Beyond Gang Truces and Mano Dura Policies : Towards Substitutive Security Governance in Latin America

With responses to urban violence receiving increasing academic attention, the literature on anti-gang efforts in Latin America has focused mainly on coercive mano dura policies and cooperative gang truces. Yet, there remains a paucity of studies going beyond such carrots-and-sticks approaches towards gangs. To fill this gap, this study investigates the possibilities and limitations of substitutive security governance across Latin America and the Caribbean. More specifically, this article looks at Disarmament, Demobilisation and Reintegration (DDR) programmes in Medellín, Armed Violence Reduction and Prevention (AVRP) efforts in Haiti and Security Sector Reform (SSR) in Guatemala and Rio de Janeiro. It will be argued that communities are driven to support gangs against the oppressive state when they are indiscriminately targeted through muscular operations. Likewise, engaging gangs in dialogue grants them legitimacy and further weakens the position of the state. Therefore, the only sustainable solution lies in substitutive security governance, which aims to replace the functions gangs fulfil for their members, sponsors, and the community in which they are nested with a modern and accountable state that is bound by the rule of law. Still, substitutive strategies vis-à-vis gangs have their own limitations, which can only be overcome by way of an integrated and coordinated framework.


Introduction
Over the last decade, the debate about the prospects and perils of engaging non-state armed groups (NSAGs) -traditionally focusing on insurgents and rebel groups -has been expanded by studies investigating how international agencies operating in cities should deal with the challenge posed by criminal gangs (Bangerter 2010;Wennmann 2014).At the same time, an additional set of studies has critically examined the consequences of increasingly militarised responses to urban violence (Graham 2011;McMichael 2015).What is more, an abundance of policy-oriented research has investigated different approaches to the response to urban violence committed by gangs, be it by way of humanitarian action, development aid, or peacebuilding interventions (Lucchi 2012;Reid-Henry and Sending 2014;Björkdahl 2013).In the context of Latin America, the literature on anti-gang efforts has focused mainly on coercive mano dura policies (Swanson 2013) on one hand, and cooperative attempts at brokering truces (Cruz and Durán-Martínez 2016) on the other.
Yet, apart from noteworthy exceptions on the pages of this journal (Muggah and O'Donnell 2015;Rodgers and Jensen 2015), little attention has been paid to what I call 'substitutive' approaches towards gangs, which are intended to replace the different functions gangs fulfil for various stakeholders -namely their sponsors, their members, and the community in which they are nested.While Disarmament, Demobilisation and Reintegration (DDR) programmes can help replace the patron-client relations between gang members and their sponsors, Armed Violence Reduction and Prevention (AVRP) efforts aim at reducing the appeal for at-risk youth to join such groups in the first place.In addition, Security Sector Reform (SSR) has the potential to cut the ties of gang members to their patrons and to their community by contributing to end the impunity for politicocriminal entrepreneurs and by supporting the state in providing security, respectively.As this article makes clear, only an integrated framework for substitutive security governance combining elements of SSR, DDR, and AVRP can overcome the challenge posed by criminal gangs in Latin America today.

Functions of Gangs for Different Stakeholders
Before discussing strategies to curb gang violence, it is imperative to understand the drivers behind their proliferation.According to Rodgers and Hazen (2014: 8), gangs are characterised 1) by 'violent behavior patterns that are considered illegal by the dominant authorities and mainstream society' and 2) by ' a measure of institutional continuity independent of its membership' -which consists predominantly of young males.This definition sets criminal gangs apart from other types of NSAGs, including insurgents and rebel groups, which are -at least rhetorically -motivated by the belief in universal principles such as communism or Islamism (Schuberth 2015a).Although they tend to be seen as having a more parochial raison d'être, gangs do not emerge in a vacuum but in an environment shaped by numerous intertwined structural, political, and socio-economic factors (Moser 2004).Hence, gangs do have a political dimension, even though they are often pulled into the political sphere by external sponsors rather than being internally motivated (Hagedorn 2008).
When gangs are pulled into the political sphere, they act on behalf of political entrepreneurs who might also happen to be leaders of such groups.Therefore, it is important to identify the different factions or 'stakeholders' of gangs, meaning different sets of actors for whom violence committed by gangs serves -to varying degrees -a primarily political, economic, or security-related purpose and who have therefore vested interest in their continued existence.Figure 1 illustrates that gangs serve primarily political purposes for their sponsors, fulfil socio-economic functions for their members, and protect their own communities as much as they prey upon them -which is the reason why community members might -at least initially -support their emergence.It must be stressed, however, that the different functions and dimensions of gangs are inherently blurred, non-exclusive, and penetrate one another.
As illustrated in Figure 2, while the motivations to support gangs overlap, their violence fulfils one primary function for each stakeholder -political in the case of patrons, economic for gang members, and security as far as the community is concerned.Distinguishing between ' organized political violence' on one hand and ' ordinary' -that is ' ostensibly private, non-political violence' -on the other hand, Green and Ward (2009: 609) argue that violence can be ' dual purpose', meaning that it fulfils two distinct yet interconnected functions at the same time.While this reasoning resonates well with our concept of gangs, we add a third function of violence -that of security provision -to the political and criminal ones alluded to by Green and Ward.Accordingly, it is more accurate to use the term 'multi-purpose' violence, underlining that one and the same gang is regularly involved in three types of violence at the same time.

Beyond Cooperative and Coercive Strategies vis-à-vis Gangs
Given the huge variety of functions of gang violence, this article makes the case that the standard carrots-and-sticks approaches consisting of mano dura policies and gang truces might exacerbate rather than relieve the challenge posed by gangs in contemporary Latin America.On one hand, mano dura or cero tolerancia policies, characterised by military raids in slums and the mass incarceration of presumed gang members, might strengthen gangs by further alienating marginalised communities from the state and by allowing gangs to position themselves as legitimate defenders of the community against abusive state security forces and foreign intruders.On the other hand, cooperative approaches consisting of the brokering of truces between criminal gangs might

Political
Security Economic equally strengthen gangs by legitimising gang leaders as trustworthy dialogue partners and by granting gangs de facto control over their turf.Instead, it will be argued that comprehensive anti-gang policies are required that focus on replacing the functions gangs fulfil for different stakeholders.More specifically, it is imperative to move beyond coercive and cooperative approaches vis-à-vis gangs towards substitutive security governance.While DDR can facilitate separating gang members from their patrons and reintegrating them into society, AVRP can help dis-incentivising at-risk sections of the population from joining gangs in the first place.Moreover, in order to replace the security function that gangs fulfil for their community, SSR -including police and justice sector reform -helps (re)establish order and the state monopoly over the legitimate use of force.Moreover, SSR can contribute to cutting the ties between gang members and their politico-criminal sponsors by ending the impunity enjoyed by the latter.Ultimately, as this article makes clear, the only viable solution to the problem posed by gangs lies in building democratic and accountable states bound by and enforcing the rule of law.The following sections will analyse the promises and perils of coercive approaches towards gangs -often labelled the War on Gangs -and of cooperative approaches -consisting largely of the brokering of truces between criminal gangsby way of concrete cases taken from cities across Latin America.In order to cut the ties of gangs to their sponsors and to the communities in which they are based, the case will be made for an integrated framework of substitutive strategies vis-à-vis gangs.Subsequently, DDR efforts in Medellín, AVRP programmes in Haiti, and SSR in Rio de Janeiro and Guatemala will be analysed before discussing the limitations of substitutive strategies.In conclusion, recommendations are given to improve substitutive security governance in Latin America and beyond.

Coercive Strategies: The War on Gangs
The coercive approach combines military raids against gangs -which effectively elevate the groups to the level of a threat to national security -on one hand, and mass incarceration of presumed gang members by law enforcement agencies on the other hand.Nevertheless, in Latin America the theoretical distinction between police and military tends to be blurred, as raids and mass arrests are often executed by militarised police forces (Ungar 2011).Reminiscent of other coercive campaigns that rely primarily on the use of force by state security agencies, for instance the War on Terror or the War on Drugs, the two strategies can be subsumed under the comprehensive term War on Gangs.Such heavy-handed responses to gangs have proliferated over the last decade across many parts of Latin America and the Caribbean (Imbusch et al. 2011).
The War on Gangs has been most visible, however, in the countries of the Northern Triangle in Central America (Jütersonke et al. 2009).Almost simultaneously, between July 2003 and January 2004, El Salvador introduced a draconian anti-gang policy known as mano dura ('firm hand' or 'iron fist'); Honduras passed a similar legislation under the name cero tolerancia ('zero tolerance') and Guatemala adopted its own -though less strict -plan escoba (' operation broomsweep') (Rodgers et al. 2009: 12f).What these repressive policies had in common was the increasing deployment of military forces for domestic law enforcement purposes, which led to increasing confrontation not only between rival gangs, but also between gangs and state security forces (Hume 2009: 139).At the same time, mano dura legislation provided for the -in part unconstitutionalmass incarceration of suspected gang members solely based on their appearance, such as for sporting gang-affiliated colours or tattoos (Gutiérrez Rivera 2010: 496).
However, mano dura policies have been all but universally condemned as 'largely disappointing', be it by donors, scholars or local NGOs (Seelke 2007: 3).To begin, a number of advocacy groups accused the anti-gang laws of violating basic human rights, especially those of children (Serrano-Berthet and Lopez 2011).Moreover, the heavy-handed approach has been judged not just 'highly ineffective' (Fogelbach 2011: 454), but essentially counterproductive, as gangs used the overcrowded prisons to recruit new members and came out 'more sophisticated' and more organised than when they had entered (WOLA 2008: 30).Lastly, the coercive approach has been criticised for its strong emphasis on shortterm security-focused interventions, to the detriment of long-term prevention-and rehabilitation-oriented programmes (USAID 2006: 35-37).
In reaction to this criticism, Central American governments shifted from what Rodgers et al. (2009: 16) called 'first-generation' aggressive anti-gang policies to more comprehensive 'second-generation' interventions that focus on prevention and follow a more cooperative or substitutive logic.However, Peetz (2010: 1485) and Fogelbach (2011: 452) observe that such mano extendida (' outstretched hand') and mano amiga ('friendly hand') policies were notoriously underfunded and lacked serious commitment from the governments in question.Thus, despite significant financial and technical support from donor countries, UN agencies, and international organisations such as the World Bank, Rodgers and Muggah (2009: 311) suspect that states were guided by 'ulterior motives' when announcing secondgeneration anti-gang policies, which were found to be 'highly cosmetic, and principally aimed at pleasing potential donors and raising international funds'.
Lacking real commitment to less coercive methods and facing some of the highest homicide rates in the world (UNODC 2014), it perhaps comes as little surprise that governments in the region reverted to first-generation mano dura policies in recent years (Ribando Seelke 2013: 10f;Lohmuller 2015).On top of that, it has been argued that populist politicians have used mano dura policies as a strategy to win elections, not to actually fix a problem (Reisman 2006: 150).In this respect, Hume (2007: 739, 743) argues that criminal gangs are constructed as a 'principal security threat in the hemisphere' through ' an "othering" and dehumanising discourse', in which fears of possible connections between Mara Salvatrucha and al-Qaeda have been stoked in an attempt to conflate the equally coercive War on Terror with the War on Gangs.
While the 'tough on crime' talk of penal populism might please certain constituencies, poor communities will feel further alienated if their experience with state agencies is limited to military raids during which civilians often become casualties of stray bullets.This damages the legitimacy of the state and is easily exploited by the political wings of gangs, which try to boost their reputation as defenders of the community against external aggressors.Seen through the lens of counterinsurgency (COIN) theory, muscular interventions can have the inadvertent effect of contributing to the emergence of ' accidental guerrillas', meaning local people who have been driven to support or join gangs 'because they are alienated by heavy-handed actions of the intervening force' (Kilcullen 2009: 38).Thus, coercive strategies relying on the ' excessive and arbitrary use of force' can hardly be seen as a panacea for relieving the challenge of gangs as the root causes leading to their proliferation are not addressed (Muggah and Aguirre 2013: 3).

Cooperative Strategies: Gang Truces
As military raids against gangs and mass incarceration of suspected gang members overwhelmingly failed as public security policy, governments across Latin America started exploring alternative ways to deal with the problem of gangs in their cities.Chief among these is the brokering of truces between rival gangs.Such initiatives have been pursued inter alia in Belize, Trinidad and Tobago, Brazil, Mexico, and Honduras, often with tacit approval by the respective governments.Yet, as the example of the 2012 gang truce in El Salvador demonstrates, the surge in cooperative strategies towards gangs across the region rests on an equally weak evidence base as the earlier wave of mano dura policies.On top of that, truces mirror the downside of coercive strategies in that they tend to equally strengthen the security and political dimensions of gangs while weakening that of the state.
Whereas Klein and Maxson (2010: 236) note a 'total absence of valid empirical demonstrations' of successful mediation between gangs, it has been found that gang truces regularly lead to a short-term reduction of homicide rates followed by middle-to long-term surges in violence (Muggah et al. 2015: 88f).This is best exemplified by the case of the 2012 gang truce in El Salvador between the Mara Salvatrucha (MS-13) and the Calle 18 or 18th Street gang (M-18).The truce initially led to a steep decline in homicides -according to unverified accounts by the Salvadoran government by almost 60 per cent (Milliken 2014: 5).However, homicide rates have since skyrocketed and even surpassed pre-truce levels, with March 2015 being the most violent month the country experienced in ten years (Gagne 2015b).Due to the rising number of killings, the truce has been declared as failed by El Salvador's Security Minister, the National Police director, and even San Salvador's Archbishop (Dudley 2014).
The maras are not the only ones to blame for the failure, however; the Salvadoran government has played an equally counterproductive role.For instance, gang members have been understandably disappointed by the government's inaction concerning legitimate demands for appalling prison conditions to be improved and for security forces to adhere to the rule of law and human rights standards (Farah and Phillips Lum 2013: 22).Moreover, the involvement of authorities has been criticised as too intransparent and secretive, leading to suspicions about their true objectives (Umaña et al. 2014: 97).
Likewise, Milliken (2014: 6) notes that the government has been sending contradictory signals over its support of, or opposition to, the truce, notably in the months before the 2015 elections.
The opaque and at times inconsistent stance of politicians can be explained by the fact that the truce has been highly unpopular among the population as well as among donors (Does and McElligott 2012;Whitfield 2013: 13).Critics pointed out that mediation efforts focused exclusively on reducing the number of homicides, whereas all other violent criminal activities have been exempt (Schwartz 2012).Thus, whereas murder rates decreased for a certain period of time, crimes such as extortion remained on an ' exorbitantly high' level and the number of other serious violent crimes -including rape -actually augmented at the time the truce was still de facto in place (Thale et al. 2013).Today, El Salvador's public security policies are back to square one, as the breakdown of the truce has been followed by an intensification of militarised suppression efforts, with some analysts estimating that as much as 30 per cent of the skyrocketing killings can be attributed to the police (Gagne 2015a).
Given the transnational character of the maras, the truce had considerable implications for other Central American countries and beyond.For one, it is suspected that the truce has created a window of opportunity for the maras to strengthen not only control over their territory, but also their ties to transnational organised crime groups such as Mexican cartels, which have been increasingly expanding southwards into the Northern Triangle (Farah 2012a: 63).Moreover, Schwartz (2012) argues that the truce increased the bargaining power of criminals in the region by sending the ' disconcerting message […] that governments are willing to address violence through compromise with perpetrators, rather than by seeking justice', which could ' encourage criminal bands elsewhere to escalate their demands in the future'.This assessment points to the longterm implications of the failed mediation attempts.The rising numbers of attacks by gangs on state security forces raise questions about the political aims of the maras and the threat they pose to the Salvadoran state (Farah 2012b).In countries with fragile state institutions like El Salvador, the balance-ofpower problem is not limited to rival gangs but extends to the dynamics between gangs and the state itself.From this point of view, the truce has in the long term 'increased the power of the gangs rather than that of the state', whereby the 'prospects for a loss of government control over security and the rule of law in areas deemed as "safe zones"' is particularly troublesome (Farah and Phillips Lum 2013: 25; Ribando Seelke 2013: 13).
Likewise, Maguire (2013: 11) raises concerns that mediation efforts with gangs might enhance their power, prestige, and legitimacy, thereby elevating their bargaining power to ' demand concessions from government leaders in exchange for promising reductions in violence'.In such cases, due to their use of violence, gangs might be given a stronger voice than nonviolent institutions which enjoy greater legitimacy within their neighbourhoods.In the worst case, they might be treated as representatives of the very residents who are most vulnerable to the criminal violence and stigmatisation resulting from the presence of gangs.
If mediation with criminal gangs seems unavoidable to immediately lower escalating levels of violence, they must be complemented with and followed up by efforts to tackle the underlying conditions that contributed to the emergence and consolidation of gangs in the first place (Muggah et al. 2015: 88).Indeed, gang truces are best seen as a last resort to facilitate the implementation of substitutive security governance aimed at replacing the functions gangs fulfil, rather than as an end in itself.As Thale et al. ( 2013) point out, 'it is important to take advantage of the (possibly temporary) low levels of violence to implement prevention and reintegration programmes in the communities most affected by gang activity'.

Towards Substitutive Security Governance
As we have seen so far, the most vulnerable communities in many mega cities in contemporary Latin America are caught between state violence fuelled by indiscriminately coercive anti-gang policies and violence from gangs, which have been empowered by cooperative policies.Hence, the only sustainable solution lies in substitutive security governance, which aims to replace gangs with a modern and accountable state that is bound by the rule of law and serves all sections of society in a fair and equal manner.Substitutive strategies must focus on breaking the patronage between gangs and their politico-criminal sponsors while simultaneously cutting the ties between gangs and the community.This can only be achieved by an integrated approach that aims to replace all three functional dimensions of gangs -the security they provide for their community, the income-generating role they play for their members, and the political function they fulfil for their sponsors.Figure 3 shows how such a coherent framework for action could look like: As illustrated in Figure 3, DDR programmes are one possible way to break the link between the politico-criminal elite and gangs by dismantling the former command structures and offering former gang members the possibility to give up their arms and start a new life.DDR can therefore be a useful tool to target those who have already joined gangs, even though preconditions for the successful implementation of DDR -such as a peace agreement between clearly defined and centrally organised conflict parties -are often absent in urban contexts.AVRP programmes such as MINUSTAH's Community Violence Reduction initiative in Haiti, by contrast, try to prevent at-risk youth from joining gangs in the first place.Through the provision of temporary employment, often in the form of cash-for-work projects, AVRP can help sway unemployed youth away from economic opportunities offered by gangs.
Figure 3 above depicts two ways in which SSR can usefully complement DDR and AVRP efforts to cut the links between gangs, their sponsors, and the community.First, SSR can tackle the patron-client relationship between gangs and their politico-criminal sponsors by contributing to end the impunity enjoyed by affluent and influential sections of the population.To this end, assistance can be given to establish special courts or hybrid national/international tribunals to deal with transnational organised crime and to handle politically sensitive cases, as happened with Guatemala's International Commission against Impunity.Second, SSR can strengthen efforts to expand state security provision into areas formerly abandoned by law enforcement agencies in which gangs have temporarily assumed the roles of informal crime control and self-defence.This can be done by enhancing the capacity and legitimacy of police and judiciary through training as well as proximity policing techniques, as exemplified by the Pacifying Police Unit in Rio de Janeiro.The following sections will critically examine cases in which substitutive strategies have already been employed to replace gangs in cities across Latin America.

DDR in Colombia
The substitutive logic of DDR becomes evident when looking closer at its constituent parts.In contrast to disarmament, which is arguably the most straightforward and technical aspect of DDR, there is general agreement that the reintegration component is both the most crucial and the most complex aspect of DDR (Özerdem 2012).Bridging the short-term goals of disarmament with the Art.17, page 9 of 20 long-term agenda of reintegration, demobilisation has been described as 'the true core of the DDR project' (Shibuya 2012: 55).A main objective of demobilisation is to replace the patron-client relationship between ' entrepreneurs of violence' within the political elite and ex-combatants.As Themnér (2011: 25) has shown, former mid-level commanders often play an important role as intermediaries between the elite and demobilised members of armed groups as they ' continue to yield much influence over their former subordinates' whom they can effectively 'remobilize'.
In recent years, DDR has been applied in a number of unconventional scenarios, notably in cities plagued by non-war, gangrelated, criminal violence (Muggah and O'Donnell 2015).Apart from a UN-led effort to disarm gangs in Haiti, most academic attention in the Latin American context has been paid to urban DDR processes in Colombia (Rozema 2008: 439; Howe 2014).This is predominantly due to efforts to disarm the Autodefensas Unidas de Colombia (AUC), which had taken control over youth gangs and vigilante groups in Medellín's poor neighbourhoods, where 13 per cent of the 31,671 demobilised combatants lived (Abello-Colak and Guarneros-Meza 2014: 3274).Among the demobilised urban AUC-affiliated blocs was the Bloque Cacique Nutibara (BCN), which had established its own criminal fiefdom within Medellín's comunas, at one point controlling more than two thirds of the city (Rozema 2008: 439).
Praised for having contributed to a significant drop in homicides in the city, DDR in Medellín is generally seen as a success and has even served as a model for national reintegration policy in Colombia (Giraldo-Ramírez and Preciado-Restrepo 2015: 6).Nevertheless, it has also come under severe criticism.To begin with, Nussio and Howe (2014: 850) raise the concern that the demobilisation of AUC-affiliated blocs left behind a power vacuum in neighbourhoods previously under their grip.This resulted not only in increasing violence rates as the informal protection system has been removed, but also in fierce battles over control of these areas (Howe 2014: 187ff).Therefore, it is indispensable to fill the 'governance void' left by demobilisation efforts with formal state security forces, which need to be strengthened through SSR initiatives.
The partial or complete absence of the state in urban areas where the AUC once reigned can be traced back to failed demobilisation initiatives implemented in the 1990s, which led to the incorporation of former militias into a newly established private security structure (Giraldo-Ramírez and Preciado-Restrepo 2015: 4f).The Security and Community Service Cooperative (COOSERCOM) established in Medellín in 1994 had demobilised militias patrol in poor neighbourhoods (Rozema 2008).However, due to concerns over human rights violations, abuses against the civilian population and fatal internal power struggles, COOSERCOM was disbanded in 1996, leaving behind a power vacuum that was later filled by the ilk of BCN (Sanín and Jaramillo 2004).Even worse, no lessons seem to have been learned from this episode: a decade later, ex-BCN members that had demobilised in 2003 were reportedly employed by private security companies (PSCs) as part of a Zonas Seguras programme, which had subsequently been abandoned due to concerns raised inter alia by Amnesty International (2008: 41).
When they are not hired by PSCs, urban ex-combatants have good prospects of being recruited by bandas criminales -also known as BACRIMs -which can afford to offer substantially higher sums than what is being paid as part of the reintegration scheme (Kaplan and Nussio 2016: 19).These criminal groups resulted from the failure to properly demobilise former mid-level commanders of AUC blocs and have been found to operate as 'umbrella organizations for different local gangs', taking advantage of the power vacuum left behind by the demobilisation of the BCN (Wienand 2015: 203).In recent years, more than 30 BACRIMS have consolidated into a handful of major ones, of which Los Urabeños operate on a national level, whereas the Paisas and the Oficina de Envigado are based in Medellín (Ince 2013: 29).
BACRIMs have been found not only to contribute considerably to the internal displacement in Colombia, but also to recruit their members specifically among Internally Displaced Persons (IDPs) in cities (Howe 2014: 190).Therefore, in addition to DDR, there's an urgent need for AVRP initiatives to help vulnerable urban dwellers, such as IDPs, to stay away from criminal groups.In light of the recently signed peace agreement between the Colombian government and the Revolutionary Armed Forces of Colombia (FARC), such an integrated approach is also needed should a DDR process be initiated vis-à-vis FARC subgroups, many of whose former members have relocated to cities where they are subject to recruitment efforts by criminal gangs (McQuinn 2016: 13).

AVRP in Haiti
Mirroring botched demobilisation efforts in Colombia, the failure to demobilise and reintegrate former soldiers of the Haitian armed forces in the 1990s contributed decisively to the political turmoil that led to the ouster of President Aristide in 2004 and the proliferation of armed gangs in the capital (von Einsiedel and Malone 2006: 162).Subsequent endeavours to demobilise armed gangs proved no less difficult, partly because the preconditions for the implementation of a traditional DDR programme never existed in Haiti.The country was not in a typical post-conflict situation as a diffuse array of paramilitary groups and politicised urban gangs were causing general instability and chaos rather than fighting for well-defined political aims or to take power (Molloy 2008: 6f).Accordingly, negotiating a peace agreement between two or more clearly defined factions with a central command was not an option in the first place.
Because of the challenges of applying traditional DDR to the Haitian context, a new approach to deal with Haiti's gangs had to be found.Therefore, MINUSTAH shifted towards a more comprehensive and preventative Community Violence Reduction (CVR) strategy trying to sway at-risk youth within marginalised communities away from the influence of gangs rather than focusing exclusively on armed members of such groups.This shift was based on the realisation that, as a Programme Manager for the IRC put it succinctly, 'with DDR you might be able to provide livelihoods for a set of members that were involved, but if the underlying causes generating the membership to gangs are not addressed, you will have a new batch.Then you're basically providing a retirement programme for gang members'. 1 The substitutive logic of CVR is evident in its objective to work 'with marginalized neighbourhoods to create economic and social opportunities in view to extract them from violence and the influence of gangs' (MINUSTAH 2012: 3).This objective is to be achieved inter alia through the labourintensive rehabilitation of community infrastructure which aim to provide ' a temporary alternative to criminality and unemployment' and the improvement of ' capacities and skill sets of at-risk youth' through professional training (MINUSTAH 2013: 2).Thus, the ultimate goal of CVR is to cut the stream of new recruits from the community to gangs by replacing the income-generating function the latter provide for their members, chiefly through the provision of legal job opportunities. 2  However, CVR's extensive focus on temporary employment through Quick Impact Projects (QIPs) has been criticised as unsustainable and prone to corruption and exploitation (Schuberth 2016).In addition to providing alternative means of income in the short term through QIPs, another expected outcome of CVR is the strengthening of community cohesion and the fostering of trust and political will to work together, which could possibly lead to disarmament in the long term (MINUSTAH 2013: 1).In this respect, a former CVR Programme Officer maintained that 'CVR is seen as a very successful experiment […] because if there is no political will, you won't disarm people.But CVR is a very flexible approach that you can use to get people to work together and foster trust.And trust ultimately creates political will'. 3  An additional long-term objective of CVR is the strengthening of the state in such a way that the responsibility over the programme can be transferred to Haitian institutions in the future.However, Haitian authorities regularly complained that state institutions were bypassed at best and sabotaged at worst (CNDDR 2011: 3).In order to address these shortcomings, more emphasis has recently been given to the strengthening of state institutions, closer coordination with the government, and the prospect of handing over the management of CVR initiatives to the state (MINUSTAH 2013: 5f).However, MINUSTAH staff voiced scepticism whether the state has the capacity to take charge of the programme. 4Accordingly, vertical integration of top-down state-building and bottom-up violence reduction initiatives at the community level has been described as ' elusive' (Donais 2015: 45).
Another common criticism is that the thematic scope of CVR initiatives is too broad and in many cases not linked clearly enough to the goal of reducing violence. 5Yet, it could be argued that the thematically broad scope of interventions in geographically well-delineated communities is exactly what sets CVR apart from earlier unsuccessful DDR experiments in Haiti.Compared to DDR, CVR has been described as ' a very broad reintegration programme' by a former Programme Officer: 'when you're looking at a DDR programme, your measurement is the individual.But in CVR the understanding is that violence happens in the community, so you have to address the community as a whole'. 6Hence, an important difference is that DDR focuses on particular groups, whereas CVR has a clearly defined geographical focus.
Moreover, the excessive focus on police reform in Haiti has come to the detriment of other parts of the criminal justice system that have become penetrated by politicocriminal entrepreneurs who continue to instrumentalise urban gangs with impunity (Schuberth 2015b).Therefore, strengthening the independence of both the judiciary and the police from the political and economic elite must be at the centre of SSR efforts in Haiti (Albrecht et al. 2009).More specifically, the International Crisis Group (2007: 1) recommends the establishment of ' a hybrid Haitian/international tribunal to deal with cases of transnational, organised crime that the country can not tackle on its own ', whereas Perito and Dziedzic (2008: 14) suggest to create a 'special-crimes courts for politically charged cases', such as 'political assassinations, intercommunal violence, and political linkages to drug trafficking, kidnapping, and gang warfare'.

SSR in Guatemala and Rio de Janeiro
While rampant impunity and the collusion between gangs and politico-criminal entrepreneurs is a challenge in many a country around the world (Schuberth 2014), the International Commission against Impunity in Guatemala (Comisión Internacional contra la Impunidad en Guatemala -CICIG) has been praised as a model for strengthening the rule of law in comparable contexts, such as Haiti (WOLA 2015).Established in 2006 as a joint hybrid commission through an agreement between the UN and the Guatemalan State, CICIG's mission is evident in the name of its predecessor, the Commission for the Investigation of Illegal Groups and Clandestine Security Organizations in Guatemala (Comisión de Investigación de Cuerpos Ilegales y Aparatos Clandestinos y de Seguridad -CICIACS) (Maihold 2016).
Since the end of the Guatemalan civil war in 1996, violent crime rates escalated in the country as urban gangs have teamed up not only with organised crime groups and Mexican cartels, but also with police officers, former military, and the political establishment in the capital, especially during electoral campaigns (Briscoe and Stappers 2012: 23ff).With government officials intimidated or assassinated and state institutions coopted by criminal groups, the CICIG proved crucial for dismantling criminal elements by supporting the investigation and prosecution of their crimes (Brands 2011: 228).Yet, external security governance interventions in the judicial sphere such as CICIG must also be supplemented with efforts to reform corrupt law enforcement agencies, a task the Pacifying Police Unit (Unidade de Policía Pacificadora -UPP) has attempted to fulfil in Rio de Janeiro.
Mirroring the War on Gangs in Central America, the purely coercive approach of fighting drug gangs by way of violent raids by the Batalhão de Operações Policiais Especiais (BOPE), a special police unit of the Military Police of Rio de Janeiro State, led to staggering numbers of so-called 'resistance killings', whereby the police accounted for the majority of intentional homicides in some of Rio's favelas in 2007 (Human Rights Watch 2009: 4).One year later, amidst national calls for the promotion of citizen security and with Brazil winning the bid to host the World Cup in 2014 and the Olympics in 2016, the governor of the state of Rio de Janeiro, Sérgio Cabral, announced the launch of UPP, a proximity policing programme aimed at replacing the functions gangs fulfil as security providers for their community by re-establishing state control over Rio's gang-ruled favelas (Muggah and Souza Mulli 2014).
Strong government support behind UPP and partnerships with the private sector set it apart from an earlier community policing programme called Police Group for Special Areas (Grupo de Policiamento em Áreas Especiais -GPAE), which ultimately failed due to corruption and politically motivated changes in the leadership of the programme (Leeds 2007: 31f).While UPP was designed to reduce violence rather than to confront criminal gangs, it must be stressed that BOPE remains an integral part of the programme.Applying to the favelas of Rio the triad of ' clear-hold-build' that is at the heart of COIN, UPP consists of three phases: a tactical intervention, followed by a stabilisation phase, and a consolidation period (Muggah and Souza Mulli 2014: 204).In fact, it is only between the second and the third phase that UPP take over from BOPE, which remains central to the first two phases (Lessing 2012: 66f).
Thus, far from being just another experiment in public security, observers and residents alike hope that in the long run, UPP can contribute to the '"pacification" of the police' themselves (Serrano-Berthet et al. 2012: 119).It has been argued that rather than constituting a major police reform by and in itself, UPP stands for ' a conduit for gradual, incremental changes that […] could effectively reform policing in Rio' (Pinto and Do Carmo 2016: 130).This is supposed to be achieved primarily by way of recruiting and training fresh police officers not yet '"contaminat[ed]" by the levels of corruption and the old doctrine of the rest of the force' (Cano and Ribeiro 2016: 369).From this perspective, UPP can be seen as an attempt to replace both gangs acting as informal security providers in Rio's favelas as well as the traditional police force which had lost its legitimacy in the eyes of many residents (Denyer Willis and Mota Prado 2014: 239) However, UPP has had its fair share of critics.To begin with, the programme has been criticised for failing to target favelas ruled by milícias, vigilante groups made up of off-duty policemen and fire-fighters that have established veritable reigns of terror and extortion in the neighbourhoods they have taken over from drug factions (Rodrigues 2014: 6).Moreover, Lessing (2012: 70) reports that parts of special forces such as BOPE are involved in milícias and that UPP troops have been accused of ' extortion and other milícialike activities'.On top of that, concerns that UPP targets primarily favelas adjacent to rich neighbourhoods and tourist spots are coupled with fears that the programme will be discontinued after the Summer Olympics in 2016 have been held (Isacson 2011).
A further point of critique is the relative negligence of UPP Social, the social development component which aims to 'integrate the favelas into the rest of the city' and sets the programme apart from purely coercive strategies (Serrano-Berthet et al. 2012: 38).Even though UPP Social is crucial to reduce violence and prevent at-risk youth from joining gangs, it has not received adequate support (Cano and Ribeiro 2016: 370).Due to political and legal disagreements between the state and the city administration, UPP Social was transformed into a municipal research institute that could not effectively fulfil its objective to coordinate the implementation of social projects (Foley 2014).Thus, it is crucial to complement UPP's police reform efforts with community-based violence reduction and prevention projects aimed at swaying at-risk youth away from the influence of gangs.
What is more, while the practice of announcing operations beforehand helped keep confrontation between police and gangs low -sometimes with no shots fired at all -it has been criticised for giving criminals the possibility to flee and continue their illegal activities elsewhere (Lessing 2012: 71).This 'balloon effect' has led to the relocation of drug-related violence into smaller municipalities outside Rio de Janeiro and further south towards Paraná state (Muggah and Souza Mulli 2012: 66).To prevent this unintended consequence without resorting to violent confrontations of earlier days, those gang members willing to demobilise, disarm and (re)integrate into their community should be given the possibility to participate in DDR programmes.

Limitations of Substitutive Strategies
As the preceding sections have shown, the substitutive strategies of SSR, DDR, and AVRP share several common difficulties when it comes to their implementation.To begin with, they are notoriously difficult to coordinate and integrate into a coherent framework for action as they involve multiple stakeholders and aim to address various interconnected problem areas in a comprehensive manner (Paris 2009).A related concern is that substitutive security governance is potentially overambitious, particularly when considering the short timeframes typically allocated for its implementation (Egnell and Haldén 2009).The difficulty of evaluating their impact or to attribute their outcomes to specific projects is an additional challenge common to substitutive programmes (Mutongwizo et al. 2015: 1).Nevertheless, SSR, DDR, and AVRP have one crucial advantage over coercive or cooperative approaches: they aim to make gangs obsolete by replacing the functions they fulfil for their community, sponsors, and members.By doing so, substitutive security governance gives legitimacy to the state, while limiting that of gangs.
Whereas coercive and cooperative strategies by themselves run the risk of strengthening the legitimacy of gangs while weakening that of the state, they can play a supportive role for the three substitutive strategies of SSR, DDR, and AVRP.Concerning coercive strategies, welldirected interventions by security forces can target exceptionally intractable gangs that cannot be dealt with by cooperative or substitutive means.In this respect, it is of utmost importance that the use of force is restricted in order to avoid civilian casualties.Moreover, cooperative strategies can facilitate safe access to neighbourhoods under the control of gangs so that international agencies can implement substitutive programmes.Still, whenever coercive and cooperative strategies are seen as an inevitable necessity in the short term to stabilise the situation and bring open hostility to an end, it is crucial to switch towards substitutive security governance with the least possible delay.

Conclusion and Recommendations
To conclude, it has been argued that standard tools to deal with criminal gangs across Latin America -coercive strategies, such as military raids in slums and the mass incarceration of presumed gang members, on one hand, and cooperative strategies including the brokering of truces between gangs, on the other hand -have failed to live up to their own promises.In their stead, the case has been made for efforts to improve security governance by way of substitutive strategies which tackle the root causes that led to the emergence of gangs in the first place by replacing the functions they fulfil for their different stakeholders.Such substitutive strategies vis-à-vis gangs must address their three dimensions -security, political, and economic -in a complementary and coordinated manner.As has been argued in this article, SSR helps end impunity and replace the security-related dimension of gangs; DDR initiatives may substitute their political dimension; and AVRP programmes can supplant their economic dimension.
Yet, to reshape institutions -let alone build them from scratch -is a notoriously lengthy, time-consuming, and complicated process.Therefore, the substitutive approach can only be implemented if substantial amounts of time and resources are invested into integrated efforts which address the structural causes for the formation of gangs, while at the same time mitigating the influence patrons exert on such groups.Development activities should focus on the generation of job opportunities for slum dwellers in order to improve their life chances and to limit the attractiveness of joining gangs.Ideally, this should not be limited to temporary cash-for-work employment schemes, but should include efforts to bring the private sector into marginalised areas that offer an untapped potential for investments but have long been left out due to their bad reputation.
What is more, instead of simply containing the urban poor, the security and judicial sectors must refocus their efforts on combatting high-scale organised crime, which requires ending the impunity enjoyed by the political elite and powerful families in control of the business sector.Thus, the state must serve the interest of all citizens by retaking control of gang-ruled areas and fulfilling its primary function as the exclusive provider of security.In this respect, it is of paramount importance that law enforcement agencies abide by human rights standards and instil trust, rather than fear, across all sections of the population.
Perhaps most importantly, the use of gangs as goon squads by political actors must be thwarted through the promotion of the core principles of democracy, which include respect for the rule of law and good governance.Especially those most affected by the violence -the urban poor -must benefit from the long-term advantages of democratisation by way of a functioning rule of law and a capable government, not only from short-term gains in the form of patronage distributed by populist leaders.Lastly, researchers and practitioners alike should draw lessons from the vast experience gathered by local initiatives from Medellín to Rio de Janeiro to Port-au-Prince, which can contribute meaningfully to the development of more effective strategies to deal with gangs in contemporary Latin America.

Figure 1 :
Figure 1: Functions of gangs for different stakeholders.

Figure 2 :
Figure 2: Functions of violence for various stakeholder of gangs (dark grey = primary function; grey = secondary function; light grey = tertiary function).