Cascades of violence and a global criminology of place

From displacement to cascades

The spirit of this article is to start with some of the propositions listed in the abstract that have a fairly high degree of acceptance with the explanation of crime and to translate them to explanations of war. In the process of translating them to more general propositions of crime–war, peace studies deliver some productive insights to criminology as well. The article argues that because violence cascades, it is imperative that we prevent it before it cascades too much. One reason is that there are tipping points in cascades of violence (Kennedy, 2009). It also follows that it is mostly a mistake to fight violence with violence; only rarely is this the best or only recourse remaining to us. We must focus on fighting it through prevention. Violent means of pursuing simple policies such as regime change in Iraq or Afghanistan must be tested critically against the question: ‘What might violence in pursuit of that regime change cascade to?’

The plan of the article is first to hypothesize that both war and crime are hot-spot phenomena subject to cascade more than to displacement. Then a case is made for addressing both forms of violence one hot spot at a time. Mass rape in the Congo is used to illustrate the dilemmas of crime–war. Cascade dynamics that explain the spread of militarized violence from one place to another in a place like Congo are then employed to develop a different kind of theory of patterns of criminal violence in societies that are not at war.

Criminologists have been insufficiently interested in cascades of violence. They have been more concerned with violence displacement. This is the idea that if we extinguish violence at one place, those who have a propensity to violence will simply move to indulge it at another place. The empirical literature mostly tends not to reveal displacement effects for violence and crime generally (Weisburd et al., 2011). If anything, when we extinguish crime at one hot spot of violence, crime also tends to decline somewhat in surrounding areas. This should have been a clue for criminologists to follow cascades of violence as a more important research question than displacement of violence. The particular kind of cascade dynamic explored in this article is violence that spreads horizontally in space from hot spot to hot spot, sometimes linking one hot spot to another as ink spots that expand and connect up. Elsewhere my research explores more vertical cascades from the centre to the periphery of the world system.

A starting point is the hypothesis that there will prove to be analytic value in not separating the study of cascades of violent crime from cascades of warfare. In the past year I have conducted the Peacebuilding Compared project in countries such as Afghanistan and the Democratic Republic of Congo (DRC) that continue to be war-torn. People say it must be dangerous. Perhaps it is not as dangerous as wandering into a hot spot of crime in a large American city. The 3% of addresses in American cities that account for 50% of their crime can be genuinely dangerous places (Sherman et al., 1989; Weisburd, 2012). Yet we wander about the other 97% of high-homicide American cities in great safety compared to other times and places in human history. Likewise in countries at war, possibly more than 97% of the locations in space–time are lovely, safe places to be. So my hypothesis is: Violent crime is highly concentrated at hot spots; warfare is concentrated at local hot spots.

Peacebuilding one spot at a time

The perceptive subtitle of Gerry van Klinken’s (2007) book on ethnic, religious and separatist conflicts in Indonesia is Small Town Wars. It is a book that provides us with useful tools to diagnose the opportunity structures of Indonesian provincial centres and how local political elites and local organized criminals manoeuvre to dominate them. A key insight of Small Town Wars is that if you diagnose warfare in the statist fashion that typifies most international relations and political science, you will miss the point of the war. Severine Autesserre (2010) has provided one of the most intellectually influential critiques of UN peacekeeping in The Problem with Congo. She scolds UN officials and the diplomats of western states alike as buying into a simplistic grand narrative of what the wars of the Democratic Republic of Congo are about. In that narrative the invasion of DRC by Rwandan and Ugandan military forces in 1996 is seen as the start of a first Congo war. Then there was a second Congo War in 1998 that was also triggered by an international intervention. As we discuss later, these international interventions were certainly important elements of the slaughter of millions in the Congo. The withdrawal of invading armies achieved by international peace agreements were critical steps toward peace. The problem diagnosed by Autesserre is that diplomats are socialized to think that these international political drivers of war are what make these wars tick. She points out that there was terrible fighting, rape and slaughter for years before the Rwandan–Ugandan invasion of 1996. This raged around many local conflicts over land and local power that at times connected up to one another to create wider wars. Combatants in one war allied with combatants in another to build their own alliance structure in order to defeat a shared or unrelated enemy.

These wars raged in such remote and geopolitically unimportant places that the international community barely noticed them, let alone attempted to understand them. But without these pre-1996 wars creating a pretext for invasion and local allies to join the invaders, the 1996 invasion would not have occurred. When international peace negotiations finally persuaded invading armies to leave the Congo, the international community deluded itself into believing that what it was doing was supporting post-conflict peacebuilding in the DRC through MONUC (The UN Mission to the Congo). Autesserre pointed out that wars continued to wage throughout the MONUC mission. They were in part the same local wars that had raged for years before 1996, or successors to them. So we can see the failure of peacemaking in the Congo as one of looking at the conflict too much through the national-international lens of UN and other diplomats. What was needed was extinguishing the conflicts that were the root causes of the 1996 invasion hot spot by hot spot.

We have learnt from Iraq that violent death rates often go up after a war ends, with only El Salvador having a higher total violent death rate than Iraq between 2004 and 2009 (Geneva Declaration, 2011: Chapter 2). It has also been discovered in a number of African and other conflicts that killing (Duffield, 2001: 188), even moreso rape and domestic violence, can increase after a peace agreement is signed, and likewise for some Latin American civil wars where a doubling of already extreme homicide rates at the end of the war delivered a higher death rate than during many of the peak years of civil war (Richani, 2007; Muggah and Krause, 2011: 180). There can be various local reasons for this. A common one is that a peace agreement between warring armies gives them security from each other’s guns which allows them to concentrate with impunity on pillage of civilian populations whose domination they divvy up in the peace deal. Below, I will discuss the mass rapes at Luvungi in which several armed groups, who once fought each other, participated at adjacent villages in a string of mass rapes.

One of the worst ways of reading the takeaway message of this essay would be that because wars must be resolved hot spot by hot spot, where each one has its own logic, that all we need to do to fight twenty-first century wars is to deploy the technocratic tools of criminology. A bit of violent crime mapping here, a bit of problem-oriented or pulling-levers policing there, a touch of restorative justice conferencing to secure reconciliation somewhere else, is not what I have in mind, though these tools probably should have more place than they do in UN peacekeeping. On this analysis, peacebuilding remains fundamentally about finding political solutions to problems through political conversations among protagonists. It is just that the most critical political conversations may not be among national leaders, but among regional clan leaders, religious or ethnic leaders, or rural warlords who have chased out the state, who are the state for all practical purposes in the region they control. A key role of UN peacekeepers becomes one of creating safe spaces where these leaders can meet with locally knowledgeable peacebuilders to advance such political conversations without fear of being killed. UN diplomats do not need to know enough about the variegated local conflicts to drive their preferred solutions to them village by village. They do at least need to know where the hot spots are. They need to know enough of the dynamics and leaderships of their local hot spot contestation to know who must be promised security to come together to lead political dialogue toward a local peace agreement, local reconciliation and reintegration of combatants.

For the policing part of the peace operation, it follows that the normal approach of dropping in thousands of international police to drive UN vehicles around the capital is not very helpful. Rural policing is normally the priority. It is not just that you cannot build rural community policing capacity in a village society with UN police sitting in patrol cars in the capital. It is also that you can no more build a police organization to reach out to rural spaces by flying in a thousand police than you can build a civil aviation system by flying in a thousand pilots to build aircraft and control towers. ¹ Operational UN police have never built a police organization in their home state. Doing that requires people with skills in project management, in founding police academies, building payroll systems and the like. Most of all it requires political sophistication in how to support local leaders in remote corners of village societies who have the language and cultural skills of their district to teach community policing. It must be community policing skills of a kind that are relevant to community problem solving at the kinds of hot spots they have.

We now turn to one case study from the Congo to help us conceive the challenge of peacebuilding as partly a challenge of building peace one diamond mine at a time, one village at a time. Our conclusion up to this point in the analysis is to have advanced this hypothesis: Peace is accomplished hot spot by hot spot rather than nation by nation. Political settlements and a politics of reconciliation are needed to resolve contemporary violence that are less national, more oriented to the local politics of hot spots.

Luvungi: Hot spot of war and crime

I was booked on a UN helicopter to go to Luvungi in Walikale, North Kivu, Democratic Republic of Congo in April, 2012. The day before at an interview with the regional UN military commander the wisdom of the visit was discussed. At the end of the discussion, the general not only counselled against my going, but limited the UN helicopter to 30 minutes on the ground. I was glad I did not go because research was impossible in 30 minutes and when the helicopter did land, local people were angry because the UN should have flown in more food, fuel for generators and fewer people. So this research is not based on observation but only interviews with people who have been there and on access to UN documents. I hope I might visit Luvungi on my next visit to the Congo. In the meantime, I want to try to tell their story as best I can.

Luvungi is a village in one area controlled by one Colonel Ntabo Ntaberi Checka and his gang, Mai Mai Checka. Hot spots of persistent warfare in the Congo often map onto sites of mines for gold, diamonds and other precious resources, but especially coltan, vital for our mobile phones, tablet computers and laptops. Colonel Checka is wealthy from control of mines such as the Bisiye mines, and extracting benefits from child slaves working in those mines. In 2010, Checka’s men allegedly committed mass rape against at least 387 people in a string of 13 villages around Luvungi with some FDLR (Democratic Forces for the Liberation of Rwanda) and other elements. Most victims were women and children, though there were 23 confirmed male rape victims. The youngest rape victim was male, a one month old baby. Beyond the 387 survivors confirmed by the UN police, there is an unknown and probably large number of even more terrified victims still hiding in the jungle, dead, or fearing the consequences of providing testimony. Almost all known victims were gang raped by two to six men or boys on multiple occasions in the course of four days. A number of terrified rape victims hiding in long grass also suffered snakebites. 116 people, including a number of children, are known to have been captured into slavery as a result of the Luvungi attack.

The order to attack was issued by Checka as part of a campaign to consolidate his control. Checka was one of many leaders of armed groups who had been offered a deal of amnesty and integration into the Congolese armed forces in the late 2000s, and earlier, in return for ceasing to add to the interminable wars that had cost millions of Congolese lives since 1996. Terms were never reached in any of these peace deals. It is easy to critique such peace proposals as unconscionable impunity. My own view is that peace deals in which most rapists get an amnesty and financial support to reintegrate into a life as peaceful farmers in return for handing in their weapons, relinquishing military control of the hot spots around mines and their slave labour are a necessary path to protecting people from violence in the many sites of mass rape in places like Luvingi. The reality is that the military capacity is not available to capture all of the Checkas in the Congo. Nevertheless, amnesty for Colonel Checka will now be shown to be something that could not be supported.

After the 2010 Luvungi mass rapes, with international assistance from European and Canadian rape investigators, warrants were issued for the arrest of Checka and seven others. Indian attack helicopters from MONUSCO, the current UN peacekeeping mission, were laid on to back up Checka’s arrest. Perhaps someone was paid to tip him off. There seemed to be good intelligence on where he could be captured, but he had fled. After the first court hearing for the only defendant who was arrested, Checka’s fighters advised victims that unless they recanted another mass rape would be repeated against the women and children of their village. This led many women to conclude that recanting was the right thing to do to protect their children.

This is actually a case where arrest and criminal prosecution may not be sufficient for the peak of the responsive enforcement pyramid approach I tend to apply to these challenges (Braithwaite, 2002: Chapter 6; 2008: Chapter 4). Indeed the want of a capability to escalate beyond attempted arrest is making things much worse for the terrified folk of Luvungi today. They are at risk of another stint as refugees struggling to survive in the bush. The need at the time of writing in mid-2012 is to escalate to deployment of credible military strike capability in Luvungi and to advise Colonel Checka that unless he surrenders to face trial he will be attacked. This is unlikely to happen. For reasons that need not detain us, the DRC government has neither the political will nor a reliable military capability to do this. The two senior local Congolese commanders, Colonels, were assassinated near Luvungi with 18 other soldiers two days before I was due to fly there in April 2012, quite probably by Checka, though we cannot be sure that some other armed group was not responsible. Nor do we know whether it had anything to do with military attempts to arrest him or to clear his control of mines.

What we probably do know is that if western special forces were deployed to the area and negotiated a deal, Checka could be brought to surrender. We know this because it has been done before. Because Checka probably has only around 150 fighters, some of them children, he could not survive credible force indefinitely. He might be able to hide from competent special forces for a time. Meanwhile it would not be a difficult military objective to clear him permanently from any control of local mines, cutting off the financial lifeblood that makes insurgency an attractive business, and protecting victims of his crimes from further intimidation. At the time of the last big crisis in this province of Eastern Congo in 2009, the UN Secretary-General called for western troops to come in to help control the situation. He issued this call because in the previous crisis when European Union troops were deployed in one part of the East for a few months, the killing was quickly and sharply reduced, though not eliminated. No western state heeded the Secretary-General’s 2009 call. All the senior western diplomats I interviewed in DRC gave the same reason why western boots on the ground to take the fight to armed groups (if surrender could not be negotiated) was not going to happen. Iraq and Afghanistan were enough; their electorates were fed up with deployments to protracted conflicts. I argued back to the diplomats that clearing weak armed groups from mines was such an easy objective compared with defeating the Taliban. This did not budge them from the political impossibility of satisfactorily protecting the people of Luvungi. They might be at such great risk because of our consumption of the minerals they dig; they might be terrorized because of attempts at criminal enforcement that had been launched at the behest of the west, but the west is then capable of washing its hands of further responsibility. Perhaps it is not fair to say washing its hands as the UN Security Council put Checka on its sanctions list in November 2011: useful, but not something that will protect victims.

For a responsive regulatory theorist like myself, this is a parable with layers of meaning. One meaning is that it can be reckless with peoples’ lives to escalate a long way up the regulatory pyramid required unless one is willing to escalate all the way. In the face of how politically hard this can be, we cannot responsibly be dismissive of impunity at the base of pyramids that reconcile with the most terrible of criminals. More escalation is so often worse than less if there is not that political commitment to escalate however far is needed to secure safety with justice.

While the worst place to be in a tough case is stuck in the middle of an enforcement pyramid without a credible capability to escalate to force the safest possible resolution, this is also a case where negotiated settlement is second-best to deployment of military power. One reason is that it is possible that asserting territorial domination and punishing villagers for past cooperation with the government are not the only reasons for Checka’s terror campaign. Another reason might be to spread terror as a bargaining chip to end the terror in return for a deal where he gets impunity, corrupt control of certain mines, and effective control of the state apparatus in this patch of the country after integration of his forces into the Congolese army.

A second argument for special forces deployment toward the peak of an enforcement pyramid here is the terrible re-victimization situation the survivors have been placed in. I was deeply impressed, moved to tears, by the integrity of the young western feminists in the UN gender affairs and sexual violence units, UNDP, and the UNPOL rape investigation team in North Kivu listening to Congolese women. There were some hard fought differences of opinion on how to handle the terrible dilemmas they manage. One was on where to hold the preliminary hearing for the one Luvungi rape defendant who was arrested. Some felt it would put survivors less at risk if the trial were not held in their own community. This also suited the court. So the hearing was held at the provincial capital, Goma. There is no road across the rough terrain between Luvungi and Goma. Some brave survivors walked 90 km to testify, traversing territory controlled by Checka. We in the west have special culpability if we (successfully in Luvungi) call for escalation only part way up an enforcement pyramid, then walk away politically if our states are asked to back up those arrests with the resources needed to take care of survivors and with the force required to protect them. Until we are ready to do that, it is ethically wrong of us to support the principle of international law that criminal prosecution is mandated for crimes against humanity such as the use of rape as a weapon of terror in war. By doing so in part measures we cause great suffering.

One reason it is right to make the special forces investment in peacekeeping to enforce the arrest of the likes of Colonel Checka is that doing so would strengthen the hand of peace negotiators at other hot spots where vulnerability to mass rape might be ended by negotiations. That is, conviction of a Checka allows negotiators to say to others that if you try to hold out instead of surrendering your weapons, the fate that awaits is that which Checka suffered when he held out and threatened re-victimization of survivors. So our policy hypothesis here is: Display a responsive enforcement pyramid that gives violent groups in control of hot spots an opportunity to negotiate a truce at the base of the pyramid. When they do, reintegration of perpetrators and shaming of their crimes can occur. When the opportunity to renounce violence is spurned, firm resolve to escalate to arrest backed by military capability in the case of armed resistance to arrest is necessary.

This policy would be not so different from gang violence reduction programs such as Operation Ceasefire in Boston (Kennedy, 2009: 3–8). These involve recognition that enforcement resources to target all gangs in all high crime areas are hard to find. However, selective convictions of leaders of recalcitrant gangs who will not renounce gun violence are used as a resource to negotiate new norms for other gangs that choose to be gun free. The limitation of these US gang ceasefire programs compared with the best peacekeeping practices is that they did not fund reconciliation conferences followed by reintegration programs that support alternative non-violent livelihoods for large numbers of perpetrators (see Braithwaite, 2002: 36–43), though they did talk about the desirability of doing that. Operation Ceasefire programs might show higher violence reduction impacts in the criminological evaluations if they delivered on that. This hypothesis is: Negotiations that persuade violent groups to relinquish control of local hot spots of criminal opportunity under threat of prosecution will prevent violence more permanently when accompanied by reconciliation and reintegration into legitimate opportunity structures.

Even so, the evidence for the effectiveness of programs of the Operation Ceasefire type at American violence hot spots is encouraging enough (Kennedy, 2009; Braga and Weisburd, 2012; but see Corsaro et al., 2012). One reason we might expect superior violence reduction in a program that shuts down Eastern Congolese hot spots of violence one by one is that these hot spots are supported by a much more lucrative opportunity structure than dominating drug markets or some other illicit market in part of a US city. When a hot spot is a hot spot because it is the site of a coltan or diamond mine, military clearing of armed groups from the mine followed by training of police to hold the mine for long enough for the military to fly back in should militants try to retake it, can more decisively remove the root cause of the hot spot than in the case of a drug market.

Thinly resourced policing can be effective for risks of great scale and complexity. I have been slow to grasp this possibility. As recently as two years ago I commiserated with a European general who had served in the Congo about how impossible his job must have been with so few enforcement resources, with mass murder and mass rape of millions occurring around him. Not at all, he replied, what was required, and what he was doing in his area of Eastern Congo in the mid-2000s, was to call all the warlords of his patch in after an incident of mass rape and explain that while he did not have the resources to pursue all of them militarily, he did have the resources to pursue the next one of them who perpetrated any atrocity of that kind in future. He had independently invented an Operation Ceasefire strategy. When I interviewed another general in the Congo in 2012, he said something rather similar. He opined that if he were allowed to hunt down these small armed groups one by one in the way he would if the problem of mass rape arose in his home country, it would take 4–6 months to clean out the warlords from the mines and to end the mass rapes. The problem was, both generals said, that their home governments and the UN did not want to risk losing peacekeepers to the cause of making the DRC safe. And such a campaign possibly could not be conducted without taking some losses.

Congo as a cascade

The Congo wars that have been more or less continuous since 1996 are perhaps the worst the continent of Africa has experienced in lives lost and rapes committed. They were not caused by armed groups fighting to control mines any more than the wars in Afghanistan were caused by fighting to control poppy fields. The coltan and the poppies contributed to keeping these wars going at the hot spots where these illicit opportunities abound.

Conflicts over land between different ethnic groups in Eastern Congo, some of whom had fled there after previous wars in Rwanda and Burundi and elsewhere, had caused low level conflict in the Congo for several decades (Autesserre, 2010). The genocide in Rwanda steeply escalated this kind of land conflict as hundreds of thousands of Hutu refugees, among them many of the perpetrators of the Rwanda genocide, cascaded across the border into the DRC. This caused even more land conflict and minor civil wars between Rwandan refugees and Congolese. Refugee camps full of desperate young men and boys also became excellent recruiting grounds for Hutu militants intent on destabilizing the post-genocide Tutsi regime in Rwanda. There were actually many cross-border cascades of this kind that fuelled regional concern about Eastern Congo becoming a lawless zone where insurgents could mount attacks against other states. For example, Uganda had this concern about Mr Kony’s Lord’s Resistance Army destabilizing Uganda from DRC. Rwanda and Uganda were so concerned at the failure of the Mobutu regime to control these armed groups that they worked with Congolese opposition groups to sweep into DRC in 1996 and drive Mobutu from power. Mass slaughter of Hutus by Rwandan forces and their proxies in DRC followed. This might be characterized as a counter-genocide.

Further wars in DRC resulted when Rwanda formed the view that the new DRC regime was not much better than the old one in bringing Hutu armed groups to heel. The second time around other regional powers sided with the DRC regime when Rwanda catalyzed another war aimed at regime change and retribution against Hutu armed groups in DRC. We need not come to grips with the enormous complexity of this war for the purposes of this essay. The lessons I wish to draw from it are that three key dynamics drove such an extreme cascading of violence. First, those displaced by violence displaced others from domination of the spaces to which they fled. Second, refugee camps became recruiting centres for those who sought to cause violence to cascade. Third, extreme destabilization of part of a country by successive waves of violence can cause certain spaces to become anomic Hobbesian vacuums that attract terrible tyrannies of domination. Colonel Checka’s domination of Luvungi illustrates that third dynamic, with the wealth of the mines simply underwriting an opportunity structure that has made his domination resilient.

It is early days in the research that seeks to refine these hypotheses inductively from the experience of cascades of violence not only across the Great Lakes region of Africa, but also elsewhere in Africa and across Central and South Asia. In Afghanistan, cascading from other regional players such as Pakistan also shows these three dynamics to be in play. Residues of previous wars left many groups angry at other groups who had occupied their lands. Refugee camps were places where the Taliban and Al Qaeda were able to recruit angry young men by giving them an education in Madrassas and a path to employment. Finally, the Taliban were only able to begin on their path to power in Kandahar in the mid-1990s by offering to establish Islamic order in Hobbesian spaces exploited by many armed gangs. The Taliban were able to show farmers that they could make it possible to get their produce to markets without being shaken down by many different armed groups along the road. The Taliban shut down their roadblocks. At one level, they made rural spaces safe for women who were being raped by armed men. Of course, the Taliban then used that domination of an anomic space as an ‘armed rule of law movement’ to impose a new form of tyranny, not only upon women. David Kilkullen (2011) has taken us on a journey across time and space to show from the writings of the ancient Greek Herodotus (1954) how military commanders with a small local base could expand that base during periods of Hobbesian (1651) anarchy by providing quality justice and security services to ever widening circles of frightened citizens.

Bangladesh is another case where we have seen an Islamist group cascade into domination of a rural space where there was a rule of law vacuum. Bina D’Costa and I are undertaking a study of armed violence in the Chittagong Hill Tracts of Bangladesh that sees it as falling at the end of a long cascade of violence that consumed all of South Asia (Braithwaite and D’Costa, 2012). Imperial domination cascaded down to a politics of separation of India from the Mughal and British Empires, which cascaded further down to a politics of separation of Pakistan from India that cost two million lives, cascading to a war of separation of Bangladesh from Pakistan, which cascaded through layers of internal conflict within Bangladesh. Modern wars cascade to the creation of Hobbesian rural spaces where women are routinely violated, revenge is indulged, rule of law is in abeyance, insurgents morph into gangs of organized criminals. These anomic spaces are in the market for a supplier of order. That supply might come from one organized crime group dominating, from an armed rule of law movement like the Taliban, UN peacekeepers, a state that supplies community policing and a rule of law, or a state military that allows enough pacification to justify its presence and enough anarchy to itself profit from organized crime. Braithwaite and D’Costa (2012) found the Chittagong Hill Tracts to fit the last description. Its descent into this condition is viewed in our research through a lens that points from the Chittagong Hill Tracts back through the layers of the above cascades. The Congo has sadly morphed into this latter kind of militarized domination of a Hobbesian vacuum. Foreign militaries from Rwanda, Uganda and elsewhere learnt from the Congo wars that there was much money to be made when they controlled Eastern Congolese mines. The successor DRC armed forces learnt that lesson as well.

Richani’s (2007) analysis of El Salvador, with comparisons to Guatemala and Lebanon, is more illustrative of the type of case where criminal gangs move in to monopolise social control at hot spots. The civil war in El Salvador caused a flood of refugees into the United States after 1980. They were mostly unskilled peasants with little or no English who experienced widespread unemployment in the US. This resulted in their recruitment into gangs and the drug trade particularly in Los Angeles. Gang wars arose from the takeover of formerly Mexican gangs by Salvadorians with fighting skills sharpened by civil war. In the years after the 1992 peace agreement in El Salvador, the US forcibly repatriated 130,000 immigrants, 43,000 of them with criminal records, back to El Salvador. Returning gang members took over certain hot spots, making a number of coastal sites in El Salvador transit points in narco-trafficking, something that never had happened before 1990 in El Salvador. El Salvador’s prisons became particularly inflamed hot spots where assassinations of gang members involved in the drug trade were rife. After the 1992 peace agreement, El Salvador’s homicide rate skyrocketed, killing more people per year than the civil war had. The homicide rate peaked at 138 per 100,000 in 1994 and 139 in 1995. Gang conflict was responsible for most of the homicide, which was concentrated at hot spots.

My research team is exploring and refining the nature of these global hot-spot-to-hot-spot dynamics in other regions of the world as well. In sum, we have so far refined our first hypothesis with three cascade dynamics: Violence cascades when:

Those displaced by violence displace others from spaces to which refugees flee.

Now I turn to applying these insights from the causes of war to a reframing of patterns of crime around the world.

Cascades of refugees and crime

Our refined set of hypotheses about violence might cause us to see patterns of crime differently through this macro-sociological lens. We might see Japan as a society that has fought wars at great cost, but has not suffered huge waves of refugees contesting the spatial control of incumbent groups over private land and public space as a result of those wars (Hypothesis 1 in the Abstract). Japan is also a highly reintegrative polity (Hypothesis 5), at least for the overwhelming majority of its population (Braithwaite, 2002: 18, 27). In this way, the hypotheses offer one account of Japan’s comparatively low level of violent crime. Of course there may be other reasons as well, as argued elsewhere (Braithwaite, 1989).

While much of South Africa is as placid as Japan, South Africa has countless hot spots of violence that are quite extreme compared to Japan. Even after a halving of South African homicide rates since 1994, that rate is almost 50 times the Japanese rate. One hypothesis to explain this could be that, in comparison, South Africa has suffered huge internal cascades of displacement in modern conflicts. First there was colonial displacement of indigenous populations for the white colonial economy, then the Boer War to assert British imperial claims to the resources (diamonds, gold again) controlled by the white colonists. Between 1910 and 1939 there was huge forced displacement of non-whites onto reserves where they would be available as reservoirs of cheap, unskilled labour for white farms, factories and mines, formation of squatter settlements in places like Soweto in the 1940s (Thompson, 2001: 163–4, 178). Then Apartheid, the Group Areas Act, and armed resistance to it produced enormous displacement of non-white peoples, driving refugee cascades internally and into neighbouring states, almost doubling black African unemployment (Thompson, 2001: 195). Furthermore South African regimes imposed centuries of domination that could take a century of reintegrative reform to repair. South Africa not only shares with the Congo suffering from recent cascades of violence, it also shares an early colonial history of some indigenous groups being enrolled against others to attack their lands and recruit them as slaves for the Americas.

What then of the lands into which these slaves cascaded? Some which received the largest numbers of slaves – such as Brazil, the Caribbean and the United States – ended with many and deep hot spots of violence. The slaves were bundled into ships with other slaves from ethnic groups unfamiliar to them, speaking languages they could not understand. This laid conditions for black on black competition (and violence) over the limited resources for which slaves could compete. New World slavery was not in the least reintegrative, even after one of the most deadly wars of the century was fought that theoretically liberated slaves in the United States. In the century after that civil war, freed slaves earned the resources to buy land in progressively more northern parts of the country with growing demand for factory labour. As the former slaves cascaded into industrial cities like Chicago and Detroit, they did not use violence to push out other immigrant groups such as Italians and Poles who lived in the cheaper parts of cities to meet the bourgeoning demand for industrial labour. They simply used their wages to rent and buy properties. But as Italian or Polish or Irish neighbourhoods gradually turned black (or Hispanic), property values fell, resentment and racism were stirred. There was no new civil war, but the violence of the riots triggered at Watts in 1965, by the police assault on Rodney King in 1992, cascaded hugely and widely. As in South Africa, there was a massive carceral response directed against the black population, a criminal justice system with a low level of reintegration.

Australia is a country where white immigration from 1788 caused a comparable displacement, decimation and stigmatization of its Aboriginal population to North America. On both continents this caused a significant number of high violence hot spots on reserves and on the margins of country towns like Alice Springs and Bourke where the numbers of survivors of disintegration were substantial and where peoples from different language groups were forced to compete with each other for meagre leftovers of land and jobs from the white settlers’ tables. Australia has delivered to these culturally decimated, displaced peoples neither reintegration policies to the white economy nor policies to restore lost identity that have succeeded in rebuilding a sense of hope in these spaces of hopelessness. One might say that the conditions of herding convicts onto ships with people they did not know had much in common with the slave trade. The thing it did not have in common is that convicts experienced conditions of extreme reintegration in a nineteenth century Australia with acute labour shortages (Braithwaite, 2001). By the 1820s, ex-convicts owned over half the wealth of the colony and three quarters of the land (Hirst, 1983: 81). The other critical difference between the Australian penal colony and King Leopold’s Congolese colony (Hochschild, 1998) that left the latter more violent is well captured by Acemoglu and Robinson’s (2012) brilliant book. Places like the Congo were bequeathed weak institutions designed to allow maximum extraction from blacks (resource extraction, extraction of human capital); white-settler colonies like Australia got institutions designed to build welfare for whites. One might have hoped that when black leaders professing an ideology of building welfare for Africans like Mobutu took over, the violence done by a politics of extraction and domination would cease. Mobutu was a good example of a leader who was attracted to seek power precisely to sustain institutions of extraction and domination. They made him one of the wealthiest people in the world. Extractive colonial institutional legacies are resilient because they bequeath criminogenic opportunity structures to successor elites that tempt those who control the guns in a society.

Much of this speculation on why some societies have more long-term violence than others doubtless seems tendentious. These are early days of a research work in progress. My objective for now is to open up for discussion a different way of seeing the global geography of violence in countries as diverse as Japan, South Africa, the Congo, Brazil, the United States and Australia, whereby war and crime are not as analytically separate as they are in extant scholarship. And my objective is to invite you to extend that way of pondering patterned difference to other societies with which you might be familiar such as Canada or New Zealand. That invitation is not only about pondering why some societies have more hot spots of violence than others, but also why the United States has more of its worst hot spots at centres of industrial capitalism like Detroit, while Australia has its highest crime at the extreme periphery of its capitalism in remote places like Bourke.

My argument is that cascades of violence drive long-term patterns of killing through both warfare and crime in places like Congo and Afghanistan. South Africa illustrates that when war stops, the violent crime abates only slowly in new conditions of peace and at some hot spots not at all if people continue to feel no less marginalized.

Cascades of refugees displacing settled normative orders, unsettling collective efficacy (Sampson et al., 1997) in the schools, parks and other recreational spaces of the neighbourhoods to which they move, have quite minor impacts on crime in societies where the refugee flows are in the range most societies experience, and for many refugee communities, by the next generation such is their sense of collective efficacy in their new communities and new opportunity structures that they actually have a lower crime rate than long-term natives of that country (Francis, 1981). Where good refugee reintegration policies are in place, in the medium term in the swings and roundabout of refugee impact, refugees do more to strengthen collective efficacy in a society than to weaken it. Cascades of refugees only become long-term causes of violence in places like Northwest Pakistan where five million refugees arrived in a short space of years. The other conditions I hypothesize for refugees contributing to long-term problems of violence are weak reintegration support and housing of refugees in concentration camp conditions. For Australia, for example, moves toward replacing detention of boat people in concentration camps with accommodation in the homes of volunteer families is, by the lights of my five hypotheses, a promising crime prevention policy.

Reversing the lens, cascades of refugees of high volume and high desperation can be a cause of war and mass rape that generous refugee policies in the west can help alleviate. Cascades of refugees are hypothesized only to have large impacts on violence in the extreme conditions we see in the refugee camps of the borderlands of the likes of Pakistan, Somalia, Sudan. My way of seeing the impact of large flows of people across borders on patterns of crime in a country like Australia is about nineteenth century flows of people that saw comparatively large scale and saw one people completely take over spaces from other peoples.

Even so, in the short term any kind of geographical mobility may make a tiny contribution to crime risks. Criminologists have known for 80 years that cities and suburbs with high rates of residential mobility have higher crime rates than spaces where these are lower (Shaw and McKay, 1969), partly because high mobility reduces collective efficacy (Sampson et al., 1997: 921). In most contexts what is going on here is minor and short-term as with those of us who have found that persuading high school children to move to another city is like convincing Lady Gaga to wear a burka. What our teenagers perceive in this context is a move to schoolyards, school socials and to nightlife that seem to them anomic. Of course they quickly learn that they are not. It is just that different peer elites set the norms, norms that are sometimes a bit different, but that can be learnt quite quickly with the help of new friends. Nevertheless, any move of a high school student is a move into a subjectively anomic space at least for a period; during that period mobility can cause significant conflict. One mobility context that might be more seriously criminogenic is forced mobility of the poor and the mentally ill in and out of prison and back into disrupted neighbourhoods (Clear et al., 2003).

The social dynamics of forced migration, anomie and conflict in western cities have elements in common with the long-term causes of asymmetric warfare in the world system. It is just that the consequences are more minor with western urban crime. For the same reason, when societies organize opportunities for young people to meet and date at school, university or church dances or socials, this is less anomic than meeting at huge city pubs and clubs that attract hundreds of patrons mostly unknown to one another. So within the framework of the theory being developed here, we might reframe the work of Ross Homel and others (Homel et al., 1997; Lincoln and Mustcin, 2000) on civilizing violence at pubs and clubs as interventions into spaces that are inherently rather anomic. The interventions that successfully reduce violence might be theoretically reframed as interventions to reduce anomie through bouncers and other club staff who regulate the space using some of the techniques the Taliban used in the mid-1990s to restore order to disordered spaces in southern Afghanistan.

Taking the theory more deeply into domestic space and away from violence, we might even see the challenge of regulating the crimes of Wall Street as one of civilizing a space with a high degree of mobility of aggressive young traders, of imbuing new norms into an anomic space. Settled rules of the game that are decent and that work well in the context of a particular hot spot that is unsettled in some way is all that is being suggested. It is not very novel. It is just a Chicago School portrait of the whole globe painted pointillist style. It is Durkheim for hot-spots, as opposed to Durkheim for so-called organic societies.

Conclusion: Contextual hot spot experimentalism

Geography matters. Most spaces are peaceful. Perhaps even more than 97% of the people of this planet live on a street where locals cannot recall anyone being murdered in recent memory. When many have been killed on a street, this usually says more about how different that neighbourhood is from others in that society than it says about their society, even in a society as violent as the Congo. There is hope in this that we can build a less violent world by expanding already large ink spots of civility until they encompass more and more of the surviving spots of incivility (Kaldor, 1999). Geography matters, we have argued, because the most local forms of politics matter. We are good at denying this because we like to pretend we understand a world that has too much local complexity for us to be capable of much understanding. One way we delude ourselves is by pretending that peacebuilding and violence control is mainly about statebuilding. If regulation of violence were about control by a strong state, Washington would have little of it.

This article is partly about peacebuilding learning useful lessons from the criminology of hot spots. Yet one reason why the United States, even after great progress during the past two decades, has been so much more unsuccessful at extinguishing hot spots than most societies is its technocratic lack of interest in learning from other societies in radically different contexts, such as peacebuilding. There is much that the criminology of hot spots can learn from peacebuilding. The prominence of DDR in peacebuilding is a good illustration. All peacekeeping operations have a DDR unit, so we might wonder why none of them have taken the idea back to establish DDR units in western police departments. DDR refers to Disarmament, Demobilisation and Reintegration of armed groups. In this essay I have emphasized reintegration as a serious neglect in the criminology of place. The May 2012 issue of Criminology and Public Policy has a series of essays on the criminology of place, including one by the author of a new book by that name, David Weisburd (2012), and another series of essays on ‘Pulling Levers Policing’ as a strategy for tackling the hot spots that Americans refer to as open-air drug markets. The special issue does not discuss reintegration of gang leaders targeted through these programs by alternative livelihoods programs, special educational programs, diverting capital accumulated through drug trading into building legitimate businesses (such as the moves to become private security contractors that we often see in war zones). Such options are not dismissed using any of a variety of good arguments that can be mounted against these forms of reintegration. None of these kinds of options are even mentioned in the special issue. Finding a reintegration strategy that will work is extraordinarily difficult. It takes a lot of contestable local knowledge. Yet when criminals continue to work at hot spots because they provide better opportunities than others currently available to them, it might seen as implausible as it is to peacekeepers that unless better opportunity structures are put in place, the violent opportunity structure of the hot spot will be resurrected.

One might say that disarmament and demobilization of gangs that control hot spots is central in that American hot spot literature. A difference from peacemaking practice is that peacemaking is more contextual whereas criminology tends to diagnose policy options like gun buy-backs as if the only way to execute them were as universal programs in a jurisdiction. Peace negotiators may agree on one buy-back fee per gun in the peace agreement it settles with one warlord, another price with another and offer nothing at all to a third. In some cases, payments will be made to the warlord (hopefully for distribution to the fighters who surrender the weapons). In others, payments are made to fighters one at a time as they surrender themselves and their weapon for destruction. Most commonly, in the holistic context of a peace deal with a warlord who controls a particular hot spot, a whole set of reintegration proposals are on the table, such as paying school fees for the children of fighters who have missed education while hiding in the hills. The deal is that none of these reintegration benefits will be provided until the guns are handed in. There are even peace agreements where destruction of weapons does not happen until a very late stage of a peace process after the state has built confidence by meeting all earlier milestones of its side of an agreement. Famously, in the Bougainville peace, rebel commanders stored guns in locked containers to which both they and the peacekeepers held a key. This prevented minions from reigniting conflict through rash discharge of firearms, yet gave leaders the option of re-arming until they could see that their enemies were honouring their side of the deal (Regan, 2010).

While I am an active supporter of the experimental criminology favoured by leading hot spot scholars Sherman and Weisburd (1995), here we see its methodological limits. If we test something like a universal gun buy-back policy with a randomized controlled trial, we might miss the contextual insights we get ethnographically in war zones: universal buy-backs are often gamed and fail, while disarmament deals that are well attuned to the local political realities of relationships between particular negotiators, particular warlords and their fighters, can work quite well. Expressed in a more scientifically cautious fashion, if we experiment non-randomly with one kind of disarmament deal after another until a peace holds, we very often get to peace. Scientifically, we do not know as we move up a regulatory pyramid whether it was the last layer of the pyramid, the last peace negotiation, that made the peace work, or some complex mix of things settled across multiple previous rounds of the negotiations. All we know is that redundantly adapting with different elements of a DDR repertoire can eventually silence the guns.

This is one reason for favouring deployment of a pyramid of strategies for regulating hot spots. The experimental approach here is about experimenting with one layer of the pyramid after the other until the violence goes away. This is not an experimentalism that validates causal universals. ² It is based on the humility of believing that most of our hypotheses are wrong most of the time, including those in this essay. Our best hope is that they will not always prove wrong in all contexts and that we will learn from many contexts to build redundancy into our strategies for violence control.