The Implications of Migration Governance and Colonial Structures in Humanitarian Organisations in Mexico

Non-government organisations (NGOs) have emerged in Mexico to support migrants while also cooperating with governmental institutions. Based on nine months of fieldwork, this paper explores the ambivalences within humanitarian organisations that aid migrants from Guatemala, Honduras and El Salvador. First, this article suggests that organisations engage in practices of mistrust and policing of migrants that do not align with their intention to safeguard them. Second, it argues that NGOs facilitate instances of paternalism to control migrants in their shelters and further subordinate them. Finally, it addresses how local organisations and shelters resolved the duality of care and control by punishing migrants. NGOs are drawn to reinforce the power structures that govern global migration. This is often in contradiction to their caring and supportive roles. As a result, NGOs replicate entrenched racialising discourse, power inequalities and control over mobile subjects.

In recent years, it has been estimated that around 400,000 migrants travelled from northern Central America, that is, Guatemala, Honduras and El Salvador (Canales 2019: 52;Rodríguez Chávez 2017: 9). Once they are in Mexico, northern Central American migrants have to travel more than 3,000 kilometres northbound to reach the US. During their journeys, migrants encounter adverse conditions as well as numerous risks that put them under constant threat (Campos-Delgado 2021), especially from the state. Between 2013 and 2015, Mexican authorities detained 175,000 migrants annually--more than the US Government detained in the same period (Canales 2019). The implementation of governmental policies in Mexico, backed up by the US, targets migrants, which, in turn, has accentuated their vulnerability (Castillo García and Nájera Aguirre 2016). As a result, the Mexican government has imposed more measures for controlling, detaining and deporting migrants trying to reach the US-Mexico border.
In contrast to governments' actions, humanitarian organisations have become 'excellent allies' (Comisión Nacional de los Derechos Humanos 2018: 22) to migrants. Local NGOs are generally known as casas del migrante (which translates as 'migrant houses') because they are considered welcoming hubs for migrants. More than 100 local humanitarian institutions are operating in Mexico (Basok et al. 2015), offering basic aid and services to migrants, who mostly come from northern Central American countries. These humanitarian organisations I refer to have institutionalised their operations at a local level. While they are diverse in size and composition, many receive external funding. According to a study by Calvillo Velasco (2018: 296-297), 42% of the funding of the NGOs examined received funds from international cooperation and agencies, while 54% of funding was sent by the government and the remaining 4% by private donations. Some casas del migrante are faith-based organisations supported by the Catholic Church, specifically by the Scalabrini order, 2 while others are run by civil society.
In this article, I will elaborate further on the rationales, mechanisms and dynamics of power that operate in humanitarian organisations. I aim to show that the effect of humanitarian support cannot be solely understood as beneficial to migrants. Nonstate actors are drawn to reinforce the power structures that govern global migration. This is often in contradiction to their caring and supportive roles. As a result, NGOs replicate entrenched racialising discourse, power inequalities and control over mobile subjects. Concretely, I build on NGOs resorting to dubious practices; by intending to safeguard people and create a familiar environment, they end up fostering migrants' subordination and employing punitive practices. The themes I draw on include humanitarianism along the US-Mexico border (Squire 2015; Vogt 2018), their underlying colonial and governance structures (Brun 2016;Pallister-Wilkins 2022;Rivera Cusicanqui 2010), and instances of immigration control and punishment of non-state institutions (Balaguera 2018; Gerard and Weber 2019).
As much as some of this research delves into NGOs' operations, my objective is not to expose them or accuse them. Some of these organisations have been through unsettling situations (Merlín-Escorza, Davids and Schapendonk 2021); nevertheless, we have to consider that they are not immune to internal and external power dynamics. Therefore, it is vital that some of these relations are recognised to improve the detrimental conditions that migrants endure. Moreover, I decided not to include the names of the organisations. This decision was not taken lightly since issues of visibility, accountability and credibility were some aspects that I considered. Nevertheless, naming them would put at risk the identities of the people who worked at the organisations and wished to remain anonymous.
The article will focus on the following sections. First, I outline the place of humanitarian organisations in migration governance and their link to colonial structures. Then, I retrace my steps for conducting semi-structured interviews. Second, I approach casas del migrante procedures to admit migrants in which they became suspects and their identities become challenged. Third, I examine the paternalistic practices that govern micro-humanitarian settings. Finally, I address some punitive practices in NGOs, such as acts of intimidation, aggression and expulsion.

Global, Local and Colonial Governance in Humanitarian Organisations
The international regulation of migrants' mobilities has been tackled by governments through their sovereignty claims and regional agreements. The United Nations' (2018) Global Compact for Safe, Orderly and Regular Migration was one of the latest concerted efforts to create a cohesive international and legal instrument to regulate international migration (Micinski 2021). Despite the diplomatic and political negotiations, throughout 2016 and 2018, many countries, including the US, ultimately refused to adopt this international agreement. Notwithstanding, the flourishing of the global governance of migration has been propelled by disparate actors, among them governments but also humanitarian organisations. Aspasia Papadopoulou-Kourkoula (2008) suggested that humanitarian organisations are mainly supporting structures dedicated to providing resources and becoming an important asset to migrants. Although one of their primary goals is to supply humanitarian relief, I contend that humanitarian institutions do not have simple structures, and their practices are entrenched in power structures.
Critical scholars (Agier 2011;Fassin 2012;Feldman and Ticktin 2010;Ticktin 2015) have addressed the humanitarian contradiction between caring for and controlling people. For instance, Didier Fassin (2012) has explored the tension between compassion and repression in immigration, which takes the form of 'ambivalent hospitality'. In his analysis of the French Red Cross, Fassin (2012: 139) examined the substitution of state functions by this NGO: 'the task the state had conferred on the humanitarian organization was primarily one of public order'. For his part, Michel Agier (2011) proposed that the controlling character of humanitarianism is not something bestowed by the state to humanitarians that they reluctantly accept; rather, it is ingrained in humanitarianism's caring prerogative, especially in Global South operations. These authors have located the conflictive and ambivalent assistance of international organisations with migrants, providing insights into the nature of humanitarian workers to sustain the management of migration.
Further, William Walters (2015) suggested that humanitarian governance is not separate from state control of immigration; instead, it is linked to its functioning through complex power operations that constrain the movement of people through state borders. The roles of international humanitarian institutions, grassroots organisations, workers and even volunteers (Pallister-Wilkins 2022) in governing state borders do not just entail a mere counter-response to state sovereignty. Instead, organisations' work has become co-constitutive to the production of migration borders, making them more aligned with states' interests in terms of controlling the mobility of people.
In the context of Northern and Central America, the discussion on humanitarian organisations' ambiguities is fairly recent. Some of the existing literature has stressed humanitarian actions as providing liberation and salvation (Hagan 2006). By examining NGOs through their caring practices, scholars have suggested that organisations are implementing 'supportive intervention' (Olayo-Méndez, Haymes and Vidal de Haymes 2014: 211-213). Other researchers have also started to analyse the operations of shelters, NGOs and refugees in Mexico (Álvarez Velasco 2016; Balaguera 2018; Merlín-Escorza, Davids and Schapendonk 2021). Among the evidence provided by these authors, instances of the maltreatment of migrants and ethical dilemmas have been considered part of these humanitarian institutions. Moreover, they have cautioned against shelters reinforcing migration management while facilitating migrants' mobility. In that sense, this growing scholarship has become more critical regarding the role of humanitarian organisations.
As much as there has been a critical appraisal of humanitarianism, especially at the US-Mexico border, there are some underlying challenges that have been downplayed. More specifically, some authors (Mayblin and Turner 2020) have suggested that the analysis of intermediates, such as humanitarian organisations, would need to explore the experiences of the racialisation and dehumanisation of migrants. Therefore, to understand humanitarian borders and the global governance of migration, I suggest that addressing colonial logic and governance structures may provide a more comprehensive account of humanitarian organisations. As Picozza (2021: 5) indicated, the understanding of humanitarian work is incomplete without a decolonial outlook: 'neither a humanitarian nor a no-border framework suffice; solidarity becomes an empty word if one does not adopt an anticolonial framework, namely a commitment to a horizon of [the] abolition of all (post) colonial relations'.
Decolonial perspectives have paid close attention to colonial and power relations in humanitarianism, especially in international organisations and agencies. Polly Pallister-Wilkins has examined the colonial implications of humanitarianism within the plight of migrants, tracing their colonial racial reproduction. Humanitarian workers secure and 'maintain colonial order' (Pallister-Wilkins 2022: 265), even as grassroots organisations and collectives fall into the lure of colonialism. Critical decolonial scholars have urged us to expand our understanding of the underlying colonial structures operating in humanitarianism and their unequal pattern of power dynamics. Silvia Rivera Cusicanqui (2010) proposed that NGOs follow an imperialistic formulation as they conceal forms of colonialism. Acts that recognise vulnerable people, such as Indigenous populations, end up hiding entrenched structural inequalities. Sirin Adlbi Sibai (2016) examined how the discourses of international, Spanish and local NGOs have affected Islamic women in Morocco. She observed the imposition of the duality of 'tradition/modern' categories on local women to legitimise the work carried out by the international cooperation that holds superior power positions. As it has been argued by Magalhães Gomes (2021), this theoretical lens can help us observe colonial structures operating in institutions such as the criminal justice system. In that sense, as much as humanitarians cannot escape border controls through 'less hierarchical forms of humanitarian assistance' (Pallister-Wilkins 2022: 163), I argue that their practices have colonial traces that cannot be entirely eradicated but may still be contested by migrants.
Overlapping practices between the government and humanitarian workers need to be further unravelled as people navigate complex institutions and processes. I would not like to suggest that humanitarianism automatically reproduces power relations. As much as state and colonial structures operate through humanitarian workers, they are neither exceptional incidents nor can be reduced to them. Consequently, in this paper, I argue that colonial logics continue to inform the global governance of migration and humanitarian borders as they become intertwined with operations of practices of care and control. In that sense, migrants have also developed other tactics (Abrego 2014), such as compliance, avoidance and defiance, to display their agency.

Conducting Fieldwork
I conducted fieldwork in Mexico for nine months from August 2018 to May 2019 in 13 organisations spread throughout three cities: Mexico City, Tapachula and Tijuana. By considering participants' comments and shadowing them in particular sites, I used the approach of 'following people' to carry out multi-site fieldwork (Díaz de León 2022;FitzGerald 2006;Vogt 2018). I found that this methodological strategy also applied to humanitarian organisations operating in various places. Although organisations were established in one location, they were not fixed to one site exclusively because they were 'opening' and 'closing' their doors in different locations. Humanitarian organisations strive to locate themselves near migrants' routes, trying to reach them (Calvillo Velasco 2018: 160), and these routes are constantly changing. Additionally, staff rotations and relocations were constant; therefore, staff members tended to change job placements and forge communication and networks across them.
For this research, I worked in four organisations in Mexico City, five in Tapachula and four in Tijuana. 3 After locating potential organisations, I conducted 22 semi-structured interviews with staff members and 16 with migrants from Guatemala, Honduras and El Salvador, which amounted to 38 semi-structured interviews in total. The recruitment of NGO staff members was based on those that worked with this group of migrants. Further, I invited workers who had more experience in the community and held semi-permanent positions in the organisation. The migrants I interviewed were members of the same organisations I worked with; thus, contacting migrants was only possible through NGOs. 4 I usually invited migrants to participate during their spare time: after communal dinners, during recreational time, during routine gatherings or in the organisations' courtyards. In some organisations, informants and gatekeepers at NGOs would introduce me to migrants who would be willing to be interviewed.
Once I finished my fieldwork, I coalesced the data from all the interviews and built a corpus through which I sought to identify relevant topics. In that sense, the thematic analysis allowed me to understand different patterns of data. I not only identified repeated words but mainly traced salient themes. Once I obtained these, I connected them and encoded the information. Next, I anchored the themes with key quotes from participants. I followed this methodological approach because I was interested in the stories that people construct: 'for us, qualitative research is about meaning and meaning-making, and viewing these as always context-bound, positioned and situated, and qualitative data analysis is about telling "stories", about interpreting, and creating' (Braun and Clarke 2019: 7). In that sense, observations were important to give a concrete meaning of the narratives, individuals and places.
It should be noted that decolonial research has significantly influenced the ethnographic accounts that played a role in imperialism and colonialism (Alonso Bejarano et al. 2019). Although this research is inspired by ethnographic research, it also takes into account the alternative of 'researching back' (Tuhiwai Smith 1999: 7-8), which entails a process of analysing the 'other sides' of colonisation-that is, individuals often excluded and marginalised. In my interpretation of this methodological program, I took into consideration both staff workers and migrants in the decisions made throughout this research. It made it possible to listen attentively to what individuals had to say. Staff members often enjoyed having a platform to disseminate their work, and the fact that I have a Mexican background eased some of the meetings and conversations. Given the broader context of anti-Central American sentiments in Mexico, migrants were overly grateful and often too cautious to speak to me. When I also introduced myself as a postgraduate researcher from a United Kingdom university, I discovered that this facilitated rapport with migrants. Migrants felt that I understood, even if just partly, the immigration experience. This had a profound resonance with me in terms of reflecting on my positionality of coming from the Global South as I was going through a migratory process of my own.
Even if I acknowledged the power hierarchy with participants, I was firmly aware that I still held a dominant position during the interviews because I had a specific interest in engaging with them. Committing to this critical perspective also made me wary of 'data extraction' from participants: 'it cracks open the oysters of other people's lives and harvests the rich goo within' (Alonso Bejarano et al. 2019: 7). Hence, I was conscious of my privileges as a social researcher delving into people's lives and work and that I could only moderate and minimise but not eliminate completely. My position as a researcher carried implicit power, so I did not try to level the power field. The best ethical position I could sum up was not to affect them by being sensitive and respectful to potential traumatic memories, not revealing their identities, offering to stop the interviews at signs of distress, and extending emotional validation and reassurance. Further, it was important to take them seriously during the research by listening to what individuals had to say without leaving aside their experiences and prioritising their understandings, concerns and words.

The Colonial Doubt About Migrants
Rodrigo worked in one of the most well-established local NGOs in Southern Mexico. As we started the conversation, he quickly pointed out that one of the central aims for that organisation was to become a safe space: 'it is about them [migrants] having a safe place, a space of trust'. The allusion to a safe space has a long history with vulnerable populations, where they could meet others away from the violations of their rights (Flensner and Von der Lippe 2019: 276). Safety concerns have become important for the growing number of northern Central American migrants in Mexico. 5 Attesting to the growth of emigration from northern Central America, the US and Mexican governments have created a zone of border contention, also known as an 'arterial border' or 'vertical border' (Reineke and Martínez 2014;Vogt 2018) that has geographically and politically further divided Central America from North America. This border expansion has especially affected northern Central American migrants who travel through Mexico and seek to cross the US-Mexico border.
Humanitarian organisations rely on this protective discourse to substantiate the provision of humanitarian relief by making their environments safe for migrants. NGOs closely associate their mission with creating a humanitarian place as numerous risk factors--such as trauma exposure, human trafficking, extreme poverty and territorial displacement--plague migrants' journeys (Olayo-Méndez, Haymes and Vidal de Haymes 2014). Eduardo was a 21-year-old Salvadoran migrant travelling through Mexico who I met at an organisation in Tijuana. He had stayed in five shelters across various towns and cities. Despite Eduardo's accomplishment in finding these organisations, his admission was not as easy as expected. As I spoke with him, he detailed how being allowed in these organisations was not just a matter of showing up and knocking on their door. It usually required being assessed as a potential candidate to be granted some degree of help. He had to go through an internal process in which his identification card and his asylum application were required. Afterwards, he was interviewed, and upon qualifying for aid, he was asked to apply for a job as soon as possible as a condition for his stay.
Similarly, Omar, a Honduran migrant, was subjected to routine and protocol screenings to detect if he had any sort of prior criminal record. Omar knew that as soon he arrived at a shelter in Mexico City, he would be asked to provide personal details: 'when one arrives here [to the shelter], she [the director] has to check them in on the computer to see if I am not a thief or a kidnapper because all shelters are linked. So, if in a shelter I were to steal something, it would show up in another shelter … so that they [criminals] will not come'. The use of personal information not only raises questions about the governance implication of migrants' registration of data (Pelizza 2020), and the use of military and security technology tools for NGOs' operations (Garelli and Tazzioli 2019), but has also become a normalised practice for migrants. Migrants whom I spoke to usually had to provide a credible story about their journeys, verifiable proof of their travels (e.g., train and bus tickets), specific locations of where they had been and their passport or any official form of identification. Those who failed to comply with these requirements were denied aid and even turned down by these humanitarian places.
NGOs' frequent assessment of migrants was justified as an attempt to identify smugglers, criminals and members of organised crime in their midst. Organisations often alluded to these as the ultimate perpetrators of violence against migrants. The effects of humanitarian organisations' caring mission, promoted as providing a necessary safe haven for migrants, has resulted in engaging in suspicion and surveillance to protect them from alleged perpetrators. 6 In this sense, I argue that NGOs were actively reinforcing discourse on the criminalisation of migrants from the Global South. NGO directors like Sara had gone to the extent of consciously patrolling migrants and assigning themselves as protective agents: My role in this civil association is to be vigilant, to prevent. So, … if I detect a mara [gang member], [or] a pollero [smuggler], I attend to them personally. The way to get them out is never by being rough. I will always be very polite. I have even given them money … I have even made a show of it.
In a theatrical and performative manoeuvre, Sara took pride in detecting potential criminals, outsmarting and expelling them. Her desire to prevent migrant smugglers and gang members from entering distorted her role. Therefore, it can be inferred that to cast out criminals, staff at NGOs were compelled to become keepers of 'social order' and promote 'hostile solidarities' based on the bond with other vulnerable individuals (Carvalho and Chamberlen 2017).
Humanitarian organisations have been vocal in their advocacy as they denounce the criminalisation of migrants at the US-Mexico border. Nevertheless, this does not exclude them from also perpetuating disbelief and policing practices by questioning migrants' stories, photographing potential suspects and assessing individuals' potential of being criminals. Wendy Vogt (2018: 375) also acknowledged that there are instances in which protection is not always guaranteed in local organisations and that there is some disconcerting lack of trust, but it is labelled 'compassionate' since this doubt is born out of a safety concern. Nevertheless, this scepticism does not only result in policing migrants out of the unrequited self-appointed task of protecting them but is actively produced by colonial legacies that racialised individuals.
The attitude of perpetual questioning and doubting people were heavily enacted on colonised people during imperial administration. Nelson Maldonado-Torres (2007: 245) examined the scepticism of colonised people in order to interrogate their humanity justified by the ego conquiro: 7 'and just like the ego conquiro predates and precedes the ego cogito, a certain skepticism regarding the humanity of the enslaved and colonized sub-others stands at the background of the Cartesian certainties and his methodical doubt'. Unlike the Cartesian doubt that gives way to rationalising the outside world and allows understanding, this colonial doubt, the 'misanthropic scepticism', questions 'how human is the "other" ', which challenges the humanity of others.
Maldonado-Torres (2007) observed how colonial mistrust is imposed over racialised individuals in postcolonial contexts. In this case, I found that the mistrust was instrumentalised by humanitarian organisations that kept migrants policed and subordinated, echoing colonial legacies in these places. As the colonial doubt was placed on migrants, it was entrenched in historical colonial legacies; it also carried a racial imprint. Humanitarian organisations arguably participate in the racialisation of migrants through the enactment of the Mexican national 'mestizo project' that excludes minority groups because they are perceived as a threat to the majority of a mixed population (with Indigenous and European heritage). The idea of the mestizo is based on a national ideology founded by the ruling elite that strives towards whiteness (Moreno Figueroa 2010) and is rooted in racism implanted during colonial administration. This results in the stigmatisation of Black, Chinese and other minorities, including Central Americans, that threaten the mestizo ideology.
While I visited these organisations, it was recurrent to hear that Central American migrants brought violence to Mexico. Even though it was acknowledged that Mexico is already subsumed under structural violence, 8 staff members resorted to perceiving migrants as bearers of violence, as Felicia, a social worker, recounted: '[violence] limits a lot [of] the actions of people, since they have lived with violence this whole time … Because they [migrants] already carry a lot of violent acts, they violate other people, perhaps, unintentionally'. Felicia was among staff members that would rely on racialising discourses upon Central American migrants. An important outcome of protecting migrants and becoming guardians of safety is that they end up stereotyping and racialising migrants' identities as violent people. This, in turn, feeds into their dehumanisation when, as individuals, they enter a process in which their humanity evolves into an animal form (Fanon 1963: 42) through instances that evoke less-than-human figures. I would also add that the humanitarian sector engages in the process of dehumanising migrants by their constant lack of trust, alluding to their inherent violence and, thus, racialising them. The next section will show further subordinate instances in which migrants are involved.

Paternalism in Humanitarian Shelters
While some migrants like Eduardo and Omar gained knowledge about these humanitarian organisations from maps, word of mouth or browsing through web pages, most are directly transferred. Migrants are frequently brought from governmental institutions such as the Mexican Commission for Refugee Assistance (Comisión Mexicana de Ayuda a Refugiados in Spanish), Grupo Beta 9 or other international humanitarian institutions. These transferals of migrants to humanitarian shelters are part of 'channellings' (or canalizaciones in Spanish), which consist of communication between government institutions, international agencies and other casas del migrante for migrants and asylum seekers to be housed and cared for during the migratory process. Additionally, these exchanges follow semi-strict protocols, where information is shared with various state and non-state actors. It is through these everyday communications, as well as their work receiving and transferring migrants, that organisations serve as points of contact with governments and other institutions.
Staff members at local NGOs occasionally suggested that the benefits and resources they offered to migrants were in substitution of the state, which reveals a gap-filling function (Griffiths, Sigona and Zetter 2004;Nyberg Sørensen 2013). The Mexican government generally outsources their obligation of providing housing for migrants to humanitarian organisations, while the latter is portrayed as an alternative to that of the state. Functioning as part of the government while being against it has been a defining feature of humanitarianism, which partakes in implicit ways with the migratory governance of migrants. As Walters (2011: 149) argued, humanitarianism is in a frequent 'state of co-option, infiltration but also provocation with the state (but also with other supranational and international entities as well)'. I observed that the effect of this contradiction was revealed in the conviviality that takes place inside humanitarian hubs. Once admitted to NGOs, migrants have a place to spend the night. Migrants at these organisations are provided with housing, meals, donated clothing and shoes, together with medical, legal and psychological support and the opportunity to participate in recreational activities (e.g., football matches, making art and watching films). As has been mentioned, 'migrant houses' not only offer resources and services, but above all, they strive to resemble real homes for people. NGOs have encouraged migrants to bond with each other and for the shelters to become families, as Santiago mentioned: We try to make the second-floor home for everyone; and for them to become friends, family, since they have gone through very similar situations. From what I can perceive, they become like family; there is support among themselves. And they manage all the upstairs--they manage their cleaning, schedules and who takes turns for things.
Social relations are produced among migrants living together in the shelter. Migrants have gone through similar experiences that bring them closer and allow them to become like a family, even if their own are far away. Díaz de León (2022: 15) has found that migrants form social communities that fundamentally serve during people's mobilities as an 'important substitute for social networks in times of violence and social breakdown'. These new close relationships can also alleviate the grief associated with losing contact with loved ones. Hence, the idea of encountering a family for migrants at casas del migrante can lessen the disconnection of the family they left behind, as these new social relationships can provide some level of belonging, affection and intimacy.
As Santiago overstated the family environment, he clarified that the management of menial tasks and cleaning schedules are 'allocated' to migrants. Migrants, along with some volunteers, are usually in charge of house chores (making the bed, cooking, cleaning and washing) and, at times, the construction and overall maintenance of these places. In most organisations I visited, it was not necessarily the case that these tasks restored the ability to rebuild homes for displaced people (Gil Everaert 2020) because it was not up to migrants to come up with the tasks; rather, they were assigned to them. Humanitarian organisations implement rigorous schedules, including specific hours to sleep, eat, clean, smoke, go out and even use their mobile phones. The norms that are established in humanitarian organisations are accompanied by power relationships that are unequal, as these are supervised by staff members.
The manufacturing of a family setting and the enforcement of chores entails a power dynamic of paternalism that guises the interaction between staff and migrants. Paternalism in humanitarianism is not rare (Malkki 2010;Sigona 2014), as it emerges from the contradiction between caring and control, depicting individuals in need of care and support and less capable of decision-making. Because migrants are deemed 'minors', this, in turn, infantilises people by considering them less mature adults and, thus, fixating people in an inferior place.
Acting in the best interest of migrants while resembling a family reinforces paternalistic attitudes, encourages claims of superiority and reiterates colonial attitudes and racialisation from the Global North: 'humanitarianism is a product of the same paternalism we can observe in the global colour line by which, in both cases, there is an assumed hierarchy between those who can care for themselves and those who cannot' (Pallister-Wilkins 2022: 73-74). Not only is this recognised, but also performed by migrants. Jose, who was travelling from Honduras, talked about the rules and family dynamics at a shelter: 'yes, there are rules … But above all, care. They treat us like kids. As if for the managers we were children to them'. While being treated as a child may provide some respite from the turmoil of the experiences of mobility, it is conditional on migrants' stay. In other words, paternalism has perpetuated migrants' subordinate positions and bolstered the surrendering of some level of autonomy. Ultimately, migrants' infantilisation reproduces a colonial logic placed on racialised people, in which they are supposedly unable to govern themselves (Chakrabarty 2000: 8). This perception of migrants as immature and unprepared to exercise full agency is put into practice in humanitarian organisations to ascertain their authority.
As paternalism roles were reinforced at NGOs, it became apparent that this was part of a wider context in which international humanitarian institutions influence and pressure local organisations. Local humanitarian organisations work in close partnerships with government offices and international agencies. In an interview with Diego, a humanitarian worker in Tapachula, he talked about how paternalism became preponderant among staff members who had close bonds with larger international humanitarian agencies. In a trickle-down effect, this paternalism is also exercised by casas del migrante towards migrants, as they are constrained by the mandates and guidelines from the international agencies in the Global North. This is not to say that staff workers mindlessly follow external agendas but that the latter overpower relevant context-specific information and knowledge (Brun 2016) because NGOs rely on their funding (Gerard and Weber 2019). Power is displayed through this specific dynamic where migrants become children, and staff members act as parental figures. In turn, this paternalistic dynamic may reinforce controlling and even punitive practices.

Punishing Migrants in Humanitarian Ways
Besides the patrolling and paternalism towards migrants, practices of control were constantly performed in shelters that restricted the agency of people. Adriana, a staff worker at an NGO in Southern Mexico, mentioned: 'sometimes [migrants] don't want to be locked up; they don't want to be controlled. They want to go out and not follow the rules. In other words, they want to live'. Adriana retells the reluctance of migrants to follow the rules and, above all, not to be subordinated and subjugated. During the interview, Adriana also acknowledged that the controlling mechanisms at that organisation oppressed migrants' lives. In that sense, humanitarian workers were, at times, self-critical of their practices. Similarly, Juana, who was working at a shelter in Mexico City, was one of these people who were in disagreement with some of the practices and attitudes yielded by the director: One time, Claudia [the shelter director] came in and kicked the door, and she said, 'whose are those clothes?' I did not say anything. I just remember that there was a boy, he smiled and I left the shelter crying. Claudia, who stands up against this issue, has this attitude towards them [migrants] … They [migrants] told us, 'if Claudia comes, I rather go. Avoid her'.
It was confusing for Juana how an advocate for migrants imposed practices of aggression and intimidation towards migrants. Although the director yielded this excessive behaviour, it was not uncommon to find authoritative attitudes among the highlevel staff at the places I visited. This further suggests that the humanitarian ambiguity inside organisations can be conducive to maintaining controlling and bordering practices. I argue that the tension between benevolence and security within NGOs is not resolved by a bad partnership with the state, as it takes a punitive dimension-that is, punishment is a way out for organisations to continue operating through dominance, control and authority. Taking into account other controlling instances, NGOs resemble disciplinary technologies used to exercise power over individuals (Foucault 1982) in a prison-like scenario.
In a similar vein, Martha Balaguera (2018) found these actions at the hands of casas del migrante that targeted migrant trans women. Her study demonstrated instances of coercion and the enactment of the carceral regime beyond the state: 'when shelters intervene so that migrants may access otherwise elusive rights and protections, they partake in a larger regime of sovereignty not reduced to the state' (Balaguera 2018: 655). As the transgender population goes through specific instances of gender normativity in casas del migrante, I also found that Global South migrants are subjected to techniques from the criminal justice system. Traces of penal power embedded in migrant shelters are impactful and widespread among individuals, especially from northern Central America. Migrants were punished if they broke the rules, did not complete their chores or were caught doing something 'wrong'. That is the reason avoidance has become a tactic for migrants, in the sense that they adopt a nonconfrontational manner to deal with the exercise of power when tension arises with high-level staff. Avoidance tends to occur in poorly organised settings as an individualised form of agency (Roscigno and Hodson 2004). From Juana's quote above, it can be observed that when migrants felt threatened and intimidated, they refrained from those situations. Indeed, most migrants I talked with wished to preserve their stay at the shelter for as long as they could, so they resorted to subtle noncompliance practices.
Yet, when migrants did not follow the rules or violated any regulations, they were asked to leave. One of the significant reprimands of casas del migrante is the expelling of migrants from their premises. This process of expulsion was especially difficult and entailed punishment for migrants who did not have anywhere else to go. The gravity of this is that migrants may go missing or disappear once they are dismissed from casas del migrante, pushing them into dangerous scenarios.
Rebeca was desperately searching for her son, who had left El Salvador, which had been her main reason for migrating: 'there are many [migrants] who are looking to fulfil their dreams, but many [of us] are looking for other things. We do not seek to emigrate … I'm looking for what I have lost'. It was difficult for her to talk about locating her son in the midst of her also being uncertain about being expelled from the shelter. Still, she believed that she had a choice in all that was happening around her and told me so: 'I believe that we act with our own mind and are masters of our own destiny'. I suggest that many of these local organisations function with punitive practices that aggravate the danger of becoming homeless. As has been explained, humanitarian organisations operate internally with a governmental and colonial logic that places migrants in a disadvantaged position that constrains and punishes those who challenge institutional structures of power and control.

Conclusion
Humanitarian assistance provided by local organisations is not exempt from global designs. On the contrary, as I have demonstrated in this paper through instances of admittance, sheltering and care, humanitarian workers have resorted to mistrust, control, paternalism and excessive control of the people they seek to help. While scholarship has critically explored how humanitarian organisations actively participate in the governance of migration, I incorporated a decolonial perspective in this paper that has sought to broaden the understanding of the humanitarian-migration nexus. I have built on power dynamics to reveal that the colonial past remains present through the constraints and control of humanitarian organisations in global and state management of migration. This article contributes to a specific and concrete case of mid-level actors that find themselves enforcing bordering and racialising practices by focusing on humanitarian ambivalences in the context of northern Central American migration in Mexico. Moreover, it would be insightful to explore the roles of migrant organisations and grassroots communities with Global South migrants to continue expanding on the complexities of non-state actors. How do they approach issues of racism, family and integration within the migrant population? Do they engage in acts of solidarity and activism? And how do they navigate gender normativity and practices of advocacy?
As has been shown, at times, humanitarian workers have displayed self-criticism and self-awareness of their work, but also, in other instances, the migration governance over Global South has intruded and merged with workers' performances of authority, power and racialisation. For their part, migrants staying at these places have contested some of the practices. While some migrants resist abuse, others seldom explicitly express disagreements that evade overly idealised accounts of migrants. Further, migrants' presence in NGOs and overall mobilities to the Global North are constantly subjected to dominant power relations and deemed a threat to global migration configuration, which ultimately adversely affects their humanity.
Correspondence: Dr Erika Herrera Rosales, Associate Fellow at the Institute of Advanced Studies, University of Warwick, United Kingdom. erika.herrera-rosales@warwick.ac.uk