Chapter 2. Violence today

Wendy MacClinchy
Independant Consultant
Clionadh Raleigh
Armed Conflict Location and Event Data Project (ACLED)

This chapter provides an overview of the main findings about the complex violence landscape in the world today. After a brief discussion of the issues involved in the measurement of violence and conflict, it outlines how violence is increasing, and shows that violence is occurring in surprising places. The chapter then looks at the complexities of violence, its multidimensional nature, and how violence evolves and adapts to changing contexts. There is discussion of violence as a regional issue, including as people are forced from their homes and across borders, and how it is often also an urban problem. The chapter looks at the role of domestic political instability in driving violence, and outlines how civilians, especially civilian women and children, are most at risk. It concludes by outlining the escalating cost of violence, and investigates opportunities in treating violence as a behavioural problem. There are case studies on the Central African Republic and Kenya.


Of all victims of lethal violence, 83% lived outside of conflict zones.

Violence is one of many factors that can contribute to fragility. However, it is not the only factor, and the presence of violence does not automatically mean that a context is fragile. States of Fragility 2016 places a spotlight on violence, in all its forms, to explore how violence can contribute to fragility, and examine what should be done about it.

Ten findings show a broader, more complex violence landscape than has been considered in development policy to date. States and societies can experience multiple forms of violence simultaneously, each caused by related issues but with different locations, triggers and impacts on fragility. Across this landscape, variables may include increases in lethal violence, the adaptability and networking capacities of organised criminal networks, deepening political instability, the emergence of fragile cities, and the rise of violent extremism. These factors can combine in a sort of “contagion” effect that has complicated the way in which the international development community must view the effects of violence in every dimension of fragility. The following findings demonstrate the complex violence landscape that has emerged, offer insights that can usefully inform analysis and provide key elements of a roadmap for building an effective set of policies in response.

Finding One: Debates over how violence should be measured distract from a better understanding of the broader violence landscape

The way violence is assessed offers only a glimpse of the problem’s scale. By most measures global violence has become markedly worse. But the debate about how to measure violence, and thus define its severity, often focuses on conflict-related violence and thus misses the most important point. The vast majority of lethal violence victims, 83%, died outside of conflict according to the definition accepted in international humanitarian law (Box 2.1). Indeed, direct conflict deaths account for only 17% of the global total (Small Arms Survey, 2016) (Figure 2.1). Therefore, focusing analysis and programming primarily on conflict violence may mean turning a blind eye to the major part of the global violence problem.

Violence manifest in armed conflict, fatalities and refugees flows is at its highest level in over two decades.

Properly measuring violence will require a broad lens and systemic data. Standard measures often fail to capture violence in all its forms, and fail to recognise how different forms and manifestations of violence are interconnected. This is particularly the case regarding large-scale criminal violence and low-level armed conflict, which may have similar levels of intensity and casualties, but are not considered within a single set of measurements (Geneva Declaration Secretariat, 2015). For example, the World Health Organization reports that more than 1.3 million people worldwide die each year as a result of self-directed, interpersonal or collective violence, accounting for 2.5% of global mortality (WHO, 2014). Other research concludes that armed conflict fatalities are at their highest numbers in over two decades (Gates et al., 2016; Uppsala University, 2016; ICG, 2016). The International Institute for Strategic Studies found that conflict-related deaths totalled 167 000 in 2015, with 55 000 of those in the Syrian Arab Republic (hereafter “Syria”) (IISS, 2016). “Battle deaths” also tripled since 2003 (Gates et al., 2016; Uppsala University, 2016). These different measurements provide at best a fragmented set of statistics and at worst a distorted picture of the reality on the ground. A coherent framework for bringing these different measures together is needed.

Box 2.1. Defining conflict

Researchers use a variety of criteria to define conflict including the nature of the conflict, the type of actors, the number of deaths and the parties’ level of organisation. In this report, while every effort has been made to provide consistent data, results cited will depend on the different definitions of conflict that are used by the different researchers.

International humanitarian law differentiates international armed conflicts between states using armed force from non-international armed conflict where hostilities reach “a minimum level of intensity” and parties show “a minimum” of organisation.

The Uppsala University Conflict Data Program (UCDP) defines armed conflict as meeting a threshold of 25 battle deaths per calendar year; it can be state-based, involving at least one state party, or non-state based. A variant is what the UCDP calls “one-sided violence”, where an organised group deliberately kills at least 25 civilians in a year.

The Armed Conflict Location and Event Data (ACLED) project uses subnational data to analyse “political violence” or “civil conflict”, which encompass “diverse but recurrent forms of violence between individuals and groups” including “organized violent crime, gang warfare, terrorism, religious and sectarian rebellions, and spontaneous riots or violent protest over state failures such as poor or absent service delivery”.

Sources: ICRC (2008); Uppsala University (n.d), ACLED (2015a).

Figure 2.1. Violent deaths, by category, 2010-15

Source: Small Arms Survey (2016).

Even data on conflicts are contested. There is debate today over whether the world is seeing fewer but more deadly armed conflicts (IISS, 2015; Geneva Declaration Secretariat, 2015) or more conflicts overall (ICG, 2016; Uppsala University, 2016). Research into conflict-related violence produces wide-ranging results, mainly due to difficulties in comparing data (Box 2.2). One set of research, for example, concludes the number of active civil and transnational wars has declined (Fearon and Laitin, 2003; Human Security Centre, 2005; Newman, 2009; Straus, 2012); other research concludes that civil conflicts1 almost tripled from 2007 through 20132 (Gates et al., 2016; Uppsala University, 2016), with an increase to 50 in 2015 compared to 41 in 2014. Similarly, while there is general consensus that traditional interstate conflicts are decreasing, there is equally strong consensus that violence driven by domestic political instability3 and social violence is increasing (Uppsala University, 2016).

Box 2.2. Difficulties in comparing and consolidating violence data

Measuring levels and dynamics of violence remains a challenge. Data may be unreliable due to underreporting and misreporting, and so require serious scrutiny. There is an urgent need for consistent, systematic and comprehensive data to reliably track variations in exposure, risk and harm across contexts and over time. This would facilitate greater comparability and analysis of trends and patterns.

Some conflict monitoring is done by the Armed Conflict Location and Event Data Project (ACLED), the Global Terrorism Database (GTD) and the Small Arms Survey, among others, and is also done in country-specific datasets such as Syria Tracker and Nigeria Watch. Social violence is often measured using homicide or violent crime statistics, with global monitoring carried out by the United Nations Office on Drugs and Crime (UNODC), the World Health Organization (WHO), and country and regional violence observatories, such as the Igarapé Institute.

Discrepancies in definitions of violence across datasets limit opportunities to combine them because they differ in the time periods covered, on fundamental definitions, on inclusion criteria and in the structure of analytical units. In addition, there is very little disaggregation of violence, and of conflict in particular, at the subnational level. Carayannis et al. (2014), in a recent literature review, found that only a small portion (3.1%) of studies on conflict included local-level empirical data. These data problems perpetuate limited understanding of the different impacts of violence across individuals, groups and communities.

The “unified approach” to measuring lethal violence, used by the Geneva Declaration Secretariat in the Global Burden of Armed Violence reports, takes a broad view in how it calculates violent deaths, consolidating normally disparate data sources on conflict, crime, homicide, interpersonal and other forms of violence together into a composite whole. In addition to offering clear advantages in overcoming data challenges, this approach also facilitates a better-informed analysis, as well as a clearer basis for monitoring indicators and identifying trends and risks over time and across contexts. It also enables a more holistic programmatic response and therefore increases prospects for effectiveness. By including a fuller range of risks, this approach facilitates the risk management approach called for in the World Development Report 2014.

Sources: Carayannis et al. (2014); World Bank (2013).

The complex and changing nature of social violence makes its forms even more difficult to define and measure. Social violence can often include a broad manifestation of some form of grievance, criminal behaviour and interpersonal violence such as homicide, gender-based violence and self-directed violence (WHO, 2002). It can be collective, for instance gangs, or individual, such as intimate partner violence. These multiple shapes of social violence make it especially difficult to measure. The Small Arms Survey has made some good progress on this issue, multisourcing a time series database on violent deaths, which informs the Global Burden of Armed Violence report (Geneva Declaration Secretariat, 2015). These data are producing a global baseline of violent deaths for Sustainable Development Goal (SDG) 16.1. Importantly, this analysis focuses on global trends in violent deaths as a composite indicator that brings together data on homicide and direct conflict deaths (Box 2.3).

A lack of systemitised data makes it difficult to assess and compare the severity of concurrent violence across the world. A broad lens is needed to capture the spillover and interactions among different types of violence and in different contexts. More data on the gender dimensions of violence and conflict are also needed, given the disproportionate impact of violence on women.

Box 2.3. Small Arms Survey data on violence
  • Over 2010-15, an average of 535 000 people were killed violently every year.

  • The number of people in conflict is growing. The annual average of recorded deaths in 2010-15 rose to 90 000 from 70 000 deaths in 2007-12. Armed conflicts in Afghanistan, Iraq and Syria are responsible for a large proportion of these deaths.

  • The global homicide rate continues to decrease, but not enough to offset the increase in conflict deaths in 2010–15.

  • The vast majority (83%) of victims of lethal violence do not die in conflict zones.

  • The global distribution of violence is becoming increasingly unequal. Fewer countries are registering high violent death rates (above 20 per 100 000 population), but their average violent death rates are on the rise.

  • In absolute numbers, more lives were lost to violence in 2015 in large countries such as Brazil and India, which were not experiencing conflict, than in war-torn Syria.

  • Of the 20 countries with high violent death rates in 2015, 11 were not affected by armed conflict.

Source: Geneva Declaration Secretariat (2015).

Finding Two: There is more violence, and it is occurring in surprising places

Violence has been on the rise in recent years. The world has become more violent over the last decade, interrupting a long-term trend of increasing peace (IEP, 2016), and with a significant increase since 2014. In terms of fatalities, 2014 and 2015 were the second-worst and third-worst years since the Cold War (Uppsala University, 2016; Melander, 2015). Over the last 15 years, 53 contexts have been or are now affected by some form of political violence. These comprise 3.34 billion people, or almost half of the world’s population.

The emergence of violence in certain locations can surprise. Many contexts now experiencing conflict fatalities have only recently become considered as fragile (e.g. Syria) (Figure 2.2).

Figure 2.2. Political violence event levels and average micromort (likelihood of violent death) selected countries, 2014

Sources: Africa and Asia data (ACLED, 2016); and ACLED Asia Version 1, 2016; Syrian Arab Republic data from Syria Tracker through 2015; Ukraine data, UN.

Violence in various forms has also reached epidemic proportions in contexts unaffected by political armed violence and not typically considered fragile (Figure 2.3). According to 2015 figures, 11 out of the 20 contexts with high lethal violence rates were not affected by armed conflict. While fewer countries show high violent death rates (above 20 per 100 000 population), average violent death rates in these countries are on the rise. This widening disparity leads to some surprising results. In 2015, more people died violently in countries outside of conflict, such as Brazil and India, than in Syria in absolute numbers (Small Arms Survey, 2016). The highest homicide and violent crime rates in the world are found in Central America and the Caribbean where urban gang violence and drug-related crime are features of everyday life (Geneva Declaration Secretariat, 2015).

Figure 2.3. Where is the risk of lethal violence highest?
Contexts where the risk of death from terrorism, conflict and homicide was highest in 2014

Source: Muggah (2016).

Globally, conflict is not the leading cause of violent death. Social violence in the form of homicide can be more deadly than war in some contexts. Of the 37 countries most affected by lethal violence in 2012, 83% were not emerging from or recently experiencing conflict (Geneva Declaration Secretariat, 2015). The sub-region4 most affected by lethal violence is Central America (with a rate of 33.6 violent deaths per 100 000 population), followed by Southern Africa5 (31.2), the Caribbean (20.5) and South America (17) (Geneva Declaration Secretariat, 2015). In fact, one-third of all homicides in the world occur in Latin America and the Caribbean (LAC), home to just 8% of the world’s population (Vilalta, 2015; Szabó de Carvalho and Muggah, 2016). In 2015, El Salvador’s murder rate climbed to 103 per 100 000 (ICG, 2016). LAC is the only region in the world where rates of lethal violence have increased since 2000 (UNDP, 2014) and where homicide rates continue to increase (Vilalta, 2015) (Figure 2.4).

A cluster of global murder “hotspots” are found in the Americas, where 46 out of the 50 most violent cities are concentrated. Beyond pockets of extreme homicidal violence, the risk of murder is also more widely distributed than violent deaths associated with terrorism or war. In fact, in the past decade, terror and conflict represented only a portion of all violent deaths: 9 out of every 10 violent deaths were homicides (Muggah, 2016).

Figure 2.4. Average homicide rates per region, 2000 and 2012

Source: Vilalta (2015).

Development and violence are not mutually exclusive. Low-and middle-income countries bear a disproportionately high share of the burden of political and social armed violence, which often impedes development gains (De Martino, 2012). In these contexts, violent conflict and political instability prevent progress towards development targets, such as in South Sudan or Yemen. However, higher-income countries are also affected by high levels of social violence including intimate partner violence, crime and organised crime (such as narcotics and trafficking). Indeed, against a 5% declining global trend in intentional homicide,6. the Americas was the only region to show a significant increase (nearly 10%) in a comparison of data for the periods of 2004-09 and 2007-12 (Geneva Declaration Secretariat, 2015). A major social issue in the United States is the high number of deaths and injuries related to armed violence. These bleak facts suggest that neither development nor wealth assures an escape from violence.

Finding Three: Violence is multidimensional, complex and evolving

Violence is shape-shifting. Even in post-conflict contexts, violence can simply change its form as settings, actors and drivers change. Its versatility in the face of shifting risk landscapes, and ranges of non-state actors, challenges the international community’s best intervention efforts. Armed groups and militia organisations, often operating on behalf of political elites, are responsible for an increasing proportion of armed violence. Often they also engage in criminal and social forms of violence, spreading fear and instability while consolidating their own power in the process. A report recently published by the UN University advised the UN to “recognize the political power of criminal groups” in order to limit the influence of organised crime in transitional political processes (Bosetti, Cockayne and de Boer, 2016). The compounding of these security risks, and their continuity in one form or another, challenge the coping capacity of states and societies.

Violence is complex, and responses should not be based on simplistic narratives. Violence does not fit neatly into customary security frameworks or conflict narratives, and is often treated subjectively within different organisations. Approaches that view violence in terms of “perpetrators versus victims” and “criminals versus innocent citizens” are not helpful for understanding the complexity of violence (Adams, 2012). Violence is tremendously versatile, transforming itself according to changing circumstances and contexts. Responses to violence, then, can inadvertently compound it. Broad punitive measures can sweep up non-violent individuals or fail to account for social norms, motivations and other factors. As a result, they can deepen marginalisation, foster mistrust for the rule of law or incite more violence.

Political transitions, even towards democracy, can provoke violence. As a state moves towards or away from democracy or devolution, the risk of different forms of violence also changes. Civil war violence often emerges from exclusive politics, where large sections of the population are disenfranchised or marginalised. In transitioning and democratising states, competition over “who sits at the table” and “who gets what” can also become violent. The risk of overlapping forms of violence also may increase during transition. The political violence of the Arab Spring was accompanied by sexual and other forms of gender-based violence (SGBV) and gendered repression. In Tunisia, for example, women protesters became the victims of sexual harassment or rape at the hands of security forces (Johansson-Nogués, 2013); in Egypt, politically motivated acts of sexual assault were witnessed in protest spaces (Tadros, 2015; Amar, 2013). Even where the trend is towards democracy, there may be an increased risk of domestic political instability (Choi and Raleigh, 2014). Both decentralisation and power distribution across parliaments, judiciaries and the military tend to increase the number and power of non-regime elites shaping government policy. They also lead to high levels of elite competition and fragmentation over access to state resources and power (Brancati, 2011). In the redistribution of political power, incumbents and opponents have incentives to design forms of violence to assure access to power (Schedler, 2006; Gandhi and Lust-Okar, 2009; Arriola and Johnson, 2012). Many agents in new democracies depend on violence to create cleavages in society, which elites can manipulate. The transition to democracy can thus dramatically increase the risk of violence, even while lowering the chance of interstate conflict. Studies looking solely at conflict ignore this.

Drivers of violence do not disappear when conflict is over. Political violence has a cyclical nature if the factors that provoked it remain unaddressed. Recent research suggests that factors affecting and producing the likely onset of civil war also influence other forms of non-political violence such as social violence (Rivera, 2016). Armed agents who engage in political violence during wars or periods of domestic instability are highly likely to be involved in organised crime in more peaceful periods, including racketeering, mercenary activity and illicit trafficking. These patterns are evident across both Africa and Asia. Violent actors in conflicts reconstitute themselves in post-conflict periods to take economic and political advantage of fragile and new political environments (von Einsiedel et al., 2014; de Boer and Bosetti, 2015).

The Central African Republic provides a good example of the complexity of violence (Box 2.4).

Box 2.4. Violence and fragility case study: Central African Republic

The Central African Republic (CAR) serves as an example of an acutely fragile state. It experiences long-running and cyclical waves of violence. As most political violence has occurred in rural locations, it has received comparatively little international attention. Conflict levels in 2015 have declined from the most intense periods during 2013 and 2014, and recent political developments (i.e. the holding of elections) may signify a move toward securing more lasting peace. Nevertheless, the underlying dynamics, drivers and agents of recent and ongoing conflict demonstrate key characteristics of violence and its associated challenges.

A wide range of violent actors and the many simultaneous forms of violence intersect to create conditions of acute instability, civilian vulnerability and sexual violence in the CAR. Actors include groups such as the Anti-Balaka and Séléka militias that mobilised domestically, as well as the regionally active Lord’s Resistance Army (LRA) and a host of communal militia groups (Figure 2.5). In 2014, ACLED recorded over 40 discrete non-state conflict agents active in the CAR (ACLED, 2015b). Violent actors include not only rebel groups and militias, but also armed criminal elements, known as zaraguinas. Such groups are not directly party to political negotiations, but have nevertheless severely affected civilian vulnerability and protection (HRW, 2009). This proliferation is a testament to the overlapping forms of violence and conflict affecting the state, and the multidimensional nature of the crisis.

A rise in diverse forms of sexual and gender-based violence by armed groups, intimate partners and international forces is further testament to the multidimensional nature of fragility in the CAR and the overlapping forms of violence there. All parties to the conflict are reported to have used sexual violence to subjugate and humiliate opponents (UN Security Council, 2016). Women and girls have been systematically targeted, with the International Rescue Committee reporting that sexual violence is the biggest fear of women in the CAR (IRC, 2014). There have also been cases of conflict-related sexual violence against men and boys. In addition, women have reported increases in domestic violence. A series of allegations of sexual and abuse and exploitation have also been made against UN and non-UN forces tasked with maintaining peace and security in the CAR, which the UN is investigating (Deschamps, Jallow and Sooka, 2015).

Violence changes form as conditions and actors change. Over time, some of the CAR’s violent actors have disbanded and reconstituted themselves. An example is the Séléka coalition. Associated conflict patterns show that violence changed considerably with these transformations. Séléka violence, for instance, has been both more frequent, and more directed at civilians, than the violence attributed to its historical predecessors (Duffy, 2015). These patterns suggest that analyses of violence in Africa need to be more sensitive to complex and evolving conditions on the ground, rather than focusing on religious or ethnic identities as drivers.

A non-neutral and weak state perpetuates fragility. Historically, the state in CAR has struggled to exert its control outside the capital (Bierschenk and de Sardan, 1997). However, fragility should not be seen simply in terms of weak state capacity but also in terms of how particular interests in the conflict benefit from that weakness. The illicit and unregulated exploitation of resources, particular in the extractive sector, has flourished, in part because the state has limited capacity to control such activities. Elites have profited directly (Dalby, 2015; Smith, 2015). The state and its elites should therefore be understood as non-neutral actors. Fragility can serve their economic interests, which in turn may result in the diversion of resources to fund anti-state violence.

Competing militia groups exacerbate fragility and weak state capacity. State security forces have been largely absent in the conflict in the CAR in recent years. In 2013, state forces were reported to be involved in just over 5% of all recorded political violence events; in 2014, they were involved in fewer than 2%, and despite a general increase in insecurity, state forces were involved in just over 3% of conflict events in 2015 (ACLED, 2016). Throughout this time, rebels, militias and international forces played a more significant role. The emergence of local vigilante or so-called “defence” militias like the Anti-Balaka is due in part to the absence or weakness of the state. But the actions of these non-state actors also speak to a wider context of impunity, elite-sponsored violence, and the targeting of civilians in the CAR and other acutely fragile contexts.

Conflict in the CAR is highly localised. The geographic profile of violence in the CAR attests to the multiple and localised dynamics of insecurity experienced within a single state. Political violence has been concentrated in the south and west of the country, primarily involving state, international, and Séléka and Anti-Balaka forces. The more limited violence in the north and east has been largely attributed to the Ugandan Lord’s Resistance Army (LRA), which has preyed on populations in eastern CAR, leveraging the limited state presence in this relatively remote region (Cakaj, 2015).

Figure 2.5. Who commits acts of violence in the Central African Republic, and where?
Number of conflict events by actor in the Central African Republic, 2013-15

Source: ACLED (2016).

The CAR conflict has impacts across the region. The successive crises in the CAR have a strong regional dimension, as do the conflicts that affect several other fragile states and contexts in Africa. Regional interest in and influence over domestic CAR politics have shaped the outcome of political contests; neighbouring powers played the role of king maker and regional stakeholders offered support to political actors within the CAR (Marchal, 2015). Additionally, over the past 15 years, the interplay of overlapping crises in the CAR, Chad, the Democratic Republic of the Congo, South Sudan and Sudan has created a dynamic, regional conflict system. The impact of population displacement, in particular, is regional. People, including combatants and former combatants, have sought refuge in the CAR from various crises in Chad, South Sudan and Sudan. As well, persistent cyclical crises in the CAR have resulted in regional and international interventions on numerous occasions (Kilembe, 2015; Olin, 2015).

The effects of the CAR crises also spillover to other countries in the region: more than half of the almost 500 000 people fleeing the most recent conflict in the CAR fled to Cameroon and more than one-fifth fled to other fragile contexts such as Chad and the Democratic Republic of the Congo (WFP, 2015).

Sources: ACLED (2015b; 2016); Bierschenk and de Sardan (1997); Cakaj (2015); Dalby (2015); Duffy (2015); HRW (2009); IRC (2014); Kilembe (2015); Marchal (2015); Olin (2015); Smith (2015); Deschamps, Jallow and Sooka (2015); UN Security Council (2016); WFP (2015).

Finding Four: Violence is increasingly a regional problem

Political armed violence can more easily spillover between states. Global communications and transboundary criminal networks connect armed actors more easily than ever before, allowing for co-operation in motives and resources. Several of the world’s most persistent conflicts are regional (Figure 2.6). These include historical conflicts in the Great Lakes region of Africa; Pakistan, India and Afghanistan; and Syria and Iraq, as well as the networked transnational diffusion of violence in northern Nigeria, Cameroon and Chad across the Sahel. Violence affects regional stability, as neighbours’ domestic instability spills over borders, driven by linkages between aggrieved or armed groups that may share a common identity, loyalty or objective. Shared ideologies also bond groups across regions. Flows of financial, logistical and troop resources facilitate the linkages, extending the reach and consequences of violence. They also prompt states and armed actors to disregard boundaries, inciting intervention in one another’s domestic politics and wreaking devastating havoc on civilians in their wake.

Figure 2.6. Conflict events, Africa 1997-2015

Source: ACLED (2015b).

Violence is networked and knows no borders. The organised crime-political violence nexus allows political armed groups to finance themselves through proceeds from criminal activities, with illegal resource exploitation and the drug trade providing revenue (Hansen, 2014). Even informally organised groups engaging in targeted, armed competition and committing violence for political elites often engage in crime to sustain their own activities. Many of these criminal networks, particularly those engaged in organised illicit trafficking, cross borders. Violence in various forms crosses borders with them. Organised criminal groups, within and among states, exploit networks to corrupt politicians and influence domestic politics. Illicit financial flows also travel across these networks, feeding this negative feedback loop, for example in West Africa (OECD, forthcoming). Connections among types of networked violence enable the continued presence of violence in a variety of contexts. For example, Mali’s political violence creates a governance and rule of law vacuum which South American drug cartels exploit for trafficking to European counterparts, prompting an increase in cocaine trafficking through the Sahel as a transit route (Ellis, 2009). This activity ignites local grievances, which Islamist extremists also exploit in the security vacuum, creating a deeper negative spiral of violence (Box 2.5). In this context, nascent lucrative “business opportunities” for violence emerge, such as human trafficking along established drug trafficking routes (Shelley, 2014).

Box 2.5. The globalisation of local conflict: The case of ISIS

The emergence of the subnational movement known as the so-called Islamic State, also referred to as “ISIS”, was made possible by pre-existing conflict and the collapse of state control in parts of Iraq and Syria. ISIS has quickly expanded its reach on a regional and international scale. The extremist group exploited possibilities to move fighters, money, oil, supplies and other lucrative resources across borders, which also led to a fundamental change in its strategy. Not only are cross-border activities essential to the group’s ability to fund its operations, but borderless information technology allows it to recruit new activists from across the world.

ISIS also strategically spread its attacks to targets outside its region of origin. The group has claimed responsibility for attacks by individuals and affiliates in Belgium, Egypt’s Sinai Peninsula, France, Lebanon, Libya, Nigeria, Tunisia, Turkey and the United States, among other places. In addition to this transnational activity, ISIS continues to control parts of Iraq and Syria, and engages in kidnappings and executions of civilians of many nationalities.

Source: Glenn (2016).

Violence drives millions from their homes, thereby extending its impacts to often overburdened neighbouring countries already in distress. Forced displacement is one of the most profound non-lethal impacts of violence. It is a direct consequence of violence in the form of war and persecution. In 2015, 65.3 million people were forcibly displaced worldwide, according to the United Nations High Commissioner for Refugees (UNHCR); that was the highest level recorded since the refugee organisation was established in 1950, and the figure continues to rise. On average, every minute 24 people had to flee their homes in 2015 (UNHCR, 2016a). Internal displacement associated with conflict and violence has been on an upward trend since 2003. There were 8.6 million new cases during 2015, or an average of 24 000 a day (IDMC, 2016). While the last two years have seen unprecedented numbers of refugees and migrants fleeing fragile and conflict-affected countries to Europe, most refugees are hosted in developing countries, straining these countries’ already overburdened capacities. Sub-Saharan Africa, for example, hosts more than 26% of the world’s refugees (UNHCR, 2016b). Many refugees flee to countries that are already fragile.

Finding Five: Violence is increasingly driven by domestic political instability

State legitimacy, authority and capacity are primary root causes of political violence. Domestic political factors are among the most challenging causes of violence to effectively analyse and address. Yet the evidence suggests that domestic political dynamics determine whether a society tips into political violence. These may include corruption, financing opportunities, and external and internal shocks such as elections and demographic shifts (Clapham, 1996; Englebert, 2000). Along with poor governance, breakdown of order (Reno, 2011) and weak institutions (Sobek, 2010; Hendrix, 2010; Thies, 2010) these factors interact, creating openings for collective armed violence.

Politically motivated violent conflict has distinct and varied manifestations which differ according to who fights, which political issues are contested, the vulnerability of governments and civilians, the feasibility of long-running conflict, and how likely conflict is to persist, recur and diffuse. Examples of active conflicts reveal significant variations:

  • An active civil and transnational war in Syria, where multiple, unaffiliated rebel organisations seek to replace the current regime. In neighbouring Iraq, political elites7 create militias to compete with other political agents. These same militias act as both security forces for some communities, and predatory forces in others.

  • Multiple, clustered political challenges in countries such as India and the Philippines, where organised armed agents fight the state, but other forms of violence are common and widespread, including local communal conflicts, organised criminal violence, and widespread rioting and protests.

  • Challenges to central government legitimacy and capacity in Pakistan and Yemen that stem from a breakdown in the political settlement8 between central government and opposition groups.

  • Violent stalemates in the Central African Republic and South Sudan, where no single territorial authority exists or has widespread control over multiple armed, organised groups.

  • Intractable conflicts including Somalia and the Democratic Republic of the Congo, where fighting has persisted for decades, and dozens of armed groups compete for local control while co-ordinating, temporarily allying or competing with larger groups.

Weak state structures are a source of violence, potentially leading to a negative feedback loop between political fragility and violence. Weak institutions or those with entrenched patronage systems can create vacuums in which elites are able to siphon off public resources with impunity while also perpetuating economic exclusion. Criminal networks and armed groups can also fill these vacuums, exploiting local grievances while enabling other forms of social violence to spread. In fragile contexts, poor or unequal service provision may fuel unrest or violent crime, particularly when coupled with economic deprivation. Weak governance can also allow non-state actors to create parallel structures, increasing the risk of widespread criminality and related violence. The nature of local authorities differs. Conflicts such as the Chad Basin (Roitman, 2001), the Ituri province conflict in the Democratic Republic of the Congo (Vlassenroot and Raeymaekers, 2004), and the onset of the crisis in northern Mali in 2011 all followed similar trajectories where the withering away of the state allowed local elites to replace state authority (Blattman and Miguel, 2010; Justino, 2012; Justino, Brück and Verwimp, 2013; Kalyvas, 2003, 2008). In parts of Africa and the Middle East, a “rentier political marketplace”, as described by de Waal, is a particular challenge (de Waal, 2014). In these instances, violence, or the threat of violence, is used as a means of political bargaining when the government or political elites do not have a monopoly on the legitimate use of force.

States can use identity politics to incite and exploit sectarian divisions. In fragile contexts, distinctions of ethnicity, religion or livelihood can be more pronounced and are often a source of political identity. As these identities are flexible, allegiances of convenience can form in a changing context, and can transfer long-standing grievances into new causes, and thereby serve as a driver of new forms of violence (Fearon and Laitin, 2003; Collier and Hoeffler, 2004).

Political inequalities drive violence. A persistent problem is the tension between “excluded” and “included” groups that have different access to, and exercise of, power, with consequences for government policies and related socio-economic inequalities (Stewart, 2011). Exclusion along ethnic lines leads to limited representation in public offices (Bangura, 2006); poorer levels of health and education; greater income inequalities (Stewart, 2008); and limited public good provision (La Porta et al., 1999).9 Members of excluded groups are more likely to engage in civil war, particularly if they have recently lost access to power (Cederman, Wimmer and Min, 2010).

Yet the terms of inclusion can also drive violence. Increasingly, the terms of inclusion within governments – that is, the distribution of positions, authority and resources among included elites – can drive domestic political instability (Fischer, 2008; Lindemann, 2008) and violence. The extent to which the state is a non-neutral, politicised and active agent in conflict is often underestimated. A focus on achieving stability through the inclusion of violent elites can reduce violence levels in the short term, but may increase incentives for violent competition in the longer term. However, new evidence suggests that the effective participation of women in peace processes has increased the likelihood that agreements are reached and maintained over time (UN Women, 2015).

Kenya provides a useful case study of national violent political competition and social violence (Box 2.6).

Box 2.6. Case study on national violent political competition and social violence in Kenya

Kenya illustrates the phenomenon of multiple, interrelated and networked forms of violence. Kenyans overwhelmingly approved a new constitution in 2010, paving the way for political devolution to redress regional inequalities and historic marginalisation as well as to, ultimately, reduce violent competition over the presidency. The country’s post-election violence in late 2007 and early 2008, which brought it to the brink of civil war, influenced the design of the new constitution. However, violence has continued to flare up in many areas of the country.

Spatial variation in subnational politics. Broadly, there are four main types of violence in Kenya: national political violence following ethno-regional splits; subnational political violence and competition around county-level political offices; political violence associated with transnational Islamist groups; and social violence manifest in endemic crime and interpersonal violence that disproportionately affects poor urban neighbourhoods. The map in Figure 2.7 illustrates that communal violence, typically associated with mobilisation along ethno-regional identities, is particularly prevalent in the far north, west and central regions of the country. It also shows that political militia violence is highly centralised, with sporadic violence along the coast. Rebel violence, associated with a violent Islamist insurgency, is primarily concentrated in the eastern region, the coastal areas and the Somali border area. Rioters and protesters are also clustered in the Nairobi area and in central and western parts of the country. State forces’ activity is dispersed throughout the country, but with a markedly lower profile in the far northwest and rural inland areas.

National-level political violence centres on competition to control the presidency. It corresponds to Kenya’s election cycles, peaking in the build-up to national elections that are held every five years. In the 1990s, officials affiliated with the former ruling party played on long-existing ethno-regional divisionsto orchestrateviolencedesigned to intimidatepolitical opponents. A presidential commission of inquiry established to investigate so-called “tribal clashes” in the lead-up to the 1992 and 1997 elections, the Akiwumi Commission of Inquiry into Tribal Clashes (Akiwumi Commission), detailed how the government helped provoke ethnic violence for political purposes and failed to prevent it from escalating into wider conflict (JCITCK, 1999). These ethno-regional tensions came to the fore in Kenya’s 2007-08 post-election violence, in which rival coalitions mobilised support from ethnically-rooted political constituencies. The widespread violence over the contested presidential result resulted in more than 1 000 deaths, the displacement of over 500 000 civilians and widespread sexual violence. According to a national, population-based, cluster survey of adults conducted in 2011, compared to pre-election, sexual violence incidents increased over 50-fold during that post-election violence period (to 1 671.8 incidents per 100 000 population from 33.3 per 100 000 population). The incidents included a sharp increase in intimate partner sexual violence, opportunistic sexual violence and, above all, politically motivated sexual violence designed to humiliate, terrorise and break ties to rival groups (Johnson et al., 2014).

Much of the post-election violence was premeditated and mobilised by political and community leaders. Kenyan police were also implicated in approximately 40% of civilian deaths (ICRtoP, n.d.). Evidence from the Commission to Investigate Post-Election Violence (Waki Commission) suggests that at least 26 % of women raped during post-election violence were raped by police officers (HRW, 2011).

Figure 2.7. Violent events by type of actor in Kenya, 2014

Source: Conflict and violence data, ACLED (2015b); crime data, Kenya Police Service (2014); population data, Kenya Bureau of National Statistics (2009).

Political violence is rooted in competition over subnational political offices, and access to economic rents associated with these offices. Kenya’s constitutional reforms in 2010 laid the groundwork for political devolution. Elections in March 2013 for new county governors, deputy governors and assembly positions then paved the way for devolved administration and governance. The political reforms were meant in part to reduce the potential for political violence concentrated around control of the presidency. However, violence has since flared in the northern and coastal areas of Kenya. The number of conflict events in northern Kenya in 2013 (176) was nearly three times the number recorded (63) for 2007 (ACLED, 2015b). Spikes of high-intensity violence in recent years have further reflected ongoing instability associated with subnational political competition. For example, in August 2012, 118 people were killed and more than 6 000 displaced by violent clashes between Pokomo farmers and Orma herders in Tana River County (HRW, 2012). It was the worst violence to affect Kenya since the 2007-08 post-election violence.

Violence from a localised, transnational insurgency is increasing. Attacks by Al-Shabaab, a Somalia-based militant Islamist organisation, and affiliated groups in Kenya have increased since Kenyan military personnel were deployed to southern Somalia in late 2011. They have killed hundreds of people, mainly in the northeast of the country and coastal areas. The April 2015 siege of the Garissa University College campus, when Al-Shabaab gunmen killed 148 students and faculty members, was Kenya’s deadliest attack since the 1998 bombing of the US Embassy in Nairobi. Al-Shabaab has shown itself adept at stoking deeply entrenched grievances among Kenya’s ethnic Somali, Muslim and coastal populations, in effect localising its transnational conflict in Kenya.

High levels of urban and social violence manifest in endemic criminal and interpersonal violence. Much of the violence in Nairobi is concentrated in its densely populated informal settlements where an estimated 60% of the city’s poorest people live on just over 8% of its land area (UNEP, 2009). Up to two-thirds of the population of these settlements and slums report that they do not feel safe in their neighbourhoods (Goodfellow and Taylor, 2009). The city’s wealthier enclaves employ as many as 100 000 private security guards; in the poor areas criminal organisations and vigilante groups purporting to offer protection have become commonplace (Ruteere et al., 2013).

State responses to violence take place in a context in which violence itself has become political currency. Different forms of violence interact in a self-reinforcing chain (Omenya and Lubaale, 2012), meaning that responses to a particular form of violence may be undermined by the persistent use of violence as a negotiation and competition strategy in the wider political environment. This is evident in proliferating subnational political violence that could lead to a worrying mobilisation of militias and gangs by political actors in the lead up to the 2017 elections. While Kenya has enacted laws related to sexual violence and protection orders in cases of domestic violence, incidents are rarely reported or investigated. The Waki Commission received evidence that 82% of victims of sexual violence did not file reports with the police. Thirty-two per cent of these victims cited as a reason that the police themselves were the perpetrators (CIPEV, 2008). Indeed, Human Rights Watch reports that perpetrators have only been prosecuted in a handful of the killings committed during the violence, and that victims of rape, assault, arson and other crimes also await justice (HRW, 2011)

Sources: ACLED (2015b); CIPEV (2008); HRW (2011, 2012); ICRtoP (n.d.); JCITCK (1999); Johnson et al. (2014); Kenya Police Service (2014); Kenya Bureau of National Statistics (2009); Omenya and Lubaale (2012); Goodfellow and Taylor (2009); Ruteere et al. (2013), UNEP (2009).

Finding Six: Civilians, and especially civilian women and children, are most at risk

In both political and social forms of violence, civilians are most at risk.10 Today 30% to 40% of political violence within states is directed against civilians (ACLED, 2015b). In political armed violence, militias are particularly likely to target civilians. These armed groups are responsible for the majority of fatal attacks on civilians (von Einsiedel et al., 2014). This violence is most acute during periods of domestic political instability

A range of politically motivated actors target civilians with violence. Domestic political instability is often dominated by militias, which seek to renegotiate or change the distribution of political power across elites but not entirely supplant the sitting regime. These differences produce divergent patterns in the targeting and nature of violence across different groups. Labelling a conflict as “ethnic”, “religious” or “resource-based” often diminishes the complexity of these collective groups. It also obscures important similarities and differences across agent type. For example, agent type affects how likely a group is to engage with state forces, target civilians or clash with other non-state groups. Across Africa and select Asian states in 2015, rebel forces typically engaged with the state in over half of their attributed violence, and targeted civilians in just over 20% of actions. In analysing patterns of violence against civilians comparatively, more than half (50.2%) of all anti-civilian violence recorded in 2015 was attributed to political militias, and just 16% of events attributed to rebel groups (ACLED, 2015b).

However, the incidence of interpersonal violence against civilians in many non-conflict contexts can be higher than in countries at war. For example, trade and turf wars between criminal gangs in cities in Brazil, El Salvador, Jamaica and South Africa result in a disproportionately high number of civilian fatalities among the high annual death toll (Banfield, 2014).

Weapons of armed violence designed to cause higher civilian casualties are increasingly available. Improvised explosive devices (IEDs) have caused more civilian deaths and injuries than any other type of explosive weapon in the last five years. In 2015, 43 786 people were killed or injured by IEDs; of those 33 307, or 76%, were civilians (Figure 2.8). This percentage rises in populated urban areas, where 92% of those killed or injured by IEDs are civilians. IED attacks are also becoming more lethal. The number of IED incidents decreased by 20% in 2015 (as compared with 2014), meaning fewer IEDs are causing more lethal casualties (AOAV, 2016). These weapons are used in 64 contexts, although their most prevalent use is in Afghanistan, Iraq, Nigeria, Pakistan and Syria. For example, in Nigeria’s violent landscape, many more civilians are killed and injured by IEDs than armed actors, and the numbers are growing (AOAV, 2016) (Figure 2.9).

Figure 2.8. People killed or injured by improvised explosive devices, 2015

Source: AOAV (2016).

Figure 2.9. Improvised explosive devices deaths and injuries by year in Nigeria, 2011-15

Source: AOAV (2015).

As these types of violence grow more lethal, they disproportionately affect women, youth and children. Adult and young men are predominantly both perpetrators of violence and its main victims. But among civilians, women and children in particular are disproportionately affected by both direct and indirect consequences. The effects of violence on women are different, which must be taken into account. More than half of all global homicide victims are under 30 years old. Much of this violence takes place in urban areas (UNODC, 2013) where poverty and marginalisation are concentrated, and more means for exploitation and violence exist.

Every five minutes, somewhere in the world, a child dies as a result of violence.

Among the most insidious forms of violence is that committed against children. Children experience multiple forms of violence inside and outside the home, with dire consequences. Every five minutes, somewhere in the world, a child dies as a result of violence (UNICEF, 2016a). According to one estimate, at least 275 million children worldwide are exposed to violence in the home with resulting physical injuries, disability or premature death (UNICEF, 2006). Research shows as many as half of all sexual assaults globally are perpetrated against girls under the age of 16 (UNFPA/UNICEF, 2011). Almost one-quarter of 15-19 year-olds have been victims of physical violence (UNICEF, 2016a).

Children bear the longest lasting, and often most severe, consequences of violence. In addition to injuries that affect millions of children and youth directly each year, murder ranks as the fourth leading cause of death among youth globally. Approximately 200 000 young people aged 10-29 years old are victims of homicide (WHO, 2015). The mental and physical health of children exposed to violence and exploitation (at home, school or work or in the community) suffers, and that deeply impacts their overall well-being and future opportunities (Box 2.7). Research shows that children who have been subjected to violence are more likely to become violent themselves (Spano, Rivera and Bolland, 2010). For these reasons, young people also hold the key to ameliorating the inter-generational effects of the same chronic violence that they witness or perpetrate (Batmanglich, 2015).

Box 2.7. The new mental health crisis among displaced Syrian children

More than half of Syrians displaced by the violent upheaval in their country are children, many experiencing mental trauma with long-lasting effects. The United Nations High Commissioner for Refugees (UNHCR) has found that among the anxiety disorders caused by exposure to violence, post-traumatic stress disorder was most common in children and affects more than 50% of them. A study for International Medical Corps found 54% of the displaced had severe emotional disorders and 27% of these children faced developmental challenges.

Source: Karasapan (2016); Save the Children (2014); UNHCR (2015a); Weissebecker and Leichner (2015).

Protracted conflicts have a profound impact on children. The escalation of conflicts in Afghanistan, Somalia, South Sudan and Syria has been especially devastating to children (UN General Assembly, 2016). In Yemen child recruitment is becoming a problem; children are often the victims when schools and hospitals are attacked. Children also are increasingly being recruited as suicide bombers. The Iraqi Independent Commission for Human Rights, for example, found that in the six months from November 2014 to May 2015, extremists trained 1 000 Iraqi children to be suicide bombers (AOAV, 2015; UNICEF, 2016b). A similar trend is seen in Syria where over half the verified cases of children recruited in 2015 were under 15 years old, some as young as 7 (UNICEF, 2016c).

Intimate partner violence is estimated to cost USD 4.4 trillion or, 5.18% of world GDP – more than political violence and interpersonal violence (homicides) combined.

Sexual and gender-based violence (SGBV), and especially intimate partner violence, is extremely prevalent across conflict, non-conflict, low-and high-income contexts. It poses an immense burden across conflict-affected contexts and societies considered “peaceful”. Globally, an estimated 35% of women have experienced either physical and/or sexual intimate partner violence or non-partner sexual violence in their lifetime (WHO, 2016). Regarding intimate partners, little change in prevalence has been observed over time and regions, although it is not reliably tracked. In countries with low rates of female homicide, intimate partners make up the majority of perpetrators, in some cases over 60% (Geneva Declaration Secretariat, 2015). Intimate partner violence is estimated to cost USD 4.4 trillion, or 5.18% of world GDP – more than political violence and interpersonal violence (i.e. homicides) combined. Interpersonal violence is estimated to cost USD 1.245 trillion, or 1.44% of world GDP (Hoeffler and Fearon, 2014).

The impact of violence on civilian women is particularly striking in Latin America and the Caribbean (LAC). The Geneva Declaration Secretariat (2015) reported that five of the seven countries with “very high” rates of homicide against women are in LAC. Most violence against women in the region is at the hands of intimate partners, and is often non-lethal but sustained over long periods of time. A comparative analysis of data from 12 countries in LAC found that 17% to 53% of women, depending on the country, experienced intimate partner sexual and physical violence. Key risk factors include controlling behaviour by the partner, alcohol consumption and experiences of domestic violence in childhood, meaning that violence begets violence (Bott et al., 2012) The Inter-American Development Bank (IADB) estimates that in countries throughout LAC, violence against women is associated with a loss of between 1.6% and 3.7% of GDP (IADB, 2016).

While sexual and gender-based violence is prevalent across “peaceful” societies, political violence can further aggravate it. Sexual violence can be employed as a weapon of war – a deliberate strategy by armed groups to torture and humiliate opponents; terrify individuals; destroy societies; incite flight from a territory; and reaffirm aggression, brutality and domination (Bastick, Grimm and Kunz, 2007). Gender norms also drive SGBV and contribute to other forms of violence and conflict (Saferworld, 2014). Although men comprise most of homicide victims (by predominantly male perpetrators), women are the majority of victims of intimate partner homicide. By highlighting these dynamics, a gender perspective points to the connections between political and social violence, between violence in times of peace and conflict, and between violence at all levels of society.

Over half of the 21.3 million refugees in the world today are children under the age of 18. In 2015, unaccompanied or separated children, many fleeing violent conflict, submitted 98 400 asylum requests – the largest number ever recorded (UN General Assembly, 2016). Displaced children are vulnerable to sexual violence, forced marriage, abduction and human trafficking (UN General Assembly, 2016). Every day children around the world are exposed to some form of violence with long-term and devastating consequences, affecting their physical and mental health, their education and their overall well-being. The effects are lasting for the child and the family. It may take generations for the family to recover.

Youth violence is driven by an interplay of risk factors most prevalent in fragile communities with the highest concentrations of poverty. Those factors include the presence of local trade in small arms and drugs, organised violent groups (e.g. gangs), high unemployment rates, and economic and political marginalisation (WHO, 2015). Youth recruitment into criminal, terrorist or armed political groups tends to be based on similar factors. It is more often a result of unemployment, self-protection, lack of respect or idleness (World Bank, 2011a) than ideology, belief in a cause or revenge. For example, research by the UN Assistance Mission in Somalia (UNSOM) revealed religion and ideology are not the main drivers for recruits to Al-Shabab in Somalia; rather it is the promise of a good salary (Anyadike, 2016). As a livelihood coping strategy, this can be directly linked to poverty but also to other forms marginalisation (Figure 2.10).

Figure 2.10. Youth violence risk factors

Source: US Centers for Disease Control and Prevention (2015).

Violent extremism and radicalisation are attributed to a combination of factors. On the one hand, there are “push” factors including unemployment, poverty, clan/social/political marginalisation, corruption and youth frustration. On the other hand, “pull” factors appeal to the individual and include access to material resources, weapons and protection, a sense of belonging and empowerment and strong governance (Glazzard et al., 2016).

Finding Seven: Urban violence is becoming the new norm

The rise in urban violence will continue. Over the past 20 years, rural violence has been declining while violence is becoming increasingly concentrated in urban areas (Raleigh, 2015). Governance, security and services have not kept pace with rapid population growth in urban areas, providing opportunity, means and space for non-state actors to vie for political power and patronage. Criminal networks can then take root more easily, enabling illicit flows of arms or drugs. Concentrated pockets of violence can appear as youth criminal gangs or militias emerge. Together, they can cause localised social discord, increased interpersonal violence, heightened risk of youth recruitment, and deteriorating overall safety and security.

Various forms of violence most strikingly collide in cities, creating a negative cycle of mutually reinforcing factors that pose the greatest risks to civilians. Political violence is more prevalent where it coincides with poverty, inequality and poor rule of law, and where its implications and risks can multiply. For example, densely populated urban areas in Africa experience almost twice the rate of political violence than rural areas and other towns. However, when factoring in social, interpersonal and criminal-related violence, 45 of the top 50 most violent cities are actually found in North, Central and South America (Muggah, 2015).

The biggest cities in fragile contexts are likely to be the world’s most vulnerable (de Boer, 2015). But the world’s biggest cities are not necessarily the most fragile (Igarapé Institute, 2016). Many cities with rapid growth rates may be at higher risk than the 35 “megacities” and 600 “large cities”. Research finds that properties of risk and resilience are present in all cities, irrespective of location; this intensifies when risks and stresses accumulate (Igarapé Institute, 2016). A map of the world’s major cities is shown in Figure 2.11.

Figure 2.11. Global city distribution, populations over 500 000, 2015

Source: Satterthwaite and McGranahan (2014); International Institute for Environment and Development (2016); IIED interactive at

Violence is especially visible in cities where a lack of growth, security and welfare is associated with higher levels of conflict and violence (Muggah, 2015; Vidal, 2015). Rapid and unregulated urbanisation, income and social inequality, concentrated poverty, youth unemployment, policing and justice deficits, and real and perceived insecurity are all factors (Muggah, 2015). Countries in Latin America and the Caribbean are among the most urbanised in the world, with more than 80% of the region’s population living in cities. This contributes to the region’s high homicide rates. The vast majority of homicides globally are concentrated in fast-growing mid- and large-sized cities of the Americas.

Individual and collective forms of violence co-exist in urban areas. This is most evident where the capacity gaps in providing basic and accountable security services are a key determinant in shaping urban violence. For example, Soares and Naritomi (2010) observed the influence of low incarceration rates, among other factors, in shaping “cultures of impunity”. Riots and protests also overwhelmingly occur in urban contexts (ACLED, 2015b), where population, power and wealth are most concentrated and municipal authorities are often ill-equipped to cope. In addition, states and cities that have failed to reform their security sectors after conflicts are exposed to complex forms of violence.

Urbanisation can act as a trigger for violence. Economic development brings rural poor to cities where they often live in slums. In the Sustainable Development Goal period a huge demographic shift will occur towards urban areas. The most populated cities in the world are also likely to be where structural inequalities and social exclusion become more apparent.

Within cities, violence is unevenly distributed, and particularly acute in lower-income informal areas (Muggah, 2012). In Bogota, for example, roughly 98% of all homicides occur in less than 2% of street addresses (Igarapé Institute, 2015). By contrast, the middle and upper classes more commonly experience abductions and disappearances, even though they often resort to private security for self-protection (Alvarado and Santiso, 2015).

Economic, political and social violence also occur in large towns and on the edges of cities. Upwardly mobile populations and marginalised social groups, breaking free of village hierarchies and gaining access to urban and peri-urban employment, often come into conflict with traditional urban elites whose dominance is threatened. In India, for example, riots and protests account for over 75% of the country’s violence (ACLED, 2016). Incidents of social instability are geographically widespread, but are especially prevalent in Jammu and Kashmir, as well as Gujarat, Punjab, Uttar Pradesh and West Bengal. A significant proportion of these incidents occur in cities with a population of larger than 1 million people.

Politically violent actors take advantage of geography for attacks, retribution and recruitment. Political violence tends to cluster in strategic and target areas where opposing forces can openly contest each other, such as large towns and cities and areas with high road mass and dense populations (Raleigh and Hegre, 2009). Cities have large, potentially aggrieved populations available to participate in and potentially support conflict (Goldstone, 2002; Urdal and Hoeschler, 2012). Peripheral and distant areas are more likely to experience the emergence of insurgent and separatist actions (Fearon and Laitin, 2003). Armed battles are also more common in rural areas, while urban areas see more riots and protests (Figure 2.12). New spaces of violence are the result of new agendas, actors (e.g. gangs and militias) and organised responses to violence (Rodgers, 2009).

Figure 2.12. Political violence and protest data by urban and rural location

Source: ACLED (2015b).

Several countries most affected by conflict will see rapid population growth in their cities before 2030. According to some projections, the urban populations of Bangladesh, the Democratic Republic of the Congo, Ethiopia and Pakistan will grow by up to 50 million by 2030 (UN DESA, 2014). Traditional aid approaches focused on armed conflict at the national level need to realign accordingly. In light of these stark assessments, it appears likely that future crises in the SDG era will be more likely to occur in cities.

Finding Eight: Violent extremism and terrorism are on the rise

Forms of terrorism or extremist acts against civilians are increasingly common features of violent conflict. Terrorism occurs far more often within established conflicts and as such, these are where most terrorism victims are found. Terrorism-related deaths rose by 61% in 2013, encompassing 18 000 victims of terrorist attacks globally; most victims were in Afghanistan, Iraq, Nigeria, Pakistan and Syria (IEP, 2015). Women and girls are particular victims of this form of violence: religious extremist groups attack the rights of women and girls, and have been associated with notable increases in forced marriage, restrictions on education and participation in public life, and systematic sexual and gender-based violence (UN Women, 2015).

The global scale of Islamist extremism is also of concern. The outbreak of violent extremist networks claiming Islamist ideology has led to politically motivated and criminal violence across a vast region of territory in the Sahel region of Africa, North Africa and the Middle East in particular. In Iraq and Syria, as well as in Mali, northern Nigeria, Somalia and the Sahel, religious extremists are sowing widespread violence, frequently using terror tactics. Although the underlying drivers of violent conflict may not be new, this ideology-based violence reflects recent shifts and adaptations (Dowd, 2015). These extremist groups have been able to mobilise large numbers of fighters under a transnational Islamist identity rather than a narrower communal, national or ethnic identity. This, in turn, has improved their ability to attract financial, logistical and troop support for global violent actions. Claiming opportunities to create alternative governance structures under Islamist rule in collapsing or profoundly fragile contexts has also been a useful mobilisation strategy.

Violence also stems from other forms of extremist ideology. Extremist groups come in many forms, whether nationalist, right wing or left wing, religious, racist or based in other ideology. They can all spread and promote hatred, and may contribute to violence by condoning it or directly advocating it. The UN High Commissioner for Human Rights, citing lessons of history, warned in September 2016 of the danger of xenophobia and bigotry becoming “weaponized” as “bigots, populists and political fantasists” use “half-truths and oversimplification to spread toxic hatred” (OHCHR, 2016).

Terror is increasingly a tactic to further a range of political aims and not a separate form of violence. While terrorism is often considered a separate and distinct form of political violence, it is more accurately understood as a tactic employed within and across different forms of conflict. Any armed, organised agents can engage in the use of terror to mount large-scale attacks on civilians. Terror tactics are commonly used in contexts where violence is already present. In spaces and periods of domestic political instability, agents employ terror tactics against local populations; in civil and transnational wars, the use of terror tactics against transnational or global targets and populations is common.

Extremist groups such as Boko Haram and the so-called Islamic State rebrand pre-existing local conflicts or grievances as global causes. The reach of extremist groups is expanding, but there is no global consensus on whether or how to engage these groups in negotiations about peaceful means of addressing grievances.

Finding Nine: The cost of violence is rising

Beyond mortality and physical injury, the effects of violence are far-reaching and can have devastating impact (Geneva Declaration Secretariat, 2015). Its costs also continue to manifest over time. The immediate consequences are mortality, injury and displacement. In the medium term violence affects the health and viability of individuals and their productive capacity. Over the longer term it throws up obstacles to social and economic development and overall well-being. Moreover, the psychological wellness of individuals and households is severely compromised with exposure to violence, as vulnerabilities are transferred to the broader society. Violence threatens the resilience capacity of a society by sapping social capital (e.g. through forced displacement) and economic capital (e.g. financial losses and security costs, damage to infrastructure). Although violence may disproportionately affect fragile contexts with high poverty rates, its high costs are also borne by middle- and high-income contexts not considered fragile.

However, poverty does have a relationship with conflict. Some evidence shows that violent conflict is most likely to break out in low-income areas, even if the state overall is not particularly poor, and in areas that diverge significantly from national income averages (Buhaug et al., 2011). Poorer areas are more likely to witness increased conflict-related deaths (Do and Iyer, 2010); and higher levels of poverty lead to higher rates of homicide (Chon, 2011; Pridemore, 2008; Moser and McIlwaine, 2006). Some studies have found a close relationship between certain forms of conflict (for example, lower-intensity conflict) and poverty, for example in Indonesia (Pradhan, Barron and Kaiser , 2004); as well as variation in the significance of poverty at various stages of conflict (Sánchez and Chacón, 2005, Murshed and Gates, 2004).

Economic opportunities and individual-level inequality show a stronger relationship in incentivising social violence participation. Evidence from Central and South America suggests that crime and theft are driven by economic motivations (Moser and Winton, 2002), and that high levels of grievances linked to economic opportunities lead to increased gang violence (Sizemore, 2014). There is also a higher likelihood that individual-level inequality is related to social violence or violent crime (Chon, 2011; Fajnzlyber, Lederman and Loayza, 2000, 2002).

The global cost of violence is staggering. The Institute of Economics and Peace (IEP), which provides a global aggregate of the costs of violence containment, reports the global economic impact of violence was a daunting USD 13.6 trillion in 2015, equivalent to 13.3% of global GDP or USD 1 876 for every person in the world (IEP, 2016) (Figure 2.13). In its World Development Report 2011, the World Bank found that “poverty reduction in countries affected by major violence is on average nearly a percentage point lower per year than in countries not affected by violence” (World Bank, 2011b). It has been estimated that the global cost of homicide in 2010 was USD 171 billion and that roughly USD 2 trillion in global violence-related economic losses could have been saved had homicides rates dropped to what are termed “normal” levels (Geneva Declaration Secretariat, 2015).

Figure 2.13. Economic cost of violence

Source: IEP (forthcoming).

Individual social forms of violence cost the world more than collective political violence. Writing for the Copenhagen Consensus, Hoeffler and Fearon (2014) estimated interpersonal violence (e.g. homicides) at USD 1.245 trillion; intimate partner violence USD 4.4 trillion; and reported sexual violence against women as USD 66.7 billion. Strikingly, intimate partner violence alone carries a greater global cost than collective political violence and homicides combined (Hoeffler and Fearon, 2014). This evidence makes a clear case for focusing on prevention, and on intimate partner violence in particular, as a development priority. Development policies and violence intervention measures without a gender component will ultimately fail to make any meaningful differences in the incidence and consequences of this costliest form of violence.

Yet development assistance invests only marginally in violence reduction outside of conflict. Programmes aimed at preventing conflict and building and sustaining peace are chronically under-prioritised and underfunded (Hoeffler and Fearon, 2014). Social violence faces even greater gaps, in part because, as Cockayne (2013) describes it, “its heaviest costs fall on those furthest from the outsiders’ gaze: the children who are killed working corners for drug gangs, and the wives and women terrorized by the violent male criminal culture”.

The long-term cost of violence far exceeds the estimated cost of prevention. Violence is costly – in terms of conflict-related GDP loss; military spending; aid to displaced people and refugees; private security for businesses and individuals; and spending on law enforcement and internal security. A recent study estimates the consolidated costs associated with various forms of violence against children, for example, to be as high as USD 7 trillion (Pereznieto et al., 2014). This startling figure reveals the magnitude of the toll incurred by children due to physical, psychological and sexual violence.11 The annual global costs from children’s association with armed forces or groups can be up to USD 144 million (Pereznieto et al., 2014). These costs far exceed the estimated investments believed necessary for the prevention of that violence. If global violence were to decrease by 10% uniformly, an additional USD 1.43 trillion would effectively be incorporated into the world economy each year. This is more than six times the total value of Greece’s bailout and loans (IEP, 2015).

Box 2.8 shows some of the key numbers involved in violence and conflict.

Box 2.8. By the numbers: Violence and conflict

Finding Ten: Violence is a behaviour reinforced by social norms which acts like a contagion

Violence is an epidemic and is contagious. Research reveals that violence behaves like an epidemic, sharing the same characteristics of clustering, spread and transmission (Slutkin et al., 2015). Violence clusters occur in “hot spots” where people have been exposed to violence – just as cholera typically clusters around water sources where people are exposed to bacteria that cause the disease. It can mimic epidemic spread across time or geographically across space, and has a transmission mechanism in which exposure correlates to risk: those exposed to violence are at increased risk of perpetuating it themselves (Spano, Rivera and Bolland, 2010).

Recurring cycles of violence persist because of regular exposure to violence. Violence of all types increases immediately following conflict when community violence and spouse and child abuse are higher (Dubow, Huesmann and Boxer, 2009). Children who are exposed to violence are more likely to engage in violence than children who have not been exposed, and the likelihood of involvement in violence increases when exposed to higher amounts of violence (i.e. “dose dependence”) (Spano, Rivera and Bolland, 2010). Because all behaviour is contagious, exposure to violence leads to unconscious modelling and adoption of like behaviour (DuRant et al., 1994; Kelly 2010). Repeated over multiple exposures and viewed collectively, this raises the risk. At the same time it creates social norms in which social pressure and a permissive environment condone, and even encourage, violent responses to even minor disagreements, particularly in areas of chronic violence (Slutkin et al., 2015). These social norms reinforce this contagion by encouraging violent behaviour to spread (Cure Violence, 2016).

Drawing insights and lessons from the health approach may be useful in guiding violence prevention efforts in conflict and non-conflict settings. Strategies associated with disease control have yielded dramatic results, detecting and “interrupting” conflicts, identifying and treating the highest risk individuals, and changing social norms in violence hotspots in North, Central and South America (Slutkin, 2015). Independent evaluations have demonstrated reductions in shootings and killings in a range of 41% to 73%. These contributed to gradual shifts in social norms and expectations (Skogan et al., 2009). The approach to violence programmes can benefit from applying the knowledge that relationships between perpetrator and victim can shift, and that violence is a learned behaviour reinforced by norms. Where perpetrators are the products of a learned undesirable behaviour in contexts where violence may have come to be accepted as “normal”, it can be effective to shift responses to treat the causes and means of its transmission, rather than its manifestations and symptoms.

Understanding violence in epidemiological terms may unlock new insights and solutions. By using the same approach that the WHO uses to stop epidemics – interrupt transmission, change behaviour, change norms – policy makers can gain more traction upstream before the violence is able to manifest. This breaks the recurrent cycle of violence, stopping its “transmission”. Treating violence with a health approach shifts the optimal point of intervention where prevention is viable and enables mitigation efforts to intercept the contagion immediately (Figure 2.14). This represents a paradigm shift in thinking about violence as an inevitable condition; as only a domestic law enforcement, political or international security problem; or as related only to underlying causes or under-development. The strategy is relevant for political or social violence contexts where community and individual resilience to violence is made possible.

Figure 2.14. Violence in disease – a control model

Source: Adapted from Cure Violence (n.d).


ACLED (2016), “Armed Conflict and Location Event Data Project”, Version 6 data,

ACLED (2015a), “Definitions of political violence, agents and event types”,

ACLED (2015b), “Armed Conflict and Location Event Data Project”, Version 5 data,

Adams, T. (2012), “‘Chronic violence’: Toward a new approach to 21st-century violence”, NOREF Policy Brief, p. 2, Norwegian Peacebuilding Resource Centre,

AOAV (2016), Unacceptable Harm: Monitoring Explosive Violence in 2015, Action on Armed Violence, London,

AOAV (2015), “2015: An epidemic of suicide bombs”, Action on Armed Violence, London,

Alvarado, N. and C. Santiso (2015), “Insécurité endémique en Amérique latine”, Le Monde Diplomatique, No. 735, p. 13.

Amar, P. (2013), The Security Archipelago: Human-Security States, Sexuality Politics, and the End of Neoliberalism, Duke University Press, Durham, NC.

Anyadike, O. (2016), “One man’s terrorist is another man’s carpenter: How DDR can shrink support for al-Shabab in Somalia”, IRIN News, 21 July 2016,

Arriola, L.R. (2009), “Patronage and political stability in Africa”, Comparative Political Studies, Vol. 42/10, pp. 1339-1362.

Arriola, L.R. and C. Johnson (2012), “Electoral violence in democratizing states”, draft paper, University of California, Berkeley, CA,

Banfield, J. (2014), Crime and Conflict: The New Challenge for Peacebuilding, International Alert, London,

Bangura, Y. (2006), Ethnic Inequalities and Public Sector Governance, Palgrave MacMillan, Basingstoke, UK.

Bastick, M., K. Grimm and R. Kunz (2007), Sexual Violence in Armed Conflict: Global Overview and Implications for the Security Sector, Geneva Centre for the Democratic Control of Armed Forces (DCAF),

Batmanglich, S. (2015), “Mapping existing thinking on criminality, urban violence and subnational-conflict”, INCAF Discussion Paper, OECD, Paris.

Bierschenk, T. and J.-P. Olivier de Sardan (1997), “Local powers and a distant state in rural Central African Republic”, The Journal of Modern African Studies, Vol. 35/3, pp. 441-468.

Blattman, C. and E. Miguel (2010), “Civil war”, Journal of Economic Literature, Vol. 48/1.

Bosetti, L., J. Cockayne and J. de Boer (2016), “Crime-proofing conflict prevention, management, and peacebuilding: A review of emerging good practice”, Occasional Paper, No. 6, United Nations University Centre for Policy Research, Tokyo,

Bott, S. et al. (2012), Violence Against Women in Latin America and the Caribbean: A Comparative Analysis of Population-based Data from 12 Countries, Pan American Health Organization and Centers for Disease Control and Prevention, Washington, DC.

Brancati, D. (2011), “Decentralization: Fueling the fire or dampening the flames of ethnic conflict and secessionism?”, International Organization, Vol. 60/3, pp. 651-685.

Buhaug, H. et al. (2011), “It’s the local economy, stupid! Geographic wealth dispersion and conflict outbreak location”, Journal of Conflict Resolution, Vol. 55/5, pp. 814-840.

Cakaj, L. (2015), “In unclaimed land: The Lord’s Resistance Army in CAR”, in T. Carayannis and L. Lombard (eds.), Making Sense of the Central African Republic, Zed Books, London.

Carayannis, T. et al. (2014), “Practice without evidence: Interrogating conflict resolution approaches and assumptions”, JSRP Paper, No. 11, The Justice and Security Research Programme, London School of Economics and Political Science, London,

Cederman, L.E., A. Wimmer and B. Min (2010), “Why do ethnic groups rebel? New data and analysis”, World Politics, Vol. 62/1, pp. 87-119.

Choi, H. and C. Raleigh (2015), “Dominant forms of conflict in changing political systems”, International Studies Quarterly, Vol. 50/1, pp. 158-171.

Chon, D.S. (2011), “Contributing factors for high homicide rate in Latin America: A critical test of Neapolitan’s regional subculture of violence thesis”, Journal of Family Violence, Vol. 26/4, pp. 299-307.

CIPEV (2008), Report of the Commission of Inquiry on Post-Election Violence,

Clapham, C. (1996), Africa and the International System: The Politics of State Survival, Cambridge University Press, Cambridge.

Cockayne, J. (2013), “Beating the crime trap’: What can we do about the feedback loop between fragility and organized crime?”, Remarks at the launch of the Global Initiative against Transnational Organized Crime, 19 September 2013, UN Headquarters, New York.

Collier, P. and A. Hoeffler (2004), “Greed and grievance in civil war”, Oxford Economic Papers, No. 56,

Cure Violence (2016), “Reducing violence through health methods: Policy implications”, presentation for OECD DAC-INCAF Violence Workshop, 27 June 2016.

Cure Violence (n.d), The Cure Violence Health Model,

Dalby, N. (2015), “A multifaceted business: Diamonds in the Central African Republic”, in T. Carayannis and L. Lombard (eds.), Making Sense of the Central African Republic, Zed Books, London.

De Boer, J. (2015), “Resilience and the fragile city”, United Nations University, Tokyo,

De Boer, J. and L. Bosetti (2015), “The crime-conflict ‘nexus’: State of the evidence”, Occasional Paper, No. 5, United Nations University Centre for Policy Research, Tokyo,

De Martino, L. (2012), “Reducing armed violence, enabling development”, Small Arms Survey Research Notes, No. 19, Graduate Institute of International and Development Studies, Geneva,

De Waal, A. (2014), “The political marketplace: Analyzing political entrepreneurs and political bargaining with a business lens”, memorandum prepared for World Peace Foundation seminar, Fletcher School at Tufts University, Medford, MA,

Deschamps, M., H.B. Jallow and Y. Sooka (2015) Taking Action on Sexual Exploitation and Abuse by Peacekeepers,

Do, Q.-T. and L. Iyer (2010), “Geography, poverty and conflict in Nepal”, Journal of Peace Research, Vol. 47/6, pp. 737-748.

Dowd, C. (2015), “Cultural and religious demography and violent Islamist groups in Africa”, Political Geography, Vol. 45, pp. 11-21.

Dubow, E.F., L.R. Huesmann and P. Boxer (2009), “A social-cognitive-ecological framework for understanding the impact of exposure to persistent ethnic-political violence on children’s psychosocial adjustment”, Clinical Child and Family Psychological Review, Vol. 12/2, pp. 113-126.

Duffy, G. (2015), Country Report: Central African Republic, Armed Conflict Location and Event Data (ACLED),

DuRant, R.H. et al. (1994), “Exposure to violence and victimization and fighting behavior by urban black adolescents”, Journal of Adolescent Health, Vol. 15/4, pp. 311-318.

Ellis, S. (2009), “West Africa’s international drug trade”, African Affairs, Vol. 108/431, pp. 171-196.

Englebert, P. (2000), State Legitimacy and Development in Africa, Lynne Rienner Publishers, London.

Fajnzylber, P., D. Lederman and N. Loayza (2002), “Inequality and violent crime”, Journal of Law and Economics, Vol. 45/11, pp. 1-39.

Fajnzylber, P., D. Lederman and N. Loayza (2000), “Crime and victimization: An economic perspective”, Economia, Vol. 1/1, pp. 219-302.

Fearon, J. and D. Laitin (2003), “Ethnicity, insurgency, and civil war”, American Political Science Review, Vol. 97/1, pp. 75-90.

Fischer, A.M. (2008), “Resolving the theoretical ambiguities of social exclusion with reference to polarisation and conflict”, DESTIN Working Paper, No. 0890, Development Studies Institute, London School of Economics, London,

Francois, P., I. Rainer and F. Trebbi (2014), “The dictator’s inner circle”, NBER Working Paper, No. 20216, National Bureau of Economic Research, Cambridge, MA,

Gandhi, J. and E. Lust-Okar (2009), “Elections under authoritarianism”, Annual Review of Political Science, Vol. 12, pp. 403-422.

Gates, S. et al. (2016), “Trends in armed conflict 1946-2014” PRIO Policy Brief, Conflict Trends, 1, Peace Research Institute, Oslo,

Geneva Declaration Secretariat (2015), Global Burden of Armed Violence 2015: Every Body Counts, Geneva,

Glazzard, A. et al. (2016), “Conflict and countering Islamist violent extremism”, Summary paper, Royal United Services Institute (RUSI) for the UK Department for International Development, London,

Glenn, C. (2016), “Timeline: Rise and spread of the Islamic State”, The Islamists blog, blogpost for The Wilson Center, 5 July 2016, Washington, DC,

Goldstone, J. (2002), “Population and security: How demographic change can lead to violent conflict”, Journal of International Affairs, Vol. 56/1, pp. 3-21.

Goodfellow, T. and W. Taylor (2009), “Urban poverty and vulnerability in Kenya: The urgent need for co-ordinated action to reduce urban poverty”, Oxfam GB Briefing Note, 10 September 2009, Oxfam, London,

Hansen, W. (2014), “The organized crime-peace operations nexus”, PRISM Security Studies Journal, Vol. 5/1.

Hendrix, C.S. (2010), “Measuring state capacity: Theoretical and empirical implications for the study of civil conflict”, Journal of Peace Research, Vol. 47/3, pp. 273-285.

Hoeffler, A. and J. Fearon (2014), “Benefits and costs of the conflict and violence targets for the post-2015 development agenda”, Post-2014 Consensus Working Paper, Copenhagen Consensus Center, Copenhagen,

HRW (2012), “Kenya: Investigate all politicians in Tana River violence”, HRW news, 13 September 2012,

HRW (2011), Human Rights Watch, New York,

HRW (2009), “Central African Republic”, Country summary, Human Rights Watch, New York,

Human Security Centre (2005), Human Security Report 2005: War and Peace in the 21st Century, University of British Columbia, Vancouver, BC, Canada, Oxford University Press, New York/Oxford,

IADB (2016), “Violence against women”, Inter-American Development Bank Diversity blog,,8014.html?isajaxrequest&actionuserstats=close&valcookie.

ICG (2016), “Easy prey: Criminal violence and Central American migration”, Report No. 57, International Crisis Group,

ICRC (2008), “How is the term ‘armed conflict’ defined in international humanitarian law?”, International Committee of the Red Cross Opinion Paper, March 2008,

ICRtoP (n.d.), “The crisis in Kenya”, International Coalition for the Responsibility to Protect, New York,

IDMC (2016), 2016 Global Report on Internal Displacement, Internal Displacement Monitoring Centre, Geneva,

IEP (forthcoming) Economic Value of Peace, Institute for Economics and Peace

IEP (2016), Global Peace Index 2016: Ten Years of Measuring Peace, Institute for Economics and Peace, Sydney/New York.

IEP (2015),

Igarapé Institue (2016), “Fragile cities”, data visualisation,

Igarapé Institute (2015), “Homicide monitor”, interactive website,

IISS (2016), Armed Conflict Survey 2016, International Institute for Strategic Studies,

IISS (2015), Armed Conflict Survey 2015, International Institute for Strategic Studies,

International Institute for Environment and Development (2016), Cities: An Interactive Visual Tool,

IRC (2014), “Sexual violence biggest fear for women in Central African Republic, says International Rescue Committee”, Press Release,

JCITCK (1999), Judicial Commission of Inquiry into Tribal Clashes in Kenya,

Johansson-Nogués, E. (2013), “Gendering the Arab Spring? Rights and (in)security of Tunisian, Egyptian and Libyan women”, Security Dialogue, Vol. 44/5-6, pp. 393–409.

Johnson, K. et al. (2014), “A national population-based assessment of 2007–2008 election-related violence in Kenya”, Conflict and Health, Vol. 8/2, pp. 125.

Justino, P. (2012), “Nutrition, governance and violence: A framework for the analysis of resilience and vulnerability to food insecurity in contexts of violent conflict”, HiCN Working Papers, No. 132, Households in Conflict Network, University of Sussex, Brighton, UK.

Justino, P., T. Brück and P. Verwimp (2013), “Research and policy implications from a micro-level perspective on the dynamics of conflict, violence and development”, HiCN Working Papers, No. 139, Households in Conflict Network, University of Sussex, Brighton, UK.

Kalyvas, S.N. (2008), “Promises and pitfalls of an emerging research program: The microdynamics of civil war”, in S.N. Kalyvas, I. Shapiro and T. Masoud (eds.), Order, Conflict, Violence, Cambridge University Press, Cambridge, UK,

Kalyvas, S.N. (2003), “The ontology of ‘political violence’: Action and identity in civil wars”, Perspective on Politics, Vol. 1/3, pp. 475-494.

Karasapan, O. (2016), “Syria’s mental health crisis”, Future Development blog, The Brookings Institution, 25 April 2016,

Kasara, K. (2007), “Tax me if you can: Ethnic geography, democracy and the taxation of agriculture in Africa”, American Political Science Review, Vol. 101/1, pp. 159-172.

Kelly, S. (2010), “The psychological consequences to adolescents of exposure to gang violence in the community: An integrated review of the literature”, Journal of Child and Adolescent Psychiatric Nursing, Vol. 23/2, pp. 61-73.

Kenya Bureau of National Statistics (2009), Population and Housing Census 2009, Government of Kenya, Nairobi,

Kenya Commission of Inquiry into Post-Election Violence (2008), Report of the Commission,

Kenya Police Service (2014), Annual Crime Report 2014,

Kilembe, F. (2015), “Local dynamics in the PK5 district of Bangui”, in T. Carayannis and L. Lombard (eds.), Making Sense of the Central African Republic, Zed Books, London.

La Porta, R. et al. (1999), “The quality of government”, Journal of Law, Economics and Organization, Vol. 15/1, pp. 222-279.

Lindemann, S. (2008), “Do inclusive elite bargains matter? A research framework for understanding the causes of civil war in sub-Saharan Africa”, Crisis States Discussion Paper, No. 15, Development Studies Institute, London School of Economics, London,

Marchal, R. (2015), “CAR and the regional (dis)order”, in T. Carayannis and L. Lombard (eds.), Making Sense of the Central African Republic, Zed Books, London.

Melander, E. (2015), “Organized violence in the world: An assessment by the Uppsala Conflict Data Program”, UCDP Paper, No. 9, Uppsala University, Uppsala, Sweden,

Moser, C. and C. McIlwaine (2006), “Latin American urban violence as a development concern: Towards a framework for violence reduction”, World Development, Vol. 34/1, pp. 89-112.

Moser, C. and A. Winton (2002), “Violence in the Central American region: Towards an integrated framework for violence reduction” ODI Working Paper, No. 171, Overseas Development Institute, London,

Muggah, R. (2016), “Terrorism is on the rise – but there’s a bigger threat we’re not talking about”, World Economic Forum International Security blog, 8 April 2016,

Muggah, R. (2015), “Fixing fragile cities: Solutions for urban violence and poverty”, Foreign Affairs blog, Council on Foreign Relations, 15 January 2015,

Muggah, R. (2012), Researching the Urban Dilemma: Urbanization, Poverty and Violence, IDRC/DFID, Ottawa/London,

Murshed, S.M. and S. Gates (2004), “Spatial horizontal inequality and the Maoist insurgency in Nepal”, Research Paper, No. 2003/43, UNU-WIDER, United Nations University, Tokyo,

Newman, E. (2009), “Conflict research and the ‘decline’ of civil war”, Civil Wars, Vol. 11/3, pp. 255-278.

OECD (forthcoming), Illicit Financial Flows: Illicit Trade and Development Challenges in West Africa, OECD Publishing, Paris, forthcoming.

OHCHR (2016), “Zeid warns against populists and demagogues in Europe and U.S.”, statement to Peace, Justice and Security Foundation gala, The Hague, 5 September 2016,

Olin, N. (2015), “Pathologies of peacekeeping and peacebuilding in CAR”, in T. Carayannis and L. Lombard (eds.), Making Sense of the Central African Republic, Zed Books, London.

Omenya, A. and G. Lubaale (2012), Understanding the Tipping Point of Urban Conflict: Violence, Cities, and Poverty Reduction in the Developing World, Manchester, University of Manchester.

Pereznieto, P. et al. (2014), The Costs and Economic Impact of Violence against Children, Overseas Development Institute, London,

Pradhan, M., P. Barron and K. Kaiser (2004), “Local conflict in Indonesia: Measuring incidence and identifying patterns”, World Bank,

Pridemore, W. (2008), “A methodological addition to the cross-national empirical literature on social structure and homicide: A first test of the poverty-homicide thesis”, Criminology, Vol. 46/1, pp. 133-154.

Raleigh, C. (2015), “Urban violence patterns across African states”, International Studies Review, Vol. 17/1. Raleigh, C. and H. Hegre (2009), “Population size and civil war: A geographically disaggregated analysis”, Political Geography, Vol. 28/4.

Reno, W. (2011), Warfare in Independent Africa, Cambridge University Press, Cambridge.

Rivera, M. (2016), “The sources of social violence in Latin America: An empirical analysis of homicide rates, 1980-2010”, Journal of Peace Research, Vol. 53/1.

Rodgers, D. (2009), “Slum wars of the 21st century: Gangs, mano dura and the new urban geography of conflict in Central America”, Development and Change, Vol. 40/5.

Roitman, J. (2001), “New sovereigns? Regulatory authority in the Chad basin”, in T. Callaghy, R. Kassmir and R. Latham (eds.), Intervention and Transnationalism in Africa Global-Local Networks of Power, Cambridge University Press, Cambridge.

Ruteere, M. et al. (2013), “Missing the point: Violence reduction and policy misadventures in Nairobi’s poor neighbourhoods”, Institute of Development Studies Evidence Report, No. 39,

Saferworld (2014), Masculinities, Conflict and Peacebuilding: Perspectives on Men through a Gender Lens, Saferworld, London,

Sánchez, F. and M. Chacón (2005), “Conflict, state and decentralisation: From social progress to armed dispute for location control, 1974-2002”, Working Papers Series 1, No. 70, Crisis States Research Centre, London School of Economics and Political Science, London,

Satterthwaite, D. and G. McGranahan (2014), “Urbanisation concepts and trends”, IIED,

Save the Children (2014), A Devastating Toll: The Impact of Three Years of War on the Health of Syria’s Children, Save the Children, Westport, Connecticut and London,

Schedler, A. (ed.) (2006), Electoral Authoritarianism: The Dynamics of Unfree Competition, Lynne Rienner Publishers, Boulder, Colorado.

Shelley, L. (2014), Human Smuggling and Trafficking into Europe: A Comparative Perspective, Migration Policy Institute,

Sizemore, K. (2014), “A peacetime crisis: Violent crime and community-oriented policing in Guatemala”, Journal of International Service, Vol. 23/1,

Skogan, W. et al. (2009), Evaluation of CeaseFire-Chicago, Institute for Policy Research, Northwestern University, Evanston, Illinois.

Slutkin, G. (2015), Violence Is A Contagious Disease, The University of lllinois at Chicago,

Slutkin G. et al. (2015), “Cure violence: An evidence based method to reduce shootings and killings”, paper for the World Bank,

Small Arms Survey (2016), “Monitoring trends in violent deaths”, Research Notes, No. 59, September 2016, Geneva,

Smith, S. (2015), “The elite’s road to riches in a poor country”, in T. Callaghy, R. Kassmir and R. Latham (eds.), Intervention and Transnationalism in Africa Global-Local Networks of Power, Cambridge University Press, Cambridge.

Soares, R. and J. Naritomi (2010), “Understanding high crime rates in Latin America: The role of social and policy factors”, in R. di Tella, S. Edwards and E. Schargrodsky (eds.), The Economics of Crime: Lessons for and from Latin America, National Bureau of Economic Research Conference Report, University of Chicago Press, Chicago.

Sobek, D. (2010), “Masters of their domains: The role of state capacity in civil wars”, Journal of Peace Research, Vol. 47/3.

Spano, R., C. Rivera and J.M. Bolland (2010), “Are chronic exposure to violence and chronic violent behavior closely related developmental processes during adolescence?”, Criminal Justice and Behavior, Vol. 37/10.

Stewart, F. (2011), “Inequality in political power: A fundamental (and overlooked) dimension of inequality”, European Journal of Development Research, Vol. 23/4.

Stewart, F. (2008), Horizontal Inequalities and Conflict: Understanding Group Violence in Multiethnic Societies, Palgrave Macmillan, Basingstoke, UK.

Straus, S. (2012), “Wars do end! Changing patterns of political violence in sub-Saharan Africa”, African Affairs, Vol. 111/443.

Szabó de Carvalho, I. and R. Muggah (2016), Latin America’s Cities: Unequal, Dangerous and Fragile. But That Can Change, World Economic Forum, 13 June,

Tadros, M. (2015), “Whose security lens on gender matters in the Arab uprisings?”, in E. Monier (ed.), Regional Insecurity After the Arab Uprisings: Narratives of Security and Threat, Palgrave MacMillan, UK.

Thies, C.G. (2010), “Of rulers, rebels, and revenue: State capacity, civil war onset, and primary commodities”, Journal of Peace Research, Vol. 47/3.

UN DESA (2014), World Urbanization Prospects, UN Department of Economic and Social Affairs, United Nations, New York,

UN Security Council (2016), “Special report of the Secretary-General on the strategic review of the United Nations multidimensional integrated stabilization mission in the Central African Republic”, S/2016/565,

UN Security Council (2015), Statement by the President of the Security Council, S/PRST/2015/23, 25 November 2015, United Nations, New York,

UN Women (2015), Preventing Conflict, Transforming Justice, Securing the Peace: A Global Study on the Implementation of United Nations Security Council Resolution 1325, UN Women, New York,

UNDP (2014), Human Development Report 2014: Sustaining Human Progress: Reducing Vulnerabilities and Building Resilience, United Nations Development Programme, New York,

UNEP (2009), Kenya: Atlas of Our Changing Environment,

UNFPA/UNICEF (2011), “Girls and young women”, Year of Youth Fact Sheet, UN, New York,

UN General Assembly (2016), “Report of the Special Representative of the Secretary-General for Children and Armed Conflict to the General Assembly”, A/71/205, 25 July 2016,

UNHCR (2016a), Global Trends: Forced Displacement in 2015, United Nations High Commissioner for Refugees, Geneva,

UNHCR (2016b), “Where we work: Africa”,

UNHCR (2015a), Culture, Context and the Mental Health and Psychosocial Wellbeing of Syrians: A Review for Mental Health and Psychosocial Support Staff Working with Syrians Affected by Armed Conflict, United Nations High Commissioner for Refugees, Geneva,

UNHCR (2015a), Mid-year Trends June 2015,

UNICEF (2016a), #ENDviolence against children,

UNICEF (2016b), Children on the Brink: The Impact of Violence and Conflict on Yemen and its Children, United Nations Children’s Fund, New York.

UNICEF (2016c), No Place for Children: The Impact of Five Years of War on Syria’s Children and their Childhoods, United Nations Children’s Fund, New York.

UNICEF (2006), Behind Closed Doors: The Impact of Domestic Violence on Children, United Nations Children’s Fund, Geneva,

UNOCHA (2016), Humanitarian Response Plans Financial Tracking Service,

UNODC (2013), Global Study on Homicide 2013: Trends, Contexts, Data, United Nations Office for Drugs and Crime,

Uppsala University (2016), Online portal to Uppsala Conflict Data Program, Department of Peace and Conflict Research, Uppsala University, Uppsala, Sweden,

Uppsala University (n.d.), “Uppsala Conflict Data Program definitions”,

Urdal, H. and K. Hoelschler (2012), “Explaining urban social disorder and violence: An empirical study of event data from Asian and sub-Saharan African cities”, International Interactions, Vol. 38/4, pp. 512-528.

US Centers for Disease Control and Prevention (2015), “National violent death reporting system”,

Vidal, J. (2015), “Why fragile cities hold the key to stability and development”, The Guardian, 26 January 2015,

Vilalta, C.J. (2015), “Global trends and projections of homicidal violence: 2000 to 2030”, Homicide Dispatch, No. 2, Igarapé Institute, Rio de Janeiro,

Vlassenroot, K. and T. Raeymaekers (eds.) (2004), Conflict and Social Transformation in Eastern DR Congo, Academia Press Scientific Publishers, Ghent, Belgium.

Von Einsiedel, S. et al. (2014), “Major recent trends in violent conflict”, Occasional Paper, No. 1, UN University Centre for Policy Research, Tokyo,

Weissebecker, I. and A. Leichner (2015), Addressing Mental Health Needs Among Syrian Refugees, Middle East Institute,

WFP (2015), “C.A.R. crisis regional impact”,

WHO (2016), “Violence against women: Intimate partner and sexual violence against women”, Fact Sheet No. 239, World Health Organization, Geneva.

WHO (2015), Preventing Youth Violence: An Overview of the Evidence, World Health Organization, Geneva,

WHO (2014), Global Status Report on Violence Prevention 2014, WHO/UNDP/UNODC,

WHO (2002), World Report on Violence and Health, World Health Organization, Geneva,

Wilmot, H. et al. (eds.) (2016), Protection of Civilians, Oxford University Press, Oxford, UK,

World Bank (2013), World Development Report 2014, Risk and Opportunity: Managing Risk for Development, World Bank, Washington, DC,

World Bank (2011a), Crime and Violence in Central America: A Development Challenge, World Bank, Washington, DC,

World Bank (2011b), World Development Report 2011: Conflict, Security and Development, World Bank, Washington, DC,


← 1. Characterised by those with over 1 000 battle deaths per year.

← 2. From 4 to 11. These are: Afghanistan, Central African Republic, Democratic Republic of the Congo, Iraq, Libya, Nigeria, Pakistan, Somalia, South Sudan, Syria and Ukraine.

← 3. In this chapter, domestic political instability refers to the contest among elites, their respective agents or non-governmental insurgent groups that seek to strengthen their position within a domestic political system.

← 4. The Geneva Declaration Secretariat defines sub-regions in line with the UN Statistics Division classification: Eastern, Middle, Northern, Southern and Western Africa; Caribbean, Central America and South America; Northern America; Central, Eastern, Southern, South-Eastern and Western Asia; Eastern, Northern, Southern and Western Europe; Australia and New Zealand, Melanesia, Micronesia and Polynesia. See

← 5. The Southern African countries are Botswana, Lesotho, Namibia, South Africa and Swaziland.

← 6. In this report, homicide refers to “intentional homicide”, which UNODC defines as “unlawful death purposefully inflicted on a person by another person”. See, for example,

← 7. A political elite is a group of people that represents an important political, social or economic interest. Their types and number vary across political contexts, and include formal government figures such as national representatives (e.g. senators and cabinet members), regional governors, mayors of large cities, traditional authorities of large ethnic communities, leaders of militias or rebel groups, top military personnel, representatives from resource wealthy areas and business leaders, and opposition party leaders. The identification of elites is possible through multiple metrics including the effective political parties, cabinet positions, traditional authorities and veto players. This information can be leveraged to determine the distribution of power and influence of subnational and regime authorities throughout the state.

← 8. A political settlement refers to the agreed power distribution between elites. It is also commonly called the “elite bargain”.

← 9. Contrary to perceptions of state inclusion, favouritism and nepotism, communities with co-ethnics in power do not consistently have disproportionate access to powerful positions or public goods over other areas without such standing (Kasara, 2007; Arriola, 2009; Francois, Rainer and Trebbi, 2014). Hence, the “exclusion” argument can be quite difficult to prove.

← 10. The international community has increasingly focused on protection of civilians (PoC) as a priority, and the UN Security Council (2015) confirmed it intends to build PoC into its mandates in contexts of political violence where civilians are at significant risk. PoC featured in UN deployments in the Democratic Republic of the Congo and South Sudan and Darfur, and in the NATO intervention in Libya, as well as in the failure of intervention in Syria (Wilmot et al., 2016).

← 11. Other associated social service, judicial or health costs are excluded from this estimate due to lack of data.