GLOBAL BURDEN of ARMED VIOLENCE 2011 ISBN 978-1-107-60679-1 Takes an Integrated an Takes 2011

1 1 1 1 VIOLENCE2 0 1 1 www.genevadeclaration.org GENEVA DECLARATION ARMED o f Lethal Encounters BURDEN GLOBAL

GLOBAL BURDEN of ARMED VIOLENCE 2011 ISBN 978-1-107-60679-1 takes an integrated an takes 2011

GENEVA DECLARATION Global Burden of Armed Violence Armed of Burden Global he he approach to the complex and volatile dynamics of armed country- comprehensive on Drawing world. the around violence

Photos level data, including both conflict-related and criminal violence, it estimatesleast thatat violently 526,000 peopledie every more year, settings.highlights that non-conflict three-quarters It in than them of the 58 countries with high rates of lethal violence account for two- thirds of all violent deaths, and shows that one-quarter of all violent Americas. the in are which of seven countries,14 just in occur deaths New research on femicide also reveals that about 66,000 women and girls are violently killed aroundthe world each year. Thisalso assessesvolume linkagesthe between violent deathrates rates homicide that demonstrating development, socio-economic and are higher wherever income disparity, extreme poverty, and hunger are high. It challenges the use of simple analytical classifications and policy responses, andoffers researchers and policy-makers new tools for studying and tackling different forms of violence. www.genevadeclaration.org

Top left: Rescuers evacuate a wounded person from Utoeya, Norway, July 2011. © Morten Edvarsen/AFP Photo Centre left: Morgue workers transport a coffin to be buried along with other unidentified bodies found in mass graves, Durango, Mexico, June 2011. © Jorge Valenzuela/Reuters Bottom right: An armed fighter walks past a burnt-out armed vehicle in the Abobo district of Abidjan, Côte d’Ivoire, March 2010. © Issouf Sanogo/AFP Photo Centre right: A protester runs from a burning police booth in front of the Interior Ministry in Cairo, Egypt, June 2011. © Mohamed Abd El-Ghany/Reuters T 1 1 1 1 VIOLENCE2 0 1 1 www.genevadeclaration.org GENEVA DECLARATION ARMED o f Lethal Encounters BURDEN GLOBAL

GLOBAL BURDEN of ARMED VIOLENCE 2011 ISBN 978-1-107-60679-1 takes an integrated an takes 2011

GENEVA DECLARATION Global Burden of Armed Violence Armed of Burden Global he he approach to the complex and volatile dynamics of armed country- comprehensive on Drawing world. the around violence

GLOBAL of ARMED BURDEN VIOLENCE

Lethal Encounters

20111111

www.genevadeclaration.org ii GLOBAL BURDEN of ARMED VIOLENCE 2011 permission of Cambridge University Press. University Cambridge of permission written the without place take may part any of reproduction no agreements, licensing collective relevant of provisions andtothe exception tory tostatu- Subject isincopyright. publication This 2011 Geneva Secretariat, Declaration © Geneva www.cambridge.org/9781107606791 Information on this title: www.cambridge.org York New Press, University Cambridge by America of States United in the Published UK 8RU, CB2 Cambridge Building Edinburgh The Press University Cambridge City Mexico Tokyo, Dubai, Delhi, Paulo, São Singapore, Town, Cape Madrid, Melbourne, York, New Cambridge, PRESS UNIVERSITY CAMBRIDGE Copyright Cambridge: Cambridge University Press. University Cambridge Cambridge: 2011: Violence Encounters. Lethal Armed of Burden 2011. Secretariat. Declaration Geneva citation: Suggested or appropriate. accurate remain, orwill is, suchwebsites on tent con- any that guarantee not anddoes publication, tointhis referred websites internet or third-party external for URLs of oraccuracy persistence the for responsibility no has Press University Cambridge Paperback 978-1-107-60679-1 ISBN Hardback ISBN 978-1-107-02307-9 ofCongress Cataloguing inPublication data Library Library British the from isavailable publication record this A catalogue for Cambridge Press, University the at Kingdom United inthe Printed 2011 published First Global

iii Foreword FOREWORD

he Global Burden of Armed Violence 2011 slower development. Armed violence undermines introduces a unified framework for under- governmental institutions, destroys human T standing patterns and characteristics of and physical capital, reduces productivity, and lethal violence across all settings, from armed redirects resources away from social and eco- conflicts to economically motivated crime and nomic investment in fighting crime and violence. interpersonal violence. Countries that exhibit high levels of income The report flies in the face of conventional wisdom inequality and unemployment as well as poor and mainstream media coverage by highlighting rule of law are more likely to experience high that conflict deaths account for a relatively small levels of armed violence. part of the global burden of lethal violence, and The picture may be bleak, but a great deal can that most places featuring high death rates are be done to fight the scourge of armed violence. not to be found in conflict zones. Only about 10 Together with more than 100 other signatories of 1 per cent of all violent deaths occur as a result of the Geneva Declaration, the Swiss Confederation 2 armed conflicts or terrorism. Roughly 90 per cent recognizes that effective prevention and reduc- 3 of victims die as a result of homicides, or from tion of armed violence is a key to sustainable deaths occurring during legal interventions in development. Around the globe, in the North as 4 non-conflict countries. in the South, states are stepping up their efforts 5 The human toll is staggering. Between 2004 and to tackle armed violence in urban contexts, along 2009, more than half a million people (526,000) borders, and among high-risk groups. National died annually as a direct result of violence. In and local governments successfully join forces addition, hundreds of thousands of fatalities with civil society to help reduce violent death probably occur without being recorded, or as an rates. A number of good practices derive from indirect result of armed conflict. Further, for every these experiences. person killed, many more are injured, or suffer The international community—including agencies permanent disability and long-lasting pain. such as the Organisation for Economic Co-operation The Global Burden of Armed Violence 2011 also and Development, the United Nations, and the examines the negative links between armed vio- World Bank—is increasingly supporting such lence and development. It shows how freedom programming efforts and providing guidelines from violence is a fundamental prerequisite of for good practices in preventing and reducing economic progress and human well-being and armed violence. In this context, the Global Burden how armed violence is strongly associated with of Armed Violence 2011 provides policy-makers iv GLOBAL BURDEN of ARMED VIOLENCE 2011 Department of Foreign Affairs Foreign of Department Federal Swiss the of Secretary State Maurer Peter violence. armed of burden global the toreduce together towork tocontinue partners society andcivil authorities, local governments, community, international the Weurge levels. andregional national, local, the at programmes and policies indesigning toevidence responding for tool atimely with stakeholders and other v Contents CONTENTS

List of Illustrations ...... vii

About the Geneva Declaration ...... xii

Acknowledgements ...... xiii

Executive Summary ...... 1

Disaggregating lethal violence ...... 4

Reducing armed violence, enabling development ...... 9

Chapter One A Unified Approach to Armed Violence ...... 11

Why a uniﬁed perspective? ...... 13

Obstacles to a uniﬁed perspective ...... 18 1

Neither war nor peace: armed violence in contemporary conflicts ...... 19 2 Neither peace nor war: violence in post- and non-conflict settings ...... 26 3 Conclusion: a unified approach to armed violence ...... 34 Abbreviations ...... 36 4 Endnotes ...... 36 5 Bibliography ...... 39

Chapter Two Trends and Patterns of Lethal Violence ...... 43

Explaining the framework ...... 44

Explaining the data sources ...... 48

Lethal violence 2004–09: a snapshot ...... 51

Trends in lethal violence, 2004–09 ...... 65

Violent deaths: the missing pieces? ...... 69

Disaggregating unintentional homicide ...... 73

Killings during legal interventions and extrajudicial executions ...... 76

Conclusion ...... 80

Abbreviations ...... 81

Endnotes ...... 81

Bibliography ...... 83 vi Chapter Three Characteristics of Armed Violence ...... 87

Contextualizing homicide ...... 89

Contextualizing homicide by region ...... 91

Homicide context and ﬁrearm use ...... 99

Homicide context and the law enforcement response ...... 101

Homicide and criminal justice system performance ...... 102

Conclusion ...... 109

Endnotes ...... 109

Bibliography ...... 110

ARMED VIOLENCE 2011 VIOLENCE ARMED Chapter Four When the Victim is a Woman ...... 113 of Disaggregating homicide by sex ...... 115

Characteristics of femicide ...... 124

Hidden forms of lethal violence against women ...... 133

Conclusion ...... 138

Abbreviations ...... 138

Endnotes ...... 138

GLOBAL BURDEN BURDEN GLOBAL Bibliography ...... 140

Chapter Five More Armed Violence, Less Development ...... 145

Conceptualizing the relationship ...... 147

Assessing the relationship ...... 151

Reviewing the relationship ...... 158

Monitoring the relationship ...... 163

Conclusion ...... 168

Abbreviations ...... 169

Endnotes ...... 169

Bibliography ...... 172

Note to reader: Cross-references to other chapters are indicated in capital letters in parentheses. For example, ‘Information was checked for consistency and categorized according to the classiﬁcation of countries presented in Chapter Two (TRENDS AND PATTERNS).’ vii List of Illustrations

Boxes

Box 1.1 Somali pirates: bandits and soldiers of convenience 16

Box 1.2 How different forms of violence are linked in Iraq 24 ILLUSTRATIONS OF LIST

Box 1.3 Deﬁning illicit trafﬁcking 29

Box 1.4 Mexico’s drug war 30

Box 2.1 Challenges in deﬁning ‘terrorism’ and recording its victims 46

Box 2.2 Why incident reporting provides conservative numbers 50 ox 2.3 B Analysis of violent deaths of Iraqi civilians 56 1 ox 2.4 Violent deaths in Central America and the Caribbean 62 B 2 ox 2.5 B Violent deaths undercounted in Yemen 63 3 ox 2.6 Estimating crude mortality rates in the Democratic Republic of the Congo 71 B 4

Box 3.1 Mob justice in Africa 92 5

Box 3.2 The homicide context and responses in the Caribbean 96

Box 4.1 Unpacking femicide: what’s in a label? 116

Box 4.2 Femicide in Ciudad Juarez: a human rights crisis 121

Box 4.3 Insecurity and impunity in Afghanistan 125

Box 4.4 Intimate partner violence 129

Box 4.5 Dowry deaths in India 134

Box 4.6 HIV and sexual violence in Kenya 135

Box 5.1 Building states and societies that are resilient to conﬂict, crime, and violence: the 2011 World Development Report 148 viii Box 5.2 The relationships between lethal violence and development 158

Box 5.3 Lethal violence and MDG attainment in São Paolo 160

Box 5.4 South Africa: violence and development 162

Box 5.5 Global trends in monitoring armed violence 164

Figures Figure 1.1 A framework of armed violence categories 14

Figure 1.2 Active pro-government militias and the number of civil wars, 1981–2008 22

ARMED VIOLENCE 2011 VIOLENCE ARMED Figure Aggregated homicide rates in 13 selected Western European countries, of 1.3 1970–2009 (base year=100) 26

Figure 1.4 Global homicide trends, 2004–09 28

Figure 1.5 Distribution of victims of organized crime-related killings by type of violence, Mexico, 2009 31

Figure 1.6 Overall homicide and femicide levels in Guatemala, 2003–09 33 GLOBAL BURDEN BURDEN GLOBAL Figure 1.7 The armed violence ‘lens’ 35

Figure 2.1 How violence is reported and recorded 44

Figure 2.2 Incident reporting of violent events leading to death 48

Figure 2.3 Countries ranked by violent death rate per 100,000 population, 2004–09 53

Figure 2.4 Civilian violent deaths of Iraqi women and children from Coalition and anti-Coalition forces, mid-March 2003–mid-March 2008 57

Figure 2.5 Woman and child dirty war index, mid-March 2003–mid-March 2008 57

Figure 2.6 Distribution of violent death rates among 186 countries, per 100,000, 2004–09 58

Figure 2.7 Average violent death rates by region, per 100,000 population, 2004–09 60

Figure 2.8 Percentage of countries per violent death rate per 100,000, 2004–09 61

Figure 2.9 Violent death rates per 100,000 in Mexico, Chihuahua, and Ciudad Juarez, 2004–09 64

Figure 2.10 Direct conﬂict death rates per 100,000 in Pakistan’s provinces, 2006–09 65

Figure 2.11 Violent death rates per 100,000 population, 2004 and 2009 67

Figure 2.12 Countries with signiﬁcant increases in violent death rates per 100,000, 2004–09 68 Figure 2.13 Countries with signiﬁcant decreases in violent death rates per 100,000 ix population, 2004–09 69

Figure 2.14 Disaggregating the global burden of lethal violence 70

Figure 3.1 Homicide context by geographic region 91

Figure 3.2 Vigilante/mob killings as a percentage of homicides in three African countries 92

Figure 3.3 Homicide rate in 12 Caribbean states, 2000–09 96

Figure 3.4 Homicide context in Jamaica and Trinidad and Tobago, 2000 and 2008 96

Figure 3.5 Homicide rate and case clearance in Jamaica, 1970–2005 98 LIST OF ILLUSTRATIONS OF LIST Figure 3.6 Homicide context and income inequality, 2010 or latest available year 99

Figure 3.7 Homicide rate and proportion of homicides committed with ﬁrearms, 2009 or latest available year 99

Figure 3.8 Proportion of gang- or organized crime-related homicides and ﬁrearm use 100

Figure 3.9 Proportion of homicides committed with ﬁrearms and cases solved 101

Figure 3.10 Possible criminal justice system attrition measures 104 1 Figure 3.11 Attrition of homicide cases in cross-national perspective 105 2 Figure 3.12 Rule of law and overall intentional homicide rates, 2009 or latest available year 107 3 Figure 3.13 Reporting of crime involving armed violence by victims, 1989–2005 108 4 Figure 4.1 Estimated average proportion of female vs. male homicide victims per year, 5 2004–09 117

Figure 4.2 Countries and territories by rate of femicide per 100,000 female population, 2004–09 118

Figure 4.3 Average homicide and femicide rates per 100,000 total population, by region, 2004–09 119

Figure 4.4 Average femicide rates per 100,000 female population in 25 countries and territories with high and very high rates, 2004–09 120

Figure 4.5 Trends in femicide per 100,000 female population in Ciudad Juarez and Mexico, 1993–2009 121

Figure 4.6 Femicide trends in ﬁve selected countries, 2004–09 (basis 2004=100) 123

Figure 4.7 Number of femicide victims for every 100 male homicide victims in 83 countries, 2004–09 124 x Figure 4.8 Victims of violent crime committed by an intimate partner in the United States, by sex and rate per 100,000 population, 2008 129

Figure 4.9 Total femicide rates per 100,000 female population and estimated percentage of IPV-related femicides per country, 2004–09 129

Figure 4.10 Deaths of men and women as a consequence of domestic violence in the European Union, 2006 130

Figure 4.11 Average percentage of femicides committed with ﬁrearms in 24 countries, by femicide rate, 2004–09 131

Figure 4.12 Percentage of male and female homicide victims killed by ﬁrearm, Mexico

ARMED VIOLENCE 2011 VIOLENCE ARMED and Ciudad Juarez, 1993–2009 132 of Figure 4.13 Percentage of femicides committed with ﬁrearms in the United States, by age group, 1999–2007 133

Figure 5.1 The widening gap between countries affected by major, minor, and negligible violence 148

Figure 5.2 Disaggregated homicidal violence for 182 countries by HDI, 1986–2009 152 GLOBAL BURDEN BURDEN GLOBAL

Figure 5.3 Disaggregated homicidal violence by income, 1986–2009 154

Figure 5.4 Disaggregating income inequality and lethal violence, 1986–2009 155

Figure 5.5 Disaggregating adolescent birth rates and lethal violence, 1986–2009 157

Figure 5.6 Homicide rate per 100,000 population in Brasília, Rio de Janeiro, São Paolo, and Brazil, 1992–2007 160

Figure 5.7 The core attributes of an armed violence monitoring system 164

Figure 5.8 Violence indicators tracked by AVMS initiatives 164

Figure 5.9 Data sources used by AVMS, as percentage 164

Figure 5.10 An inter-agency armed violence monitoring group 166

Maps Map 1.1 Organized crime-related homicide rates, Mexico, 2007 31

Map 1.2 Organized crime-related homicide rates, Mexico, 2008 31

Map 1.3 Organized crime-related homicide rates, Mexico, 2009 32

Map 2.1 Average annual violent death rates per 100,000, 2004–2009 52 Map 2.2 Average annual violent death rates per 100,000 in Central America xi and the Caribbean, 2004–09 62

Map 4.1 Rates of female homicide victims per 100,000 female population, 2004–09 118

Map 5.1 Geography of armed violence and human development, 2009 153

Map 5.2 Homicide rates per 100,000 population by district, São Paolo, 2009 160

Tables

Table 2.1 Reported number of victims of terrorism or one-sided violence 47 LIST OF ILLUSTRATIONS OF LIST Table 2.2 Iraqi civilian deaths by type of perpetrator and method, mid-March 2003–mid-March 2008 56

Table 2.3 Intentional versus unintentional homicide: a sample 74

Table 2.4 Killings during legal interventions: A sample 78

Table 5.1 The relationship between development indicators and lethal violence 158 Table 5.2 MDG indicators and violence in two district clusters in São Paolo, 2006 161 1 Table Violence and development indicators in South Africa, various years 162 5.3 2

5 xii About the Geneva Declaration

he Geneva Declaration on Armed Violence Core group members include Brazil, Colombia, and Development, endorsed by more than Finland, Guatemala, Indonesia, Kenya, Morocco, 100 countries, commits signatories to the Netherlands, Norway, the Philippines, Spain,

ARMED VIOLENCE 2011 VIOLENCE ARMED T supporting initiatives intended to measure the of Switzerland, Thailand, and the United Kingdom. human, social, and economic costs of armed Afﬁliated organizations include the United Nations violence, to assess risks and vulnerabilities, to Development Programme (UNDP), the Organisation evaluate the effectiveness of armed violence for Economic Co-operation and Development (OECD), reduction programmes, and to disseminate and the Quaker United Nations Ofﬁce (QUNO). knowledge of best practices. The Declaration calls upon states to achieve measurable reduc- For more information about the Geneva Declara-

GLOBAL BURDEN BURDEN GLOBAL tions in the global burden of armed violence and tion, related activities, and publications, please tangible improvements in human security by 2015. visit www.genevadeclaration.org. xiii Acknowledgements

he second edition of the Global Burden of Gilgen, Matthias Nowak, and Katherine Aguirre Armed Violence report (GBAV 2011 hereafter) Tobón of the Small Arms Survey led the process

T combines the efforts of the staff of the Small of building the extensive GBAV 2011 database by ACKNOWLEDGEMENTS Arms Survey and several dozen specialists around collecting information on conflict deaths, terrorism the world who have contributed input, reviews, data, victims, intentional and unintentional homicide, case studies, and critical assessments of the and deaths occurring during legal interventions. methodology. The Small Arms Survey was respon- The United Nations Office on Drugs and Crime sible for conceiving, researching, commissioning, (UNODC) provided comprehensive national hom- and editing the report. The editors of the report icide statistics; the World Health Organization are Keith Krause, Robert Muggah, and Elisabeth (WHO) shared preliminary 2008 estimates on Gilgen; all three are with the Small Arms Survey. interpersonal and collective violence; the Global Burden of Disease Injury Expert Group provided 1 Alessandra Allen, the Small Arms Survey’s publi- and analysed data stemming from vital registration cations manager, coordinated the production of 2 sources; and the Bogotá-based Conflict Analysis the GBAV 2011; Fabio Dondero, Sarah Hoban, Resource Center (CERAC) was responsible for the 3 Mihaela Racovita, and Pilar Reina fact-checked dataset on conflict deaths. the report; Jillian Luff produced the maps; Richard 4 Jones provided the design and the layout; Tania Numerous country representatives and national 5 Inowlocki copy-edited and Donald Strachan institutions, such as national police or statistics proofread the report; and Margaret Binns com- offices, provided essential clarifications and piled the online index. Olivia Denonville helped contributions regarding data on violence. They with photo research. Katherine Aguirre Tobón include representatives from Anguilla, Antigua coordinated the production of the methodologi- and Barbuda, Australia, Chile, France, Liberia, cal annexe. John Haslam and Carrie Parkinson of Mexico, Nauru, Peru, the Seychelles, Spain, Cambridge University Press provided support Suriname, and the United Kingdom. throughout the production of the report. Carole In early 2011, an expert meeting was held in Touraine, Benjamin Pougnier, Cédric Blattner, and Geneva to review the methodological approaches David Olivier provided administrative support. to measure overall violent deaths around the The GBAV 2011 assembles research based on a globe. The editors are extremely grateful for the substantial collection of data on lethal forms of valuable feedback gathered during this meeting violence around the globe and generated by a and would like to thank all the participants: wide range of institutional partners. Elisabeth Kidist Bartolomeus (WHO), Kavi Bhalla (Harvard xiv The principal chapter authors were assisted by Global Burden of Armed Violence 2011 many in-house and external contributors who are Editors acknowledged in the relevant chapters summaries, Keith Krause, Robert Muggah, and Elisabeth Gilgen below. While the report represents a collective

Coordinator effort, the editors wish to recognize the following Elisabeth Gilgen experts and institutions. Our apologies to anyone

Publications Manager we may have inadvertently omitted. Alessandra Allen

Designer Richard Jones, Exile: Design & Editorial Services A Unified Approach to Cartographer ARMED VIOLENCE 2011 VIOLENCE ARMED Armed Violence

of Jillian Luff, MAPgraﬁx

Copy-editor The lines between categories of violence are Tania Inowlocki blurry, allowing various types of violence to reinforce each other in a vicious circle. These traits Proofreader Donald Strachan thwart simple classiﬁcations and policy responses. In this chapter, Keith Krause and Matthias Nowak Principal chapter authors provide an overview of cases and arguments

GLOBAL BURDEN BURDEN GLOBAL Chapter One for a uniﬁed approach to armed violence. By Keith Krause and Matthias Nowak acknowledging the blurring of boundaries, this Chapter Two approach seeks to facilitate the design of policies Elisabeth Gilgen and interventions that effectively address the Chapter Three serious challenges posed by armed violence to Steven Malby the safety and well-being of individuals and com-

Chapter Four munities. Jennifer Hazen, Sabine C. Carey, and Anna Alvazzi del Frate Neil J. Mitchell provided insight into the evolving

Chapter Five nature of armed groups; Germán Lugo of the Robert Muggah and Jorge Restrepo Sistema Nacional de Seguridad Pública supplied information regarding deaths due to organized crime in Mexico; and the International Maritime University), Alex Butchart (WHO), Koen De Groof Bureau provided assistance in the use and inter- (Bureau for Crisis Prevention and Recovery, UN pretation of yearly piracy reports. Development Programme), Elizabeth Gurian (UNODC), James Harrison (Flinders University, Australia), Lisa Knowlton (Harvard University), Steven Malby (UNODC), Colin Mathers (WHO), Trends and Patterns of Richard Matzopoulos (Medical Research Council), Lethal Violence Andrés Rengifo (Harvard University), Jorge Instead of conforming to boundaries between Restrepo (CERAC), and Sebastian Taylor (Action organized, political, and criminally or economi- on Armed Violence). cally motivated violence, this chapter presents comprehensive data on violent deaths at the support with data collection, communications xv global, regional, and national levels. Authored with national governments, and database by Elisabeth Gilgen, it introduces the GBAV management. 2011 database. Jean-Marc Flückiger (consultant, Switzerland) contributed background research on the complexities of defining terrorism and counting its victims, and Paul Smit (consultant, When the Victim is a Woman Netherlands) explored the difficulties of counting One useful way of measuring the extent of lethal victims of violent deaths categorized as ‘inten- violence perpetrated against women is by disag- tional homicides’. Gavin Hales (consultant, UK) gregating homicide statistics by sex. In this offered valuable input on legal definitions of chapter, Anna Alvazzi del Frate presents a global ACKNOWLEDGEMENTS intentional and unintentional homicides. Andrés dataset on femicide, which she compiled using Rengifo (Harvard University) provided guidance mostly publicly available information. Steven Malby on deaths due to police killings; and Raza Shah (UNODC), Mario Arroyo (Instituto ciudadano de Khan, executive director of the Sustainable Peace estudios sobre la inseguridad, Mexico), Mumbi and Development Organization, offered insight Machera (University of Nairobi,), Sami Nevala into counting conflict casualties in Pakistan. (European Union Fundamental Rights Agency), Michael Spagat (University of London), John Soula McFarlane (Australian Bureau of Statistics), Sloboda (University of London and Iraq Body Maria Giuseppina Muratore (Istituto nazionale di Count), Madelyn Hsiao-Rei Hicks (King’s College), statistica, Italy), Henriette Jansen (consultant on 1 and Hamit Dardagan (Iraq Body Count) provided violence against women, Switzerland), and Natalie an analysis of violent deaths of Iraqi civilians. Jaynes (Open Society Foundation, South Africa) 2 provided valuable input. 3

4 Characteristics of Armed Violence Data on intentional homicide has become increas- More Armed Violence, 5 ingly detailed in many countries in recent years. Less Development Making use of data and analysis provided by the There is a widespread consensus that armed vio- United Nations Office on Drugs and Crime, this lence and underdevelopment are connected. And chapter moves beyond a description of overall yet there is comparatively little empirical evidence homicide rates to present what is known about the demonstrating the precise nature of the relation- characteristics of homicides and state responses ship. Written by Robert Muggah and Jorge Restrepo, to it. At UNODC, the Statistics and Surveys Section this chapter disaggregates the specific dynamics undertook research and analysis. The work was of that association with an aim to better inform led by Steven Malby under the overall supervision the policy decisions of multilateral and bilateral of Angela Me. The chapter benefitted from the aid agencies that are investing in armed violence data collection system managed by Catherine prevention and reduction activities. Mayra Iglesias Pysden and assisted by Ali Saadeddin. Interna- and Manual Moscoso of CERAC and Katherine tional consultants Wilfried De Wever, Lievine Aguirre Tobón of the Small Arms Survey generated Prince, and Elizabeth Gurian provided valuable additional statistical input. Further contributions xvi were provided by Gary Milate (World Bank), Renato Sérgio de Lima (Brazilian Forum on Public Safety), Achim Wennmann (Geneva Peacebuilding Platform), and researchers at the Centre for the Study of Vio- lence and Reconciliation in South Africa.

Finally, special thanks must be extended to Luigi de Martino, coordinator of the Geneva Declara- tion on Armed Violence and Development, and to Ambassador Claude Wild of the Swiss Federal Department of Foreign Affairs, chair of the Geneva Declaration Core Group. Together with its part- ARMED VIOLENCE 2011 VIOLENCE ARMED

of ners from the UN Development Programme, the Quaker United Nations Ofﬁce, the Organisation for Economic Co-operation and Development, and others, the Core Group provided important strategic guidance on key issues concerning armed violence. Together with Siro Beltrametti and Julien Thöni, the contribution of the Government of

GLOBAL BURDEN BURDEN GLOBAL Switzerland must also be specially recognized. The report is an independent contribution of the Small Arms Survey to the Secretariat of the Geneva Declaration on Armed Violence and Development; as such, it does not necessarily reflect the views of the Government of Switzerland or any other signatory state of the Geneva Declaration. While the report is a collective effort, the editors are responsible for any errors and omissions of fact or judgement. distinctions give the misleading impression that that misleading impression the give distinctions Yet these and policies. programmes reduction violence ortoplan violence of levels overall toassess them use world the around institutes andresearch organizations, non-governmental agencies, multilateral Governments, acts. violent behind motivations andthe organization of level the tocapture intended are distinctions These criminal between and interpersonal organized between drawn are tions distinc- common most two The perpetrator. the of intentions orunderlying context to aparticular according categories distinct into violence armed compartmentalize often analyses Conventional humanbeings. fellow of harming erate delib- the involves it isuniquebecause violence But people. disasters—kill tonatural epidemics tragedies—from Many home. insidethe violence of forms various endure must more Still ties. andcommuni- homes their toleave forced are others year. Countless every lives their or lose injuries suffer people of thousands of hundreds crime, organized ortransnational trafficking drug with associated andkillings violence gang of or uprisings, orrebel conflicts of context in the Whether forms. multiple take can violence armed Contemporary violence. armed of and outcomes T approach to understanding the origins origins the tounderstanding approach Armed Violence edition he 2011 (economically motivated) violence. violence. motivated) (economically conflict Executive Summary (politically motivated) and motivated) (politically (individual) violence, and violence, (individual) adopts an integrated an integrated adopts Global Burden of of Burden Global the of (collective) (collective) Key findings of the report are: report the of findings Key with each other. interact may violence of forms different and how contexts, indifferent itself manifests violence how of understanding our anddeepening refining further for foundation asolid provides it violence, armed of forms orinterpersonal criminal, conflict, to exclusively analysis its confining than Rather violence. of forms different across death violent of overview aglobal andprovides violence armed to approaches suchcompartmentalized lenges Violence Armed of Burden Global 2011 The categories. andseparate neat into fit violence of andincidents forms different The proportion of homicides related to related homicides of proportion The death violent exhibit countries Fifty-eight eachyear killed are people 526,000 than More Europe. Homicide rates related to related rates Homicide Europe. or inAsia than America andSouth Central or organized crime andJamaica. Iraq by followed in2004–09, violence lethal by affected most country the was Salvador El deaths. allviolent of two-thirds almost for account countries These 100,000. 10.0per above rates year. every occur homicides intentional 396,000 while activities, terrorist orduring settings conflict inso-called occurs world the around deaths violent allreported of ten inevery One violence. lethal of aresult as is significantly higher in higher is significantly robbery or robbery chal- gangs 2 3 4 5 1 EXECUTIVE SUMMARY 1 2 tend to be higher in countries with greater theft Photo A police officer takes notes following an incident in which income inequality. one gang member was killed and two injured after they shot a bus driver dead in San Salvador, El Salvador, September 2010. The proportion of homicides related to intimate © Luis Romero/AP Photo partners or the family represents a significant proportion of homicides in some countries in Europe and Asia. Roughly 66,000 women are violently killed around the world each year, accounting for approximately 17 per cent of total intentional homicides. Lethal violence is strongly associated with ARMED VIOLENCE 2011 VIOLENCE ARMED

of negative development outcomes in various ways and is accompanied by low levels of overall achievement of the Millennium Development Goals.

Chapter One (A Uniﬁed Approach to Armed Violence) shows high levels of gang violence in

GLOBAL BURDEN BURDEN GLOBAL Guatemala or Honduras, vigilante justice in post- war and fragile states such as Liberia or Timor- Leste, post-election violence in Côte d’Ivoire or Kenya, and high levels of urban crime in cities such as Kingston or Rio de Janeiro amply demonstrate how the lines between armed conflict and criminal violence are increasingly blurred. In Iraq since 2003, for example, the targeting of non-combatants by insurgents, militias, and sectarian groups may seem chaotic or random at first glance, yet a closer look at underlying patterns of violence suggests that seemingly arbitrary or criminal violence may also serve political purposes in line with the goals of armed groups. In many places, non-conflict violence is linked to highly organized criminal activity, or to different forms of ‘political violence’, either targeting political opponents or government officials (such as mayors, teachers, police officers, or journalists), or seeking to influence and modify government policies through corruption and use of force. In these contexts, the label ‘homicide’—which implies ostensibly 2 3 3 4 5 1 EXECUTIVE SUMMARY 4 apolitical interpersonal and criminal violence—is These recurring characteristics—the multiple, slightly misleading. simultaneous, and shifting motivations of violent actors, and the links between different forms of The violent activities of organized criminal groups violence—demand more than simple analytical frequently have broader political consequences, classifications and policy responses. They require even if their main motivation remains profit- new ways of understanding the relationships seeking. Criminal activities such as trafficking between what were previously held to be distinct in drugs or other illegal goods have also been forms of armed violence. The Global Burden of used to finance war efforts in places such as Armed Violence 2011 offers a preliminary roadmap Afghanistan, Bosnia and Herzegovina, Colombia, to do precisely this. and Liberia. The operations of organized crime groups, and especially the trafficking of illicit ARMED VIOLENCE 2011 VIOLENCE ARMED

of narcotics, are frequently accompanied by high levels of violence. Such groups have shown an Disaggregating lethal violence extraordinary capacity for blurring the boundaries The intensity and organization of violent killings between criminal and political types of violence, provides a critical indicator of a state’s—and its as evidenced by the drug wars in Mexico and the population’s—relative insecurity. From a statisti- rest of Central America, the Caribbean, and certain cal perspective, violent deaths tend to be more Andean countries. Drug cartels are locked in battle systematically recorded than other crimes and GLOBAL BURDEN BURDEN GLOBAL for control over the flow of narcotics while gov- human rights violations. Based on data on lethal ernments in countries across these regions have violence from established administrative sources mobilized their armies to boost a faltering war on in the criminal justice, health, and conflict stud- drugs. Illicit trafficking of drugs is increasingly ies sectors, Chapter Two (Trends and Patterns of recognized as a threat to international, regional, Lethal Violence) finds that an average of 526,000 and national security, as well as public safety. people died violently per year between 2004 and 2009. The estimate includes civilian conflict Figure Disaggregating the global burden of 2.14 deaths, battle deaths, and victims of terrorism lethal violence (combined as direct conflict deaths), intentional Legend: and unintentional homicide, and legal interven- Direct conflict deaths (55,000; 10.4%) Intentional homicide (396,000; 75.3%) tions in non-conflict settings (see Figure 2.14). Unintentional homicide (54,000; 10.2%) While war casualties are frequently featured in Legal intervention killings (21,000; 4.1%) media headlines, their actual number is far lower than that of victims killed in many ostensibly non- conflict countries. Roughly three-quarters of all violent deaths are the result of intentional homicide, while approximately 10 per cent are direct conflict deaths. This translates into 396,000 intentional homicide victims and 55,000 direct conflict deaths per year. Map 2.1 presents a snapshot of the global distribution of direct conflict and inten- Source: GBAV 2011 database tional homicide death rates per 100,000 population. capacity to monitor incidents. capacity andinstitutional reporting quality on depends system harvesting data any obvious: are this for reasons The situation. given inany victims of number the undercount often sincethey estimates conservative yield typically which databases, and systems reporting incident from is derived data the of Most deaths. allviolent of cent 4 per ayear, or victims 21,000 least at for accounts interventions— legal during category—killings remaining The homicide. unintentional of a result as dieviolently deaths) allviolent of cent 10 per than (more people additional 54,000 estimated An Source: 2.1 Map L Per 100,00population EGEND GBAV 2011 database 2011 GBAV No data <3 3–10 10–20 20–30 >30 : Ave rage annual violent death rates per 100,000, 2004–2009 100,000, per rates death annualviolent rage encing extremely high violent death rates—more rates—more death violent high encing extremely experi- are countries 14 them, Among countries. in these eachyear killed violently are people 285,000 estimated An deaths. allviolent of (63 cent) per two-thirds almost for and account violence armed of rates high andvery high exhibit population— global the of cent per 18 or roughly people billion 1.2 some countries—comprising world’s the of one-quarter findsthat It population. 100,000 10per than more of combined) homicides andintentional deaths conflict (direct rates death violent experiencing are that states 58 the in on Violence Armed of Burden Global 2011 The zooms zooms 2 3 4 5 5 1 EXECUTIVE SUMMARY 6 Figure 2.3 (detail) Countries with average annual violent death rates of more than 30 per 100,000 population, 2004–09

El Salvador Iraq Jamaica Honduras Colombia Venezuela Guatemala South Africa Sri Lanka Lesotho Central African Republic Sudan

ARMED VIOLENCE 2011 VIOLENCE ARMED Belize of Congo, Democratic Republic of the

0 10 20 30 40 50 60 70 80 90 100 Source: GBAV 2011 database

than 30 violent deaths per 100,000 people (see population, for example, the state of Chihuahua Figure 2.3); these comprise 4.6 per cent of the experienced a rate of 108 per 100,000 in the global population and account for an estimated same year. Understanding what is behind such GLOBAL BURDEN BURDEN GLOBAL 124,000 violent deaths. In other words, 25 per extreme sub-national variations in the incidence cent of violent deaths occur in just 14 countries, of armed violence is a prerequisite for designing which are home to less than 5 per cent of the and administering effective violence prevention world’s population. Of these 14 countries, seven and reduction programmes. are in the Americas. The Global Burden of Armed Violence 2011 also As these findings reveal, armed violence is highly unpacks the diverse contexts and settings in concentrated in specific regions and in a compara- which intentional homicides occur. It examines tively small number of countries. The regions most how intentional homicide may arise in the con- affected by lethal violence include Latin America text of violent operations by gangs or organized and the Caribbean, and Central and Southern criminal groups, premeditated or unplanned Africa. At the country level, El Salvador experi- crimes of passion committed against intimate enced the highest overall annual average violent partners or family members, or other crimes, such death rate between 2004 and 2009, followed by as robbery or theft. Chapter Three (Characteristics Iraq and Jamaica. of Armed Violence) looks at the trends and patterns of these different forms of intentional Lethal violence is not only distributed unevenly homicidal violence across different situations and across states or regions, but also within states. geographic settings. While specific municipalities, cities, or neighbourhoods may be highly affected by criminal Such disaggregration of data is important for violence and armed conflict, other areas may be policy and programmatic reasons. For example, comparatively peaceful. Whereas Mexico’s vio- while countries in Asia and Europe show a com- lent death rate in 2009 stood at 18.4 per 100,000 paratively high proportion of intimate or family- general availability of weapons to civilians, as as tocivilians, weapons of availability general tothe is related shift This groups. criminal ized andorgan- gangs youth for choice of weapons the as andbluntobjects knives supplanting increasingly are firearms that demonstrates Three Chapter addition, In rates. death violent overall higher experience also firearms with committed homicides of proportions high with societies that indicate findings andother These population100,000 or above. 20 of per rates homicide high disproportionately show afirearm with out carried are homicides of cent 70per than more inwhich countries the of cent) per (78 four-fifths however, firearm; by out carried homicides of proportion ahigh have rates homicide high with allcountries Not firearm. by committed homicides of proportion andthe rates homicide intentional overall between tionships rela- the untangling aimof the with is available data accurate which for 104countries of review a presents Three Chapter efforts. and reduction prevention violence armed andfocus to refine ways highlight also can inhomicide used are they frequently andhow how of inspection and aclose violence, inlethal role animportant play Firearms inequality. income greater with countries in higher tobe tend ortheft torobbery related rates homicide how notes it time, same the At America. inLatin incountries higher nificantly issig- crime ororganized gangs with associated homicides intentional of proportion the that finds also Three Chapter violence. lethal of forms these at initiatives andprevention reduction lence vio- andlocal aimingresearch of importance the underlines andEurope inAsia countries in many homicides orfamily-related intimate of proportion high the Nonetheless, Americas. the such as regions, other of those than lower significantly overall homicides), their total of cent per 30 (around homicides related

homicide rates are are rates homicide homicides in general, there appears to be some some tobe appears there ingeneral, homicides with as But malevictims. with inhomicides than in femicides common isless firearms of use The women. than homicide from todie likely more times ten than more are men where andVenezuela, Rico, Puerto Colombia, inBrazil, case isthe This malevictims. with cides homi- of rates of afraction just represent rates femicide rates, homicide high experiencing tries incoun- spectrum, the of end other the At 1 to1. approaches ratio victim female–male the rare, are homicides where incountries Indeed, texts. con- violent inmore than ishigher male victims to compared victims homicide female of centage per- the orSwitzerland, Norway, Japan, Austria, in as rare, relatively are homicides where tries incoun- how highlights 83countries from data of Areview variations. considerable shows victims maleandfemale of proportion the of inspection comparative adeeper time, same the At girls. and women of numbers higher also but numbers, inhigh dying are who men young just isnot it andGuatemala, Salvador El suchas countries In rates. femicide higher experience typically also rates homicide high comparatively feature that Countries homicides. intentional 396,000 total the of cent 17 per approximately for accounts femicide deaths, violent of victims of proportion larger the make men up eachyear. While world the around killed violently are andgirls women 66,000 that estimates conservatively chapter the available, is that data sparse the on Based data. of city scar- of because andinterpret tomonitor difficult especially are Trends infemicide awoman. of killing intentional ‘femicide’—the on a spotlight Violence Armed of Burden Global 2011 the In offirearms. smuggling and trafficking illegal of presence the as well Chapter Four (When the Victim IsaWoman), Victim the (When shines shines 2 3 4 5 7 1 EXECUTIVE SUMMARY 8 ARMED VIOLENCE 2011 VIOLENCE ARMED of GLOBAL BURDEN BURDEN GLOBAL be distributed in May 2010. © Trevor Snapp ©Trevor 2010. May in distributed be to aid for wait Sudan, South Tambura, near Army Resistance Photo Women displaced by recent attacks by the Lord’s Lord’s the by attacks recent by displaced Women enabling development enabling violence, armed Reducing firearms. of use of percentage lower toa correspond frequently rates femicide low firearms: with committed femicides of centage per- andthe rates femicide between relationship negatively and significantly linked to changes in tochanges linked andsignificantly negatively are rates homicide words, other In violence. lethal of levels lower toexhibit likely most also are HDI intheir animprovement register that Countries progress. development constrains violence lethal Violence Armed of Burden Global 2011 the for conducted Research indicators. (MDG) Goal Development Millennium andthe (HDI) Index Development Human the by measured as progress anddevelopment violence lethal between relationship the of ananalysis presents opment) Chapter Five andimpunity. violence increasing of ing aspiral enter- risk may andJamaica, Salvador El such as factors, of combination this showing Countries agencies. enforcement law by solved cases of proportion andalow firearms, with committed homicides of proportion ahigh rates, homicide high between isanexus there time, same the At homicide. intentional of levels lower show broadly systems—also justice criminal including effective law— of rule the for respect higher with Countries suffer. development andeconomic social, human, security, Without andunderdevelopment. curity inse- between association isastrong there that and capacities, institutional andfragile violence armed of levels higher between a relationship is there that accept widely now and governments agencies Aid anddevelopment. violence armed between relationship complex the considers Violence Armed of Burden Global 2011 The (More Armed Violence, Less Devel- Less Violence, Armed (More suggests that that suggests also also 9 2 3 4 5 1 EXECUTIVE SUMMARY 10 a country’s HDI rating. Yet whether levels of vio- research institutions, can provide crucial infor- lence cause lower scores is difficult to determine. mation on the scale and distribution of lethal vio- Country data for 2000 to 2009 indicates that lence. This data is indispensable in unpacking the the greater the income disparity, the higher the complex relationships between armed violence homicide rates. The inverse is also true: societies and factors such as unemployment, inequality, reporting less severe income inequality report much the presence of illicit markets, corruption, weak lower levels of homicidal violence. The findings rule of law, and impunity. The ability of the inter- are aligned with and confirm the body of research national community and national as well as local that identifies a robust relationship between governments to design appropriate policies and income inequality and violent criminality. programmes for armed violence prevention and reduction depends critically on an integrated More positively, the 2011 Global Burden of Armed ARMED VIOLENCE 2011 VIOLENCE ARMED and comprehensive understanding of the distri-

of Violence ﬁnds that a reduction in a country’s bution and dynamics of lethal (and non-lethal) incidence of armed violence corresponds with violence worldwide. improved MDG performance. High rates of intentional homicide are accompanied by signiﬁcantly higher levels of extreme poverty and hunger (MDG 1), lower primary school enrolment (MDG 2), Abbreviations higher infant mortality and adolescent birth rates HDI Human Development Index

GLOBAL BURDEN BURDEN GLOBAL (MDGs 4 and 5), and higher youth unemployment. MDG Millennium Development Goal The same relationship is found between direct conflict deaths and MDG progress. Higher rates of direct conflict deaths are correlated with higher rates of poverty (measured as the population living below USD 1 per day); a lower share of women in wage employment in the non-agricultural sector; lower enrolment in primary education and a lower ratio of girls to boys in primary education; and last, but not least, lower HDI. These findings reveal a broad set of linkages between armed violence and development outcomes and represent a solid basis for further research at the local and national levels.

Containing and reducing the incidence of armed violence requires a proper diagnosis of its causes and consequences. Many governments affected by high levels of armed violence—as well as many others that are not affected—have initiated comprehensive armed violence monitoring systems. Such ‘observatories’, especially when administered in partnership with civil society and reliable takes two further steps, both of which have have which of both steps, further two takes edition This (homicide). violence interpersonal and conflict from death violent of levels regional andsub- andglobal violence, armed of burden global overall the estimating for framework basic a out sets in2008, published edition, first The consequences. andits causes, its violence, armed Armed Violence of Burden Global the of edition second This orcompare. tomeasure difficult are that ways in andsocieties communities of fabric economic and political, social, the andcorrode insecurity, and fear orinjury—spread indeath result not do they when violence—even armed of levels High humanbeings. fellow of orharming killing ate deliber- the involves it inthat isdistinct violence yet people’s lives, andaccidents—claim ters, disas- natural anddiseases, including epidemics tragedies— other Many violence. and sexual famine, tosickness, exposed and communities, homes their toleave forced been have others less count- while lives, their orlost injuries suffered have inMexico—millions war drug tothe Lagos and Angeles inLos violence street and from d’Ivoire, orCôte inLibya conflicts contemporary to Africa inSub-Saharan colonialstruggles the from andAfghanistan, inIraq interventions the A of settings—from the 20 the settings—from of avariety In world. the around people of millions directly affected the lives of hundreds of of hundreds of lives the affected directly manifestations—has and contemporary rmed violence Chapter A Unified Approach to Violence Armed report takes a unified view of of view aunified takes report —in both its historical historical its both —in th

-century world wars to wars world -century One armed groups, and persistently high levels of of levels high and persistently groups, armed andnon-state gangs, crime, organized national trans- involving networks regional of growth the wars, of dimensions Economic violence. motivated and economically motivated politically andbetween andcrime, conflict armed between line the blurred has violence armed porary contem- of nature changing the that noted have analysts Numerous contexts. of range in awide andappears forms many takes violence Armed violence. of armed and distribution scale inthe variations national and regional of comparisons fine-grained more allows also It estimate. nuanced amore provide andto 2004–07) than rather (2004–09 period time alonger at tolltolook death global the of 2011 Violence Armed of Burden the isthat analysis of level andthe ments, instru- data, of refinements these of result One with each other. interact may violence of forms different and how contexts, indifferent ismanifest violence how of understanding our anddeepening refining further for basis the provides thus It violence. armed of forms andinterpersonal criminal, conflict, between todistinguishing opposed as allsources, from death violent of overview aggregate first the topresent sources multiple from data available andanalyses synthesizes it Second, victimization. violent for estimates level national- comparable in’topresent ‘zooms it First, implications. andprogramming policy important revisits estimates estimates revisits Global Global 11 2 3 4 5 1 A UNIFIED APPROACH 12 interpersonal violence—whether in conflict and sources of violence and insecurity before post-conflict situations or in settings that have focusing on specific drivers or manifestations not experienced war—make clear that armed of violence. violence is a complex phenomenon to untangle. Drawing sharp boundaries around the organiza- The following chapters explore different elements tion, nature, and purpose of different violent of these general arguments. Chapter Two unpacks acts is unhelpful in developing responses to the national- and regional-level data on rates and diverse manifestations of violence around the levels of armed violence around the world, focusing world. Following the usage in the first Global on the 58 most violence-affected states, all with Burden of Armed Violence report, this volume an overall violent death rate exceeding 10 per defines armed violence generally as ‘the inten- 100,000. It shows that violence in non-conflict ARMED VIOLENCE 2011 VIOLENCE ARMED tional use of illegitimate force (actual or threat- settings (intentional homicide) is responsible for of ened) with arms or explosives, against a person, the vast majority of killings (slightly more than group, community, or state, that undermines 75 per cent of all deaths), while conflict-related people-centred security and/or sustainable violence accounts for only 10 per cent of all vio- development’ (Geneva Declaration Secretariat, lent deaths. The remaining deaths are attributed 2008, p. 2).1 to killings during legal interventions (4 per cent) and unintentional homicide (just over 10 per cent).2 This chapter presents an overview of the main

GLOBAL BURDEN BURDEN GLOBAL Chapter Three focuses on violent victimization and themes of the Global Burden of Armed Violence the instruments of violence involved in so-called 2011, focusing in particular on the reasons for— non-conflict settings. It finds that gang- and organ- and challenges to—adopting a unified approach ized crime-related homicides are highly concen- to contemporary armed violence. Its key conclu- trated in Central and South America, and that sions are that: deaths related to robbery are higher in countries with pronounced inequalities. It also points towards The intensity and location of conflict and non- the critical nexus that may exist between high conflict armed violence has changed signifi- homicide rates, a high proportion of homicides cantly over recent decades. committed with firearms, and a low proportion of Conventional typologies of armed violence cases solved by law enforcement. Chapter Four based on the context, intention, and type of examines global patterns of violence against actor have limitations for both research (data women. The final chapter studies the links between collection) and policy-making (prevention and armed violence and development by considering reduction programmes). the impact of armed violence on progress towards The boundaries between political, criminal, achievement of the Millennium Development and intimate or gender-based violence have Goals and other development indicators. It finds become increasingly blurred, as revealed in that countries with high and very high levels the cases of Iraq, Mexico, and Somalia. of violence are concentrated in the low human Effective violence prevention and reduction development and low-income categories, and that programmes and policies need to start with a there is a persistent link between poverty, armed unified assessment of the scope, scale, and violence, and development. gender-based violence. and and interpersonal violence, motivated cally criminal and economi- organized violence, terrorist and state, tointer-communal, war with ciated asso- violence large-scale the from range These actor. violent the of orintentions organization of level tothe according violence and criminal conflict of forms different todelineate attempt that andcategorizations typologies of range awide lies distinctions general these Beyond disputes. interpersonal or example) for robberies, during (murders gain economic impersonal by ismotivated it whether of 2008), regardless andWelch, (Riedel law by sanctioned isnot that violence armed of use considerations. topolitical subordinate are that ends particular toachieve way inacalibrated force and uses organized, ishighly means, other by struggles political of continuation is the forms) various its (in ‘war’ towhich according logic, Clausewitzian a follows violence Conflict acts. violent the behind, motivations of, and the organization of level the capture distinctions These violence. and motivated) andbetween violence, vidual) organized between drawn are distinctions common most two The perpetrator. the of intentions underlying orthe context the as tosuchfactors according categorized be can that types distinct of series a as treated isconventionally violence Armed Why perspective? aunified Figure 1.1 The ‘macro’ distinction is between is between distinction ‘macro’ The 1.1 Figure in presented violence armed of picture general tothe rise gives classification of sort This orrape. assault, kidnapping, extortion, violence, gang terrorism, such as acts violent specific into disaggregated be can (collective) and (collective) 3 criminal Criminal violence is simply the the issimply violence Criminal (economically motivated) motivated) (economically 4 Each of these categories categories these of Each interpersonal conflict (politically (indi- 1,000 victims 1,000 than more the for case the was as groups, nized orga- loosely by inmassacres ‘combat’ formal of outside killed often more are violence, organized of contemporary of victims proportion significant widely different estimates of the burden of conflict conflict of burden the of estimates different widely andleadto conflict of victims of majority the for Violence Armed of Burden Global the of edition inthis briefly only with dealt are conflict—which from deaths non-violent Indirect conflict. armed of aconsequence as populations civilian by suffered mortality orexcess deaths violent) (non- indirect the as well as groups), and armed states by victimized ordirectly zones, conflict in trapped (either andcivilians combatants both of deaths violent the through violence armed of burden global tothe contributes violence Conflict consequences. lethal the counting of ways different the captures category last the involved; andmotives actors of types different the captures distinction second-order the violence; of organization of levels different (AI, 2009, p. 22; Foglesong and Stone, 2007, and Stone, p.18). p.22;Foglesong 2009, (AI, deaths violent of cent 13–43 per for account can they however, orNigeria, Jamaica suchas places in inallcountries; statistics inhomicide counted systematically not are killings judicial andpolice extra- example, For levels. different between ‘translation’ isanimperfect there violence, armed of incidence actual the of picture an adequate toprovide likely orless more are violence lethal counting of ways various the While forces. security state by killings or‘legal’ extrajudicial as well as (manslaughter), killing der, andunintentional ormur- homicide intentional of interms captured iscommonly violence ornon-conflict Criminal in2011. East Middle andthe Africa North of parts across visible was as demonstrators, unarmed against violence orinstate in 2008, 5 Civilian non-combatants, who form a form who non-combatants, Civilian 6 of post-election violence in Kenya inKenya violence post-election of —can account account —can 7

13 2 3 4 5 1 A UNIFIED APPROACH 14 Figure 1.1 A framework of armed violence categories

ORGANIZED (COLLECTIVE) VIOLENCE INTERPERSONAL VIOLENCE

LEVEL OF ORGANIZATION

Violent (State) violence Terrorist Communal Gang and Domestic and Inter-state and assaults and against attacks and and non-state organized crime intimate partner civil war armed civilians violence violence violence violence ARMED VIOLENCE 2011 VIOLENCE ARMED robberies of

LETHAL OUTCOMES GLOBAL BURDEN BURDEN GLOBAL CONFLICT DEATHS Victims of extrajudicial killings NON-CONFLICT DEATHS

CIVILIAN Battle-related deaths CONFLICT Civilian deaths DEATHS Victims of terrorism INTENTIONAL HOMICIDES

Victims of legal INDIRECT CONFLICT interventions DEATHS UNINTENTIONAL HOMICIDES

deaths. The Democratic Republic of the Congo is Report 2009/2010 revises these figures and estab- an interesting, yet controversial case. The Inter- lishes an estimate of up to 2.4 million deaths national Rescue Committee, for example, finds (HSRP, 2010, part II, p. 38). The first edition of that up to 5.4 million people died between 1998 the Global Burden of Armed Violence suggests a and 2007, of which only 10 per cent were victims global average ratio of four indirect deaths for of violence (Coghlan et al., 2006, p. 44; IRC, 2007, every direct (violent) death due to armed conflict, p. ii). On the other hand, the Human Security although this depends heavily on the nature of the to which Iraqis have been subjected. been have Iraqis to which violence’ ‘everyday tothe compared as especially victimization, violent for estimates leads tolow agenda— apolitical with group armed recognized a by claimed are that attacks deaths’—violent ‘conflict counting on focus anarrow example, for 2010). Iraq, In (Stepanova, actors same the by perpetrated andbe simultaneously, present be also can violence armed of forms Different 483). pp.480, 2003, (Kalyvas, resources and land over ordisputes feuds, orclan family rivalries, local tosettle violence touse tions motiva- own their with parties external co-opted actors local cases, both In bombed. be would targets these that so presence Taliban orAl-Qaeda alleged about USforces informed factions local inAfghanistan, villagers; local by provided lists name on based perpetrated were killings instance, for war, civil Guatemalan the During overlap. can motives how of examples striking provides denunciation of practice The situation. violent inone present be can andrespect) resources, revenge, (suchas motives andother ideological, identity-based, economic, political, different motives; andoverlapping multiple have can violence Yet armed boxes. neat the one) of only (and in one fits incident violent particular any that misleadingimpression the give they is that first The settings. indifferent violence to armed responses andprogrammatic practical effective todevelop ability our hinder distinctions sharp and classifications convenient why reasons four are There picture? this with iswrong what So p.32). 2008, Secretariat, Declaration (Geneva response humanitarian andthe conflict based conflict, non-state conflict, and one-sided andone-sided conflict, non-state conflict, based state- datasets—covering (UCDP) Program Data Conflict Uppsala combined The inIraq. alone 2006 for killed 27,000 civilians records example, for Project, Count Body Iraq the by provided 8 The data data The similar findings appear throughout the 20 the throughout appear findings similar 1927); (Abbott, waves’ ‘crime non-conflict into over spill can violence wartime that note War, analysts USCivil the as back Yet far as causes. underlying similar share ormay linked, be may violence of forms different inwhich ways the tosee difficult is it isthat result One andconditions. survivors, victims, perpetrators, of system aparticular within self-contained as violence armed of forms different treat categorizations andexclusive rigid Finally, 1.1. inBox alysed an- Somalia—is of coast the off the piracy rise of example— recent, more Asecond, p. 294). 2010, (Malaquias, violence protracted tothe a support provided trade diamond coldwar, the the of end the at smallarms of availability unprecedented an with combination In mining. diamond on based activity predatory more favoured that constraints andstrategic political faced in1992,UNITA process electoral failed the after resumed war civil the Once p.347). 2009, Beck, (Koloma arena coldwar inthe USsupport from benefited it such, as region; inthe expansion’ Russo–Cuban the ‘against struggling as itself defining tone, aadopted clear anti-communist group the 1970s, p. 2010, 296). Inthe (Malaquias, group ethnic largest Angola’s Ovimbundu, the to avoice providing movement a revolutionary as emerged first UNITA example, for Angola, In actors. violent of andcapabilities motives of ity fluid- the reflect that andlulls peaks with tempo ashifting following often time, over to another form one from change also can violence Armed Colombia, the Democratic Republic of the Congo, Congo, the of Republic Democratic the Colombia, andHerzegovina, Bosnia as in suchplaces women against violence of andlevels patterns the that torecognize begun have researchers recently, 1976). More andGartner, (Archer tury the same period. for intotal deaths 4,261 only violence—register 9 th cen- 15 2 3 4 5 1 A UNIFIED APPROACH 16 Box 1.1 Somali pirates: bandits and soldiers of convenience Photo Armed Somali pirates prepare a skiff in Hobyo for future attacks, Somalia, January 2010. © Mohamed Dahir/AFP Photo Armed violence has been pervasive in Somalia since the 1991 collapse of the state and the ousting of Siad Barre’s government. Thousands of people, especially civilians, have suffered directly and indirectly from the armed confrontations among the numerous factions at war in the country. The heavy human toll of the early years of conflict resulted both directly from armed violence and reprisals against civilians, and indirectly from the devastation of farmland, which brought starvation upon thousands of Somalis and displaced tens of thousands of people.

The violent collapse of the Somali state is also linked in complex ways to the increasing acts of piracy in the Gulf of Aden. Initially, with the disappearance of any state surveillance of Somalia’s maritime waters, ﬁshermen took up ARMED VIOLENCE 2011 VIOLENCE ARMED arms to oppose illegal industrial ﬁshing and international waste disposal of off the coast. These armed groups rapidly realized that unarmed commercial vessels represented a convenient opportunity for enhanced income generation (Lennox, 2008, p. 9). Initial small-scale attacks quickly grew more sophisticated, peaking with the seizure of the Sirius Star, a tanker with an estimated USD 100 million worth of cargo, and the seizure of the Faina and its cargo of 33 Russian T-72 tanks, weapons, and ammunition (Balakrishnan, Rice, and Norton-Taylor, 2008; Höges, Klussmann, and Knaup, 2008).

The scope and nature of the attacks are also influenced by the availability GLOBAL BURDEN BURDEN GLOBAL of weapons and the environment of pervasive insecurity. The International Maritime Bureau finds that the 213 reported attacks by Somali pirates in 2010 was double the figure for 2008 (and four times higher than the 2007 figure). Furthermore, Somali pirates perpetrated 48 per cent of worldwide reported attacks. Indeed, for the years 2009 and 2010, Somali pirates were responsible for an average of 80 per cent of all attempted attacks, and an average of 25 per cent of successful acts of piracy. Overall, they were responsible for half of the attempted and actual attacks worldwide (IMB, 2010; 2011).10 The Somali pirates were also more violent than their peers: although they account for only half of the attacks worldwide, Somali pirates accounted for 86 per cent of the hostage-taking, all of the deaths, and 78 per cent of the attacks involving guns in 2010 (IMB, 2010; 2011).

Piracy is by definition a violent act serving private economic interests, but in practice Somali pirates are entangled in local conflict dynamics.11 Though pirates have kept some distance from the ongoing civil conflict, recent reports show that these groups have bolstered their armaments, and that local government officials as well as opposing militias are increasingly relying on the pirates’ firepower and strength for carrying out protective and predatory tasks (Gettleman, 2010).

The involvement of piracy—commonly associated with international organized crime—in the Somali conflict illustrates how blurry the distinction between criminally motivated and ideologically motivated violence has become. Though Somali pirates have so far not been directly involved in killings in the ongoing conflict, the fact that they are linked to the warring factions represents a serious threat to safety and security in the region. 17 2 3 4 5 1 A UNIFIED APPROACH 18 El Salvador, or Iraq may be conditioned by the arresting youths suspected of gang membership’ experiences of war and deeply entrenched con- (Rodgers, Muggah, and Stevenson, 2008, p. 16; flict dynamics.12 And the high suicide rates among Bateson 2009, p. 7). Yet data in El Salvador, for members of the US armed forces in the past example, shows that only around 13 per cent of decade are linked to wartime experiences in com- homicides in 2008 were attributed to gangs plex ways (USDoD, 2010). (IML, 2009, p. 70). The focus on gangs, however, means that other violent acts are left aside, such These four aspects—the multiple, simultaneous, as the 93 extrajudicial killings reported in 2006, and shifting motivations of violent actors, and the and the more recent escalation of drug-related links between different forms of violence—con- violence linked not to gangs, but to organized found simple classifications and policy responses. crime (Aguilera, 2008, p. 134). Facing criticism Rigid distinctions and categorizations lead to policy ARMED VIOLENCE 2011 VIOLENCE ARMED and lack of effect, heavy-handed approaches

of stovepipes, in which policies and programmes to have started to be combined with mano amiga deal with one sort of armed violence (gang violence, (Friendly Hand) and mano extendida (Extended for example, or conflict prevention strategies) are Hand) interventions, which focus more on incen- developed in a narrow fashion that disregards the tives for demobilizing gangs and the establish- way in which different forms of armed violence ment of stricter controls on small arms (Rodgers, can be closely linked. In Liberia, for example, Muggah, and Stevenson, 2008, p. 16). Finally, much of the post-war effort focused on more policy stovepipes can also lead to category GLOBAL BURDEN BURDEN GLOBAL traditional demobilization, disarmament, and errors, in which programmers misunderstand reintegration, as well as security sector reform. or mistake the form of or motivation behind the Post-war Liberia, however, faces other serious violence encountered, leading to inappropriate challenges; high unemployment, a large youth programmes or responses. population, and severe development needs all represent sources of discontent and risk factors for potential armed violence. To date, post-war security promotion efforts have not tackled Obstacles to a unified perspective these issues as a way of reducing the risk of the Achieving a unified perspective on armed more criminalized forms of violence (Small Arms violence is difficult. Policy-makers require Survey, 2011a, p. 2; 2011b). boundaries within which to structure practical Policy stovepipes also result in blind spots, in programmes. Public health practitioners, for which the most important types of violence may example, rely upon the World Health Organiza- be neglected due to the policy predispositions or tion’s epidemiological model, which focuses on orientations of major donors and stakeholders. risk and resilience factors while seeking to identify For example, strategies to deal with homicidal factors that can be addressed at the individual, violence in Central America have focused on community, and societal level (WHO, 2002). The violent gangs, neglecting the range of other vio- World Bank’s World Development Report 2011: lent acts and actors in the region. The initial Conflict, Security, and Development relies mainly response to violence in the region ‘can be char- on data relating to conflict, and only to a much acterized as enforcement-first’ measures, ranging lesser extent on data and analysis relating to from tougher prison sentences to ‘aggressively homicide and organized crime (World Bank, 2011). These challenges can be overcome through through overcome be can challenges These andconditions. contexts tospecific and adapted migrated successfully be can practices and best solutions promising which tolearn platforms in‘multi-stakeholder’ of actors range a wide together tobring policy-makers challenges goal this Furthermore, communities. and between within andinsecurity violence of patterns larger into fits interest particular their where on spective per- andgain focus, their widen predispositions, their aside toset communities research different requires It challenge. andananalytical practical a both isthus perspective aunified Achieving (IISS). Studies Strategic for Institute International and the Ploughshares, Project UCDP, suchas centres, research society and civil academic of aseries by isconducted conflict on research andCrime; Drugs on Office Nations United the suchas andorganizations nologists crimi- of realm isthe example, for homicide, on Research purposes. different andfor standings under- todifferent according violence armed of anddistribution scope the on data collects them eachof but base, andstatistical literature asolid produced have communities these years, the Over toanthropology. toeconomics relations international andfrom health, topublic nology crimi- from ranging disciplines, of array in awide isabundant women against andviolence lence, vio- gang crime, conflict, on Research disciplines. between communication little with violence, armed of manifestation oranother one understanding in specialize andanalysts researchers Similarly, erupted. has conflict’ ‘armed formal a inwhich contexts few those on mainly focuses community and peacebuilding prevention conflict the And programming. andreduction prevention violence other from inisolation works generally women against andviolence violence based gender- of issues on working community The 14 violence (Gilgen and Tracey, 2011, pp.38–39). andTracey, 2011, (Gilgen violence and reduce toprevent andstrategies issues on discussions andcatalyses community, the into information this brings hotspots, violence and crime identifies It sources. local from data and data crime police with hospitals from data links example, for Observatory, Crime Jamaican The processes. topolicy-making directly analysis and linkresearch which observatories, violence or systems monitoring violence armed effective argue that non-violent means of resolving con- resolving meansof non-violent that argue Others institutions. multilateral of efforts building andpeace- peacekeeping ofthe success steady but slow tothe linked are declines these that argue observers Some pp.148–51). 2005, Gleditsch, and Lacina 2011; Bank, (HSRP,war World 2010; major of end the toherald andeven war of costs human shrinking the to highlight analysts some led has This conflicts. of lethality overall the by measured as globe, the around conflict armed of intensity the as well as incidence inthe decline areported witnessed also have decades past The were these of All in2009. active conflicts armed state-based 36 with stabilized, has ongoing are and that states involve that conflicts armed of number the Sincethen, 1990s. early inthe peaking increased, steadily wars andcivil conflicts armed internal of number the while decades, recent over declined have states between warfare of forms Traditional decades. two past the over documented well been have violence armed of forms organized highly most the of andintensity scope, inlocation, Shifts violence in contemporary conflicts Neither war nor peace: armed side states (Pettersson and Themnér, 2010, p.16). 2010, andThemnér, (Pettersson side states out- from combatants involved them of and seven internal conflicts, conflicts, 15 19 2 3 4 5 1 A UNIFIED APPROACH 20 ARMED VIOLENCE 2011 VIOLENCE ARMED of GLOBAL BURDEN BURDEN GLOBAL

Photo Supporters of the opposition take part in post-election protests in Kibera slum, Nairobi, Kenya, January 2008. © Shaul Schwarz/Getty Images pp. 312, 342). 2008, (CIPEV, killed) persons (744 overall victims of numbers largest the suffered also deaths), which allfirearms-related of cent or47per (194 victims Valley Rift inthe highest were victims gunshot of numbers the allkillings); of cent per (35.7 Kenya across cases in405 death of cause the as identified were orgunshots firearms addition, p.237). In 2008, (CIPEV, in2007–08 violence post-electoral the during abuses sexual and other rapes perpetrated reportedly andpolice army Kenyan the p.200). Furthermore, 2008, (Mueller, game’ ethnic zero-sum in a‘winner-takes-all organized been have traditionally politics party where inanenvironment violence inthe pated Mungiki—partici- Taliban Kikuyu Luo andthe the as gangs—such aligned) (and ethnically Active estimated 350,000 persons (CIPEV, 2008, p.272). 2008, (CIPEV, persons 350,000 estimated an anddisplaced lives 1,000 than more claimed 2007 of elections disputed the after erupted that clashes violent the example, for in Kenya, 2008 In state. the of involvement direct the out with- population, the of segment aspecific target orthat eachother, confront that groups armed organized formally orless more of ing number increas- isthe captured easily isnot that conflicts armed contemporary of feature important One violence. tive collec- of outbreaks slideinto can countries how andof change, political of nature divisive the of examples recent two only are 2010 November of elections presidential contested d’Ivoire’s Côte after clashes armed 2011of andthe Spring Arab violent The tenuous. remains wide world- violence conflict-related inlethal drop the explanation, the Whatever interactions. violent replacing slowly are states indemocratic flicts or militias) as well as the Kenyan security forces. security Kenyan the as well as or militias) gangs, groups, ethnic (suchas groups non-state identified eight least at involved violence The 17 16

21 2 3 4 5 1 A UNIFIED APPROACH 22 This rise in armed group activity can be illustrated economies are all indicators of the complex web more generally in several ways. One way is to of violent actors involved in different forms of vio- track the number of non-state armed actors around lence at the local, national, and regional levels. the world, which, according to some estimates, The heightened activity among organized armed may reach more than 1,000 different groups, groups can also be illustrated by the number of large and small, around the globe (Carey and armed conflicts that do not involve governments. Mitchell, 2011, p. 1; IISS, 2009, pp. 465–74; Although the figures fluctuate widely from year 18 UCDP, 2011). Figure 1.2 illustrates that while the to year, the number of non-state armed conflicts overall number of international and civil wars has reached 35 in 2008—more or less the same as been stable or declining since the mid-1990s, the number of armed conflicts involving states the number of active pro-government militias (HSRP, 2010, ch. 11, pp. 5–22).19 Although mostly ARMED VIOLENCE 2011 VIOLENCE ARMED around the globe increased steadily from the of small-scale, this form of violence poses signifi- early 1980s through 1990s, only to decline since cant security and development challenges; the 2003. The increasingly complex power struggles 2008–08 post-election violence and political and the atomization of actors and political con- uncertainty in Kenya, for example, had a direct figurations have been noted in conflict settings impact on the economy.20 After several years of such as Darfur, Iraq, and Pakistan (Stepanova, high annual GDP growth (more than six per cent 2010). Predatory groups, counter-insurgency per year in 2006 and 2007), growth dropped to

GLOBAL BURDEN BURDEN GLOBAL operations undertaken by states, insurgent 1.6 and 2.6 per cent in 2008 and 2009, respec- actions, and criminal activities linked to war tively (World Bank, n.d.).

Figure 1.2 Active pro-government militias and the number of civil wars, 1981–2008 Number of pro-government militias (PGM) Number of civil wars Number of countries with at least one active PGM

Number

120

110

100

1981 1982 1983 1984 1985 19861987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Source: Carey and Mitchell (2011) than during a preceding armed conflict, as in as conflict, armed apreceding during than higher or even high as rates mortality to reach violence peacetime for uncommon isnot It 1.2). Box (see Iraq of case the by illustrated as wrong, are beliefs three these of ormore one contexts inmany But violence. andnon-conflict conflict of forms between links strong no are there that and settings, inconflict higher are victimization and violence of levels that improves, and safety andsecurity decline violence of levels overall terminate conflicts when that beliefs: three on based isgenerally violence andnon-conflict flict con- between divide andprogrammatic analytic sharp The settings. post-conflict in so-called occur can that violence armed of transformation the recognize must analysts conflicts, armed contemporary of nature changing tothe addition In p.13). 2010, p.225; UN, 2005, andStuddard, (Nitzschke agreements or peace cease-fires tonegotiate interventions targeted narrowly tolaunch difficult ismore it complex, more become them between links the and as increases, context inaparticular actors violent of andnumber type the As opposition). suppress (to orstate-led groups), armed (formal predatory and organized highly orresources), land over fight groups neighbouring (aswhen organized loosely and opportunistic be may violence The conflict. and insecurity of context overall tothe another or way in one linked are they but organization, formal of levels different have may groups the violence; of pockets of eruption the by terized charac- fire toabrush compared be can groups armed non-state involving conflicts Rather, gain. oreconomic change political ameanstoachieve as violence uses that goal andcommon ideology lar aparticu- with actor homogeneous aunitary, as group anarmed ideaof the question into calling andstrategies, goals collective defined easily less andhas stable isless group non-state A violent in many ways. in many linked be can violence andpost-conflict Conflict pp.37, andTobón, 2011, 46). (Restrepo police Guatemalan the by reported were cide victims homi- 6,498 year. per Yet alone, deaths in2009 3,508–5,800 into translates which killed, were people 119,300–200,000 war, anestimated civil 36-year the During Guatemala. contemporary interact in complex (and poorly understood) ways. understood) (and poorly incomplex interact and causes root share can violence of ifestations man- anddifferent blurred, become may motives andcriminal political time, Over p.141). 2010, Survey, (SmallArms situation security the of tion aggrava- toanoverall contribute also can settings inpost-war weaponry accessible Easily p. 308). 2009, andMuggah, (Rodgers time over existence groups’ these tosustain as well as level individual the at generation income for anoption become tohave known are activities p. 130). Criminal 2008, Secretariat, Declaration (Geneva violence from communities ortheir themselves to protect groups vigilante-like into themselves organize to people push also can uncertainty economic andsocio- insecurity heightened of situations to combatants former demobilized of return The conflict. the of settlement the after nal violence crimi- of continuation tothe also often but itself, conflict armed of duration tothe only not tribute con- economies towar related networks illegal importantly, Most andLiberia. Guatemala such as incountries andvigilantism lynching included has that justice andrough policing toinformal led have capacities policing effective of lack and the institutions state of breakdown the Similarly, inslumsorshantytowns. cleansing operations andsocial killings extrajudicial of incidence toahigher haveled andMozambique Salvador, El Colombia, suchas incountries conflicts during groups paramilitary of formation andthe sector 21 The militarization of the security security the of militarization The 23 2 3 4 5 1 A UNIFIED APPROACH 24 Box 1.2 How different forms of violence are linked in Iraq In post-invasion Iraq, the widely reported targeting of non-combatants by insurgents, militias, and sectarian groups may have appeared chaotic or random at first glance. Yet a closer look at underlying patterns and motives of violence suggests that this seemingly arbitrary or criminal violence may also serve ‘clear political purposes’ aligned with the political goals of these armed groups (Green and Ward, 2009, p. 612).

ARMED VIOLENCE 2011 VIOLENCE ARMED Such ‘dual-purpose violence’ in Iraq is charac- of teristic of both the politically motivated violence of insurgents and organized criminality since 2003. Individuals, often with a criminal background, ‘preﬁnanced’ future insurgent activities by participating in the generalized looting shortly after the fall of the Ba’ath regime, and impover- ished looters targeted the homes of the political elite ‘in acts of political revenge but also to satisfy GLOBAL BURDEN BURDEN GLOBAL long accumulated material needs’ (Green and Ward, 2009, p. 618). Notorious organized crime ﬁgures have reportedly helped insurgent cells to fund their activities through kidnappings, bribery, and highway robberies (p. 619).

Violence against women has also been used to serve sectarian or political ends, demonstrating how common criminality, individual motives, and collective violence interact in Iraq. Under the Ba’ath regime the hijab (women’s traditional dress code) was forbidden and women’s rights were a ‘bargaining chip’ for the political elite. While advancing a rather progressive position in terms of women’s rights and participation in society and the workforce, the Ba’ath regime ‘accepted tribal practices’ in return for loyalty from local leaders (Green and Ward, 2009, p. 614).

Since 2003, traditional or sectarian practices against women have been used as part of political struggles. When a faction takes control of a

territory, the imposition of the veil or a strict dress Photo A woman stands next to the grave of her daughter, a Sunni code for women is usually among the first meas- who was killed by Shia militiamen, Baghdad, March 2008. ures announced as part of a wider ‘campaign of © Ghaith Abdul-Ahad/Getty Images cially through ‘honour’ killings and rape. and killings ‘honour’ through cially espe- violence, of levels high suffer and groups different among polarization increased of ments instru- and victims both are women case, Iraqi the In conflict. ongoing overall the of umbrella the under women against violence in engage to actors various for opportunities increased has Iraq in conflict The violence. for motives private and personal as well as political modate accom- acts violent individual how demonstrate women against violence of forms various These p.16). 2008, UNAMI, 2007; (USDoS, including self-immolation incidents, related ‘honour’- in burnt 150women and killed women Kurdistan Regional Government 56 reported the meanwhile, behaviour; un-Islamic allegedly their for dumped bodies their and killed were 2007,women In 57 Kurdistan. Iraqi of region northern the and Basra in concern particular of been have killings ‘Honour’ 2008). (Sarhan, murders sectarian like appear them make and killings out carry to gunmen professional hired had members family reported, lawyer Iraqi an cases, some 2007. 47in In versus murdered, been had 81women 2008, November to January From rise. the on were Basra of city southern the in ings UK-based the 2008, In p.615). 2009, Ward, and (Green opponents their upon harm and inflict intimidate maliciousdenunciationto have used conflict in families of members that reported been also has it but kin, their upon shame brought have who women denounce often members Family p.15). 2008, (UNAMI, militias different with affiliated are who guards university of attention the attracting avoid to order in families their by codes dress conservative adopt to pressed are students female many that reports Iraq for Mission Assistance Nations United The terms. sectarian in ‘dishonour’ define that groups armed by trated perpe- increasingly also are killings ‘Honour’ p.615). 2009, and Ward, (Green areas in militia-controlled terror’ Guardian reported that ‘honour’ kill- ‘honour’ that reported 25 2 3 4 5 1 A UNIFIED APPROACH 26 Neither peace nor war: violence in ward trend in Western Europe since the beginning of the early modern period. Homicide rates across post- and non-conflict settings Western Europe—in what are today Belgium, The incidence and severity of lethal violence in England, Germany, Italy, the Netherlands, Scan- non-conflict settings has also undergone signifi- dinavia, and Switzerland—declined by roughly cant changes over the past few decades, yet the half from the early 17th to the early 18th century, picture is more complex than that presented for and by the 19th century, they had dropped three armed conflict. The limited empirical data on to five times further (Eisner, 2001; Gurr, 1981; historical homicide rates reveals a regular down- Monkkonen, 2001).22 Although the exact timing

Figure3 1. Aggregated homicide rates in 13 selected Western European countries, 1970–2009 (base year=100) ARMED VIOLENCE 2011 VIOLENCE ARMED

of Change (basis year 1970=100)

200

180

160

140

120

100 GLOBAL BURDEN BURDEN GLOBAL

1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986

Change (basis year 1987=100)

200

180

160

140

120

100

1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Note: The rate for the initial year is indexed at 100 to facilitate trend comparison. Source: Small Arms Survey elaboration based on UNODC (n.d.b.) one homicide per 100,000 persons. 100,000 per homicide one around of level isalow all13 countries for rate average the 1987. of 2009, rate For homicide initial the below rates Spain) of exhibit exception the (with review under 13 the countries of majority the 2009, in 1970. By than in1986 rate lower a have andSpain Finland, Austria, only rates; homicide recorded inthe increases severe suffer countries several 1970 and1986, Between worthy. note- also are trends level, country the At years. six injust cent 40per decreased rates homicide in the early 2000s. early in the todecline begin and eventually stabilize, then 1990s, early the until toincrease continue Rates trends. on amajorimpact have can cide defined orhomi- isrecorded data inhow small changes meanthat inEurope rates homicide low very the since caution some with considered tobe have figures these although andSwitzerland, in Norway seen are increases highest The 100,000. per cides homi- three around of arate reaching and 1986, 1970 between cent per to80 close by increased gradually rates homicide that reveal figures The trends. of comparability the to ensure year first the for to100 set are rates country All difficult. very andrates trends of analysis renders and observed countries the throughout rates of fluctuation inasignificant isvisible which in1987, homicide of recording statistical the in change duetoanapparent and 1987–2009, 1970–86 periods, two into isdivided data The Systems Justice Criminal of Operations the the by collected data on based states, 13 European in rates homicide aggregated the of evolution the 1.3 illustrates Figure significantly. fluctuate do rates homicide however, term, medium the In states. European within violence inlethal decline long-term the about doubt islittle there place, to place from varies decline the of and scope United Nations Survey of Crime Trends Crime of and Survey United Nations 23 Between 2003 and 2009, and2009, 2003 Between (CTS).

2000s, reporting decreased first but then stabi- then but first decreased reporting 2000s, the of beginning the In improving. simply ing was report- infact when increasing, were rates global though as seem madeit andUkraine) Estonia, Colombia, (suchas rates homicide high orvery high relatively with states from data more of inclusion progressive the consistently; reporting were 67 states about 1990s, the By period. in this under-represented especially was which Africa, Sub-Saharan except regions allother from vided pro- were samples though Europe, from were most reported; 50states around only example, for 1970s, early the In rates. global the changes radically states of andgroups states of (and exit) entry the andsince survey, inevery report states no almost since trends global togenerate used be cannot Yet data this Trends Surveys. 11 Crime of series its from 1970–2008 period the for 163 countries covering data published has andCrime Drugs on Office Nations United The analysis. trend for basis the as serve could that data historical reliable of isalack there level, global the At p.210). 2008, (Spierenburg, wounds andother gunshot from mortality reduce that interventions medical emergency improved as including suchfactors emerged, have drop the for explanations several 2007); Zimring, 2010; Tseloni al., et 2004; (Aeby, crime violent on inliterature recognized largely been also has mid-2000s inthe andagain mid-1980s sincethe rates homicide of trend downward The 2002). andDrass, (LaFree crimes tocommit ties opportuni- resulting andthe wealth increasing of manifestation a‘normal’ orconsidered in Europe, change andsocial modernization torapid variously attributed been has 1980s 1970s andearly the in trend upward The term. andshort medium the in violence lethal influence factors what about questions raise fluctuations these term, longer the over low remain rates European Although 27 2 3 4 5 1 A UNIFIED APPROACH 28 lized at a higher level, with around 95 countries There are several possible explanations for these reporting homicide data by 2005–06. different—and fluctuating—trends at the European and global level. Most analysts point to the pro- It is thus difficult to derive long-term trends from gressive development of modern state institutions, global averages based on partial and shifting and the expansion of the state’s practical monopoly data alone. Coverage has been comprehensive over the legitimate use of force through security enough to present trend data since 2004, reveal- institutions such as law enforcement and national ing an upward trend in global homicides in the defence as one explanation for declining homicide second half of the 2000s (see Figure 1.4). As (and overall crime) levels. Not only has the state noted above, Western European countries have been able to intensify its presence over its terri- a decreasing homicide rate over this period, but tory, but scientific progress and institutional in other parts of the world, entire regions were ARMED VIOLENCE 2011 VIOLENCE ARMED reorganization of police forces have also improved

of suffering from generally increasing rates in homi- its capacity to contain crime and apprehend per- cidal violence. Despite the rather small magnitude petrators (Spierenburg, 2008, pp. 169–70). The of the increase since 2004 (about 5 per cent), the low levels of violence in Europe also seem to reflect global impact in terms of human lives lost is sig- changing normative understandings of the legiti- nificant. The difference between the low point of mate use of violence and the importance of personal 2006 and the figure for 2009, drawing on data security.27 Violent practices that were commonplace from the Global Burden of Armed Violence data- a century or so ago—such as public executions, GLOBAL BURDEN BURDEN GLOBAL base, represents more than 54,000 additional torture, and lynchings—have today been stigma- homicide deaths (and an increase of 24,000 tized to the point of near extinction in the West, since 2004). although there are notable exceptions. Similarly, the everyday use of violence to resolve conflicts has Figure 1.4 Global homicide trends, 2004–09 been condemned, even in what was hitherto con-

Homicide rate per 100,000 population (index year 2004=100) sidered the ‘private sphere’, where intimate partner

120 violence existed beyond the reach of the state.

110 Behind these ﬁgures lies a more complex reality,

100 one in which lethal non-conﬂict violence unfolds

90 with different levels of organization. Although non-conﬂict lethal violence is generally counted as 80 ‘homicide’, it is often linked to highly organized 70 criminal activity or to different forms of ‘political 60 violence’ that either target political opponents or 50 civil servants such as mayors, teachers, or police 40 ofﬁcers, or that seek to modify government poli- 30 cies. In these contexts, the term ‘homicide’ is a 20 slightly misleading term since it conjures up images 10 of individual inter-personal acts of violence.

2004 2005 2006 2007 2008 2009 One result is that the distinction between the Source: GBAV 2011 database activities of politically motivated armed groups power-sharing offers may be less effective as as effective less be may offers power-sharing suchas compromises political criminalized, become have groups armed where circumstances in implications: policy important has shift This groups. criminal organized of model business self-perpetuating the resemble activities their and fade goals ideological purpose, political ent appar- without violence predatory its maintains Army Resistance Lord’s Ugandan orthe financing, self- for need andthe opportunities increased of because activities criminal towards turn Colombia in FARC the suchas groups motivated politically p.17). ostensibly Yet 2009, when (Steenkamp, funds’ toraise activity incriminal engage allsides on protagonists political as actors andpolitical criminal between relationship knit isa‘close- result One efforts. war their sustain to income generate groups armed how of ples exam- classic are narcotics of and distribution production 1.3). Box The see 2007; Wennmann, p.234; 2010, (UNODC, andLiberia Colombia, andHerzegovina, Bosnia Afghanistan, such as inplaces effort war tofinancethe used been have ingoods, trafficking suchas activities, Criminal p.42). 2010, (Stepanova, goods luxury such as ends ductive unpro- for profits use andoften business their of expansion the on focus hand, other the on tions, organiza- criminal change; andsocial political toachieve troops) their for andcare weapons (tobuy resources tomobilize activities criminal use Rebels activities. these from arising profit the of use planned inthe lies maindifference the tooperate, money require both p. 210). Although 2007, (Cornell, profit’ ‘monetary pursuing latter andthe causes’ higher ‘self-defined pursuing first the idealtypes’, ‘opposing as seen been have groups armed motivated and economically politically Traditionally, spectrum. the of ends both from blurred has organizations and criminal the struggle against the drug traffickers. drug the against struggle the abandoning into authorities the andtocoerce tions institu- local isto weaken goal sinceone political, and criminal isboth cause andthe frequent, are officers andpolice mayors suchas servants civil and politicians of Killings territory. inits drugs on awar towage army the mobilized has state Mexican the while north, tothe narcotics of flow the over control for inabattle locked are cartels Drug 1.4). Box (see inMexico war drug the by denced evi- as violence, of types andpolitical criminal between boundaries the blurring for capacity amarked shown andhas violence of levels high by isaccompanied narcotics, illicit of trafficking the andespecially crime, Organized p.303). 2003, Kemp, 2010; (UNODC, security ornational regional, tointernational, athreat as recognized increasingly resources—are environmental goods, counterfeit smallarms, humans, drugs, trafficking—in of forms Different profit-seeking. remains motive main ifthe even andconsequences, implications political have can groups criminal organized of activities violent the angle, another from Seen phenomena. separate than rather acontinuum as of thought best are groups andpolitical criminal between interactions the that imply also developments 2008). These (Goodhand, operations criminal established their from profit that entrepreneurs’ ‘violence toremain interesting more findit groups some terrorist financing (Jojarth, 2009, p.8). 2009, (Jojarth, financing terrorist as such activities, illicit to destined but legally obtained and legal are that goods and goods; stolen or diamonds conflict as such ways, illicit in obtained or processed goods narcotics; as such goods, illicit covers ‘trafficking’ term The B ox 1. 3 Defining illicit trafficking illicit Defining 29 2 3 4 5 1 A UNIFIED APPROACH 30 Box 1.4 Mexico’s drug war Violence and insecurity in the wake of President Felipe Calderón’s war on drugs have reached extremely high levels in several regions of Mexico. Official estimates set the human cost of Mexico’s drug war at around 35,000 dead from 2006 to 2010 (Booth, 2010; Turbiville, 2010, p. 124; Mexico, n.d.a).28

Mexico’s overall levels of violence have long been steady—with an average violent death rate of 11.5 per 100,000 persons in 2004–09 (ICESI, n.d.). But this national rate, while demonstrating that most parts of the country are generally safe, masks the bitter reality that some cities and regions suffer from extraordinarily high levels of violence, higher than found in many war

ARMED VIOLENCE 2011 VIOLENCE ARMED zones (see Figure 2.9, TRENDS AND PATTERNS). of At the state level, Chihuahua—home of Ciudad Juarez—had a rate of organized crime-related deaths of 98.6 per 100,000 in 2009 (with an overall homicide rate of 108.0); it is followed by Durango, which has a rate of 43.5 per 100,000 (Mexico, n.d.a).29 By contrast, Mexico City had an overall murder rate of 8 per 100,000, and a drug-related homicide rate of 1.5 per 100,000. Yucatán and Puebla are among the states witnessing the lowest incidence of drug-related violence and overall homicide rates, with rates below 1 per 100,000 for drug- GLOBAL BURDEN BURDEN GLOBAL related violence and ranging from 2 to 7 for overall homicide rates. See Maps 1.1–1.3 for the distribution of organized crime-related violence in Mexico (ICESI, n.d.; Mexico, n.d.a).30

Not surprisingly, the violence is concentrated where the drug cartels are most active. Four groups reportedly control most of the drug trade and other illicit activities in Mexico, with their influence reaching far into the United States. Several splinter groups and factions have also become central players in Mexico’s drug war; these are the Sinaloa cartel, the Gulf cartel, the Juárez cartel, and the Tijuana cartel, but also the Zetas, the Beltrán-Leyva organization, and the cartel Pacifico Sur, among others (Bunker, 2010, p. 11; Stratfor Global Intelligence, 2010).31 As Maps 1.1–1.3 illustrate, violence in Mexico is highly concentrated, although it also fluctuates and spreads. Thus states not affected by drug-related violence in one year can have very high figures the next. For example, Tamaulipas recorded a total of 90 drug-related deaths in 2009, yet one year later this figure climbed to 1,209.

A war on two fronts is taking place in the regions where the cartels are active. On the one hand, they are in violent confrontation with each other over the control of the lucrative trafficking routes between the Andean regions and the United States, as well as other forms of revenue generation such as extortion, Photo Mexican soldiers burn marijuana seized during a clash kidnapping, and human trafficking (Bunker, 2010, p. 11). On the other hand, with drug traffickers in Tijuana, Mexico, October 2010. the Mexican government has openly declared war upon drug traffickers, with © Francisco Vega/AFP Photo record specifically targeted violence against state state against violence targeted specifically record hand, other the on Aggressions, forces. security and enforcement law with confrontations violent as well as groups criminal between battles gun both register Confrontations deaths. drug-related of level alesser for only account hand, other the on aggressions, and Confrontations body). the with isleft amessage when example, (for cartel a drug of isamember perpetrator suspected or known the when violence drug-related as coded also are they but members, cartel drug at directed typically are events These body. the of mutilation and torture, of traces gunshots, multiple as such violence, of levels high involving killings intentional captures which ‘executions’, as iscategorized violence recorded of majority vast the that reveals figure The violence. of types by 2009 in tribution by disaggregating the drug-related deaths dis- this shows 1.5 Figure p.12). 2009, Intelligence, Global violencemore (Stratfor episodes of overt even to way given have power of readjustments and struggles resulting the divided, and weakened been have cartels drug the While results. ambiguous Legend: 2009 Mexico, violence, of type by killings crime-related organized Figure Source Executions (8,906; 93%) (8,906; Executions 6%) (596; Confrontations 1%) (111; Aggressions Mexico (n.d.a) : Mexico 1.5

Distribution of victims of of victims of Distribution XX

Map 1.1 Map 1.2 L California California California population Per 100,00 EGEND 40.93 L population Per 100,00 Baja Baja Baja EGEND <3 3–10 10–20 <3 3–10 10–20 20–30 >30 : California California :

Baja Baja Organized crime-related homicide rates, Mexico, 2007 Mexico, rates, homicide crime-related Organized Sur Sur Organized crime-related homicide rates, Mexico, 2008 Mexico, rates, homicide crime-related Organized Sonora Sonora Sinaloa Sinaloa Sin aloa Chihuahua Chihuahua Durango Durango Michoacán Michoacán Michoac Michoac Guerrero Guerrero án án 63.04 31 2 3 4 5 1 A UNIFIED APPROACH 32 Map 1.3 Organized crime-related homicide rates, Mexico, 2009 Zetas, a non-family-based group that originated with defectors from the military elite forces. First active as the Gulf cartel’s enforcement arm, the unit became increasingly independent, sub-contracting to other groups, until it broke its relations with the Baja Gulf cartel and became an independent and fully California Sonora established actor among the other cartels in Mexico

Chihuahua 98.63 (Killebrew and Bernal, 2010, p. 21).32

Baja California Reports on several carefully planned and executed Sur 39.93 SinaloaSinaloa raids requiring high levels of intelligence, resources,

ARMED VIOLENCE 2011 VIOLENCE ARMED Durango 43.47 and weaponry have appeared over the past few of LEGEND: years. These attacks are frequently accompanied by Per 100,00 unusual and gruesome displays of violence, regu- population larly including decapitations, mutilations, mass >30 20–30 MichoacánMichoacán executions, and extreme torture (Quinones, 2009; 10–20 Guerrero González, 2009). 3–10 <3 Politicians, law enforcement agents, civil society GLOBAL BURDEN BURDEN GLOBAL members, and their families and friends suffer the consequences of criminal violence. For example, Source for Maps 1.1–1.3: Mexico (n.d.a) Carlos Reyes López, a policeman of the state of authorities, law enforcement agents, and security forces by drug cartel members, Tabasco, was shot with ten members of his family resulting in the death of the targeted person (Mexico, n.d.b). in February 2009 (Wilkinson, 2009). In another case in Michoacán, the public safety minister, Minerva In response to the war on drug traffickers, cartels have implemented bolder Bautista, escaped death in an attack that killed four and more professionalized violent tactics. Over the past decade, drug-related of her colleagues and bodyguards in April 2010 violence has ‘acquired an increasingly organized and paramilitary character’, (Wilkinson, 2010). In April 2011, the Mexican poet making a ‘transition from the gangsterism of traditional narco hit men to para- Javier Sicilia lost his son, who was killed together military terrorism with guerrilla tactics’ (Turbiville, 2010, p. 124). Occasional with six of his friends in what seemed to be ‘collateral large-scale firefights, targeted assassinations, and military-style raids are damage’ of the drug war (Miglierini, 2011). common manifestations of the improved tactical competence and effectiveness of the drug cartels. These strategic improvements in the cartels’ capacities are Low incomes and the ready availability of money reportedly linked to the sweeping Mexican military reforms of the 1990s. In from drug cartels are important contributing factors response to low-intensity guerrilla activities and growing drug trafficking vio- to Mexico’s spiral of violence. Members of the state lence at that time, the Mexican state trained new elite commando units and or federal police may earn between USD 350 and assault troops, which were deployed in Chiapas, Guerrero, Puebla, and other USD 1,000 per month. Yet when 93 police officers regions of the country. A lack of careful recruitment, low salaries, and continuous were arrested in June 2009 for charges of corruption efforts by criminal organization to recruit security officers led to a substantial in the Mexican state of Hidalgo, some officers had number of defections by military personnel, who brought their skills and capaci- revenues reaching as high as USD 225,000 per ties to their new employers (Turbiville, 2010, pp. 128–32). An example is the month (Nagle, 2010, p. 100). Sources 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 Total homicide victims Figure clear distinctions between various types of vio- of types various between clear distinctions todraw impossible isnearly it involved, motives and actors violent different many the of view In p.56). 2010, Stepanova, 2008; (Ellingwood, conflict amajorarmed of andintensity scale the resembles that violence of inasurge resulted and officers, police federal 5,000 and anadditional soldiers 40,000 about of deployment the involved inMexico cartels drug against fight The 2006. inDecember traffickers drug on war declared and crime organized on down tocrack chose Calderón Felipe President violence, drug-related and corruption increased of challenges the Facing p.87). 2010, (UNODC, cartels drug Colombian the of andfragmentation dismantling the following position of Mexican drug cartels during the 1990s, the strengthened andMexico America Central towards routes trafficking drug of shift The reveals. Mexico of case the as however, zones, conflict to limited not are trafficking drug of quences conse- violent 2010). The Feldab-Brown, p. 209; 2007, (Cornell, andMyanmar Colombia, stan, inAfghani- conflict torecurring linked commonly andare ratio’ ‘value-to-size ahigh have Drugs Female victims victims Female 2003 PDH (2011); UNDP (2007) UNDP (2011); : PDH 1.6

Overall homicide and femicide levels in Guatemala, 2003–09 2003–09 inGuatemala, levels andfemicide homicide Overall Total victims Total 2004 2005 stood, but as Figure 1.6 shows, the increase increase the shows, 1.6 Figure as but stood, under- poorly are killings these of and causes nature p.79). The 2011, (PDH, in2003 killed edly report- women 383 the double isalmost figure that assaulted; orsexually tortured been having after many in2009, killed were women 720 an estimated example, for Guatemala, In pp.60–61). 2011, Bank, (World andRwanda Guatemala, Congo, the of Republic Democratic the as insuchplaces explored tobe beginning women—is against and violence violence orcriminal) (political organized between suchlinkage—that One violence. of forms other between mechanisms andtransmission interactions possible the highlight also violence and political criminal organized between intersections The bloodshed. the spared civilians innocent are nor immune, not are officials and government soldiers, officers, police killings, gang-on-gang be may deaths the of most While forces. armed andthe personnel military of deployment the through involved isdirectly it means, coercive over and monopoly authority challenge to its andtheir cartels drug of expansion the counter to attempts state Mexican the As inMexico. lence 2006 2007 2008 Totalfemale victims 2009 800 600 900 200 400 300 500 700 100 33 2 3 4 5 1 A UNIFIED APPROACH 34 appears to track the overall rise in criminal and Conclusion: a unified approach to gang-related violence in the country, which implies some connection between these different forms armed violence of violence (PDH, 2011).33 Different forms of armed violence interact in ways that go well beyond the simple dichotomy In Rwanda, an estimated 350,000 women were between political and criminal violence, or between raped during the genocide, translating to about conflict and non-conflict violence. As noted in 8,972 rapes per 100,000 women for the whole this chapter, the boundaries between violence adult female population; the risk of being raped categories are blurry and overlap, and they can for a Tutsi woman during this period was around reinforce each other in a vicious circle in multiple 80 per cent (Bijleveld, Morssinkhof, and Smeulers, ways. Violence can be ‘dual-purpose’, as high- 2009, p. 219). No comparable or reliable figures

ARMED VIOLENCE 2011 VIOLENCE ARMED lighted in the Iraqi and Somali examples; moreover, are available for the Democratic Republic of the of it can serve both ‘individual and organizational Congo, but the incidence of rape—as well as other goals’ (Green and Ward, 2009, p. 611). The con- forms of sexual and gender-based violence—is widespread (HHI, 2009, pp. 7–9). The prevalence cept of dual-purpose violence seems to apply to of rape is also anecdotally linked to the dynamics many settings in which large-scale acts of organ- of the conﬂict, including the exploitation of min- ized political violence coexist with individual and eral resources (Mukengere Mukwege and Nangini, criminal forms of violence, and where rape, looting, trafﬁcking, personal revenge, and other forms

GLOBAL BURDEN BURDEN GLOBAL 2009, p. 3). Although poorly understood, the linkages are important in policy terms. As Eriksson of opportunistic and criminal violence can be Baaz and Stern point out: observed.

the specific, often exclusive, focus on sexual Recognizing the multiple, simultaneous, linked, violence [. . .] hampers our understanding of the and changing forms that armed violence takes relationship between sexual violence and other is one step towards a unified approach to armed (supposedly) ‘ungendered’ violence . . . These violence prevention and reduction. An additional forms of violence are, to a large extent, mani- step is captured by the idea of a ‘system of vio- festations of the same systemic failures and lence’, in which high levels of violence and crime mechanisms as those contributing to [sexual in post- and non-conflict settings can be attrib- and gender-based violence] (Eriksson Baaz and uted to a series of intersecting factors present in Stern, 2010, p. 12). any given setting (Richani, 2007, p. 4 5). These These linkages—between organized political and factors include: organized criminal violence and between conflict low coercive and distributive state capaci- and sexual violence—highlight the ways in which ties, measured in terms of law enforcement armed violence can assume multiple, simultaneous, capacities (impunity and attrition rate) and and shifting forms that vary from place to place. Understanding how different forms of violence are the government’s social expenditures; linked is crucial to assessing the global burden low opportunity costs of crime derived from a of armed violence, and to developing effective lack of law enforcement and a lack of revenue- armed violence prevention and reduction strate- generating activities (such as education and gies and policies. job opportunities); 7 Figure opportu- new as catalysts—such specific in which context the provide acrime committing for costs opportunity andlow opportunities, economic weak enforcement, law for capacities state of lack the when occurs This together. come factors these when emerge can violence’ of A ‘system Source particular contingencies, such as the repatria- the suchas contingencies, particular Includes the unregulated availability and distribution distribution and availability unregulated the Includes (Richani, 2007, pp.4–5). (Richani, andAfghanistan inIraq suchas interventions orinternational andMexico, America Central to Caribbean the from routes trafficking drug of shift andthe America, toCentral States United the from immigrants illegal of tion of small arms and light weapons, mines, explosive explosive mines, weapons, light and arms small of remnants of war, and factors affecting their supply their affecting factors and war, of remnants OECD (2009, p. 50) p. (2009, : OECD 1.

The armed violence ‘lens’ violence armed The ntuet Agents Instruments informal (traditional and cultural) norms, rules, rules, norms, and cultural) (traditional informal Both formal institutions of governance and and governance of institutions formal Both societies affected by armed violence armed by affected societies Individuals, communities, and communities, Individuals, Institutions and practices People cycles of violence (World Bank, 2011). Bank, (World violence of cycles Report 2011 Bank’s World inthe lighted allhigh- are factors andcontingent underlying 2007, These p.5). (Richani, violence’ perpetuates and institutionalizes ‘that relationship dynamic a form which entrepreneurs, andpolitical crime), andorganized (gangs groups criminal sector, rity secu- violence’, the of suchas specialists the are units maininteracting ‘the inwhich emerge can violence of asystem 2002). suchasituation, In (Richani, spark the diamonds—provide such as goods’ ‘conflict exploitable oreasily trafficking, todrug linked generation income illicit for nities Perpetrators of armed violence and motivations for for motivations and violence armed of Perpetrators as playing an important role in repeated inrepeated role animportant playing as acquisition and misuse of arms and ofacquisition misuse arms (demand factors) World Development Development World National Regional Local Global 35 2 3 4 5 1 A UNIFIED APPROACH 36 Seeing different forms of armed violence as dual- This chapter has highlighted the way in which purpose, or as part of a broader system, opens the many different contemporary manifestations new opportunities for evidence-based research of armed violence blur the line between conflict and policy-making in various settings, ranging and non-conflict contexts. These various mani- from Jamaica to Afghanistan, Kenya to Haiti, and festations call out for a unified approach to Venezuela to Nepal. Piracy and warlordism amid measuring and monitoring armed violence that civil war and state collapse (Somalia and Afghan- captures the wide variety of actors, contexts, istan), drug-related violence and its political motivations, and consequences. The next chap- implications (Central America), shadow networks ter takes up this challenge by taking a broader in war economies (West Africa), as well as sexual perspective and presenting comprehensive and gender-based violence during conflict (Demo- national-level data that reflects how these

ARMED VIOLENCE 2011 VIOLENCE ARMED cratic Republic of the Congo) can all be better different forms of lethal violence are distributed

of understood by taking into account the complex around the world. It highlights not only that con- interactions between different forms of violence. ﬂict deaths are a relatively small part of the global burden of armed violence, but also that the One potentially useful framework for designing majority of the most violent places on earth are policies and programmes is the Organisation for not found in conﬂict zones. Subsequent chapters Economic Co-operation and Development’s armed focus on how the armed violence lens can be violence ‘lens’ (OECD, 2009, p. 50). The lens pro- employed to map global and regional differences

GLOBAL BURDEN BURDEN GLOBAL vides a flexible and unified framework for appre- in patterns of homicidal non-conflict violence hending the contexts, motives, and risk factors (CHARACTERISTICS OF ARMED VIOLENCE), on associated with armed violence (see Figure 1.7). the state of knowledge about violence against Its three legs provide different entry points for women (WHEN THE VICTIM IS A WOMAN), and armed violence prevention and reduction policies, on the negative links between armed violence focusing on the perpetrators of armed violence and development outcomes (MORE VIOLENCE, and their motives, the instruments of armed vio- LESS DEVELOPMENT). lence, and the wider institutional environment that enables or protects against armed violence. Although the lens is not in itself a policy or programmatic tool, it does highlight that a variety Abbreviations of factors can or do come together in different IISS International Institute for Strategic Studies situations to create an enabling environment in FARC Fuerzas Armadas Revolucionarias de Colombia which violence can occur and escalate, in both UCDP Uppsala Conflict Data Program conflict and non-conflict contexts. It also avoids UNITA União Nacional Para a Independência Total de debates about the specific categories or types of Angola armed violence; what counts, in the end, is to be able to assemble evidence regarding the various enabling factors of armed violence in order to design policies and interventions that effectively Endnotes address the serious challenges posed by armed 1 The definition also focuses on the physical use of force and violence, and deliberately excludes such concepts as violence to the safety and well-being of individuals structural, cultural, and psychological violence, however and communities. important they may be in other contexts. According totheUppsala Conflict Data Program, inthe 8 Countries such as Argentina, Colombia, and Costa Rica 7 oftheCommiss int Thereport ion ofInquiry 6 5 See the Thesetypes of violence, though by nomeans exhaustive, 4 According to von Clausewitz’s dictum, ‘war is thecon- 3 Thesefigures donot includetheglobal burdenof indirect 2 force between to anincompatibility, but must result fromuseof armed In territory). related totheoverarchin counted only if they result fromuseof armedforce directly case of state-based conflicts, battle-relat defence. See Table 2.4(TRENDS ANDPATTERNS). they arec offence. Nordoes theUnited States includethem,since by police duringconfrontations orcommitment of an such as France andNigeria, donot count civilians killed include extrajudicial andpolice (CIPEV, 2008,pp. 308–09). victims, of which 11werechildren and78werewomen Election Violence finds that the violence caused 1,133 (Geneva Declaration Secretariat, 2008,pp. viii–ix). for measuring thedifferent of direct andindirect conflict deaths andmethodologies for 2008(WHO, 2011). global burdenis estimated does not includeself-directed 6–7). For thepurposeof this research, ‘armed violence’ Moser andRodgers (2005,p. 5),and WHO (2002, pp. sources, includingHazen andHorner(2007, pp. 56–61), draw onthetypologies of violence published by several and Münkler (2003). armed violence, see Brzoska (2007, of the‘conflict’ lens as way toassess organized political Clausewitz, 197 tinuation of politics by other(i.e. violent) means’ (von AND PATTERNS). and extrajudicial killings, see Tables 2.3 and2.4(TRENDS homicide counts anddeaths dueto legaltions interven PATTERNS). For adetailed discussion onunintentional the humanitarian response; seeBox 2.6 (TRENDS AND status of theaffectedpopulation, andtherapidity of tion displacement, thebaseline health andnutrition conflict toconflict, dependingontheseverity of popula- conflict deaths, althoughthis rate varies widely from hovers around roughly facilities, food,andclean water dueto violence probably malnourishment, andageneral lack of access tohealth conflict deaths stemmingfrom conflict-related deaths. The estim Global Burden ofArmed Violence lassified as non-state conflict, thedeaths are warring factions 6). Ontheimplicatio ‘justifiable homic four times thenumberof direct g incompatibility (government or impacts of armedconflicts at 782,000 victims worldwide violence (suicide), whose (such as inMexico, where preventable disease, killings, whileothers, ated burdenof indirect app.2C),Ka ns andlimitations ide’, ed deaths are fordefinitions o thePost as isself- not linked ldor ( 1999), 1999), 3 In2008,2,831 homicides wereregistered by theInstituto 13 Eriksson Baaz and Stern (2009); Greenand Ward (2009); 12 Piracy is definedas ‘any illegal acts of violence ordeten- 11 In2010, Somali pirates perpetrated 148attempts and 10 9 The data 16 The report oftheCommiss int Thereport ion ofInquiry 16 These seven—called ‘internationalized intrastate conflicts’ 15 4 See,forexample, UNODC (n.d.a) andthesources forthe 14 605 victims of non-state armed violence (best estimates). For 2006,UCDPcounts 3,656 battle-related deaths and Count Project websites, res (IML, 200 a significant share of thesecases may begang-related for 1,910 cases in2008, maras, organg-related violence. It shouldbenotedthat in El Salvador. Of these cases, 354wereattributedto de Medicina Legal—the Instit Hume (2008); Sanchez et al.(2011); Wood (2003; 2006). private aircraft’ (Anyuand Moki, 2009, p. 95). ends by thecreworpassenge tion, orany act of depred (IMB, 2010;2011). representing 50percent of a 2009,January Somali pirates have perpetrated 430 attacks, amounted to169 attempts and48actual attacks. Since 65 successful attacks, whilein2009, thesenumbers Sundberg, 2010). version 1.3 (Eck andHultman, 2007; Eck, Kreutz,and Conflict Database version 2.3 andOne-sided Violence UCDP (n.d.); data examples herestemfromNon-state Luminoso inPeru). See thecodebooks foreach dataset at al-Qaeda attack of September 11thor victims of the Sendero formally organized groupagain use of armedforce by thegovernment of astate orby a one-sided violence registered between1993 and2008).Inthecategory of 751 deaths that wereduetointer-cartel warfare were 600,000 (CIPEV, 2008,p. 293; 2010, UNGA, p. 14). indicates 350,000 the figuresfordi are prone;meanwhile, IISS li under-counting towhich global incident systems reporting 258 fromnon-state violence), illustrating thedegreeof base only records 445deaths (187 fromone-sided violence; Election Violence lists 1,133 deaths, whilethe UCDPdata- ments involved, not thelocation of thefighting. Uganda, andtheUnited State —were Afghanistan, Algeri logical Annexe. direct conflict deaths database intheonline Methodo- was r 9, p. 70). etrieved formtheUCDP splaced persons,the wher , killings arerecorded incases of eas theUnitedNatio the motive ation, committed forprivate pectively (UCDP, n.d.; IBC,n.d.). a, Iraq, Rwanda, Somalia, sts 1,500 deaths. Regarding ll piracy attacks intheworld ute forForensic Medicine— s. Names reflect thegovern- st civilians (such as the rs of aprivate shipora was ‘unknown’; Commiss ion report and theIraqBody o thePost ns reports ns reports 37 2 3 4 5 1 A UNIFIED APPROACH 38 17 For data on identified state and non-state actors involved 26 On the early efforts to compile global data on crime, see in Kenya’s post-election violence, see the datasets on non- UNCJIN (n.d.). state conflict and on one-sided violence at UCDP (n.d.). 27 This point is explored in Norbert Elias’ thesis of ‘civilizing 18 These databases count 283 pro-government armed groups processes’, which links the general decline of lethal vio- (Carey and Mitchell, 2011) and anywhere from 500 to more lence to increasing degrees of ‘self-control’, based on than 900 non-state armed groups (IISS, 2009; UCDP, n.d.). individuals’ more detached and rational understanding 19 See data at HSRP (n.d.). of the world around them (Elias, 1994). 20 The report of the Commission of Inquiry into the Post 28 Estimates of the numbers have been rising significantly Election Violence lists 1,133 deaths, while the UCDP data- as well. In 2009, Felipe Calderón estimated that around base only records 445 deaths (187 from one-sided violence; 9,000 narco-linked murders occurred between 2006 and 258 from non-state violence), illustrating the degree of April 2009, while other estimates place the figure at around undercounting to which global incident reporting systems 11,000 people killed by early 2010 (Turbiville, 2010, p. 124). are prone; meanwhile, IISS lists 1,500 deaths. Regarding In April 2010, the Washington Post cited a confidential the figures for displaced persons, the Commission report report that estimates the human cost of drug-related

ARMED VIOLENCE 2011 VIOLENCE ARMED indicates 350,000 whereas the United Nations reports violence at 22,000 victims since 2006 (Booth, 2010). of 600,000 (CIPEV, 2008, p. 293; UNGA, 2010, p. 14). By early 2011, official government data counted 35,000 deaths, of which more than 15,000 occurred in 2010 21 On the complexity and multiplicity of links between alone. See Mexico (n.d.a) for the dataset on organized conflict and post-conflict violence, see Aguilera (2008), crime-related homicide victims. The Washington Post Cornell (2007), Restrepo and Tobón (2011), Jojarth (2009), subsequently indicated that Mexican news media counts Killebrew and Bernal (2010), Nitzschke and Studdard had surpassed the 40,000 threshold during the first half (2005), Rodgers and Muggah (2009), Small Arms Survey of 2011 (Washington Post, 2011). (2011), and Steenkamp (2009). 29 Drug-related violence is fluid and spreads easily. In 2010, 22 The homicide rate in England dropped from about 23.0 Chihuahua state exhibited the worst concentration of th th

GLOBAL BURDEN BURDEN GLOBAL per 100,000 in the 13 and 14 centuries to 4.3 per violence, with a rate of 129 per 100,000, whereas Sinaloa 100,000 by the end of the 17th, and 0.8 per 100,000 by followed in second place with a rate of 68 per 100,000. the first half of the 20th century. In the Netherlands and Belgium, contemporaneous figures are 47.0, 9.2, and 1.7 30 These numbers stem from two different datasets. One per 100,000; in Germany and Switzerland, rates fell from covers organized crime-related deaths (Mexico, n.d.a); 43.0 per 100,000 to below 2.0 for the 20th century (Eisner, the second reflects overall homicide deaths in Mexico 2001, pp. 618–38; Gurr, 1981, pp. 295–353; Monkkonen, (ICESI, n.d.). 2001, pp. 5–26). 31 The Sinaloa and Gulf cartels alone count an estimated 23 Trends in homicide rates can be considered in the light of 100,000 foot soldiers throughout the region, including the findings of LaFree and Drass (2002), who identify their enforcers or gang contractors (Bunker, 2010, p. 13). ‘crime booms’ between the late 1950s and the late 1990s 32 Increasing violence in Tamaulipas state, for example, is in more than one-third of their sample, which includes 21 linked to the battles opposing the Zetas to their former European nations. They define ‘crime booms’ as taking ally, the Gulf cartel. The Zetas are also known to have place in countries with (i) increasing homicide rates, expanded operations into Guatemala, forming alliances (ii) individual average annual growth in these rates of at with local gangs (Beittel, 2011, p. 10). least 10 per cent for any three successive years, and 33 See also Chapter Four on patterns and the extent of female (iii) sustained changes in the direction of the homicide victimization and resulting policy implications (WHEN THE trends (LaFree and Drass, 2002, pp. 780–81). By eliminat- VICTIM IS A WOMAN). ing countries that show some increase at some point, and by including reverse trends to evaluate crime booms, the authors still find that the period in question witnessed significant crime booms, though these were more pronounced in industrializing countries than in industrialized ones. 24 Recent improvements in data collection for violent crime have reduced the statistical errors that may be responsible for some previously unclear or fluctuating patterns in homicide trends. 25 See UNODC (n.d.b) for criminal justice data from 1970 to 2008. Carey, Sabine andNeil Mitchell. 2011. Bunker, 2010. Robert. ‘Strategic Threat: Narcos andNarcotics Brzoska, Michael. 2007. ‘Coll Booth, William. 2010. ‘Drug Violence Appears inMexico’s North- Bijleveld, Catrien, Aafke Morssink Beittel, June. 2011.Mexico’s Drug Trafficking Organizations: Source Bateson, Regina. 2009. ‘ThePolitical Consequences of Crime Balakrishnan, Angela, Xan Rice, andRichard Norton-Taylor. Archer, Gartner. DaneandRosemary Anyu, J.Ndumbeand Samuel Moki. 2009. ‘Africa: The Piracy Hot- AI (Amnesty International). 2009. Aguilera, Gabriel.2008. ‘Enfrentar la violencia con “mano dura”: Aeby, Marcelo. 2004. ‘Crime Trend Abbott, Edith. Bibliography Militias? Working paper. June. March, pp. 8–29. Overview.’ University Press, pp. 94–106. andInternational SecurityDisarmament, Peace Research Institute. Definition of ArmedConflict.’ In Stockholm International article/2010/04/20/AR2010042004961.html> Rwandan Genocide.’ 2009. ‘Counting theCountless: Rape Victimization during Washington, DC:Congressional Research Service. January. and Scope of the RisingViolence. CRS Report forCongress. vanderbilt.edu/files/kyHPZ6/Bateson_CPW_April_14.pdf> Comparative Politics Workshop. 14April. Guardian. 19 November.

Cornell, Svantee. 2007. ‘Narcotics andArmedConflict: Interaction Coghlan, Benjamin, et al.2006. ‘Mortality intheDemocratic von Clausewitz, CIPEV (Commission of Psychogenetic Investigations Criminology, cide Rates in Theoretical Perspective.’ Violence: The Long-Term Dynamics of European Homi- and Practitioners.’ NewOrleans, 17 February. annual meetingof ‘Theory vs. Policy? Connecting Scholars UCDP Non-State Conflict Dataset.’ Paper presentedat the of Peace Research Civilians in War: Insights fromNewFatality Data.’ Armed Violence. Geneva: Geneva Declaration Secretariat. RWP07-019. Cambridge, Mass.: University. Harvard April. Dominican Republic of CrimeandViolence: Les Contribution ofCriminalJustice Systems to the Control Washington, DC:FDDPress, pp. 90–112. Pakistan Theater: Militan Gartenstein-Ross andClifford May, eds. Asia: Beyond Taliban Profits andAfghanistan.’ InDaveed and theNordic Africa Institute. Working Paper onGender-based Violence. Uppsala: SIDA Violence in the Democratic Republic ofCongo (DRC) Vol. 53, No.2,pp. 495–518. Forces intheCongo (DRC).’ Rape? Masculinity, Violence, and Sexuality intheArmed latimes.com/2008/jun/0 for Control.’ Carl. h, pp. 207–27. Los Angeles Times Vol. 41, No.4,pp. 618–38. 1976. Inquiry intothePost Inquiry . Nairobi:CIPEV. , Vol. 44,No.2,pp. 233–46. . KSG Faculty Research Paper Series, On War. Translated by Michael Howard Sundberg. 2010. ‘Introducing the t Islam, Security andStability 3/world/fg-border3> Stern. 2009. ‘Why Do Soldiers sons from Jamaica and the International Studies Quarterly Princeton University Press. . Oxford: Blackwell Publishers. . 3June. The New York Times. 1 September. 31 December 2009. London: International Chamber of Gilgen, Elisabeth and Lauren Tracey. 2011. Armed Violence Moni- Commerce. January. toring Systems: Contributing Evidence to Programming. —. 2011. Piracy and Armed Robbery against Ships: Annual Report, Geneva: Geneva Declaration Secretariat. 1 January–31 December 2010. London: International González, Franciso. 2009. ‘Mexico’s Drug Wars Get Brutal.’ Chamber of Commerce. January. Current History, Vol. 108, No. 715, pp. 72–76. IML (Instituto de Medicina Legal). 2009. Epidemiología de los Goodhand, Jonathan. 2008. ‘Corrupting or Consolidating the homicidios en El Salvador: Período 2001–2008. San Peace? The Drugs Economy and Post-conflict Peacebuilding Salvador: Corte Suprema de Justicia. October. in Afghanistan.’ International Peacekeeping, Vol. 13, No. 3, IRC (International Rescue Committee). 2007. Mortality in the pp. 405–23. Democratic Republic of Congo: An Ongoing Crisis. Green, Penny and Tony Ward. 2009. ‘The Transformation of Vio-

ARMED VIOLENCE 2011 VIOLENCE ARMED lence in Iraq.’ British Journal of Criminology, Vol. 49, No. 5.

of September, pp. 609–27. Jojarth, Christine. 2009. Crime, War, and Global Trafﬁcking. Gurr, Ted Robert. 1981. ‘Historical Trends in Violent Crime: Cambridge: Cambridge University Press. A Critical Review of the Evidence.’ Crime and Justice, Kaldor, Mary. 1999. New Wars and Old Wars: Organized Violence Vol. 3, pp. 295–353. in a Global Era. Stanford: Stanford University Press. Hazen, Jennifer and Jonas Horner. 2007. Small Arms, Armed Kalyvas, Stathis. 2003. ‘The Ontology of “Political Violence”: Violence, and Insecurity in Nigeria: The Niger Delta in Action and Identity in Civil Wars.’ Perspectives on Politics, Perspective. Occasional Paper 20. Geneva: Small Arms Vol. 1, No. 3. August, pp. 475–94. Survey. Kemp, Walter. 2003. ‘Breaking the Crime–Conﬂict Nexus: A Chal-

GLOBAL BURDEN BURDEN GLOBAL HHI (Harvard Humanitarian Initiative). 2009. Characterizing lenge for the OSCE.’ In Institute for Peace Research and Sexual Violence in the Democratic Republic of the Congo: Security Policy at the University of Hamburg, ed. OSCE Profiles of Violence, Community Responses, and Implica- Yearbook 2003. Baden-Baden: Nomos Verlagsgesellshaft, tions for the Protection of Women. Final Report for the pp. 301–18. Open Society Institute. Cambridge, Mass.: HHI. August. Killebrew, Bob and Jennifer Bernal. 2010. Crime Wars: Gangs, Höges, Clemens, Uwe Klussmann, and Horand Knaup. 2008. Cartels, and U.S. National Security. Washington, DC: ‘Pirates versus Weapons: Dealers Looking for the Good Center for a New American Security. Guys off the Somali Coast.’ 10 June. in the Angolan UNITA.’ Contemporary Security Policy, HSRP (Human Security Report Project). 2010. Human Security Vol. 30, No. 2, pp. 343–55. Report 2009/2010: The Causes of Peace and the Shrinking Lacina, Bethany and Nils Petter Gleditsch. 2005. ‘Monitoring Costs of War. Pre-publication text. Vancouver: HSRP. Trends in Global Combat: A New Dataset of Battle Deaths.’ —. n.d. ‘Graphs and Tables: Report 2009/2010.’ European Journal of Population, Vol. 21, Nos. 2–3, LaFree, Gary, and Kriss Drass. 2002. ‘Counting Crime Booms Hume, Mo. 2008. ‘The Myths of Violence: Gender, Conflict, and among Nations: Evidence for Homicide Victimization Community in El Salvador.’ Latin American Perspectives, Rates.’ Criminology, Vol. 40, No. 4, pp. 769–800. Vol. 35. No. 5, pp. 59–76. Lennox, Patrick. 2008. Contemporary Piracy off the Horn of IBC (Iraq Body Count). n.d. ‘Iraq Body Count.’ Website. Africa. Calgary: Canadian Defense and Foreign Affairs Institute. ICESI (Instituto ciudadano de estudios sobre la inseguridad— Malaquias, Assis. 2010. ‘UNITA’s Insurgency Lifecycle in Angola.’ Citizen Institute for Studies on Insecurity). n.d. ‘Estadísti- In Klejda Mulaj, ed. Violent Non-State Armed Actors in cas oficiales.’ Mexico. n.d.a. ‘Base de datos de fallecimientos ocurridos por IISS (International Institute for Strategic Studies). 2009. The presunta rivalidad delincuencial.’ Mexico City: Presiden- Military Balance 2009: The Annual Assessment of Global cia de la República. Accessed 23 June 2011. Riedel, Marc and Wayne Welsh. 2008. —. 2007. ‘Systems of Violence andtheirPolitical Economy inPost- Richani, Nazih. 2002. Systems ofViolence: ThePolitical Economy Quinones, Sam. 2009. ‘State of War.’ Pettersson, Therése andLotta Themnér, eds.2010. PDH (Procurador delos Derechos Humanos). 2011. OECD (Organisation forEconomic Co-operation andDevelopment). Nitzschke, Heiko andKaysie Studdard. 2005.‘TheLegacies of Nagle, Luz.2010. ‘Corruption of Po Monkkonen, Eric.2001.‘New Standards forHistorical Homicide Miglierini, Julian.2011.‘Mexico Poet Javier Sicilia Leads Anger at Mexico. n.d.b. ‘Base dedatos dehomicidios presuntamente Münkler, 2003. Herfried. ‘The Wars of the21 Mukengere Mukwege, Denis andCathy Nangini.2009. ‘Rape Mueller, Susanne. 2008. ‘ThePolitical Economy of Kenya’s Moser, Caroline andDennis Rodg versity Press. Causes, andPrevention Conflict Situations.’ Unpublished paper. March. of New York Press. of War andPeace inColombia. New York: State University Research, Uppsala University. Conflict 2009. Uppsala: of Department Peace andConflict Derechos Humanos. Humanos enGuatemala Circunstanciado—Tomo I: Situación delos Derechos Enabling Development.Paris: OECD. 2009. Conflict andFragility—Armed Violence Reduction: No. 2,pp. 222–39. and Peacebuilding.’ War Economies: Challenges andOptions forPeacemaking pp. 95–122. Border.’ Small Wars andInsurgencies and theJudiciary inMexico andComplicity Across the tional Review of the Red Cross Research.’ uk/news/world-latin-america-13141263> Drug Violence.’ BBC News.22April. 2011. tute, SIPRIY Violence.’ In Stockholm Internation London: I.B. Tauria &Co. Ltd. tion: Managing Insecurity in the Aftermath ofPeace Accords Polity Press. in Europe from the Middle Ages to the Present Arms Survey.’ Liberia. LiberiaArmed Violence Assessment. Geneva: Small Armed Violence Assessment. Geneva: Small Arms Survey. and Guns.Cambridge: Cambridge University Press. Honour Killings Rise.’ Guardian. 30November. Rape andOther Violence: Leave My Body Out of War. Colombian Armed Conflict, 2001–2009 of Sexual Violence against Women in the Context of the Small Arms Survey. Costs, andInterventions rary Security Policy, Vol. 30, No.2.August,pp. 301–17. Armed Groups: The Central American Case.’ tion Secretariat. and Geneva: Editorial Panamericana andGeneva Declara- cijada: Panorama deuna violencia transformada pp. 123–44. Small Wars andInsurgencies Tactical Forms of Cartel Violence and Their Underpinnings.’ pp. 37–60. International Security Vol. 7, No. 5,pp. 375–94. Decline inCrime Rates.’ 20100802_mexico_security_memo_aug_2_2010> 20 December. atfor.com/an Guatemala enlaencru- al Peace Research Insti- . Bogotá: Campaign can_drug_cartels_ Contempo- ations: sassinations: alysis/ . Cambridge: . Bogotá Liberia Liberia . ,

41 2 3 4 5 1 A UNIFIED APPROACH 42 UCDP (Uppsala Conflict Data Program). 2011. ‘UCDP Actor Data- Wennmann, Achim. 2007. ‘The Political Economy of Conflict set v. 2.1-2010.’ Uppsala: Uppsala University. Financing: A Comprehensive Approach Beyond Natural WHO (World Health Organization). 2002. World Report on Vio- —. n.d. ‘UCDP Data.’ lence and Health. Geneva: WHO. —. 2011. ‘Deaths Estimates for 2008 by Cause for WHO Member UN (United Nations). 2010. Second Generation Disarmament, States: By Sex.’ Accessed June 2011. Demobilization, and Reintegration (DDR) Practices in Peace on Challenges and Opportunities for UN Peacekeeping. Wilkinson, Tracy. 2009. ‘Police Officer, Ten Relatives Killed in New York: Department of Peacekeeping Operations, Attacks in Tabasco.’ Los Angeles Times. 16 February. United Nations. Human Rights Report: 1 January–30 June 2008. ARMED VIOLENCE 2011 VIOLENCE ARMED —. 2010. ‘Mexican State Security Minister Can’t Trust Her Own

of Branch). n.d. ‘Data Set of the 1st and 2nd United Nations Wood, Elisabeth Jean. 2003. Insurgent Collective Action and Civil World Crime Surveys.’ Vienna: United Nations Office in War in El Salvador. New York: Cambridge University Press. Vienna. —. 2006. ‘Variation of Sexual Violence During War.’ Politics & UNDP (United Nations Development Programme). 2007. Informe Society, Vol. 34, No. 3, pp. 307–41. estadístico sobre violencia en Guatemala. Guatemala World Bank. 2011. World Development Report 2011: Conflict, City: UNDP. GLOBAL BURDEN BURDEN GLOBAL Security, and Development. Washington, DC: World Bank. UNGA (United Nations General Assembly). 2010. National Report —. n.d. ‘Indicators: GDP Growth (Annual %).’ Accessed on Submitted in Accordance with Paragraph 15 (a) of the 24 May 2011. Annex to Human Rights Council Resolution 5/1: Kenya. A/HRC/WG.6/8/KEN/1 of 22 February. Zimring, Franklin. 2007. The Great American Crime Decline. UNODC (United Nations Office on Drugs and Crime). 2010. The New York: Oxford University Press. Globalization of Crime: A Transnational Crime Threat Assessment. Vienna: UNODC. June. —. n.d.a. ‘UNODC Homicide Statistics.’ —. n.d.b. ‘United Nations Surveys on Crime Trends and the Opera- tions of Criminal Justice Systems (CTS).’ USDoD (United States Department of Defense). 2010. The Chal- lenge and the Promise: Strengthening the Force, Preventing Suicide and Saving Lives. Final Report of the Department of Defense Task Force on the Prevention of Suicide by Members of the Armed Forces. Washington, DC: USDoD. USDoS (United States Department of State). 2007. Iraq. Washington, DC: Bureau of Democracy, Human Rights and Labor, USDoS. Washington Post. 2011. ‘Debate Mounts over Death Toll in Mexico’s Drug War, as Government Refuses to Release Figures.’ 2 June. and rural communities seized by armed violence. armed by seized communities and rural towns, cities, inmost supply inshort a service assistance, medical of andquality availability the on depend often andrecovery Survival orminutes. centimetres mere of a question by determined be may outcome the victim, the T and ‘intentional homicides’ arising in non-conflict innon-conflict arising homicides’ and ‘intentional wars during occur that deaths’ ‘conflict between distinguish typically analysts violence, lethal andrecording to documenting comes it When violence. orsexual abuse, assault, of forms many than death intentional documenting on apremium of more place traditions and oral dispatches, media registries, church records, legal systems, registration Vital crimes. other than systematically more recorded tobe tends sinceit insecurity, of index animportant is also killing perspective, anadministrative From ment. and measure- toexamination amenable readily more it renders which inallsocieties, seriously istreated Killing victimization. of forms other of andscale scope the for measure approximate an as serve can violence lethal violence, armed of anddynamics characteristics the shape factors many although But violence. armed of andconsequences intensity the of indicator one only istherefore deaths of number The wounds. logical andpsycho- physical prolonged or experience injured are more many killed, eachperson for And or long-lasting pain to minor injury. For For injury. tominor pain or long-lasting disability topermanent death from range of he consequences Trends and Patterns of Lethal Violence Chapter

Two armed violence violence armed

years 2004–09, this chapter findsthat: chapter this 2004–09, years the covers which violence, lethal on database 2011 GBAV comprehensive the on Drawing levels. andnational regional the at trends of reading comprehensive amore andfacilitates violence armed of burden global the of portrait anoverall for allows anapproach Such total. asingle generate to violence lethal on information combines it data, disaggregated on draws it Although all settings. across violence lethal on data available lapses col- chapter this categories, two the between distinction artificial the retaining of Instead meaningless. ‘non-conflict’ and ‘conflict’ between distinction simple binary a renders intensity varying of conflicts armed with violence criminal organized of merging the andNigeria, andYemen toMexico Afghanistan from ranging countries In orneither. both of result the be may objectives oreconomic political by motivated tobe believed are that Killings APPROACH). UNIFIED (A motivations or economic ortopolitical violence, orinterpersonal organized to exclusively attributed be can adeath whether todetermine contexts andpost-conflict fragile in difficult isoften it however, One, in Chapter shown As p.3). 2004, andHoeffler, (Collier settings At least 526,000 people are killed each year each year killed are people 526,000 least At 21,000 killings during legal interventions. legal during killings 21,000 and homicides, ‘unintentional’ so-called 54,000 homicides, intentional 396,000 deaths, conflict direct 55,000 an estimated includes This violence. lethal of aresult as 2 43 2 3 4 5 1 TRENDS AND PATTERNS 44 One in every ten of all reported violent deaths The rates of both intentional homicide and around the world occurs in so-called conflict direct conflict deaths are volatile. In 2006, the settings or during terrorist activities. number of victims of intentional homicides The average annual global violent death rate dropped to around 368,000, while in 2009 between 2004 and 2009 was 7.9 per 100,000. this figure increased to around 423,000. After a dramatic increase in direct conflict deaths At least 58 countries exhibit violent death between 2005 and 2007—roughly 66,000 rates above 10.0 per 100,000. These countries people died directly in armed conflict in account for almost two-thirds of all violent 2007—the figure dropped to roughly 58,000 deaths (63 per cent) or 285,000 deaths. in 2009. El Salvador was the country most affected by

ARMED VIOLENCE 2011 VIOLENCE ARMED lethal violence in 2004–09, followed by Iraq of and Jamaica. Middle and Southern Africa, Central America Explaining the framework and the Caribbean, and South America are Whether focusing on the global or the national the regions that exhibit highest levels of level, a comprehensive estimate of lethal vio- lethal violence. lence necessarily relies on multiple data sources.

Figure 2.1 How violence is reported and recorded GLOBAL BURDEN BURDEN GLOBAL

CONFLICT DEATHS Victims of extrajudicial killings NON-CONFLICT DEATHS

CIVILIAN Battle-related deaths CONFLICT Civilian deaths DEATHS Victims of terrorism INTENTIONAL HOMICIDES

Victims of legal interventions

UNINTENTIONAL HOMICIDES

Victims of legal Intentional homicides: Direct conflict deaths: interventions: violent Unintentional homicides: deaths as a result of inter- Note: deaths as a result of deaths of civilians by law The circles only deaths as a result of personal violence, gang approximately reflect the armed conflicts, political enforcement and state ‘accidental’ killings— violence, economically proportions of violent violence, and terrorism— security forces during 54,000 deaths per year motivated crime—396,000 deaths in each category. 55,000 deaths per year legal interventions— deaths per year 21,000 deaths per year Source: GBAV 2011 database governmental organizations, and the media are are media andthe organizations, governmental non- institutions, academic authorities, national bodies, international exceptions, some With data. adequate given reported, often cide are homi- andintentional deaths conflict direct how demonstrates It violence. armed of burden global the estimating for aroadmap 2.1 provides Figure much messier. course isof violence lethal of reality the recorded, and reported are acts violent inwhich way the acknowledges figure the While exclusive. ally mutu- tobe categories the assume often which organizations, andnon-governmental ernments gov- by done routinely as categories, analytical particular into compartmentalized deaths violent shows figure the Nevertheless, tation. interpre- of amatter isfrequently categories two the between distinguishing that fact the as well as dichotomy inthis inherent complexity the betrays deaths andnon-conflict conflict between differentiation sharp The limitations. afew from suffers framework the violence, lethal of aspects some clarifies it While literatures. studies flict andcon- sciences, health justice, criminal the in categories established within and reported isrecorded violence lethal how 2.1 outlines Figure violence. lethal of estimate integrated and acomprehensive develop inorder violence lethal of forms different between distinguish that sources data with works chapter this approach, anintegrated achieving of challenges the lights high- One Chapter While violence. lethal standing under- for framework aunified introduces thus The lence. vio- lethal of burden overall the of image partial a gather only frequently they eachanother, from inisolation work typically researchers the Because studies. andconflict epidemiology, criminology, including disciplines, many from gathered deaths on information statistical reconciling requires It Global Burden of Armed Violence 2011 Violence Armed of Burden Global

thus facilitating cross-national comparisons. cross-national facilitating thus countries, different across shared generally are categories broad inboth orexcluded included are events of kinds what of determinations The andspace. time across phenomena two these andreporting defining at adept reasonably killings during legal interventions. legal during killings and homicides unintentional on data gathering in involved challenges the of analysis detailed a offers chapter this of finalsection The caution. with treated tobe have figures these ventions, inter- legal during occurring anddeaths homicide unintentional on data of reliability low and the reporting, of quality poor the definitions, parable com- of absence the Given violence. lethal of count overall the into incorporated are they although volume, inthis indetail analysed not are they aresult, As statistics. inhomicide included ally gener- not are interventions legal during killings and homicides unintentional hand, other the On 2.1). Box see pp.72–73; 2006, (Friedrichs, contexts cultural and social, historical, indifferent andgroups tries coun- of range toawide applied been has concept The entails. actually terrorism what of definition agreed auniversally of absence the of aresult as partly arise andinjuries deaths terrorism-related incounting Challenges zones. inconflict terrorism of includevictims already deaths conflict direct on databases most andbecause war ing from oremerg- by affected incountries occur them of sincemost category death’ conflict ‘direct in the included are deaths terrorism-related example, For together. merged are death violent of categories certain how demonstrates also framework The (suicide).violence self-directed cover it does nor conflict, armed of aconsequence as populations civilian on inflicted mortality orexcess deaths (non-violent) indirect include not does chapter inthis analysis The 3

45 2 3 4 5 1 TRENDS AND PATTERNS 46 Box 2.1 Challenges in defining ‘terrorism’ and recording its victims Defining what is terrorism and who is a terrorist is a delicate matter. For more than two millennia, the term ‘terrorist’ has been applied to a disparate assortment of groups—including the Sicarii in first-century Palestine, the Assassins in the 13th-century Middle East, and the so-called ‘Thugs’, who were active from the 7th to the 19th century in India, and enjoyed a heyday in the 13th century (Rapoport, 1984). More recently the label ‘terrorism’ was applied to the actions of the French ARMED VIOLENCE 2011 VIOLENCE ARMED revolutionaries at the end of the 18th century of and to Russian and European anarchists during the 19th century. Likewise, a wide range of actors have assigned the label ‘terrorist’ to the activities of totalitarian states, such as Nazi Germany and the Soviet Union, anti-colonialist and leftist groups of the 1960s and 1970s, and extremist religious and environmental groups since the 1970s (Rapoport, 2002). GLOBAL BURDEN BURDEN GLOBAL There is no international consensus on how ‘terrorism’ should be defined. In the 1970s the UN General Assembly’s Ad Hoc Committee on International Terrorism failed to arrive at a common definition, only to abandon its efforts in 1978 (UNGA, 1972). The goal of arriving at a universal definition was revived in 1996 with a new Ad Hoc Committee established by the General Assembly (UNGA, 1996).4 Despite a major effort to draft a comprehensive convention on international terrorism, the Committee has also failed to issue a definition or set of parameters that satisfy all members, and its draft definition remains provisional (UNGA, 2010a).

Despite the absence of an agreed deﬁnition, a number of databases provide annual estimates of terrorist victims, which vary widely in their estimates (see Table 2.1). In this context, the US-based National Counterterrorism Center (NCTC) reports are the most comprehensive as they speciﬁcally focus on terrorism rather than other kinds of political violence. Yet the NCTC also counts ‘military personnel and assets outside war zones and war-like settings’ as terrorism victims (NCTC, 2010, p. 4); this approach is

Photo S Forensic experts investigate a bomb-destroyed bus following a coordinated attack on London’s public transport system, July 2005. © Kai Pfaffenbach/Reuters the battle-death dataset of the Uppsala Conflict Data Data Conflict Uppsala the of dataset battle-death the Likewise, network. al-Qaeda the against forces Coalition its and States United the between conflict armed the of aresult as occurring ‘battle-deaths’ contains it instead, however; terrorism, of victims all record not does base data- This terrorism/al-Qaeda’. ‘international on category a maintains Database Conflict Armed (IISS) Studies Strategic for Institute International London-based The and inclusion data exclusion. for criteria different apply which of all and 1970s the to back dates which of one databases, different three of up ismade it since sistency, incon- data from suffers also GTD the Moreover, attacks. terrorist of victims civilian and military counts GTD the because increases also over-counting of risk the NCTC, the with case isthe As insurgencies. to respect with of violence either, and otherforms including terrorism between distinguish always not does Maryland, of versity Uni- the of out run (GTD), Database Terrorism Global The They write: Iraq. of case the in reports 2009 and 2008 NCTC the of authors the by acknowledged been has point This 2011). (Flükiger, them kill or injure may terrorist as nated desig- organizations though even terrorism, of victims as excluded be should soldiers and combatants rorism, ter- of characteristic adefining as non-combatants and civilians of targeting the view who analysts many the to According zones. war outside personnel military ing by includ- victims over-counting sinceproblematic it risks Sources Table GTD IIS NCTC UCDP one-sided violence S international terrorism/al-Qaeda 2009, p. 4; 2010, p.5). 2010, p.4; 2009, (NCTC, insurgency former-regime-elements the as well as networks terrorist Sunni the both in participated Iraqis as years, previous in challenging especially was Iraq in gency insur- and terrorism between distinction The 2.1 GTD (n.d.); IISS (n.d.); NCTC (n.d.); UCDP (n.d.b) UCDP (n.d.); NCTC (n.d.); IISS (n.d.); : GTD

Reported number of victims of terrorism or one-sided violence orone-sided terrorism of victims of number Reported 2004–09 were attacked in a ‘main armed conflict’ in a ‘main armed attacked were 2004–09 period the for NCTC by reported terrorism of victims all of cent per 98.2 example, For settings. conflict in killed are terrorism of casualties of majority vast the that finds datasets violence one-sided UCDP and NCTC, GTD, the of A review in 2007–09. in Mauritania Maghreb Islamic the in al-Qaeda of attacks the as such incidents lethal—terrorist small-scale—but account into take not does violence one-sided on dataset UCDP the instance, For victims. afew only claim they if especially deaths, flict con- direct as defined been not have and ‘terrorism’ as count could incidents violence armed of Amultitude ayear’. in 25 deaths least ‘at claim that conflicts includes only it Second, victims). (52 bombings London 2005 the of victims the count not does but (191 victims) bombings Madrid 2004 the includes it since however, consistent, entirely isnot dataset p.2). The 2008, (Kreutz, excluded’ are custody in killings Extrajudicial 25 deaths. least at in results which civilians against group organized aformally by or astate of government the by force armed of use ‘the as violence one-sided defines It dataset. violence’ sided ‘one- UCDP isthe terrorism of victims capture may that dataset additional An IISS. by defined as conflict active in engaged not is that acountry in occurs act aterrorist if only terrorism international of egory cat- the in exclusively counted are victims These occur. they which in countries the in deaths battle ‘regular’ as counted also are IISS by listed terrorism’ tional ‘interna- of victims the that note to isimportant It Yemen. and Arabia, Saudi Pakistan, Afghanistan, as such countries in deaths records dataset IISS The Source: Flükiger (2011) Flükiger Source: 60 people killed in Amman in 2005 (Povey et al., 2009, p.10). 2009, al., et (Povey 2005 in Amman in killed people 60 the and 2005; in el-Sheikh Sharm in victims 88 the 2004; in bombings Madrid the in killed people 191 the example, for include, UCDP. and They NCTC, GTD, by provided information the of areview of basis the on included are terrorism of victims conflicts, main armed Outside in 2011 GBAV the deaths’ dataset. ‘battle have not added to been regular the datasets three inlisted these victims terrorism counting, double To avoid Pakistan. or Iraq, Afghanistan, in does not use the term ‘international terrorism’ (UDCP, n.d.c). (UDCP, terrorism’ ‘international term the use not does it although conflict, armed this of aresult as deaths records (UCDP) Program 0420 0620 082009 2008 2007 2006 2005 2004 ,2 ,0 ,8 260873n/a 8,753 12,620 8,988 6,007 5,520 7,826 ,7 3892,5 2791,0 15,311 15,709 22,719 20,453 13,889 7,474 5 2 0 ,6 ,4 n/a 1,243 1,760 709 620 550 ,1 ,9 ,6 ,6 n/a 3,264 4,161 4,096 3,919 7 such as in the armed conflicts conflicts armed the in as such 6

47 2 3 4 5 1 TRENDS AND PATTERNS 48 Explaining the data sources tries legally require that every death be certified and registered by the public health authorities. The Global Burden of Armed Violence 2011 relies The original data for such death certificates is on multiple data sources to measure the scale typically sourced from hospitals, health clinics, and magnitude of lethal violence. Most of the emergency rooms, mortuaries, or autopsy reports data is derived from incident reporting systems. Incident reporting encompasses passive surveil- of forensic institutes. In the best case, data is lance of the number of people reported to have integrated into a national vital registration system died in violent events through hospital, mortu- that codes the causes of deaths according to the ary, police, or criminal justice data collection.8 International Classification of Disease (ICD), cur- Figure 2.2 provides a graphic illustration of a rently in its tenth revision (WHO, n.d.a). At the common incident reporting data collection pro- international level, national data is aggregated ARMED VIOLENCE 2011 VIOLENCE ARMED through systems such as the World Health Orga- of cess, which typically results in three different types of databases: one for criminal justice sta- nization’s Mortality Database (WHOMDB), the tistics, one for public health data, and another single largest dataset on causes of death reported focusing on direct conflict deaths. by national vital registration systems.

The most reliable incident reporting mechanisms The criminal justice system is another signiﬁcant are frequently those connected to a country’s source of data on violent deaths in non-conﬂict public health system. This is because most coun- settings, since these typically constitute the GLOBAL BURDEN BURDEN GLOBAL

Figure 2.2 Incident reporting of violent events leading to death

Reporting of event Police Criminal justice system to police investigations (intent of killing)

Body in hospital Forensic medicine National criminal UNODC Crime Trend Death event or mortuary (autopsy report) justice statistics Survey (CTS)

Monitoring through media, governments, Death certiﬁcate Vital registration NGOs, and interna- (cause of death) system tional organizations

Databases on direct National public WHO Mortality conflict deaths health statistics Database (WHOMDB) East Asia, and South-east Asia, the vital registra- vital the Asia, andSouth-east Asia, East Asia, South Africa, Sub-Saharan of including most regions, inseveral But Caribbean. andthe America inLatin notably regions, andmiddle-income low- andseveral regions inallhigh-income available is data registration vital and comprehensive Sophisticated world. the around widely vary tems sys- reporting incident of andcoverage quality The n.d.). (UCDP, IBC, Count n.d.a; Body andIraq UCDP by together put databases different the ing are report- incident of make use that databases death conflict of examples Prominent reporting. media and NGO, humanrights, health, from drawn often deaths, conflict on data of collection the to dedicated iscurrently institutions research andindependent academic of cadre a growing 2011). Likewise, (Muggah, soldiers own their on especially conflict, armed of effects lethal the understanding with preoccupied been long have strategists anddefence military Indeed, century. 20th the middleof sincethe recognized been has lacuna data This altogether. orabsent equate, inad- unreliable, isoften data justice criminal and health public settings, conflict-affected In andquality. availability intheir ments improve- significant seen have years few last the orpoliticization, coverage partial totheir owing scepticism with approached routinely are tistics sta- homicide intentional while And process. this of out emerge usually homicides unintentional and intentional on Statistics killing. the of intent the investigate system justice criminal and the police the reports, autopsy the from information Usingforensic police. the by reported are cides homi- settings, p.68).most In 2008, Secretariat, Declaration (Geneva person’ another by son aper- on inflicted death an‘unlawful as defined be chapter,can ahomicide this of purposes the illegal killings of persons by other persons. For For persons. other by persons of killings lence) for the years 2004 and 2008 (WHO, n.d.b). (WHO, and2008 2004 years the lence) for vio- (collective and‘war’ (interpersonal) ‘violence’ on estimates country-level provides It patterns. tical models to estimate broad cause-of-death statis- developed WHO gaps, data chronic these for To compensate data. comparable reliable, toprovide weak too issimply infrastructure tion It is important to stress that that to stress isimportant It (UNODC). andCrime Drugs on Office Nations Systems Trends ofCriminal Justice andOperations Crime of Nations United current the suchas tives initia- gathering collection data crime national cross- of aresult as isavailable data of type This VIOLENCE). ARMED OF (CHARACTERISTICS analysis afine-tuned andgenerate sources totriangulate possible isnevertheless it differ, often a‘homicide’ constitutes what of nitions p.10).defi- 2010, While andMalby, Heiskanen, (Harrendorf, world inthe countries most able for isavail- data homicide intentional comparable Africa, inSub-Saharan countries of a number of exception notable the with Indeed, countries. across comparisons facilitate also can data justice criminal on draw that systems reporting Incident cross-sectional snapshot of a given situation. situation. agiven of snapshot cross-sectional anarrow provides often data survey household Rather, countries. selected oracross able within avail- data survey iscomparable cases unusual in Only indicators. andother morbidity, tality, mor- on data toprovide used often are surveys household randomized violence, of rates high by characterized insettings especially aresult, As met. be not may requirements basic these cases, tain cer- In incidents. tomonitor capacity institutional andaminimum reporting quality on depends system Any obvious. be may this for reasons 2.2). Box The (see situation given inany deaths violent of number the undercounts typically (UN-CTS), administered by the United United the by administered (UN-CTS), incident reporting incident Survey Survey

9 49 2 3 4 5 1 TRENDS AND PATTERNS 50 Box 2.2 Why incident reporting provides Since the main characteristic of a homicide is a conservative numbers dead body, homicides are rarely recorded when the body is not found. Yet people who are reported In theory, hospital records should be among the most reliable sources of information on violent deaths. ‘missing’ may well be victims of a homicide or forced In practice, however, particularly in countries with disappearance, though other explanations may limited ﬁnancial resources and in those affected by apply. Between 1980 and 2010, the UN Working acute violence, health facilities rarely systematically Group on Enforced or Involuntary Disappearances capture the causes of injury and deaths. Rather, the transferred a multitude of cases of forced disap- priority is placed on the treatment of patients (Holder pearances for clearance to governments such as et al., 2001). Further, in countries with rudimentary Algeria (2,950 cases), Argentina (3,449), Colombia public health surveillance systems, many deaths (1,236), El Salvador (2,662), Guatemala (3,155), Iraq might not be recorded as ‘assault’10—which covers (16,545), Peru (3,009), and Sri Lanka (12,230) (UNGA,

ARMED VIOLENCE 2011 VIOLENCE ARMED intentional homicides and ‘injuries inflicted by 2011, pp.135–39). of another person with intent to injure or kill’ (WHO, Between November 2009 and 2010 the Colombian gov- n.d.a)—but rather as ‘events of undetermined intent’,11 ernment cleared 211 cases, while non-governmental which may or may not include intentional homicides. sources cleared the status of another 68 missing The latter category covers all deaths for which ‘avail- people. Of the 279 cleared cases, 35 per cent were able information is insufficient to enable a medical confirmed as deaths (98 persons). In Mexico, an even or legal authority to make a distinction between higher proportion of missing people were confirmed accident, self-harm and assault’ (WHO, n.d.a). dead. Of the 134 cases cleared by the Mexican govern- While often more easily available, criminal justice data ment and the 24 cleared by NGOs, 63 were confirmed GLOBAL BURDEN BURDEN GLOBAL is significantly more vulnerable to undercounting as deaths. In Guatemala, 25 per cent of the missing than public health data. Criminal justice statistics persons were confirmed dead (63 out of 256 cleared on intentional homicides frequently capture only cases) (UNGA, 2011, pp. 136–37). events that are considered unlawful. Yet not all coun- The report by the UN Working Group does not tries share categories of what is lawful. Indeed, legal indicate whether all of these victims were intention- definitions of homicide vary across countries and may or may not include assault leading to death, ally murdered. Nor does it specify whether these euthanasia, infanticide, or assisted suicide. What is deaths were later added to homicide statistics. more, criminal justice data often records homicide However, the figures in the report highlight that if events. In other words, it does not account for the the tally of ‘missing’ who have been killed were to number of victims but rather the number of incidents. be included in homicides statistics, the number of If several people are killed in one event, the number documented homicides would increase significantly. of victims is frequently undercounted. Even in countries such as the Netherlands, ‘if all persons who were still missing after a year were victims Ultimately, the precision and reliability of criminal of a homicide, the total number of homicides would justice data—including homicide—is hostage to the increase by 5–10 per cent’ (Smit, 2011, p. 2). willingness of people to report their experiences to the police or other authorities. If a person does not In conflict settings the limitations of incident reporting trust the government, he or she may be unlikely to are even more pronounced. Studies of undercounting report an event. Obviously, only relatives and wit- in specific conflicts reveal that the number of direct nesses—rather than the victims themselves—can conflict deaths could, in extreme cases, be between report a homicide. If relatives act as perpetrators in two and four times the level actually captured by a homicide—as may be the case in ‘honour’ killings, passive incident reporting systems.12 This partly for example—they are not likely to report the incident explains the dramatic rise in field-based surveys to the authorities (Geneva Declaration Secretariat, in a growing number of countries affected by and 2008, p. 120). emerging from war. were added. were deaths conflict anddirect homicides intentional of number the step, asecond In gaps. to fillkey selectively used were WHO by generated mates esti- and2008 2004 Likewise, sources. national from data justice andcriminal health public of acombination on draws that tree a decision of basis the on determined was rates homicide specific of selection The systems. reporting dent 186 countries, 186 of atotal for collected was 2004–09) (for tion informa- of confidence. Time-series degree high acomparatively with coverage wide provides lence 2011 GBAV resulting The so that rates would not skew the data. the skew not would rates that so Region Micronesia andthe Region Antilles Lesser the into together grouped 100,000—were than less of populations with Pacific—many South andthe Caribbean inthe states island smaller of Anumber statistics. annualpopulation of basis the on calculated population—was 100,000 per year Violence the of edition this For reduction efforts. and prevention violence armed of impact the ure andmeas- todesign practitioners allow that ways in developed isseldom anddata to evaluate difficult more thus are violence Trends inarmed of Armed Violence Armed of the of edition first inthe sented pre- homicides on figures the than lower what some- are homicides intentional for figures the 2.2), Box (see undercounting of risk ahigher runs which data, justice criminal of use increased the of aresult As estimates. andWHO data health public on less relies analysis this availability, data justice incriminal improvement remarkable toa Owing globe. the across violence armed of andtrends patterns changing examining for , data was tabulated from a range of inci- of arange from tabulated was , data 13 The final violent death rate—per rate—per death finalviolent The 15 providing a useful starting point point starting auseful providing . Global Burden of Armed Armed of Burden Global database on lethal vio- lethal on database Global Burden Global 14 of the Congo, Iraq, Sri Lanka, andSudan). Lanka, Sri Iraq, Congo, the of Republic Democratic the Colombia, (namely year inanaverage deaths conflict 1,000 than more have five only population), 100,000 per ing 30 armed violence (with violent death rates exceed- by affected most states 14 top the Of conflicts. in killed number the than greater far Africa—is Sub-Saharan of andinparts Caribbean, and the America andSouth inCentral as settings—such non-conflict inso-called violently dying people of number the headlines, inmedia featured has Lanka toSri Afghanistan from inwars rienced expe- violence while isthat tonote thing first The deaths). conflict (excluding only homicides intentional for 100,000 per 6.0 andaround interventions, legal during andkillings homicide ing unintentional exclud- 100,000 per 6.8 deaths), around violent of categories allfour 7.9 (including 100,000 per isroughly rate death violent global overall the that inmind tobear useful be may it comparison, of abase As violence. lethal of levels recorded est high- the experiencing countries 58 top the ranks 2.3 andFigure population, 100,000 per deaths these of distribution global the 2.1Map shows (451,000). deaths homicide tional andinten- conflict direct of number total the on inparticular focusing deaths, these of distribution andnational regional the of asnapshot presents section This and2009. 2004 between each year interventions legal during andkillings homicide, unintentional homicide, intentional conflict, of aresult as violently died people 526,000 mated anesti- average, on that, 2004–09—highlights years the covering violence lethal on database comprehensive 2011 database—a GBAV The snapshot a 2004–09: violence Lethal 51 2 3 4 5 1 TRENDS AND PATTERNS 52 Map 2.1 Average annual violent death rates per 100,000, 2004–2009 ARMED VIOLENCE 2011 VIOLENCE ARMED of

LEGEND: Per 100,00 population

GLOBAL BURDEN BURDEN GLOBAL >30 20–30 10–20 3–10 <3 No data

Source: GBAV 2011 database

One-third of all countries with rates of lethal vio- In the 186 countries under review, roughly 12.2 lence above 10 per 100,000 population (16 out of per cent of the lethal violence occurred in armed 58) are either experiencing a ‘main armed conflict’ conflict settings and 87.8 per cent in non-conflict or emerged from one between 2004 and 2009.16 settings. This translates to 55,000 direct conflict Yet in only six of these countries—Afghanistan, deaths and 396,000 intentional homicide victims Iraq, Lebanon, Palestine, Somalia, and Sri Lanka— per year. The 55,000 average deaths per year in do direct conflict deaths constitute the majority armed conflicts around the world can be com- of all violent deaths. In the majority of the coun- pared to the estimated 48,800 people who die tries (ten) experiencing or having emerged from violently on average each year in Brazil. armed conflict, the incidence of homicide is actu- Three Central American countries—El Salvador, ally greater than the number of direct conflict Honduras, and Guatemala—display among the deaths. From among the 16 countries, three highest rates of lethal violence in the world. With countries are considered post-conflict settings: a violent death rate of 61.9 per 100,000 in 2004– Burundi, Côte d’Ivoire, and Lebanon.17 09, the people of El Salvador were more at risk of Source Madagascar Peru Lebanon Uganda Mongolia Burundi Kazakhstan Mexico Togo Guinea Cameroon Nicaragua Mali Panama Guinea New Papua Benin Paraguay Gabon Mauritania Federation Russian Botswana Tanzania Eritrea KoreaNorth Ethiopia Côte d’Ivoire Guinea-Bissau Guyana Mozambique Afghanistan Bahamas Equatorial Guinea Puerto Rico Region Lesser Antilles Ecuador Chad Namibia Republic Dominican Palestine Malawi Brazil Somalia Congo, Republic of the Swaziland Republic ofCongo, Democratic the Belize Sudan Republic African Central Lesotho Lanka Sri South Africa Guatemala Venezuela Colombia Honduras Jamaica Iraq El Salvador Figure GBAV 2011 database 2011 : GBAV 2.3

Countries ranked by violent death rate per 100,000 population, 2004–09 population, 100,000 per rate death violent by ranked Countries 0 02 04 06 08 90 80 70 60 50 40 30 20 10 100 53 2 3 4 5 1 TRENDS AND PATTERNS 54 Photo The bodies of unidentified homicide victims are buried in a mass grave at a cemetary in Tegucigalpa, November 2010. © Edgard Garrido/Reuters ARMED VIOLENCE 2011 VIOLENCE ARMED of GLOBAL BURDEN BURDEN GLOBAL 18,200, Colombia 17,500, Colombia 18,200, China 18,700, Africa South 20,700, Federation Russian the 32,700, India 48,800, recorded Brazil rates. homicide low relatively with times at although inIraq, deaths conflict direct of number the than higher were that homicides intentional from deaths violent of annualnumbers average recorded countries eight period, same the During deaths. civilian Iraqi include the only 2008) andmid-March 2003 mid-March between lence vio- armed of aresult as occurring deaths violent 2.3 inBox (92,614 figures (n.d.). The iCasualties by recorded forces Coalition the of fatalities and the Count Body Iraq by recorded deaths civilian the totaling by estimated are Iraq for deaths Conflict inIraq. violence lethal with associated istics character- the of 2.3 some Box deaths. discusses conflict direct 15,900 of annualaverage estimated an with Iraq, was in2004–09 deaths conflict of number highest the recorded that country The 100,000. per 30 than more of rates violence lethal have countries 14 Overall, 100,000. 43 per above rates death violent with Guatemala—all and Venezuela, Colombia, Honduras, Jamaica, are order, they descending Caribbean—in and the America inLatin countries other five by lowed fol- are andIraq Salvador El 100,000. per 59.4 of rate death hadaviolent Iraq and2009, 2004 between year inanaverage comparison, In world. the around population any than violently dying according to violent death rates. The main risk, mainrisk, The rates. death toviolent according countries 2.3, ranks which inFigure presented data the with associated uncertainties lining the under- isworth it trends, tobroader turning Before totals. global overall inthe heavily weigh still may violence lethal of rate low arelatively but population alarge with countries that demonstrate merely figures These victims. homicide intentional 16,800 reported both States United andthe Congo the of Republic 19 and the Democratic Democratic andthe 55 2 3 4 5 1 TRENDS AND PATTERNS 56 Box 2.3 Analysis of violent deaths of Iraqi civilians single method and perpetrator in events lasting under two days, 10,599 (18 per cent) were directly attributed to small arms and a Detailed analysis of civilian deaths during armed conflict can further 19,691 (34 per cent) were executions, of which the vast improve our understanding of the effects on civilians and specific majority were perpetrated using small arms (see Table 2.2). vulnerable subgroups in the population, including women and children. A 2011 assessment of the 92,614 Iraqi civilian deaths As shown in Figure 2.4, the researchers also calculated the number that occurred as a result of armed conflict from mid-March 2003 of women and children killed; Figure 2.5 shows the proportion of through mid-March 2008, developed from the Iraq Body Count women and children among all civilian deaths identified as men, dataset, represents the most in-depth such study to date (Hicks women, or children. Known as the ‘woman and child dirty war et al., 2011, p. 1). Iraq Body Count is a non-governmental project index’ (DWI), this indicator reflects the degree of indiscriminate that collates media reports of deaths of Iraqi civilians and cross- lethal effects on a civilian population, from indiscriminate weap- checks these reports against data from hospitals, morgues, ons or from the indiscriminate use of weapons in a conflict. The NGOs, and government bodies. DWI scale ranges from 0 (no indiscriminate lethal effects) to 100 (extreme indiscriminate lethal effects). The most indiscriminate ARMED VIOLENCE 2011 VIOLENCE ARMED The study finds that most of these violent deaths were inflicted by of effect from weapons was from unknown perpetrators firing mor- unknown perpetrators and consisted primarily of extrajudicial tars (DWI = 79). Air attacks by Coalition forces (DWI = 69) and executions of captured individuals. Unknown perpetrators also non-suicide vehicle bombs by unknown perpetrators (DWI = 54) frequently used small arms, suicide bombs, vehicle bombs, and also had highly indiscriminate effects on women and children. mortars, which had highly lethal and indiscriminate effects on Indeed, ‘Coalition forces had higher Woman and Child DWIs than Iraqi civilians. Most of the Iraqi civilians who were killed by Coali- Anti-Coalition forces, with no evidence of decrease over 2003– tion forces died during air attacks without ground fire, while fewer 2008, for all weapons combined and for small arms gunfire, died from small arms gunfire. Of the 58,251 deaths attributed to a specifically’ (Hicks et al., 2011, p. 1). GLOBAL BURDEN BURDEN GLOBAL

Table 2.2 Iraqi civilian deaths by type of perpetrator and method, mid-March 2003–mid-March 2008

Unknown perpetrator only Anti-Coalition perpetrator only Coalition perpetrator only Method Civilian Mean death/ Civilian Mean death/ Civilian Mean death/ Total deaths event deaths event deaths event Execution, any 19,321 7 (0.2) 316 7 (1.2) 54 5 (2.2) 19,691 Execution with torture 5,697 8 (0.4) 60 7 (1.6) 0 0 5,757 Small arms gunfire 8,086 2 (0.03 1,526 2 (0.1) 987 2 (0.1) 10,599 Suicide bomb 5,363 19 (2.3) 3,333 8 (0.5) 0 0 8,696 Suicide bomber in vehicle 3,029 19 (3.7) 2,370 7 (0.5) 0 0 5,399 Suicide bomber on foot 2,320 19 (2.4) 963 11 (1.5) 0 0 3,283 Vehicle bomb 3,748 7 (0.5) 1,612 5 (0.5) 0 0 5,360 Roadside bomb 1,561 2 (0.1) 1,293 2 (0.1) 0 0 2,854 Mortar fire 1,763 3 (0.1) 289 3 (0.2) 19 2 (0.6) 2,071 Air attack without ground fire 0 0 0 0 2,384 9 (0.9) 2,384 Bombs only 0 0 0 0 479 17 (3.6) 479 Missiles only 0 0 0 0 353 8 (2.4) 353 Air attack with ground fire 0 0 0 0 213 13 (3.2) 213 Totals for single perpetrators, any method* 44,750 4 (0.1) 9,511 4 (0.1) 3,990 4 (0.3) 58,251

Note: * The total ﬁgures include deaths from events involving ‘other’, ‘unknown’, and ‘combined’ methods if attributable to a single perpetrator; these criteria are not shown in the single-method rows.

Source: Hicks et al. (2011, p. 5) the Bahamas. Some states that are not in the inthe not are that states Some Bahamas. the in than inAfghanistan prevalent is muchmore violence lethal that possible isalso It list. in the Somalia follows which Brazil, of that than higher possibly and quite Republic, African Central the for reported that than ishigher Somalia for rate death violent the that possible entirely is it example, For fall. rates their tosee likely be would few very but ranking, upinthe move might states several data, better With systems. collection data andinternational national equate orinad- deaths conﬂict of reporting in incident inherent isundercounting 2.2, inBox outlined as 20 40 60 80 100 120 140 20 40 60 80 100 120 140 Figure 2008 2003–mid-March mid-March Figure DWI Coalition forces forces Coalition DWI forces Coalition by killed children and Women March March 2003 2003 2.5 2.4

Woman and child dirty war index, mid-March 2003–mid-March 2008 2003–mid-March mid-March index, war andchilddirty Woman Civilian violent deaths of Iraqi women and children from Coalition and anti-Coalition forces, forces, andanti-Coalition Coalition from andchildren women Iraqi of deaths violent Civilian Trend line line Trend March March 2004 2004 DWI anti-Coalition forces Women and children killed by anti-Coalition forces anti-Coalition by killed children and Women Trend line Trend March March 2005 2005 for the Bahamas and Brazil—as well as for most most for as well andBrazil—as Bahamas the for probably muchhigher. probably are countries inboth rates death violent overall the Somalia, for 100,000 per and1.5 Afghanistan for 100,000 2.6per were estimates WHO 2008 the low. While relatively andare figures WHO on based are andSomalia Afghanistan both for rates homicide data, justice orcriminal health public of lack consequent andthe systems surveillance In absence the of administrative 2.5). Box (see available were data upifbetter move probably would orYemen, Nigeria, Haiti, suchas ranking, top March March 2006 2006 20 In contrast, the figures figures the contrast, In March March 2007 2007 March March 2008 2008 57 2 3 4 5 1 TRENDS AND PATTERNS 58 of the states with high levels of violence—are largely reliable. So while the overall rankings might shift with better data, the reported rates for most states would not. The online methodological annex discusses in detail the confidence with which to read particular figures.

Figure 2.6 provides the overall distribution of countries according to their rates of lethal violence. Not surprisingly, a large number of countries— 77 in all—have low rates of lethal violence (less than 3 per 100,000). Most of these countries are ARMED VIOLENCE 2011 VIOLENCE ARMED

of in Europe, Northern Africa, and East Asia. One noteworthy exception in Europe is the Russian Federation, which reported annual average violent death rates of 15.0 for 2004–09. Rates of violent death above 10 per 100,000 characterize 58 states—mostly in Middle and Southern Africa, and in Latin America and the Caribbean—with GLOBAL BURDEN BURDEN GLOBAL 14 countries featuring very high levels of armed violence (more than 30 per 100,000). These 58 countries account for around two-thirds of all

Figure 2.6 Distribution of violent death rates among 186 countries, per 100,000, 2004–09

Number of countries

100

<3 3–10 10–20 20–30 >30 Violent death rate per 100,000 population

Source: GBAV 2011 database © Mikhail Metzel/ AP Photo AP Metzel/ © Mikhail 2009. January Moscow, downtown in victim Photo An investigator works near the body of a homicide ahomicide of body the near works investigator An 59 2 3 4 5 1 TRENDS AND PATTERNS 60 violent deaths (63 per cent); an estimated 285,000 rates for 186 countries into regional groupings.21 people died violently each year in these coun- The regions most affected by lethal violence are tries. The 14 countries with annual violent death Central America, with an average regional rate of rates above 30 per 100,000 population account 29.0 per 100,000, followed by Southern Africa for an estimated 124,000 deaths. In other words, (27.4) and the Caribbean (22.4). more than one-quarter of all deaths (27.5 per The variation between regions is also accompa- cent) occurred in 14 countries, where less than 5 nied by considerable intra-regional heterogeneity per cent of the world’s population lives. Targeted in lethal violence. This is hardly surprising given efforts to prevent and reduce the lethal impact countries’ different historical, political, economic, of armed violence in these countries could sig- and social experiences. Some regions show com- niﬁcantly reduce the global burden of armed mon patterns of lethal violence across countries ARMED VIOLENCE 2011 VIOLENCE ARMED violence.

of while others exhibit a wide disparity in rates of The uneven distribution of lethal violence is espe- lethal violence among states. Figure 2.8 displays cially apparent at the regional level. Figure 2.7 the proportion of countries by region according aggregates the average national violent death to the categories of levels of lethal violence.

Figure 2.7 Average violent death rates by region, per 100,000 population, 2004–09

GLOBAL BURDEN BURDEN GLOBAL Central America

Southern Africa

Caribbean

Middle Africa

South America

Eastern Africa

Southern Asia

Western Africa

Western Asia

Northern Africa

Central Asia

Eastern Asia

Eastern Europe

Northern America

South-eastern Asia

Oceania

Northern Europe

Southern Europe

Western Europe

0 5 10 15 20 25 30 35 40 45 50 Violent death rate per 100,000 population

Source: GBAV 2011 database South America most countries display elevated elevated display countries most America South and Africa Middle In 2.4). Box (see in Jamaica 100,000 per to 58.1 Bahamas in the 100,000 per 19.7 from ranging rates death violent with lence, vio- lethal of levels high tovery high from suffer countries other All 100,000. 10per below rates death violent experience andHaiti Cuba only region, Caribbean andinthe 100,000, 10 per than more of rates death violent exhibit countries American allCentral Rica, Costa of exception the With population. 100,000 10per above rates death annualviolent allfeature Africa Southern in countries ﬁve the example, For countries. across homogenous violence is comparatively lethal of incidence the regions certain Within Source Europe Western Southern Europe Africa Northern Europe Northern Oceania Asia Western Eastern Asia Eastern Europe South-eastern Asia America Northern Southern Asia Asia Central Africa Western Eastern Africa South America Caribbean Middle Africa America Central Southern Africa Figure <3 3–10 GBAV 2011 database 2011 : GBAV 2.8 10–20 10–20

Percentage of countries per violent death rate per 100,000, 2004–09 100,000, per rate death violent per countries of Percentage 20–30 20–30 01 >30 02 03 0 lethal violence with a rate just below 20 per 100,000 100,000 20 per below just arate with violence lethal of levels high records also Afghanistan 100,000; per 30 than more of rates from suffers Lanka Sri although 100,000, 10per below rates death violent have countries numerous where Asia, is Southern inpoint Acase countries. among violence lethal of distributions unequal highly feature regions some homogenous, relatively are regions many While 100,000. 10per below rates death violent have allcountries America, andNorthern Asia, in Southern and Europe, Northern South-east Similarly, population. 100,000 3per below rates death annualviolent experience countries European allWestern spectrum, the of end other the age). At aver- global (abovethe violence armed of levels Percentage ofcountries per region 40 50 60 70 80 90 100 61 2 3 4 5 1 TRENDS AND PATTERNS 62 (a figure that probably suffers from undercounting Box 2.4 Violent deaths in Central America and the Caribbean of homicides due to limited institutional capacity, The intensity of armed violence across Central America and the Caribbean as noted above). Likewise, with the exception of is several times the global average, with a few exceptions. The extent of Sudan, Northern Africa exhibits low levels of armed lethal violence in these states has led to concerns that it is contributing to violence. The so-called ‘Arab Spring’ that swept the erosion of the rule of law and social order. Many analysts underline the across Bahrain, Egypt, Libya, Syria, Tunisia, and fact that homicide casualties as a result of urban violence in El Salvador, Yemen in 2011 may yet reveal dramatic changes in Guatemala, Honduras, and Jamaica have surpassed the number of deaths reported violence and victimization. In any case, in classic conflicts (Seligson and Booth, 2010; Rodgers, 2010; Zinecker the considerable variation between and within 2008); some have actually likened the characteristics of violence to that regions—and the different factors that may account of an outright armed conflict (Manwaring, 2007). Indeed, governments in for it—highlight the importance of not generating North America and Western Europe have started to initiate military and ARMED VIOLENCE 2011 VIOLENCE ARMED overly simplistic conclusions from the data.

of security support strategies to these countries to prevent the deepening of armed violence and contagion across borders. Although the data for countries with high levels of violence is relatively robust, a number of

Map 2.2 Average annual violent death rates per 100,000 in Central America and the Caribbean, 2004–09 GLOBAL BURDEN BURDEN GLOBAL

BAHAMAS CUBA

MEXICO

HAITI DOMINICAN PUERTO RICO REPUBLIC JAMAICA L e s s BELIZE e r

HONDURAS l

GUATEMALA e s

EGEND L : R

EL SALVADOR i Per 100,00 population o >30 NICARAGUA n 20–30

10–20 COSTA 3–10 RICA PANAMA <3 No data

Source: GBAV 2011 database being recorded (UCDP, n.d.c). even or attention public’s the catching without often intensity, low at simmered have Yemen in violence armed of forms different lives, 1,500 estimated an formal state mechanisms. state formal than rather customary by with dealt are cases such importantly—many most First—and disputes. water or land to related deaths undercounting for reasons three suggested official government AYemeni tries. coun- or regions some in violence conflict-related and interpersonal both of underreporting significant isprobably there that suggests example Yemeni The statistics. homicide nor conflict armed neither in counted are which but violence, armed on dataset and inclusionin one or close attention another deserve that violence of manifestations iswitnessing Yemen Clearly, database. 2011 GBAV the into integrated not are they thus and conflicts armed on statistics international common most the in appear deaths these of none 2010b); Survey, Arms (Small vations moti- economic and political underlying with disputes’ ‘land to due year each killed are people 4,000 of age aver- an that acknowledges government Yemeni the of report unpublished an importantly, More n.d.). (CSO, period same the for reported are that murders’ 163 ‘unintentional average additional an are figure this in Missing n.d.). (CSO, 2009 and 2004 between deaths) (14 branch’ the kill origin ‘the and deaths), (30 death’ to leading ‘assault deaths), (875 murder’ ‘intentional to due year per deaths 919 of average an records data justice and security available Publicly involved. become to states other and Arabia, Saudi Kingdom, United the States, United the prompted has which terrorism, and war civil all-out to protesters and forces security state between clashes political separatist from ranging conflicts, armed interlocking of number a by afflicted isalso It turmoil. political and social in isembroiled Yemen state, aunified as 20years After B make it impossible to identify who killed whom, whom, killed who identify to impossible it make can conflicts such many of dynamics and nature tribal the Finally, actions. state of impact the about sensitivities raising intervene, to isused military the ox 2.5 Violent deaths undercounted in Violent Yemen 22 Since the 1994 civil war, which claimed claimed which war, civil 1994 the Since 24 Moreover, in many cases in many cases Moreover, 23 Sources: Somalia. in reported 100,000 per 26.0 of rate the to comparable be would which 100,000, per 26.2 to increase would rate death violent overall country’s to added Yemeni the the were deaths data, conflict’ ‘land 2010b). 4,000 the If Survey, Arms (Small involvement prosecutorial of likelihood the reducing inspects his damaged residence in Sana’a, June 2011. © Mohammed Huwais/AFP Photo Huwais/AFP © Mohammed 2011. June Sana’a, in residence damaged his inspects Photo S Small Arms Survey (2010a; 2010b) (2010a; Survey Arms Small An armed tribesman guards Yemeni opposition tribal chief Sheikh Sadiq al-Ahmar as he as al-Ahmar Sadiq chief Sheikh Yemeni tribal guards opposition tribesman armed An 63 2 3 4 5 1 TRENDS AND PATTERNS 64 countries are not included in Figure 2.3 although conflict deaths reported by the news portal INSEC they have either recently experienced armed and the IISS Armed Conflict Database (CBS, 2009, conflicts or are believed to have high levels of s. 17.4; INSEC, n.d.). For 2004–09, Nepal—along violence. Countries such as Burundi, Haiti, Kenya, with Afghanistan, Iraq, Palestine, Somalia, and Liberia, Nepal, Nigeria, Pakistan, and Yemen Sri Lanka—reported more direct conflict deaths have each featured chronic and acute outbreaks than intentional homicides. In war-affected soci- of violence in their capitals or other prominent eties, administrative sources typically lack the cities in the prelude and wake of elections or are capacity to record all intentional violent deaths. affected by armed conflicts. And yet all of these As such, there are reasons to suspect that the countries actually report lethal violence rates information reported by Nepal’s Central Bureau below the global average. of Statistics undercounts the number of victims. ARMED VIOLENCE 2011 VIOLENCE ARMED Another case of potential undercounting—Yemen— of For example, despite recently emerging from a is highlighted in Box 2.5. decade-long civil war in 2006, Nepal’s average national violent death rate is reported at 6.2 per Lethal violence is unevenly distributed not only 100,000; it thus figures in the category of coun- across countries, but also within them. Both tries with a medium level of armed violence. The Mexico (with an estimated 74,000 violent deaths country’s lethal violence rate is estimated on the between 2004 and 2009) and Pakistan (with basis of data on intentional homicides reported 90,000 victims of homicide and armed conflict GLOBAL BURDEN BURDEN GLOBAL by Nepal’s Central Bureau of Statistics and direct over the same period) have medium lethal violent death rates. Mexico’s annual violent death Figure 2.9 Violent death rates per 100,000 in Mexico, Chihuahua, and rate averaged 11.5 per 100,000 between 2004 and Ciudad Juarez, 2004–09 2009, although some states have much higher Mexico Chihuahua Juarez rates; in 2009 the violent death rate in the most

Violent death rate per 100,000 population affected region, Chihuahua state (with three

180 million inhabitants), was 108.0 per 100,000 (see Box 1.4, A UNIFIED APPROACH). Armed violence is

150 even more concentrated in certain cities. In Ciudad Juarez, a city of just over 1.3 million inhabitants, 2,399 people were killed in 2009, which trans- 120 lates into a murder rate of 170.4 per 100,000 (see Figure 2.9). 90 As in Mexico, the number of violent deaths in Pakistan appears to have increased signiﬁcantly 60 since 2004. Owing in part to the escalation of violence mainly on the Afghanistan–Pakistan 30 border, the number of annual violent deaths in Pakistan has increased from around 10,500 in 2004 2005 2006 2007 2008 2009 2004 to 24,500 in 2009. Despite this escalation,

Source: Small Arms Survey elaboration based on INEGI (n.d.) the national violent death rate in Pakistan in an see Figure 2.10). Figure see (GoP, 167.0 in2009 mately 1998; 100,000 per approxi- of rate death conflict adirect represent would (1998 census),figure people this million 3.1 over just of apopulation With alone. in 2009 deaths conflict direct had5,304 reportedly stan andAfghani- Pakistan between region western north- inthe Areas Tribal Administered Federally figure increased to423,000. increased figure these in2009 while in2006, to368,000 2004 397,000 in from dropped homicides intentional of victims of number The countries. particular andin anannual basis on dramatically fluctuate can violence lethal of rates time, of periods er long- across stable rather seemingly While 2006. in and6.4 in2004 100,000 per to6.8 compared as 7.0 at 100,000, stood per in2009 rate death violent global the First, patterns. possible some out totease ispossible it analysis, trend tailed de- for enough isnot data of years six Although 2004–09 violence, Trends lethal in between 2006 and 2009 inPakistan. and2009 2006 between rates death conflict direct the 2.10 highlights Figure others. than affected more are country inthe regions andcertain distributed unevenly is violence the though 100,000, per 8.6 only still roughly was and2009 2004 between year average described in greater detail below. detail ingreater described all Lanka, and Sri Somalia, Pakistan, in Iraq, conflicts armed of andflow ebb the of a result largely are deaths conflict indirect changes The 58,000. toroughly again dropped they 2009 In in2007. to66,000 increased deaths conflict direct in2005, to40,000 in2004 46,000 from decrease a After volatile. more are deaths conflict Direct 26

25 The 30 60 90 120 150 180 Violent death rate per 100,000 population review whose violent death rates for 2009 are are 2009 for rates death violent whose review under 40countries the of ten 2.12 tracks Figure 2004. of rates the twice least at were countries inthese rates death violent The andUganda. Peru, Panama, Palestine, Pakistan, Mexico, Iran, Honduras, Afghanistan, were and 2009 2004 between shifts upward hadsignificant that countries year. Other that conflict armed intense duetothe mainly since2004, increase greatest the and rate death violent highest the experienced section. inthis attention of focus the are andthey and2009, 2004 between year given inany 100,000 10per than higher rates death violent with 40countries only for exists data series time andcoherent comprehensive Sufficiently and GoP (1998) Source in Figure 2.11. in Figure shown as and2009, 2004 between rates death inviolent changes hadsignificant countries these provinces, 2006–09 provinces, Figure Balochistan Balochistan Small Arms Survey calculation based on conflict deaths figures provided by SATP (2011) (2011) SATP by provided figures deaths conflict on based calculation Survey Arms : Small 2.10 2006 Khyber Pakhtunkhwa Pakhtunkhwa Khyber

Direct conﬂict death rates per 100,000 in Pakistan’s inPakistan’s 100,000 per rates death conﬂict Direct 28 It reveals that in 2009 Sri Lanka Lanka Sri in2009 that reveals It 2007 Federally Administered Tribal Areas Areas Tribal Administered Federally 27 A number of of Anumber 2008 Punjab Punjab 2009 Sindh Sindh 65 2 3 4 5 1 TRENDS AND PATTERNS 66 ARMED VIOLENCE 2011 VIOLENCE ARMED of GLOBAL BURDEN BURDEN GLOBAL

Photo S Tamil families use old artillery shell boxes to navigate their way through mine ﬁelds as they return to homes they were forced to leave due to ﬁghting, Palampiddi, Sri Lanka, July 2010. © Patrick Brown/Panos Pictures Burundi Nepal Lithuania Mongolia Suriname Kazakhstan Rica Costa Federation Russian Peru Seychelles Uganda Iran Somalia Paraguay Pakistan Guyana Iraq Mexico Ecuador Puerto Rico Republic Dominican Panama Region Lesser Antilles Bahamas Afghanistan Palestine Sudan Belize South Africa Colombia Guatemala Venezuela Jamaica Honduras El Salvador Lanka Sri 2009 and 2004 Figure Source 2004 2004 GBAV 2011 database 2011 : GBAV 2009 2.11

Violent death rates per 100,000 population, population, 100,000 per rates death Violent 0 02 04 06 009 100 90 8070 60 50 40 30 20 10 Violent death rate per 100,000 population 67 2 3 4 5 1 TRENDS AND PATTERNS 68 more than two times higher than the lowest In contrast, a cluster of countries reported dra- rate in any given year between 2004 and 2009. matic decreases in lethal violence between 2004 It shows that the violent death rate in 2009 in and 2009. Figure 2.13 shows six of the 40 coun- Sri Lanka was more than ten times higher than tries under review whose reported violent death that reported for 2004. The increase in violent rates in 2009 were less than half the highest rate deaths is largely a function of extensive military recorded for any given year between 2004 and operations in the final phase of the 26-year civil 2009. The country exhibiting the largest propor- war. In 2009, the armed conflict cost the lives tional decline was Lebanon. After a sharp increase of at least 15,500 people.30 Meanwhile, a number during the 2006 conflict between Hezbollah and of Central American countries experienced fast Israel (resulting in a violent death rate of 33.1 per and steady increases in homicidal violence. 100,000), the violent death rate dropped to a

ARMED VIOLENCE 2011 VIOLENCE ARMED Between 2004 and 2009, violent death rates in reported level of just 2.7 per 100,000 in 2009. of Honduras more than doubled, from 31.9 to 70.6 The country now appears to have low levels of per 100,000. Perhaps somewhat paradoxically, armed violence, although the completeness of and despite concerns about underreporting national reporting is questionable (Small Arms with respect to dramatic declines in lethal vio- Survey, 2011). lence, upward trends are usually based on fairly Between 2007 and 2009, violent death rates in robust data. Somalia dropped more than sixfold, from 76.0 per GLOBAL BURDEN BURDEN GLOBAL 100,000 in 2007 to 12.5 in 2009. Lethal violence Figure 2.12 Countries with significant increases in violent death rates in Iraq also plummeted. In 2006, Iraq experienced per 100,000, 2004–09 an estimated overall violent death rate of 105.6 Sri Lanka Afghanistan Palestine Iran Panama per 100,000, but by 2009 the annual toll was down Honduras Mexico Pakistan Peru Uganda to around 5,400 people—or a rate of roughly 17.6 Violent death rate per 100,000 population per 100,000—largely attributed to a combination 110 of military ‘surge’ activities and declining insur- 100 gent violence. 90 All of these six countries were affected by an 80 armed conflict at some point between 2004 and 70 2009. While post-conflict settings are sometimes 60 associated with rising criminal armed violence, a 50 number of post-conflict countries seem to exhibit 40 reductions in violent death rates.29 Burundi and 30 Nepal, for example, report a considerable decline

20 in lethal violence. These declines over a short

10 period are mainly attributed to peace processes and conflict termination, and the consequent 2004 2005 2006 2007 2008 2009 reduction in direct conflict deaths. Note: For 2009, these countries recorded violent death rates that were twice as high as the lowest rate recorded for any given year between 2004 and 2009. In Burundi, the number of reported direct conflict Source: GBAV 2011 database deaths dropped from 820 in 2004 to 17 in 2009, regional, national, and local levels. andlocal national, regional, the at violence of dynamics shifting the assess to isneeded analysis fine-grained amore stable, relatively remain may trends global while that highlight however, do, They countries. particular in violence armed of patterns changing driving factors the about anything reveal themselves by not anddownward—do upward shifts—both these analysis, contextual of absence the In in2009. and2.8 in2006, 2.2 2004, in 100,000 per 3.4 between only varying period, six-year the over stable comparatively remained inNepal rates homicide reported the however, Unlike inBurundi, 100,000. per 8.9 to1.0 from rate death conflict direct the of adecline into translates This in2009. 300 than tofewer 2004 in 2,380 significantly—from decreased deaths conflict direct of number the inNepal Likewise, n.d.). (UNODC, year same the for to UNODC 7.6 of 100,000 arate per reported officially Police National Burundi the p. 26); meanwhile, 2008, anddeTessières, (Pézard 2008 for 100,000 per 12.3 aloneat andexplosives weapons, bladed guns, with committed homicide of rate the places Violence Armed on Observatory Burundian the by collected data example, For murdered. people of number actual the undercount may statistics as cautiously interpreted be must figures official poor; remain Burundi suchas settings in post-conflict capabilities collection data official war, however, the during infrastructure surveillance-related of destruction tothe inpart Owing in 2009. 100,000 per to3.2 in2005 12.2 from fell edly report- inBurundi rates homicide and2005, 2004 between increase asharp After conflict. armed the of end sincethe considerably dropped to have reported also was inBurundi rate homicide the time, same the At 100,000. per 11.5 to0.2 from rate death conflict direct inthe adrop into translating Source: 2009. and 2004 between year given any for recorded 10 20 30 40 50 60 70 80 90 100 110 Violent death rate per 100,000 population 13 2004–09 population, 100,000 per Figure improvements in the availability of criminal justice justice criminal of availability inthe improvements recent given especially straightforward, paratively iscom- deaths and conflict homicide Tabulating puzzle. the of part isanimportant it but time, orover country by down broken be stage, this at cannot, data This interventions. legal during andkillings homicides unintentional for data includeaggregate also that totals overall presents chapter this Consequently, violence. armed of burden total the of muchless violence, lethal of magnitude actual the of picture partial a only provides data this insight, ing important offer- While deaths. conflict anddirect homicide intentional of review asystematic from is derived above analysis andtrend regional, national, The pieces? missing the deaths: Violent Note: Lebanon For 2009, these countries recorded violent death rates of less violent death rate rates than half recorded the highest countries these 2009, For GBAV 2011 database 2. 2004 Somalia Somalia

Countries with significant decreases in violent death rates rates death inviolent decreases significant with Countries Burundi Burundi 2005 Iraq Iraq Georgia Georgia 2006 Nepal 2007 082009 2008 69 2 3 4 5 1 TRENDS AND PATTERNS 70 data. This edition of the Global Burden of Armed while just over 10 per cent are direct conflict Violence estimates an average of approximately deaths. Unintentional homicide also accounts for 396,000 intentional homicide victims per year just over 10 per cent of all violent deaths, while between 2004 and 2009, a lower figure than the killings during legal interventions account for 490,000 homicide victims estimated in the 2008 slightly more than 4 per cent of all violent deaths. report. The latter figure draws more extensively The 2011 Global Burden of Armed Violence does on public health data that reports violent deaths not focus on indirect or ‘excess’ conflict deaths, independent of the intent behind the killing, or which is certainly the largest portion of the burden that relies on modeling estimates. In addition, of conflict deaths. Individuals dying in conflict this edition estimates a small increase in annual zones due to easily preventable diseases such direct conflict deaths, yielding an annual average as dysentery, measles, hunger, and malnutrition ARMED VIOLENCE 2011 VIOLENCE ARMED of 55,000 deaths between 2004 and 2009 (com- of are a major contributor to the overall burden. The pared to the 52,000 estimated in 2008 for the 2008 Global Burden of Armed Violence report period 2004 to 2007). estimates a ratio of 4:1 indirect to direct conflict Adding the figures for the total number of violent deaths. This conservative ratio was used to deaths from unintentional homicide (54,000) and facilitate estimates of the overall excess death from killings during legal interventions (21,000) rate in conflict-affected countries; it was gener- to these two subtotals—the 396,000 intentional ated from a review of mortality rates in 13 different GLOBAL BURDEN BURDEN GLOBAL homicide victims and 55,000 direct conflict deaths— conflicts around the world (Geneva Declaration yields an estimated 526,000 violent deaths each Secretariat, 2008, p. 42). Applying this same year between 2004 and 2009. As Figure 2.14 shows, ratio to the estimated 55,000 direct conflict slightly more than 75 per cent of these violent deaths in an average year between 2004 and deaths are the result of intentional homicide, 2009 would imply an excess death toll of 220,000. If indirect conflict deaths are added to Figure 2.14 Disaggregating the global burden of the total number of violent deaths, the total glo- lethal violence bal burden of armed violence reaches 746,000 Legend: deaths per year for 2004–09. Direct conflict deaths (55,000; 10.4%) Intentional homicide (396,000; 75.3%) Nevertheless, fundamental disagreements persist Unintentional homicide (54,000; 10.2%) over the methodologies used to measure direct Legal intervention killings (21,000; 4.1%) and indirect conflict deaths.31 There remains considerable debate over whether conflict deaths have been increasing or decreasing in recent decades, and over whether estimates of indirect deaths and excess mortality are accurate. The controversy over efforts to measure the burden of violence in the Democratic Republic of the Congo highlights the challenges of estimating excess mortality in complex emergencies (see

Source: GBAV 2011 database Box 2.6). 1,000 per month for the five areas surveyed. The The surveyed. areas five the for month per 1,000 per deaths excess 3.7 of estimate IRC the yielding figure, this from subtracted then was month per 1,000 per 1.5 of CMR baseline pre-war The month. per population 1,000 per deaths 5.2 of rate mortality age aver- an estimated IRC 2007, the and 2004, 2002, in out carried surveys nationwide three of basis the On figures. final the on impact amajor has rate abaseline of choice the situation, acrisis in CMR the and CMR baseline the between difference isthe mortality excess Since conflict. armed an of absence the in rate mortality natural baseline DRC’s the over adisagreement of isaresult estimates low and Report Human Security the with associated Researchers p.45). 2010, (HSRP, deaths conflict indirect or deaths’ ‘excess so-called to attributed be can estimate IRC the of is and even more significant onlydifference one-third the surveys, three last the for that claims further report The IRC. the by asserted than lower cent per 60 almost isactually survey first the in estimate deaths’ 2010 Report The rates. mortality calculating to approaches survey-based of use the and figure this challenged have organizations of A number survey. household randomized autopsy’—a a‘verbal was toll death the estimate to used approach primary 2007, p.2). The al., et (Coghlan conflict the of aresult as died people million 5.4 of atotal 2007 and 1998 between that estimates IRC 2007, the and 2006 between ducted con- surveys additional on based report, a2007 In 2004. and 1998 between died people million 3.9 2004—that and 2000 between conducted were that surveys four on estimates—based IRC the report, cited awidely In (DRC). Congo the of Republic Democratic the in conflict armed of costs human the understand better to effort amajor launched In 2000 the (IRC) Rescue International Committee B the Democratic Republic of the Congo the of Republic Democratic the ox 2.6 contend that the difference between high high between difference the that contend Estimating crude mortality rates in rates mortality crude Estimating , for example, claims that the ‘excess ‘excess the that claims example, , for crude mortality rate (CMR)—the (CMR)—the rate mortality crude Human Security Human Security rates deteriorated. rates where areas neighbouring included improvement an demonstrating sites several since identified be could pattern geographical clear no but trend), tive posi- most the shows that region (the DRC eastern in located mainly were areas latter The deterioration. a witnessed districts six and unchanged, almost remained three improvements, showed districts four 2008–10, and 2006–07 period the during that found CRED available, was data survey Where p.30). 2011, (CRED, level acommunal on years five last the in CMRs of volatility observes CRED context, this In CMRs. deteriorating show communities neighbouring while years, five last the over trends improving show nities commu- some regions, certain Within CMRs. in tions fluctua- highlights study CRED the time, same the At 70). (p. IRC the by used day, per 10,000 per 0.5 or month, per 1,000 per 1.5 of rate baseline a with consistent are day) per 10,000 0.3 per 1.1 and (between CMR post-conflict the for figures its but CMR, abaseline of estimate an provide not does report CRED The Maniema. and Katanga, Kivu, South and including North provinces, clear in eastern the larly isparticu- trend positive The decade. past the over stable or remained decreased has inces in DRC the prov- all in CMR overall the that shows report CRED The p.12). 2011, (CRED, month per people 1,000 per 3deaths roughly into translates which day, per people 10,000 per 1death of threshold emergency the below aCMR reveal DRC the in 2010 and 2000 between conducted surveys mortality most that concludes CRED p.8). 2011, (CRED, tolls’ death estimated around debates raging to therefore and deaths of under-estimation or over- to lead can which bias, of sources many to subject be ‘can they as CMRs estimating of complexity the highlights countries African eight in situations emergency plex com- in health on (CRED) Disasters of Epidemiology the on Research for Centre the by report A new p.33). (HSRP, 2010, percent’ 60 some ‘by reduced be thus would toll mortality excess ate; the appropri- more be would month per 1,000 per deaths 2.0 of CMR abaseline that contends report HSRP 71 2 3 4 5 1 TRENDS AND PATTERNS 72 ARMED VIOLENCE 2011 VIOLENCE ARMED of GLOBAL BURDEN BURDEN GLOBAL

Photo S Displaced people, who fled their homes due to fighting, line up for food at a distribution centre in Kibati, DRC, November 2008. © Les Neuhaus/Reuters culpability andintent’. culpability degree, of interms defined are they inhow ries andterrito- states the across differences are there offences, generic fairly are andmanslaughter murder ‘while that confirms Statistics of Bureau Australian The system. federal aseparate tomention not unit, eachadministrative for definitions offence and legislation, of sets systems, justice different slightly eight are there aresult, law. As criminal aseparate features unit eachadministrative that means which level, andterritory state the at cated adjudi- often are offences criminal example, for Australia, In countries. within) even sometimes (and across vary frequently ahomicide stitutes con- what of definitions legal isthat reality The 4). (p. excluded also is homicide incomplete but Attempted suicide. andassisted abortion are as isexcluded, driving dangerous through death Causing p.4). 2010, stat, (Euro- andinfanticide’ euthanasia manslaughter, includingmurder, aperson, of killing ‘intentional the as homicide defines Eurostat agency The ‘manslaughter’. as described andcommonly tal’ ‘acciden- generally are homicides’ ‘Unintentional p.349). 2009, (Krause, perpetrator’ the of part the on intentionality or responsibility diminished implies that is‘acategorisation hand, other intend to cause death. purposefully perpetrator the that requires ically typ- ‘murder’, toas referred commonly homicide, intentional hand, one ‘homicides’. the On record and toclassify how over disagreement siderable con- isstill there above, noted As task. daunting a remains homicide intentional beyond lence vio- lethal non-conflict of analysis the Extending homicide unintentional Disaggregating categories. England and Wales, for example, example, for andWales, England categories. sub- several into manslaughter divide countries some intentionality, of To degrees for account 32 33 ‘Manslaughter’, on the the on ‘Manslaughter’, 73 2 3 4 5 1 TRENDS AND PATTERNS 74 Table 2.3 Intentional versus unintentional homicide: a sample

Country* Latest available year Intentional homicide Unintentional homicide**

Argentina 2008 2,305 613

Burkina Faso 2008 222 98

Costa Rica 2009 525 5

France 2009 682 137

Fiji 2008 20 3

Germany 2009 706 0 (341)

ARMED VIOLENCE 2011 VIOLENCE ARMED Ghana 2005 383 4 of India 2009 32,369 3,930

Italy 2008 611 372

Kenya 2009 1,203 41

Mexico 2009 16,117 2,763

Nepal 2009 806 12

GLOBAL BURDEN BURDEN GLOBAL Nicaragua 2008 693 22

Nigeria 2008 1,956 17

UK 2009 638 0 (194) (England and Wales)

Total sample 59,236 8,017

Global total 396,000 54,000 (rounded)

include murder, section 2 manslaughter (accepting involuntary manslaughter is defined as homicidio diminished responsibility), other manslaughter, culposo. In many of these countries, homicidio and cases of infanticide in homicide records culposo also includes the killing of one person (Smith et al., 2011). Complicating matters is the by another as a result of a road accident. This can fact that in almost all other languages besides lead to serious misunderstandings and errors English, there is no exact equivalent of the word when working on comparing homicides statistics ‘homicide’. Some languages do not even feature across countries and language groups (Smit, 2011). a generic term, while others use a generic term At a minimum, it is critical that policy-makers, for intentional killings only. In many countries, practitioners, and researchers recognize these unintentional homicide also includes deaths disparities and reconcile them where possible. resulting from car accidents. In Mexico and numerous other Latin American countries, murder is Table 2.3 presents a sample of countries that dif- commonly referred to as homicidio doloso while ferentiate between intentional and unintentional Germany: The 365 cases of murder ( murder of cases 365 The Germany: 125). p. 2011, (FIBOS, not is ‘manslaughter’ statistics; homicide intentional in included is ‘Murder’ Fiji: homicide statistics (DCPJ, 2009, p. 11). France: 8). p. 2009b, lesiones arma de fuego (PJCR, injuries) categories the include only year per victims five The accidents. car of victims to refer deaths these of Some Rica: Costa statistics; volontaires Homicides Faso: Burkina cidios culposos en tránsito are included in intentional homicides statistics. statistics. homicides inintentional included are that homicides of number the shows table The 2.5). inYemen Box (see and20009 2004 between year inanaverage occur that homicides tional 163 the uninten- excludes it Likewise, violence. duetoarmed certainly are which of 2010), some (SAPS, Africa inSouth in2008 occurred that cides’ ‘culpable homi- 13,184 than more includethe not does table the aresult, As list. the from excluded havebeen accidents road from resulting and those violence interpersonal of aresult as homicides unintentional between differentiate not do that Countries accidents. car excluding while homicide, hom intentional on statistics the in brackets) in (shown homicides unintentional include Wales) and (England UK the and Germany homicide notes ** Unintentional 36). p. 2011, al., et (Smith statistics homicide intentional in included are cases All 1infanticide. as well as 2mansla manslaughters, section 28 other 166 and murders, 284 are these Among aconviction. reported cases 479 homicides, 638 the Of Wales): and (C (England UK ‘manslaughter’ of 17victims the not but ‘murder’, of victims 1,956 the include statistics homicide intentional The Nigeria: 66). p. 2008, (PN, statistics homicide intentional homicido culposo categories The asesinatos to refer statistics homicide intentional in included are that victims 693 The Nicaragua: ( 1). p. murder’ 2009, with (CBS, excluded ‘dacoity of are cases 12 murder) the while statistics, homicide intentional in included are ‘murder’ of cases 806 The Nepal: 1). p. 2009, (SESNSP, it as accidents car of table the in victims listed not includes 2009—is in probably people causes’—12,665 ‘other of died who victims homicide unintentional of group largest The causes: following the from died who p. 19). 2010, Police, (Kenya Mexico: included not is ‘manslaughter’ statistics; homicide intentional in included are infanticide’ and ‘Murder Kenya: stradale ( 2009, Italy: (NCRB, not is murder’ to amounting not homicide ‘culpable statistics; homicide intentional in included is ‘Murder’ India: 18). p. 2005, (GST, not is ‘manslaughter’ statistics; homicide intentional in included is ‘Murder’ Ghana: 131). p. 2010, (BKA, statistics homicide intentional in included all are accid traffic of aresult as deaths to refers homicide of group largest The not. are means) other using homicides (unintentional Argentina: notes * Country Table omicidi colposi Omicidi volontari consumati Omicidi volontari Coups et blessures volontaires, suivis de mort de suivis volontaires, blessures et Coups 2.3 ) (ISTAT, 2008). ) (ISTAT, Homicidios dolosos coups mortels Homicidos dolosos Homicidio doloso Homicidio

Notes ) are not included. The 372 victims refer to unintentional homicides without victims of car accidents ( accidents car of victims without homicides unintentional to refer victims 372 The included. not ) are :

(mortal blows) are not included (INSD, 2010, p. 144). p. 2010, (INSD, included not are blows) (mortal ), which is also excluded from intentional homicide statistics (MJSDH, 2008, p. 1). p. 2008, (MJSDH, statistics homicide intentional from excluded also is ), which are included in intentional homicide statistics. The 2,763 cases of of cases 2,763 The statistics. homicide intentional in included are is included in homicide statistics (PJCR, 2009a, p. 5); homicidio culposo 5); p. 2009a, (PJCR, statistics homicide in included is (intentional homicides) are included in intentional homicide statistics; statistics; homicide intentional in included are homicides) (intentional and homicido and preterintencional (voluntary homicides, completed) are included in in in included are completed) homicides, (voluntary Mord : et por arma de fuego, por arma blanca arma por fuego, de arma por

assassinats ) and the 341 victims of manslaughter or killing on request ( request on killing or manslaughter of victims 341 the ) and (voluntary homicide and assassinations) are both included in intentional homicide homicide intentional in included both are assassinations) and homicide (voluntary (voluntary blows and injuries, followed by death) are not included in intentional intentional in included not are death) by followed injuries, and blows (voluntary (felony homicide), accounting for 22 victims in 2008, are not included in in included not are 2008, in victims 22 for accounting homicide), (felony sin datos sin , and statistics on intentional homicides. intentional on statistics in homicides unintentional Germany—include or andWales England as countries—such few Table as 2.3 a only cases; inmany lence shows, vio- lethal of rates overall of to underestimating lead may homicide andunintentional intentional between Distinguishing andMexico. Italy, India, Germany, France, Faso, Burkina include Argentina, homicides intentional of total the of cent 10 per than more are that rates homicide unintentional with countries Sample included. were homicides ifunintentional substantially increase would rates homicide overall cases insome that suggests It tentional homicide statistics; unintentional homicides homicides unintentional homicide statistics; tentional (with a firearm, with a bladed weapon, and no data). data). no and weapon, abladed with afirearm, (with homicidios culposos (assassinations) and (assassinations) (unintentional homicide) is not included. homicidios culposos por otros hechos Totschlag und Tötung auf Verlangen auf Tötung und Totschlag refer only to the victims victims the to only refer di cui da incidente incidente da cui di homicidio doloso LEEN, n.d., p. 1). p. n.d., LEEN, banditry with banditry ents ( ents p. 220). p. ughters, ughters, (firearm (firearm icides. homi-

) . 75 2 3 4 5 1 TRENDS AND PATTERNS 76 A review of the available ﬁgures suggests that the rate of homicide would increase by around Photo A woman holds her head in her hands following an 13.6 per cent if unintentional homicides were explosion at Assumption Church, Kathmandu, in which two people were killed, May 2009. © Shruti Shrestha/Reuters included. Put another way, if this proportion were applied to the estimated 396,000 intentional homicides, the global burden would increase by an additional estimated 54,000 deaths, yielding an estimated 450,000 annual homicide deaths (intentional and unintentional). This is roughly consistent with (although somewhat lower than) the estimate for homicide deaths put forward in the

ARMED VIOLENCE 2011 VIOLENCE ARMED ﬁrst edition of the Global Burden of Armed Violence, of which relied extensively on WHO estimates.

Killings during legal interventions and extrajudicial executions

GLOBAL BURDEN BURDEN GLOBAL Another category of lethal violence often not captured by homicide statistics consists of deaths occurring during legal interventions and extrajudicial killings. The UN Special Rapporteur on extrajudicial, summary or arbitrary executions, whose office was established in 1982, defines extrajudicial executions and unlawful killings as ‘killings that violate international human rights or humanitarian law’ (UNGA, 2010b; UN-ECOSOC, 2005, para. 6). Such a broad interpretation opens the door to a wide range of categories of lethal violence. For example it would include killings by law enforcement officials or other security forces; killings during armed conflict; killings during counterterrorism operations; killings by non- state actors; and deaths in custody and due to the death penalty (UNGA, 2010b).34

It is currently impossible to verify or validate the annual global distribution and burden of extrajudicial executions. There are no reliable monitoring mechanisms and many governments are not prepared for full disclosure or may lack the of Armed Violence Armed of 2008 inthe reported As killings. andunlawful executions extrajudicial on report to seeking actors of set only the represents often sector humanrights The investigations. necessary toundertake andresources capacities occurring during legal interventions (killings of of (killings interventions legal during occurring Violence Armed of Burden the of edition This judicial executions. andextra- homicides, intentional deaths, conflict direct between overlaps potential many are There executions. extrajudicial of aresult as annualdeaths 1,900 least at of anestimate yields 3 category 2and51 for category 5for of pliers multi- conservative Applying year). per 50 deaths than (more frequent are killings and unlawful executions extrajudicial where countries and 3) year); (1–49 place per take deaths killings sional occa- where 2) countries occur; killings unlawful or executions noextrajudicial where 1) countries categories: three into countries divides project The as: project the by defined killings, extrajudicial on data country-level tive compara- togenerate attempts Project Data Rights Human Cingranelli–Richards the Nevertheless, or others (Cingranelli and Richards, 2008, p.7). 2008, Richards, and (Cingranelli others or prisoners, detainees, suspects, criminal against whether state the of agents other or forces, rity secu- police, the by force lethal of use excessive and illegal, deliberate, the from result may ings p.132). 2008, Secretariat, Declaration (Geneva institutions relevant of legitimacy the of afear and practices reporting about awareness of alack or them report to isnobody there that reason simple the for unreported, go frequently killings unlawful and executions extrajudicial groups groups private by murder include They law. of process due without officials government by killings if instigated by government. These kill- These government. by instigated : examines only deaths deaths only examines Global Burden Global Global Global 77 2 3 4 5 1 TRENDS AND PATTERNS 78 Table4 2. Killings during legal interventions: A sample

Country* Latest available year Intentional homicide Killings during legal interventions

Argentina 2008 2,305 (52)

Colombia 2009 15,817 (363)

Costa Rica 2009 525 (1)

Croatia 2009 58 1

Czech Republic 2009 79 1

Finland 2009 99 1

Honduras 2008 4,473 (54) ARMED VIOLENCE 2011 VIOLENCE ARMED

of India 2009 32,369 644

Netherlands 2009 145 2

Nigeria 2008 1,956 967

Spain 2009 314 1

Sweden 2009 51 1

United States 2009 15,241 454 GLOBAL BURDEN BURDEN GLOBAL Uzbekistan 2009 619 2

Venezuela 2009 13,985 2,685

Total sample 88,036 4,759

Global total 396,000 21,000 (rounded)

civilians by law enforcement officials, or killings during legal interventions is predictably scarce of law enforcement officials on duty). Often, these and often anecdotal. deaths are referred to as deaths as a result of legal While severely underreported, killings during actions. The Special Rapporteur routinely encoun- legal actions appear to be surprisingly routine. ters what are effectively ‘intentional homicides’ in According to a recent report from Jamaica, one in which ‘police shoot to kill alleged criminals with- five killings in the country is committed by secu- out resort to other appropriate measures’ (UNGA, rity forces, yet these do not appear in the official 2010b, p. 8). The lack of respect for principles on national homicide record (Sunday Herald, 2011). the use of force and firearms while arresting a A study on the criminal justice systems in Jamaica suspect or the indiscriminate force in a riot- and the Dominican Republic estimates that control context are ‘often due to poor training, approximately 200 police killings took place in inappropriate “use of force” regulations and 2007 (Foglesong and Stone, 2007, p. 18); in the resource deficiencies’ (p. 8). Owing to significant Dominican Republic, there have been reports of political sensitivity, reporting on deaths occurring up to 58 police killings per month (p. 17). Table 2.4 interventions. Reports published by the non- the by published Reports interventions. legal during occurring anddeaths cial killings extrajudi- andcounting inidentifying inherent es challeng- the highlights example Venezuela The killings. include these already homicides intentional on statistics national the and Honduras, Rica, Costa Colombia, Argentina, such as countries, inseveral that shows It countries. inselected tions interven- legal during killings of examples provides ( resistance during killed Humanos) were Derechos en civilians 2,685 that Educación–Acción de reports Venezolano (Programa Provea organization non-governmental Venezuelan The Venezuela: vio of aresult as killed 619people and Y35) (ICD-10: X85–Y05) (WHO, n.d.c). (ICD-10: interventions legal during killed persons two lists DMDB The Uzbekistan: offic enforcement law other or police involve that killings as chapter this 2of endnote in defined narrowly qua not do they interventions, officers, legal enforcement law of involvement the Without either. manslaughter non-negligent or murder as counted not tabl the in included Not (the citizens killing of homicides a the by commission during private felon, justifiable of a 2010d). felony, citizen), by a alt private (FBI, 2009 in duty on killed feloniously were officers enforcement law 48 additional An 2010e). non- and murder of cases 15,241 are dete 2010c) since the killings these (FBI, manslaughter negligent from separately listed are killings These 2010a). (FBI, duty) of line the in officer enforcement a law 406 reports Bureau of Investigation Federal United The States: n.d.c). (WHO, X85–Y05) intervention one legal Sweden: during killed DMDB person lists The n.d.c). (WHO, X85–Y05) oneduri killed Spain: person lists DMDB The year same that policemen 110 10). p. killed n.d., robbers (CLEEN, armed that and 2009 in robbers 857 killed police that reports Foundation CLEEN The Nigeria: o aresult as killed people 164 and Y35) assaults (ICD-10: (ICD-10: X85–Y05) (WHO, n.d.c). interventions legal during killed persons two lists DMDB The Netherlands: The 567). (p. accident acar of aresult as died who those duty—excluding on killed als otherwise they 563); p. officers 2009, (NCRB, firing police in killed policemen 131 and civilians 184 list India in statistics crime The India: 3). p. 2009, (IUDPAS, figures homicide intentional homicides 54 as a ofall result lists police of action, which in are included Honduras in th Observatory Violence The Honduras: n.d.c). (WHO, X85–Y05) Y35) and 99 people as a killed (ICD-10: of result violent interventions legal oneduring killed person lists DMDB The Finland: v of aresult as killed people 79 and (ICD-10: Y35) X85–Y05) (WHO, n.d.c). (ICD-10: interventions legal during killed person one lists DMDB The Republic: Czech and Y35) (ICD-10: interventions legal during killed 58 X85–Y05) people n.d.c). as killed a (ICD-10: of result (WHO, violent assaults person one lists WHO the of (DMDB) Database Mortality Detailed European The Croatia: du on while perpetrators by ( committed homicides of acategory include 2009 for statistics homicides intentional The Rica: Costa ( seguridad forces de yorganismos pública security fuerza por official by abatidos killed were who elements’ ‘anti-social 363 the include statistics homicide intentional The Colombia: en servicio 4 ( homicide intentional of cases 52 The Argentina: notes * Country Table en cumplimiento del deber 2. ). Another 50 homicides were committed by police officers who were not on duty or by other security personnel (MJSDH, 2008, p. 8 p. 2008, (MJSDH, personnel security other by or duty on not were who officers police by committed were homicides 50 ). Another

Notes :

) (PJCR, 2009a, p. 34). p. 2009a, ) (PJCR, ng legal interventions (ICD-10: Y35) and people 314 as killed a (ICD-10: of result violent a ng interventions legal homicidio doloso ) (Espino-Duque, 2010, p. 75). p. 2010, ) (Espino-Duque, resistencia justifiable homicides byjustifiable a law enforcementagent (the killing of ) include police officers who committed killings while being on duty ( duty on being while killings committed who officers police ) include rmined through law enforcement investigations to be justifiable law investigations enforcement through rmined s (ICD-10: Y35) ands 51 (ICD-10: people as killed a of result violent a ) in 2009 (Provea, 2010, p. 418). p. 2010, (Provea, 2009 ) in ‘executions’ by security providers (on or off duty) duty) (on oroff providers security by ‘executions’ annual hundred several on data provides also Provea However, p.4). 2010, (INCOSEC, homicides intentional on statistics in government included not are deaths these Security, Citizen on Institute Research Venezuelan tothe p. 418). According 2010, (Provea, authorities’ ‘resisting while year each killed were people 2,685 anestimated 2009 in that indicate Provea organization governmental o report 329 police police 329 o report hough they are are they hough assaults (ICD-10: ssaults (ICD-10: ials (FBI, 2010b). (FBI, ials ssaults (ICD-10: e are the 261 261 the e are iolent assaults iolent assaults lent assaults lent assaults e overall e overall a felon by by a felon f violent antisociales antisociales (FBI, (FBI, lify as lify ty policía ). 79 2 3 4 5 1 TRENDS AND PATTERNS 80 who open fire with the intent to kill; a lack of agencies are requiring a more determined focus information makes it impossible to determine on data and analysis to shape programme whether they are included in the intentional design, implementation, and monitoring and homicide statistics (Provea, 2010, p. 417). evaluation. It is unsurprising that statistics— including data related to lethal violence—are A review of this sample of 15 countries indicates profoundly shaped by political and economic that an additional 4,759 deaths would have to be interests. This is especially the case if reputa- added to the number of intentional homicides for tions and the flow of aid dollars are even partly states that do not already include them in the determined by factual evidence. Specialists homicide statistics. The number of intentional agree that these challenges are particularly pro- homicide victims would thus increase by about nounced in relation to trends in homicide, conflict

ARMED VIOLENCE 2011 VIOLENCE ARMED 5.4 per cent if killings during legal interventions

of deaths, and other forms of lethal violence (Andreas were included. When applied to the estimated and Greenhill, 2010). 396,000 intentional homicides, this percentage implies that at least 21,000 people are killed The chapter has also demonstrated that existing during legal interventions every year. administrative data on lethal violence must be cautiously and critically interpreted. On the one hand, reported increases in particular trends—such as intentional homicide or unlawful killings—may GLOBAL BURDEN BURDEN GLOBAL Conclusion imply a genuine escalation of armed violence. On The production and dissemination of reliable, the other, such increases can also imply increased comprehensive, and cumulative data is essential faith or trust of citizens in government institutions to promoting a better understanding of and more and therefore increased reporting rather than appropriate responses to trends and patterns changes in the underlying phenomenon itself.35 of lethal violence. This chapter has taken an What is more, decreases in violent mortality could important step towards providing such a picture, imply improvements in health care provision, through a careful integration of data from diverse policing, or other unrelated phenomena. sources and a holistic approach to counting lethal violence. Yet the gaps in the data are many, allow- This chapter presents estimates that will continue ing for only a partial picture. A more accurate to be refined and enhanced over time. By draw- description of the overall global burden of lethal ing attention to the most violent contexts world- violence will require continued commitment to wide, and to the importance of armed violence in building global and national administrative and so-called non-conflict settings, it widens the lens analytical capabilities, legal frameworks, and for policy-makers, practitioners, and research- political will to present the facts on the ground ers. While the overall number of people dying (Harrendorf, Heiskanen, and Malby, 2010). in armed conflicts is at historic lows, in several The international development community appears regions the burden of armed violence remains to be recognizing the importance of evidence frighteningly high, with ripple effects on the as a driver of effective policy and programming. prospects for local, regional, and global security A growing number of multilateral and bilateral and development. Seetheonline methodological annexe at 3 Killings duringlegal int 2 Ontherelationship betweenlethality of gunshot wounds 1 Endnotes CMR Abbreviations WHOMDB WHO UNODC UN-CTS UCDP NCTC IRC IISS ICD GTD DWI DRC DMDB CRED International Rescue Committee Rescue International International Classification of of Classification International Dirty war index war Dirty International Institute for Strategic Strategic for Institute International Democratic Republic of the Congo the of Republic Democratic Global Terrorism Database Terrorism Global

World Health Organization Health World www.genevadeclaration.org. ofﬁcials by civilians duringlegal actions. actions, orthekillings of police or otherlaw enforcement quelling disturbances, maintaining order, orotherlegal ment ofﬁcials inthecourse of arrestinglawbreakers, killings of civilians attribut (2002) andJarmanet and medical technology, see,forexample, Harris et al. National Counterterrorism Center Counterterrorism National

Uppsala Conflict Data Program Data Conflict Uppsala European Detailed Mortality Mortality Detailed European United Nations Office on Drugs Drugs on Office Nations United Survey of Crime Trends Crime of and Survey World Health Organization’s Organization’s Health World Mortality Database Mortality and Crime Systems Justice Criminal of Operations Studies Database Disease of Disasters Epidemiology the on Research for Centre rate mortality Crude al. (1999). al. erventions aredefinedhereas the ed topolice orother law enforce- 5 The database includes the189UNmemberstates (all but 15 TheLesser Antilles re 14 For details, seetheonline methodological annexe at 13 See Obermeyer, Murray, andGakidou (2008)andAlkhuzai 12 TheICDc 11 TheICDc 10 See‘Deaths Estimates f 9 Foradetail 8 For adefinitionof ‘main armedconflict’, seetheonline 7 Authorcommunication with theInternational Institute for 6 The1996AdHoc For moredetails, seeAdHoc Committee (n.d.). 5 4 Y10–Y34 (WHO, n.d.a). of any country, orthedelimitation of its boundaries. of any opinion of the editors concerning the legal status countries and theirnames does Region andMicronesia Region). The presentation of tries, 15wereamalgamated into2regions (Lesser Antilles Ireland, and Northern Scotland). Of thetotal of 199 coun- Rico,Puerto Reunion, andtheUK’s England and Wales, members (Anguilla, Bermuda, Guam, territories intheGBAV 2011database), and 10 non-UN San Marino, Tuvalu, andtheUK, Marshall Islands, Nauru,andPalau. Federated States of Micronesia, Guam, Kiribati,the and Tobago. The Micronesia Region is comprised of the Nevis, St. Lucia, St. Vincent andtheGrenadines, Trinidad Barbuda, Barbados, Dominic www.genevadeclaration.org. et al.(2008). (WHO, n.d.a). injuries andoperations duetolegal intervention of war States’ in WHO (n.d.b). see Gilgen and Tracey (2011,annexe 2). methodological annexe at www Strategic Studies, 19 May 2011. 2005,pp.(UNGA, 9–10). See alsoCNS (2010). organization todoorabstain from doingany act’ lation, ortocompel aGovernment oraninternational conduct, by its nature orcontext, is tointimidateapopu- result inmajor ec paragraph 1(b)of resultingorlikely thepresent article to to property, places, facilitie infrastructure facility ortotheenvironment; or(c)Damage or government facility, apublic transportation system, an private property, includingaplac toanyinjury person;or(b) Serious damage topublic or fully andintentionally, causes: (a)Death orserious bodily present Convention if that person,by any means, unlaw- person commits anoffence within themeaning of the 2of thedrafttion. Presentedin2005,Article reads: ‘Any lassific lassific ed anal ation for‘eventof ation for‘ onomic loss; whenthepurposeof the Committee haspr ysis ofdata gion includes Anguilla, Antiguaand or 2008by Cause for WHO Member assault’ is X85–Y09; itexcludes s orsystems referredtoin a, Grenada, St. Kitts and sources onarmed violence, .genevadeclaration.org. not imply theexpression which is split intothree e of public use,a State undetermined oduced defini- adraft Hong Kong, Palestine, intent’ is 81 2 3 4 5 1 TRENDS AND PATTERNS 82 16 The 16 countries are Afghanistan, Burundi, Chad, Central 28 The violent death rates of Georgia and Lebanon were African Republic, Colombia, Côte d’Ivoire, the Democratic below the threshold of 10 per 100,000, both in 2004 and Republic of the Congo, Ethiopia, Iraq, Lebanon, Palestine, 2009. Nevertheless, both countries witnessed violent the Russian Federation, Somalia, Sri Lanka, Sudan, and death rates of more than 10 per 100,000 in a given year Uganda. between 2004 and 2009 (GBAV 2011 dataset). They are 17 Post-conflict settings are defined according to the UCDP discussed in detail below. terminology of ‘termination of the use of armed force’. 29 For a more detailed discussion of post-conflict violence, This occurs when the incompatibility is solved either by see the first edition of the Global Burden of Armed Violence an agreement or by a victory; when a party ceases to exist; (Geneva Declaration Secretariat, 2008, ch. 3). or when the use of armed force does not meet the 25 30 The UN Panel of Experts on Accountability in Sri Lanka battle-related deaths criteria (Kreutz, 2010). reports an estimated 7,721 civilians killed between August 18 The 55,000 direct conflict deaths figure also includes 2008 and May 2009. In the limited surveys that were the 1,100 direct conflict deaths that occurred in Jammu- carried out in the aftermath of the conflict, a high per- Kashmir; Central Asia (Fergana Valley, which is split centage of people reported dead relatives. A number of

ARMED VIOLENCE 2011 VIOLENCE ARMED between Kyrgyzstan, Tajikistan, and Uzbekistan); and the credible sources have estimated that there could have of conflict between Armenia and Azerbaijan. In the analysis been as many as 40,000 civilian deaths in a similar time of the national violent death rates, however, these figures period (UNSG, 2011, pp. 40–41). are not included because they cannot be attributed exclu- 31 HSRP (2010); Murray et al. (2002); Obermayer, Murray, sively to one country. and Gakidou (2008); Spagat et al. (2009). 19 The statistics on intentional homicides in Colombia include 32 ‘Situations where the perpetrator is reckless or grossly extrajudicial executions and unlawful killings, as well as negligent, or where the perpetrator kills in self-defence, people killed as a result of the armed conflict. are therefore usually excluded from the category of inten- 20 WHO estimates of 2008 ‘war deaths’ in Afghanistan (83.6 tional homicide’ (Geneva Declaration Secretariat, 2008,

GLOBAL BURDEN BURDEN GLOBAL per 100,000) and Somalia (129.0 per 100,000) are much p. 68). higher than those used in the GBAV 2011 dataset. 33 Author communication with the Australian Bureau of 21 The regions are labeled according to geographical regions Statistics, 13 May 2011. defined by the UN Statistical Division (UNSD, n.d.). 34 See also PEE (n.d.). 22 Saudi Arabian military forces entered the war in Sa’dah in 35 See Marenin (1997) and Soares (2004a; 2004b). early November 2009, while both the United States and the United Kingdom have supported the development of Yemen’s counterterrorism capacity (Aljazeera.net, 2009; The New York Times, 2009; BBC, 2010). 23 The UCDP battle-related deaths database lists an estimated 1,491 victims of the 1994 civil war in Yemen (best estimate). See UCDP (n.d.c). 24 Crimes in Yemen and elsewhere are generally formally recorded only once they are ‘detected’, such as once a suspect has been identified and a prosecutor has received the relevant file. 25 The figures do not include direct conflict deaths that occurred in the disputed area of Jammu–Kashmir. 26 Part of this fluctuation may be due to a greater reliance on WHO estimates in 2004 and again in 2008 and 2009; these years have slightly higher rates than do years for which only reported administrative data was available. 27 Criminal justice data on homicides in Latin America and the Caribbean is relatively accessible and comprehensive; an analysis of trends in armed violence in these regions can be conducted across all countries. Most countries in Sub-Saharan Africa were excluded. CSO (Central Statistical Organisation of Yemen). n.d. ‘Central CRED (Centre forResearch on ofthe Epidemiology Disasters). Collier, Paul andAnke Hoeffler. 2004. Coghlan, Benjamin, et al.2007. CNS (Center forNonproliferation Statistical Organisation.’ Countries 2011. Health inComplex Emergencies: Trends in8African Economies, University of Oxford. Paper No.2004-10. Oxford: Centre forthe Study of African economic Determinants ofHomicideandCivil War International Rescue Committee. Republic ofCon pdfs/intlterr.pdf> Crime Statistics inNigeria 2008. short_descriptions.pdf> The Cingranelli– New York: litics/8437757.stm> . Working

Rate per100,000 Inhabitants, 1990–2009.’ Homicide by Weapon, Private Citizen, 2005–2009.’ offenses/expanded_inform ment, 2005–2009.’ DCPJ. Geneva: Geneva Declaration Secretariat. toring Systems: Contributin Violence. Geneva: Geneva Declaration Secretariat. Vol. 19, pp. 69–91. tional Terrorism.’ The Political Struggle behindtheLegal DebateonInterna- Series. RWP07-019. Cambridge, University. MA:Harvard Dominican Republic. Crime andViolence: Lesson Contribution ofCriminalJustice Systems to the Control of Arms Survey. Victims. Unpublished background paper. Geneva: Small 12.3%20crime%20cases%20recorded.pdf> 2011: CrimeCases Recorded. Suva: FIBOS. Statistics onCrimeand the CriminalJustice System: Leiden Journal ofInternational Law ov/ucr/ciu The DeﬁnitionofTerrorism andIts Faculty Research Working Papers y Killed.’ g Evidence to Programming. ation/data/shrtable_14.html> s from Jamaica and the s2009/offenses/expanded_ s2009/offenses/expanded_ ernational Public Enemy: Key Statistics March Measuring the Criminalité Criminalité , Vol.52, . Paris: ,

83 2 3 4 5 1 TRENDS AND PATTERNS 84 GoP (Government of Pakistan). 1998. ‘Population by Province/ Jarman, Brian, et al. 1999. ‘Explaining Differences in English Region since 1951.’ Islamabad: Population Census Organi- Hospital Death Rates Using Routinely Collected Data.’ zation. British Medical Journal, Vol. 318, pp. 1515–20. GST (Ghana Statistical Service). 2005. Ghana in Figures. Accra: Kenya Police. 2010. Annual Crime Report for the Year 2010. GST. resources/Crime%20report%202010.pdf> GTD (Global Terrorism Database). n.d. ‘Global Terrorism Database.’ Krause, Keith. 2009. ‘Beyond Definition: Violence in a Global Perspective.’ Global Crime, Vol. 10, No. 4, pp. 337–55. Harrendorf, Stefan, Markku Heiskanen, and Steven Malby, eds. Kreutz, Joakim. 2008. UCDP One-sided Violence Codebook. 2010. International Statistics on Crime and Criminal Justice. (HEUNI) and United Nations Office on Drugs and Crime —. 2010. ‘How and When Armed Conflicts End: Introducing the Harris, Anthony, et al. 2002. ‘Murder and Medicine: The Lethality UCDP Conflict Termination Dataset.’ Journal of Peace

ARMED VIOLENCE 2011 VIOLENCE ARMED of Criminal Assault 1960–1999.’ Homicide Studies, Vol. 6, Research, Vol. 47, No. 2, pp. 243–50.

of No. 2, pp. 128–66. Manwaring, Max. 2007. A Contemporary Challenge to State Hicks, Madelyn Hsiao-Rei, et al. 2011. ‘Violent Deaths of Iraqi Sovereignty: Gangs and Other Illicit Transnational Criminal Civilians, 2003–2008: Analysis by Perpetrator, Weapon, Organizations in Central America, El Salvador, Mexico, Time, and Location.’ PLoS Medicine, Vol. 8, No. 2, pp. 1–15. Jamaica, and Brazil. Carlisle: Strategic Studies Institute, Holder, Yvette, et al. 2001. Injury Surveillance Guidelines. Geneva: United States Army War College. HSRP (Human Security Report Project). 2007. Human Security Marenin, Otwin. 1997. ‘Victimization Surveys and the Accuracy Brief 2007. Vancouver: Simon Fraser University. and Reliability of Ofﬁcial Crime Data in Developing

GLOBAL BURDEN BURDEN GLOBAL Countries.’ Journal of Criminal Justice, Vol. 25, No. 6, —. 2010. The Shrinking Costs of War. In HSRP. Human Security pp. 463–75. Report 2009. Vancouver: Simon Fraser University, part II. MJSDH (Ministerio de Justicia, Seguridad y Derechos Humanos). IBC (Iraq Body Count). n.d. ‘Documented Civilian Deaths from Violence.’ 2008. Reporte Anual del Sistema de Alerta Temprana: Homicidios. Buenos Aires: Dirección Nacional de Política iCasualties.org. n.d. ‘Iraq Coalition Casualties: Fatalities by Criminal, MJSDH. Argentina2008_homi.pdf> IISS (International Institute for Strategic Studies). n.d. ‘Armed Muggah, Robert. 2011. ‘Measuring the True Costs of War: Conflict Database.’ Consensus and Controversy.’ PLoS Medicine, Vol. 8, INCOSEC (Instituto de Investigaciones de Convivencia y Seguridad No. 2. February. Ciudadana). 2010. La Situación de la Seguridad en Murray, Christopher, et al. 2002. ‘Armed Conflict as a Public Health Venezuela. NCRB (National Crime Records Bureau). 2009. Crime in India INEGI (Instituto National de Estatistica y Geographia). n.d. 2009. New Delhi: Ministry of Home Affairs. ‘México en cifras.’ INSD (Institut National de la Statistique et de la Démographie). NCTC (National Counterterrorism Center). 2009. 2008 Report on 2010. Annuaire Statistique 2009: Burkina Faso. Terrorism. Washington, DC: NCTC. —. 2010. 2009 Report on Terrorism. Washington, DC: NCTC. INSEC (Informal Sector Service Centre). n.d. ‘Trend Analysis.’ gov/FederalDiscoverWITS/index.do?N=0> ISTAT (Istituto Nazionale di Statistica). 2008. Delitti denunciati New York Times, The. 2009. ‘Now Yemen.’ 30 December. istat.it/Nemesis/jsp/NewDownload.jsp> Obermeyer, Ziad, Christopher Murray, and Emanuela Gakidou. IUDPAS (Instituto Universitario en Democracia, Paz, y Seguridad). 2008. ‘Fifty Years of Violent War Deaths from Vietnam to 2009. Mortalidad y otros: Boletín Enero–Diciembre 2008. Bosnia: Analysis of Data from the World Health Survey Tegucicalpa: Observatorio de la Violencia. Programme.’ British Medical Journal, Vol. 336, pp. 1482–86. PJCR (Poder Judicial deCosta Rica). 2009a. AnuarioPolicial de SESNSP (Secretariado Ejecutivo del Sistema Nacional de Seguridad Seligson, Mitchell andJohnBooth.2010. ‘Crime, Hard Times, SATP (SouthAsian Terrorism Portal). 2011.‘Pakistan Assessment SAPS (SouthAfrican Police Service). 2010. Culpable Homicide Rodgers, Dennis. 2010. ‘Slum Wars of the21st Century: Gangs, —. 2002. ‘TheFour Waves of Modern Terrorism.’ InAudrey Cronin David. 1984.Rapoport, ‘Fear and Trembling: Terrorism in Three Provea (Programa Venezolano deEducación–Acción enDerechos Povey, David, et al.2009. PN (Policía Nacional). 2008. —. 2009b. AnuarioPolicial deCosta Rica, 2009: Homicidios Pézard, Stéphanie and Savannah de Tessières. 2008. ‘ PEE (Project onExtrajudicial Executions). n.d. Website. New York: Policial_%202009/elemen judicial.go.cr/planificacion/Estadisticas/Anuarios/Anuario_ Costa Rica, 2009: Homicidios Dolosos Burundi. Geneva: Geneva Declaration Secretariat. est uneautre guerre’: Analysedela violence arméeau 2009. Mexico City: SESNSP. Pública). 2009. IncidenciaDelictiva Nacional y por Estado April, pp. 123–35. and Discontent.’ index.htm> 2011.’ Nicaragua. Estadisticas/Anuarios/Anuario_Policial_%202009/ Culposos University School of Law. Center forHuman Rights andGlobal Justice, New York . Caracas: Provea. .

Situación delos Derechos Humanos en Homicides, Firearm Offences and Anuario Estadístico 2008 American Political Science Review, Attacking Terrorism: Elements of . HomeOffice Statistical Bulletin. tos/principal.html> stics/reports/crimestats/ e 2009/Septiembre . SecretariadoEjecutiv org/application/media/61%20Comm%20HR%20SR%20 22 December 2004. news&Itemid=671> jamaica-committed-by-security-forces&catid=1:latest- content&view=article&id=711:one-in-five-killings-in- Committed by Security Forces.’ 16January. No. 6,pp. 934–50. of Contestation.’ Development Economics, Vol. 73,No.1,pp. 155–84. Vol. 52, No.4,pp. 851–71. the International Differences inCrime Rates.’ (Jamaica). 2011.‘OneinFiveKillings inJamaica ated_deaths_dataset/> Journal ofConflict Resolution Homicides, Firearm Offences andInti- o/Incidencia_Delictiva_Nacional_ Economic Development andCultural Including the Questions ofDis- ided_violence_dataset/> ting as aMeasure of Institu- . London: HomeOffice. . E/CN.4/2005/7of Journal of , Vol.53, .’ 85 2 3 4 5 1 TRENDS AND PATTERNS 86 UNGA (United Nations General Assembly). 1972. Resolution —. n.d.c. ‘European Detailed Mortality Database (DMDB).’ 3034: Measures to Prevent International Terrorism which Copenhagen: Regional Office for Europe, WHO. Endangers or Takes Innocent Human Lives or Jeopardises Fundamental Freedoms, and Study of the Underlying Zinecker, Heidrun. 2008. Violence in a Homeostatic System: The Causes of those Forms of Terrorism and Acts of Violence Case of Honduras. PRIF Report No. 83. Frankfurt: Peace which Lie in Misery, Frustration, Grievance and Despair, Research Institute Frankfurt. and which Cause Some People to Sacrifice Human Lives, Including Their Own, in an Attempt to Effect Radical Changes. 18 December. —. 1996. Resolution 51/210. A/RES/51/210 16 January 1996. —. 2005. Letter Dated 3 August 2005 from the Chairman of the Sixth Committee Addressed to the President of the General ARMED VIOLENCE 2011 VIOLENCE ARMED Assembly. A/59/894 of 12 August 2005. of —. 2010a. Report of the Ad Hoc Committee Established by General Assembly Resolution 51/210 of 17 December 1996. A/65/37 of 12–16 April 2010. —. 2010b. Report of the Special Rapporteur on Extrajudicial, GLOBAL BURDEN BURDEN GLOBAL Summary or Arbitrary Executions. A/HRC/14/24 of 20 May 2010. —. 2011. Report of the Working Group on Enforced or Involuntary Disappearances. A/HRC/16/48 of 26 January 2011. UNODC (United Nation Office on Drugs and Crime). n.d. UNODC Homicide Statistics. UNSD (United Nations Statistics Division). n.d. ‘Composition of Macro Geographical (Continental) Regions, Geographical Sub-regions, and Selected Economic and Other Groupings: Geographical Region and Composition.’ UNSG (United Nations Secretary-General). 2011. Report of the Secretary-General’s Panel of Experts on Accountability in Sri Lanka. New York: United Nations. 31 March. WHO (World Health Organization). n.d.a. International Statistical Classification of Diseases and Related Health Problems, 10th Revision. —. n.d.b. ‘Health Statistics and Health Information Systems: Disease and Injury Country Estimates.’ A the wider institutional environment that enables, enables, that environment institutional wider the and violence, armed of instruments the motives, andtheir violence armed of perpetrators the consequences), the suffer that andsocieties ties communi- wider andthe victims immediate (both violence armed by affected people tothe tion atten- draws lens violence armed The APPROACH). INTEGRATED 2009; (OECD, purpose very this for ‘lens’— violence armed framework—the ceptual acon- proposes andDevelopment Co-operation Economic for Organisation The categories. arate sep- into violence homicidal disaggregating is by suchinformation obtaining tofacilitate way One andinjury. death intentional reduce andeffectively prevent that interventions andimplementing to designing iscentral factors these on Information violence. tohomicidal deterrents topotential point sarily neces- it does Nor circumstances. what or under how, where, whom, orby killed, isbeing exactly who reveal not may data relevant others, than dangerous more tobe known are areas some While policy-making. inevidence-based use its undermine that gaps knowledge leaves routinely it violence, armed of forms certain for a proxy as serves data Yet homicide while collection. revenue ultimately, and, tourism, investment, lost of aresult as suffer often rates homicide high exhibit that andcities Countries insecurity. incidence of homicide a fair index of of index afair homicide of incidence the consider public general and the world the round Characteristics of Armed Violence Armed of Characteristics Chapter

Three , politicians, investors, investors, , politicians,

The main findings of the chapter are: chapter the of mainfindings The data. justice andcriminal enforcement law on based lence state the explores context the on data available examining by gap information this addressing towards step afirst takes chapter This violence. homicidal of and characteristics context the concerning details additional provide disaggregation requires lens the view, of point astatistical From violence. of acts andrepeated specific of extent and nature the about information vital capture to fail may rates homicide monitoring on sively exclu- rely who andplanners decision-makers that suggests It homicide. intentional such as indicator, aggregate asingle of limitations the to attention draws lens violence armed The violence. non-lethal and andtolethal violence armed conflict-related andnon- conflict- toboth applied be can lens the principle, In violence. armed against, or protects Homicide rates related to related rates Homicide to related homicides of proportion The The to vary according to geographical context. togeographical according to vary tend to be higher in countries with greater greater with incountries higher to be tend orEurope. Asia of in those than America andSouth inCentral countries or organized crime situational context in which homicides occur. It also occur. also It homicides inwhich of ‘total offence data’ in order to inorder data’ offence ‘total of response response is significantly higher in higher issignificantly of homicide events tends tends events homicide of to non-conflict vio- to non-conflict robbery ortheft robbery gangs

87 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 88 income inequality, including countries located A nexus appears to exist between high homi- in the Americas. cide rates, a high proportion of homicides The proportion of homicides related to intimate committed with firearms, and a low proportion partners or the family represents a significant of cases solved by law enforcement. Countries proportion of homicides in some countries in showing this combination of factors risk enter- Europe and Asia. ing a spiral of increasing violence and impunity. The relative weakness of a country’s rule of Measuring the effectiveness of the criminal law is broadly linked with higher overall homi- justice system response to homicide and Photo T Members of a citizen vigilante group cide rates. Yet particular challenges—such as crime involving armed violence requires more patrol the streets of gang activity, a history of conflict, or high than a simple calculation of conviction rates Barcenas, Guatemala income inequality—may lead to high homicide or ‘case attrition’ rates. The deterrent effect

ARMED VIOLENCE 2011 VIOLENCE ARMED City, September 2007. rates even in societies with comparatively of police and justice institutions is also linked of © Carlos Javier Ortiz/ Laif/Redux strong institutions and rule of law. to public perception and the quality of justice. GLOBAL BURDEN BURDEN GLOBAL according to: according cide events. homi- tocharacterizing approaches many also are homicide’, there an‘intentional constitutes what of definitions legal different are there as Just homicide. of anddrivers context the of understanding toincrease inorder measures performance system justice criminal effective of development inthe investment greater for need the highlights It death. violent on data to collecting approaches varying widely the considers chapter This system. justice criminal the of responses andthe deaths violent surrounding stances circum- specific the toassessing comes it when compounded are challenges These enterprise. isacomplex rates homicide national on data of andinterpretation collection, identification, the confidence may figures, be placed in these of degree acertain While people. 100,000 per arate as population—usually the of proportion a as homicide present frequently media the and groups, non-governmental Governments, homicide Contextualizing the characteristics of the offence, such as such as offence, the of characteristics the night); evening, (daytime, offence of time the busi- street, (home, event the of location the (intimate relationship victim–offender the the threatened or actual use of a weapon aweapon of use oractual threatened the demo- on based andperpetrators victims (sharp object, blunt object, firearm [by type]); type]); [by firearm object, blunt object, (sharp status); marital income, education, race, age, (sex, graphics membership; organg alcohol, drugs, of involvement the rural); urban, premises, ness victim); tothe unknown perpetrator acquaintance, friend, member, family partner, 1 These include classifying events events includeclassifying These countries to classify homicides underlines the the underlines homicides toclassify countries indifferent used categories of diversity sheer The perpetrators. suspected ing andlocating identify- of likelihood andthe evidence, forensic of and confidentiality availability the toness talk, willing- witnesses’ adeath, surrounding stances circum- the regarding clarity of degrees varying have will cases homicide Suspected nationally. andcross- acountry within both considerably, vary can recorded detail of andlevel quality the information, more offer generally records police 1999). While (Riedel, elements and situational victim the on information limited only reveal may and offender the of characteristics the on data include usually not do they but death, of cause andprobable location, age, sex, race, victim’s a on details provide may certificates death ners’ andcoro- records examination medical Likewise, comparison. cross-national tofacilitate designed than rather needs, local tospecific tailored tobe tend context event the of categorization of methods isavailable, it Where statistics. national of part form necessarily not and need purposes oroperational investigation for only cumstances. andcir- elements andoffence offender, victim, for systems andrecording coding systematic to inaddition work, investigative police of level aconsistent require also cide circumstances homi- on statistics comparable level, municipal the at as just national the At toobtain. easy dom issel- people andaffected institutions, actors, instruments, the on information detailed more countries, most for accessible be may homicides of incidence the on data superficial Although any other contemporaneous offences, such offences, contemporaneous other any religious, racial, suchas factors, motivating as robbery or theft. ortheft. robbery as and tensions; or communal 2 Such information may be recorded recorded be may information Such 89 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 90 difficulties in developing a standardized approach offences, including drug trafficking. The group to data collection and analysis. National homicide can be randomly formed for the immediate categories are often developed according to local commission of an offence. It does not need to needs and rarely with international comparison in have formally defined roles for its members, mind. While detailed local typologies may be use- continuity of its membership, or a developed ful for individual study, the identification of generic structure. categories for national and regional comparison The intimate partner or family category includes represents a significant challenge. Successful ‘intimate’ individuals such as current or former translation of national data into cross-national intimate partners; a ‘family’ is defined as categories requires both a bottom-up approach persons living in the same household as the (working with the data that is available) and a top- victim, in addition to blood relatives. ARMED VIOLENCE 2011 VIOLENCE ARMED down approach that defines common standards.

of Robbery refers to the theft of property from a The relatively low response rate to previous data person by using force to overcome resistance collections on homicide typology3 and the recent or by threatening the use of force. Theft involves attempt to develop a European-level classiﬁcation the removal of property without the property system of offences to be used in the context of owner’s consent. This category includes crime statistics have shown that striking the right muggings, domestic burglary, and house- 4 balance in this domain is a delicate exercise. breaking as well as theft of a motor vehicle.

GLOBAL BURDEN BURDEN GLOBAL This chapter draws from three categories of hom- The abovementioned categories do not exhaust icide, derived from a review of available national the range of possible homicide situations and data and the requirements of developing—as far are not strictly mutually exclusive. Nevertheless, as possible—comparable definitions for cross- they do reflect commonly used typologies and national comparison. Such an approach treats capture the predominant contexts in which non- the homicide incident holistically, as a complete conflict lethal violence may occur (Mohanty, 2004). composite of offender, victim, and offence as In particular, the term ‘gang’ incorporates a wide well as temporal and spatial elements (Miethe range of contexts and groups, including youth and Regoeczi, 2004). Data in this chapter is gangs, street gangs, drug gangs, motorcycle gangs, therefore disaggregated into organized crime skinheads, and prison gangs (Small Arms Survey, and gang-related homicides, robbery- and theft- 2010, ch. 5). The term ‘organized criminal group’ related homicides (based on contextual variables), is defined in the United Nations Convention and intimate partner- or family-related homicides against Transnational Organized Crime; however, (related to the offender–victim relationship). In the definition is diffuse and does not require a addition, the chapter considers the category of developed group structure (UNODC, 2004).5 homicides committed with firearms. The categories are forged on the basis of the following definitions: Gangs may include organized criminal groups whose members act in concert to commit serious A gang is defined as a structured group of offences, yet some organized criminal groups— three or more persons existing for a period such as drug cartels with high levels of organiza- of time and acting in concert with the aim of tion—are not necessarily gangs. This chapter committing one or more serious crimes or presents classifications that are as broad as organized crimeandgan of categories tothe addition In n.d.a.). UNODC, annexe; methodological online (see statistics homicide UNODC of ananalysis on based Europe and Asia, Americas, inthe 26 countries for text con- homicide of categories four shows 3.1 Figure region. geographic by vary and the the rates—both homicide overall to addition that—in suggests data available with incountries context homicide the of Analysis PATTERNS). AND (TRENDS Europe in Western and1.1 inOceania population 3.3 100,000 per to Africa, and27.4 inSouthern America Central in population 100,000 per 29.0 from range rates homicide national average The regions. geographic across significantly vary rates homicide Overall region by homicide Contextualizing na of arange of account totake inorder possible ‘robbery offences’ (Germany). offences’ ‘robbery to related as statistics innational recorded homicides includes category andtheft robbery The (Japan). orfamily’ and‘intimate (Uganda), violence’ States),‘domestic (United orgirlfriend’ boyfriend family, other sister, brother, daughter, son, father, mother, wife, husband, by perpetrated ‘homicide as statistics innational recorded homicides includes category andfamily partner intimate The (Korea). interaction’ and‘gang Republic); (Dominican disputes’ ‘drug Honduras); gangs, and tions) (Jamaica); called togangs linked are that includehomicides records police reviewed from Examples definitions. tional police-recorded data that employs local local employs that data police-recorded tional pandillas situational context context situational robo relacionado con mara relacionado (paid assassina- modalidad sicariato (paid (Argentina), ‘robber’ (India), and (India), ‘robber’ (Argentina), (Panama); ‘drugs’ or ‘gangs’ or‘gangs’ ‘drugs’ (Panama); g-related homicides of lethal violence also also violence lethal of 6 (related to (related use of weapons use of ,

Source categories. These may include homicides related include homicides may These categories. three above the of toone not belonging identiﬁed as positively allhomicides contains category ‘other’ mate andfamily-related partner- homicides homicides andtheft-related robbery- 9 countries inEurope 9 countries inAsia 6 countries Americas inthe 11 countries Figure UNODC elaboration based on UNODC (n.d.a) UNODC on based elaboration : UNODC 3.1

Homicide context by geographic region geographic by context Homicide , and , the , the inti- Legend: Legend: Legend: Intimate partner or or partner Intimate (11) theft or Robbery organized or Gangs Unknown (23) Other (38) or partner Intimate (4) theft or Robbery organized or Gangs Unknown (9) Other (31) or partner Intimate (18) theft or Robbery organized or Gangs Unknown (21) Other (35) family (8) (26) crime family (27) family (6) crime family (28) (14) crime 91 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 92 to ﬁghts, arguments, or sex offences. While of intimate partner- or family-related homicides results must be interpreted with caution due to in Asia and Europe (around 30 per cent). the limited number of countries for which data is It is not currently possible to present an equivalent available, two patterns are particularly prominent: chart for countries in Africa. Even if better data the high proportion of gang- or organized crime- existed, the exercise would still be particularly related homicides in countries of the Americas challenging. While homicide categories, such as (more than 25 per cent) and the high proportion ‘gang- and organized crime-related’, have relatively

Box 3.1 Mob justice in Africa Power and security vacuums are not unusual in post-conﬂict settings. When ARMED VIOLENCE 2011 VIOLENCE ARMED they occur, vigilante groups and mob violence can replace formal state-led of policing and justice:

Disillusionment with the governing authority’s ability to thwart these crimes may result in the population relying on parallel power structures such as militias, which may be seen as able to protect people from the effects of serious crimes (Rausch, 2006, p. 7).

In post-conﬂict Liberia, for example, violence related to community and infor-

GLOBAL BURDEN BURDEN GLOBAL mal justice and policing continue to be reported: ‘Communal and mob violence continue, often emanating from tensions between ethnic groups and communities over land disputes’ (UNSC, 2009). The incidents of vigilante attacks and mob justice—though rare—reveal a lethal potential whose effects can be destabilizing, especially in the absence of strong law enforcement institutions (Small Arms Survey, 2011).

In the aftermath of presidential elections in Kenya in December 2007, for example, violence ﬂared between rival factions of the country’s main ethnic groups—the Kikuyu, Luos, and Kalenjin. In addition to large-scale rioting and clashes between protestors and the police, the post-election period saw a signiﬁcant rise in ‘mob’ or ‘vigilante’ killings, as long-held tensions over land, the economy, and political power spilled into lethal violence. An estimated 1,100 people were killed and some 350,000 displaced in violence following the

election (CIPEV, 2008, pp. 272, 308). XX

Figure 3.2 Vigilante/mob killings as a percentage of homicides in three African countries

Uganda (national, 2008) Tanzania (Dar es Salaam, 2000–04) South Africa (6 survey sites, 2000–05)

0 5 10 15 20

Sources: Uganda Police (2008); Ng’walali and Kitinya (2006); CSVR (2008) Africa, homicides may take place in contexts as as incontexts place take may homicides Africa, in environments post-conﬂict In p.142ff.). 2010, Survey, (SmallArms region African inthe plex com- isespecially groups criminal and organized warlords, local bandits, rebels, insurgents, groups, armed gangs, between nexus the and Europe, Asia, Americas, the across meanings consistent in 2007, to 368 cases in 2008’ (Uganda Police, 2008, p.12). 2008, Police, (Uganda 2008’ in cases 2007, 368 in to cases 184 ‘from increased action mob from killings of number the that suggests report 2008 The available. publicly information such make to countries African © Rodger Bosch/AFP Photo Bosch/AFP © Rodger 2008. May Johannesburg, Park, Reiger in attacks violence. andmob looting, criminal retribution, individual clashes, inter-communal politicized as diverse Photo the in reported are action’ ‘mob by Killings 2006). Kitinya, and (Ng’walali execution of methods common most the as reported were burning and Stoning 2004. to 2000 period the ing dur- deaths suspicious of cent per 12 around at estimated killings, action mob of rate ahigher reveals Tanzania, Salaam, es Dar in records hospital on based Research 2008). (CSVR, dockets homicide the of cent 3per around to attributed positively be can alone actions vigilante homicides; recorded all of cent 7 per around to up add together killings revenge and justice vigilante that ﬁnds sites survey six at records police of examination an Africa, South For Uganda. and Tanzania, Africa, South countries: African three for estimates shows 3.2 Figure islimited. countries African in actions mob or vigilante to relate that homicides of proportion the on data Nonetheless, 2009b). (UN, executions arbitrary or summary, extrajudicial, on Rapporteur Special Nations the of concern the to due least not years, recent in level international the at attention received has justice mob and killings vigilante of issue The Violence: Election Post the into Inquiry of Commission the to According considers one type of homicide common toa common homicide of type one considers analysis in this area highly demanding. highly area inthis analysis makes recording, data for capacity justice criminal andlimited law of rule weak with combined often p. 278). 2008, (CIPEV, camps the within security provide to groups vigilante as to referred iscommonly what in mobilised were men Young matters. other and security services, access to committees representative into themselves organised camps most in persons] displaces [internally

A young boy walks past the wreckage of burnt-out shacks, following xenophobic xenophobic following shacks, burnt-out of wreckage the past walks boy A young 7 This situational complexity, which is which complexity, situational This Annual Police Reports Reports Police Annual 8 Box 3.1 3.1 Box of Uganda, one of the few few the of one Uganda, of United

93 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 94 number of post-conflict or transitioning African unplanned and often represent spontaneous countries, that of vigilante killings or ‘mob and emotion-laden acts (Gillies, 1976). While justice’. men are generally more likely to be victims of homicide than women, women are generally It should be noted that a high proportion of a more likely to be killed by someone they know.9 particular type of homicide in Figure 3.1 does not necessarily correspond to a high rate of that type Yet patterns of intimate partner and family-related of homicide. While countries in Asia and Europe homicides, particularly in Asian countries, are show a comparatively high proportion of intimate complex. Underlying reasons for lethal violence partner- and family-related homicides (around are heavily influenced by local traditions, family 30 per cent), for example, overall homicide rates and social structures, and levels of education. in Asia and Europe are significantly lower than On the one hand, patriarchal societies in Asian ARMED VIOLENCE 2011 VIOLENCE ARMED

of in the Americas. This difference results in corre- countries can present a particular risk factor for spondingly lower intimate partner and family- lethal violence against women who are considered related homicide rates for Asia and Europe than property and whose subjugation is a symbol of male for the Americas. power and status (UNFPA, 2003). On the other hand, Nonetheless, the signiﬁcant proportion of intimate some observers argue that the Asian emphasis partner- and family-related homicides in many on a collectivistic culture, with a high degree of countries in Asia and Europe does highlight the interdependence among partners and within the GLOBAL BURDEN BURDEN GLOBAL importance of this homicide context as a focus family, makes lethal violence less likely to involve for violence reduction initiatives in these regions. an offender and victim with close ties (Salfati and These types of killings are more likely to be Park, 2007; WHEN THE VICTIM IS A WOMAN).

Photo X A man involved in an attempted robbery is transported to hospital following an exchange of gunfire with police, Karachi, June 2010. © Athar Hussain/Reuters Central and South America, are well-known well-known are America, and South Central particularly Americas, in the countries Other A APPROACH). UNIFIED 1.3, Box (see trade drug the for strategic are that andcities inprovinces seen deaths violent of majority the with localized, similarly to be tended have cartels drug toMexican linked Killings p.134). 2010, Survey, (SmallArms deaths violent of proportion greater afar represent homicides gang-related however, and neighbourhoods, of total homicides). total of cent per 5.5 (around level national the at and low constant reasonably stayed States United the in homicides gang-related to2009, 1998 From 1993). and Block, (Block contradictory be can todrugs related are homicides gang-motivated towhich extent the on evidence example, for drugs, illegal of sale and distribution inthe involved been long have gangs USstreet Although toallgangs. central necessarily not are suchphenomena gangs, many among commonplace are andviolence use drug 2010). (Covey, While crime property profitable on focus that groups structured clearly more to ties ethnic along based often associations loose from range example, for States, United the in Gangs Americas. inthe forms many take can crime andorganized togangs related Killings homicide. crime-related and organized gang- particular, categories—in inother falling homicides of number ahigh of result the rather but se, per homicides orfamily partner intimate of rate alow of indicative not isprobably rate this humandevelopment, of levels lower with countries with As cent). 10per under (just low iscomparatively Americas inthe incountries cides homi- andfamily partner intimate of proportion the andAsia, inEurope tocountries contrast In 10 At the level of individual cities cities individual of level the At of countries from South America, inparticular. America, South from countries of inclusion further by affected be may homicides crime-related andorganized gang- of percentage overall the that noted be should It years. ten last inthe increased Tobago significantly have and Trinidad and inJamaica homicides related gang- that highlights 3.2 Box Caribbean. and the America, Central America, North from largely drawn are 3.1 inFigure included and therefore isavailable data which for countries the Overall, groups. other with conflicts inturf killing of modes tosuch resort gangs when unclear, particularly madeis are suchdistinctions onwhich basis The 2008). (Honduras, assassins hired of involvement the homicides), implying total of and homicides) con mara relacionado as separately homicides report Honduras from statistics 2007). Available (Panama, activity gang orgeneral genus gang toaparticular refers term this isunclearwhether it pandillas, although to related as homicides intentional of 19 cent per record Panama from statistics example, for 3.1, inFigure represented Americas inthe countries the Of statistics. inpolice involvement ing gang identify- of andmethods definitions different by complicated isfurther comparison Cross-national 2007).(UNODC, however considerably, vary countries in these deaths violent for responsible are such groups towhich degree the of 2009). Estimates Rodgers, and Muggah, (Jütersonke, respectively States, United the from deportees around and formed neighbourhoods American inCentral originating gangs, of types distinct two are and Honduras, maras, and Rica, andCosta inNicaragua primarily found violence. andgang togangs homes present mostly in El Salvador, Guatemala, Guatemala, Salvador, inEl mostly present (36 per cent cent per modalidad sicariato (36 (under 1 per cent of total total of cent (under 1per Pandillas , 95 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 96 Box 3.2 The homicide context and responses in the Caribbean

Overall, intentional homicide rates in the Caribbean have increased dramatically in recent years. Figure 3.3 shows a signiﬁcant increase of intentional homicide rates for 12 countries in the Caribbean in the ten-year period from 2000 to 2009. In 2009, the average national rate stood at 24.7, which is more than double the rate of 13.2 in 2000.11 Evidence from at least two Caribbean countries—Jamaica and Trinidad and Tobago—suggests that this increase is largely due to an increase in gang-related killings. In order to better understand what is driving these trends, it is essential to understand the dynamics of gangs themselves (Townsend, 2009).

In Jamaica a distinction can be made between ‘corner crews’ (or simply ‘crews’) that hang out on street corners they consider their turf, area ‘dons’ who serve as

ARMED VIOLENCE 2011 VIOLENCE ARMED role models for younger gang members, and highly active criminal gangs (Leslie, of 2010). These groups have different levels of involvement in crime, ranging from petty offences to car-jacking, marijuana or cocaine trafﬁcking, and, in some cases, white- collar crime. Crews often have violent conﬂicts with rival crews, leading to revenge murders and fuelling persistent violence (Covey, 2010). Community responses to gangs in Jamaica are comparatively tolerant, due perhaps to the fact that many

gangs represent a source of income from drug activities for low-income families. XX

Figure 3.3 Homicide rate in 12 Caribbean states, 2000–09 GLOBAL BURDEN BURDEN GLOBAL Homicide rate per 100,000 population

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Source: GBAV 2011 database

Figure 3.4 Homicide context in Jamaica and Trinidad and Tobago, 2000 and 2008 Gang Other

Jamaica – 2000 Jamaica – 2008 Trinidad and Tobago – 2000 Trinidad and Tobago – 2008

0102030405060 Homicide rate per 100,000 population

Source: UNODC elaboration based on UNODC (n.d.a) © Hans Deryk/Reuters lord. drug alleged an of supporters and forces security Jamaican between clashes of victims for Photo A cemetery employee walks away from caskets caskets from away walks employee Acemetery 97 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 98 Figure 3.4 shows an increase in the overall homicide rate in Jamaica Just under 20 per cent of gang-related homicides were reported from just over 30 in 2000 to almost 60 per 100,000 population in as cleared up. 2008. Similarly, the rate in Trinidad and Tobago increased from less Almost 60 per cent of domestic-related homicides and just than 10 in 2000 to more than 40 per 100,000 in 2008.12 Strikingly, over 50 per cent of homicides related to another criminal act the proportion of gang- or organized crime-related killings increased were reported as cleared up. from 14 per cent to 45 per cent in Jamaica and 24 per cent to 69 per cent in Trinidad and Tobago in 2008. In both countries, this rise Clearance rates for ﬁrearm homicides were just 30 per cent, accounts for almost all of the increase in overall homicide. The compared with more than 40 per cent for homicides committed pattern of homicides is tightly linked to the use of ﬁrearms. Guns using any other weapon (JCF, 2010). were used in approximately 80 per cent of homicides in both Jamaica and Trinidad and Tobago in 2008 (UNODC, n.d.a). This triple com- Limited police resources may still be a restraining factor in homi- bination of high overall levels of homicide, a high proportion of cide case clearance. Figure 3.5 shows that, in years when homicide ARMED VIOLENCE 2011 VIOLENCE ARMED

of firearm homicides, and a high degree of gang- or organized crime- rates dropped, clearance rates rose. This suggests either a brief related homicides presents a significant challenge to an effective respite for overwhelmed law enforcement officials and/or a drop state response. in harder-to-solve, gang-related homicide cases in those years.

Figure 3.5 shows that, as homicide rates have relentlessly risen in The crime and security situation in Jamaica and, increasingly, in Jamaica since the 1970s, the police ‘clearance rate’ for reported other Caribbean countries threatens to reach endemic proportions. homicides has gradually fallen, from more than 80 per cent of A recent national victimization survey in Jamaica found that more cases in 1970 to around 40 per cent in 2005. The drop in the clear- than 7 per cent of all respondents had a relative or close friend ance rate may be partly due to an increasing proportion of cases who had been murdered in the past year, and almost 33 per cent GLOBAL BURDEN BURDEN GLOBAL that are difﬁcult to solve. Detailed police statistics reveal the reported having witnessed the murder of a relative or close friend in challenges in solving drug or gang-related homicides. In 2008: their lifetime (Wortley, 2009, p. 24). Analysis of the homicide context and responses indicates that armed violence reduction must begin No drug-related homicides were reported as cleared up by with addressing the issues of gang membership and impunity from the Jamaican police. justice for members of criminal groups who engage in violence.

Figure 3.5 Homicide rate and case clearance in Jamaica, 1970–2005 Clearance rate Homicide rate

Clearance rate (%) Homicide rate per 100,000 population

90 90

80 80

70 70

60 60

50 50

40 40

30 30

20 20

10 10

’70’75’74’73’72’71 ’76 ’77 ’78 ’79 ’80 ’81 ’82 ’83 ’84 ’85 ’86 ’87 ’88 ’89 ’90 ’91 ’92 ’93 ’94 ’95 ’96 ’97 ’98 ’99 ’00 ’01 ’02 ’03 ’04 ’05

Source: UNODC elaboration based on JCF (n.d.) Homicide rate per 100,000 population year available orlatest 2010 inequality, Figure Source 0.5 1.0 1.5 2.0 2.5 3.0 0.40–0.60). (Gini index inequality income higher with tries coun- andten 0.20–0.40) (Gini index inequality income lower with 15 countries for homicide andtheft-related robbery- of rates and average homicide andfamily partner intimate of rates average shows 3.6 Figure inequality. income of degree by clustered are countries when emerge also difference of patterns context, inhomicide variations toregional addition In tions to use violence, particularly once a criminal acriminal once particularly violence, touse tions motiva- heighten may andtension inequality, social inincome, differences Rather, context. homicide across uniform be not may in homicide suchincreases that countries) of group a limited of basis the on (albeit suggests it nuance inthat provides 3.6 ﬁndinginFigure 2002); the Loayza, and Lederman, (Fajnzylber, homicides more tohave tend inequality income higher with tries coun- that time some for recognized been has It members. or family partners intimate against inviolence than theft, or robbery suchas crime, inproperty force lethal of use the on impact agreater tohave appears Intimate partner- and family-related homicide and family-related partner- Intimate homicide theft-related and Robbery- index 0.20–0.40 (15countries) Lower income inequality: Gini : UNODC calculations 3.6

Homicide context and income andincome context Homicide 13 Increasing income inequality inequality income Increasing index 0.40–0.60 (10countries) Higher income inequality:Gini Homicide rate per 100,000 population 100,000 per 34 around of rate had ahomicide which Africa, isSouth inpoint Acase homicides. firearm of proportion ahigh have necessarily not do population 100,000 20 per exceed that rates homicide with Countries police. national by recorded as firearm by committed homicides those of proportion andthe population 100,000 per rate homicide overall the between relationship the demonstrates region, by organized countries, 104 from Data rates. homicide on violence armed the of impact the shows 3.7 Figure Homicide context and use firearm LESS DEVELOPMENT). VIOLENCE, (MORE indicators development and violence homicidal between relationships the of assessment astatistical of findings the analyses Five Chapter upon. embarked been already has ortheft) robbery (suchas endeavour Source 10 20 30 40 50 60 70 80 90 100 year available orlatest 2009 firearms, Figure Americas 20304050607080901000 UNODC elaboration based on UNODC (n.d.a.) UNODC on based elaboration : UNODC 3.7 10 Asia

Homicide rate and proportion of homicides committed with with committed homicides of andproportion rate Homicide Europe Oceania % homicides with firearm instruments of 99 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 100 population in 2009, although less than 50 per cent hand, gun usage in homicide is probably linked to of those homicides were carried out by firearm.14 the comprehensiveness and application of regu- Of the countries where more than 70 per cent of lations (regarding purchase, carrying, licensing, homicides are carried out by firearm, however, weapon type, and access to ammunition). On the some four-fifths have homicide rates of 20 or other, it is also potentially connected to the rela- higher (UNODC, n.d.a). tive presence of gang and organized crime structures that are more likely to misuse weapons than Regionally, countries in the Americas tend to show the general population. Indeed, Figure 3.8 dem- both significantly higher homicide rates and higher onstrates the close links between the proportion proportions of homicides committed with firearms of homicides committed by firearm and the pro- than countries in Asia, Europe, or Oceania. For portion of homicides attributable to gangs or example, the nine Caribbean countries included ARMED VIOLENCE 2011 VIOLENCE ARMED organized crime activity for 2009 or the latest of in their analysis show an average homicide rate available year for 17 countries in the Americas, of around 25 per 100,000 population, with an Asia, and Europe. average of 60 per cent of homicides committed by firearm. Ten countries in South America show Despite the lack of reliable, comparable, cross- an average homicide rate of around 18 per 100,000 national data on the sub-types of gang or organized population, with 60 per cent of homicides com- crime groups responsible for lethal violence in mitted by firearm. These figures stand in stark the Americas, the ubiquitous involvement of fire- GLOBAL BURDEN BURDEN GLOBAL contrast to the average rates of below 5 per 100,000 arms in such killings is well established. Figure 3.8 population for certain countries in Asia and Europe, shows medians and quartiles for the proportion where 22 per cent and 24 per cent of homicides of gang or organized crime homicides in ten are committed by firearm, respectively. As dis- countries with less than 50 per cent of homicides cussed below, this discrepancy may be linked to committed with firearms and seven countries with the higher proportion of gang- or organized crime- more than 50 per cent of homicides committed related homicides in a number of countries in the with firearms. It demonstrates that in the former Americas region, particularly in Central America. Figure 3.7 does not include information from African Figure 3.8 Proportion of gang- or organized countries, for which there is insufficient data. crime-related homicides and firearm use Available data on civilian firearm ownership—part Legend: the outlier (star), the median (horizontal bar), and the upper and lower quartiles (top and bottom of each bar) of the ‘instruments’ element of the armed violence lens—shows no obvious relationship with % gang or organized crime homicides overall homicide rates. It is thus not possible to 50

assess the impact of given numbers of civilian 40

ﬁrearms on the pattern shown in Figure 3.7. 30

Nonetheless, the presence of guns clearly repre- 20 sents an increased risk of lethality of violent 10 encounters. A number of researchers have dem- Countries with less than 50% of Countries with more than 50% of onstrated that weapons in the home constitute homicides with ﬁrearm (n=10) homicides with ﬁrearm (n=7)

more of a health risk than a benefit.15 On the one Source: UNODC calculations lent events. Figure 3.9 shows the percentage of of percentage the shows 3.9 Figure events. lent vio- orenable incombiningtorestrain violence of agents and the violence, of instruments the system), justice andcriminal enforcement law the (including context institutional the of nature interconnected the linkhighlights this lens, lence vio- armed the through Viewed cases. homicide of ‘solvability’ inthe amajorfactor isalso it but groups, criminal ororganized gangs of role the and homicide of levels overall with linked appear inhomicide smallarms of use the does only Not enforcement response Homicide context and the law weapons (Small Arms Survey, 2010, p.115). 2010, Survey, (SmallArms weapons orsteal share ormay anintermediary, or through legally guns purchase may tojoiningagang, prior guns owning of ahistory have may members Gang country. given inany ownership civilian normal of rate the times four around be could ownership gun gang that suggest andWales England, States, United the from data survey on primarily based Estimates smuggling. firearm of ease and the society incivilian availability weapon to general according significantly tovary appears however, firearms, orcollectively) (individually possess groups andcriminal gangs towhich extent The crime. inorganized engaged andgroups gangs of choice of weapon the increasingly are Guns cent). 20 per than more at group the half with cent, per (a30 of median higher tobe tends homicides crime-related ororganized gang- of proportion average the firearm, by committed are cides allhomi- of half than more where countries in the contrast, In cent). 10per under group the and half cent 4per of amedian (with low comparatively is homicides crime-related ororganized gang- of proportion average the countries, of group % homicides committed with firearms reprisals), homicide cases involving a firearm are are afirearm involving cases homicide reprisals), of dueto fear forward tocome witnesses some of reluctance the (such as homicides gang-related of clearance surrounding issues tothe addition In Jamaica. and Salvador El suchas prevalent, particularly are homicides gang-related inwhich Americas inthe includethose firearms with out carried homicides of proportion ahigh with Countries homicides. of proportion alower solve police the increases, firearms with committed cides homi- of percentage the as that toassert possible is it Nonetheless, results. of interpretation the in exercised be thus should Caution meanings. different have may solved’ ‘cases concerning countries by reported anddata limited, is very isavailable data which for countries of set The year. available closest orthe 2006 for isavailable data which for tries coun- 33 for solved cases homicide of percentage the against firearms with committed homicides 10 20 30 40 50 60 70 80 90 100 cases solved Figure Source 20304050607080901000 UNODC elaboration based on CTS-10 data (UNODC, n.d.c) (UNODC, data CTS-10 on based elaboration : UNODC 3.9 10

Proportion of homicides committed with firearms and firearms with committed homicides of Proportion Jamaica 16 Paraguay % homicide cases solved El Salvador Sweden Georgia Philippines United States Belize Ukraine Italy Vietnam Serbia Sl Austria ovenia Denmark 101 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 102 less likely to be solved than those involving a officers per recorded homicide. In contrast, knife or other forms of physical contact between countries with higher rates of solved cases have the victim and the offender. Many countries have between 50 and 500 police officers per recorded limited forensics capacity for detailed ballistics homicide.18 analysis, and the lack of close contact between There is not enough data to ascertain whether the victim and the offender means that few types this pattern holds true for other countries with of physical evidence (such as the offender’s hair, high homicide rates. Nonetheless, at least in some blood, or fingerprints) are left behind (Addington, countries, it may be the case that law enforcement 2006; Riedel, 2008). officers are unable to devote sufficient resources A study on the decline in the percentage of homi- to individual case investigation and clearance cide cases solved in the United States from more when faced with especially high levels of lethal ARMED VIOLENCE 2011 VIOLENCE ARMED

of than 90 per cent in 1960 to 61 per cent in 2006 violence. Looking through the armed violence finds that homicides involving weapons other than lens, however, does not help to determine the firearms were related to higher clearance rates: direction of cause and effect. Law enforcement institutions, for example, may lack the requisite The reason that weapons other than firearms are resources to fully investigate a large number of cleared more quickly is that forensic evidence is homicides, which leads, in turn, to a culture of [. . .] not available with firearms, especially impunity, little deterrent effect, and further handguns that kill at a distance and are widely GLOBAL BURDEN BURDEN GLOBAL increases in levels of armed violence. Whether available (Riedel, 2008, p. 1157). impunity from law enforcement drives high levels Furthermore, a number of countries at the top left of armed violence or vice versa remains an open of Figure 3.8 (high percentage of firearms, low question. However the problem is viewed, it is clearance) also have higher overall homicide rates, clear that strengthening the effectiveness of the while countries towards the bottom right (low criminal justice system represents a key entry percentage of firearms, high clearance) tend to point for armed violence reduction and an essen- be those with lower overall homicide rates. For tial component for violence prevention. example, Jamaica had a homicide rate of 58 per 100,000 in an average year between 2004 and 2009, with 75–80 per cent of homicides committed using a firearm and typically less than 40 per Homicide and criminal justice cent of homicides ‘cleared up’ by police.17 This system performance may indicate a connection, in some contexts, A criminal justice system may take many forms, between high rates of homicide, use of firearms, both in terms of both its structural organization low police performance, and, potentially, limited and its aims. It may choose a strong deterrent police resources. Countries that report low rates approach through the extensive use of punitive of solved cases also show comparatively low sanctions, such as imprisonment. It may alterna- numbers of police per recorded homicide. In tively (and usually for crimes less serious than particular, Jamaica and El Salvador (two of the homicide) seek restorative outcomes through countries whose homicide rates are among the resolution between the victim and perpetrator. highest in the world) have less than ten police The law enforcement component of the system homicide. The chapter also suggests that, where where that, suggests also chapter The homicide. intentional of levels lower with linked broadly system—is justice criminal the of operation the in law—including of rule the for respect below, shown as system, the of nature the Whatever courts. state by and enforceable recognized are decisions whose elders village or tribal by mediated systems justice informal of case inthe as control, state central of pendent inde- andlargely semi-formal be even may itself system The element. policing and community prevention crime astrong on focus may it hand, other the On offenders. of arrest andthe crime reported of solving the towards oriented be may Common approaches include the examination of: of: examination include the approaches Common varied. highly are violence armed involving crimes to andresponding inpreventing effectiveness system’s justice criminal the measuring for Methods levels. andcross-national national the both at system, justice criminal the of ness effective- the of understanding toincrease need anurgent highlights analysis Such time. over significantly increase rates homicide in which violence of toa spiral contexts) some (in tribute con- may this exists, crime serious for impunity criminal justice system resources system justice criminal and financial); and financial); (both human human (both Photo Reuters Marchante/ ©Rafael 2005. October Spain, Malaga, in trial stands suspect

S A homicide 103 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 104 system workload (such as the number of brought before the criminal courts, persons ofﬁcials per recorded case); prosecuted, or persons arrested) but always case progress (such as ‘case solved’ rates or uses the end point of conviction by a competent the likelihood that a suspected offender will court or tribunal. In contrast, an ‘attrition rate’ be convicted of an offence, known as case usually describes the ‘exit’ of persons from the ‘attrition’); criminal justice system at various stages, whether at the police, prosecution, or court levels. Attrition the impact of justice on the individual offender rates may be calculated separately for different (such as recidivism rates); and stages of the system or for the system as a the quality of justice (including the extent to whole, usually with reference to the number of which the system is perceived to lead to persons brought into initial formal contact with ‘wrongful’ convictions).19 ARMED VIOLENCE 2011 VIOLENCE ARMED the police.20 Figure 3.10 depicts the different of stages of the criminal justice system and possible This chapter has already brieﬂy considered pos- attrition measures. sible connections between police resources and case solved rates. It now focuses on measures of Such measures can provide quantitative infor- the likelihood that an offender will be brought to mation about the progress of cases through the justice. A common approach to this issue for crime criminal justice system and reveal areas where involving armed violence is the use of case ‘attri- system performance—and the resulting deter-

GLOBAL BURDEN BURDEN GLOBAL tion’ rates. rent effect—could be improved. A significant drop in case load from police to prosecution, for Criminal justice system case attrition rates can be example, may indicate difficulties in obtaining measured in a number of ways. The exact name sufficient evidence to link suspects to offences given to the measure often depends upon the in order to proceed to court trial. reference points within the system from which it is derived. A ‘conviction rate’, for example, may At the international level, the collection and inter- use different starting points (such as persons pretation of such data is extremely challenging.

Figure 3.10 Possible criminal justice system attrition measures

System stage Persons Police attrition measure Police Persons brought into initial formal contact Prosecution attrition measure Overall police to prosecution Police/prosecution Persons formally charged attrition measure Possible conviction rate Prosecution Persons prosecuted measures

Courts Persons brought before the criminal courts

Courts Persons convicted at ﬁrst instance 11 Figure averaged for a (minimum) two-year period between between period a(minimum) for two-year averaged countries, in24 homicide intentional of crime ofCriminal JusticeOperations Systems the to reported data available presents 3.11 Figure time. over measurements of averaging by extent, for, tosome accounted be can effect Source 7 countries inAsia 3 countries intheAmericas 14 countries inEurope years. three than more taking cases of cent 50per than more year, one with than more took cases of cent per 80 in2007,ending than more trials criminal for that, suggests example, for India, from Data consequences. signiﬁcant tially poten- with next, tothe year one from over carried cases by isaffected it instead, cases; of ‘cohort’ same the represent not does data compared the since imprecise, tobe likely are measures Such that year. convicted persons of number andthe prosecuted, persons of number the homicide, for arrested persons of number the against year inone police the by recorded offences homicide of number the between example, for made, be can A comparison process. justice criminal the of stages different at data suspect total from measure attrition mate anapproxi- of calculation isthe approach One Convicted Convicted United Nations Survey of Crime Trends Crime of and Survey United Nations UNODC elaboration based on CTS-11 and CTS-12 data (UNODC, n.d.c) (UNODC, data CTS-12 and CTS-11 on based elaboration : UNODC 3. Prosecuted

Attrition of homicide cases in cross-national perspective incross-national cases homicide of Attrition Suspected Suspected 02 04 06 08 01010101010150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0 Offences for the the for 21 This This 2003 and 2009, presented by region. by presented and2009, 2003 offences. This fits with the pattern of higher of fire- pattern the with fits This offences. recorded of number the of that two-thirds than isless suspects homicide of number average the however, isavailable, data which for Americas the in countries three the In offences. homicide than suspects homicide more average, on show, Asia and inEurope Countries regions. ance indifferent perform- system justice criminal of indicative be indeed may that features some reveals 3.11 Figure exercise, amethodological as Nonetheless, years. of anumber for open been have that files of orcounting years previous from cases’ of over ‘carrying includingthe institutions, between vary may practices Recording inpractice. patterns attrition case typical reflect not do that results create may countries across attrition case of averaging the data, inthe variance tohigh Due suspects. as system the entering initially sons per- of number the of apercentage as presented prosecuted persons of ofpercentage the number of suspected persons of number cent’. The per to‘100 isset offences homicide annualrecorded of number the figure, % attrition ofhomicide cases of homicide is presented as a as ispresented homicide of and convicted offences are then then are . Numbers 22 In the the In 105 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 106 arm homicides and resultant lower case clearance in the first place, nor in correctly identifying rates in some countries in the Americas. perpetrators or in reforming and successfully reintegrating convicted criminals into society. At first glance, countries in Europe and the Indeed, the core aim of the criminal justice system Americas show—paradoxically—more persons should be to uphold the rule of (criminal) law. ‘prosecuted’ for intentional homicide than persons ‘suspected’. This may be due to the inclusion of This means that acts of armed violence must be countries, particularly in Eastern Europe and Latin carefully defined and prohibited by clear criminal America, with legal and criminal justice systems law. Such laws must be publicly promulgated, that allow cases to be initiated at the prosecution equally and fairly enforced, and independently level or in which serious charges, such as homi- adjudicated. The criminal justice system must cide, are only assigned at the prosecution stage,

ARMED VIOLENCE 2011 VIOLENCE ARMED itself be accountable to procedural law, must

of following investigation by the prosecution service. demonstrate equality of persons before the law, Finally, data from countries in the Americas shows show fairness in application of the criminal law, particularly low numbers of persons convicted ensure procedural transparency, avoid arbitrari- for homicide compared to persons suspected ness, and—importantly—operate in a manner and prosecuted. This may be due to the use of consistent with international human rights different definitions at the prosecution and court standards and norms (UN, 2010). Indeed, case levels or different counting rules. Nonetheless, attrition and conviction rates contain no informa- GLOBAL BURDEN BURDEN GLOBAL the pattern is striking and may hint at limited tion about the quality of justice. In that context, criminal justice system capacity and resultant the United Nations Human Rights Committee has minimal deterrent effect. expressed concern, for example, about ‘extremely High case attrition in the form of low conviction high conviction rate[s] based primarily on confes- rates should not automatically be associated sions’ (UN, 2008, para. 19). In one survey of US with poor criminal justice system performance, law enforcement and criminal justice system however. Experts argue that there is a spectrum professionals, respondents said that wrongful of criminal justice system approaches. At one end felony convictions occurred in 1–3 per cent of all is a large law enforcement–punishment appara- felony cases (Ramsey and Frank, 2007). tus designed for public order maintenance, with One element that is inherent to the rule of law is high rates of arrest, prosecution, conviction, the quality of justice, including the idea that all and incarceration. At the other end is a system persons are ‘accountable to laws that are publicly commonly found in liberal democracies, which promulgated, equally enforced and independently emphasizes the protection of human and due adjudicated’ (UN, 2004). Thus, while the rule of process rights, invests more heavily in the judici- law demands strong state institutions endowed ary, and accepts a higher rate of case acquittals with clear legal powers, it also requires that insti- as a price for overall legally defensible convictions tutions act fairly and in a manner consistent with (Sung, 2006). international standards. The presence of weak Thus, while a criminal justice system can be effi- state law enforcement and criminal justice institu- cient at convicting persons for criminal offences, tions may create space for violent crime to prosper it is not necessarily effective in preventing crime (UN, 2009a). Conversely, where institutions are -3 bean countries—including St. Kitts and Nevis, andNevis, Kitts St. countries—including bean Carib- particular, In outliers. noticeable are that countries of groups three identifies 3.12 Figure rule of law. poorer with linked broadly are rates homicide higher that demonstrates figure The left. to the numbers negative isassigned law of rule weaker x-axis; the on right tothe numbers positive year). available orlatest (2009 rates homicide tional inten- overall against Index, Law Ruleof Bank World the by measured as law, of rule plots which 3.12, in Figure shown are homicide intentional and law of rule weak between links broad The (FBSP, p.32). 2010, countries European Western inmost rate homicide overall the than higher rate adeath for accounted example, for 2009, in deJaneiro inRio police andmilitary civil with inconfrontations killed Persons crime. violent fuel help can inturn, This, trust. public of loss and animosity generating risk they law, of rule the of principles respect not do but strong basically Homicide rate per 100,000 population Source Figure Somalia 23 UNODC elaboration based on CTS-10 data (UNODC, n.d.c) (UNODC, data CTS-10 on based elaboration : UNODC Stronger rule of law is expressed with with isexpressed law of rule Stronger 3.12 Group 3

Rule of law and overall intentional homicide rates, 2009 or latest available year available orlatest 2009 rates, homicide intentional andoverall law Rule of Afghanistan -2 Iraq Venezuela Weaker ruleoflaw Guatemala Group 2 Honduras -1 El Salvador Jamaica examples of particular contexts in which inwhich contexts particular of examples represent may but geographically linked not are countries These andVenezuela). Jamaica, Honduras, Guatemala, Salvador, El (including curve general the from expected than homicide of rates higher significantly but law of rule level tolow- mid- with countries of 2consists Group institutions. ernment gov- influence strong) of (reasonably preventive potential the notwithstanding violence, armed lethal of toacts contribute that trafficking, drug of presence orthe conflict gang suchas factors, additional of presence the suggest finding does this Nonetheless, involved. deaths of number smallabsolute corresponding andthe tions popula- duetosmallcountry unstable somewhat be can region inthis rates homicide that noted be should law. of It rule of levels solid paratively com- the from expected be may than homicide intentional of rates higher —show 3.12) in Figure 1 (Group Rico andPuerto Dominica, Grenadines, andthe Vincent St. Lucia, St. Bahamas, the 100 80 60 90 20 40 30 50 70 10 0 Dominica St. Kitts andNevis Bahamas Puerto Rico Group 1 St. Lucia St. VincentGrenadines and the +1 Stronger ruleoflaw +2 +3 107 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 108 comparatively weak rule of law is combined with The complexities of unpacking the two-way challenges, such as a past history of conflict, relationship between violence and institutional gang or organized crime activity, or high income responses suggests that information, such as inequality. The presence of such underlying driv- statistics on the number of offenders suspected, ers of violence with the absence of an effective prosecuted, or convicted, is unlikely to provide a deterrence system appears to be responsible complete picture of the probable impact of the for the especially high homicide levels in these law enforcement or criminal justice system. Rather, countries. information is also required on factors such as the degree of public confidence in the system Finally, countries in Group 3 (Afghanistan, Iraq, and the professionalism and integrity of criminal and Somalia) show much lower levels of homicide justice system professionals. Crime victimization

ARMED VIOLENCE 2011 VIOLENCE ARMED than expected from very weak rule of law institu-

of surveys and specialized access-to-justice surveys tions. This is probably due to limited institutional are one way of obtaining such information. Figure recording of violent deaths and particular diffi- 3.13 shows, for example, the extent to which culties in distinguishing between conflict- and victims reported experience of an assault with a non-conflict-related deaths. None of the three weapon to police or to other authorities. The data countries provide any homicide statistics; the is derived from the International Crime Victim estimates on homicide rates are all based on esti- Surveys conducted in urban areas in 74 countries mates of deaths caused by interpersonal violence GLOBAL BURDEN BURDEN GLOBAL between the years 1989 and 2005 (ICVS, n.d.). provided by the World Health Organization (TRENDS AND PATTERNS). These estimates of In countries in East Africa, victims of assault with deaths caused by interpersonal violence are far a weapon reported the crime to the police in just lower than in many other countries (Afghanistan: 31 per cent of cases. In 23 per cent of cases, the 2.6, Iraq: 2.2, and Somalia: 1.5 per 100,000). crime was reported to a different authority. This One reason for the low levels is that the pro- compares to only 12 per cent of victims reporting gramme categorizes many of the deaths that to other authorities and 48 per cent to the police occur in these conflict-affected countries as in Western and Central Europe. Low levels of case collective violence. reporting, particularly of serious crime types such as assault with a weapon, may be symptomatic of distrust or even of fear of law enforcement Figure 3.13 Reporting of crime involving armed violence by victims, authorities. Where victims turn to authorities 1989–2005 Reported to the police Reported to other authorities other than the police, the role of these authorities must be considered in the broader armed East Africa violence prevention context. In Europe, ‘other Southern Africa North America authorities’ could include private security guards, South America neighbourhood watch coordinators, or commu- Western and Central Europe nity support organizations. In many countries in

0 10203040506070Africa, traditional dispute resolution structures, % of victims who reported assault such as clan elders or councils of elders, have his-

Source: UNODC calculations from ICVS (n.d.) torically functioned and still function as relevant rates at the international level. international the at rates homicide ‘headline’ of publication frequent the in resulted has anddevelopment violence armed between links inthe interest Heightened Goals. Development Millennium the of attainment the and todevelopment iscritical that an indicator as recognized isincreasingly homicide Intentional Conclusion compensation. of payment as such penalties, imposing of capable authorities and combating non-conflict armed violence. armed non-conflict and combating measuring for tool policy akey as rates homicide inpromoting effective been have observatories andviolence security, crime, of a proliferation levels, andnational regional the At PATTERNS). AND (TRENDS Caribbean andthe America, and South Central Africa, andSouthern inMiddle highest rates homicide with and2009, 2004 between year inanaverage occurred homicide intentional from deaths 396,000 anestimated findsthat report 2010). andLever, Spuy der (van techniques policing including coercive to adopt ahave law begun enforcementrole, Americas andthe inAfrica countries of number ina organizations sector security private though, ational context. context. ational situ- its andof homicide of causes underlying the of anunderstanding towards step alimited only represent they andsocieties, states on have can violence lethal impact negative the highlight suchcomparisons While Governance. African of Index Ibrahim andthe Index, Peace Global 2007/8 Report Development UN’s includinginthe years, in recent settings indevelopment applied been increasingly has levels homicide of comparison Cross-national , the Economist, the 24 25 Increasingly, Increasingly, Chapter 2 of this this 2of Chapter Human ’s ’s 26

SeeBijleveld and Smit (2006); Dek 1 Endnotes See, forexample, FBI(2009). 2 cases encountered and responses delivered. delivered. andresponses encountered cases of monitoring andcomplete timely of purposes the for systems information justice and criminal of law enforcement area include strengthening the inthis data of improvement the towards steps Key torespond. astate of capacity existing the and on violence (and non-lethal) lethal of context the of a clear understanding upon dependent cally iscriti- needed, where support, technical deliver andto violence armed non-conflict of reduction the for policies andregional national coherent todevelop governments, andlocal national as well as community, international the of ability The DEVELOPMENT). (MORE LESS VIOLENCE, inparticular violence armed of complexities andthe homicides of anddistribution scale the understand tobetter tool animportant become since have that systems monitoring prehensive com- toestablish started have violence armed of levels high by affected governments Many occurrence. its andreduce toprevent action different requires context homicide Each ortheft. robbery as such crime, asecondary of context inthe killings to members, orfamily partners intimate against committed passion of crimes tounplanned groups, criminal ororganized gangs by out carried ings kill- violent from range may These and contexts. actions of range toadiverse correspond actually homicide intentional as andrecorded classified population’—events 100,000 per homicide tional ‘inten- measuring of approach indicator’ ‘single tothe contrast that—in suggests data Available Roberts (2009). er (1993); Killias (1993); 109 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 110 3 See, for example, the annexe of the Tenth United Nations committed the offence has been identified (regardless of Survey on Crime Trends and Operations of Criminal Justice whether he is in custody, on provisional release, still at Systems, or CTS-10 data (UNODC, n.d.c). large, or dead), or (iv) Police investigations reveal that no 4 See De Bondt and Vermeulen (2010). penal offence was in fact committed’ (UNODC, n.d.c). 5 The United Nations Convention against Transnational 17 See the crime review period 1 January 2008–31 December Organized Crime defines an ‘organized criminal group’ 2008 in JCF (2010). as ‘a structured group of three or more persons, existing 18 UNODC calculations based on CTS data (UNODC, n.d.c). for a period of time and acting in concert with the aim of 19 Aromaa and Heiskanen (2008); Riedel (2008); Ramsey committing one or more serious crimes or offences estab- and Frank (2007); Smit, Meijer, and Groen (2004). lished in accordance with this Convention, in order to 20 See, for example, Aromaa and Heiskanen (2008). obtain, directly or indirectly, a financial or other material benefit’ (UNODC, 2004, art. 2(a)). A ‘structured group’ 21 UNODC calculation based on NCRB (n.d.). means ‘a group that is not randomly formed for the 22 The Eleventh and Twelfth United Nations Surveys of Crime immediate commission of an offence and that does not Trends and Operations of Criminal Justice Systems contain

ARMED VIOLENCE 2011 VIOLENCE ARMED need to have formally defined roles for its members, time series for the period 2003 to 2009. Recorded inten- of continuity of its membership or a developed structure’ tional homicide offences, persons suspected of intentional (art. 2(c)). homicide, persons prosecuted for intentional homicide, 6 Data sources used, including the GBAV 2011 dataset, are and persons convicted of intentional homicide were aver- provided in the online methodological annexe to this aged, by country and by two-year groups, across the time chapter; see www.genevadeclaration.org. period. Only countries with a consistent two-year percentage change across the variables were included in the 7 See, for example, Muggah and Krause (2009). analysis (UNODC, n.d.c). 8 See Small Arms Survey (2009, ch. 7). 23 The World Bank Rule of Law Index consists of a composite 9 See, for example, Home Office (2010), FBI (2009), and measure of both representative and non-representative GLOBAL BURDEN BURDEN GLOBAL Salfati and Park (2007). survey-based sources that attempt to capture perceptions 10 This rate is the average percentage of homicides classified and experience of concepts related to the rule of law as ‘gangland killings’ or ‘juvenile gang killings’ (informa- (World Bank, n.d.). The index represents a general rule of tion on ‘murder circumstances’) among total homicides law measure and not an assessment of the effectiveness for the period 1998 to 2008, as reported by the Federal of the criminal justice system alone. Some correlation Bureau of Investigation in the United States Uniform Crime between the Rule of Law Index and intentional homicide Reports (FBI, n.d.). rates is to be expected, insofar as the index includes per- 11 The countries included are: Antigua and Barbuda, the ceptions of crime and measures of victimization, giving Bahamas, Bermuda, Cuba, Dominica, Dominican Republic, rise to some degree of cross-linking between the two Grenada, Jamaica, St. Kitts and Nevis, St. Lucia, St. Vincent datasets. and the Grenadines, and Trinidad and Tobago. (TRENDS 24 See, for example, Kinyanjui (2009). AND PATTERNS). 25 See, for example, Geneva Declaration Secretariat (2010). 12 This chapter treats Trinidad and Tobago separately, 26 At the regional level, see OIS (n.d.). whereas Chapter Two lists it as part of the Lesser Antilles Region (TRENDS AND PATTERNS). 13 The Gini coefficient of inequality assigns values between 0 and 1 to each country, with 0 representing absolute equality and 1 representing absolute inequality. Bibliography 14 See Minnaar (2008). Addington, Lynn A. 2006. ‘Using National Incident-Based Report- 15 See Hemenway (2011). ing System Murder Data to Evaluate Clearance Predictors.’ 16 The Tenth United Nations Survey on Crime Trends and Homicide Studies, Vol. 10, No. 2, pp. 140–52. Operations of Criminal Justice Systems defines a case as Aromaa, Kauko and Markku Heiskanen, eds. 2008. Crime and ‘solved’ according to the following criteria: ‘(i) The police Criminal Justice Systems in Europe and North America, are satisfied of a suspect’s guilt because there is a cor- 1995–2004. Helsinki: HEUNI. roborated confession and/or because of the weight of the Bijleveld, Catrin and Paul Smit. 2006. ‘Homicide in the Netherlands: evidence against him, or (ii) The offender was caught in On the Structuring of Homicide Typologies.’ Homicide the act (even if he denies all guilt), or (iii) The person who Studies, Vol. 10, No. 3, pp. 195–219. Hemenway, David. 2011.‘Risks andBenefits of aGun inthe Gillies, H. —. 2010. Measuring andMonitorin Geneva Declaration Secretariat. 2008. GlobalBurden ofArmed FBSP (FórumB FBI (Federal Bureau of Investigation). n.d. ‘UniformCrime Reports.’ Fajnzylber, Pablo, Daniel Lederman, andNorman Loayza. 2002. Deker, Scott. 1993. ‘Exploring Victim–Offender Relationships in De Bondt, Wendy andGert Vermeulen. 2010. CSVR (Centre forthe Study of Violence andReconciliation). 2008. Covey, 2010. Street Herbert. Gangs throughoutWorld the CIPEV (Commission of Block, Carolyn andRichard Block. 1993. JCF (Jamaica Constabulary Force). 2010. CrimeStatistics Analysis tion System. Inquiry intothePost Inquiry , Vol. 10, No.4,pp. 585–612. Segurança Púb g Armed Violence atistical Bulletin: Homicides, s/cjis/ucr/ucr> e, ‘Achieving theMillennium wnload datafiles.htm> Journal ofLaw and Street GangCrimein lica). 2010. Post Election Violence Election Violence). Revolutions inEU . Year 4. São Paulo: British Journal wnload?func 2February. : Goals, Anuário do . Killias, 1993. Martin. ‘International Correlations between Jütersonke, Olivier, Robert Mu —. n.d. ‘Crime Statistics.’ Ng’walali, Paul andJames Kitinya. 2006 NCRB (National Crime Records Bureau). n.d. ‘Crime inIndia.’ Muggah, Robert andKeith Krause. 2009. ‘Closing theGap Mohanty, ManojKumar. 2004. ‘Variants of Homicide: AReview.’ Minnaar, Anthony. 2008. ‘Firearms andMurderin South Africa, Miethe, Terance and Wendy Regoeczi. 2004. RethinkingHomicide: Leslie, Glaister. 2010. Kinyanjui, Sarah. 2009. ‘Restorative Ramsey, Robert andJames Frank. 2007. ‘Wrongful Conviction: Panama. 2007. New Delhi:Ministry of HomeAffairs, Government of India. Peacekeeping, Towards a Violence Reduction Agenda.’ Between Peace Operations andPost-Conflict Insecurity: Journal ofClinical Forensic Medicine media/conferences/2008-homicide/minaar> gov.au/events/aic%20upcoming%20events/2008/~/ Paradise, Queensland, 3–5December. l/cpo_observatorio.asp> Enabling Development , Vol. 40, Nos.4–5, , Vol. 11,pp. 214–18. Vol. 148,No.10, International Tribal Law Panama Panama , Vol.6, . 111 2 3 4 5 1 CHARACTERISTICS OF ARMED VIOLENCE 112 Rausch, Colette, ed. 2006. Combating Serious Crimes in Post- —. 2010. Drug Control, Crime Prevention and Criminal Justice: conflict Societies: A Handbook for Policymakers and Practi- A Human Rights Perspective—Note by the Executive Direc- tioners. Washington, DC: United States Institute of Peace. tor. E/CN.7/2010/CRP.6-E/CN.15/2010/CRP.1 of 3 March. Riedel, Marc. 1999. ‘Sources of Homicide Data.’ In M. Dwayne Vienna: Commission on Narcotic Drugs and Commission Smith and Margaret A. Zahn, eds. Homicide: A Source- on Crime Prevention and Criminal Justice. —. 2008. ‘Homicide Arrest Clearances: A Review of the Literature.’ UNFPA (United Nations Population Fund). 2003. Violence against Sociology Compass, Vol. 2, Iss. 4, pp. 1145–64. Women in South Asia: A Regional Analysis. Kathmandu: UNFPA. Roberts, Darryl. 2009. ‘Intimate Partner Homicide: Relationships to Alcohol and Firearms.’ Journal of Contemporary Criminal UNODC (United Nations Office on Drugs and Crime). 2004. United Justice, Vol. 25, pp. 67–88. Nations Convention against Transnational Organized Crime and its Protocols. Salfati, C. Gabrielle and Jisun Park. 2007. ‘An Analysis of Korean

of Publications/TOC%20Convention/TOCebook-e.pdf> Violence, Vol. 22, No. 11, pp. 1448–70. —. 2007. Crime and Development in Central America: Caught in Small Arms Survey. 2009. Small Arms Survey 2009: Shadows of the Crossfire. Vienna: UNODC. War. Cambridge: Cambridge University Press. —. n.d.a. ‘UNODC Homicide Statistics.’ Cambridge: Cambridge University Press. —. n.d.b. ‘Criminal Justice Assessment Toolkit.’ —. 2011. Perceptions of Security in Liberia: A Legacy of War? Small Arms Survey. GLOBAL BURDEN BURDEN GLOBAL —. n.d.c. ‘United Nations Surveys on Crime Trends and the Smit, Paul, Ronald Meijer, and Peter-Paul Groen. 2004. ‘Detection Operations of Criminal Justice Systems (CTS).’ Rates: An International Comparison.’ European Journal on National Perspective: From Crime Control to Due Process.’ UNSC (United Nations Security Council). 2009. Nineteenth Annals of the American Academy of Political and Social Progress Report of the Secretary-General on the United Science, Vol. 605, pp. 311–37. Nations Mission in Liberia. S/2009/411 of 10 August. Townsend, Dorn. 2009. No Other Life: Gangs, Guns, and Govern- van der Spuy, Elrena and Jeffrey Lever. 2010. ‘Crime Prevention ance in Trinidad and Tobago. Working Paper No. 8. Geneva: and Community in Africa.’ Working Paper No. 6. Santiago Small Arms Survey. de Chile: Global Consortium on Security Transformation. Uganda Police. 2008. Annual Crime Report 2008. August. info.worldbank.org/governance/wgi/mc_countries.asp> UN (United Nations). 2004. The Rule of Law and Transitional Wortley, Scot. 2009. 2009 Jamaican National Crime Victimization Justice in Conflict and Post-conflict Societies: Report of the Survey: Preliminary Results. —. 2008. Consideration of Reports Submitted by States Parties under Article 40 of the Covenant—Concluding Observations of the Human Rights Committee: Japan. CCPR/C/JPN/CO/5 of 18 December. —. 2009a. Promoting Development through the Reduction and Prevention of Armed Violence: Report of the Secretary- General. A/64/228 of 5 August. —. 2009b. Report of the Special Rapporteur on Extrajudicial, Summary or Arbitrary Executions. A/64/187 of 29 July. 2007 (CDC, n.d.a). (CDC, 2007 and 1999 15–24 between aged allraces of women for death of leadingcause second the reportedly was homicide example, for States, United the In women. against committed violence homicidal of gravity the tounderestimate however, take, amis- be would It 2009). ABS, 2010; FBI, 2009; (INEGI, inAustralia cent and29 per States, United inthe cent 23per inMexico, victims homicide of cent 10per approximately constitute women while homicide, of victims the of percentage high adisproportionately represent generally Men conflict. armed andfollowing during women against committed violence sexual as well awoman—as of killing ‘femicide’—the on focuses It women. against directed violence armed of patterns global unpacks chapter This wanting. found are generalizations crude these Yet inspection closer on households. headsof single as andoften survivors, victims, as ways: indifferent affected are said, is often it Women, killed. andbe tokill likely most to be lence conceal complex patterns of victimization victimization of patterns complex conceal lence vio- andsexual femicide of estimates statistical Yet 2005). BJS, 2008; Secretariat, Declaration (Geneva partners or former current frequently are women against violence serious of trators perpe- the violence; andsexual homicide trating inperpe- involved commonly more also are Men T especially young men—are determined determined men—are young especially Men— terms. ingeneral described often of he demographics When the Victim is aWoman Chapter 1

Four armed violence are are violence armed challenge for development. for challenge aserious women—constitutes against violence especially violence—and armed broadly, More 1978). (Pearce, insociety women marginalizing further poverty’, of ‘feminization tothe contribute can violence of aculture that evidence is growing Yet there considered. rarely are market labour wider and workforce female the on violence and sexual femicide of implications the contrast, In 2006). Florquin, 2008; Secretariat, Declaration (Geneva economy formal inthe men wage-earning of tivity produc- lost the highlight violence armed of costs andeconomic social the examining assessments Many andcommunity. family inthe role former victim’s the reflect that repercussions reaching far- generates it effects; knock-on traumatic with isatragedy individual any of killing violent The pp.12–13). 2008, (MOWA, marriage forced including violence, andsexual physical, logical, psycho- of forms tosevere isattributed women amongAfghan suicideandself-immolation reported in increase sharp The misery. their toend order suicidein commit also relationships and violent abusive endure who women Many and bystanders. witnesses, includingchildren, others, of lives the taking also suicidewhile commit sometimes tors perpetra- casualties; indirect frequently also are there iskilled, awoman when Moreover, comes. out- leadtofatal cases insome that aggressions escalating of culmination the isoften It incident. anisolated as occurs seldom women against violence that given especially and suffering, 113 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 114 While femicide and sexual violence are prevalent text of widespread tolerance of sexual violence in the domestic sphere, this chapter finds that and a high level of impunity (Peterman, Palermo, women are exposed to many other forms of vio- and Bredenkamp, 2011).2 Indeed, a study conducted lence—from gang violence to robberies and stray by the World Health Organization in selected bullets. They are also often singled out for hate countries reveals that adolescent girls frequently crimes, particularly when they are categorized as experience the practice of forced sex initiation migrants and refugees (Freedman and Jamal, 2008, (WHO, 2002, p. 153); this finding suggests wide- pp. 13–19). Guatemala and Mexico’s Ciudad Juarez spread acceptance and impunity of violence exhibit staggering numbers of women victims and against women and girls. exceptional brutality. In these contexts, intimate Femicide is an important component of armed vio- partner violence accounts for a relatively small lence and includes violence in the domestic sphere,

ARMED VIOLENCE 2011 VIOLENCE ARMED proportion of femicides (Suarez and Jordan, 2007); such as that perpetrated by intimate partners and of many women are victims of the increasingly wide- strangers. This chapter seeks to disaggregate the spread violence related to organized crime and demographics of armed violence and capture the narco-trafficking, which affects the entire popu- ways in which women of different ages are at risk. lation (Molloy, 2010). Furthermore, according to Specifically, the chapter finds that: a group of Mexican NGOs:

impunity and government permissiveness, which In the 111 countries and territories under review, an annual average of 44,000 women

GLOBAL BURDEN BURDEN GLOBAL serves as a crude expression of institutional violence, have led to a multiplication in the became homicide victims in 2004–09. number of women murdered throughout the Roughly 66,000 women are violently killed country and this can be attributed to a lack of around the world each year, accounting for due diligence (RNOCDH, 2010, p. 4). approximately 17 per cent of total intentional homicides. In such areas, an exclusive focus on femicide may limit the understanding of a broader picture of On average, men are killed approximately five extreme violence, which reveals major ‘systemic times more frequently than women. failures’ (Eriksson Baaz and Stern, 2010, p. 12; Femicides generally occur in the domestic A UNIFIED APPROACH). sphere; the perpetrator is the current or former partner in just under half of the cases. The social, cultural, and political risk factors for femicide and sexual violence are widely debated. Countries featuring high homicide rates in Analysts frequently point to cultures of ‘machismo’ the male population also typically experience that can distort traditional gender roles and high femicide rates. encourage constraints on the freedom of girls High levels of femicide are frequently accom- and women, misogynist behaviour, and recurring panied—and in some cases generated by—a violence with impunity. In many war zones, rape high level of tolerance for violence against has been used as ‘a weapon of mass destruction’ women. (UNDP, 2008, p. 2). Assessments of sexual violence In countries where violence is widespread, in the Democratic Republic of the Congo indicate the rate of victimization of women reaches that mass rapes and atrocities committed against levels far above the average risk of domestic women during periods of fighting occur in a con- violence. post-conflict settings. post-conflict and HIV/AIDS inconflict contracting of risk the as well as witchcraft, with associated practices and lethal killings, ritual killings, ‘honour’ deaths, includingdowry women, against violence lethal and non- lethal of forms other examines section third The victims. female involving homicide of anddynamics characteristics the considers tion sec- second The used. andinstruments offender, and victim between relationship the incidence, includingthe statistics, available of basis the on femicide of anoverview provides section first The violence. andsexual shapingfemicide and risks settings particular the considers chapter This remains scarce. remains data comparable internationally information, sex-disaggregated tomaintaining committed are countries of number agrowing Although occurred. event agiven inwhich context the on information orcritical victim the of sex the record isfactorily sat- not may andfiles reports police is pursued, suchanalysis when Even data. homicide overall tosupplement information additional generate and toidentify required be may analysts settings, inmany available readily not issimply such data Yet since sex. by statistics homicide gregating disag- isby women against perpetrated violence lethal of extent the measuring of way useful One sex by homicide Disaggregating defines femicide broadly as ‘any homicide with a with homicide ‘any as broadly femicide defines chapter This scope. geographical possible largest the covering femicide on statistics national piles Violence Armed of Burden Global 2011 The In some countries that exhibit low homicide homicide low exhibit that countries some In similar to that of malevictims. of tothat similar is victims female of percentage the rates, 3

com- the following: including violence, andsexual femicide on tics statis- disaggregated ingenerating challenges numerous face Countries standards. national inter- of short falls toviolence—still in relation life—including andprivate public of areas in many statistics gender andreliable valid of production the improvements, widespread Notwithstanding and parliamentary representation (UN, 2010a). (UN, representation and parliamentary employment, schoolenrolment, population, to relating sectors inkey available increasingly are statistics sex-disaggregated Fortunately, offenders. to andrespond violence toprevent measures of effectiveness the toassess which against indicators includingbaseline women, against violence of incidents on information objective of provision the for calls context This law. national inter- under violence from free tolive right their of inrecognition allwomen, on afocus for allows deﬁnition wider The sex. their of because females of targeting the suchas intent, of interpretation narrow anexclusive, avoiding thus victim’, female the reliance on mixed-quality data from dif- from data mixed-quality on reliance the leads statistics inavailable detail of lack the in toinvest andresources capacity limited standards international agreed of absence the concepts, basic of under-development the mines validity (Alvazzi del Frate, 2010). Frate, del (Alvazzi validity mines under- data) andmortuary andhealth sources, justice criminal andother police data, survey and administrative (suchas sources ferent to under-diagnosis; outputs; qualitative on inanover-reliance results collection data comparability; reduces systems and coding collection; data limits andmethods deﬁnitions,

115 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 116 Investments in generating accurate sex-disaggre- Box 4.1 Unpacking femicide: what’s in a label? gated data in the crime and justice sectors appear 5 When it was coined by the feminist movement in the 1970s, the term to be growing. Several initiatives have already ‘femicide’ implied the killing of women specifically because they were started to generate a wealth of data, albeit not women. As such, it was intended to convey ‘the misogynous killing of always comparable across time and space.6 women by men’ and to capture ‘the proportion of female deaths that Owing to the absence of agreed definitions, many occurred due to gender-based causes’ (Radford and Russell, 1992, p. 3; institutions and researchers collect data and Bloom, 2008, p. 178). The overall femicide concept emerged as an expres- develop datasets based on incompatible working sion of the feminist movement to politicize and contest male violence against women. definitions. The term ‘femicide’ is a case in point (see Box 4.1). Some analysts describe femicide as Diana Russell, an architect of the femicide concept, argues that the notion ‘the proportion of female deaths occurring due th ARMED VIOLENCE 2011 VIOLENCE ARMED has been in use for centuries. She traces its origins to early 19 -century

of to gender-based causes’ (Bloom, 2008, p. 178). Britain, when it was used to describe ‘the killing of a woman’ (Russell, 2008, p. 3).4 She also acknowledges that the term later emerged as a symbol of the Others contend that femicide refers to gender- battle to emancipate women and free them from violence in the 20th century. disaggregated data on homicide or ‘murder of women’, as indicated by the former Special Rap- From the beginning, the idea of femicide was designed to account for a range of specific forms of violence. These include dowry and ‘honour’ kill- porteur on Violence against Women, Its Causes ings, intimate partner or spousal violence, murder with rape, the killing of and Consequences (UN, 2008, p. 18). prostitutes, female infanticide or selective abortion, and other deaths that, This chapter draws on a combination of data GLOBAL BURDEN BURDEN GLOBAL according to forensic reports, occur as a result of women or girls being sources to generate a profile of femicide and targeted on the basis of their sex (for example, victims of a serial killer who specifically targets women) (Bloom, 2008, p. 178). sexual violence. These sources include reports and surveillance data from national statistical While the concept has drawn attention to the particular ways in which institutions, law enforcement and public health women are selectively targeted, the definition has become progressively agencies, and research institutions. The resulting diluted and confused. Indeed, it is now often conflated with a broader understanding of violence against women (such as any killing of a woman) GBAV 2011 femicide database entails the follow- and has thus lost much of its original political connotation. The broaden- ing three sets of data, covering 111 countries and ing of the definition may be connected to a growing interest in generating territories—56 per cent of the world’s female quantitative information of violence against women to facilitate compara- population—for the period 2004–09:7 bility across countries and jurisdictions.

Today, most of the related literature applies a broad interpretation of femi- Femicides in general (homicides with female cide. Yet there are also several qualitative studies and data collection victims): drawing on the general femicide initiatives assessing femicide in the stricter sense, examining the intent of database alluded to above, information was the perpetrator and thus separating femicides from other killings of women. checked for consistency and categorized These projects are especially common in Latin America and include, for according to the classification of countries example, the Registro de Feminicidio del Ministerio Público, hosted by the presented in Chapter Two (TRENDS AND Crime Observatory of the Ministerio Público in Peru; the Observatorio de PATTERNS).8 The examination therefore Muertes Violentas de Mujeres y Femicidio in Honduras; the Observatorio de Femicidios in Argentina; the Observatorio Ciudadano Nacional del includes a total of 104 countries or territo- Feminicidio in Juarez, Mexico; and the Banco Datos Feminicidio for Latin ries, providing information on most world America, based in Chile. regions (with the exception of Middle and Western Africa). global level. global the at year every killed women 66,000 mately approxi- yields estimate The data. no with regions two the for rate global the andusing information missing with countries of populations cide tothe femi- of rates regional applying by calculated was women against violence lethal of extent global The femicide?of 2004 and 2009. and 2004 between review under andterritories countries 104 inthe reported were femicides 44,000 about of annualtotal anaverage findsthat analysis The Two. inChapter presented data of that than andcomparable reliable isless information sex-disaggregated of quality overall the ards, stand- international relevant of absence to the Due excluded. been thus have andthey women, of deaths conflict-related direct on information insufficient was There femicide. torecord structed con- ever largest the isamong database femicide 2011 GBAV the femicide, of burden overall the under-representing andthus exhaustive not While male homicide victims than female ones. female than victims male homicide more times five are there average, on means that, ratio this population; total 100,000 per killed men five vs. woman one of inarate results This 4.1). Figure (see year inagiven deaths violent of number total the of cent 17 per approximately sent

tal’ violence, may be used in some settings. insome used be may tal’ violence, or‘mari- ‘spousal’ suchas definitions, varying however, partners; intimate orformer current by committed tohomicides refers dataset IPV the territories. and countries 54 of sample asmaller yielded dataset this (IPV), violence intimate-partner to attributable femicides on information of femicides: Intimate partner-related sources for 24 countries andterritories. countries 24 for sources of avariety from information collates set firearms: with committed Femicides 11 Female victims thus appear torepre- appear thus victims Female 9 In principle, all data included in included alldata principle, In 10 What, then, is the global extent extent global isthe then, What, consisting consisting this data- this vs. male homicide victims per year, 2004–09 per victims malehomicide vs. Legend: Figure assessments of interventions. assessments and policy-making effective more for allow will data accurate that conviction the on based Africa across systems instatistical invested have cies agen- development of asmallnumber however, Goals, Development Millennium the of lishment estab- Sincethe PATTERNS). AND (TRENDS rates homicide high featuring those of including many countries, andAsian African most for unavailable is information andsex-disaggregated common are gaps major data shows, mapclearly the As isavailable. information reliable which for tries incoun- rates femicide specific reveals 4.1 Map Sources surveillance capacity can be established. be can capacity surveillance acomprehensive before taken be still must steps major developed, being are systems monitoring andnon-governmental public both that shows Twenty-two countries feature ‘low’ rates (1–2) rates ‘low’ feature countries Twenty-two population). female 100,000 (<1 per rates low’ ‘very exhibit 41countries findsthat It available. is data which for andterritories in 104countries population female 100,000 per femicide of sity inten- the of distribution the considers 4.2 Figure and Europe. Americas the from isavailable coverage best the Female victims (66,000; 17%) Male victims (330,000; 83%) 4.1 GBAV 2011 database and femicide database femicide and database 2011 : GBAV

Estimated average proportion of female proportion average Estimated 12 While evidence evidence While 13 To date, To date, 117 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 118 Map 4.1 Rates of female homicide victims per 100,000 female population, 2004–09 ARMED VIOLENCE 2011 VIOLENCE ARMED of

LEGEND: Per 100,00 women >6 3–6 2–3 GLOBAL BURDEN BURDEN GLOBAL 1–2 <1 No data

Source: GBAV 2011 femicide database

Figure 4.2 Countries and territories by rate of femicide per 100,000 while 16 countries have ‘medium’ rates (2–3). female population, 2004–09 The ﬁgure shows, however, that 13 countries suffer from ‘high’ rates (3–6) and another 12 Number of countries countries report a ‘very high’ intensity of femicide 90 (>6). All in all, 25 countries, or nearly one-quarter 80 of all reviewed countries, exhibit ‘high’ or ‘very 70 high’ rates of femicide—that is, rates above the 60 global average. The sum of femicides committed 50 in these 25 countries represents almost half 40 (47 per cent) of the total number of femicides in 30 the database. 20 At the regional level, the highest rates of femi- 10 cide are found in regions that also exhibit very

Very low (<1) Low (1–2) Medium (2–3) High (3–6) Very high (>6) high rates of homicide. Figure 4.3 compares the Rate of femicide per 100,000 female population average rates of total homicide and femicide

Source: GBAV 2011 femicide database across countries pooled by region.14 Spreading Western, and Southern European countries. countries. European andSouthern Western, Northern, of majority inthe rates homicide overall than higher times up tofive are rates femicide the countries, these allof In femicide. of rates high extremely register andHonduras Guyana addition, In PATTERNS). AND (TRENDS ingeneral violence andlethal femicide inboth first ranks Salvador El Indeed, orabove. population female 100,000 per 10femicides of rates allreport Africa and South Jamaica, Guatemala, Salvador, El level, country the At homicide. of rates high tofeature tend also femicide of rates high exhibit that Countries Femicide bycountry rates Africa, order—Southern in descending namely— ranking, femicide the of top the at ture fea- Two also inChapter rates homicide highest the with regions five of out Four higher. are rates where larger appear that discrepancies shows tion popula- total the across femicides of number the (CHARACTERISTICS OF ARMED VIOLENCE). ARMED OF (CHARACTERISTICS 1:1 approaches therefore ratio victim to-female male- resulting The victims. the frequently are women inwhich violence ordomestic passion’ of ‘crimes of proportion includeahigher ably prob- ingeneral—homicides violent less thus are that rates—and homicide low exhibit that tries incoun- that be may explanation One population. total 100,000 per 0.4 only by rate femicide the exceeds that rate ahomicide with difference, est small- the records Europe Western isminor; rates andhomicide femicide between difference the (<1 100,000), per rates low very with regions In list. the of bottom the at appears Africa Northern femicide, of rates lower exhibit that regions Among ingeneral. homicides to respect with high disproportionally appear Europe inEastern rates femicide meanwhile, America; andCentral Caribbean, the America, 15 South South rate of 5.0 per 100,000. per 5.0 of rate ahigh has Belize 141,000, approximately of tion popula- andafemale homicide of victims female seven of count atotal with example, For rate. femicide high arelatively generate women of ings kill- afew even smallcountries, andother these In individuals. 500,000 than fewer of populations hometofemale andGuyana—are Belize, Bahamas, the as countries—such Some inAfrica. and one inAsia are three countries, remaining the Among Europe. inEastern andfour inNorthern three inEurope: located are rates femicide high or very high reporting countries Seven 4.4). Figure (see America inSouth andsix America, in Central four Caribbean, in the four Americas: inthe are (14) cent 50per than more rates, femicide high andvery high feature that 25 countries the Of Northern Africa Northern Southern Europe Asia Eastern Oceania Europe Northern Europe Western Asia Southern America Northern Eastern Africa Asia South-Eastern Asia Western Asia Central America Central Caribbean South America Eastern Europe Southern Africa 2004–09 region, by population, Figure Source Femicide Femicide GBAV 2011 database and femicide database (unweighted regional averages) regional (unweighted database femicide and database 2011 : GBAV 4.3 Total homicide

Average homicide and femicide rates per 100,000 total total 100,000 per rates andfemicide homicide Average 0 16 51015202530 40 Rate per 100,000 total population 54 50 45 35 119 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 120 Figure 4.4 Average femicide rates per 100,000 female population in 25 countries and territories with high and very high rates, 2004–09

El Salvador

Jamaica

Guatemala

South Africa

Russian Federation

Guyana

Honduras

Azerbaijan

Lesser Antilles Region ARMED VIOLENCE 2011 VIOLENCE ARMED

of Colombia

Bolivia Very high Bahamas

Lithuania High Belarus

Venezuela

Latvia GLOBAL BURDEN BURDEN GLOBAL

Belize

Kazakhstan

Moldova

Brazil

Kyrgyzstan

Ukraine

Ecuador

Dominican Republic

Estonia

0123456 7 8 91011 12 Rate per 100,000 female population

Source: GBAV 2011 femicide database

Countries in Latin America and the Caribbean on femicide at the sub-national level reveals even appear to share a particularly high incidence of higher rates than those recorded as national aver- femicide. A comparison reveals that not only young ages. In Mexico’s Ciudad Juarez, for example, the men, but also a disproportionately high number 2009 femicide rate was 19.1 per 100,000 female of women and girls, are dying in high numbers. population (see Box 4.2). In the state of Espirito Killings appear to be most frequent in specific Santo in Brazil, the government authorities docu- states and cities of these countries. Indeed, data mented a femicide rate of 10.9 per 100,000 in 2008.17 Source 5 10 15 20 25 30 Rate per 100,000 female population Figure Ciudad Juarez Ciudad Juarez 9319 95199 1995 1994 1993 level of savagery often confined to war zones. war to confined often savagery of level aparticular betrays killed, being before tortured are whom of many women, of executions brutal the Nevertheless, victimization. male of rate the in trend asimilar by ismirrored 2009 in observed peak the 2008, to 2007 from climb a radical Following 2009. 1993– during levels city and national the at femicide in Source: Small Arms Survey elaboration of Arroyo (2011) Arroyo of elaboration Survey Arms Small Source: serious human rights crisis. a constitutes 2009—and and 1990 between 669 counts issignificant—INEGI city this in killed girls and women of number The 2009). (INEGI, 2009 in population 100,000 per homicides 170 some corded re- It Mexico. in rates homicide highest the of one exhibits currently state Chihuahua in Juarez Ciudad the national average. above also were (3.7) Nayarit and (3.8), Michoacán (3.8), Tamaulipas Oaxaca (6.0), (4.5), (5.1), Sonora (7.1), Sinaloa Durango in (10.1). Rates Guerrero and (10.1), California Baja population), female 100,000 per (13.1 Chihuahua average: national the above well rates exhibited states three 2009 in violence; escalating by affected equally are Mexico of areas all Not 2009. by 100,000 per 3.5 to increase rapid a by 2007, in followed population female 100,000 per 2.0 of low ahistorical at rates femicide reported (INEGI)— yGeografía Estadística de Nacional Instituto Mexico—the of authority statistical national The crisis a humanrights B Femicide in Ciudad Juarez: Juarez: inCiudad Femicide ox 4.2 Small Arms Survey elaboration of Arroyo (2011) (2011) Arroyo of elaboration Survey Arms : Small 4.5

Trends in femicide per 100,000 female population in Ciudad Juarez and Mexico, 1993–2009 1993–2009 andMexico, Juarez inCiudad population female 100,000 per Trends infemicide Mexico 9719 1999 1998 1997 6 18 Figure 4.5 shows trends trends 4.5 shows Figure 0020 0220 0420 0620 082009 2008 2007 2006 2005 2004 2003 2002 2001 2000 Mexico, June 2009. © Teun Voeten/Panos Pictures Voeten/Panos ©Teun 2009. June Mexico, Juarez, Ciudad in women of disappearance and murder the Photo Armed police stand behind pink crosses marking marking crosses pink behind stand police Armed 121 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 122 As noted above, high femicide rates are often Fears of reprisal during the war partly explain the accompanied by high levels of tolerance to vio- silence regarding contemporary forms of violence lence against women among the wider population. and crime, including violence against women. Analysts claim that such behaviour is shaped by Certain members of the Salvadoran state had levels of gender inequality and norms that discrimi- employed local militia to perpetrate violence nate against the status of women—norms that are during the civil war, highlighting how conﬂict- often shared by women themselves. One widely related violence can affect the domestic and cited study reports that about one in four women private spheres. Thus: surveyed in 33 countries agreed that it was justi- despite important legislative changes, normative ﬁable to be hit or beaten for arguing or refusing to notions of appropriate behavior for men and have sex with one’s husband (UN, 2010a, p. 137). women still make violence ‘acceptable’ in certain ARMED VIOLENCE 2011 VIOLENCE ARMED

of Additional research has demonstrated that in contexts to the point that it is not always recog- settings with high rates of femicide the criminal nized as violence (Hume, 2008, p. 64). justice response may be substantially slower and less efﬁcient than for homicide more generally. Trends in femicide Cases may not be investigated and consequently It is possible to examine time-series trends in not prosecuted, resulting in very low clearance femicide and violence against women for coun- and conviction rates. For example, a recent study 21

GLOBAL BURDEN BURDEN GLOBAL tries with available data. Attributing the value finds that between 2008 and 2010 in Honduras, 100 as a starting point for all countries in 2004 only 211 of 1,010 reported cases of femicide were allows several different patterns emerge. These heard in court and only 56 sentences were passed include countries experiencing rapid increases, (Sánchez, 2011, p. 40). Similar patterns have rapid decreases, fluctuating trends, and more emerged in other regions where femicide is poorly stable rates. Among 22 countries with trend investigated and countered.19 series available, only seven exceeded the starting Armed conflicts in particular can reconfigure point rate in 2008 or 2009; the countries are—in gender relations. Research tracking the popular descending order of femicide rates—Honduras, usage of violence in everyday speech in post- Mexico, Finland, El Salvador, Azerbaijan, Guatemala, conflict El Salvador identifies men as having and the Dominican Republic. In some countries— ‘more “right” to use violence than women’ such as Azerbaijan and Finland—rates that had (Hume, 2008, p. 66). Indeed, prolonged repres- initially climbed may have dropped towards the sion and everyday violence affect the lives of end of the observed period. The majority of coun- many long after a conflict ends. In particular, tries under review show stable or decreasing Hume finds that: rates. Figure 4.6 displays trends in femicide rates for a sample of five countries with data for the individuals and communities learned that silence entire 2004–09 period. was the only option when no one could be trusted and violence was an ever-present possibility. They testified to feeling afraid of the orejas Male versus female victims (informers), who were often neighbors or family Available data shows that there is a correlation members (Hume, 2008, p. 71). between the rates of homicides with male victims dor, order. inthat Salva- and El Guatemala, Colombia, Venezuela, in observed being differences largest the with increase, violence overall of levels the when pan. andJa- Norway, Switzerland, inAustria, such as rare, isrelatively homicide where case the also is It Darussalam. Brunei suchas populations, male large disproportionately with countries in case the ismostly This men. for observed tothat close very appears women for rate victimization rates—the homicide low presenting those countries—especially insome Indeed, and those with female victims. female with and those the population structure of eachcountry. of structure population the reflecting partly variation, considerable shows ineachcountry ratio victim homicide to-female male- the of inspection adeeper Nevertheless, tomalevictims. respect with rates homicide high show which Islands, Virgin US andthe Venezuela, inColombia, observed be can opposite the while rates, femicide high tively rela- feature andGuyana Federation Russian The victims. maleandfemale toboth respect cide with homi- of rates highest the exhibit and Guatemala Salvador El femicide. of rates high report also victims malehomicide of levels high featuring tries Source 50 100 150 200 250 300 Femicide rate (index year 2004=100) Figure El Salvador 2004 24 GBAV 2011 database : GBAV femicide This gender gap becomes more significant significant more becomes gap gender This 4.6

Guatemala Femicide trends in ﬁve selected countries, 2004–09 (basis 2004=100) (basis 2004–09 countries, selected inﬁve trends Femicide Honduras Honduras 2005 22 Mexico In general, coun- general, In United States 23

2006 countries. Even if the proportion of women killed killed women of proportion ifthe Even countries. toother compared victims homicide female maleand of proportions higher have also rates homicide highest the exhibit that countries The review. under period the during highest the were rates femicide inwhich countries the also were they that given elsewhere, than countries inthese safer were women meanthat not does This ones. female of proportion the than greater times ten than more is victims malehomicide of proportion the Brazil, and Rico, Puerto Colombia, inVenezuela, example, For ismuchhigher. malevictims of proportion andthe ishigher, sexes between disparity the higher, are rates homicide Where killed. women of proportion andthe rates homicide between tion correla- isanegative There rates. high very with inthose cent and12.5 per rates high with tries 16.3 men: incoun- for cent per rates homicide of afraction just represent rates femicide cide rates, homi- high andvery high experiencing countries in that shows also 4.7 Yet 4.7). Figure Figure (see men for rates homicide than lower one-third just average, on are, rates femicide countries, these In smallest. appears andwomen men for rates between difference the population, total cide per homi- of rates low very feature that countries In 2007 20 2008 2009 123 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 124 Figure 4.7 Number of femicide victims for every 100 male homicide while female perpetrators represent a residual victims in 83 countries, 2004–09 portion.26 Since femicide often occurs in the family or in restricted circles close to the victim, the per- Number of femicides for every 100 male homicide victims petrator is likely to be found in most cases. Often 90 there is a previous connection with the victim, 80 most frequently at the family or intimate partner 70 level. The broad category of intimate partners 60 includes all sexual partners—such as current 50 and former spouses or partners—as well as other 40 close family members if involved in an intimate 30 relationship with the victim. Non-intimate part-

ARMED VIOLENCE 2011 VIOLENCE ARMED 20 ners include friends and acquaintances, as well of 10 as known strangers and family members.

Very low (<1) Low (1–3) Medium (3–10) High (10–30) Very high (>30) The perpetrator remains unknown if the investiga- 13 countries 30 countries 20 countries 13 countries 7 countries tion does not succeed in identifying the offender, Homicide rate per 100,000 overall population preventing the case from being ‘solved’. This is Sources: GBAV 2011 database and femicide database frequently the case with respect to deaths occurring between armed groups, during armed con- GLOBAL BURDEN BURDEN GLOBAL may appear small with respect to that of male flicts, and as a result of robberies, gang activity, homicide victims within a country, the actual shootings in public places, and other incidents femicide rate remains high. A possible explana- of a similar nature. In addition, the capacity and tion for these sustained high rates is the involve- resources necessary for effective investigations may ment of organized crime and gangs in committing also have an impact on the number of ‘unsolved’ violent acts against women and girls, especially cases (CHARACTERISTICS OF ARMED VIOLENCE). femicide.25 This is not the case in the majority Some researchers contend that women who kill of countries, where femicides apparently occur themselves in the context of protracted violence predominantly within the family or in the context or aggression should be counted as victims of of an intimate relationship (CHARACTERISTICS femicide.27 Indeed, the category of ‘forced suicide’28 OF ARMED VIOLENCE). is the frequent outcome of strong social pressure and ‘honour’-based violence, such as reported in Iraqi Kurdistan (KWRW, 2008). The consequences Characteristics of femicide of structural repression and social stigmatization in such settings can lead to enduring physical In order to better understand the factors shaping harm and serious psychological hardship.29 femicide it is important to disaggregate the actors, causes, and circumstances shaping the killing of Another lethal scenario is the killing of a rape women. Specifically, the characteristics of the victim in order to restore the family ‘honour’. For perpetrator are a central component of this clas- example, in Libya women and girls who become sification system. It is widely accepted that male pregnant through rape run the risk of being mur- perpetrators comprise the vast majority of offenders dered by a family member in so-called ‘honour’ United Nations Assistance Mission in Afghanistan: in Mission Assistance Nations United the by reported As killed. and harassed, threatened, been have positions public in women Numerous roles. traditional their from away breaking as seen are who women target persistently groups Armed 2007). (Azarbaijani-Moghaddam, society civil astrong of absence the in change social meaningful about bring to struggling been since have agencies Both 2001. in Conference Bonn the at Affairs Women’s of Ministry the and Commission Rights Human Independent Afghanistan the of creation the by evidenced as rights, women’s regarding preoccupation international growing witnessed has period Taliban post- The impunity. and widespread discrimination, conflict, armed of decades by hampered been has girls and women Afghan of security personal The B school, especially daughters’ (MOWA, 2008, p.11). 2008, (MOWA, daughters’ especially school, to children their sending fear ‘parents attacks, such of aresult As destruction. total and burning, including alone, 2007 looting, their and February 2005 July between groups insurgent by schools on 192 attacks recorded Affairs Women’s of Ministry The violence: with targeted explicitly also are girls Young p.11). 2009, violence (UNAMA, domestic women dealing with police- ten of aunit leading was she because allegedly 2008, in September killed in Kandahar, was officer police female highest-ranking the Kakar, Malalai ox 4.3 flee the country (UNAMA, 2009, p.10). 2009, (UNAMA, country the flee to had one and murdered been have three country, their reconstructing in arole take to wanted who women Afghan of stories covering in interested anewspaper by 2005 in interviewed women high-profile five Of 2011, p. 15). p. 15). 2011, (UNESCO, of Afghanistan in province Helmand schools of 70% over of closure the in resulted have fears Security particular. in schools girls’ and general in infrastructure education attacked repeatedly have groups Insurgent Insecurity and impunity inAfghanistan andimpunity Insecurity 30 however. victim, the to rights few provide mechanisms These reports: Mission UN the As intercourse. sexual extra-marital of because prosecution including criminal victimization, further risks arape reports who woman Any cally. specifi- rape and general in women against violence of incidents to widespread factor contributing amajor isconsidered impunity of level high The 21). (p. husband the of member afamily by raped being of risk at cially espe- are marriages forced in girls young very that suggests research the Nevertheless, p.21). 2009, (UNAMA, marriage to relation in especially lence’, vio- ‘sexual or ‘rape’ as such terms use to difficult’ ‘extremely it found Mission Assistance UN the stan, women oninof violenceAfghani- research against suggesting significant undercounting. In the course reported, frequently less were rape of Incidents 19). 1–2, pp. 2006, (UNIFEM, were physical followed attack, by forced marriage violence of incidents common most The cent). (82 per cousin or son, father-in-law, ahusband, as such them, to close someone by and homes their within perpetrated were women against attacks most that highlights data the of analysis An respondents. 818 from cases 1,327 recorded project the 2005, June and 2003 January Between n.d.). Afghanistan, (UNIFEM women against violence on adatabase established Afghanistan, together with participating agencies, participating with together Afghanistan, UNIFEM 2006, In women. against directed lence vio- domestic extensive from suffers Afghanistan p.8). 2011, (UNESCO, boys 100 every for enrolled are girls 66 only Today, 2009, p. 25). or ajirga through level, community the at or privately case the resolve to attempt often families rape, of stigma social the and society Afghan Coupled with the conservative nature of shura shura [a council of elders] (UNAMA, [a council of elders] (UNAMA, [a gathering of tribal elders] elders] tribal of [a gathering 31

125 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 126 ARMED VIOLENCE 2011 VIOLENCE ARMED of GLOBAL BURDEN BURDEN GLOBAL © Ash Sweeting/Panos Pictures Sweeting/Panos © Ash assassinated violence. was Kakar domestice of victims from evidence gathers Academy, Police Photo Malalai Kakar, the first woman to graduate from Kandahar Kandahar from graduate to woman first the Kakar, Malalai by the Taliban in 2008. these cases to: attributes largely Commission 4.3).Box The (see inAfghanistan andwomen girls among infrequent isnot self-immolation Commission, Rights Human Independent Afghanistan to the according example, For consequences. lasting ates gener- andgirls women of repression place, take not may suicideandfemicide forced where Even 1998; Faqir, (Ruggi, 2001). ‘honour’ family the of indefence killed be may victim rape a that confirm studies 2011). Other (Harter, killings includes suicide as a key category among seven: among category akey suicideas includes andZimbabwe— Zambia, Swaziland, Africa, South Africa—Botswana, in Southern countries infive out carried astudy from femicide of A classification assumed. often than complex more suicide are and femicide between connections the Indeed, pp. 9–10). 2010, (Psytel, sample inthe women the of cent per 42 of death cause the suicideas identifies violence spousal of context inthe femicides and homicides of a sample considers that study Union European 1996). Arecent and Flitcraft, (Stark relationship abusive their of termination the orafter during point some at suicide attempt one least madeat violence ofdomestic tims vic- of cent and40per 35 between that reveals States United inthe undertaken Research flict. con- orarmed violence homicidal chronic with confronted tocountries limited suicide isnot and violence domestic between correlation The price) and family problems (AIHRC, 2004, p.32). 2004, (AIHRC, problems family and price) as such practices customary war, of 25 years of impact psychological the rights, women’s of awareness societal of lack the tices, prac- discriminatory other as well as marriages forced marriage, premature marriage and multiple Tuyana (bride 127 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 128 intimate femicide involving women killed by IPV-related femicides current and former partners; Femicide is often linked to situations involving femicide by other (male) family members; intimate partner violence, between either spouses sexual femicide such as rape followed by or partners. An intimate partner may be defined murder; as a person with whom the victim had a physical intimate relationship, either at the time of the witch femicides; femicide or in the past. In this context, it is irrel- ritualistic femicides involving women and girls evant whether they were ever married or lived ritually killed to cut out their genital organs; together. For this reason, the term ‘intimate partner women killed by thieves or robbers; and violence’ is more apt than ‘spousal’ or ‘marital’ Photo T A victim of domestic violence takes suicides by women experiencing violence, violence. All these categories, however, may be ARMED VIOLENCE 2011 VIOLENCE ARMED

of shelter at a safe house in including cases in which there is strong evi- captured in the wider definition of ‘domestic vio- Nevada County, California, dence that women kill themselves to escape lence’. As noted in Chapter Three, the proportion August 2010. © Rich Pedroncelli/ intolerable levels of violence (Watts, Osam, of homicides occurring in the domestic sphere is AP Photo and Win, 2001, p. 91). different for males and females and is not the same GLOBAL BURDEN BURDEN GLOBAL 100,000 population, 2008 per andrate sex by States, United the in partner anintimate by committed the victim. the over power exercise or dominate, coerce, to violence use who siblings other and parents by committed to include acts couples butextends ex- or couples to limited isnot which violence, domestic of apart considered be may IPV lence. vio- severe more to serious less from escalating often incident, asingle than rather a pattern as represented isfrequently IPV 2002). (WHO, couples same-sex in other the against partner one by and men against women by violence of cases to refers also it women; against men by committed violence to limited isnot IPV partners. dating and partners, live-in spouses, former and current to refers generally partner’ ‘Intimate Figure 2005). (BJS, male victims of with cent homicides by anper intimatefive partner, comparedonly to are perpetrated femicides of cent per 30 about Furthermore, is80. men for rate the while is 430, women for rate the population, 100,000 per 260 is rate total the while average; the to respect with victims IPV the of share large a disproportionately represent women 4.8, Figure in shown as Indeed, Figure 4.8). (see partner intimate an by committed crime of victims be to men than likely more times five than more are States United the in women tics, Statis- Justice of Bureau US the to According IPV. of risk at group largest the represent Women femicide. to factor contributing amajor as violence partner intimate identify studies Many B aevcis Femalevictims Male victims Intimate partner violence Intimate partner ox 4.4 Homicide rate per 100 4.8 32

Victims of violent crime crime of violent Victims 100 200 300 400 500 ,000 malepopulation Malta Poland Sweden France Portugal Cyprus Denmark Latvia Italy Austria Serbia Finland Puerto Rico Switzerland Republic Czech Luxembourg Croatia New Zealand UK (England, Wales) Guatemala Canada Belgium Romania Hungary Ecuador Republic Dominican Peru United States Greece Norway Slovakia UK (Scotland) Australia Nicaragua Bulgaria Region Lesser Antilles Panama Chile Uruguay Slovenia Netherlands Argentina Estonia Germany Russian Federation Paraguay Spain Costa Rica Honduras Ireland Mexico Lithuania Colombia El Salvador Figure Source 2004–09 country, per femicides IPV-related of percentage estimated GBAV 2011 database : GBAV femicide 4.9

Total femicide rates per 100,000 female population and population female 100,000 per rates Total femicide 0 1020304050607080900 Total femiciderate per 100,000 f 5 Percentage ofIPV-related femicides 10 emale population (based on average) Percentage of IPV-related femicides femicides IPV-related of Percentage Total femicide rate rate femicide Total 15 20 100 25 129 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 130 in all countries. Statistically, women run a much being physically abused by the same perpetrator higher risk of being killed by their partners than (McFarlane et al., 1999; Campbell et al., 2003). do men (WHO, 2002; UNIFEM, 2006; see Box 4.4). Understanding these and other risk factors is critical when considering measures to prevent Figure 4.9 shows total rates of femicide and the and reduce femicide. estimated percentage of IPV-related femicides based on the sample of 54 countries and territo- As noted above, the total number of femicide ries for which data is available. There is a weak victims may be higher if suicides and indirect negative correlation between the two indicators casualties are also included. For example, chil- (–0.431, n=54), reﬂecting that the higher the rate dren, relatives, or other witnesses may be killed of femicide, the lower the proportion of IPV-related as a result of IPV. A study carried out in the 27 femicide. This corroborates the conclusion that countries of the European Union analyses 3,413 ARMED VIOLENCE 2011 VIOLENCE ARMED

of higher rates of victimization of women may accom- deaths resulting from domestic and spousal con- pany widespread violence rather than domestic flicts in 2006, taking into account both male and or intimate partner violence. In countries where female victims (Psytel, 2010, p. 9; see Figure 4.10). femicide rates are high, women also run a higher The findings are instructive: risk of becoming targets of violence outside the private sphere. Incidents of femicide were most common, accounting for 41 per cent of all cases. Violence committed by current and former intimate GLOBAL BURDEN BURDEN GLOBAL partners represents a risk across all countries and Some 8 per cent of the victims were men killed cultures. Strikingly, the vast majority of women by their female partners. killed in the context of IPV had previous experi- Women who committed suicide after enduring ence of domestic violence or stalking, including long-lasting domestic or spousal violence represent 30 per cent of the sample. Figure 4.10 Deaths of men and women as a In 16 per cent of the cases, perpetrators killed consequence of domestic violence in the themselves in ‘homicide–suicide’ scenarios. European Union, 2006 Collateral victims who happened to be involved Legend: in domestic or spousal incidents represent Male homicide victims (272; 8%) Male collateral homicide victims (186; 5%) 5 per cent of the sample. Male suicides after femicide (536; 16%) Femicide (1,409; 41%) In some cases—known as ‘extended suicide’, Female suicides after violence (1,010; 30%) ‘familicide’, or ‘homicide–suicide’—the perpetrator may kill his or her entire family before committing suicide. The vast majority of these incidents involve a male perpetrator who commits ‘femicide– suicide’ or, in some cases, ‘femicide and attempted suicide’; they tend to take place in the home and guns represent the instrument of choice (Auchter, 2010; Liem and Nieuwbeerta, 2010). Homicide– suicide represents an important component of Source: Small Arms Survey elaboration of Psytel (2010) lethal domestic violence. Research on this issue arm is more likely to be used to threaten and injure andinjure tothreaten used tobe likely ismore arm afire- indeed, andgirls; women against violence lethal facilitate also may home inthe weapon a of presence the 2009). Likewise, and Glass, Webster, (Campbell, afirearm with threatened being reported hadpreviously femicide of victims Many use. actual of isa predictor firearms displaying p.17). 1998, (Meloy, Nevertheless, victims’ injure than rather andintimidate control ‘to them and use weapons own may stalkers Specifically, stalking. of case inthe present frequently also are Weapons offenders. ormore one with tocomply victims tocoerce used be may Firearms 2000). Azrael, and (Hemenway victims to shoot than rather tothreaten merely used often more are Firearms 2008). al., et (Mathews higher suicideisalso commit will femicide of perpetrator the that probability the isused, arm afire- when inIPV. Moreover, especially women, against inviolence role animportant play firearms Nevertheless, frequent. isless infemicide arms fire- of use the men, involving in homicides used frequently are firearms While p.5). 2008, Secretariat, Declaration (Geneva afirearm with committed are homicides of cent per 60 approximately that mates Violence Armed of Burden Global 2008 The in femicides offirearms use The 2010). andMathews, Jewkes, Abrahams, 2010; al., et (Roberts murder the of aweek suicide within commit femicide of perpetrators of 19 cent per mately approxi- that highlighting countries, in high-income found tothose similar patterns identified has Africa inSouth Yet research countries. middle-income and low- from isavailable muchinformation Not 2003). andLehtia, Kivivuoria 2008; (Oberwittler, society of strata lower inthe than middleclass the among frequent ismore incident of type this where States, United andthe inEurope advanced is most esti- by rate of femicide per 100,000 female population. population. female 100,000 per femicide of rate by firearms with committed femicides of distribution average the shows 4.11 Figure afirearm. of use the involved in2004–09 perpetrated femicides of cent per 60 than andHonduras—more dor, Guatemala, Salva- El Colombia, Brazil, as rates—such homicide high with incountries contrast, of way By VIOLENCE). ARMED OF 3.7, CHARACTERISTICS Figure (see firearms of use rare the with correspond quently fre- rates homicide Low firearms. with committed femicides of percentage andthe rates femicide between correlation isadirect there available, is data adequate which for countries 24 the In (RCMP, 2010). period asix-year over killed women of number the of astudy on based tragically, ending disputes domestic of risk the reduced has inhouseholds tofirearms access limiting policy gun that demonstrated has study p.7). ACanadian 2011, (Hemenway, intruders from home the toprotect than members family Source 10 20 30 40 50 60 70 80 90 100 Percentage offemicidescommitted with afirearm 2004–09 rate, femicide by countries, 24 Figure GBAV 2011 database : GBAV femicide Very low(<1) 6 countries 4.11

Average percentage of femicides committed with firearms in firearms with committed femicides of percentage Average 7 countries Low (1–2) Femicide rate per 100,000 femalepopulation Medium (2–3) 4 countries 2 countries High (3–6) Very high(>6) 5 countries 131 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 132 Figure 4.11 also reveals that the use of firearms is the national average for both male and female much more frequent in countries that exhibit very victims. In 2007–09, the percentage of femicides high femicide rates. Among all countries under committed with firearms increased dramatically, review, an average of one-third of all femicides reaching 82 per cent in 2009 and thus nearing were committed with firearms. the percentage observed in homicides with male Figure 4.12 highlights the difference between the victims (89 per cent). use of firearms in homicides with male vs. female In the United States, data from the Centers for victims in Mexico on the whole and in Ciudad Juarez Disease Control and Prevention shows that firearms from 1993 to 2009. On average, approximately were used in approximately half of all recorded 60 per cent of homicides involving a male victim femicides in 1999–2007 (CDC, n.d.a; see Figure were committed with a firearm; in contrast, fewer 4.13). Most of the femicides by firearm occurred ARMED VIOLENCE 2011 VIOLENCE ARMED than 40 per cent of femicides involved the use of a

of in the age group of 15–24-year-olds, in which they gun. Since 2005, there has been a marked increase account for 59 per cent of the cases. This suggests in the proportion of homicides committed using that firearm use may be linked to the age of the a firearm, with more than two-thirds of male vic- victim. Firearms were also used to kill more than tims killed by firearms in 2009. one-third of the victims aged 5–9 and almost half The extremely high levels of violence in Ciudad of those aged 10–14. Indeed, ‘children aged 5 to Juarez, capital of Chihuahua state, are reflected 14 years in the United States have 11 times the

GLOBAL BURDEN BURDEN GLOBAL in the very high proportion of homicides commit- likelihood of being killed accidentally with a gun ted with ﬁrearms, which is generally higher than compared with similarly aged children in other

Figure 4.12 Percentage of male and female homicide victims killed by firearm, Mexico and Ciudad Juarez, 1993–2009 % of male homicide victims killed with a firearm in Mexico % of female homicide victims killed with a firearm in Mexico % of male homicide victims killed with a firearm in Ciudad Juarez % of female homicide victims killed with a firearm in Ciudad Juarez

Percentage of homicides committed with a ﬁrearm

100

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Source: Small Arms Survey elaboration based on INEGI (2009) across Asia and Africa—and among diasporas— among andAfrica—and Asia across countries different inmany occur practices related andother These 4.5). cide (HRCP, Box see 2011; homi- as seriousness same the with punished or procedures justice tocriminal subject be not may deaths anddowry killings ‘honour’ Specifically, level. statistical national andeven aglobal at ture tocap- difficult be may that violence non-lethal and lethal of examples considers section This 2000). 1995; 1994; (UN, in2000 andSecurity’ Peace and 1325 ‘Women on Resolution Council Security inUN culminating Declaration, Beijing and the Women against Discrimination of Forms All of tion Elimina- the on Convention the with continued it 1993); (UN, Women against Violence of Elimination the on Declaration the with started only process slow.Asteady quite been has violence from them andtoprotect women for rights equal tee toguaran- debate aninternational of development The misbehaviour. awoman’s perceived sanction to violence of use the foresees that inplace tion legisla- have still countries Some and condoned. istolerated andgirls women against violence texts, con- some In systems. reporting of accessibility poor andduetothe toreport victims of reluctance the of a undocumented as result goes violence frequently such isthat fact The women. against and violence femicide of andbreadth extent the underestimates above presented data the that doubt islittle There women against Hidden oflethal forms violence p. 358).2003, andCasteel, Moracco, (Coyne-Beasley, a firearm with behaviour reckless of aresult as often dence, insidearesi- place took frequently most firearm a by killed girls involving incidents Furthermore, p.2). 2011, (Hemenway, countries’ developed have damaged the reputation of the household, household, the of reputation the damaged have to judged behaviour of arange for punishment as malerelatives by eachyear killed cases some andin toviolence, exposed abused, are girls and women 5,000 around that estimates Fund Population ‘honour’. UN damaged family the The or customs social on infringed allegedly they because relatives by orkilled toviolence jected sub- being women of examples myriad are There 62–63). 2010b, paras. (UN, killings’ orwitchcraft violence femicide, domestic honour killings, concerning ‘whether andkillings, violence gender-based linkbetween the on research more encourages Executions or Arbitrary Summary Extrajudicial, on Rapporteur Special the report, inhis2010 gap this Acknowledging acrime. constitutes even circumstances certain suicideunder tocommit them orforcing women killing that consider not may tions jurisdic- 2008). Some Watch, (Gendercide Americas andthe inEurope countries incertain as well as 10 20 30 40 50 60 70 80 90 100 Percentage offemicidescommitted with afirearm 1999–2007 group, age by States, United Figure Source Small Arms Survey elaboration of CDC (n.d.a) CDC of elaboration Survey Arms : Small 4.13 1–4

Percentage of femicides committed with ﬁrearms in the inthe ﬁrearms with committed femicides of Percentage 5–9 1–41–42–435–44 25–34 15–24 10–14 Age group 133 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 134 clan, or tribe (UNFPA, 2003). Women who demon- Box 4.5 Dowry deaths in India strate intolerance towards traditional restrictions, The home of the in-laws in South Asia can be a particularly dangerous especially as regards the choice of friends or part- environment for women and girls. Survey data suggests that the vast majority ners and sexual behaviour, are especially at risk. of reported murders of women in India are incidents of husbands killing their wives, with 85 per cent of female victims killed in their own home (Mohanty Another scenario in which women and children et al., 2004; UNFPA, 2003). may be killed is when they are believed to be Marriages involving dowry and other demands from the husband are not practising witchcraft. According to a 2002 report only a risk factor for domestic violence, but also a direct cause of violent by Radhika Coomaraswamy, the then Special death for women. While dowry payment is illegal in India, it remains common practice, with related disputes or violent demands for more money Rapporteur on Violence against Women, Its frequently leading to the death of brides or brides-to-be (Ash, 2003). Causes and Consequences, the practice is found

ARMED VIOLENCE 2011 VIOLENCE ARMED Of India’s reported 32,369 homicide victims in 2009, around one-quarter were ‘mainly in Asian and African communities’ (UN, of female (NCRB, n.d.b). ‘Dowry deaths’ under Section 304B of the Indian Penal 2002, p. 16). Code are recorded separately.33 In 2009, 8,383 cases were ﬁled under that heading, which may include more cases of violent deaths of women and While victims can be male or female, the majority girls than those recorded under the homicide category. Taken together, of witch killings across societies appear to target dowry killings may represent up to 26 per cent of violent deaths and more women. Reports of some 50 ritual killings in Papua than 50 per cent all violent deaths of women and girls in India in 2009.34 New Guinea in 2009, for example, provide accounts The number of police-recorded dowry killings in India has risen by 25 per of young and old victims, mostly women and girls cent since 1999 (NCRB, n.d.b, p. 58). Whether the rise is due to an increase GLOBAL BURDEN BURDEN GLOBAL in police and law enforcement attention to the issue or a real underlying (AI, 2009; Parry, 2009). In Tanzania, up to 1,000 increase is unclear. The latter explanation is supported by a retrospective persons are reportedly killed every year based hospital study of female homicide victims, which estimates that 30 per on allegations of witchcraft; the majority of the cent of those deaths are dowry-related (Mohanty et al., 2004, p. 153). victims are women above the age of 50 (HAI, 2008, Source: Malby (2011) p. 7). According to a study carried out in Ghana, being aware of the status of women and girls in a society is crucial to understanding witch-related femicides (Adinkrah, 2004). The Special Rappor- teur also points out that witch killing is often the result of highly unequal gender relations in a society (UN, 2002, p. 17).

Research suggests that witch murders are often linked to poverty and situations of economic despair. One study carried out in Tanzania assesses the number of witch murders in connection with extreme rainfalls that lead to droughts or floods. It finds that the link is significant: ‘There are twice

Photo W A newlywed stands in a shelter for victims of dowry violence next to a poster highlighting the economic demands some new brides face from their husbands and in-laws, New Delhi, June 2004. © Elizabeth Dalziel/AP Photo stoning, orbeating. stoning, includingburning, forms, ritual take may Killings ways. of inavariety killed are witches Alleged suffering: economic of result the be well may they that gest sug- inIndia killings witch of reports Similarly, p.1153). 2005, (Miguel, years’ inother as fall rain- extreme of inyears murders witch many as grammes and interventions. andinterventions. grammes pro- prevention of effectiveness the and limits HIV/AIDS of spread tothe contributes assistance andmedical information of Alack them. against violence further use may who partner, the of cion suspi- raise would care medical for asking which in relationships inabusive live Many incidents. orreport help seek rarely violence of Victims problem. the of extent the toquantify difﬁcult 2010). isextremely It Moreno, andGarcía (Dunkle because of the likelihood of tissue laceration transmission virus of chances the increases sex forced Organization, Health World tothe According virus. the contracting of risk high at are violence sexual of victims andwomen transmission HIV of majormode isthe sex HIV/AIDS. Unprotected with infection isthrough killed unintentionally and intentionally are women that way Another are abducted and killed in order to sell their organs. their tosell inorder andkilled abducted are victims suchscenarios, 2009). In (Kelly, in organs trafﬁcking for covers as acting rituals witchcraft of risk tothe attention called have reports eral the problems (UN, 2002, p.17). 2002, (UN, problems the iscausing that spirit evil the of rid be to death, through elimination their suggests and ‘dayans’ or witches be to women or awoman declares usually [...]He isapproached. doctor magic local the calamity, isanatural there if even or failure, isacrop there if or die, cattle or family, the in isadeath there if or water, drinking of a lack is there if or illnesses, from suffer people When 35 In the past few years, sev- years, few past the In 36

enced physical violence since they turned 15 turned they since violence physical enced experi- have they say who women of proportion 2008–09 The Moreno, 2010). García and (Dunkle status HIV their to due violence experience to likely more are HIV with women living In addition, decision-making. sexual healthy assert or infection HIV from selves them- protect to adequately women’s ability hamper can violence of threat the and Violence 2009). (Machera, partners their by abuse of forms many to subjected are women HIV-positive that reveals Kenya in out carried study A recent infection. HIV of source isamajor partner, male the by initiated usually violence, sexual spousal that reality: along-hidden confirm ings find- These 2009). NACC, (Kenya partnerships regular and unions existing within sex erosexual het- to attributed be can incidence HIV national of cent per 44.1 that finds study Arecent ences. such andviolence experi- silent to about remain domestic rationalize even and tolerate, accept, to socialized are women elsewhere, as Kenya, In high. remained women for risk the subsided, violence election post- the after Even (COVAW, 2008). rape gang of incidence elevated the to due people placed dis- 350,000 approximately the among serious extremely was situation the Specifically, others. than transmission HIV/AIDS and violence sexual to vulnerable more were women of groups Some 2008). (Holmes, assistance ing seek- victims rape of number the in surge a rapid with confronted soon were Hospitals rape. mass of in suddenly, waves ushering erupted Violence 2008. early and 2007 December of violence post-election the during HIV/AIDS contracting of risk high especially at were Kenya in girls and women that is evidence There pandemic. HIV/AIDS the of vector akey also but problem, health apublic only isnot women against Violence B ox 4.6 Kenya Demographic and Health Survey Survey Health and Demographic Kenya HIV and sexual violence inKenya violence andsexual HIV documents a sizeable reduction in the the in reduction asizeable documents XX

135 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 136 years old (KNBS, 2010). The survey finds a reduction from 49 per cent in 2003 to 39 per cent in 2008–09. Even so, at least 12 per cent of women aged 15–49 reported that their first sexual intercourse was forced against their will. Moreover, overall, one in five Kenyan women (21 per cent) experienced sexual violence. Analysis across provinces indicates that the two provinces with the highest proportion of women experiencing physical violence (Nyanza and Western provinces) also featured the highest proportion of women experiencing sexual violence.

Among surveyed women who had been married at least once, sexual violence was reportedly perpetrated mainly by current and former husbands and partners. Those who never married said that violence was committed ARMED VIOLENCE 2011 VIOLENCE ARMED

of mainly by boyfriends, although almost one in ﬁve never-married women (19 per cent) was violated by a friend or acquaintance and almost as many by a stranger (17 per cent). The likelihood of experiencing either physical or sexual violence increases with the age of the women. That said, women with secondary or higher education and those in the top two wealth quintiles are less likely to experience sexual violence than other women.

GLOBAL BURDEN BURDEN GLOBAL In the vast majority of cases, sexual violence is perpetrated by persons known to the victims (Machera, 2009). Indeed, strangers commit only 6 per cent of recorded sexual violence. About 37 per cent of women who experienced sexual violence reported current husbands or partners as the perpetrators, followed by current or former boyfriends (16 per cent) and former husbands or partners (13 per cent). Women who have experienced both physical and sexual violence are more likely to seek help than those who experienced only one or the other (KNBS, 2010). Older women are more likely than younger women to seek help to stop the violence.

To reduce sexual violence the Kenyan government has enacted the Sexual Offences Act No. 3 of 2007 (KNBS, 2010). The law has been lauded as a move in the right direction, although its implementation remains slow. For example, the cases brought by women and girls who were raped during the post-election violence in 2007–08 have not yet been addressed by the courts. Mechanisms for retrieving evidence and tracking down perpetrators are generally substandard. Consequently, the law has not yet succeeded in deterring rapists.

Source: Machera (2011)

Photo X A woman stands outside an office dealing with child abuse and violence against women at a police station in Mombasa, Kenya, 2005. © Sven Torfinn/Panos Pictures 137 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 138 The level of brutality has been heightened through of women and pervasive gender inequality. In the systematic use of rape as a weapon of war these places, women and girls cannot enjoy a safe and perpetrated by soldiers who carry the HIV/ or secure lifestyle. The chapter calls attention to AIDS virus. Indeed, the widespread and system- incidents occurring inside as well as outside the atic targeting of civilians and the use of rape is a domestic sphere. Women are vulnerable to vio- striking aspect of recent armed conflicts in Sub- lence committed by strangers, but more frequently Saharan Africa (Elbe, 2002, p. 168; see Box 4.6). they are unsafe in their own homes. In other parts of the world, rape has also been A key conclusion, then, is that there is a need to used as a systematic weapon of terror leading to enhance reporting and analysis of data on lethal the spread of HIV/AIDS. During the military rule and non-lethal violence against women—both cross- between 1991 and 1994 in Haiti, for example, nationally and sub-nationally. This goal could be

ARMED VIOLENCE 2011 VIOLENCE ARMED women were raped because of the alleged politi- achieved by undertaking steps such as those pro- of cal activities of their husbands. The perpetrators posed by the Geneva Declaration study on Tackling were reportedly ‘police, soldiers and criminal Violence against Women: From Knowledge to Prac- gangs operating with impunity’ (Bastick, Grimm, tical Initiatives (Milliken, Gilgen, and Lazarevic, and Kunz, 2001, p. 79). Twenty per cent of police 2011). Practical recommendations include: officers reportedly suffered from HIV/AIDS in Haiti at the time. As a consequence, in 2001 Haiti had supporting international initiatives to track the highest prevalence of HIV/AIDS in the Western violence against women globally; GLOBAL BURDEN BURDEN GLOBAL hemisphere (p. 79). the promotion of field-based research on mapping violence against women, including surveys to measure violence committed by Conclusion intimate partners and strangers; and The critical role of collecting data on lethal vio- the development of costing tools to improve lence against women cannot be overstated. And estimates of the effects of violence against while the evidence base is growing to demonstrate women on development, including direct and the scale and distribution of femicide and violence indirect costs of violence against women. against women, this chapter has underlined critical information gaps, especially across Africa and Asia. The fact is that the quantity and quality of Abbreviations data on femicide are very poor and characterized INEGI Instituto Nacional de Estadística y Geografía (Mexico) by incomplete geographical coverage. Reliable IPV Intimate partner violence and valid information on violence according to sex, age, relationship to the perpetrator, and instrument used is crucial to designing effective violence prevention and reduction strategies. Endnotes In settings where reported levels of violence are 1 Figures based on a report generated at CDC (n.d.b). high, femicide levels are also likely to be high. 2 The findings are based on a 2007 survey that estimates that between 1.7 and 1.8 million women were raped in Similarly, these environments are likely to be their lifetime in the Democratic Republic of the Congo characterized by the systematic discrimination (Peterman, Palermo, and Bredenkamp, 2011). 3 See Small Arms (2006)foranassessmentSurvey of injury 13 For example, the World Bank recently provided funding 12 For methods of calculation, seetheonline methodological 11 Femicide rates represent anaverage over theperiod 10 Thefact that this dataset contains asmaller numberof 9 As inChapter Two, several smaller Caribbean islands have 8 For 7 See,forexample, thecompilation of data at UNECE (n.d.) deta See,forexample, UNECE (n.d.). 6 5 Russell alsonotes that certain dictionaries define‘femicide’ 4 Thedatapresent 3 collective violence. due to‘external’ cause light thescale anddistribution of mortality andmorbidity preventionontheAfricaninjury continent and tohigh- of individuals andinstitutions for Africa. The Initiativ surveillance systems fr prevention surveillance systems inAfrica. see Zavala andHazen (2009) about theroleof injury for therestructuring of theNigeri annexe at www.genevadeclaration.org. gaps indata series. and lowpoints as well as thedistortions resultingfrom 2004–09. This ‘smoothing’ of data reduces extremehigh the information,higher collection methods indifferent co due totheuseofFurthermore, varying definitions ordata is extremely scarce at theint the capacity orresources tocompile such information, which on thecircumstances of homicide. Many countries may lack the difficulty inherent incapturing additional information countries than does theonef the Maldives,Monaco, andQatar. from theanal countries that nofemale reported victims areexcluded and theGrenadines, and Trinidad and Tobago. The five nica, Grenada, St. Kitts andNevis, St. Lucia, St. Vincent they areAnguilla, Antiguaand been groupedtogether as the‘Lesser Antilles Region’; declaration.org. see theonline methodological annexe at www.geneva and that presentedin etSanmartín al.(2010). as ‘thekilling of awoman’ (Russell, 2008,p. 3). Union (UNECE, n.d.). statistics, covering 29 countrie Economic Comm 44 countries et (Sanmartín al.,2010);c)theUnitedNations carried out by theQueen Sofia Center in Spain, covering age andsex (Bhalla et al.,2011);b)astudy onfemicide public health data on violent deaths disaggregated by University,Harvard coverin database developedby theHomicide Advisory Groupat obtained fromthefollowing international sources: a)the ils onpopulationdataand ysis; thecou ission forEu ed inthischapter hasla e wasfou s, includinginterpersonal and om the Injury PreventionInitiative om theInjury ntries areAndorra, g 96countries andproviding ernational comparative level. or general femicide reflects rope databa s, mostly fromtheEuropean involved in violence and Barbuda, Barbados, Domi- nded in1997asa untries, the more detailed untries, themore detailed risk of incomparability. a Bureau of Statistics; regional c rgely been rgely been se ongender lassific Liechtenstein, network ations, ations, 1 Anotherchallenge f 21 Inordertocompare trend 20 See,forexample, Suar 19 Seethe2009 18 Adetailed breakdown of femicide rates at thestate and 17 Seetheonline methodological annexe at www.geneva 16 15 The anal 4 Regional homicide rates presentedheredonot correspond 14 4 Thesecountries exhibit someof thelowest homicide 24 Thepopulation sex ratio varies significantly across coun- 23 Whilethereis of adearth homicide data onwomenin 22 each country over theperiod2004–09. in the year 2004. Lines showperc corresponding tothenumberof femicides inthecountry terns over fromanormalized timestarting value of 100, interference of thedifferent levels,thechart shows pat- pp. 21–23). in thecountry andespecially inCiudad Juarez (UN,2011, Mexico respondtocontinued violence against women Causes andConsequences, Rashida Manjoo, that Special on Rapporteur Violence against Women, Its (SRE, 2009, p. 1);seealsother disappearances andkillings of womeninCiudad Juarez Human Rights holdingMexico responsible forunsolved trend data is frequently availab which limitsthepotential base includes fewcountriesvery with complete timeseries, represented by timeseries. The (2010). Ministry of Justice in Waiselfisz (2011). municipal levels inBrazil is presentedby theBrazilian declaration.org. Southern Africa. AND PATTERNS). calculated based onalarger numberof countries (TRENDS ter Two, which alsoincludedir to rates of violent deaths perregioninFigure2.3 inChap- small numberof cases. rates, inwhich there may belarge fluctuations duetothe AND PATTERNS). Switzerland (0.83), Norway (0.69), andJapan (0.45) (TRENDS rates intheworld: Austria (0.68 per100,000 population), (92) (UNdata, n.d.) (86 menper100women),Belarus (87), andCape Verde with alarger female population areLatvia andEstonia men per100women,respectively. Incontrast, countries of male population, with ratios of 146, 135, 121, and106 Arabia, andBrunei Darussalam have astrongprevalence female population. For example, Kuwait, Bahrain, Saudi larger differences, showing anexcess of eithermale or women for2010(UNdata, n.d.). Some countries have tries. The global ratio is estimated at 101.7 menper100 also limited. many countries, relevant information onmale victims is ysis inc decision oftheInt ludes datafromonlyin onecountry or statistical ez andJordan(2007) andRNOCDH oftrendsan s across countries without the ect conflict deaths andare le forcountries with low eiterated requests of the analysis GBAV 2011femicide data- er-American Court of entages of change for alysis. Fu of femicideis rthermore, rthermore, 139 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 140 25 See, for example, Suarez and Jordan (2007) on the involvement of organized crime in femicide in Guatemala. Bibliography Abrahams, Naeemah, Rachel Jewkes, and Shanaaz Mathews. 26 Statistically, female authors of femicide, either in same- 2010. ‘Guns and Gender-based Violence in South Africa.’ sex couples or in other circumstances, represent a very small portion of the total, for example less than 1 per cent South African Medical Journal. September, pp. 586–88. in the United States. See Glass et al. (2004). 27 Sexual abuse is a strong predictor of suicide; see, for example, McFarlane and Malecha (2005). ABS (Australian Bureau of Statistics). 2009. ‘4510.0—Recorded Crime: Victims, Australia, 2009.’ 28 The definition of ‘forced suicide’ is commonly used and has been adopted in the context of violence against women its Causes and Consequences. See UN (2008). Adinkrah, Mensah. 2004. ‘Witchcraft Accusations and Female 29 Any form of violence against women is likely to be reiter- Homicide Victimization in Contemporary Ghana.’ Violence ated and to have a long-lasting effect on the victim. This against Women, Vol. 10, No. 4, pp. 325–56.

ARMED VIOLENCE 2011 VIOLENCE ARMED generates psychological effects ranging from low self- ncjrs.gov/App/publications/Abstract.aspx?id=205819>

of esteem to the development of aggressive or criminal AI (Amnesty International). 2009. ‘Papua New Guinea: Authori- behaviour and, very frequently, self-inﬂicted violence. ties Must Act Now to Prevent More Witch Hunts.’ When the violence comes from the surrounding community See, for example, Ho (2008), Dubow, Huesmann, and Boxer (2009), and Leslie (2000) on the long-term psychological AIHRC (Afghanistan Independent Human Rights Commission). effects of violence against women in conﬂict settings. 2004. Annual Report.

GLOBAL BURDEN BURDEN GLOBAL 31 The agencies are the Ministries of Women’s Affairs, Interior, Alvazzi del Frate, Anna. 2010. ‘Crime and Criminal Justice Statis- Health, Education, and Justice, as well as the courts, tics Challenges.’ In Stefan Harrendorf, Markku Heiskanen, women’s shuras (councils of elders), provincial councils, and Steven Malby, eds. International Statistics on Crime the Afghanistan Independent Human Rights Commission, and Justice. Helsinki: European Institute for Crime Preven- referral centres, Save the Children, legal aid providers, tion and Control (HEUNI) and United Nations Office on and the UN Assistance Mission in Afghanistan. Drugs and Crime, pp. 167–75. 32 See Catalano et al. (2009). Victimization rates are per Arroyo, Mario. 2011. Femicides in Mexico and Ciudad Juarez. 100,000 persons age 12 or older. The difference between Unpublished background paper. Geneva: Small Arms Survey. male and female intimate partner victimization rates is Ash, Lucy. 2003. ‘India’s Dowry Deaths.’ BBC News. 16 July. significant at the 95 per cent confidence level. ‘where the death of a woman is caused by any burns or bodily injury or occurs otherwise than under normal cir- Auchter, Bernie. 2010. ‘Men Who Murder Their Families: What the cumstances within seven years of her marriage and it is Research Tells Us.’ NIJ Journal, No. 266. June, pp. 10–12. shown that soon before her death she was subjected to cruelty or harassment by her husband or any relative of Azarbaijani-Moghaddam, Sippi. 2007. ‘On Living with Negative her husband for, or in connection with, any demand for Peace and a Half-Built State: Gender and Human Rights.’ dowry, such death shall be called “dowry death” and such International Peacekeeping, Vol. 14, No. 1, pp. 127–42. husband or relative shall be deemed to have caused her death’ (India, 1860). Bastick, Megan, Karin Grimm, and Rahel Kunz. 2001. Sexual Violence in Armed Conflict: Global Overview and Implica- 34 NCRB (n.d.a; n.d.b). tions for the Security Sector. Geneva: Geneva Centre for 35 See Watts, Osam, and Win (2001, p. 91) about the stoning the Democratic Control of Armed Forces. of accused witches in Zimbabwe. See Schnoebelen (2009) for details on the beating and burning of so-called witches Bhalla, Kavi, et al. 2011. Country-level Estimates of Homicides in Angola. from Death Registration Data. Unpublished background paper. Geneva: Small Arms Survey. 36 In the last few years, the connection between trafficking in women and girls and trafficking in human organs has BJS (Bureau of Justice Statistics). 2005. ‘Homicide Trends in the moved up on the international agenda. See Pearson (2004) U.S.: Intimate Homicide.’ Geneva: World Health Organization andUNAIDS. Violence against Women andHIV/AIDS: What Works? and Family Psychology Review on Children’s Psychosocial Adjustment.’ Impact of ExposuretoPersi Cognitive-Ecological Framew April, pp. 355–60. Archives ofPediatrics &Adolescent Medicine, Vol.157. 2003. ‘Adolescent NGO/Disp-VAW_RapidAssess_COVAW_2008.pdf> Kisii, January–February 2008. IDPs Camps inWestern Kenya—Kisumu, Kakamega and Rapid Assessment of the Situation ofWomen andGirls in Violent Death Reporting System (NVDRS).’ ment of Justice. Victims ofViolence Bureau ofJustice Statistics—Selected Findings:Female org/cgi/content/abstract/93/7/1089> No. 7. July, pp. 1089–97. Chapel Hill, NC:MEASURE Evaluation. A Compendium ofMonitoring

American Journal ofPublic Health ding Causes ofDe Violence against Women andGirls: Control andPrevention).n.d.a. ‘National . Washington, DC:United States Depart- Femicide: APopulation-Based Study.’ a GarcíaMoreno.2010. gov/content/pub/pdf/fvv.pdf> gov/content/pub/pdf/cv08.pdf> Webster,and org/cgi/reprint/157/4/355.pdf> stent Ethnic–Political Violence ork forUnderstanding the

and Evaluation Indicators , Vol. 12,No.2,pp. 113–26. assets/file/Resources/ ath Reports, ath Reports, ehealth/publications/ rs/nvdrsDisplay.jsp> Nancy G Vol. 24,No.4, Clinical Child Addressing lass. 1999–2007.’ Kenya: , Vol.93, . Elbe, Stefan. 2002. ‘HIV/AIDS andtheChanging Landscape of War ECOSOC (Economic and Social Council). 2004. Resolution 2004/22 FBI (Federal Bureau of Investigation). 2010. ‘ExpandedHomicide Faqir, Fadia. 2001.‘Intrafamily Femicide inDefence of Honour: Eriksson Baaz, Maria and Maria Stern. 2010. Human Organs. Africa Institute. American Journal ofLifestyle Medicine BBC. 14June. ngos/HAITanzania41.pdf> armssurvey.org/fileadmin/docs/A-Yearbook/2006/en/ Oxford University Press, pp. 189–213. e_against_migr ov/ucr/cius200 Third World Quarterly ating andPunishing Trafficking in ide.org/case_honour.html> ant_women_35725562 a-portal.org/smash/record. . Copenhagen: Euro- , Vol. 27, No.2,pp. 159–77. GlobalBurden ofArmed orld-africa-13760895> Matters:Assessing the 9/offenses/expanded_ . SIDAWorking Paper Violence against . The Complexity of , Vol. 22,No.1, 0.pdf> 141 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 142 — and Deborah Azrael. 2000. ‘The Relative Frequency of Offensive Liem, Marieke and Paul Nieuwbeerta. 2010. ‘Homicide Followed and Defensive Gun Use: Results of a National Survey.’ by Suicide: A Comparison with Homicide and Suicide.’ Violence and Victims, Vol. 15, pp. 257–72. Suicide and Life-Threatening Behavior, Vol. 40, No. 2. Ho, Joyce. 2008. ‘Community Violence Exposure of Southeast Asian American Adolescents.’ Journal of Interpersonal Violence, Vol. 23, No. 1, pp. 136–46. Machera, Mumbi. 2009. Linkages between Violence against Women and HIV/AIDS among WOFAK Beneficiaries. Nairobi: Holmes, Stephanie. 2008. ‘Gang Rape Spirals in Violent Kenya.’ Women Fighting AIDS in Kenya and ActionAid International. BBC News. 23 January. Study_by_AAK_2009.pdf> HRCP (Human Rights Commission of Pakistan). 2011. State of —. 2011. ‘Gender-based Violence in Kenya.’ Unpublished back- Human Rights in 2010. Lahore: HRCP. April. ground paper. Geneva: Small Arms Survey. Malby, Steven. 2011. ‘Dowry Deaths in India.’ Unpublished back- Hume, Mo. 2008. ‘The Myths of Violence: Gender, Conflict, and ground paper. Geneva: Small Arms Survey. ARMED VIOLENCE 2011 VIOLENCE ARMED Community in El Salvador.’ Latin American Perspectives, of Vol. 35, No. 5, pp. 59–76. Mathews, Shanaaz, et al. 2008. ‘Intimate Femicide–suicide in South Africa: A Cross-sectional Study.’ Bulletin of the World Independent (UK). 2008. ‘Women Who Took on the Taliban—and Health Organization, Vol. 86, pp. 552–58. McFarlane, Judith and Ann Malecha. 2005. ‘Sexual Assault among Intimates: Frequency, Consequences & Treatments.’ India. 1860. ‘Section 304 B in the Indian Penal Code, 1860.’ Washington, DC: United States Department of Justice. INEGI (Instituto Nacional de Estadística y Geografía). 2009.

GLOBAL BURDEN BURDEN GLOBAL McFarlane, Judith, et al. 1999. ‘Stalking and Intimate Partner Exchange of Data on Homicides in Mexico. Unpublished Femicide.’ Homicide Studies, Vol. 3, No. 4. November, background paper. Geneva: Small Arms Survey. pp. 300–16. Kelly, Annie. 2009. ‘Child Sacrifice and Ritual Murders Rise in ed. The Psychology of Stalking: Clinical and Forensic Kenya NACC (National AIDS Control Council). 2009. Kenya: HIV Perspectives. San Diego: Academic Press, pp. 2–27. Prevention Response and Modes of Transmission Analysis. onepage&q&f=false> Kivivuoria, Janne and Martti Lehtia. 2003. ‘Homicide Followed Miguel, Edward. 2005. ‘Poverty and Witch Killing.’ Review of by Suicide in Finland: Trend and Social Locus.’ Journal of Economic Studies, Vol. 72, pp. 1153–72. Vol. 4, No. 2, pp. 223–36. Milliken, Jennifer, Elisabeth Gilgen, and Jasna Lazarevic. 2011. KNBS (Kenya National Bureau of Statistics). 2010. Kenya Demo- Tackling Violence against Women: From Knowledge to graphic and Health Survey 2008–09. Calverton, MD: KNBS Practical Initiatives. Geneva Declaration Working Paper. and ICF Macro. Geneva: Geneva Declaration Secretariat. KWRW (Kurdish Women’s Rights Watch). 2008. ‘Reports of Honour- based Murders and Suspected Murders.’ April. Mohanty, Manoj Kumar, et al. 2004. ‘Victimologic Study of Female Homicide.’ Legal Medicine, Vol. 6, No. 3, pp. 151–56. Leslie, Helen. 2000. ‘Conceptualising and Addressing the Mental Molloy, Molly. 2010. ‘Juárez Murders: Impunity Regardless of Health Impacts of Gender Roles in Conflict and Peace- Gender.’ Grassroots Press. 12 May. making.’ Development Bulletin, No. 53, pp. 65–69. Afghanistan: Baseline Statistics on Gender. Kabul: MOWA RNOCDH (Red Nacional deOrganismos Civiles de Derechos RCMP (Royal Canadian Mounted Police). 2010. Radford, Jill Psytel. 2010. Estimation delamortalitéliéeaux violences conju- Peterman, Amber, Tia Palermo Pearson, Elaine. 2004. Pearce, Diana. 1978. ‘TheFeminization of Poverty: Women, Work Parry, Richard Lloyd.2009. ‘“Witches” Put toDeath inPapua Oberwittler, Dietrich. 2008. —. n.d.b. ‘Table 3.3: Age Group Wise Victims of Murder(Section NCRB (National Criminal Records Bureau of India).n.d.a. ‘Table Pursuant to Article40of the I Consideration as Part ofits Fifth Periodical Report onMexico ‘All Rights for All’) to the UNHumanRights Committee for (National Network ofHumanRights Civil Organizations Humanos ‘Todos los Derechos para Todas y Todos’ the Red NacionaldeOrganismos Civiles deDerechos Information Presented by Organizations FormingPart of Humanos ‘Todos los Derechos para Todas y Todos’). 2010. 2007/DAP-1/140. scientifique. Rapport gales enEurope: ‘IPVEU Mortality Journal ofPublic Health Women intheDemocratic Republic of Congo.’ ‘Estimates andDeterminants of Sexual Violence against 2004.pdf> 13 February. 302 IPC)during2009.’ Amounting toMurder(C.H.)during2009.’ Afghanistan. Femicide: ThePolitics The Times(UK). . February. RCMP Canadian American Vol.11, Eschborn: UN (UnitedNations). 1993. Jordan.2007.Suarez, JulieandMarty Stark, Evan andAnneFlitcraft. 1996. SRE (Secretaría deRelaciones Small Arms Survey. 2006. Schnoebelen, Jill. 2009. José,etSanmartín, al.,eds.2010. Sánchez, Jessica. 2011.InformeFinaldeFemicidios enHonduras Russell, Diana. 2008. ‘Femicide: Politiciz Ruggi, Suzanne. 1998. ‘Commodifying HonorinFemale Sexuality: Roberts, Kerry, et al.2010. ‘Homicide–Suicide inDurban, South yearbook/small-arms-survey-2006.html> RighttoLive/ThreethousandandCounting,AReportonViolence USA. General deDerechos Humanos y Democracia. informativo No.160. 24December. Mexico City: Dirección Humanos: Agenda Internacional bodies/hrc/docs/ngos/RNOCDH_mexico98.pdf> Civil andPolitical Rights. for Refugees. Vol. 16. Washington, DC.May. Meeting on Strengthening Paper presentedat theInter pp. 12–15. Honor Killings inPalestine.’ pp. 877–99. Africa.’ Journal ofInterpersonalViolence, Witchcraft Allegations, RefugeeProtection Small Arms Survey 2006: Unfinished Declaration on the Elimination of emala Hum Exteriores). 2009. ‘Derechos e against Women inGuatemala. . Thousand Oaks,London, CA, sh/law/eliminationvaw.htm> Three Thousand andCount- nasofia.es/publicaciones. Women at Risk:Domestic Nations ing theKilling of Females.’ deMexico.’ Boletín an RightsC High Comm issioner . Tegucigalpa: Vol. 25, No.5, . Research ommission/ , No.206, 143 2 3 4 5 1 WHEN THE VICTIM IS A WOMAN 144 —. 1994. Convention on the Elimination of All Forms of Discrimi- UNDP (United Nations Development Programme). 2008. ‘Sexual nation against Women (CEDAW). A/RES/49/164. Violence against Women and Children in Armed Conflict.’ Background document presented at the 2008 Parliamentary —. 1995. Report of the Fourth World Conference on Women— Hearing at the United Nations. New York, 20–21 November. Beijing, 4–15 September 1995. A/CONF.177/20. —. 2000. Security Council Resolution 1325 on ‘Women and peace UNECE (United Nations Economic Commission for Europe). n.d. and security’. S/RES/1325 of 31 October 2000. ‘UNECE Statistical Database: Gender Statistics–Crime and Violence.’ Violence against Women, Its Causes and Consequences, Ms. Radhika Coomaraswamy, on ‘Cultural Practices in the UNESCO (United Nations Education, Scientific and Cultural Family that are Violent towards Women.’ E/CN.4/2002/83 Organization). 2011. The Hidden Crisis: Armed Conflict ARMED VIOLENCE 2011 VIOLENCE ARMED and Education. Paris: UNESCO.

of of 31 January 2002. Geneva: UN. Women in South Asia: A Regional Analysis. Kathmandu: —. 2008. Promotion and Protection of All Human Rights, Civil, UNFPA. Right to Development: Report of the Special Rapporteur on UNIFEM (United Nations Development Fund for Women). 2006. Violence against Women, Its Causes and Consequences, Uncounted and Discounted: A Secondary Data Research Yakin Ertürk—Indicators on Violence against Women and Project on Violence against Women in Afghanistan. State Response. A/HRC/7/6 of 29 January 2008. Geneva: UN. Kabul: UNIFEM. GLOBAL BURDEN BURDEN GLOBAL ESA/STAT/SER.K/19. New York: Department of Economic Waiselﬁsz, Julio Jacobo. 2011. ‘Caderno Complementar 2: and Social Affairs, UN. Homicídios de Mulheres no Brazil.’ In Mapa da Violência and Instituto Sangari. Summary or Arbitrary Executions, Philip Alston. A/HRC/14/24. Watts, Charlotte, Susanna Osam, and Everjoice Win. 2001. Geneva: UN. Roberta Harmes, eds., pp. 89 –99. —. 2011. Report of the Special Rapporteur on Violence against WHO (World Health Organization). 2002. World Report on Violence Women, Its Causes and Consequences, Rashida Manjoo— and Health. Geneva: World Health Organization. Addendum: Communications to and from Governments. Zavala, Diego and Jennifer Hazen. 2009. Understanding Violence: A/HRC/17/26/Add.1 of 18 May 2011. Geneva: UN. The Role of Injury Surveillance Systems in Africa. Geneva: UNAMA (United Nations Assistance Mission in Afghanistan). 2009. Silence Is Violence: End the Abuse of Women in Afghanistan. Kabul and Geneva: UNAMA and Ofﬁce of the United Nations High Commissioner for Human Rights. UNdata. n.d. ‘Population Sex Ratio (Males per 100 Females).’ New York: United Nations Statistics Division. reinforces the notion that the rule of law and law of rule the that notion the reinforces andresolutions statements of array increasing An A prevention andreduction. prevention violence inarmed investments toguide standards out set has (OECD–DAC) Committee Assistance Development and Development’s Co-operation Economic for Organisation the Moreover, tings. set- andviolence-affected infragile priority top a as security promote Bank—increasingly World andthe (UNDP) Programme Development UN the agencies—including Nations United squandered. are indevelopment andinvestments opportunities need, of inareas security adequate without that understand They and underdevelopment. insecurity between association strong the accept and recognize aiddepartments and government agencies development Today, international most (MDGs). Goals nium Development Millen- the of achievement andthe life of quality people’s for consequences negative generates violence armed allcases, almost In opment. anddevel- growth andstable predictable ensure to required institutions andsocial economic, political, the anddistorts disrupts it aresult, As life. aproductive for necessary andassets skills the andundermines capital, social erodes zons, planning hori- shortens violence or poor, armed Rich andtrauma. suffering, pain, resulting the of aware too only are violence andphysical shootings ones? The poor who witness or survive or survive witness who poor ones? The wealthier than violence armed by affected lower-income re More Armed Violence, Less Development Less Violence, Armed More Chapter

Five 1 populations morepopulations

ute to wider human poverty andmisery. humanpoverty towider ute andcontrib- andinvestment, productivity reduce capital, humanandphysical destroy governance, undermine can andinsecurity instability tent persis- that acknowledged islikewise 2008). It Secretariat, Declaration Geneva 2009; 2008; (UNGA, states member its andamong UN the within routine now are suchviews Indeed, lence. vio- of andpersistence onset tothe contribute andlawlessness governance, weak stagnation, economic impoverishment, that reverse: of the warn also texts same these of Many well-being. andhuman progress, economic governance, good of requisites are violence from freedom ticularly susceptible to disproportionately high high todisproportionately susceptible ticularly par- are underdevelopment with plagued states how demonstrate assessments macro-level and economic terms. and economic insocial tounderperform war—tend especially violence— armed of forms certain by affected andcommunities households, individuals, that evidence is mounting there level, micro the At orviolence. development, security, by is meant what on agreement any without presumed ships are relation- these cases many In association? the of andstrength direction isthe What inevitable? linearand relationship Isthe violence? enable always underdevelopment Does in allcases? development violence undermines that armed case the isit Speciﬁcally, interrogated. is seldom andunderdevelopment insecurity between tionship rela- two-way the acknowledged, widely Although 3 Similarly, a number of of anumber Similarly, 2

145 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 146 rates of violence.4 And yet these relationships towards achieving specific MDGs. By unpacking are complex and seldom as straightforward as the correlates between lethal violence and they are often made out to be. specific forms of development achievement, the chapter intends to assist policy-makers and prac- In order to test these and related claims it is first titioners in better understanding the reasons for critical to determine whether and to what extent investing in violence prevention and reduction. lower-income countries experience more (or less) violence than middle- and upper-income countries. Among the key findings of the chapter are: A comprehensive appraisal using a long time series and a comprehensive definition of devel- Lethal violence is strongly associated with opment is critical in order to assess whether and negative development outcomes in various how armed violence restricts poorer countries ways and is accompanied by low levels of ARMED VIOLENCE 2011 VIOLENCE ARMED

of from achieving their full development potential, overall MDG achievement. including attainment of the MDGs. Moreover, such The higher the level of lethal violence recorded an assessment is crucial if policy recommenda- in a country, the larger its gap with respect to tions are to correspond to actual needs and target other countries in terms of development. problems effectively. A reduction in a country’s incidence of lethal When examined in the aggregate, it is obvious that violence corresponds with improved MDG the global burden of armed violence is weighted performance across most indicators. GLOBAL BURDEN BURDEN GLOBAL unfavourably against the poor. The large majority High rates of intentional homicide are accom- of the estimated 526,000 people directly killed panied by significantly higher levels of extreme each year as a result of armed violence reside in poverty and hunger (MDG 1), lower primary low- and medium-income settings (TRENDS AND education enrolment (MDG 2), and higher PATTERNS).5 More than two-thirds of them die as infant mortality and adolescent birth rates a result of homicidal violence—not on the battle- (MDGs 4 and 5). field or in the midst of war. A smaller proportion Countries that report proportionately lower of those dying directly and indirectly from violence levels of income inequality and unemploy- can be attributed to conflict-related incidents and ment exhibit comparatively lower levels of easily preventable illnesses in war zones. While homicide. these figures offer a global profile of the distribu- States that feature lower levels of human tion of armed violence, they also obscure more development and income almost always report complex socio-economic trends and patterns. high and very high levels of lethal violence. This chapter gathers extensive and publicly avail- Monitoring of armed violence should be inte- able statistical data to examine the relationships grated into routine MDG progress assessments 6 between development and lethal violence. It and more investments are required in national demonstrates how countries that exhibit high data gathering capacities and observatories. intentional homicide rates also commonly register low levels of development. It finds that coun- In order to clarify the basic terms of the debate, tries affected with above-average rates of lethal the chapter first reviews a number of core con- violence also tend to report slower progress cepts that are central to empirically evaluating other more subjective factors, such as fear. suchas factors, subjective more other and deprivation, material andharassment, harm andpsychological includingphysical violence, armed of manifestations many course, of are, There deaths. conflict and direct homicide intentional violence—including lethal of indicators on marily pri- draws chapter This violence. armed define and toconceptualize ways of anumber are There confusion and evidence gaps. andevidence confusion conceptual by hampered isoften development and violence armed between relationships tical statis- the of asimpleaccounting reasons, other and these For andstates. cities, households, individuals, affecting outwards, toripple tend they since complicated ismore effects precise its down Yet pinning households. of compounds walled the or in hinterlands, inpastoral andtowns, villages inperi-urban neighbourhoods, city ininner ders, bor- international along and physically—whether geographically occurs it towhere according traced be can violence armed hand, one the On tations. manifes- intheir andheterogeneous dimensional multi- are consequences andits violence Armed relationship Conceptualizing the MDG assessments. country of evaluation aroutine into integrated be could violence armed of monitoring how of treatment acursory with concludes chapter The literature. science social inthe supported are assessment statistical the of outcomes the towhich extent the considers section third the findings, scholarly recent on Drawing indicators. key of areview from findings statistical the of view over- ashort presents two Section development. and violence armed between relationships the more consequential than may be assumed; recent recent assumed; be may than consequential more often are violence armed of dimensions intangible 7

8 The The for human health and vice versa. andvice humanhealth for consequences andpsychological physiological generates violence physical inwhich ways the highlight studies andbehavioural neurological opment metrics, and one UNDP indicator. UNDP andone metrics, opment devel- Bank World three indicators, MDG specific to21 according isdisaggregated development chapter, this of purposes the For necessary. be may they though even development, of expressions insufficient capita)—are per product or national domestic gross (including growth and economic income as proxies—such narrow that concede generally specialists Development definition? choices. people’s enlarging of aprocess as development’ ‘human defines UNDP,development. example, for andclassify todefine ways many as are There categories. between discrimination requires nevertheless research blurry, are violence armed of types different between boundaries indicators of humanwell-being. of indicators andother health, population ineducation, trends measuring for and codes categories competing andpotentially confusing concern challenges Other countries. inmany infrastructure vesting har- data registration vital andweak information andmissing toincomplete andrelate chapters inprevious length at treated are deficiencies data Routine violence. armed chronic by affected andsocieties settings andmedium-income low- fragile of case inthe especially toacquire, difficult frustratingly isoften information Such countries. of cross-section alarge across data longitudinal and valid requires relationship development violence– the of accounting A comprehensive available. isconsistently years) various (across data series time enough which for indicators only the are they Moreover, income. on focus singular a than progress developmental wider gauge tively effec- more they because selected are indicators 11 What are the metrics of such a broad suchabroad of metrics the are What 13 9

Although the 10

12 These These 147 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 148 Box 5.1 Building states and societies that are resilient to conflict, crime, and violence: the 2011 World Development Report The World Bank’s mandate has evolved from a narrow focus on facilitating reconstruction in post-war Europe to a wider agenda of poverty reduction. Recently, the organization returned in some ways to its roots of reconstruction. The 2011 edition of the Bank’s flagship publication—the World Development Report—is devoted to highlighting the plight of roughly 1.5 billion people affected by conflict, crime, and violence, and those living in ‘fragile’ situations (World Bank, 2011).

The World Development Report 2011 begins by considering the dire consequences of fragility and violence on human development. It then explores the vicious cycles that result in repeated and evolving manifestations of violence in fragile states ARMED VIOLENCE 2011 VIOLENCE ARMED and the virtuous cycles of conﬁdence building and institutional transformation of necessary to effect long-term change. Finally, the report concludes with practical steps that national and international actors can take to support these initiatives and reduce the stresses that may interrupt the development of resilient states.

The report ﬁnds that people living in fragile situations are more likely to experience undernourishment and impoverishment than those living in stable and peaceful developing countries. Child mortality is twice as high in fragile and conﬂict-affected states than in states not affected by violence. Moreover, affected

GLOBAL BURDEN BURDEN GLOBAL populations are less likely to have access to basic services—including education and water. Most telling from the report is the ﬁnding that no low-income fragile country has made any signiﬁcant progress towards attaining a single MDG, even if they have made progress in the right direction. Repeated cycles of

violence over the past decades are linked to high poverty rates; in countries XX

Figure 5.1 The widening gap between countries affected by major, minor, and negligible violence

Countries with negligible or no violence Countries affected by minor violence Countries affected by major violence

Percentage of population living in poverty

20 Photo Newly arrived refugees run away from a dust cloud at 10 the Dagahaley refugee camp in Dadaab, near Kenya’s border with Somalia, July 2011. Many of them were forced to leave 1981 1984 1987 1990 1993 1996 1999 2002 2005 their homes in Somalia in order to survive. Source: World Bank (2011, p. 60) © Thomas Mukoya/Reuters 149 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 150 experiencing ‘major’ violence at any point during the period 1981 The report argues that success in building resilient states requires to 2005, poverty rates are, on average, 20 per cent higher than confidence in national reform in fragile environments where trust in countries that were minimally or not affected by violence (see is often lacking. This frequently involves sufficiently inclusive Figure 5.1). pacts committed to the reform process. These actors can build trust in the reform process through mechanisms that create cred- Consistent with the 2008 and 2011 Global Burden of Armed ible commitments, such as appointing members of the opposi- Violence reports, the World Bank finds that conflict, crime, tion, improving transparency and accountability, revising discrimi- and violence significantly reduce the growth rate of a country’s natory laws, and reforming security. To deliver early results that gross domestic product (GDP). Beyond the destruction brought build confidence, successful reform efforts often involve ‘best fit’ on by conflict and violence, political instability and conflict rather than ‘first best’ technocratic solutions and aim for a limited dissuade economic investment and can lead to unproductive number of achievable outcomes. spending on security measures. For example, in Guatemala,

ARMED VIOLENCE 2011 VIOLENCE ARMED violence is estimated to have cost the country at least 7 per Investment in renewing social norms, bolstering the rule of law,

of cent of GDP in 2005—much more than the damage wrought by and rehabilitating basic services is critical. The report emphasizes Hurricane Stan. that weak governance and rule of law and high rates of corruption are correlated with a 30–45 per cent higher risk of experiencing a The report also points to the damaging consequences of ‘spill- civil war and a significantly higher risk of extreme criminal vio- over’ effects of armed violence on neighbouring countries, lence than that of other developing states. However, institutional where annual growth can also drop by as much as 0.7 per cent. development is slow, incremental work. Institutions can be dis- A particularly vivid manifestation of these contagion effects re- mantled in days, but it takes generations to build or repair them. lates to the costs of anti-piracy measures. At least 27 countries The World Development Report 2011 estimates that the 20 fastest have spent USD 1.3–2 billion each year on interventions in the GLOBAL BURDEN BURDEN GLOBAL reforming countries in the 20th century took between 15 and 30 Gulf of Aden and in the Indian Ocean. The World Bank estimates years to raise their institutional performance from very fragile to that these costs may have risen as high as USD 4.5 billion in more resilient levels. On average, it took roughly 17 years to reduce 2010 if all regions are considered, with the economic burden of military interference in politics and 27 years to establish effective paying out ransoms and related deterred investment running at controls on corruption. USD 5.7–11.2 billion. The report concludes that development of states that are resilient Another manifestation of conflict, crime, and violence is mass to violence is necessarily a nationally owned enterprise; resilience displacement, including both refugees and the internally dis- cannot be imposed from the outside. However, the international placed. While generating strains on neighbouring host countries, community can take steps to assist countries that are trying to the exodus of human capital has both short- and long-term impli- escape fragility. The World Development Report 2011 recommends cations for domestic development. According to the most recent that international actors: calculations released by the UN High Commissioner for Refugees, more than 43.3 million people had been forcibly displaced by the 1) commit to better coordinated programmes across the devel- end of 2009, including 27 million internally displaced—the highest opment, security, political, and humanitarian spheres, with numbers of people since the mid-1990s.14 Many have fled to urban fewer priorities and more reasonable expectations; centres, where social tension, crime, and communal violence 2) reform internal policies to better identify and manage risks, may escalate. including the risk of inaction in fragile situations; The repeated and varied forms of violence that fragile states 3) reduce external stresses through regional and global action; experience are driven by a vicious cycle of 1) elite pacts that do and not reform institutions and 2) experience of new stresses and 4) promote South–South learning and experience sharing, external shocks that can plunge a society into conflict. To break collaboratively with emerging powers and regional this cycle of fragility and violence, a society needs to build confi- institutions. dence in a national reform project and undertake institutional transformation. Source: Gary Milante others. than points data more feature indicators opment devel- some that fact isthe stressing worth Also territories. and 170 countries only for isavailable ables vari- development andrelevant violence lethal on data however, indicators; development or violence homicidal on available be may data reliable which for andterritories 239 countries of atotal are there Taken together, indicators. anddevelopment violence lethal both for tion isinforma- there unless undertaken be cannot analysis a correlation that stressing isworth It analysis. coefficient and correlation variables between inter-relationships the of tion includeaninspec- used Methods associations. of andstrength direction the on focus specific for statistically significant relationships, with a andsearches mid-1980s sincethe rates homicide pools analysis this Methodologically, above. cited indicators development andthe population) 100,000 (per rates homicide intentional between correlates the examines It underdevelopment. and violence link lethal that pathways causal cific andspe- relationships on focuses chapter This relationship the Assessing sets. data- multiple of iscomposed which database, 2011 GBAV includethe These and development. violence inlethal trends documenting datasets of inahost assembled figures reliable on draws ter chap- this limitations, these acknowledging While armed violence and development (see Box 5.1). Box (see anddevelopment violence armed Report 2011 Bank’s World the on and progress, anddevelopment achievements MDG tracking databases Bank-managed World 15 The chapter also draws on UNDP- and UNDP- on draws also chapter The in order to clarify the linkages between between linkages the toclarify inorder 17 The lack of complete temporal and temporal complete of lack The World Development Development World 16 both

states and territories. andterritories. states under-represented from information better sharing and collection data of improvement agradual by corrected be eventually could biases resulting 18.57), and very high (>18.57) rates. (>18.57) high 18.57), andvery (7.25– high population), (<7.25low 100,000 per intervals: tothree according categorized homicide of thresholds displays also figure The countries. HDI high very and38 countries, HDI 45 high tries, coun- HDI medium 75 countries, HDI low 24 were categories. development (>0.78) high andvery (0.66–0.78), high (0.48–0.65), medium (<0.47), tolow according classified 182countries of anoverview provides 5.2 Figure violence. tolethal inrelation scores (HDI) Index Development Human totheir according countries of distribution the considering isworth it follows, that analysis statistical the tosituate order In lethal violence and income, Human development, characterized by a dearth of information. of adearth by characterized are andOceania Asia, Africa, whereas Europe, and Americas the from isavailable data reliable ample development, of assessment statistical any virtually for case isthe As and inresults. sample inthe biases generate might variables development core for coverage data geographic homicidal violence. violence. homicidal of levels lower proportionately feature also ment humandevelop- of levels high andvery high both registering countries of majority vast the that shows figure the contrast, In average. long-term the above rates homicide feature human development medium exhibiting allcountries of half almost and countries humandevelopment low of thirds two- almost that reveals 5.2 Figure violence. more exhibit humandevelopment of levels lower register that countries terms, proportional In 19 In 2009, there there 2009, In 20

18 The The 151 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 152 Figure 5.2 Disaggregated homicidal violence for 182 countries by HDI, or ‘deteriorating’ across homicide and development 1986–200921 indicators reveals how countries that register a Low homicide rate High homicide rate Very high homicide rate crude improvement in their HDI are also most likely to exhibit lower levels of lethal violence. Percentage of total number of homicides That is, homicide rates determine negatively and 100 3 11 4 37 significantly the presence of any improvement of 90 15 12 21 a country’s HDI rating, constraining development 80 progress. As such, the statistical assessment con- 70 firms that higher homicide rates are associated 60 23 43 26 with lower HDI rankings. 50 Taken together, the findings confirm that less ARMED VIOLENCE 2011 VIOLENCE ARMED 40

of 9 developed countries experience more lethal vio- 30 lence than medium- and upper-income countries. 20 Moreover, they demonstrate that lethal violence 10 hampers development. They also reveal a strong and negative association between the levels of Low human Medium human High human Very high human development development development development lethal violence in a country and the degree or (24 countries) (75 countries) (45 countries) (38 countries) extent of its development. The higher the level of

GLOBAL BURDEN BURDEN GLOBAL Human Development Index category lethal violence recorded in a country, the larger Note Numbers in the bars indicate how many countries fall into each of the three homicide rate : its gap with respect to other countries in terms categories (low, high, or very high). The figure for the top segment in the right-hand bar is ‘1’. of development.24 Source: GBAV 2011 database and selected development and violence indicators It is possible to geographically chart the rela- As can be seen in Figure 5.2, countries that exhibit tionships between lethal violence and human high and very high homicide rates are concentrated development according to thresholds of homi- in the low human development band. Yet a signifi- cidal and conflict-related violence. As signalled cant number and proportion of countries in the in Chapter Two, countries featuring high homicide medium and high human development category rates appear to be located predominantly in regions report severe homicide levels. While lethal violence of Latin America and the Caribbean, Central and is concentrated primarily among underdeveloped Southern Africa, and particular areas of Central countries, it is thus not the exclusive preserve of and Southern Asia (TRENDS AND PATTERNS). the poor. Indeed, taken together, approximately one- Countries featuring armed conflicts are highly fifth (19 per cent) of the world’s population resides concentrated in lower-income settings throughout in lower- and medium-income countries experienc- Africa and Asia, although some other regions are ing high and very high levels of lethal violence.22 also represented (see Map 5.1).25

Just as alarming are the ways in which low levels A regional analysis reveals a statistically signifi- of development are in most cases correlated with cant inverse correlation between higher human higher violence over time. In other words, lethal development and lower homicide rates in North violence appears to constrain development prog- America, most of Asia, and Western, Eastern, and ress. Classifying countries as either ‘improving’ South-eastern Europe. Even across these regions, human development are affected by conflicts of of conflicts by affected are development human high relatively feature that countries three only ismore, What conflict. armed ongoing an by isaffected country the isthat less it likely the humandevelopment, of level the higher the that appears thus It conflict. anarmed encing experi- orare experienced have development human low report that countries the half Nearly violence. homicidal of dence inci- ahigher with associated or coincidentally arbitrarily isnot development lower that clusion con- the reinforcing thus human development, of levels lower comparatively with correlate also rates homicide high report that countries Source: 5.1 Map L HDI classification(2009) EGEND No data Low Medium High Very high armed conflict Country with population) (> 7.25per100,000 homicide rate Country withhigh GBAV 2011 development and and violence database selected indicators’ : Geography of armed violence and human development, 2009 andhumandevelopment, violence armed of Geography indicators income by replaced are indicators development human holdswhen lence and underdevelopment vio- lethal between relationship negative The and Turkey. Federation, Russian the kinds—Colombia, various high or very high homicide rates. homicide high orvery high exhibit countries non-OECD afew andonly lence, vio- homicidal of rates low allreport OECD the of states member wealthy 5.3). Moreover, Figure (see violence homicidal of level reported the higher the country, agiven by registered income the analysis. in this included are which all of countries, 207 for income on . 26 The World Bank reports information information reports Bank World The 27 Specifically, the lower lower the Specifically, 28 153 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 154 Figure 5.3 Disaggregated homicidal violence by income, 1986–2009 and reductions in hunger. When lethal violence Low homicide rate High homicide rate Very high homicide rate is correlated across these variables, it appears that countries registering higher poverty rates also Percentage of total number of homicides record proportionately higher levels of homicide. 100 377327It is important to stress that these findings do 16 90 4 15 not necessarily account for a causal relationship: 15 80 30 all countries that are poor are not necessarily 70 more predisposed to high levels of lethal violence.31

60 32 Even so, the statistical analysis detects a direct 24 50 24 relationship between poverty levels (measured

40 in terms of the percentage of the population ARMED VIOLENCE 2011 VIOLENCE ARMED 30 living under USD 1 (PPP) per day) and homicidal of violence. 20

10 Speciﬁcally, proportionately higher poverty levels tend to go hand in hand with higher levels of Low income Lower middle Upper middle High income High income (43 countries) income income (37 non-OECD (27 OECD lethal violence. A similar negative relationship (54 countries) (46 countries) countries) countries) holds for the poverty gap ratio, which is a simple Income category measure of inequality. This suggests that lethal

GLOBAL BURDEN BURDEN GLOBAL Note: Numbers in the bars indicate how many countries fall into each of the three homicide rate violence is not only correlated directly with pov- categories (low, high, or very high). erty measured as income, but also with poverty Source: GBAV 2011 database and selected development and violence indicators measured as inequality. Indeed, the higher the concentration of income among the rich, the The MDGs and lethal violence higher the total levels of homicidal violence.32 It also appears that higher levels of lethal violence The fact that countries registering greater are statistically correlated with lower levels of MDG income inequality also regularly exhibit a higher incidence attainment.29 This statistical analysis considers of lethal violence is illustrated in Figure 5.4. seven of the eight MDGs and 21 indicators, included Specifically, the figure depicts the relationship according to their availability.30 For MDG 1 (eradi- between homicide rates (squares) and direct cate extreme poverty and hunger) ten variables conflict deaths (triangles) and the proportion of were considered. In the case of MDG 2 (achieve income earned by the lowest 20 per cent of the universal primary education), MDG 3 (promote population in all countries for which data is gender equality and empower women), MDG 4 available.33 The figure shows that countries with (reduce child mortality), and MDG 5 (improve higher levels of lethal violence (approaching the maternal health), just two indicators (each) were right end of the horizontal axis) coincide with used. Finally, for MDG 6 (combat HIV/AIDS, malaria, higher inequality, measured as a low proportion and other diseases) and MDG 7 (ensure environ- of income earned by the poorest 20 per cent of mental sustainability), one and three variables the population (low on the vertical axis). The were included, respectively. inverse also applies: countries experiencing low MDG 1 can be parsed into at least four specific levels of lethal violence feature a higher propor- targets focused on poverty, income, employment, tion of participation of this same income group. Source Note Poorest quintile’s America), it applies generally to all regions. toall regions. generally applies it America), Latin (suchas countries of clusters certain for strong isespecially correlation negative this while And 5.4. Figure of right bottom tothe left top the from concentrated are that points of ‘cloud’ the by discerned be can correlation the of strength the Indeed, violence. andlethal inequality income of relationship apositive into translates variables two the between relationship anegative that Table (see 5.1). Note strong chance)by andvery (not significant as well as negative tobe found was relationship The tests. correlation tical statis- using scrutinized was violence and lethal inequality income between relationship This 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Figure Asia 20304050607080901000 Squares correspond to homicide rates and triangles represent direct conflict deaths. conflict direct represent triangles and rates homicide to correspond : Squares GBAV 2011 database and selected development and violence indicators violence and development selected and database 2011 : GBAV Europe 5.4

Disaggregating income inequality and lethal violence, 1986–2009 violence, andlethal inequality income Disaggregating 10 share in the natio Africa Oceania nal income or Americas Homicide anddirect conflict death consumptio n, percentage it is not statistically robust. statistically isnot it five), under children underweight of prevalence and violence lethal between is an association there while Further, bias. agender of absence the indicating strong, remains rates andhomicide women young unemployed between correlation the sex, by aggregated isdis- indicator this When rate. homicide lower a display also will country the that probability the stronger the rate, unemployment male youth acountry’s lower the particular, In violence. lethal of levels lower toreport tend also age) of years (15–24 andwomen men young of unemployment register that countries meantime, the In rate (per 100,000 population) (as measured by the the by (asmeasured hunger 34 low 155 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 156 With respect to MDG 2—the achievement of uni- MDG 5 considers the progress in relation to versal primary education—the statistical analysis maternal health, which is often measured as a considers the net enrolment ratio in primary function of declines in maternal mortality, a lower school and literacy rates of 15–24-year-olds.35 adolescent birth rate, and the quality of prenatal A close inspection finds that higher homicide and natal care. Key indicators examined as part levels tend to occur in countries that register low of this assessment include the adolescent birth primary education enrolment ratios (see Table 5.1). rate and the proportion of births attended by A potential causal link for this relationship can skilled health personnel. Figure 5.5 reveals a be hypothesized. For example, the inability of a very strong positive correlation between coun- society to keep its youth in the education system tries exhibiting a high adolescent birth rate and during a particularly risk-prone age can make high rates of lethal violence, a finding also ARMED VIOLENCE 2011 VIOLENCE ARMED presented in Table 5.1. Specifically, each point of them more predisposed to violence. Specifically, they may be more susceptible to recruitment shown in Figure 5.5 represents adolescent birth rates per thousand women (on the vertical axis) into armed groups, such as gangs or guerrilla and lethal violence (on the horizontal axis) in factions. In turn, this trajectory would deny them every country for which data is available. the productive capacities required to enter the labour market, thus further contributing to a Figure 5.5 shows that countries exhibiting lower downward spiral. rates of lethal violence (on the horizontal axis) GLOBAL BURDEN BURDEN GLOBAL also register lower levels of adolescent pregnancy The assessment also considers MDG 3—the pro- (on the vertical axis) and tend to cluster towards motion of gender equality and empowerment of the bottom left of the graph. Countries featuring women—and its statistical relationship with higher levels of lethal violence exhibit higher lethal violence. A modest positive association levels of adolescent pregnancy and tend to be exists between the share of women in wage found away from the origin, or intersection of the employment within the nonagricultural sector x and y axes. A potential causal explanation for and the ratio of girls to boys in primary and this is that lethal violence interrupts access to secondary education, in particular, and lethal health care and that large households tend to be violence (see Table 5.1). less capable of investing in education and pre- With respect to MDG 4—the reduction in child paring children to anticipate risks of violence mortality—the mortality rate of children under later in life. The statistical analysis also detects five is also significantly and positively associated a strong positive correlation between infant mor- with homicide rates (see Table 5.1). However tality rates and homicidal violence as well as a rather than indicating a direct causal link, it is strong negative association between the share possible that this finding is simply underlining of births attended by skilled personnel and the fact that less developed countries (most of homicidal violence. which feature inadequate water, hygiene, and With respect to MDG 6—combating HIV/AIDS— health systems) may witness an increase in a significant relationship exists between lethal so-called excess deaths among the most vulner- violence and HIV/AIDS (see Table 5.1).36 Specifically, able—as well as rising rates of lethal violence. countries featuring a high percentage of people the World Bank (2011). Bank World the by afindingechoed rates, homicide and higher inslums living people of proportion the between association isanegative there Likewise, facilities. andsanitation water todrinking access improved experience violence lethal of levels lower with countries that suggests It violence. and lethal 7—ensuring MDG between arelationship uncovers also analysis The andAsia. Americas, the inAfrica, significant isespecially correlation positive This rates. homicide higher experience HIV/AIDS 15–49) (ages to tend with also living Source Note Adolescent birth rate (per 1,000 women) 25 50 75 100 125 150 175 200 225 250 Figure Asia 20304050607080901000 Squares correspond to homicide rates and triangles represent direct conflict deaths. conflict direct represent triangles and rates homicide to correspond : Squares GBAV 2011 database and selected development and violence indicators violence and development selected and database 2011 : GBAV Europe 5.5

Disaggregating adolescent birth rates and lethal violence, 1986–2009 violence, andlethal rates birth adolescent Disaggregating 10 Africa environmental sustainability environmental Oceania Americas Homicide anddirect conflict death — above. cited tothose results similar reveals progress andMDG rates death conflict direct between relationships the of analysis aseparate Finally, deaths are statistically correlated with: with: correlated statistically are deaths lower HDI. andthe education inprimary enrolment lower employment inwage women of share alower pop- the as (measured poverty of rates higher rate (per 100,000 population) ratio of girls to boys in primary education; and and education; inprimary toboys girls of ratio sector; non-agricultural in the ratio); gap 1andpoverty USD below ulation 37 Indeed, higher reported direct conflict conflict direct reported higher Indeed, 38 157 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 158 Reviewing the relationship ox 5.2 The relationships between lethal violence B The above analysis draws attention to a number and development of unsettling negative relationships between Table 5.1 reviews correlations of key development indicators and lethal vio- lethal violence and underdevelopment. Some of lence. The signs located in the right-hand column—positive and negative— indicate the direction and intensity of each association. The larger the sign, these associations are more robust than others. the closer the association between the two variables. The shades reveal In some cases, the availability of additional data the statistical significance of the association, from the 1 percent (dark, most would allow for a more exhaustive assessment significant) to the 5 percent significance level (light, less significant). of patterns and trends. Nevertheless, the basic Table Table1 5.5.1 The relationship between development indicators and claim that high rates of lethal violence have lethal violence negative implications across an array of MDG indicators is confirmed, offering a more nuanced ARMED VIOLENCE 2011 VIOLENCE ARMED Development indicator Relationship to of assessment than has been provided elsewhere, armed violence including in the 2011 World Development Report. Adolescent birth rate (per 1,000 women) Moreover, this chapter finds that these relation- Births attended by skilled health personnel (percentage) ships are not always straightforward: they are Children under five severely underweight (percentage) frequently complex, heterogeneous, and of vary-

Children under ﬁve mortality rate (per 1,000 live births) ing intensity. Employment-to-population ratio, both sexes (percentage) A number of established development agencies GLOBAL BURDEN BURDEN GLOBAL Infant mortality rate (0–1 year, per 1,000 live births) are convinced that violence has deleterious effects

People living with HIV, 15–49 years old (percentage) on MDG progress. For example, the OECD–DAC’s

Poorest quintile’s share in national income or consumption International Network on Conﬂict and Fragility (percentage) --- acknowledges how ‘external and internal risk

Population below USD 1 per day (PPP, percentage) factors, alongside the continuum of conflict, armed violence and insecurity makes the MDGs Poverty gap ratio at USD 1 per day (PPP, percentage) more difficult to achieve’ (OECD, 2010a, p. 4). In Proportion of the population using improved drinking water addition, the Dili Declaration on Peacebuilding sources (total) --- and Statebuilding underlines how ‘conflict and Proportion of the population using improved sanitation fragility are major obstacles for achieving the facilities (total) --- MDGs’ and recognizes that ‘it will be extremely Share of women in wage employment in the non- difficult to achieve the MDGs in most fragile and agricultural sector conflict affected states by 2015’ (IDPS, 2010, p. 1).39 Slum population as percentage of urban population (percentage) Multilateral and bilateral aid agencies are pooling Total net enrolment ratio in primary education (both sexes) their investments where violence appears to be

Youth unemployment rate, aged 15–24 (both sexes) concentrated, including so-called fragile states

Youth unemployment rate, aged 15–24 (men) (OECD, 2010b). The OECD estimates having chan- nelled in excess of USD 34 billion in aid dollars Youth unemployment rate, aged 15–24 (women) towards these regions in 2009—more than one- Human Development Index third of total global spending on development.40 Home to more than 1.5 billion people, these fragile is acause former the whether around torevolve tend lence vio- andcriminal conflict armed of and severity onset andthe inequality income over Debates 1996, p.43). 1996, (Machel, infrastructure physical of destruction the andthrough andstudents personnel, trative adminis- teachers, of anddisplacement trauma, killing, the through largely network, school mary pri- the of cent 45 per anestimated demolished 1980s the during war civil The 2001). FitzGerald, and (Stewart services educational compromise can violence conflict-related how demonstrate to presented isroutinely of Mozambique case The extensive. as just arguably are prospects development for violence conflict-related of tions implica- longer-term the considerable, are deaths conflict direct of costs short-term the While ing countries. develop- infour toone compared and hunger—as poverty halving of goal the toachieve is expected states fragile inten one just p.29).2010, Indeed, andKlasen, (Harttgen progress MDG of terms in countries developing stable more behind lagging generally are states and conflict-affected high levels of inequality (Waiselfisz, 2010, p.148). 2010, (Waiselfisz, inequality of levels high with correlated similarly are rates homicide high that evidence provides inBrazil violence homicidal of inequality and homicide rates. andhomicide inequality income between associations positive examined has research andcross-national comparative 5.3). example, Box For (see correlations causal robust identify frequently also crime and violent inequality income between linkages the gating investi- studies Empirical 2005). and Murshed, Gates 2001; Stewart, 2000; (Collier andHoeffler, significant is less relationship the that claim others while factor causal isastrong inequality income that contend specialists conflict Some , an 42 outcome 41

, or both (Gates et al., 2010). 2010). al., et (Gates , orboth 43 One recent review review recent One © Caetano Barreira/Reuters © Caetano 2007. March Paolo, São in hotel Hilton the of front in Photo X

A view of the slum slum the of A view 159 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 160 Box 5.3 Lethal violence and MDG attainment in São Paolo Map 5.2 Homicide rates per 100,000 Although it has registered rapid economic growth and has consolidated its population by district, São Paulo, 2009 democracy, Brazil is confronting one of the highest rates of homicidal violence in the world. While there appear to have been some important reductions in recent years, homicide is especially prevalent in the country’s major cities—Rio de Janeiro, São Paolo, and Brasília (see Figure 5.6). In response, and in view of the much-anticipated 2014 World Cup and 2016 Olympics, political and economic investment in integrated violence prevention and reduction activities is increasing.44

Made up of some 96 districts, São Paolo is one of the world’s largest cities, ARMED VIOLENCE 2011 VIOLENCE ARMED with an estimated 20 million inhabitants. It is also known for its extreme rates of of violence and inequality. Levels of homicidal violence in São Paolo are also higher than the global average: 34 districts feature high homicide rates of 10 per 100,000 and the remaining 62 register high homicide rates of approximately 19 per 100,000 (see Map 5.2).45

A number of clear trends emerge from an examination of the relationship between homicidal violence and MDG achievement in São Paolo for the year 2006 (see LEGEND: GLOBAL BURDEN BURDEN GLOBAL Table 5.2). For example, in districts that exhibited higher homicide rates, a Per 100,00 higher percentage of the population lived on less than half the minimum wage. population Moreover, residents of poorer slum areas or favelas were three times more likely > 10 to be living in districts experiencing high homicide rates. Likewise, districts < 10 reporting higher homicide rates also experienced slightly higher unemployment rates among young men and worse sanitary conditions.46 Source: Muggah and Wennmann (2010, p. 26)

Figure 5.6 Homicide rate per 100,000 population in Brasília, Rio de Janeiro, São Paolo, and Brazil, 1992–200747 Brasília São Paulo Rio de Janeiro Brazil

Homicide rate per 100,000 population

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007

Source: SIM (n.d.) Author Source 2006 Paolo, São Table year-olds 15–24- literate ofPercentage of 15–24-year-olds Unemployment rate force labour productive the economically 15-year-olds in ofPercentage cent 20 per poorest the income held by of Proportion capita per minimum wage the of cent per a Population with Sex indicators MDG access to sanitation population without ofPercentage the supply awater to access population without ofPercentage the slums inhouseholds ofPercentage n income of 50 50 of n income 5.2 adapted from Muggah and Wennmann (2010, p. 27) p. (2010, Wennmann and Muggah from : adapted Renato Sérgio de Lima : Renato Sérgio

MDG indicators and violence in two district clusters in clusters district intwo andviolence indicators MDG oe 223. 31.0 35.7 32.2 Women e 8910098.5 100.0 98.9 Men e 281. 24.0 19.4 81.5 85.5 22.8 81.4 89.9 81.5 86.8 Men Women Men oe 9010098.7 100.0 99.0 Women São Paulo Districts Districts São Paulo 051. 24.3 11.4 20.5 12.4 14.9 0.8 . . 4.0 2.1 3.3 homicide of rates with lower 6.3 0.3 5.5 homicide of rates with higher Districts 18.5 15.3 1.0 Bourguignon, 1999). 2007; and UNODC, Bank (World Caribbean andthe America inLatin especially ment, tounemploy- inrelation effects knock-on creating growth, GDP hamper in turn cide can Bank have homi- of stressed rates high how World andthe andCrime Drugs on Office UN violence, intensify. relationships these of prevalence ahigh by affected countries 2010). In Jaffe, 2010; Survey, Small Arms 2001; andMoser, (McIlwaine groups armed andother gangs into recruitment youth of risk the enhance reportedly factors other and unemployment countries, other In 5.4). Box (see shows Africa South of case the as andidleness, frustration growing totheir ute tocontrib- perceived men—is young among larly unemployment—particu- rising hand, one the On violence. armed of andintensity incidence andthe unemployment between association the on research considerable isalso There (Neumayer, 2005). countries of toasmallnumber restricted isartificially sample and2) the for trolled con- not are effects if 1)country-specific determinant significant astatistically is not inequality that argues researcher One crime. of violent determinant significant statistically a isnot inequality that contending ings, find- latter these contest Yet scholars some pp.10–11). andÖzler, 2002, (Demombynes other the on androbbery hand andhomicide one the on rates growth economic andlow ity inequal- ineconomic increases a linkbetween observes assessment one 40states, almost for data panel on p.509). Drawing 2002, andMenéndez, Loayza, (Lederman, countries most across crime inviolent to increases tocontribute appear also rates growth lower and inequality income time, same the At 49

48 The The 161 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 162 Box 5.4 South Africa: violence and development the country as symbols of social power and dominance (CSVR, Despite the momentous decline in political violence after South 2008). Violent crime is typically characterized by encounters Africa’s transition to democracy in 1994, the country continues to between people who know each other as friends, relatives, 50 experience one of the highest murder rates on the planet. Indeed, or acquaintances. a series of factors place South Africa at statistical risk of high levels A comparison of the relationships between lethal violence and of homicidal violence. These include a low HDI score (ranked 129 underdevelopment across South African provinces reveals out of 189 countries in 2010), persistent income inequality, high various links. As shown in Table 5.3, the metropolitan areas of levels of youth unemployment, high rates of HIV/AIDS (one of the Gauteng and Western Cape are plagued by high levels of inequal- highest in the world at more than 18 per cent), and high adoles- ity and high levels of violence. When compared, both provinces cent birth rates. exhibit a relatively low level of absolute poverty—measured in During 2008–09 the South African Police Service reported a terms of people who live below ZAR 283 (about USD 37) per ARMED VIOLENCE 2011 VIOLENCE ARMED

of homicide rate of 37.3 per 100,000 (SAPS, 2009, p. 5). With young month but experienced a significant increase in inequality over men making up the bulk of offenders and victims of criminal the period 1995–2005. Such a trend is also apparent in Limpopo violence, firearm homicide has been singled out as the leading province, which features one of the highest rates of poverty in the cause of death for young men aged 15–21 (Fleshman, 2001). country but one of the lowest rates of murder, while inequality More than half of all homicides in the country are committed levels declined between 1995 and 2005 (Muggah and Wennmann, with firearms—instruments described routinely by scholars in 2010, p. 33).

Table 5.3 Violence and development indicators in South Africa, various years GLOBAL BURDEN BURDEN GLOBAL

Province Intentional People living Gini coefﬁcient53 Unemployment homicide rate below ZAR 283 199554 200555 2008, in %56 2009, in % per 100,000 (USD 37) per population month (2008, (2008)51 in %, rounded)52

Eastern Cape 49.5 29.0 0.65 0.64 25.2 27.0

Free State 31.6 16.0 0.66 0.65 22.6 25.3

Gauteng 37.2 6.0 0.54 0.65 20.7 25.7

Kwazulu-Natal 47.0 33.0 0.63 0.67 20.8 19.2

Limpopo 14.2 34.0 0.63 0.58 28.9 26.9

Mpumalanga 25.1 28.0 0.58 0.67 23.1 26.6

Northern Cape 36.5 27.0 0.65 0.62 21.6 24.9

North West 27.4 23.0 0.63 0.64 25.7 27.0

Western Cape 44.6 9.0 0.58 0.69 16.9 21.5

South Africa 37.3 22.0 0.64 0.69 21.9 24.3 (national)

Source: Muggah and Wenmann (2010, p. 33) It is also potentially controversial in some coun- insome controversial potentially isalso It agenda. challenging, yet isanessential, outcomes and progress development on violence armed of effects the of accounting andaccurate ongoing An relationship the Monitoring actions’ (Putnam, 1993). coordinated facilitating by society of ciency effi- the improve can that andnetworks norms trust, suchas organization, social of ‘features toencompass isheld it but defined, variously is capital Social capital’. ‘social of stock ety’s asoci- by mediated isoften underdevelopment and violence armed between relationship the of extent the scientists, social tosome According for building meaningful human development. humandevelopment. buildingmeaningful for essential relations of formation the thwarting capital, social of stock the intransforming role detrimental andoften acritical play can violence Armed ways. destructive andoften in harmful andstrengthened reconfigured be nomic elites—can andeco- topolitical andgroups individuals ties capital—which social ‘linking’ Moreover, rapidly. erode together—can orindividual groups parate dis- otherwise links that kind capital—the social ‘bridging’ But groups. homogenous relatively within ties tostrong referring capital, social ing’ ‘bond- of level high acomparatively register may communities violence-plagued hand, one the On and Moser, 2001). (Mcllwaine action andcollective reciprocity, nity commu- networks, social on violence armed of effects negative the of evidence anthropological isconsiderable organizations—there mafia-like and cartels, membership, gang of form in the perverse be even can is‘positive’—some capital allsocial not While action. collective enable that relationships social of toaspects refers tal 57 Ultimately, social capi- social Ultimately, maternal health data on a routine basis. aroutine on data health maternal and education, income, poverty, basic analyse and tocollect andsystems tools monitoring andmunicipal national, regional, international, in investment renewed of importance and the inknowledge gaps persistent the highlighted have group monitoring MDG the of meetings repeated example, For MDGs. the in monitoring challenges encountered has itself sector ment develop- The politicization. tointense and subject sensitive acutely remains topic the where tries, (UNGA, 2009, p. 19). The Secretary-General also also Secretary-General p.19). 2009, The (UNGA, 2015 until violence armed andmeasure monitor to andindicators targets, goals, the standardize Violence Armed of tion andPreven- Reduction the through Development UN Secretary-General’s on report the new. Specifically, not are achievement MDG and violence between relationship the and track monitor tobetter amechanism for calls Public 5.5). Box up(see andscaled replicated be can that systems of examples important offer they mechanisms, surveillance injury or systems, warning early conflict observatories, crime as described are they whether of Regardless violence. inarmed trends nisms toappraise mecha- andmonitoring reporting and integrated comprehensive of examples impressive are there Fortunately, vital. isnevertheless it consuming, time- andbe investment considerable require may suchsupport of provision the While violence. criminal of rates andhigh conflict armed from andrecovering by affected incountries systems collection anddata surveillance and municipal national of strengthening isthe priority A first countries. andmiddle-income low- among challenges key highlight constructively could suchevaluations as assessments, MDG national and international periodic into violence armed integrating madefor tobe case astrong is also emphasizes the need to need the emphasizes Promoting 58 There There 163 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 164 Box 5.5 Global trends in monitoring armed violence prevention orientation and most are intersectoral or interdisciplinary in approach. All of them tend to Public health and development agencies are devoting increasing attention to feature at least three basic characteristics: 1) they evidence-based approaches to armed violence prevention and reduction. The routinely gather timely data on key variables; 2) they World Health Organization has released several seminal studies since 2002 systematically analyse data over time; and 3) they through its Violence and Injury Prevention programme (WHO, n.d.b).59 Moreover, disseminate analysis with a view to informing policy the WHO-supported Violence Prevention Alliance, a network of more than 50 and programming (Gilgen and Tracy, 2011, p. 12; see public health and community development agencies, has also shed light on Figure 5.7). opportunities and challenges associated with interventions to promote safety.60 More recently, in 2009 and 2011, the OECD–DAC drew attention to the Most AVMS coordinate and harmonize information wide range of related direct and indirect programmes and projects under way from disparate public, private and non-governmental around the world.61 entities (Carrière, 2008; CISALVA, 2008). In best- case scenarios they can facilitate coordination and There are many different kinds of armed violence monitoring systems (AVMS)

ARMED VIOLENCE 2011 VIOLENCE ARMED communication between elected ofﬁcials, police, operating across the globe. In many countries, both public entities and non- of health, educational and social services, researchers governmental organizations have established surveillance and survey-based and activists, and civil society organizations. A recent monitoring systems, often in partnership.62 Many have a public health or crime study shows that most AVMS-type efforts collect data on common indicators of armed violence—including Figure 5.7 The core attributes of an armed violence monitoring system lethal violence (Gilgen and Tracey, 2011). While mortality is the most widely monitored indicator, Armed violence AVMS systems generally collect and analyse data on monitoring system (AVMS) non-fatal injuries (including sexual violence) as well as

GLOBAL BURDEN BURDEN GLOBAL incidents arising from common assault and road trafﬁc Ongoing data collection Systematic data analysis Informing programming and policy accidents (see Figure 5.8). It is important to note that most AVMS collect data from at least one government Intersectoral system source, such as criminal justice and vital registration records from hospitals and mortuaries (see Figure 5.9). Thus, the effectiveness of accurate data collection depends in large part on a government’s willingness and ability to collect and disclose accurate ﬁgures. Figure 5.8 Violence indicators tracked by AVMS initiatives63 Mortality Injuries Sexual violence Crime (drugs, robberies, theft) Figure Road accidents Suicide Other 5.9 Data sources used by AVMS, as percentage Percentage Legend: 100 Vital registration data Police and forensic sources 90 Health and hospital data Morgues Media reports Community groups Other 80

40 30 20 10

Injury surveillance Violence Crime Other systems observatories observatories Source: Gilgen and Tracey (2011) are taking part in an effort to relocate them, April 2011. © Cheryl Ravelo/Reuters ©Cheryl 2011. April them, relocate to effort an in part taking are Photo S Following a ﬁre, informal settlers of the Laperal compound in Manila throw stones and bottles as they clash with police and mem and police with clash they as bottles and stones throw Manila in compound Laperal the of settlers informal aﬁre, Following bers of a demolition team who 165 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 166 Figure 5.10 An inter-agency armed violence monitoring group Data from public and private partners Data from national and municipal governments Data from international organizations Data from research institutes and universities

Direct conﬂict deaths database Complex Emergency Internal Database International Displacement OECD–DAC Crime Victim Monitoring Surveys Centre (IDMC) ARMED VIOLENCE 2011 VIOLENCE ARMED

of Geneva Declaration Interpol and Secretariat Europol Global Burden of UNODC UNHCR Armed Violence Crime Trends and the Online Population Operations of Criminal Database Justice Systems

National criminal justice data (police reports, crime GLOBAL BURDEN BURDEN GLOBAL statistics, court UNICEF records, etc.) Multiple Indicator Cluster Survey Regional, Armed violence national, local crime observatories monitoring group Gallup World Poll

Armed violence/safety World Bank surveys Living Standards Measurement Study

Health WHO surveillance data (hospital records, death UNECE Global Burden Indicators on Violence of Disease certiﬁcates, emergency against Women room records, etc.)

Demographic and health surveys

Source: adapted from Gilgen, Krause, and Muggah (2010, p. 24) © Dieter Telemans/Panos Pictures Telemans/Panos © Dieter education. an receive can children disabled where Burundi in schools few the of one Bujumbura, in Institute Photo ground. the on needs practical by overtaken being fast is model This Nations. United tothe reporting andperiodically independently ices) working serv- andsocial police, health, (suchas agencies public on in autonomous violencerelies trends andmeasuring monitoring of model current The ground. the on actors society andcivil governments assist effectively more can it that so system amonitoring tostructure way appropriate most the on however, issilent, He 19). (p. MDGs the toachieve efforts ongoing into themes security-related tointegrate opportunity an offers process review MDG the how stresses decision-makers, programme managers, and managers, programme decision-makers, for service aninvaluable as serve and do can systems these Taken together, injury. external of forms or other violence inintentional trends to track inorder nodes sub-national of anetwork through operate others while concerns safety national overall address actors these Often interventions. policy public andinform invictimization trends document tobetter inaneffort levels municipal and national, regional, the at established been already have systems monitoring violence armed Many world. the around systems data-gathering credible of number awide on tobuild need would mechanism monitoring international A functional 2010). andMuggah, Krause, (Gilgen, issues other and welfare, social health, education, secondary and inprimary trends on cases, inmany but, zation, andvictimi- violence of issues on exclusively not focused activities collection data require will It approach. andintersectoral inter-agency an ongoing for call will attainment andMDG violence armed between relationships speciﬁc the totrack effort

X A student writes on the blackboard at the St. Kizito Kizito St. the at blackboard the on writes Astudent

As demonstrated above, a more effective effective amore above, demonstrated As 167 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 168 practitioners who work in the field of crime and violence prevention and reduction programmes. violence prevention. Building on these and other ongoing data collection frameworks, the working group would issue One way to strengthen global armed violence routine outputs and highlight critical information monitoring efforts and improve standards in gaps and needs on the way armed violence is data collection would be through the establish- affecting MDG progress (Gilgen, Krause, and ment of a global inter-agency armed violence Muggah, 2010, p. 25). monitoring group or task force. States already play a central role in global efforts to collect information on core indicators such as homicide through their departments of statistics; they Conclusion would need to play a primary role in any monitor- ARMED VIOLENCE 2011 VIOLENCE ARMED ing group. But to generate more comprehensive The costs and consequences of armed violence on of and refined information and analysis, the group development are considerable and wide-ranging. would also need to forge partnerships with agen- They include direct visible impacts such as death, cies from the health, criminal justice, conflict injury, and damage to assets and property as studies, and development sectors. An example well as indirect effects such as the increased, of an ideal type multi-nodal global monitoring recurrent costs of law enforcement and justice group is set out in Figure 5.10. delivery, disruption of social services, lost economic opportunities, and the undermining of GLOBAL BURDEN BURDEN GLOBAL Such a monitoring group would be responsible governance more generally (Skaperdas et al., for tracking the incidence, causes, and trends in 2009, p. 17; Geneva Declaration Secretariat, 2008, victimization and development. It would also pp. 89–108). usefully establish and disseminate standards and support networks in collecting data. Likewise, The analysis presented in this chapter demon- technical experts from international and non- strates emphatically that the residents of low- governmental agencies would need to play a and medium-income countries bare a grossly core role in gathering and consolidating data and disproportionate share of the global burden of undertaking analysis. The participation of orga- armed violence. More specifically, lethal vio- nizations such as UNDP, the UN Office on Drugs lence in particular is associated with low and Crime, the World Health Organization, and the achievement of human development and MDGs World Bank are critical in this regard. Meanwhile, over time. The chapter thus draws attention to established research institutions—from think how a large cluster of countries is effectively tanks to academic research centres—are also trapped in cycles of armed violence and under- invaluable partners in such an endeavour. development.

Data collection would be directed on the basis The chapter makes a strong case to better inte- of a consolidated list of targets and indicators. grate the issues of armed violence into routine Minimum indicators to measure the dependent MDG assessments. Indeed, a statistical assess- variable—armed violence—include the number ment shows how countries exhibiting high and of direct and indirect conflict deaths, the preva- very high levels of homicide report comparatively lence of homicide, and the percentage of change low gains in relation to reducing extreme poverty in bilateral development assistance to armed and hunger (MDG 1), achieving universal primary See,forexample, World Bank (2011),HiCN (n.d.), and SeeCHS (2003) and Sen (2008). 3 SeeOECD (2009). 2 1 Endnotes UNDP OECD–DAC MDG HDI GDP AVMS Abbreviations of urgency. amatter as violence armed of and reduction prevention the topromote action totake sector development tothe call toaclarion finding amounts straightforward This anddevelopment. violence armed between relationship negative two-way andpersistent isacompelling There emerging. isnevertheless picture anirrefutable limitations, todata owing cases inmost remain, gaps edge knowl- While acknowledging. isincreasingly sector development the that insight—one apowerful are violence of rates andlow indevelopment ments invest- positive of effects reinforcing mutually The humanwelfare. of areas andother health, education, nourishment, reduction, to poverty inrelation gains higher disproportionately achieve to appear also violence of rates low register that countries Moreover, violence. armed of rates lower experience countries high-income that observes also chapter the optimistically, More 5). (MDG health maternal andimproving 4), (MDG 2), childmortality (MDG reducing education Human Index Development Gross domestic product domestic Gross Millennium Development Goal Millennium Development United Nations Development Programme Armed violence system monitoring Armed MICROCON (n.d.). Organisation for Economic Co-operation and for Economic Co-operation Organisation Development–Development Assistance Committee Committee Assistance Development–Development 2 There are60official indicator 12 See,forexample, UNDP(1990, p. 10). 11 Moser(2004), forexample, 10 Thereis considerable liter 9 See,forexample, Ayres (1998, whodescribes p. 24), See,forexample, OECD (2009). 8 7 This chapter builds on initial findings presentedinMuggah 6 Theannual estimate is based onglobal data fortheperiod 5 See,forexample, Bozzoli, Brück, and Sottsas (2010)and 4 and onacontinuum. she envisionsthese manifestations. Ratherthan ascribing afixed typology, social—and alexicon of attrib ries of violence—political, institutional, economic, and youth female (percentage of female labour force ages age of male labour force ages 9)unemployment, 15–24), force ages 8)unemployment, 15–24), youth male (percent- 7) unemployment, youth total (percentage of total labour 6) children under fiveseverely five moderately orseverely population ratio (15+,total, percentage), 5)children under income orconsumption (percentage), employment-to- 4) (PPP, percentage), 3)poorest quintile’s share innational (PPP, percentage), 2)poverty gap ratio at USD1perday this analysis at UNSD(n.d.c). The 21 specific MDG variables usedin (n.d.b). The entireMDGdatabase can bedownloaded into 171 specific variables, able at UNSD(n.d.a). These 60indic all eight MDGs. The official list of MDGindicators is avail- by Miller (2009). by Mariet al.(2008)andforhomicide survivors specifically populations arealsoaddressed inurban violence cases logical distress, depression, andanxiety indifferent aggressive behaviour andcognitive impairment.Psycho- the effects of political violence onthedevelopment of and Tremblay, Pedersen, andErrazuriz (2009) focus on by Dubow, Huesmann, andBoxer (2009). Cairns (1994) of violence, includingonchild subject to violence. The politic ioural, andpsychological impacts of witnessing orbeing assessmentsparticipatory of violence. also MoserandHoll violence as ‘theundue exercise of physical force’. See and Wennmann(2010). 2004–09. Stewart andFitzGerald (2001). are: 1)populationbelo and (1997)Moser(2004) for violence categories asoverlapping ature onthecognitive, behav- which areavailable at UNSD underweight (percentage), identifies at least fourcatego- al andpsychological impacts ren and youths, arereviewed s formonitoring progress of uted characteristics and underweight (percentage), ators aredisaggregated w USD1perday 169 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 170 15–24), 10) literacy rate (youth total, percentage of people cide rates in 1986–2009. For additional information on the ages 15–24), 11) total net enrolment in primary education, descriptive statistics of homicide rates, see the online 12) share of women employed in the non-agricultural methodological annexe at www.genevadeclaration.org.

sector (percentage of total non-agricultural employment), 21 This graph excludes 12 UN member states (for which HDI 13) ratio of girls to boys in primary and secondary educa- classiﬁcation is lacking): Iraq, Kiribati, Marshall Islands, tion (percentage), 14) mortality rate (under ﬁve, per 1,000), Micronesia, Monaco, Nauru, North Korea, Palau, San 15) mortality rate (infant per 1,000 live births), 16) adoles- Marino, Somalia, Tuvalu, and Zimbabwe. cent fertility rate (births per 1,000 women ages 15–19), 22 Overall, 18 countries exhibit very high homicide rates and 17) births attended by skilled health staff (percentage 49 feature high homicide rates. The countries with very of total), 18) prevalence of HIV (total, percentage of the high homicide levels are: Belize, Brazil, Central African population ages 15–49), 19) improved water source Republic, Colombia, Democratic Republic of the Congo, (percentage of the population with access), 20) improved Dominican Republic, El Salvador, Equatorial Guinea, sanitation facilities (percentage of the population with Guatemala, Honduras, Jamaica, Lesotho, Malawi, Republic

ARMED VIOLENCE 2011 VIOLENCE ARMED access), and 21) slum population (percentage of the of the Congo, St. Lucia, South Africa, Swaziland, and of urban population). Venezuela. The countries with high homicide levels are: Three of the World Bank’s World Development Indicators Albania, Angola, Bahamas, Barbados, Belarus, Benin, are used: 1) gross national income per capita, 2) poverty Botswana, Burundi, Cameroon, Chad, Côte d’Ivoire, headcount ratio at the national poverty line, and 3) ratio Dominica, Ecuador, Eritrea, Estonia, Ethiopia, Gabon, of girls to boys in primary and secondary education. Georgia, Guinea, Guinea-Bissau, Guyana, Haiti, Kazakh- The UNDP indicator used is the Human Development Index. stan, Kyrgyzstan, Latvia, Lithuania, Mali, Mauritania, For details on indicators and definitions used, see the online Mexico, Mongolia, Mozambique, Namibia, Nicaragua, methodological annexe at www.genevadeclaration.org. Panama, Papua New Guinea, Paraguay, Peru, Philippines, GLOBAL BURDEN BURDEN GLOBAL Russian Federation, St. Kitts and Nevis, St. Vincent and 13 See Alda (2010), Altbeker (2005), LaFree (2000), Leggett the Grenadines, Sudan, Suriname, Tanzania, Thailand, et al. (2005), and Muggah (2005). Timor–Leste, Togo, Trinidad and Tobago, and Uganda. 14 See UNHCR (n.d.). 23 These findings emerge from a regression logit analysis in 15 See the online methodological annexe at www.geneva which the dependent or left-hand side variable is a dichoto- declaration.org. mous variable, taking the value of one when a country 16 The following section applies a range of statistical methods shows an improvement in the indicator during the 2000– such as correlation analysis and econometric regressions. 09 period. In this way, it is possible to capture the size For details, see the online methodological annexe at and statistical importance of potential factors affecting www.genevadeclaration.org. the likelihood of HDI change. The model parameters were 17 For example, gross national income per capita is widely estimated using maximum likelihood procedures. Robust available, with 173 countries reporting 1,573 yearly data heteroscedasticity-corrected standard errors were also used. points. Other development indicators that are readily 24 By using a zero-score regression analysis, it is possible to accessible include adolescent birth rate and employment- gauge whether the normalized homicide level has an to-population ratio. By way of comparison, only 52 coun- effect on the normalized HDI level after controlling for tries reported on the ratio of girls to boys in primary other influences. A z-score regression was used on the education, with only 822 data points. Further, the poverty normalized HDI level on the levels of lethal violence as headcount ratio and the national poverty line data is only measured in quintiles. available for a limited number of countries. 25 The GBAV 2011 database features 29 ‘main’ armed conflicts 18 See the online methodological annexe at www.geneva in countries that were selected on the basis of an average declaration.org. of the levels and rates of direct conflict deaths between 19 The classification of countries according to HDI is available 2004 and 2009. Consult the online methodological annexe at UNDP (n.d.). at www.genevadeclaration.org for more information. 20 These ranges correspond to the mean and the mean plus 26 This study uses the World Bank’s classification of countries one standard deviation of the world distribution of homi- by income group. See World Bank (n.d.a). 36 The anal 35 In contrast, withres Thenegative correlation betw 34 Field(2010, p. 6)detects asignificant negative relation- 33 The World Bank alsoconfirms therelationship between 32 For example, countries as varied as theGambiaand 31 30 The anal Correlations arecalculat 29 TheUnited States is aninterestingcase: thecountry 28 27 The World Bank lethal violence andmalaria does not yield any direct cal tests. evident by visual inspection andis confirmed by statisti- tion percentage’. ‘poorest quintile’s share innational income orconsump- ship betweenanincrease inhomicide rates andthe correlation with homicidal violence (World Bank, 2011). (percentage of population) reveals apositive significant headcount ratio according tothenational poverty line poverty andhomicide levels.Correlating thepoverty violence. Solomon Islands donot exhibit above-average rates of for MDG8(developaglobal significant findings arereported. Spearman’s rho,andKendall’s tau. Only statistically OECD countries displaying highhomicidevery rates. Antilles, Rico andPuerto aretheonly high-income non- high homicide rate andEquatorial Guinea, theNetherlands and Tobago arenon-OECD high-income countries with a other hand, theBahamas, Barbados, Estonia,and Trinidad rateinjury of 23.6 per100,000 (FICAP, 2009, p. 5).Onthe 10.2 per100,000 population with anestimated non-fatal 2006 age-adjusteddeathra ‘external’ mortality after firearm-related injuries are thesecond leading cause of ally between1980 and2006. This figuresuggests that experienced roughly 32,300 firearm-related deaths annu- Marino, ortheUS New Caledonia, Mariana Islands, Northern Palau, San Greenland, Kosovo, Macao, Mayotte, Netherlands Antilles, American Samoa, Aruba,Cayman Isl able for207 of them.Homicide data is not available for more than 30,000 (215 total); income information is avail- all otherstates andterritories (28)with populations of ysis doesnotco ysis doesnotcons classifiesa Virgin Islands. pect tolit ed using Pearson’s correlation, nsider therel motor vehicle accidents.motor vehicle The robust statistical findings. or otherdiseases. Since malaria ider anyindicators ll partnership for development). fordevelopment). partnership te fromfirearm injur member countries eracy rate een thetwo variables is ands, French Polynesia, ationships between s, thean or variables alysis (187) y was and 9 TheDiliDeclaration represents aconsensus between 39 All of thesefindings coincide with thoserelated tohomi- 38 37 The anal 44 43 See Neapolit Not all children wereaffectedinthesame way by the 42 See,forexample, World Bank (2011). 41 SeeOECD (2010a;2010b).It is noting,however, worth that 40 46 There were no reported There werenoreported 46 With reference tothedistribution of homicide committed 45 See Muggah and Wennmann (2011). in 2004–09 (GBAV 2011database). national lines concrete commitments forgovernments andinter- goals forpeacebuilding andst and fragile states, theDiliDeclaration identifies seven effective andtailored tothec right track towards development responses that areboth To buildconsensus andput international actors onthe come of ameetingheldin Timor–Leste on9–10April 2010. both developinganddevelopedcountries andis theout- various government andcivil society representatives from effects. general; as such, they arenot cidal violence andprobably outside of those zones. countries, cross-country comparisons arenot possible rates per100,000) areonly co indicators (such as malaria See IDPS (2010, p. 1). WHO (2008). been found betweeninequality andhealth. See also pletely cut off fromany schooling (Hanemann, 2005). internally displaced orhandicapped children werecom- educational services than duringpeacetime. Incontrast, international relief organizations had abetteraccess to war. Those living inrefugee camps administrated by and Iraq accounted formost increases since 2000. the total population of thesesame countries. Afghanistan 43 ‘fragile states’, repre some 51percent of births registering low birth weight. registeringbirths lowbirth infant mortality, percentage of teenage mothers,and other indicators, includingHIV/AID low andhighhomicide rate districts whencompared with 7.0 and12.4per100,000, respectively. with firearms, thetwogroups of assist ysis isba an (199 ance toimprovesuppor sed ondatafor29 thisspendingwasconc 9). Asimil senting less than of three-quarters or significant variations between per 100,000 andmalaria death reinforce theseeffects in ontext of conflict-affected presentedhereas separate llected inmalaria-affected ate-building; it alsoout- arly robustre districts featured rates of S-related mortality, ‘main’armedconflicts t intheseprocesses. entrated in6of lation ship has 171 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 172 47 Consult SIM (n.d.). Bibliography 48 For example, private firms are frequently less inclined to Alda, Erik. 2010. Preventing Crime and Violence in Sub-Saharan invest and to generate new jobs where crime rates are Africa: A Silent Emergency. Unpublished paper. Washington, high (Krkoska and Robeck, 2006; Mihalache, 2008). DC: World Bank. 49 The negative effects of armed conflict on employment through the destruction of industry and infrastructure, Altbeker, Antony. 2005. ‘Positive Trends: The 2004/05 Crime as well as through the displacement of people, are Statistics.’ SA Crime Quarterly, No. 14. December. uncontested. 50 Recent statistics indicate that up to 82 per cent of murders Ayres, Robert. 1998. Crime and Violence as Development Issues and 59 per cent of attempted murders occur between in Latin America and the Caribbean. Washington, DC: people who know each other. See SAPS (2007, p. 3). World Bank. 52 Data drawn from South Africa (2009, p. 26).

ARMED VIOLENCE 2011 VIOLENCE ARMED Bhorat, Haroon and Carlene van der Westhuizen. 2008. Economic 53 The Gini coefﬁcient of inequality assigns values between of Growth, Poverty and Inequality in South Africa: The First 0 and 1 to each country, with 0 representing absolute Decade of Democracy. Paper commissioned by the Presi- equality and 1 representing absolute inequality. dency of the Republic of South Africa. 54 Data drawn from Bhorat and van der Westhuizen (2009, p. 16). Bourdieu, Pierre. 1986. ‘The Forms of Capital.’ In J. G. Richardson, 56 Data drawn from SAPS (2009). ed. Handbook of Theory and Research for the Sociology of

GLOBAL BURDEN BURDEN GLOBAL 57 See also Coleman (1988) and Bourdieu (1986) for a Education. Westport, CT: Greenwood, pp. 241–58. thorough review of social capital. Bourguignon, François. 1999. ‘Crime, Violence, and Inequitable 58 In its Millennium Development Goals Report 2010 , the Development.’ In Boris Pleskovic and Jospeh Stiglitz, eds. United Nations asserts: ‘Improved data and monitoring Annual World Bank Conference on Development Economics. tools are crucial for devising appropriate policies and Washington, DC: World Bank, pp. 199–220. interventions needed to achieve the MDGs. Although some and the challenge of building in-country capacity to produce better policy-relevant data is enormous’ (UN, 2010, p. 74). Bozzoli, Carlos, Tilman Brück, and Simon Sottsas. 2010. ‘A Survey of the Global Economic Costs of Conflict.’ Defence and 59 See WHO (n.d.a) for a review of Violence and Injury Pre- Peace Economics, Vol. 21, No. 2, pp. 165–76. vention publications. activities. Cairns, Ed, ed. 1994. ‘Children and Political Violence: An Overview.’ 61 See Muggah and Wennmann (2011). International Journal of Behavioral Development, Vol. 17, 62 Better-known observatories include the Violence Preven- No. 4, pp. 669–74. tion Alliance’s crime observatory in Jamaica; the Madrid —. 1996. Children and Political Violence. London: Wiley-Blackwell. security observatory; the Bogotá observatory; municipal observatories in El Salvador, Guatemala, and Panama; the Capobianco, Laura. 2005. Sharpening the Lens: Private Sector regional observatory on security policies in Italy; Central Involvement in Crime Prevention. Montreal: International America’s violence observatory; and the Observatoire Centre for the Prevention of Crime. National de la Délinquance in France. See Gilgen and 63 See Gilgen and Tracey (2011, p. 24). Carrière, Jean. 2008. Observatoires de la criminalité: Recension 64 On the involvement of the private sector in crime preven- d’experiences internationales. Montreal: International tion, see Capobianco (2005). Centre for the Prevention of Crime. Coleman, James.1988. ‘Social Capital intheCreation of Human CISALVA. 2008. Sistematización deExperiencias sobre Sistemas Bradford: University of Bradford. The Impact ofArmed Violence on chs-security-may03.pdf> files/resources/91BAEEDB New York: CHS. 794/Rendered/PDF/multi0page.pdf> Server/WDSP/IB/2002/12/14/000094946_02120504013 No. 2925. Washington, DC: World Bank. Inequality inSouthAfrica. pp. 16–38. Clinical Child Family Psychology Review, Research onChildrenand from a Social Ecological Perspective: Implications from Rates the Occurrence ofMurder inSix Areas withHighMurder Streets ofPain, Streets ofSorrow: TheCircumstances of Journal ofInternational Development GDH/Campagne_RC/War.pdf> Papers, Vol. 50, No.4,pp. 563–73. 0011/Rendered/PDF/multi_page.pdf> Server/WDSP/IB/2000/06/17/000094946_0006020542 Washington, DC: World Bank. Civil War. World Bank Policy Research Working Paper 2355. . Pretoria: Secretariat . Cali: Universidad del Valle. dbank.org/servl dbank.org/servl Policy Research Working Paper Families Ireland.’ inNorthern for Safety and Security. A50C6907C1256D19006A9353- —Observatorios deViolencia: a-content/docs/programmes/ Greed andGrievance in Poverty andDevelopment , Vol. 15,pp. 397–412. Human Security Now , Vol. 94,pp. 95–120. Crime andLocal et/WDSContent et/WDSContent Vol. 12,No.1, . . Gates, Scott and S. Mansoo Gates, Scottand S. Fleshman, Michael. 2001.‘Small Arms inAfrica: Counting the Field, Chris. 2010. Statistical Report onMDGandViolence Indi- (Firearm Center &Injury FICAP at Penn). 2009. ‘Firearm inthe Injury Dubow, Eric,L.Rowell Huesmann, andPaul Boxer. 2009. Harttgen, Kenneth and Stephan Klasen. 2010. Hanemann, Ulrike. 2005.Literacy inConflict Situations — andLauren Tracey. 2011. Gilgen, Elisabeth, Keith Krause, andRobert Muggah. 2010. Geneva Declaration Secretariat. 2008. GlobalBurden ofArmed —, et al.2010. %20Resource%20Book%20Upd pp. 113–26. Child andFamily Psychology Review Violence onChildren’s Psychosocial Adjustment.’ ing theImpact of Exposureto ‘A Social-Cognitive-Ecological Declaration Secretariat. Contributing Evidence to Programming UNESCO Institute for Indicators/Metrics_Paper.pdf> Consequences ofCivil Conflict. .prio.no/CSCW/Research-and-Publications/ . Oslo/New York: Ministry of Norwegian b Murshed.2005.‘Spatial–Horizontal Armed Violence Monitoring Systems: Education. ation.org/fileadmin/docs/ ation.org/fileadmin/docs/ Persistent Ethnic–Political ns Development Programme. Framework forUnderstand- ated%202009%20 Violence: Goals,Targets , Vol. 12,No.2, Vol. 15,No.4. . Geneva: Geneva Background paper Fragility andMDG EUI Working . Hamburg: . Hamburg: Clinical 173 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT 174 HiCN (Households in Conflict Network). n.d. ‘About HiCN.’ PENNSTATE_gamblingonconflict.pdf> IDPS (International Dialogue on Peacebuilding and Statebuilding). Miller, Laurence. 2009. ‘Family Survivors of Homicide: I. Symptoms, 2010. Dili Declaration: A New Vision for Peacebuilding Syndromes, and Reaction Patterns.’ American Journal of and Statebuilding. Paris: OECD. Family Therapy, Vol. 37, No. 1, pp. 67–79. informaworld.com/smpp/content~content=a907420726> Jaffe, Rivke. 2010. Fragmented Cities in the Caribbean: Violent Moser, Caroline. 2004. Urban Violence and Insecurity: An Intro- Crime and Sociospatial Cohesion in Jamaica and Curaçao. ductory Roadmap. Environment and Urbanization Brief Unpublished paper. Leiden: Research School CNWS 10. London: International Institute for Environment and Leiden University. Development. pagePK=64193027&piPK=64187937&theSitePK=523679 ARMED VIOLENCE 2011 VIOLENCE ARMED Krkoska, Libor and Katrin Robeck. 2006. The Impact of Crime on &menuPK=64187510&searchMenuPK=64187283&site of the Enterprise Sector: Transition versus Nontransition Name=WDS&entityID=000009265_3970716145117> Countries. Working Paper No. 97. London: European Bank Muggah, Robert. 2005. Armed Violence in Africa: Reflections for Reconstruction and Development. on the Costs of Crime and Conflict. Geneva: Small Arms Survey and United Nations Development Programme. Rates.’ Encyclopedia of Criminology and Deviant Behaviour, Vol. 2. London: Taylor and Francis. — and Achim Wennmann. 2010. More Violence, Less Development: Examining the Relationship between Armed Violence and Lederman, Daniel, Norman Loayza, and Ana María Menéndez.

GLOBAL BURDEN BURDEN GLOBAL MDG Achievement. Geneva Declaration Working Paper. 2002. ‘Violent Crime: Does Social Capital Matter?’ Economic Geneva: Geneva Declaration Secretariat. Development and Cultural Change, Vol. 50, No. 3, pp. 509–39. _3ay_3a2002_3ai_3a3_3ap_3a509-39.htm> — , eds. 2011. Mapping Security Promotion: Assessing Global Legget, Ted, et al. 2005. Why Fighting Crime Can Assist Develop- Armed Violence Reduction and Prevention. Paris: OECD. ment in Africa. Pretoria: Institute for Security Studies and United Nations Office on Drugs and Crime. Nafziger, E. Wayne, Frances Stewart, and Raimo Varynen. 2000. Emergencies. Oxford: Oxford University Press. Neapolitan, Jerome. 1999. ‘A Comparative Analysis of Nations Machel, Graça. 1996. Promotion and Protection of the Rights of with Low and High Levels of Violent Crime.’ Journal of Children: Impact of Armed Conflict on Children. Note by Criminal Justice, Vol. 27, No. 3, pp. 259–74. the Secretary-General. A/51/150 of 26 August. 3ay_3a1999_3ai_3a3_3ap_3a259-274.htm> Mari, Jair de Jesus, Marcelo Feijó de Mello, and Ivan Figueira. Neumayer, Eric. 2005. ‘Inequality and Violent Crime: Evidence 2008. ‘The Impact of Urban Violence on Mental Health.’ from Data on Robbery and Violent Theft.’ Journal of Peace Revista Brasileira de Psiquiatria, Vol. 30, No. 3, pp. 183–84. Research, Vol. 42, No. 1, pp. 101–12. 44462008000300002&script=sci_arttext> OECD (Organisation for Economic Co-operation and Development). McIlwaine, Cathy and Caroline Moser. 2001. ‘Violence and Social 2009. Armed Violence Reduction: Enabling Development. Capital in Urban Poor Communities: Perspectives from Paris: OECD. Development, Vol. 13, No. 7, pp. 965–84. —. 2010a. Accelerating Progress Towards the Millennium Devel- MICROCON (A Micro Level Analysis of Violent Conflict). n.d. opment Goals in Countries Affected by Conflict, Violence ‘MICROCON.’ and Insecurity. Document JT03281230. 2 April. Paris: Mihalache, Andreea. 2008. Gambling on Conflict: Profiling International Network on Conflict and Fragility, OECD. Investments in Conflict Countries. Paper presented at the —. n.d. ‘TheHuman Development Index (HDI).’ UNDP (UnitedNations Development Programme). 1990. UN (UnitedNations). 2010. Tremblay, Jacques, Duncan Pedersen, andConsuelo Errazuriz. — and Valpy FitzGerald. 2001. Sen, Amartya. 2008. ‘Violence, Identity andPoverty.’ —. 2009. AnnualReport—CrimeStatistics: April2008–March 2009 SAPS (SouthAfrican Police Service). 2007. CrimeSituation inSouth Putnam, Robert.1993. —. 2010b. Ensuring Fragile States Are Not Left Behind. Development Report low%20res%2020100615%20-.pdf> goals/pdf/MDG%20Report%202010%20En%20r15%20- Report 2010 of SocialPsychiatry Violence andCivil Unrest inPeru.’ 2009. ‘Assessing Mental Health Outcomes of Political Oxford University Press. Economic andSocial Consequences ofConflict Ethnicity, University of Oxford. Centre forResearch onInequality, Human Security and Dimension ofDevelopment gov.za/view/DownloadFileAction?id=109313> Presidency, Republic of South Africa. OPMENT/Resources/244362-1239390842422/6012763- de informações sobr Peace Research, Vol. 45,No.1,pp. 5–15. reports/crimestats/2009/crime_stats.htm> Johannesburg: SAPS. Information An Africa: 2006/2007 FinancialYear in Modern Italy.Princeton: Princeton University Press. ataoecd/38/33/44822042.pdf> . Working Paper 1.Oxford: The Costs ofViolence e mortalidade). n.d.e mortalidade). ‘Sistema . Johannesburg: Crime International Journal aps.gov.za/statistics/ ies: ANeglected . Pretoria: Paris: OECD. . Oxford: Journal of . Human . ns Statistics Di UNSD (UnitedNationsStatistics —. n.d.b. ‘Data Availability by Series andMDGRegion.’ ‘The Refugee in Story Statistics.’ a MDG Indicators.’ c5c8 dos Homicídios noBrasil Caribbean-study-en.pdf> DC, and Vienna: World Bank andUNODC. Policy Options in the Caribbean Crime, Violence, and Developmen PK:34480~pagePK:34370~theSitePK:4607,00.html> NEWS/0,,print:Y~isCURL:Y~contentMDK:20042303~menu 2011. commission/finalreport/en/index.html> WHO. sions ofInterpersonalViolence Handlers/ExportHandler.ashx?Type=Csv> spx?Content=Indicators/OfficialList.htm> 52575610076cc98/2d75d3e8dbc4 World Development Report:Conflict, Security, lassify Countries.’ lassify ank.org/about/country- Mapa daViolência 2010: Anatomia . Saõ Paolo: Istituto Sangari. vision). n.d.a.‘Officia . Report 37820. Washington, g/fulltext> . Geneva: WHO. t: Trends, Costs, and ata-and-an 1cb18 classific . Geneva: 525762 alysis/ l Listof ations> 3004 175 2 3 4 5 1 MORE VIOLENCE, LESS DEVELOPMENT