Introduction

We experience the normalization of the expansion of sophisticated migration control and border infrastructure not only in the North but also, more and more, in the Global South. In order to prevent Africans from reaching the Mediterranean Sea and heading to Europe, the current focus of the European Union’s externalization policy is on the continent, trying to involve partners in border and migration management and exporting the migration risk narrative through collaboration. European-financed infrastructures for the control of mobility in general, and of migration suspected of being directed towards Europe in particular, in some ways perpetuate the colonial logic of regulating flows of people and goods. As under colonial conditions, control over these infrastructures is marked by power inequalities. Similar to Distretti’s (2021) study of the coloniality of the Libyan Coastal Highway, the functions and effects of today’s infrastructures of migration prevention have a long history that has not yet been sufficiently studied. Various authors point both to the different means, such as development cooperation or cooperation agreements, which the EU draws on in this process to incentivize African partners, and to far-reaching and destabilizing political and economic effects on local populations in West Africa (Brachet 2018, Frowd 2018, Galya and Volker 2021).

In the Kantchari/Makalondi borderland between Burkina Faso and Niger, (im) mobility is mediated through various infrastructures that are shaped by armed conflict, militarization, and securitization of movement. The Kantchari border is one of those sites, where funds available through, inter alia, the Programme d’Appui à la Gestion Intégrées des Frontières (PAGIF) of the EU allowed the governments to rebuild border posts, to send additional staff, and to boost the presence of security forces in order to prevent smuggling, human trafficking, crime, terrorism, and migration denoted as irregular. However, Anderson (2016) has shown that “irregular migration” is a rather new phenomenon that has started to be discussed only in the context of the more restrictive migration policies in the EU since the 1990s. Alongside the discursive framing of migration as a security problem, control infrastructures at external European borders were reinforced (Gabrielli 2014). This was followed by the externalization of migration and border control beyond EU territory, starting with the Eastern European states. Currently, the EU’s externalization policy focuses on West Africa and promotes the internalization of the desired border and migration management by the targeted countries (Bisong 2020). Post-imperially, through incentives at various levels and especially in the framework of development cooperation and through cooperation agreements, the narrative of migration risks is exported.

At a macro level, these exported migration and border management practices and narratives, or their respective translations by West African states, have a strong impact on regional migration dynamics and ECOWAS institutional frameworks (Galya and Volker 2021). On a local level, negative economic effects of these mobility restrictions on a population that depends largely on cross-border activities are an important driver for violence against the security forces (Donko et al. 2021). Prior to an armed attack on the new border post in Kantchari in April 2020, in which the buildings were burnt down and completely destroyed, the post was also equipped with the border management information system (MIDAS) of International Organization for Migration (IOM) that automatically captures biographical and biometric data of travelers using document and fingerprint readers. Kantchari is a strategic site for regional cross-border mobilities, linking the ports of Lomé, Accra, and Abidjan with Niger. After the outbreak of armed conflict in Mali in 2012, and associated restrictions on travel opportunities, the Ouagadougou-Kantchari-Niamey route became increasingly important for international migrants on their way to Agadez and to Libya.

The passage of travelers, mainly from Gambia, Côte d’Ivoire, Mali, Senegal, Ghana, Togo, and the three Guineas, had led to a vibrant migration economy in Kantchari. This not only attracted business people and traders from outside the region but also, primarily, benefited the local population who opened hotels, shops, and restaurants, serving as mediators between border officials and migrants, and bringing them uncontrolled across the border on their motorbikes. However, the economic consequences resulting from the implementation of migration control in 2018 are massive and many of the restaurants, bars, and shops are now struggling with the problems that came with the restrictions on cross-border mobility. However, despite migration control, exacerbated by further restrictions on mobility in the course of the fight against terrorism and measures to contain the corona pandemic, there are still important migration flows and, as a trader confirmed our observation, the infrastructures that enable and restrict mobility exist as well.

This town lives from migration. Or rather, I should say it has lived from migration. Now it still does, but not like before. Now it looks different. But we still have everything the migrants need.... And the state actually also has everything it needs to make their lives difficult and to earn money from them. And so people still get their share out of it (interview with a trader, Kantchari, 20.9.2020).

By infrastructure, we mean the material, technological, and social conditions, and their interrelationships and effects, through which mobility is regulated and/or immobility is enforced in this borderland (Collier and Ong 2003). Empirically, we deal with three forms of migration infrastructure in particular: buildings and roads, social relations, and technologies. What is common to all of these is that they are crucial components of the local migration economy and are, as such, used both to prevent and to enable migrant mobilities. So, in this paper we study infrastructures that condition (im)mobility. We analyze infrastructures that both support migrant mobility and serve the European idea of keeping Africans in their place, but also analyze how these infrastructures are subverted, challenged, and used by the ones who are supposed to be immobilized by them in order to remain mobile. Hence, we develop a relational view of the infrastructure of (im)mobilities as a dynamic network of social organization, economic order, and technological integration.

This introduction is followed by a theoretical discussion of infrastructures (of migration) that outlines our relational approach. After a note on methodology, we present our empirical data on infrastructures that are involved in the production of migrant (im)mobilities and conclude thereafter.

Im/mobility and Infrastructure Multiple

Mobility as a context-specific ensemble of movement, representations of this movement, and concrete mobile practices is political (Cresswell 2010). In the Kantchari borderland, the current global redefinition of the politics of (im) mobility translates into new modes of border control, government-imposed mobility restrictions, and multiple responses to these interventions. We still know relatively little about how migrants in Africa perceive and experience the restrictions of the new (im) mobility regimes. Usually, the concept of mobility regime summarizes the governance of international migration that is characterized by the sharp contrast of free movement in and from the Global North on the one hand and, on the other hand, sedentarist imaginations relating to a Global South where movement is controlled, sanctioned, and generally undesired, especially if these movements target the North (Glick Schiller and Salazar 2013). We propose to downscale this concept and to understand (im) mobility regimes as constituted by a specific set of infrastructures and infrastructuring practices through which movement is contained or, more generally, through which (im)mobilities are governed, and the ways the targeted subjects respond. Thinking of the relations of mobility and stasis as both an outcome of and the shaping of a regime allows us to reveal the entangled power relations at work in the respective infrastructures.

This is neither the place to summarize the whole range of networked, hidden, everyday, large-scale, evental, and further conceptualizations of infrastructure (Barua 2021; Furlong 2020; Lin 2020), nor to overuse the thousand-fold cited “Politics and Poetics of Infrastructure” (Larkin 2013) as is so often done in many papers on infrastructures. “Infrastructures are built networks that facilitate the flow of goods, people, or ideas and allow for their exchange over space” (ibid 328). However, Achille Mbembe (2021), in his keynote lecture during the Eurafrican Legacies Conference, underlined that infrastructures do not call for definition but for mapping and surveying. The definitions in Larkin’s seminal text are appealing because they are simple, but on the other hand, they do not capture the potential for surprises and apparent paradoxes. Imagine a house for migrants without passports who are waiting to continue their journey. The house offers shelter and preparation for a risky border crossing, but it can equally well be the site of a raid by the police or immigration service, who will send the inhabitants back to their respective countries. So, we suggest following a train of thought in the theorization of infrastructures that is significant for our analysis: relationality. Starr (1999: p. 380) has shown that infrastructures as a “system of substrates” can only be understood through relations. We do not claim this is new; other scholars have discussed the relationality of infrastructures epistemologically (Niewöhner 2015) and/or used it for the abstraction of their ethnography (Harvey 2012, Williams et al. 2019). But we consider that with a relational approach the multiple outcomes of infrastructural work in terms of migrant mobilities such as stasis, blockage, or movement are better understood.

Why “infrastructure multiple?” The simple core argument is that infrastructure, understood as the conditions of possibility, is multiple since it is something that is performed in a variety of practices (Mol 1999). There are different options of infrastructures and it is crucial to understand which specific options are enforced and get powerful. There are no contradictory infrastructures but rather multiple practices of dealing with them. However, contrasting versions of infrastructures that are practiced do not imply pluralism, understood as the idea that reality (and thus infrastructure) is composed of separable entities that may be added together. It is not plural, but multiple. Taking a biometric identification infrastructure as an example: We will show that such technologies produce very different migrant (im)mobilities depending on the mode of adaptation to the local context. In one case, it can be used to prevent undesired migrants from crossing the border, whereas in another case a traveler takes advantage of the fact that the document scanner cannot read Arabic script and enters into negotiations with immigration policy officers as in the time before biometric control. Or, as Bourdieu (1998: p. 15) has put it: “The real is relational.” We understand multiplicity as constantly updated products of processes of relating, as context-specific relations, and as relational constellations that classify infrastructure as multiple. Multiplicity refers to the relational production of multiple phenomena and forms of being that are multiple because they are themselves produced by multiple relations. Relationality refers to the dynamic, multidirectional, simultaneous, continuous, and reflexive processes of relating and the emergent character of the results they generate. These relations are social, economic, and material and may variously involve individuals, social groups, objects, and ideas. As an analytical tool, relationality thus apprehends the wide array of ways of infrastructure ontologies, while acknowledging power structures and inequalities (Spies and Seesemann 2016).

In this way, we do not study infrastructures as things, but as the multiple relations between institutional and technical structures, the routines of the heterogeneous actors involved, and supra-local overarching organizational resources. These relations are reproduced in everyday practice and in their ability to produce a specific, constantly renegotiated (im) mobility regime in the borderland.

Fieldwork: a Note on Methods and Research Ethics

This article is based on ethnographic research in a borderland that is, just as others, a hotspot of still significant state territorialities that are reshaped by supranational forms of regulation of, for example, human mobility. We identified different sites of infrastructuring in order to then conduct interviews and make observations. In order to understand infrastructure phenomena as relational practices, we have conducted multimodal research that includes dense descriptions of practices as well as narratives of different actors and the analysis of material artifacts and technologies. We spoke to migrants, travelers, the border police, immigration officers, transporters, brokers, owners of shops and restaurants, IOM staff, and local officialsFootnote 1. We made observations at different places at and along the border, conducted interviews using the visible border as a stimulus and prompt, and observed interactions between those who currently experience the borderFootnote 2. Our objective was to highlight the variety of (im) mobility infrastructures and border-related activities and experiences. Therefore, we chose our informants according to the different border activities and attitudes that we observed while strolling near the border. We draw on information gathered in several periods between March 2018 and April 2022, both on site and through remote research via phone. The paper therefore covers the period immediately after the introduction of MIDAS, the periods of increased attacks on the security authorities, and of the pandemic-related mobility restrictions and the post-corona phase. If there are, as laid out above, different options of infrastructures, then the first empirical question that follows asks what the places and sites are where infrastructures are made in which situations and moments. A second one is where these options do clash, merge, collaborate, and even depend on one another or simply co-exist. In order to make the work of MIDAS visible, which is otherwise invisible to those who want to cross the border, we went behind the scenes to observe the users and talk to them about their work. The security situation in such militarized borderlands involves very specific challenges both for field research and for the responsible handling of the sensitive data of our interview partners. During our research, we ensured the data security of the people we work with. Data was immediately anonymized and we did not collect audio-visual data without consent.

In the next three sections, we analyze different but interlinked types of infrastructures of (im) mobility and the social and economic practices inherit within them: houses and roads as built infrastructures, the Migration Information and Data Analysis System (MIDAS) as the technological manifestation of an externalized European migration control, and social infrastructures that are linked to, but also go beyond, the first two types. Although our focus is different, ours being on infrastructures that both enable and constrain mobility, we thus cover almost those dimensions identified by Xiang and Lindquist (2014: p. 124), who have defined infrastructures of migration as “systematically interlinked technologies, institutions, and actors that facilitate and condition mobility”: commercial, regulatory, technological, humanitarian, and social. Infrastructures in this sense also include actors (e.g., owners of restaurants where migrants eat) or technologies (mobile phones) that are not explicitly oriented towards migrants. We will not go into these in more detail here but will concentrate empirically on the broader complexes.

“Ghettos” and Roads

Houses, roads, and dirt tracks in and around Kantchari are infrastructures of arrival, stasis, and renewed departure, either for onward journey or, in the case of (temporarily) failed migration, for the way back. They serve to enable migrant mobility and at the same time allow for forced immobility. In Kantchari, “ghettos” are the most important arrival infrastructures (Meeus et al. 2019), shelters in which people are initially accommodated and in which their (im)mobilities are negotiated and produced (Bredeloup 2021; Lecadet 2017). Here, migrants pay to stay for two or more days before crossing the border into Niger, the length of stay depending on what possibilities of continuing their journey are allowed for by their financial state. They are places to rest, to wait for money transfers from relatives or friends in Europe or their home country, and to exchange travel experiences, strategies, and routes. In parallel to the migrants who are only stopping over, there are also those who have been stopped at the border and sent back by the officials on either the Burkinabe or Niger side.

Around these accommodations, there are related infrastructures such as bars, restaurants, cigarette shops, money transfer facilities, airtime sales, and also brothels. Shop and restaurant owners recruit mainly female migrants from the ghettos to work for them, those who are in search of additional resources for their onward journey. A rest stop before the onward journey, with a place to sleep and the possibility of exchanging information, have money sent, and prepare for the border crossing: At first glance, the “ghettos” seem to be one of the classic infrastructures that enable migrant mobility. However, as we observed ourselves and as was confirmed to us in many conversations, the concentration of migrants in the ghettos is also used by the security and border authorities as a reason to search individual houses and even entire house complexes from time to time for “candidats à la migration” and to first arrest them if there are the slightest irregularities in their identity papers and then to ban them from Kantchari. The term “candidats à la migration” is used to describe all those who are assumed to be on their way to the Mediterranean. The same applies to the Niger side of the border. In 2019, we witnessed police arresting about 20 migrants from West Africa, including their traffickers, in a ghetto in Makalondi. They had entered Niger illegally from Kantchari and, after a few days in prison, they were deported back to Kantchari where they were again housed in the ghettos. The smugglers were taken to a prison in Téra, Niger.

These crackdowns, which we observed several times between 2018 and 2020, are still being carried out today, as a former Ghanaian migrant who has settled in Kantchari and now works as a carpenter told us:

The ghettos are not always a quiet place to rest; the security services are also very present here and get a lot of information. When negotiations with them don’t work, they get more savage and often start picking everyone upFootnote 3.

The negotiations with the security services emphasized here point to the importance of social infrastructures of migration, which we will return to later. For while raids, bans, or even deportation are the most brutal, they are not the most common practices of the security and immigration services, which also act more subtly. For example, travel documents, destinations, and the financial means of travelers in the ghettos are spied on behalf of the security services by alleged helpers or street vendors going from house to house, and the information is later used in the inevitable negotiations to cross the border.

In a context of mobilizing borders to immobilize people and the related spatial ambiguity of borders and questioning of established relationships between state, border, and territory (Szary and Giraut, 2015; Yuval-Davis et al. 2018), the same non-linear, relational effects apply to roads. Numerous police checkpoints have been set up on the main roads leading to Kantchari, spatially stretching the border hundreds of kilometers away from the border post itself. Shortly after Ouagadougou, for example, the identity papers of the passengers of overland buses are checked and those suspected of being candidats à la migration are prevented from continuing their journey. This procedure is repeated at 13 other posts on the 380-km route from the capital to Kantchari. However, here too, migration control is not simply accepted and endured, but circumvented and sabotaged by both officials, migrants, and their helpers.

Already in Ouaga and Fada, we ask the passengers who have no or, shall I say, strange travel documents to contact us. They give us 1,000 to 2,000 CFA francs per checkpoint and we give them to the police officers each time. Those who refuse to give us the money to negotiate with the security forces defend themselves at each checkpoint. But this is usually more expensive and sometimes ends badlyFootnote 4.

So, while the main roads can be considered important infrastructures of (negotiable) control, the countless tracks and sneak paths leading through the bush in the borderland to Niger are successfully used for undocumented border crossing. Some of the dirt roads and tracks in the borderland of Kantchari and Makalondi are very old and are kept in good condition by their users. These are, on the one hand, the smugglers, traffickers, and ordinary borderlanders who do not want to make a diversion for a documented border crossing. On the other hand, there are also representatives of the state security authorities who are equipped with all-terrain motorbikes, as well as Koglweogo groups. The Koglweogo, the “bush defenders,” are an armed self-defense group that provides security on a local level (Frowd 2022). They have enjoyed an enormous boom over the last 10 years as a result of the general insecurity in Burkina Faso and were, at least initially, able to quickly bring the situation in the borderland around Kantchari under control with brute force.

Of course the small trails are better for us, but you know, the Koglweogo are hunters and the hunters know everything here. Even at night they don’t get lost. So if you don’t have an agreement, you have to watch out for the police, the army, the customs officers and the Koglweogo. Even if they do the same you know. If they stop you, it will be expensive and if you don’t have money, you’ll go to prison. The migrants and me tooFootnote 5.

An important characteristic of vigilante groups is a certain ambivalence regarding their very objectives (Bateson 2021). In Kantchari, the Koglweogo support the state’s efforts to control the border but at the same time work as traffickers, when paid, or else blackmail those involved. In so doing, they themselves after a while contribute to local insecurity just like other groups.

MIDAS

Today, biometrics, blockchain-based technology, artificial intelligence, and machine learning are employed worldwide for so-called migration management. The Migration Information and Data Analysis System (MIDAS) is particularly used by West African states and has become a central component of migration control in the region (Zandonini 2019). MIDAS makes the border one of many biopolitical sites where mobile bodies are treated primarily as carriers of controllable identifiers (see Amoore 2006) and signifies sovereign territorial power that enables membership in the international community of technically biometrically capable states. The IOM is the developer of MIDAS and presents it as a user-friendly and low-cost Border Management Information System (BMIS) that automatically captures biographic and biometric data through the use of document readers, webcams, and fingerprint readers (IOM 2018b). IOM claims that running a BMIS is necessary to ensure regulated and safe mobility (IOM 2018a).

In Burkina Faso, IOM has deployed MIDAS at 14 locations, including fixed and mobile border posts, and continues to work on linking MIDAS with the Personal Identification Secure Comparison and Evaluation System (PISCES), which was initiated by the US Department of State Terrorist Interdiction Program in the late 1990s and operates at numerous capital airports in West Africa, including OuagadougouFootnote 6. From 2018 to 2020, MIDAS in Kantchari also consisted of a computer, a scanner, a reader for passports and other identity documents, the fingerprint reader, and a webcam on a tripod. In addition to biographical and biometric data, IOM-trained immigration officers collected and checked travel documents, entry and exit data, visa data, and the data of vehicles crossing the border. Via the main server, based at the General Directorate of Border Police in Ouagadougou, it was connected to the European Border and Coast Guard Agency (FRONTEX) in Warsaw. MIDAS is thus part of a colonial logic in current border control and migration policy. Gross-Wyrtzen and Gazzotti (2021) have pointed out the manifold colonial formations in the negotiation of (im)mobility between the formerly colonized and colonizers.

We wanted to go to Niamey to work with our brother but they don’t let me go. They say that there is a problem with my fingerprints and that we want to go to Europe. They call us ‘candidats à la migration’ you know. And now they don’t want to give me back my passport. We tried everything and also called our brother in Niamey. But they say it could be anyone and that they don’t believe us.Footnote 7

As this quote from an interview with one of two brothers from Guinea illustrates, in individual cases MIDAS has been used effectively to prevent migrant mobility. However, our empirical data indicate that the opportunities and incentives to circumvent MIDAS, or to use it in a way that was not intended by IOM, were too great for both immigration officers and travelers for it to have a significant impact on cross-border mobility. These two brothers, for example, after several days of waiting to be able to travel on, ultimately paid a broker to get their passports back from the border officials, as well as to be taken on a motorbike across the border irregularly, about 5 km north of the post. On several occasions, immigration officers complained that MIDAS was consuming too much electricity and that the solar panel could not therefore power the air conditioning sufficiently. In fact, the ventilators were only circulating hot air and officers became increasingly aggressive in checking the occupants of packed coaches with MIDAS. As one of the officers put it:

We become slaves to a system that, far from making our job easier, causes us other problemsFootnote 8.

We have also regularly observed that MIDAS was claimed not to be working. This could not be verified in each case, but everyone confirmed that MIDAS was not running smoothly and, in particular, the scanner often failed. “La machine est en panne” (“the machine is out of order”) became a catchphrase, also because regular external maintenance servicing was not provided. Infrastructures require active adaptation to the local context. In this perspective, in a place such as Kantchari with fragmented power and accountability structures, both real and artificially created technical problems with the system gave the immigration officers the excuse to return to handwritten registers and so regain their previous agency: they can let people without adequate travel documents pass the border in exchange for a small bribe.

However, the border-crossers have also quickly come to terms with MIDAS and have even been able to use the weaknesses of the system, if you will, to facilitate border crossing. For example, like the border officials, MIDAS cannot read Arabic script. This means that identity documents from Mauritania and other Arab countries were merely photographed after a tedious dialog between French-speaking officials and Arabic-speaking travelers, and the holders were then let go. This is the reason why Mauritanian passports are popular and are also used by migrants of other nationalities. Furthermore, veils and head coverings make biometric facial recognition impossible. For supporters of full-face veils, to force a veiled woman to show her face is an unforgivable offense. Tuareg and men from Mali and Niger, for example, also often vehemently refused to remove their turbans at the border, even temporarily, as this is perceived as dishonor. As a rule, these border-crossers were only allowed through after a manual check of their papers. In a region where 95% of the population is Muslim, head and face coverings are also an easy way to circumvent a culturally non-adapted MIDAS system.

Technological infrastructures require active adaptation to the local context and there is no such thing as straightforward implementation. They do not function in the exact same way as intended but are rather modified and molded to different contexts and change their function. Digital border controls are a prerequisite for fulfilling the current norm of sovereign territoriality through biometric statehood (Frowd 2018). Border Management Information infrastructures promise a techno-neutral separation between desirable mobility and undesirable mobility, often classified as a threat. However, by analyzing how these notions of modernity and biometric statehood are undermined by the actors’ infrastructural practices, we intend to fill the “empirical gap between discourse of biometric capability and operational realities” (Singler 2021: p. 463). MIDAS, as a heavily subverted techno-solutionist intervention, serves at most as a showcasing of territorial statehood. The ideas that are inscribed into that MIDAS technology are obviously those of a seemingly universal, but in fact selective sedentarism-implementing border and migration management system: movement as a threat, with control and stasis as a desired state; biometric citizenship as a prerequisite for the right to move (more or less) freely. MIDAS is characterized by the persistence of the preceding structure, carries within it its strengths and weaknesses. As an example, it shows that infrastructures both shape and are shaped by the conventions of a community of practice.

We have seen in the two empirical sketches of the ghettos and MIDAS that the production of migrant (im)mobilities mediated by infrastructures requires people who are in relation to each other in their infrastructuring work. The following section will therefore be devoted to social infrastructures and the concrete acts and contexts of social collaboration, for which Abdou Malique Simone (2004) proposed the concept of people as infrastructures.

Brokers, Traffickers, and Spies: Social Infrastructures

Xiang and Lindquist (ibid: S133) noted that in some cases “…the social directly takes an infrastructural form.” Referring to Simone (ibid.), the authors cite the lack of physical infrastructure and of job opportunities as examples. We argue that the Kantchari borderland in its demonstration of uncertainty, negotiability, and volatile state regulation of (im)mobility is also such a case. In contrast to Kleist and Bjarnesen (2019: p. 8), who claim for West Africa a “relative absence of brokerage as a constitutive dimension of migration infrastructure,” brokerage is of the utmost importance for infrastructural practices in Kantchari. There are multiple intermediaries on various levels who profit either from enabling or disabling migrant mobilities. Brokerage is the work that mediates between migrants and the migration and border regime and was already addressed early on in the course of approaches to the so-called migration industry (Hugo 1996) or migration channels (Findlay and Li 1998).

I was already in contact with my passeur via voice messages since KétaoFootnote 9. I sent him 40,000 francsFootnote 10 via Flooz to Orange Money so that he would organise my journey to Ouagadougou and from Ouagadougou to the border. He put me in contact with the driver who was to leave from Cinkassé for Ouaga two days later. The evening before, I took a taxi to arrive before 7am as planned. Idrissa then drove me to the main station in Ouagadougou and put me in another taxi heading for Kantchari. I then paid nothing during the whole journey until Kantchari. It was the driver who spoke to his mates from the police at every checkpointFootnote 11.

As this quote from an interview with a migrant from Ketao (Togo) and our observations while traveling from Ouagadougou to Kantchari make clear, people can be relied on as infrastructures long before the border. Voice messages, mobile money, and a well-coordinated team make it possible for the organizer of the journey in Kantchari (“passeur” in French) to remain in the background for the time being. The drivers who travel the route on those days when they meet their contacts at the checkpoints on the way to the border usually facilitate a safe journey.

Once in Kantchari, the passeur accommodates his client in one of the ghettos and discusses with him the best way of crossing the border and the costs involved. Whether someone is simply accompanied across the official border crossing, or brought to Niger by secret means, depends on the migrant’s financial means and the risk he or she is ready to take. The first option is safer and hence more expensive. People pay between 40,000 and 50,000 francs to cross the border without identity papers or for not being registered. Especially during the MIDAS period, identity was hidden in order to avoid being deported to the country of origin in case of a possible later failure of the migration project. Of the 50,000 francs paid for legal border crossing, half remains for the passeur. The other half is divided among the officials, including those who are not on duty, and their superiors in Ouagadougou and FadaN’Gourma. As we were able to observe time and again at the border crossing, and as was confirmed to us many times, further journeys announced and prepared in advance by the passeurs are rarely a problem. Even when MIDAS was still functional, the customer usually just stayed in the taxi or on the bus and waited for the controls to be completed. These checks are often lengthy because migrants traveling on their own are often detained for hours, questioned, and prevented from continuing their journey. But even in such cases, ad hoc interventions by brokers are still possible in principle. Border police and immigration authorities also repeatedly use mobile traders or coffee sellers to spy on travelers and then stop them if they want to.

One of the border officials told us:

The area is too big and the tracks from one hamlet to the next are not safe. We are mostly not from this area and cannot say that we have the area completely under control ... You have to turn a blind eye sometimes and not complain too much. The Koglweogo are useful in fighting illegal migration. When we get pressure from our superiors, we are also forced to tighten the crackdown because the authority wants to see results. Otherwise, we are human beings and you cannot refuse someone who wants to afford a better life abroad to leave the country. We try to come to an agreement with the people, but we also have to do our work. So, you have to combine that. You can’t question the sovereignty of the country, but you also can’t work on the border without earning something. In the past, people used to say: “Two years on the border and you can build a house”. But it’s not that simple. The bosses who sent us here also want a shareFootnote 12.

Crossing the border through the bush is less expensive and costs between 25,000 and 35,000 francs. Smuggling migrants through the bush is less costly because the distribution chain is shorter. For example, the passeur can decide whether to give police officers or Koglweogo 5000 francs each or simply nothing. However, there is the risk for both migrants and their helpers of being caught by the Burkinabe and Niger border patrols, or by the Koglweogo, as well as the military, which, for example, in view of the frequent armed attacks in the tense security situation, suspects all motorcyclists encountered during the night of terrorism.

For example, if you have to take migrants across the border and you call the policeman who tells you that it is too hot or who consistently rejects your call. Then you shouldn’t dare to take your clients to the border. You have to wait when everything is relaxed.... The policeman himself calls later to give a time. There are times when the policemen tell us not to take certain routes because the patrols will be tough. We then wait. And then there are the Koglweogo. No problem on the Burkina side. But in Niger they are tough. They always want to gain bigFootnote 13.

Brokerage infrastructures enable migrants in Kantchari to find their way through complex systems that were initially designed to prevent mobility, and also enable a wide range of actors to make a living out of(im)mobility.

You have to know how to conjugate the verb to eat. I eat, you eat, we eat, you eat.... In what we do, you have to be very careful and know the codes, otherwise you’ll get caught. The police need us and we need themFootnote 14.

Brokers negotiate on behalf of border officials and traffickers, arranging places to sleep and even temporary work. Migrants’ mobilities depend on the transfer of documents, data, trust, and money in the interplay of state regulation and profit-oriented brokers. As Rai (2020) has shown for India, brokers are part of social infrastructures of migration in which the boundaries between altruistic support and exploitation cannot be clearly drawn.

Conclusion

Infrastructures and infrastructuring practices in the Kantchari borderland produce both site-specific logistics and logics of (im)mobility. We understand the externalization of border control and the relocation of migration management from Europe to Africa as being constituted in and through multidirectional relations between new and existing infrastructure practices of border control and migration management, and relations between people, practices, and infrastructures. Hence, rather than considering it as a discrete, given entity, migration must be understood through the processes by which migrant (im)mobilities are made through such relational practices. The modes of relating that feature in our empirical work include hierarchy, parallelism, adaptation, convergence, acceptance, appropriation, rejection, and resistance through forms of non-relation such as detachment or withdrawal.

It is not a new insight that migrants are not only victims but also have agency and various resources to keep moving (Scheel 2018). They not only create and mobilize infrastructures to facilitate their mobility en route, but also take advantage of those that are supposed to stop them. However, there are always unintended and unpredictable effects. Migrants, traffickers, local vigilante groups, or border officials make alternative uses of infrastructures that never really work as the planners intend. While migrants appropriate infrastructures of migration control to maintain their mobility, immigration and security services use the infrastructures that are primarily intended to help migrants to enforce immobility. As the “autonomy of migration” approach has pointed out, migration control is not simply accepted (Scheel 2019). Rather, those targeted by these measures develop a specific way of dealing with the infrastructures implemented for this purpose. Depending on the situation, they are avoided, tricked, subverted, appropriated, and sabotaged not only by migrants but by all actors involved.

With a focus on infrastructuring as a practice that brings together actors, technologies, and political and economic orders, we revealed the different constituents that it consists of and the effort involved to keep it running. As pointed out by Xiang and Lindquist (2014), the infrastructures studied here also overlap in many ways in their production of migrant (im)mobility. There is not only MIDAS, but technology is always involved, and economic aspects and social networks run through all infrastructures, which interrelate, overlap, and clash.

Like other infrastructures, those of migration (control) are dynamic socio-technical formations, which, under the particular spatialities of a borderland, connect the economy, politics, and society (Amin and Thrift, 2017). There are many ways (im)mobilities are brought into being by these infrastructures. By studying the processes of making “irregular migrants,” “human traffickers,” or “corrupt immigration officers” through diverse relational infrastructures, we have shown that migration is not a discrete, given entity. We have also shown how established binaries such as mobility/immobility, functional/dysfunctional, crime/law enforcement, state/non-state, or control/freedom collapse (Tazzioli 2019). Rather than reifying precisely demarcated groups, people, and interventions, we focused on fluid and entangled sets of actors and relations. Entangled (im)mobilities served as an analytical lens to grasp different modes of how actors, practices, powers, and knowledge around physical, technological, and social infrastructures relate in this specific (im)mobility regime of the borderland. They also served to make visible a specific effect of these relations that we consider a kind of acquiescent subversion of migration control.