The New Wave: Forced Displacement Caused by Organized Crime in Central America and Mexico

Forced displacement generated by organized crime is a little-studied and poorly understood phenomenon. Based on field research carried out in 2013, this article redresses this situation by analysing the broad dynamics of an alarming new wave of forced displacement sweeping El Salvador, Guatemala, and Honduras – the countries of the Northern Triangle of Central America – and Mexico. It focuses specifically on the role played by three of the main types of organized criminal groups in the region – mara street gangs, Central American drug transporters, and Mexican drug cartels – in provoking this displacement. Structural differences between these groups are shown to influence both the forms of displacement that they produce and the resulting patterns of movement by displaced persons. Consideration is then devoted to the implications for scholarship and humanitarian practice of this new wave of forced displacement generated by organized criminal groups.

Of course, this wave of displacement is but one element in the consolidation of "new" forms of power and violence in societies that are still negotiating recent experiences of political turbulence and civil war. One base expression of these new practices is found in the astronomical homicide rates registered in parts of the Northern Triangle and Mexico, which are on par with some warzones. Leaving aside broader questions of continuity and rupture with the past, however, the present study tackles the more modest aim of describing and analysing how the activities and strategies of these new organized armed actors have provoked the displacement of other inhabitants. In so doing, it addresses a question that has thus far received sparse scholarly attention.
The article opens by exploring the limits of current knowledge on the relationship between organized crime and forced displacement. It argues in favour of taking the criminal groups, rather than their victims, as the starting point for analysis. As such, the study focuses on three types of criminal groups as agents of displacement in Mesoamericamara street gangs, drug transporters, and drug cartels -showing how differences between these groups influence both the forms of displacement that they generate and the patterns of movement undertaken by displaced persons in consequence. The role of the State in addressing this forced displacement and its humanitarian consequences is considered, after which certain scholarly and practical implications of this new wave of displacement are identified.

Crime and displacement: framing the debate
Forced displacement caused by organized crime is -despite its alarming proportions -almost entirely absent from official discourse in Mesoamerica. One reason for this is that refugee movements, and even internal displacement, continue to be conceptualised in these countries through the lens of their experiences of political conflict in the 1980s. Moreover, going beyond the security-oriented discourse that presently dominates current discussion of organized crime to acknowledge the humanitarian problems entailed by the new wave of displacement presents a conceptual challenge in what is already a field dominated by political sensitivities. This challenge is exacerbated by the fact that the current displacement is so empirically complex and its character and dimensions are poorly understood not only by officials but also by civil society, academics, and the general public.
The Mesoamerican region is not exceptional in this regard. An extensive academic literature is dedicated to the study of forced migration resulting from state repression, war, development, disaster, and climate change around the world. In contrast, serious studies of the role played by organized crime in provoking such population movements are few and far between. 2 Yet this is not because the phenomenon does not exist elsewhere: it has been documented in sites across the world, even if particularly heightened in Mesoamerica. 3 Through its study of a region where the problem is especially acute, this article aims to help conceptualise how forced displacement may be generated by organized crime.
The existing literature on crime and migration is not extensive. Indeed, interest in criminal organizations and migration has thus far tended to concentrate on their role in facilitating the smuggling and trafficking of migrants across international borders. A significant body of law, policy and research exists on this issue, globally and in relation to Latin America. 4 In parallel, an important body of policy and research has examined the legal basis for persons fleeing from criminal groups to be granted asylum under refugee law. 5 This emerged in the context of the growing numbers of asylum claims presented in the US by Central Americans fleeing gang violence in the late 1990s. 6 Yet neither body of literature addresses the empirical patterns of displacement caused by such groups.
More recently, a few researchers have begun to try to quantify levels of forced internal displacement in Mesoamerica. In Mexico, the most visible scenario of violence, preliminary data indicates that some 2 per cent of the population living in the country -translating to 1.65 million persons -changed residence in the five years between 2006 and 2011 owing to the threat or risk of violence 7 suggesting a mean average of 330,000 persons internally displaced per year. Indeed, more robust survey data indicate that in one of the most conflictive urban settings -Ciudad Juárez in Chihuahua state -around 230,000 Mexicans fled their homes between 2007 and 2010, approximately half of this total crossing the nearby border to the US. 8 Although somewhat preliminary and fragmented in nature, these data suggest that the scale of forced displacement in Mexico is significant. 3 Ibid. More surprising are the fragmentary quantitative data on the situation in the Northern Triangle. In the one country for which a relatively robust national survey exists -El Salvador -it is reported that 2.1 per cent of persons changed their place of residence in just one year (2012) as a result of threats. 9 Almost onethird of these had displaced two or more times. 10 Scaled up to the population living in the country in that year, 11 it implies that around 130,000 persons were displaced one or more times within El Salvador owing to threats during 2012 alone. No data exist to shed light on whether this astonishing scale of internal displacement is an anomaly for that year or whether El Salvador is an anomaly in comparison with neighbouring countries. Nonetheless, the fact that the high homicide rates for Guatemala and Honduras are broadly comparable to those in El Salvador may suggest that such forced internal displacement is taking place on a similar scale across the Northern Triangle. 12 The estimated scale of forced internal displacement in these countrieswhether seen as an absolute figure or an annual rate proportional to the national population -is thus highly significant. Yet, in comparison with other situations of acute violent displacement in the Americas, 13 our understanding of the wave of displacement in Mesoamerica is decidedly elementary. Indeed, alongside the few quantitative studies described above, only one substantial qualitative study of the topic exists. 14 Undertaken by a local civil society organization and funded by the United Nations High Commissioner for Refugees (UNHCR), it highlights a lack of state attention to the problem and has forced the issue more firmly onto the agenda of governments in the region. Overall, however, the existing literature says little about the nature or pattern of this displacement.
Against this backdrop, the present article seeks to more fully and systematically analyse the forced displacement provoked by organized criminal groups in Mesoamerica. The research is based on fieldwork by the author in Panama, Costa Rica persons, mostly knowledgeable local informants or displaced persons, which is triangulated here with other sources. Methodological challenges of access to perpetrators and displaced persons were heightened by the insecurity in some field locations and distrust of institutions on the part of many interviewees. Measures to ensure the safe storage of data and anonymity of interviewees were taken and are thus reflected throughout this article.

Criminal groups as agents of displacement
Existing studies treat the new wave of displacement in the Northern Triangle and Mexico as if it were a relatively undifferentiated phenomenon with a single unitary cause. Some refer to "generalized criminal violence", 15 whereas others refer specifically to "Organised Crime" understood as: a whole structure organised and coordinated within itself, which includes drug-trafficking networks, gangs, maras and criminal groups that operate from the local to the transnational level. 16 Regardless of the terminology, the impression is that organized crime is essentially monolithic. The present study departs sharply from this approach and instead takes as its starting point the significant body of existing literature dedicated to analysing the differences that do in fact exist between the various organized criminal structures present in the region. 17 The fieldwork that forms the basis for this study suggests that the new wave of forced displacement is produced principally by three types of organized group: street gangs, transporters, and cartels. 18 By describing the criminal activities of each, along with their structure and modus operandi, the present section lays the groundwork for consideration of how such differences impact on the form and patterns of displacement generated. This focus on the agency of the criminal groups themselves also permits the study to describe more completely how other relevant actors -including displaced persons and state officials -situate themselves in relation to the practices of these armed organizations.

"Mara" street gangs
Street gangs (pandillas) exist across Latin America in a wide variety of forms. 19 Yet in few places is their impact on public security as great as in the Northern Triangle. Particularly in El Salvador and Honduras, and to a lesser extent in Guatemala, 20 violent Californian-style gangs affiliated with the rival maras of Barrio-18 (B-18) and Mara Salvatrucha-13 (MS-13) have come to dominate the gang scene and related criminal activities. 21 This dominance and violence, 22 especially in the last two years, 23 makes them an important case study of how street gangs provoke displacement. However, although this analysis focuses on the maras, it should be recognized that other gangs also cause displacement, as with the powerful gangs in Mexico that are closely linked to the drug-trafficking cartels. 24 As a consequence, it cannot be assumed that all aspects of the present analysis will necessarily be directly applicable to other such types of gangs.
Originating in California, the B-18 and MS-13 maras are now entrenched across the countries of the Northern Triangle of Central America. 25 Yet each mara (whether B-18 or MS-13) operates separately in each country and there is no formal hierarchy between the maras of different countries. 26 Moreover, although each mara is structured at the national level principally via a prisonbased "council" of leaders, this provides only broad direction to the local clikas (cliques) on the street. Its role is also often complicated by dissent and internal rivalries. 27 In reality, therefore, each clika operates with a large degree of autonomy in its own barrio or colonia. The implication if we wish to understand 19 For example, even just in one country such as Honduras, a 2006 study identified at least six different types of street gangs: Californian-style gangs; satellites of Californian-style gangs; traditional independent gangs; "rich kid" gangs; professional mafia gangs; and school gangs (T. Andino Mencía, Las maras en la sombra: ensayo de actualización del fenomeno pandillero en Honduras, Tegucigalpa, UCA, 2006, 9). 20 Interview 60. 21  October 2011 by C. Martínez and J.L. Sanz under the heading "El Barrio Roto". The truce between B-18 and MS-13 in El Salvador also appears to have generated some degree of fragmentation and disorder within those maras, such that there is less control and discipline as certain clikas choose not to comply (Interviews 40, 42).
such localised phenomena as forced displacement by maras is that analysis should focus principally at the level of the clika.
The maras operate principally in urban and sub-urban areas. 28 Whether composed of a dozen or several hundred members, each clika seeks to exercise exclusive control over a specific and well-demarcated territory that it defends against incursions by rivals. For the purpose of understanding displacement, each territory can be divided into a clika's "core" and "extended" zones of operation. 29 The "core" zone is ordinarily located in one or more of the marginal and poorer neighbourhoods of the major urban conurbations. Members of the clika will live here and the zone also functions as a base from which it carries out localised criminal activities such as drug-dealing. Especially in the last few years, these are the zones where extreme levels of violence have clustered. 30 The "extended" zone is one that the clika enters more sporadically to extort businesses and sometimes even residents. As well as some poorer neighbourhoods, the extended zone may also encompass less marginal and more middle-class ones with the presence of public or private security forces.
The maras' capacity to affect the lives of local inhabitants varies between these zones. In "core" zones, the clika's very survival depends upon a high level of support or acquiescence from the population that is achieved through tacit codes of conduct imposed on inhabitants. 31 Alongside a basic requirement that inhabitants "look, listen, and shut up", de facto curfews 32 and other more intrusive norms may also exist. 33 Observance of the rules is backed up by violence and sometimes other control mechanisms. 34 Local supporters -such as family members and halcones (lookouts) -act as additional eyes and ears for the clika. In contrast, in wealthier parts of the "extended" zones, the presence of public or private security forces prevents the maras influencing the lives of residents to quite the same degree. Nonetheless, extortion -the lifeblood of the maras -can 28 Interviews 35, 48. Nonetheless, specifically in the case of El Salvador, it is important to note that the maras have a marked operational presence in a number of semi-rural zones, reflecting in part the small size of the country's territory and the spread of the gang phenomenon there. 29 Interviews 24, 48. 30 Interviews 21, 22. 31 Until the early 2000s, it was not uncommon for clikas to dispense free "justice" within their community and win its sympathy. At least in Honduras, with the diffusion of extortion and greater integration of the maras into serious crime dynamics, protection of any subject not paying extortion money became less common (Interview 33). In some B-18 neighbourhoods in Honduras, though, there are clikas valued by the local population for providing "security" by killing common thieves and drug addicts (Interview 28). The maras also kill "independent" extortionists who operate in their territories, regardless of whether they themselves are extorting the population (Interview 43). 32 Interviews 23,28. 33 Common examples include tacit prohibitions on any act that shows a lack of respect to a marero, may imply contact with a rival clika or the authorities (Interviews 43, 57), or which prevents mareros from hiding in the person's house when the police or some other danger passes ("Hay medio millón de hondureños secuestrados por maras", La Prensa, 6 May 2013). These are in addition to internal rules governing membership and behaviour within the mara. 34 For instance, in mara "core" zones of Tegucigalpa, Honduras, local clikas have installed security cameras in some neighbourhoods, as well as checkpoints to which persons entering or leaving must justify themselves ("Con cámaras de seguridad y trancas se cuidan mareros en Honduras", La Prensa, 6 May 2013).
reach extraordinary levels even in these zones. 35 Such differences between the zones help to shape the resulting forms and patterns of displacement. Finally, it should be emphasised that the extent of violence varies between clikas. Although latitude for criminal action is greater in countries with a weaker security apparatus, B-18 is also generally recognized as more trigger-happy and unpredictable than the cold and calculating MS-13. 36 Less violence is applied where inhabitants of a "core" zone are family or are otherwise tolerant of the clika, or where individual mareros are "calmer" in character. 37 Moreover, confrontations with rival gangs, other criminal groups, or state security forces may not only trigger increased violent activity by the clika but also lead to changes in overall modus operandi, as happened in response to the mano dura operations of Northern Triangle governments. 38 Overall, though, the clikas of the present can be characterised as highly localised and largely urban criminal organizations that swiftly resort to the use of violence against inhabitants.

Central American " transporters"
Transporter organizations (transportistas) of varying sizes have long smuggled illicit goods across the porous borders of Central America. Increasingly, this has involved moving cocaine northwards on behalf of first Colombian and now Mexican cartels, 39 although a range of other goods are also smuggled. Each group is contracted independently by the goods owner to move the product through a specific part of one country, and the more established operators tend to be organized around a prominent local family. 40 The lucrative nature of the transnational trade affords the more powerful transportista groups a level of social and political influence in these regions and they tend to be relatively well-organized and disciplined. Forced displacement in these zones is thus rather different in form and scale to that of territories where the maras operate.
Whereas maras live from localised criminal activities in densely-populated zones, the transportista trade is better suited to more sparsely-populated 35 For instance, in poverty-stricken Honduras, where even priests report being extorted, it has been estimated that the maras receive 1.2 billion lempiras -approximately 60 million US dollars -each year in extortion money ("Extorsiones dejan al año L1,200 millones a mareros en Honduras", La Prensa, 7 May 2013). 36 This is the case for Guatemala (C. Martínez & J.L. Sanz, "Los dos caminos de las hermanas", El Faro, 13 Nov. 2012), El Salvador (Interview 43) and Honduras (Interview 28). 37 Interviews 28, 48. 38 The pressure exerted through security operations, and associated extrajudicial killings, contributed to a change among those mareros who did not leave the mara or "calm down" towards more covert forms of operating. This involved the adoption of more subtle forms of identification -rather than the highly visible tattooing, graffiti, clothing, and mode of behaviour of past years -as well as more selective recruitment, particularly focused on youths in schools, and economic diversification, including investment in licit businesses (see, for example, PNPRRS, "Situación de maras"). 39 See Dudley, "Drug Trafficking Organizations", 20-26. 40 For instance, in Guatemala, three traditional families have long dominated the transportista business -the Mendozas in Petén; the Lorenzanas in the central highlands and eastern border; and the Leones in Zacapa (ibid., 26). and peripheral parts of the country, such that their respective zones of operation are usually distinct. Moreover, unlike the criminal activities of the maras, the smuggling business has not historically depended on exclusive control of these extensive rural territories but instead on the ability to move through them unimpeded. Traditionally, a live-and-let-live attitude thus prevailed among transportistas. 41 A further distinction is that their relationship with the population of these territories is based -in the first instance at least -less on fear and violence and more on buying the tolerance of inhabitants and officials. 42 In these poor communities, working for the transportistas represents a scarce source of income and the latter can also be generous in their provision of material support for the community. 43 In consequence, the population of transportista areas of operation is somewhat less exposed to the extreme violence and extortion of the disorderly cities. 44 Indeed, maras and violent street gangs that independently set up operations in these outlying areas are often killed by powerful and well-armed transportista groups. 45 Forced displacement from these areas thus appears less pervasive and more targeted in form than in the mara zones. 46 Yet this relative "stability" should not be taken for granted, especially in light of the increasing activity of Mexican cartels in the Northern Triangle. For instance, the Zetas' aggressive campaign to control smuggling routes in Guatemala has disrupted the existing status quo among local transportistas. 47 For the moment, though, transportistas may still be characterised as comparatively more disciplined and less aggressive criminal groups that operate across broad swathes of the backwoods parts of these countries.

Mexican cartels
Drug-smuggling organizations also have a long history in Mexico. Traditionally, like Central American transportistas, the Mexican cartels were rooted in strategically-important areas of the country and led by particular local families. Yet, from the 1990s, a process of increasing fragmentation and militarisation has produced a new modus operandi in which each cartel seeks to establish exclusive control over territories through which drugs are trafficked (plazas), on which they then 41 See O. Martínez, "Guatemala se escribe con zeta", El Faro, 13 Mar. 2011. 42 Interviews 38, 60. 43 For instance, in parts of Guatemala, such groups fund the fiestas patronales (patron saint festivities) and even gift families a bag of food each month (Interview 67). Migration to such parts of Guatemala and Honduras exists because of their "artificial prosperity" and the quantity of work available (Interviews 38, 60). 44 For instance, transportistas do not usually practice extortion. However, some of the armed gangs linked to them have begun to do so for big businesses, but not for small businesses and residents like in the cities, in areas such as Olanche, Honduras (Interview 38). 45 Interview 67. 46 Interview 38. See further below. 47 For a recent analysis, see J. López, "Guatemala: La cambiante cara del narco", Plaza Pública, 18 Jul. 2013. levy a tax (piso). 48 As well as moving drugs through Mexican territory, these cartels have increasingly assumed a dominant regional role as drug owners and managers. 49 Many -especially the newer cartels -are also diversifying their interests in controlled territories to include extortion and charging piso on other local criminal activities. 50 This new mode of operations appears to have provoked forced displacement on a significant scale since the mid-2000s.
The wave of violence experienced in Mexico over the past decade results largely from disputes for the control of plazas by these ruthless and heavilyarmed criminal organizations. In affected parts of the country, much of the intense violent confrontation occurs outside the major cities, in the rural zones through which drug transportation takes place. Rural zones in states such as Sinaloa are also a focal point for armed dispute over the production of heroin, marijuana, and methamphetamines there. 51 However, the confrontations are not exclusively confined to rural areas but have increasingly extended to nearby cities, which provide attractive opportunities for diversifying into extortion and control of the local drug-dealing market. 52 In the last few years, disputes over control of drug-smuggling routes have also spread with the cartels to Mexico's southern neighbours. 53 The growing militarisation of the Mexican cartels has not only exacerbated their fragmentation, but also altered the way in which they interact with inhabitants of such territories. Most notably among the newer cartels, a bloody and uncompromising mind-set prevails in which intimidation and extreme spectacles of violence are used to control inhabitants (and officials) or to dominate new territories. 54 The deployment of such tactics has raised the stakes for other cartels, which have not hesitated to respond in kind. In urban areas, 55 violent Mexican street gangs are also sometimes employed by rival cartels as a means of waging war by proxy, thereby further fracturing the control and discipline of the cartels. 56 While the cartels' extensive territories are comparable to those of transportistas, their pursuit of exclusive territorial control via intimidation and extreme violence is thus more similar to the strategy now favoured by the maras. Yet their 48  Crime. 54 Dudley, "Emergence of Mexico's Small Armies". 55 The trend of using gangs as muscle is less apparent for cartels in the rural zones, where they instead favour connections with the municipal governments and police (Interview 94). 56 For an example, S. Dudley, "Juarez After the War", InSight Crime: Organized Crime in the Americas, 13 Feb. 2013.
power, resources, and positioning in the regional drug trade put their capacity for violence in a league far above that of other criminal organizations in the region.

Criminal groups: the present and the future
The foregoing analysis shows that the three different types of criminal groups should not simply be lumped together under the heading of "organized crime" if we are to properly understand the forms and patterns of forced displacement in the current regional context. Rather, each type of organization is distinct in terms not only of the zones where it operates but also the scale of its structure, the focus of its criminal activities and its modus operandi vis-à-vis inhabitants of those zones. Moreover, within each type of criminal group, it is important to appreciate that a degree of variety exists -as, for example, among the different kinds of street gangs or the different tendencies towards diversification of illicit activities and use of violence among the various Mexican cartels. Moreover, it is clear that the above analysis captures these forms of criminal organization only at a particular moment in time. None of the three forms is immutable but rather each evolves rapidly in response to wider contextual factors. These involve not only the changing patterns of alliance and enmity between clikas or among cartels -and between the different types of criminal groups -but also fundamental shifts in their strategies and structure. This raises the question of whether the particularly acute form of criminal violence that has afflicted the Northern Triangle and Mexico over the past few years is a merely temporary phenomenon, reflecting certain contemporary motors of instability, or whether it points instead to a more permanent change in how the maras and cartels operate. With this question in mind, we now turn to consider the forms of displacement generated by the criminal groups.

Forms and grounds of forced displacement
Despite research into the functioning of the three types of criminal groupmaras, transportistas and cartels -the present study constitutes a first attempt to systematically analyse how they have produced the forced displacement of other inhabitants. The starting point is to analyse the distinct root causes of the displacement generated by the presence and activities of these groups. Towards this end, the present section identifies the common "grounds" of displacement in this context and shows that they cannot be treated as equivalent in their character or implications. In this respect, a broad distinction is drawn here between two distinctive "forms" of displacement: those that result from the "everyday" activities of criminal groups; and those resulting from the periodic violent disputes between them.

" Everyday" displacement
We turn first to consider the constant trickle of forced displacement produced by the "everyday" activities of these criminal organizations. Yet this form of displacement is not uniform in either scale or root causes but rather reflects the different criminal activities and modus operandi of each group as they impact on the population where it operates. From the standpoint of the criminal group, the following broad grounds of displacement may be identified (even if there is some degree of conceptual overlap between them): betrayal or enmity; resistance; land appropriation; and insecurity.

Betrayal or enmity
All three types of criminal groups produce displacement as a result of their perception that the person is an enemy or traitor. Various factors may lead to the formation of such a judgment. Cooperation with the authorities is one example. Policemen and other investigators -especially if they are seen as zealous -may thus attract enmity of the criminal group, 57 as will other persons who denounce the group to the authorities. This includes criminal turncoats -such as the pecetas hated and hunted down by the maras 58 -and inhabitants who report crimes as either a victim or witness. 59 Paranoia about informants is such that the mere act of speaking to a policeman can arouse suspicions. Other examples involve persons suspected of betraying the group to rival criminal organizations or stealing from it. 60 The maras take loyalty to such an extreme that mareros or their partners who leave the mara without permission will be deemed traitors. 61 More aggressive clikas may even label some forms of "resistance" by local inhabitants as a betrayal in these terms. 62 The common factor is the consequence of such labelling, which usually amounts to a death sentence for the person concerned. 63 Forced displacement is not used here as a strategy to deal with the person since there is no interest on the part of the criminal groups in allowing him/her to live elsewhere. In many instances, the family of the person is also a target for vengeance or a means to exert pressure on the individual. Occasionally, this extends also to other associates or even the whole community. 64 63 The Mexican cartels (and, in the Northern Triangle, some mara cliques) are particularly swift in their recourse to violence. There is little flexibility or room for negotiation -if they have suspicions about a person, they will simply go ahead and kill him or her. There is no argument or investigation (Interview 97). 64 For instance, a case was reported in Olancho, Honduras, where one youth in a group had problems with a transportista group and all 12 members of the group were killed one-by-one (Interview 38). In urban El Salvador, there are cases where a violent clika's suspicion of an informant in their neighbourhood led them to say "we will finish with community" and unleashed series of killings as a result of which 20-30 families fled (Interview 48).
entirely a pre-emptive strategy by such persons to avoid this fate, although it may be preceded by attacks on the person and his or her family. These displacements exist in all of the zones where the criminal groups act. The overall flow of persons is not large and is constituted by atomised individuals or families who leave quickly and invisibly owing to their acute safety concerns both pre-and postflight. 65

Resistance
Resistance to the demands of criminal organizations also generates forced displacement. This ground of displacement depends intrinsically on the character of the activities carried out by each criminal group, such that it is more varied in nature than that based on enmity or betrayal. Nonetheless, the scenario of resistance to systematic extortion demands offers an important case in point. Whether called impuesto de guerra (war tax) as in Honduras, renta (rent) as in El Salvador, or piso (tax) as in Mexico, such extortion is paid to provide protection against harm by that criminal group, and often also against harm by other criminals. 66 In light of the above analysis, it will be apparent that this sort of demand and the resulting displacement is most common in the urban and sub-urban zones where the maras, other gangs, and some of the newer Mexican cartels operate, although it is not exclusive to them. The extensive extortion in these areas is directed principally towards businesses, particularly in the transport sector, and extends even to street vendors. 67 For wealthier targets, the kidnapping of family members is sometimes used as a means of ensuring the payment of extortion. In recent years in Honduras and Mexico, residents have also been systematically extorted, not only in the poorer areas that do not benefit from the presence of public or private security forces but also in some wealthier areas. 68 Even so, in mara "core" zones, the scope of residential extortion varies considerably. In some, it is not practised at all, 69 or it is used solely as a means to punish residents for perceived "disloyalty". 70 In 65 For instance, a youth from Guatemala City who fled the country owing to such threats by the maras was killed upon repatriation (Interview 65). 66 Interviews 33, 43, 85. 67 Interview 33. 68 In Honduras, extortion in the last two years has reached levels where even priests and school children are being extorted (Interview 33). Families try and live discreetly, e.g. refraining from painting the exterior of their walls, so that they do not become a target for extortion. More prosperous families avoid even letting people know where they live (Interview 28). See also A. Arce, "Gangs Extort Cash from Honduran Homeowners", Associated Press, 8 Aug. 2012. 69 In El Salvador, for example, businesses aside, there appeared to be little extortion of residents of the "core" zones where the maras lived (Interview 57). A similar pattern was reported for Guatemala (Interview 67). 70 For instance, in Honduras, a clika might begin to ask for extortion money from the inhabitants of a whole street of its core zone in which it lacked trust (Interview 28).
others, it is applied to all those with income from a job or overseas remittances. 71 Much depends on the character of the clika and the economic pressures that it faces at any particular point in time.
Of course, failure to pay extortion is not the only act that may be constructed as "resistance" by such groups. For instance, the code imposed by maras on "core" zone inhabitants seeks to eradicate alternative power structures and may thus designate a range of activities as signifying "resistance". 72 Much depends on the particular circumstances, but common acts include: boys refusing to join the clika; 73 girls rejecting the attentions of a marero; 74 refusing, arguing, or looking askance at a marero; 75 or attending a school in a zone controlled by the opposing mara. 76 The resulting form of retaliation also depends a great deal on the character of the local clika: whereas some treat such persons and even their families as "traitors", 77 others seem to leave the family alone after the individual "resister" has fled the zone. 78 A similar dynamic prevails with certain cartels operating in Mexico. 79 It is important to appreciate that a wider spectrum of "resistance" thus exists.
Again, forced displacement is not used as a strategy in this scenario. Extortion is an important source of revenue and the threat of violence is usually intended to secure payment rather than provoke displacement. 80 Indeed, these persons usually leave their homes quickly and quietly in order to avoid fatal retribution for the loss of funds "due" to the group. Displacement is thus pre-emptive and based on insight into the consequences of failing to pay, whether as a result of general knowledge or direct experience of escalating threats and attacks against the family, which is often used as leverage. The movement is also atomised -i.e., usually individuals and families rather than en masse. In the areas where extortion is practised, this form of forced displacement is much more 71 Interview 33. Despite the high levels of remittances received in Honduras, inhabitants of such zones avoid using them to buy clothes or other high-value items since it will make them a target for extortion (Interview 23  77 In these circumstances, this ground of displacement shades into the first ground of "betrayal and enmity" (see above). 78 Interview 35. 79 For instance, in the tierra caliente (hot-lands) area of Michoacán, extortion of businesses and subordination to the "rules" of the Caballeros Templarios is reportedly of a similar form (Interviews 85-90). There are also cases where rural populations flee the consequence their resistance to sowing drugs for one of the cartels (A. Nájar, "Los desplazados de la guerra contra narcotráfico en México", BBC Mundo, 19 Oct. 2012). 80 The gangs thus calculate the person's approximate income and make the extortion demands accordingly. The exception is where the gang wishes to take over the business. This reflects a trend in certain mara zonesparticularly those of M-13 -where the bankrupt business is taken over by the maras once the owner has fled and run as a legitimate concern with the income going to the clika (Interview 33). This appears to be part of the increasing sophistication of these groups. prevalent than that based upon enmity or betrayal, although the protection concerns post-flight appear less pressing.

Land appropriation
Increasing interest among criminal groups in the acquisition of lands has generated a rather different ground of displacement. This tendency is particularly pronounced in rural areas of Honduras and Guatemala, where transportistas force small and medium landowners to sell lands in zones strategic for cross-border smuggling. 81 In some cases a small fortune is offered and in others the offer is risible -yet any refusal to sell is met by the threat of violence. 82 In contrast, in Mexico, some cartels have also used violence in order to shift whole communities from rural lands in areas rich in natural resources or good for drug production. 83 An urban manifestation of this phenomenon has appeared in the last few years for some mara "core" zones in Honduras and El Salvador: whereas the maras have long used the houses abandoned by displaced families for their own ends, 84 certain clikas are starting to deliberately displace families in order to take over their strategically-located houses. 85 Some transfers are even formalised by a lawyer brought in by the gang. 86 This ground of displacement reflects wider trends towards increasing concentration of land ownership across Latin America. Common to all three scenarios is that forced displacement is used as a distinct tactic to deprive people of their properties, and violence is but secondary and subservient to that aim. The only real difference between the scenarios is that in the Northern Triangle this strategy appears to be used more subtly and operates at the level of individual families, whereas in Mexico entire communities are displaced wholesale through the direct application of violence. In all three cases, the criminal groups do not appear to maintain an interest in the displaced persons once they have gone. The exception is where the displaced resist removal from their lands, denounce the eviction or agitate for their return, in which cases, violent retaliation ensues. 83 Such cases appear particularly common in Sinaloa and Guerrero States (Interviews 94, 105). 84 These are used by the gang as infrastructure, escape routes and storage for drugs, weapons and kidnap victims (Interviews 24, 28, 33). 85 Interviews 24, 28, 58. 86 Interview 28. with the group. Rather, it flows from more diffuse fears for the future, such as the concern of an urban mother that her growing children may attract the attention of the local clika, or simply frustration with the increasing levels of crime and violence. Yet, even here, some specific event often serves as the trigger to displace: the arrival of a criminal group to the neighbourhood, the removal of a police station, or the witnessing of a crime may not pose an immediate risk but simply be the last straw in an already difficult context. This form of displacement is the most difficult to capture analytically, since it comprehends a range of more-or-less remote fears. Nonetheless, displacement on this ground is apparently widespread, pre-emptive in nature and tends to involve individuals and families rather than whole communities.

" Dispute" displacement
The second form of forced displacement is generated by the intense violent disputes that sometimes arise from the shifting patterns of cooperation and competition between criminal groups. Where these concern the control of territory, they often serve to exacerbate "everyday" grounds of forced displacement. Yet they also have the potential to generate their own grounds of forced displacement, owing to the manner in which inhabitants are integrated into the activities of these groups. Since the armed disputes between transportistas appear to have relatively little negative impact on the population, 87 the present section focuses upon analysing, first, gang warfare in the Northern Triangle and, secondly, conflict between the Mexican cartels.

Gang warfare in the Northern Triangle
Among the maras of the Northern Triangle, the territory of each clika is wellestablished and mareros caught in the territory of a rival can expect violent consequences. Even if the wholesale invasion of the territories of rival clikas is less common than might be expected, 88 this should not distract from the fact that episodes of acute confrontation between and within maras do erupt periodically in each of these countries. Such gang warfare occurs not only between the rival B-18 and MS-13 maras, but also between mara and non-mara gangs, 89 and even between clikas from a single mara movement. The participants in these 87 Nonetheless, it should be noted that there are occasional cases where one band arrives and attacks or threatens a few key families that support another group thereby causing their displacement, as reported for Olancho, Honduras (Interview 36) and Tecún Uman, Guatemala (Interview 67). 88 Such "invasions" may be relatively rare since the maras are focused on surviving and preventing others from muscling in on their own territories (Interviews 28, 33, 42, 48). Alongside the current B-18/MS-13 truce in El Salvador, more tacit national accords on this point are said to exist in other countries (for the example of Honduras, see "La guerra volvió a Chamelecón, las maras rompen la tregua", La Prensa, 7 May 2013). 89 For example, B-18 sometimes invaded and took over the barrios of non-mara gangs in Tegucigalpa (Interview 28).
disputes sometimes extend to include clikas from nearby countries, 90 as well as other criminal actors allying themselves with one or other side. Gang warfare often leads to an exacerbation of "everyday" displacement dynamics. This may take the form, for example, of a general increase in insecurity owing to greater violence, 91 or a "hardening" of mara attitudes towards the population, such that extortion quotas are raised or those who do not pay are killed immediately or after one warning rather than after the traditional three warnings. 92 Gang disputes also generate their own grounds of displacement. Indeed, it is one of the few scenarios in which an incumbent clika will directly order forced displacements, usually of any family believed to have sympathies with the invading gang. 93 In extreme cases, the incumbent clika may even order any inhabitants who do not have family members in the clika to leave or be killed. 94 However, where the incumbent clika is defeated, its family members and other local supporters have to flee the neighbourhood in order to avoid reprisals from the gang now controlling the zone. 95 Although the quantity of displacement in these scenarios may be elevated above "everyday" levels and entire rows of houses abandoned, the families displaced tend to be targeted on an individual rather than collective basis.

Cartel conflict in Mexico and Central America
Mexico is presently much more turbulent than Central America and patterns of forced displacement are considerably more marked by such conflictive dynamics. Much of this war between Mexican cartels revolves around the control of plazas and other strategic locations in areas which also tend to be the organizational headquarters or heartlands of one or other of the cartels. 96 The situation is complicated by the trend of cartel fragmentation that may weaken the command chain in affected cartels. 97 These dynamics are further heightened by the waging of war through proxies -both gangs and smaller cartels -and the rapidly shifting 90 As when Salvadorian MS-13 clikas dissenting from the truce in El Salvador arrived recently in Tegucigalpa invited by MS-13 in Honduras to invade the neighbourhoods of local non-mara gangs (Interview 28). 91 For instance, the shoot-outs between maras may be so frequent that people living in the area dispense with glass in their windows and use wooden boards instead (Interview 67). 92 As an example of the former, the expansion of a B-18 clika in Tegucigalpa, Honduras, began with extorting the houses at the entry to the neighbourhood as a means of determining loyalty and killing people if they did not pay (Interview 28). An example of the latter comes from Guatemala City, where open hostilities with a rival mara led to the "normal" process of extortion of inhabitants -involving three warnings -being circumvented such that those who did not pay were killed immediately (Interview 67). 93 See, "La guerra volvió a Chamelecón". 94 For an example of this, see D. Valencia Caravantes, "La legión de los desplazados", El Faro, 1 Oct. 2012. 95 The perception of collaboration can extend even beyond residents to include persons from outside the neighbourhood who entered merely to buy drugs from the defeated clika. Such persons will not be able to go back to the neighbourhood (Interview 28 alliances between groups. This can lead to confusing situations such as where an attack by one cartel on another in one part of the country is actually in response to fighting between other apparently unconnected groups elsewhere in Mexico. 98 The armed structures of the Mexican cartels are extensive and highly developed, with access to powerful weaponry and logistical support. As a result, the intense confrontations between them exacerbate the "everyday" grounds of displacement. On the one hand, the gun-battles between cartels -especially in more remote rural zones -dramatically increase insecurity for inhabitants, causing massive displacements of whole villages or sectors. For instance, in the rural tierra caliente (hot-lands) of Michoacán state, the entry of armed elements from the Cartel Nueva Generación Jalisco to do battle with the Caballeros Templarios cartel recently caused mass displacements. 99 On the other hand, fighting may also further fragment cartel control over their local armed proxies, thus leading to increased and more unpredictable forms of extortion for inhabitants. 100 In turn, these latter dynamics provoke additional displacements by individuals and families.
It is notable that the cartels also employ forced displacement as a specific strategy during territorial disputes. Mass displacements of whole communities or sectors are used as a means of undercutting local sympathy -and thus logistical and intelligence support -for the rival and capturing it for the invading cartel. This can be seen in the forced displacement of entire villages over the past two years across the rural sierra (mountain) zone of Sinaloa state -the cradle of the Sinaloa cartel -by invading armed groups linked to the Beltrán Leyva cartel. 101 The Zetas have also used this strategy to displace whole towns assumed to support the Gulf Cartel along the drug-trafficking corridor running through Nuevo Leon state on the US border. 102 In the neighbouring state of Tamaulipas, Gulf Cartel action to "liberate" certain towns from the Zetas by taking away individuals suspected of links with their rivals led to a substantial number of families fleeing their homes. 103 A similar form of displacement occurred in parts of Ciudad Juarez, Chihuahua state, when the armed structures of the rival cartels began to carry off local inhabitants. 104 Finally, the increasing penetration of the Northern Triangle by Mexican cartels raises the probability that these strategies of violence and displacement will be exported to disputes in the rural areas of those countries. Traditionally, violent territorial disputes between transportistas have been rare. However, the presence of Mexican cartels is increasingly generating negative effects for local 98 Interview 99. populations. For instance, the Zetas have tried to secure exclusive control of territories in Guatemala through violence towards populations in Péten and Alto Verapaz and inhabitants of certain towns in Huehuetenango have been threatened with death if they do not collaborate, thereby generating a real risk of heightened displacement from these rural areas. 105

Conclusions
The various interests and strategies of the different types of criminal groups impact on the lives of other inhabitants and structure the forms and grounds of forced displacement canvassed in this section. A common factor is that both the "everyday" and the "dispute" forms of displacement provoke population movements that are essentially reactive to the violence used by these groups in pursuing certain activities. Yet the latter scenario is distinguished in part by the criminal groups' strategic use of forced displacement in a manner similar to that seen in war zones such as Colombia. 106 It is important to separate out these different forms and grounds of forced displacement generated by the criminal groups, since this variation influences the resulting patterns of displacement, as well as the protection profile of those displaced.

Resulting patterns of displacement
The relationship between criminal organizations and the populations of the territories in which they operate can be complex, varying according to the former's core activities and the manner and extent to which they use violence. Thus, at one extreme and particularly where the criminal group is oppressive and not open to dialogue, individuals and even whole communities fight fire with fire, organising and arming themselves to fight against the criminal groups, 107 or approaching death squads that kill suspected criminals for a fee. 108 Likewise, in the face of excessive extortion, whole communities have allied themselves with a rival criminal group to get rid of the extortionists. 109 State security forces may also take action against the criminals, especially if the victim is wealthy or influential. 110 The agency of inhabitants in responding to the violence of criminal groups thus in responding to the violence of criminal groups thus should not be underestimated, including in its potential to take violent and confrontational forms. Yet such agency equally often manifests itself in the decision by inhabitants to flee their homes. The present section builds upon our understanding of the distinct forms of displacement outlined above in order to examine the patterns of population movement that result from the exercise of this agency by those fleeing the criminal groups. Such agency is shown not only in the decision to flee, and how to leave, but also in the choice of where to go and what to do. In this regard, it will be shown that such decisions are hedged by wider sets of opportunities and constraints that reflect not only the distinct forms of displacement identified above but also empirical differences between the affected populations in terms of social position and identity.
The present section aims to show how these differences between affected populations generate different displacement streams, within which a variety of profiles are in turn encompassed. In order to do so, it focuses analytical attention on the "hotspots" of forced displacement in Mexico and the Northern Triangle. This labelling draws attention to the fact that not all parts of these countries are equally affected by this wave of forcible displacement. Rather, the exodus takes place principally from those areas of the country where organized criminal groups wield the greatest influence. This suggests the need for a more nuanced, localised approach to understanding the resulting displacement patterns.
The forced displacement hotspots of the Northern Triangle countries and Mexico where one of the three types organized criminal groups operates can be differentiated in broad terms as follows: (1) poorer urban areas -the analysis focuses upon the mara "core" zones in the Northern Triangle, although a similar analysis might be applicable to some gang territories in Mexico; (2) wealthier urban areas -these include mara "extended" zones in the Northern Triangle and some cartel and gang territories in Mexico; and (3) rural areas -these are transportista zones of operation in the Northern Triangle and cartel trafficking and drug-production territories in Mexico. Addressing each in turn, the analysis seeks to illuminate not only the distinct displacement patterns from these areas but also how they link into other patterns of internal and international movement.

Poorer urban areas -mara core zones
The violence in the Northern Triangle is concentrated most acutely in poorer colonias and barrios (neighbourhoods) that fall within the "core" zones of the territories claimed by maras. As a result, particularly over the past two years, these marginal urban and sub-urban areas have come to represent one of the 110 Interview 48.
principal hotspots of forced displacement in these countries. Anecdotally, the scale of displacement seems at its highest in these areas, even if there is some variation between neighbourhoods depending on the circumstances of the local clika. 111 The displacement pattern from these territories appears predominantly urban-urban in character. 112 Indeed, relatively few reports exist of inhabitants from these mara core zones fleeing to rural areas. 113 Nonetheless, escape to the rural areas may be an option for those with strong subsisting connections to the countryside 114 and is reportedly more common in Guatemala among persons with strong ties to an indigenous community. 115 In general, therefore, the tendency is that displaced persons from these poorer urban areas go to other urban areas, often in the same city but also in other cities of the country, i.e., intra-and inter-urban internal displacement. This displacement stream is comprised both of families from these zones who have had to leave together and of individual youths who have been sent away by their family to other parts of the country. 116 Displaced persons largely aim for a similar class of neighbourhood where they will be able to continue their usual economic activities. 117 Families with income from employment commonly try to rent a house in a different neighbourhood of the same city. 118 The determining factor here is continued access to their source of employment, even if higher rents in non-mara colonias often mean the family must adjust accordingly. 119 In contrast, poorer families or individuals without the resources to rent a new home stay with family in the same city or a different one, even though conditions may be very crowded. 120 At least initially -until they can rent a house -urban family links thus influence their location of flight. However, those persons without resources or family support can end up living in vulnerable and degrading conditions on the street or in distant squatter zones. 121 For them, contacts among networks of displaced 111 Knowledgeable informants estimated that an average of some 5 to 10 per cent of the total number of families in a relatively violent mara core zone could be forcibly displaced each year as a result of the activities of the group (Interviews 48, 60). Some such houses remained abandoned, others are sold or rented by their former owners, and the remainder are occupied by the gang or by newcomers to the neighbourhood. 112 Interviews 28,33,40,42,57,60,65,67. 113 Interviews 21, 57. However, there are reports indicating that those fleeing justice, including mareros, sometimes fled to rural areas (Interviews 33, 38). 114 Aside from any reporting bias, this may reflect the fact that rural work is not plentiful and, in any event, the families of these zones may have been living in urban zones for a generation or two. Not only are these urban-dwellers -particularly the youth -not accustomed to the rural lifestyle or working in rural industries but connections with family in those zones may be now very attenuated, especially following the protracted civil disturbances in the 1980s which hastened the dispersal and atomisation of many families (Interview 57 former neighbours appear to be a factor in determining where they go as well as in accessing basic living opportunities there. 122 Overall, the often abrupt forcible displacement of persons from these poorer neighbourhoods puts an additional strain on their existing means of subsistence, especially if they have displaced on grounds disclosing an immediate threat and not had time to collect their belongings. Displaced families that own the houses that they left behind often try to recoup some of their value by selling them or renting them out. As a result, many houses advertised for sale in the local newspapers of Northern Triangle cities are located in the colonias where the maras operate and rows of abandoned houses can be seen in some of these marginal urban zones. 123 Families displaced within the same city also try to keep informed about the situation in their old neighbourhood, with a view to returning if the gang becomes inactive. 124 Some displaced families are even able to return to their homes after a period of time so long as the grounds for their displacement did not involve direct or serious problems with the incumbent clika. 125 For those who have no option but to move to neighbourhoods where another mara clique operates, these subsistence challenges are compounded by the potential for renewed security problems, such as efforts by the resident clika to recruit youths. Youths with gang tattoos are particularly vulnerable: the clique in the new area rapidly investigates them and will kill them if they are deserters from their own mara or were members of the rival mara. 126 The clika from which the person has fled also remains a threat and its members may seek to locate and pursue the displaced across the city if they are seen as having seriously resisted its authority. In the extreme case of perceived traitors or enemies -such as witnesses, former mareros or former partners of mareros -such pursuit can even extend between cities. 127 In such circumstances, displaced persons may cut all ties with their original neighbourhood in order to try and minimise the possibilities of being traced by their persecutors. 128 The extent to which such displaced persons end up leaving their country is questionable. In general, the trend appears largely to be one of internal displacement. 129 Nonetheless, there are exceptions. On the one hand, the vehemence with which traitors and enemies are pursued by the maras pushes some to leave the country. Often they seem to go to neighbouring or nearby countries of Mesoamerica, especially if they have family members there. 130  immediacy of the threat for this class of persons are such that ex-mareros repatriated to Honduras have been killed within days of their return. 131 On the other hand, youths at risk from gangs seem more usually to be sent north to join family in the US. 132 Sometimes they go with the help of a coyote (agent) if the family is able to raise the many thousands of dollars that this implies. 133 Otherwise, and particularly in the case of resource-strapped Honduras, youths make this dangerous journey across Mexico alone. 134 However, some cases also exist of youths relocating within Mesoamerica accompanied by their whole family. 135

Wealthier urban areas -Northern Triangle and Mexico
The other main urban displacement hotspots are comprised by those more prosperous areas that, in Northern Triangle countries, fall within the mara "extended" zones of operation or, in Mexico, within which gangs and newer cartels operate. The grounds of displacement in these urban areas tend to revolve around extortion of businessmen and inhabitants' perceptions of insecurity arising from the operations of the criminal groups. The character of the displacement produced, as with the poorer urban areas, is determinedly urban-urban. However, the greater resources available to these middleand upper-class families give these population movements several distinctive features.
A proportion of this displacement is certainly internal in character. In the Northern Triangle, the tendency seems to be for those wealthier families who remain in the country to rent a house in another part of the same city rather than changing cities. 136 These are families that prefer to remain close to their existing social networks and sources of income and do not perceive a great risk of being pursued across the city. Moreover, the challenging economic circumstances in the countries of the Northern Triangle may mean that relocating to another city is simply not seen as an attractive option by such families, especially when considered in the context of their greater ability to leave the country should circumstances push them to flee the home city.
In Mexico, in contrast, there is a considerable degree of movement between Federal states. Alongside relocation to the capital, Mexico City, the displaced also move to urban areas of other states currently less affected by intense cartel violence, including southern states such as Chiapas. 137  locations of "out-of-towners" from violent northern states has become notable in the last few years. 138 However, these migratory flows also include natives of these states who migrated away in years past and are now returning to their "safe" state as a result of the violence elsewhere. 139 The much greater expanse of territory that Mexico covers -in comparison to the Northern Triangle countries -and its more vibrant economy appear to facilitate such internal migrations between cities. Where the extortion is generalised in nature, such as in urban Michoacán, some of these displaced are able to return to their homes without problems after being away for a while. 140 This displacement stream also possesses a distinct and important international component, owing to the greater means available to these families. It is expressed principally through relocation to the US, although other destinations exist, and encompasses both formal and informal channels for migration. Even if such forced displacement takes place across all of the wealthier urban areas of Mesoamerica where organized crimes groups act, it seems particularly pronounced in northern Mexican cities. 141 During the peak of the violence in states such as Guerrero and Chihuahua, many wealthier Mexican families or those with dual nationality moved to the other side of the frontier. Some took their businesses, while others commuted to work in Mexico during the day and returned to the safety of the US at night. 142 In 2012, with the decrease in overt violence in Chihuahua, some of these families began to return to live again on the Mexican side of the border. 143

Rural areas -Mexico
Forced displacement in Mesoamerica is not only an urban phenomenon. Indeed, one of the most important displacement hotspots is comprised by those rural and semi-rural areas of Mexico where inhabitants suffer as a result of a violent dispute between two or more cartels or due to efforts by one or other cartel to appropriate their lands. This is one of the few contexts in the region where forcible displacement sometimes takes on an en masse character alongside the more persistent drop-by-drop flight of individuals and families. 144 Although forced displacement in these areas may take various forms, substantial similarities are apparent in the resulting migration patterns.
Persons displaced from these zones are mostly farmers, labourers, and small businessmen and women affected by the dynamics of cartel violence. At least in the first instance, they tend to move towards nearby towns and population centres. 145 However, secondary movements by some displaced persons also take place towards bigger cities -even outside the region -where the presence of family members and a potentially wider range of employment opportunities may facilitate reestablishment. 146 Aside from in certain states along the US border, 147 there are relatively few reports of these displaced persons leaving Mexico. In general, the movement is thus from rural to increasingly urban contexts and concentrated predominantly in the region of origin.
Moreover, particularly in relation to forced displacements with an en masse character, return to the home area by displaced persons may begin to take place relatively quickly if some level of calm is re-established in the zone. 148 This sometimes gives rise to the perception on the part of the authorities that such displacement has only a temporary character. However, it is important to emphasise that not all of the displaced return following such episodes of displacement, leading to a gradual emptying out of the countryside. Among those who do return, and based on the initial grounds of displacement, renewed threats may present themselves. For instance, where displacement has been used as a strategic tool by one cartel, returns have been met with violent retaliation, thereby dis-incentivising further returns. 149 In some mass displacements from rural and semi-rural areas of Mexico, and due to the perception of their temporary nature, local authorities have provided accommodation to the displaced in the reception location. 150 Nonetheless, especially where return is dangerous, it appears that many of the rural displacedwhether displaced individually or in mass events -end up living in marginal zones of the towns and cities of the region. 151 Some stay with family members and a few rent a property if they have the resources. 152 Interviews 86,88,90,101. 152 In some towns close to rural zones affected by these grounds for displacement, the high level of demand among displaced persons for rental properties has pushed up rent costs exponentially (J. Estrada, "Tamaulipas: un albergue, la pequeña ciudad de los desplazados por el narco", CNN México, 13 Nov. 2010).
for informal work, while trying to maintain a low profile for fear of their persecutors. 153

Rural areas -Northern Triangle
Forced displacement also occurs from rural areas of the countries of the Northern Triangle, particularly Honduras and Guatemala. However, not only is the scale less striking but the movement patterns reflect two distinct root causes or grounds of displacement. First, there is a small stream of individuals who leave the rural areas -often taking their families with them -owing to the fact that they are considered as traitors or enemies by one of the transportista groups. These displaced persons usually migrate to the capital cities where they hope either to access state protection or to lose themselves among the multitude. 154 The movement pattern is thus directly rural-urban. However, owing to the continuing risk of assassination, most of these internally displaced persons are keen to rapidly leave the country and travel to some safe country -usually the US -although not all have the means to do so. 155 Secondly, a distinct dynamic of relatively localised rural-rural displacement is produced as a result of the expropriation of local lands by transportista groups. Peasants and other small landowners displaced on this ground move to the fringes of larger towns in the region or invade other rural lands -such as forestry reserves -and seek to earn a living as local labourers. 156 Yet even those who were able to sell their lands to the criminal groups for a reasonable price find that the dubious provenance of the cash sums that they have been paid cannot be channelled easily through the formal financial system, thereby complicating the process of buying a new home etc. 157 Unless these displaced persons seek to return to their lands or denounce the case to the state authorities, they appear to be left alone by the transportistas following displacement and not molested further.

Conclusions
The forced displacement hotspots of Mesoamerica are located across a range of urban and rural contexts. The resulting displacement patterns are also relatively differentiated, reflecting differences in the grounds of displacement generated by the various criminal groups and the type of population affected. The protection and subsistence threats and needs faced by the displaced persons are equally variable. It is also clear that these displacement streams blend into other patterns of internal and international population movement. For instance, within the 153  rural areas of Northern Triangles countries, small-holders displaced by transportistas for their lands mix in with others dispossessed from the same territories by multinationals that accumulate lands for mega-crop exploitation rather than smuggling. 158 Similarly, it would seem that the patterns of emigration from these same countries for work and/or family reunification in neighbouring States, Mexico and the US may -over the last two years -increasingly include persons whose migration is motivated, at least in part, by organized criminal violence. 159 In this context, it is important to end by observing that the number of Northern Triangle citizens claiming asylum in Central America or Mexico remains low (albeit more significant as a percentage of the small numbers of overall claims for asylum received in these countries). 160 As an absolute figure, the number of claims for asylum lodged by Northern Triangle nationals in the US and Canada is substantially higher, albeit low in relation to the apparent scale of internal displacement within the Northern Triangle countries. 161 These facts may support the contention that the displacement is predominantly internal in character or it may imply that those who flee the country due to criminal violence rarely apply for asylum. 162 Even if numbers of asylumseekers may thus represent a poor source for estimating the scale of external displacement, the data from Mexico and Central America do illustrate two important points. First, there has in fact been a substantial proportional increase in the number of such claims over the last two-three years. 163 Secondly, the general profile of the applicants has shifted dramatically in this period from 158 Martínez, "Ser un nadie en tierra de narcos". persons connected to the maras to whole families of their victims. 164 These trends broadly mirror those detected in respect of forced internal displacement.

State responses
We turn finally to the State and its response to this crisis. In general, it is almost a truism that the degree of latitude with which organized criminal groups are able to operate in Mesoamerica reflects the relative weakness and, in some cases, infiltration of governmental institutions in the region. The impact of recent hard-line policies on organized crime -such as the various permutations of mano dura (firm hand) in the Northern Triangle countries and the War on Drugs in Mexico -is well documented. At times, the implementation of these policies has served to displace the criminal groups from particular locations. 165 However, equally often, it seems to result merely in a re-ordering of the criminal underworld or a shift in modus operandi by the criminal group. 166 Of course, such state operations produce their own dynamics of forced displacement. 167 The humanitarian aspects of the present situation receive comparatively little attention by Mesoamerican governments, especially outside the penal sphere. As a result, it is hardly surprising that, at the time of fieldwork in 2013, not one single Mesoamerican government had yet formulated a coherent national policy to respond to the epidemic of forced displacement. 168 Even in Mexico, where the government has demonstrated increasing interest in attending to the victims of organized criminal groups within a broader reparations framework, 169 the issue of forced displacement and its humanitarian consequences 164 Interviews 11, 66, 69, 70. 165 One example from Mexico may be that of the former "Town of Death" San Fernando, Tamaulipas (G. Moore, "Ending the Zetas Killing Spree: An Invisible Success Story", InSight Crime: Organized Crime in the Americas, 22 Sep. 2011). 166 The returning calm in Juárez appears, in part, to be an example of the former (Dudley, "Juarez After the War"), whereas the shift of the maras towards a less visible modus operandi is an example of the latter (see section above). 167 Detailed consideration of displacement produced as a result of the execution of these policies falls outside the scope of this article. Nonetheless, to give but one example from the Northern Triangle countries, it is very clear that mano dura State operations against the maras -and extralegal activities carried out by State agents -have resulted in the displacement of mareros from these zones (Interview 28). It would appear that the families of such persons, as well as other youths living in these zones but not connected to the maras, are also sometimes caught up in this pursuit. 168 Subsequently, Honduras has taken the first step in establishing a Comisión Interinstitucional para la Protección de las Personas Desplazdas por la Violencia (Inter-Institutional Commission for the Protection of Person Displaced by Violence) through a decree adopted in Nov. 2013 ("Comisión atenderá a personas desplazadas por la violencia", La Tribuna, 6 Nov. 2013). 169 See, for instance, the Ley general de víctimas (General Law on Victims) adopted by Mexico on 9 Jan. 2013 and amended on 3 May 2013.
remain largely ignored. 170 In part, this may reflect the fact that public attention to the humanitarian aspects of the present criminal violence in Mexico remains focused exclusively on murder and forced disappearance. 171 At present, the field of criminal law -and witness protection programmes in particular -represents the main legal or policy framework within which the authorities respond to forced displacement. These witness protection programmes generally provide for relocation of at-risk witnesses and some timeand resource-limited form of economic support. 172 However, the relevant laws usually require formal denunciation of a criminal act by the victim as a condition for entry into such programmes. 173 Although this requirement is sometimes waived in practice as a result of the lack of any other available protection framework for displaced persons, the principle remains problematic since many displaced persons are not victims of a cognisable crime. Moreover, the act of denunciation immediately risks converting the person -even if initially displaced on some other ground -into an "enemy" to be actively pursued by the criminal group. Finally, the infiltration of such programmes by criminal groups and other institutional deficiencies in practice render them largely unsuitable and even dangerous as a general framework for the protection of displaced persons. 174 Another form of engagement by government authorities with the humanitarian needs presented by forced displacements is the opening of hostels or other local accommodation in territories that have become a destination for displaced persons. 175 No examples of this practice were encountered in the comparatively resource-poor and institution-weak countries of the Northern Triangle. However, there are various cases in Mexico. 176 It appears to be used principally in response to episodes of mass displacement of whole communities rather than the arrival of individual families, even where the resulting number may be high. In some cases, the authorities have provided their own economic support and even attempted to help resettle some displaced persons. 177 However, this measure is taken exclusively by local or state authorities rather than national ones. In some cases, such support is reportedly provided because the local administration is infiltrated by a cartel that wishes to ensure attention to members of its social base who have been forcibly displaced by a rival cartel. 178 that the study will prove useful also in serving to push discussion of the new wave of forced displacement afflicting Mesoamerica further onto the public agenda.
From the preliminary data presently available, the scale of forced internal displacement in Mesoamerica across the past few years seems unprecedented. In the case of El Salvador, the annual rate of forced internal displacement as a proportion of that country's population (2.1 per cent) is of the same magnitude as those during the conflict of the 1980s or even contemporary armed conflicts such as Colombia, widely considered to be "one of the world's most dramatic humanitarian emergencies". 184 Nonetheless, it is apparent that such exceptional levels of displacement are a relatively new development of the past two-three years in the Northern Triangle countries and slightly longer in Mexico. Given that the creation of the organized criminal groups far predates the epidemic of forced displacement, the latter seems rather to reflect changes in the nature of these groups in recent times.
With this in mind, the study took as its starting point an explicit focus upon the agency of three of the main types of organized criminal groups in the region. In so doing, it drew attention to the limits of treating the category of "organized crime" as monolithic, unitary, or unchanging through time. Instead, it pointed to important differences between -and within -various types of organized crime groups presently operating in Mesoamerica in terms of structure, criminal interests, zones of operation, and modus operandi. Even if all of these groups tend to avoid direct armed confrontations with the State's armed forces, the ways in which they engage with, or impose upon, local populations in their zones of operation reflects such differences in their contemporary make up.
The study also makes clear that the current wave of forced displacement is not a phenomenon that equally affects all parts of these countries. In general, the phenomenon is concentrated in certain "hotspot" areas where the criminal groups operate. However, the differences between these criminal groups not only contribute to giving the phenomenon of forced displacement varying dimensions and forms in these territories but also help shape the resulting patterns of movement by affected persons. A useful analytical distinction in this regard can be drawn between the forms of displacement resulting from "everyday" criminal activities and those arising as a result of sporadic violent disputes between criminal groups.
This approach allows us to ask whether the forced displacement currently afflicting Mexico is in fact an anomaly resulting from a long-running but ultimately exceptional violent free-for all between cartels. If so, then the intensity 184  of the phenomenon should reduce if and when the Mexican underworld reaches a general accommodation. Even in this scenario of where some "normal" criminal order is re-established, less visible forms of violence would likely continue as would the ensuing displacement on "everyday" grounds. The other option for the future is that such violent confrontation actually represents the new modus operandi of Mexican cartels, in which case the current forms of displacement will persist until the nature of these groups again shifts. A similar question might also be posed for the maras in the Northern Triangle: is their excessive extortion and brutalising of some populations in recent years -and the apparently high levels of resulting displacement -the result of an exceptional set of circumstances or do they represent a relatively stable new modus operandi for these gangs? Another important aspect of the current wave of forced displacement in Mesoamerica highlighted by the study is that it appears to be predominantly internal in character. Certainly, there seems to be a large amount of displacement within the borders of these countries. Nonetheless, further academic work is required in order to quantify and map not only the internal aspect of these migrations but also the extent to which they have an external component not captured by the numbers of claims for asylum registered. In this regard, the willingness of certain Mesoamerican governments to facilitate claims for asylum outside the country by their co-nationals is also a noteworthy development in international practice.
Alongside whatever penal or military measures are used by such States to confront organized crime on their own territories, specific policy is also required from the governments of Mexico and the Northern Triangle countries to respond to the consequences of the internal displacement generated by criminal groups. This is not to suggest that such a displacement policy is required in all countries of the world or even Latin America where organized crime exists or even where it results in sporadic or isolated episodes of forced displacement. Nonetheless, the scale and intensity of forced displacement at the hands of organized criminal groups in Mesoamerica leaves little doubt of the need for a humanitarian policy framework to address the consequences of this present wave.
In designing a policy to respond to the needs of persons forcibly displaced as a result of organized criminal groups in Mesoamerica, the humanitarian character of such a response should be foremost. In other words, the criminal character of the agents of displacement should not distract attention from the humanitarian nature of the needs of their victims. For instance, requiring that victims of displacement at the hands of criminal groups formally denounce a crime before they receive government assistance for the humanitarian needs is both impractical and, in the context of infiltration of some state structures, a factor that may expose the victims to unnecessary risk.
For formulating a humanitarian policy for displaced persons, a differentiated approach may also be useful to the extent that this study has shown the existence of various profiles of displaced person with differing sets of needs. For example, the position and needs of a wealthy urbanite changing home due to perceived insecurity are distinct from those of an impoverished youth fleeing from one marginal gang-infested neighbourhood to another. Moreover, in the context of the severe resource scarcity of certain States in the region, the reality is that it may be necessary to direct priority attention to the most vulnerable first, at least at the outset.
A question implicit throughout the study has been the degree of equivalence between the new scenarios and their displacement antecedents during conflict in the region. This has implications for humanitarian practice as well as for academic comparisons. On the one hand, the modes of social control applied by some criminal groups to local populations seem similar to those used by parties to the civil wars in Latin America, such as Colombia. 185 On the other hand, there seem to be fewer points for political or humanitarian interlocution with the criminal groups and their swift recourse to killing rather than displacement as a modus operandi further complicates such efforts. 186 In general, it is thus an open question whether humanitarian frameworks and practices developed in conflict scenarios like Colombia could be applied usefully to the new forms of forced displacement in Mesoamerica. 187 Finally, the national authorities of Mesoamerican countries should not expect to have to deal with this new wave of forced displacement alone. Indeed, as we approach the 30th anniversary of the 1984 Cartagena Declaration on Refugees, 188 it is appropriate to ask whether a similar regional initiative could benefit the victims of the new wave of displacement in the region. In addition to helping to develop national policy and international protection understandings sensitive to the new situation, such an initiative might usefully consider whether certain profiles of at-risk displaced Mesoamericans could be resettled elsewhere on the continent. Governments in the Americas have the opportunity again to continue their trail-blazing example in developing frameworks for the protection of refugees and other displaced persons.