB! Survey Political Participation With regard to the most recent, 2009 national elections, which statement is true for you? AB! Survey Age How old are you? Did not vote: 0 Voted: 1 018-081 Male Respondent’s gender Male: 1 Female: 0 AB! Survey Lived Poverty, 5 Item Index, One unrotated factor explains 69% of the variance (eigenvalue=3.439), alpha=.872 Food Over the past year, how often, if ever, have you or Never: 0 anyone in your family gone without: Enough food to Just once eat? or twice: 1 Clean Water Several Over the past year, how often, if ever, have you or times: 2 anyone in your family gone without: Enough clean Many Medicine water for home use? times: 3 Always: 4 Over the past year, how often, if ever, have you or anyone in your family gone without: Medicine or Fuel medical treatment? Cash Income Over the past year, how often, if ever, have you or anyone in your family gone without: Enough fuel to cook your food? Over the past year, how often, if ever, have you or anyone in your family gone without: A cash income? ! 264! Source AB! Survey AB! Survey Table 8.2 (cont’d) Variable Name Race Question Text Respondent’s race Education ! Source AB! Survey Categorical Black White Coloured Asian Categorical City Employed Codes Do you have a job that pays a cash income? Is it full-time or part-time? What is the highest level of education you have completed? AB! Survey Cape Town Johannesburg Durban Yes: 1 No: 0 No formal schooling Informal schooling only (including Koranic schooling), Some primary schooling Primary school completed Some secondary school/ high school Secondary school completed/high school completed Post-secondary qualifications, other than university e.g. a diploma or degree from polytechnic or college Some university University completed Post-graduate 265! AB! Survey 0 1 2 3 4 5 6 7 8 9 AB! Survey APPENDIX F: QUESTION WORDING AND CODES FOR TABLES 5.7 AND 7.1 Table 8.3: Question Wording and Codes for Tables 5.7 and 7.1 Variable Name Question Text Codes State Legitimacy, 3 Item Index, One unrotated factor explains 65% of the variance (eigenvalue=1.962), alpha=.733 Please tell me whether you Original values: agree or disagree: 0=Strongly disagree The police always have the 1=Disagree Police Legitimacy right to make people obey the 2=Neither agree nor law? The courts have the right disagree Court Legitimacy to make decisions that people 3=Strongly agree always have to abide by? The 4=Agree Tax Legitimacy tax department always has the right to make people pay Created an index using the taxes? mean of 3 items then, Used the index to created a binary variable with values<3=0, values equal to or >3=1 Non-State Security If you were a victim of crime in this country, who, if anyone, would you go to first for assistance? Contact Victimization Categorical The police=0 Security service/traditional leader/ street committee/ powerful local person or gang/personally take revenge/join with others to take revenge/own family or friends/family of perpetrator)=1 During the past year, have you No: 0 or anyone in your family: been Yes, once: 1 physically attacked? Yes, twice: 2 Yes, three or more times: 3 Economic Performance Evaluations, 4 Item Index, One unrotated factor explains 61% of the variance (eigenvalue=2.436), alpha=.783 ! 266! Table 8.3 (cont’d) Variable Name Creating jobs? Keeping prices down? Narrowing gaps between rich and poor? Ensuring everyone has enough Question Text How well or badly would you say the current government is handling the following matters, or haven’t you heard enough to say: Codes Very badly: 0 Fairly badly: 1 Fairly well: 2 Very well: 3 Creating jobs? Keeping prices down? Narrowing gaps between rich and poor? Ensuring everyone has enough to eat? Institutional Trust, 2 Item Construct, One unrotated factor explains 64% of the variance (eigenvalue=1.293), alpha=.453 How much do you trust each Not at all: 0 of the following, or haven’t Just a little: 1 you heard enough about them Somewhat: 2 to say: A lot: 3 ! Trust Police The police? Trust Courts Transgression of Rule of Law Courts of law? In your opinion, how often, in this country: do officials who commit crimes go unpunished? Satisfaction with Democracy Overall, how satisfied are you with the way democracy works in South Africa today? Political Participation With regard to the most recent, 2009 national elections, which statement is true for you? 267! Never: 0 Rarely: 1 Often: 2 Always: 3 Not at all satisfied: 0 Not very satisfied: 1 Fairly satisfied:2 Very satisfied: 3 Did not vote: 0 Voted: 1 Table 8.3 (cont’d) Variable Name Age Male Question Text How old are you? Respondent’s gender Codes 018-087 Male: 1 Female: 0 Lived Poverty, 5 Item Index, One unrotated factor explains 58% of the variance (eigenvalue=2.938), alpha=.815 Over the past year, how often, if ever, have you Never: 0 or anyone in your family gone without: Just once or twice: 1 Food Enough food to eat? Several times: Clean Water Enough clean water for home use? 2 Medicine Medicine or medical treatment? Many times: 3 Fuel Enough fuel to cook your food? Always: 4 Cash Income A cash income? Race Respondent’s race Categorical Do you have a job that pays a cash income? Is it full-time or part-time? Black White Coloured Indian Yes: 1 No: 0 Employed Province Categorical Eastern Cape Western Cape Northern Cape North West Gauteng Mpumalanga Limpopo Free State KwaZuluNatal ! 268! Table 8.3 (cont’d) Education ! What is the highest level of education you have completed? 269! No formal schooling Informal schooling only (including Koranic schooling), Some primary schooling Primary school completed Some secondary school/ high school Secondary school completed/high school completed Post-secondary qualifications, other than university e.g. a diploma or degree from polytechnic or college Some university University completed Post-graduate 0 1 2 3 4 5 6 7 8 9 APPENDIX G: QUESTION WORDING AND CODES FOR TABLES 6.1-6.2, AND 7.2-7.3 Table 8.4: Question Wording and Codes for Tables 6.1-6.2, and 7.2-7.3 Variable Name Question Text Joining Now I am going to read out a list of groups that people join or attend. For each one, could you tell me whether you are an official leader, an active member, an inactive member, or not a member: Some other voluntary association or community group? Codes 0=Not a member or inactive member 1=Official leader or active member Collective action, 2 Item Construct, One unrotated factor explains 85% of the variance (eigenvalue=1.702), alpha=.825 Here is a list of actions that people sometimes take as 0=No, would citizens. For each of these, please tell me whether you, never do this personally, have done any of these things during the past 1=No, year. If not, would you do this if you had the chance: Would if had Attended a community meeting? Got together with others the chance to raise an issue? 2=Yes, Once or twice 3=Yes, Several times 4=Yes, Often Contacting During the past year, how often have you contacted any of the following persons about some important problem or to give them your views: A local government councilor? A Member of Parliament? Protesting, 5 Item Index, One unrotated factor explains 61% of the variance (eigenvalue=3.166), alpha=.850 Here is a list of actions that people sometimes take as citizens. For each of these, please tell me whether you, personally, have done any of these things during the past year. If not, would you do this if you had the chance: Refused to pay for services provided by government like water, electricity or property rates? Refused to pay a tax or fee to government? Attended a demonstration or protest march? Gone on strike in order to demand a higher salary or better working conditions? Used force or violence for a political? ! 270! 0=Never 1=Only once 2=A few times 3=Often 0=No, would never do this 1=No, Would if had the chance 2=Yes, Once or twice 3=Yes, Several times 4=Yes, Often Table 8.4 (cont’d) Variable Name Voting Question Text With regard to the most recent national election in 2009, which statement is true for you: You were too young to vote You were not registered to vote You voted in the elections You decided not to vote You could not find the polling station You were prevented from voting You did not have time to vote You did not vote because you could not find your name in the voters’ register Did not vote for some other reason Non-State Security 1= You voted in the elections The police=0 Security service/traditional leader/ street committee/ powerful local person or gang/personally take revenge/join with others to take revenge/own family or friends/family of perpetrator)=1 State Legitimacy, 3 Item Index, One unrotated factor explains 65% of the variance (eigenvalue=1.962), alpha=.733 Please tell me whether you agree or Original values: disagree: 0=Strongly disagree The police always have the right to 1=Disagree Police Legitimacy make people obey the law? The 2=Neither agree nor courts have the right to make disagree Court Legitimacy decisions that people always have to 3=Strongly agree abide by? The tax department always 4=Agree Tax Legitimacy has the right to make people pay Created an index using the taxes? mean of 3 items then, used the index to created a binary variable with values<3=0, and values equal to or >3=1 ! If you were a victim of crime in this country, who, if anyone, would you go to first for assistance? Codes 0= You were too young to vote/You were not registered to vote/You decided not to vote/You could not find the polling station /You were prevented from voting/You did not have time to vote /You did not vote because you could not find your name in the voters’ register /Did not vote for some other reason ? 271! Table 8.4 (cont’d) Variable Name Contact Victimization Question Text During the past year, have you or anyone in your family: Been physically attacked? Property Victimization During the past year, have you or anyone in your family: Had something stolen from your house? How much do you trust each of the following types of people: Your neighbors? Over the past year, how often, if ever, have you or anyone in your family: Felt unsafe walking in your neighbourhood? Social Trust Personal Insecurity Political Order Codes 0=No 1=Yes, once 2=Yes, twice 3=Yes, three or more times 0=No 1=Yes, once 2=Yes, twice 3=Yes, three or more times 0=Not at all/Just a little 1=Somewhat/ a lot 0=never 1==just once or twice, 2=several times 3=many times 4=always 0=Very Badly 1=Fairly Badly 2=Fairly Well 3=Very Well How well or badly would you say the current government is handling the following matters, or haven’t you heard enough to say: Reducing crime? Performance Evaluations: Economic Goods, 4 Item Index, One unrotated factor explains 61% of the variance (eigenvalue=2.436), alpha=.783 How well or badly would you 0=Very Badly say the current government is 1=Fairly Badly handling the following matters, 2=Fairly Well or haven’t you heard enough to 3=Very Well say: Creating jobs? Keeping prices down? Narrowing gaps between rich and poor? Ensuring everyone has enough to eat? Political Efficacy, 2 Item Construct, One unrotated factor explains 83% of the variance (eigenvalue=1.661), alpha=.796 How much of the time do you 0=Never think the following try their best 1=Only sometimes to listen to what people like you 2=Often have to say: Members of 3=Always Parliament? Local government councilors? ! 272! Table 8.4 (cont’d) Variable Name ANC Partisan (Tables 6.1-6.2 only) Discuss Politics (PS1) Interest in Public Affairs (PS2) Question Text Which party is that? (The Preceding question asks ”Do you feel close to any particular political party? If the respondent answered “yes” they were asked this question next). Codes 1= African National Congress (ANC) African Christian Democratic Party (ACDP) African Muslim Party (AMP) African National Congress (ANC) Azanian People’s Organization (AZAPO) Congress of the People Democratic Alliance (DA) Freedom Front Plus Vryheidsfront Plus (VF Plus) Independent Democrats (ID) Inkatha Freedom Party (IFP)/Minority Front (MF) National Democratic Convention (NADECO) New National Party Nuwe Nasionale Party (NNP) Pan Africanist Congress (PAC) United Christian Democratic Party (UCDP) United Democratic Movement (UDM) United Independent Front (UIF) When you get together with your friends or family, would you say you discuss political matters: Frequently, occasionally, or never? 0=All others How interested would you say you are in public affairs? 0=Not at all interested 1=Not very interested 2=Somewh at interested 3=Very interested 0=Never 1=Occasio nally 2=Frequent ly Media Exposure, 4 Item Index, One unrotated factor explains 42% of the variance (eigenvalue=1.695), alpha=.543 ! 273! Table 8.4 (cont’d) Variable Name Question Text How often do you get news from the following sources: Radio? Television? Newspapers? Internet? Age How old are you? Male Respondent’s gender Codes 0=Never 1=Less than once a month 2=A few times a month 3=A few times a week 4=Everyday 018-081 0=Female 1=Male Lived Poverty, 5 Item Index, One unrotated factor explains 59% of the variance (eigenvalue=2.938), alpha=.815 Over the past year, how often, if 0=never ever, have you or anyone in your 1==just once or twice, family gone without: Enough 2=several times food to eat? Enough clean water 3=many times for home use? Medicines or 4=always medical treatment? Enough fuel to cook your food? A cash income? Race Respondent’s race Categorical Black White Coloured Asian Categorical Province ! Urban Urban or rural Employed Do you have a job that pays a cash income? Is it full-time or part-time? 274! Eastern Cape Western Cape Northern Cape North West Gauteng Mpumalanga Limpopo Free State KwaZulu-Natal 0=rural 1=urban 0=No: 0 1=1 Table 8.4 (cont’d) Variable Name Education ! Question Text What is the highest level of education you have completed? 275! Codes 0=No formal schooling 1=Informal schooling only (including Koranic schooling) 2=Some primary schooling 3=Primary school completed 4=Some secondary school/ high school 5= Secondary school completed/high school completed 6= Post-secondary qualifications, other than university e.g. a diploma or degree from polytechnic or college 7= Some university 8= University completed 9= Post-graduate REFERENCES ! 276! REFERENCES Adler, G., and J. Steinberg. 2000. "Introduction: From comrades to citizens." From Comrades to Citizens: The South African Civics Movement and the Transition to Democracy:1-25. Almond, Gabriel A, and Sidney Verba. "ur.(1963) The civic culture: Political attitudes and democracy in five nations." Boston, itd: Little, Brown and Co. Altbeker, A. 2005. "Puzzling statistics: Is South Africa really the world’s crime capital." SA Crime Quarterly 11:1-8. Altbeker, A. 2007. A country at war with itself: South Africa's crisis of crime: Jonathan Ball Pub. Anderson, CJ. 2000. "Economic voting and political context: a comparative perspective." Electoral Studies 19 (2-3):151-170. Authority, Private Security Industry Regulatory. 2010/2011. "Private Security Industry Regulatory Authority Annual Report." Baker, B. 2002. "Living with non-state policing in South Africa: the issues and dilemmas." The Journal of Modern African Studies 40 (1):29-53. Baker, B. 2008. "Multi-choice policing in Africa." Bartels, Larry M. 2000. "Partisanship and voting behavior, 1952-1996." American Journal of Political Science:35-50. Bateson, R. 2012. "Crime Victimization and Political Participation." American Political Science Review 106 (3):570-587. Batley, Richard, and Claire Mcloughlin. 2009. "State capacity and non-state service provision in fragile and conflict-affected states." Bayart, J.F., S. Ellis, and B. Hibou. 1999. The criminalization of the state in Africa: International African Institute. Bayley, D.H., and C.D. Shearing. 1996. "The future of policing." Law and society review:585606. Blattman, Christopher. 2009. "From Violence to Voting: War and political participation in Uganda." American Political Science Review 103 (2):231-247. Booth, J.A., and M.A. Seligson. 2005. "Political legitimacy and participation in Costa Rica: Evidence of arena shopping." Political Research Quarterly 58 (4):537-550. ! 277! Booth, John A, and Mitchell A Seligson. 2009. The legitimacy puzzle in Latin America: Political support and democracy in eight nations: Cambridge University Press Cambridge. Bratton, M, and ECC Chang. 2006. "State building and democratization in sub-Saharan Africa: forwards, backwards, or together?" Comparative Political Studies 39 (9):1059. Bratton, M., M. Coulibaly, and F. Machado. 2002. "Popular views of the legitimacy of the state in Mali." Canadian Journal of African Studies:197-238. Bratton, M., and R. Mattes. 2001. "Support for democracy in Africa: Intrinsic or Instrumental?" British Journal of Political Science 31 (3):447-474. Bratton, Michael, Ravi Bhavnani, and Tse-Hsin Chen. 2012. "Voting intentions in Africa: ethnic, economic or partisan?" Commonwealth & Comparative Politics 50 (1):27-52. Bratton, Michael, and Mwangi S Kimenyi. 2008. "Voting in Kenya: putting ethnicity in perspective." Journal of Eastern African Studies 2 (2):272-289. Bratton, Michael, Robert B. Mattes, and Emmanuel Gyimah-Boadi. 2005. Public opinion, democracy, and market reform in Africa. 1 vols. Cambridge ; New York: Cambridge University Press. Brogden, M. 2004. "Commentary: community policing: a panacea from the West." African Affairs 103 (413):635. Brogden, M., and C.D. Shearing. 1993. Policing for a new South Africa: Psychology Press. Bundy, C., G. Adler, and J. Steinberg. 2000. "Survival and Resistance: Township Organizations and Non-Violent Direct Action in Twentieth Century South Africa." From Comrades to Citizens: The South African Civics Movement and the Transition to Democracy:26-51. Burns, Nancy, Kay Lehman Schlozman, and Sidney Verba. 2001. The private roots of public action: Gender, equality, and political participation: Harvard University Press. Burris, S., P. Drahos, and C. Shearing. 2005. "Nodal governance." Austl. J. Leg. Phil. 30:30. Buur, L. 2005. "The sovereign outsourced: local justice and violence in Port Elizabeth." Sovereign Bodies: Citizens, Migrants and States in the Postcolonial World:153-171. Buur, L., and S. Jensen. 2004. "Introduction: vigilantism and the policing of everyday life in South Africa." African Studies 63 (2):139-152. Buur, Lars. 2010. "Domesticating Sovereigns: The Changing Nature of Vigilante Groups in South Africa." In Domesticating Vigilantism in Africa, ed. T. K. a. T. Gratz. Woodbridge: Boydell and Brewer. ! 278! Campbell, James E, Mary Munro, John R Alford, and Bruce A Campbell. 1986. "Partisanship and voting." Research in micropolitics 1:99-126. Carothers, T. 1998. "Rule of Law Revival, The." Foreign Aff. 77:95. Cho, Wonbin, and Michael Bratton. 2006. "Electoral institutions, partisan status, and political support in Lesotho." Electoral Studies 25 (4):731-750. Davidson, William B, and Patrick R Cotte. 2006. "Sense of community and political participation." Journal of Community Psychology 17 (2):119-125. Delhey, Jan, and Kenneth Newton. 2003. "Who trusts?: The origins of social trust in seven societies." European Societies 5 (2):93-137. Diamond, L., and L. Morlino. 2004. "The quality of democracy." Journal of Democracy 15 (4):20-31. Dixon, Bill, and Lisa-Marie Johns. 2001. "Gangs, PAGAD & the state: Vigilantism and revenge violence in the Western Cape." Dixon, Bill, Wilfried Schärf, and Elrena Van der Spuy. 2003. Justice gained? : crime and crime control in South Africa's transition. Cullompton: Willan. Dunning, Thad, and Lauren Harrison. 2010. "Cross-cutting cleavages and ethnic voting: An experimental study of cousinage in Mali." American Political Science Review 104 (1):119. Easton, David. 1957. Approach to the Analysis of Political Systems: Cambridge Univ Press. Easton, David. 1975. "A re-assessment of the concept of political support." British Journal of Political Science 5 (4):435-457. Emmett, Tony, and Alexander Butchart. 2000. Behind the mask : getting to grips with crime and violence in South Africa. Pretoria: HSRC Publishers. Englebert, P. 2002. State legitimacy and development in Africa: Lynne Rienner Pub. Fernandez, K., and M. Kuenzi. 2008. Crime, Institutional Trust and Support for Democracy in Africa. Ferree, Karen E. 2006. "Explaining South Africa's racial census." Journal of Politics 68 (4):803815. Finkel, Steven E. 1985. "Reciprocal effects of participation and political efficacy: A panel analysis." American Journal of Political Science:891-913. ! 279! "Flamethrower Now an Option on South African Cars." 1998. CNN. Gibson, J.L., and G.A. Caldeira. 2003. "Defenders of democracy? Legitimacy, popular acceptance, and the South African Constitutional Court." Journal of Politics 65 (1):1-30. Gibson, J.L., and A. Gouws. 1997. "Support for the rule of law in the emerging South African democracy*." International Social Science Journal 49 (152):173-191. Gilley, Bruce. 2006. "The Determinants of State Legitimacy: Results for 72 Countries." International Political Science Review 27 (47):24. Gottschalk, Keith. 2011. "Vigilantism v. the State: A case study of the rise and fall of Pagad, 1996–2000." Harris, B. 2003. "Spaces of violence, places of fear: Urban conflict in post-apartheid South Africa." paper presented on the Conflicts and Urban Violence panel, Foro Social Mundial Tematico:16-20. Henn, Matt, Mark Weinstein, and Dominic Wring. 2002. "A generation apart? Youth and political participation in Britain." The British Journal of Politics & International Relations 4 (2):167-192. Herbst, Jeffrey Ira. 2000. States and power in Africa : comparative lessons in authority and control. Princeton, N.J.: Princeton University Press. Hero, Rodney E. 1986. "Explaining Citizen-Initiated Contacting of Government Officials: Socioeconomic Status, Perceived Need, or Something Else?" Social Science Quarterly 67 (3):626-635. Hirschmann, David. 1991. "Women and political participation in Africa: broadening the scope of research." World Development 19 (12):1679-1694. Hobbes, T. 1998. Leviathan: Oxford University Press, USA. Howe, Paul. 2006. "Political knowledge and electoral participation in the Netherlands: Comparisons with the Canadian case." International Political Science Review 27 (2):137166. Huntington, Samuel P. 1968. Political order in changing societies. New Haven,: Yale University Press. Hydén, Göran. 2006. "African politics in comparative perspective." Cambridge ; New York: Cambridge University Press. ! 280! Irish, J. 1999. Policing for Profit: the future of South Africa's private security industry: Institute for Security Studies. Iwanaga, Kazuki, ed. 2008. Women's Political Participation and Representation in Asia: Obstacles and Challenges. Copenhagen: NIAS. Jackman, R.W. 1993. Power without force: The political capacity of nation-states: University of Michigan Press. Jackson, RH, and CG Rosberg. 1982. "Why Africa's weak states persist: the empirical and the juridical in statehood." World Politics: A Quarterly Journal of International Relations:124. Jensen, S. 2008. Gangs, politics & dignity in Cape Town: James Currey Oxford. Johnston, L., and C. Shearing. 2003. "Governing security." Explorations in Policing and Justice, London. Kaase, Max. 1999. "Interpersonal trust, political trust and non‐institutionalised political participation in Western Europe." West European Politics 22 (3):1-21. Kempa, M., R. Carrier, J. Wood, and C. Shearing. 1999. "Reflections of the Evolving Concept of'Private Policing'." European Journal on Criminal Policy and Research 7 (2):197-223. Kenski, Kate, and Natalie Jomini Stroud. 2006. "Connections between Internet use and political efficacy, knowledge, and participation." Journal of Broadcasting & Electronic Media 50 (2):173-192. Kramon, Eric. 2009. "Vote-Buying and Political Behavior: Estimating and Explaining VoteBuying’s Effect on Turnout in Kenya." Afrobarometer Working Paper. Kuenzi, Michelle T. 2006. "Nonformal education, political participation, and democracy: Findings from Senegal." Political Behavior 28 (1):1-31. Kynoch, G. 2005. "Crime, conflict and politics in transition-era South Africa." African Affairs 104 (416):493. Leighley, Jan E. 1995. "Attitudes, opportunities and incentives: A field essay on political participation." Political Research Quarterly 48 (1):181-209. Lemanski, C. 2004. "A new apartheid? The spatial implications of fear of crime in Cape Town, South Africa." Environment and Urbanization 16 (2):101. Levi, M., A. Sacks, and T. Tyler. 2009. "Conceptualizing legitimacy, measuring legitimating beliefs." American Behavioral Scientist 53 (3):354. ! 281! Lewis-Beck, MS, and M Paldam. 2000. "Economic voting: an introduction." Electoral Studies 19 (2-3):113-121. Lijphart, Arend, and Carlos H. Waisman. Institutional design in new democracies : Eastern Europe and Latin America. Westview Press 1996 [cited. Linz, Juan J., and Alfred C. Stepan. 1996. Problems of democratic transition and consolidation : southern Europe, South America, and post-communist Europe. Baltimore: Johns Hopkins University Press. Lipset, SM. 1959. "Some social requisites of democracy: Economic development and political legitimacy." The American Political Science Review:69-105. Locke, J. 1966. The second treatise of civil government and a letter concerning toleration: Oxford, Blackwell. Louw, A. "Citizen Perceptions of Crime and Policing in South Africa." Johannesburg:197. Louw, A. 1997. "Surviving the transition: Trends and perceptions of crime in South Africa." Social indicators research 41 (1):137-168. Lowndes, Vivien. 2004. "Getting on or getting by? Women, social capital and political participation." The British Journal of Politics & International Relations 6 (1):45-64. Lubbe, Gerrie. 2008. "Victims of Xenophobia: African Immigrants in South Africa." MacLean, Lauren M. 2011. "State retrenchment and the exercise of citizenship in Africa." Comparative Political Studies 44 (9):1238-1266. Marks, M. 2005. Transforming the robocops: Changing police in South Africa: Univ of Natal Pr. Matland, Richard E. 2005. "Enhancing women’s political participation: legislative recruitment and electoral systems." Women in Parliament: Beyond Numbers:93-111. Mayekiso, M. 1996. Township politics: Civic struggles for a new South Africa: Monthly Review Pr. McLeod, Jack M, Dietram A Scheufele, and Patricia Moy. 1999. "Community, communication, and participation: The role of mass media and interpersonal discussion in local political participation." Political Communication 16 (3):315-336. Miller, Arthur H, Patricia Gurin, Gerald Gurin, and Oksana Malanchuk. 1981. "Group consciousness and political participation." American Journal of Political Science:494511. ! 282! Mondak, J.J. 1993. "Institutional legitimacy and procedural justice: Reexamining the question of causality." Law and society review:599-608. Morales, Laura, and Marco Giugni, eds. 2011. Social Capital, Political Participation and Migration in Europe: Palgrave Macmillan. Moran, Dominique, and Richard Batley. 2004. "Literature review of non-state provision of basic services." paper commissioned by DFID from Governance Resource Centre, University of Birmingham, UK. Mosselson, Aidan. 2010. "‘There is no difference between citizens and non-citizens anymore’: Violent Xenophobia, Citizenship and the Politics of Belonging in Post-Apartheid South Africa." Journal of southern African studies 36 (3):641-655. "National Crime Prevention Strategy." 1996. ed. D. o. S. a. Security. Neocosmos, M. 1998. "From People’s Politics to State Politics: Aspects of National Liberation in South Africa’." The Politics of Opposition in South Africa:195-241. Newman, Gareth. 2011. "Crime and Policing in South Africa." In Institute for Security Studies Seminar. Nie, Norman H, G Bingham Powell, and Kenneth Prewitt. 1969. "Social structure and political participation: developmental relationships, Part I." The American Political Science Review 63 (2):361-378. Norris, Pippa. 1999. Critical citizens : global support for democratic government. Oxford [England] ; New York: Oxford University Press. Norris, Pippa. 2004. Electoral engineering : voting rules and political behavior. Cambridge, UK ; New York: Cambridge University Press. Norris, Pippa, and Robert Mattes. 2003. "Does ethnicity determine support for the governing party? The structural and attitudinal basis of partisan identification in 12 African nations." Olson, Mancur. 1971. The logic of collective action; public goods and the theory of groups. Rev. ed. New York,: Schocken Books. Ostrom, Elinor. 1990. Governing the commons : the evolution of institutions for collective action. Cambridge ; New York: Cambridge University Press. Paige, Jeffery M. 1971. "Political orientation and riot participation." American Sociological Review:810-820. ! 283! Pauw, J. 1991. In the heart of the whore: The story of apartheid's death squads: Southern Book Publishers Johannesburg. Pavanello, Sara, and James Darcy. 2008. "Improving the provision of basic services for the poor in fragile environments: International Literature Review-Synthesis Paper." Peltier, Jean-Philippe. 2007. State Legitimacy in Sub-Saharan Africa, Political Science, Michigan State University, East Lansing. Pérez, O.J. 2003. "Democratic legitimacy and public insecurity: Crime and democracy in El Salvador and Guatemala." Political Science Quarterly 118 (4):627-644. Platt, Matthew B. 2008. "Participation for What? A Policy-Motivated Approach to Political Activism." Political Behavior 30 (3):391-413. Plutzer, Eric. 2002. "Becoming a habitual voter: Inertia, resources, and growth in young adulthood." American Political Science Review 96 (1):41-56. Portney, Kent E, and Jeffrey M Berry. 1997. "Mobilizing Minority Communities Social Capital and Participation in Urban Neighborhoods." American Behavioral Scientist 40 (5):632644. Powell Jr, GB, and GD Whitten. 1993. "A cross-national analysis of economic voting: Taking account of the political context." American Journal of Political Science:391-414. Print, Murray. 2007. "Citizenship education and youth participation in democracy." British journal of educational studies 55 (3):325-345. Putnam, Robert D. 2001. Bowling alone: The collapse and revival of American community: Simon & Schuster. Putnam, Robert D., Robert Leonardi, and Raffaella Nanetti. 1994. Making democracy work : civic traditions in modern Italy. Princeton, N.J.: Princeton University Press. Reconciliation, Center for the Study of Violence and. 2009. "Why Does South Africa Have Such High Rates of Violent Crime?". Reilly, Ben. 2001. Democracy in divided societies : electoral engineering for conflict management. Cambridge, UK ; New York: Cambridge University Press. Resnick, Danielle, and Daniela Casale. 2011. "The political participation of Africa's youth: Turnout, partisanship, and protest." Working paper//World Institute for Development Economics Research. Rose, Pauline M. 2007. "Supporting non-state providers in basic education service delivery." ! 284! Rose, R., and D.C. Shin. 1999. Democratization backwards: The problem of third wave democracies: Cambridge Univ Press. Rosenstone, Steven, and John M Hansen. 1993. "Mobilization, participation and democracy in America." Rotberg, RI. 2003. When states fail: causes and consequences: Princeton Univ Pr. Sacks, Audrey. 2012. "Can donors and non-state actors undermine citizens' legitimating beliefs?" World Bank Policy Research Working Paper (6158). Salamon, Lester M, and Stephen Van Evera. 1973. "Fear, apathy, and discrimination: A test of three explanations of political participation." The American Political Science Review 67 (4):1288-1306. Savas, E.S. 2000. Privatization and public-private partnerships: Chatham House New York. Saward, M. 1992. Co-optive politics and state legitimacy: Dartmouth Pub Co. Schaar, J.H. 1981. Legitimacy in the modern state: Transaction Pub. Schärf, W., and D. Nina. 2001. The other law: non-state ordering in South Africa: Juta. Schatzberg, M.G. 2001. Political Legitimacy in Middle Africa: father, family, food: Indiana Univ Pr. Schlozman, Kay Lehman, and Henry E Brady. 1995. Voice and equality: Civic voluntarism in American politics: Harvard University Press. Schönteich, M., and A. Louw. 2011. "Crime in South Africa: A country and cities profile." Seekings, J. 2000. "After apartheid: Civic organizations in the “new” South Africa." From comrades to citizens: The South African civic movement and the transition to democracy:205-224. Sharp, Elaine B. 1982. "Citizen-initiated contacting of government officials and socioeconomic status: Determining the relationship and accounting for it." The American Political Science Review:109-115. Shaw, M. 1995. Partners in crime?: crime, political transition, and changing forms of policing control: Centre for Policy Studies. Shaw, M. 2002. Crime and policing in post-apartheid South Africa: Transforming under fire: Indiana Univ Pr. ! 285! Shearing, C.D. 1992. "The relation between public and private policing." Crime and Justice:399434. Sil, R, and C Chen. 2004. "State> Legitimacy and the (In) significance of Democracy in PostCommunist Russia." Europe-Asia Studies 56 (3):347-368. Singh, A.M. 2008. Policing and crime control in post-apartheid South Africa: Ashgate Pub Co. Tyler, T.R. 2003. "Procedural justice, legitimacy, and the effective rule of law." Crime & Just. 30:283. Tyler, T.R., and J. Fagan. 2010. "Legitimacy and Cooperation." Race, Ethnicity, and Policing: New and Essential Readings 6:84. Uhlaner, Carole J, Bruce E Cain, and D Roderick Kiewiet. 1989. "Political participation of ethnic minorities in the 1980s." Political Behavior 11 (3):195-231. Uniform Crime Reporting Program, FBI. 2011. "Crime in the United States." Verba, Sidney, and Norman H Nie. 1972. "Participation in America: Social equality and political democracy." New York: Harper& Row. Wagle, Udaya R. 2006. "Political participation and civic engagement in Kathmandu: an empirical analysis with structural equations." International Political Science Review 27 (3):301-322. Weber, M. 1984. "Legitimacy, politics and the state." Legitimacy and the State:32-62. Weingast, BR. 1997. "The political foundations of democracy and the rule of law." American Political Science Review:245-263. "White Paper on Safety and Security." 1998. ed. D. o. S. a. Security. Wood, J., B. Dupont, and MyiLibrary. 2006. Democracy, society, and the governance of security: Cambridge Univ Press. Young, Crawford. 2012. The Post-Colonial State in Africa: Fifty Years of Independence, 19602010. ! 286!