Channeling Contraband: How States Shape International Smuggling Routes

Abstract Although smuggling is commonly assumed to happen in remote and difficult-to-access borderlands, in reality, smuggling is most prevalent in areas that states tightly control, including at formal border crossings. To understand this puzzle, this article explores the relationship between states and smugglers at international borders. Based on extensive empirical research in various borderlands in North Africa and Southeast Asia, it argues that different kinds of smugglers prefer different types of relationships with the state. The article outlines six ideal types of such relationships. It contends that these types of relationships are the dominant factor in how different smuggling networks choose routes along a border. The findings have implications for our understanding of smuggling and policies that aim at addressing smuggling, especially regarding the effects of border fortifications and corruption prevention.

For over 400 kilometers, the border between Tunisia and Libya stretches across deserts, lakes, and mountains. The border provides a source of income for one of the largest smuggling economies in North Africa, with trading in everything from gasoline, fabrics, and electronics to cigarettes and narcotics. Its length and terrain make the border difficult to control for both Tunisian and Libyan security forces, and offer plenty of opportunities for the thousands of smugglers who cross the border every day to stay undetected or leverage their superior knowledge of the terrain in highspeed car chases with soldiers and customs officers. And yet, most of these smugglers have chosen two particular points in the border to conduct their business: the formal border crossings as Ras Jedir and Dhiba. In fact, different smuggling networks prefer different routes: while those trading in mobile phones, bananas, or alcohol have typically preferred to play cat and mouse with the soldiers through the desert, those trading in gasoline, carpets, or microwaves have preferred the official border crossings.
Similarly, the border between Thailand and Malaysia stretches across mountains and rough terrains, making large parts difficult to control and monitor. An ongoing violent conflict on the Thai side of the border adds another layer of complexity when navigating the territory. However, some parts of the border, such as in the province of Narathiwat, are more easily accessible and well connected on both the Thai and Malaysian sides of the border. Because of the ongoing conflict and accessibility factors, the border here is very securitized, with an elevated presence of soldiers, police officers, and border guards. Nonetheless, this part of the border is a particularly popular spot with smugglers, preferred over the less controlled but more difficult to access paths in the countryside. At the Golok River, which defines the border between Thailand and Malaysia in Narathiwat, large warehouses are a prominent sight on both banks. Small ships, heavily loaded with smuggled T-shirts, rice, cooking oil, cigarettes, alcohol, and in some cases drugs, go back and forth between the warehouses in the two countries, often only meters away from the official border crossings. Although security forces are very present, they do not intervene; the owners of the warehouses pay a monthly Flat Rate "fee," which allows them to smuggle as much as they can.
What makes some smuggling networks choose to operate at points of maximum state control, while others prefer to stay off its radar? This article explores how the relationship between smuggling networks and state structures shapes smuggling routes. We argue that beyond geography or border security infrastructure, the nature of the interaction between smugglers and state structures is the most powerful predictor of the routes through which different smuggling networks prefer to operate. At the heart of our argument stands the observation that most smuggling networks do not desire to operate in an environment where third-party rule enforcement by a local monopolist of power is entirely absent, as they rely on some of the services that the monopolist provides. However, as we argue, different forms of interactions between smugglers and state structures typically generate different costs and risks for smuggling networks and have crucial implications for the respective smuggling economy's market structure. Due to different organizational capacity, access to financial or social capital, features of the goods they transport, and risk aversion, networks have heterogeneous preferences regarding their interaction with state structures along borders, thus determining their choice of route.
This discussion implies significant stakes both for the study and practice of border security today. Globally, the last two decades have seen an increasing focus on border security as a critical security challenge in the context of a globalization of crime and the risks of globally operating armed groups. Traditional thinking about border control within security studies and policymaking typically relies on at least two strong assumptions that position smugglers and state structures as antagonists. One makes the interdiction of smuggling primarily dependent on state capacity and the minimization of individualized corruption. The other interprets smuggling as inherently subversive and the existence of large-scale smuggling as an indication of state weakness and fragility. In policy practice, both are associated with heavier investment in border security as a way to strengthen states, especially in developing countries. 1 The existence and frequency of a wider set of interactions between state structures and smuggling networks, and these interactions' role in shaping smuggling routes, therefore, have critical consequences for the analysis not just of border security in particular but also "state capacity" and "state fragility" more broadly. Here, it connects to a wider set of questions about how states secure borders that scholars of smuggling have highlighted in recent years. Critically it notes that global investments into border infrastructure need to be analyzed not necessarily as measures that interdict smuggling but as measures that change the structure and routes of smuggling streams in more complex ways.
To explore this argument, we utilize empirical data from a variety of borderlands to develop six ideal types of interaction between smuggling networks and state structures: Genuine Enforcement, Toleration, Flat Rate, States as Smugglers, Cat and Mouse, and Petty Corruption. We analyze these interactions with respect to their desirability for different network types. While this represents a first step in the study of typologies of smuggler-state relationships, it already allows us to draw conclusions regarding the effectiveness of border security infrastructure and anti-corruption measures on combating smuggling, the features of point-of-entry smuggling, and the effects of armed conflicts on smuggling routes.
Our analysis is built upon extensive empirical fieldwork and interviews with both smugglers and state agents we conducted in nine different borderlands between 2014 and 2018. Through its empirical strategy, the paper aims to fill a crucial gap in the literature on border security within security studies and the literature on smuggling more widely. Work that rests on single case studies or regional analyses dominate these literatures, but analyses that bring together borderlands across regions and continents, to build toward larger, global arguments and theories, are extremely scarce.
The rest of the paper is divided into five sections. The first reviews some of the existing literature on the interaction between state structures and smuggling networks. The second presents the methodology and empirical basis of this paper. The third section outlines a typology of different interactions between smuggling networks and state structures. The fourth discusses a range of implications based on this typology, and the fifth section summarizes our overall conclusions.

Smuggling in security studies
Although the degree to which states have historically exerted control over the porosity of their borders is commonly overstated, it has come to symbolize a crucial element of state sovereignty. 2 Few topics have been so dominant in contemporary political discourse on state sovereignty and security as the securitization-and walling-of borders.
A common assumption is that a key reason why states try to control their borders is that they want to stop smuggling and enforce custom regulations. The relationship between states and smugglers is viewed as one of antagonists, in which states try to enforce the laws and arrest and prosecute smugglers. Meanwhile, smugglers are assumed to be trying to evade prosecution and are therefore forced to operate in the shadows. If smuggling does nonetheless occur, according to this logic, it is the consequence of a state lacking the capacity to enforce its rules.
This understanding of the state-smuggler relationship dominates the security studies literature; it commonly is an implicit assumption, but in some cases is stated explicitly. For example, scholars such as Louise I. Shelley argue that transnational crime and terrorism flourish due to a "decline in state capacity" and a "retreat of the state." 3 Along similar lines, Robert I. Rotberg describes "smuggling" and "corruption" as indicators of "state fragility," assuming states are subverted and hollowed out through their interaction with organized crime groups, and Ulrich Schneckener considers smuggling to be a main "risk area." 4 Similarly, much traditional scholarship on global illicit trade has framed the relationship between smugglers and states as one of enforcer and evader. Smuggling is presented as part of an economy that is "underground," 5 hidden, or in the "shadows," 6 and seeks to evade interactions with state structures. Another set of studies characterizes the interaction between smuggling and states as one primarily created by corruption within the security services, giving rise to relationships that are typically clientelistic. From this literature has emerged a "crime-fragility rationale" 7 similar to the one in security studies.
In the policy world, based on the same understanding of the relationship between states and smugglers, international organizations and bilateral development programs aim at curbing smuggling and illicit trade by strengthening the capacity of states considered to possess limited control of their borders and weak enforcement of their customs regulations. As Ron E. Hassner and Jason Wittenberg note, while investments into border fortifications designed to limit uncontrolled cross-border movements have expanded dramatically in recent years, political science has largely neglected the politics of border control and border porosity. 8 As Peter Andreas points out, siloization of scholarship between security studies, international relations, and international political economy has contributed to a lack of a unified debate, divergent conceptual approaches, and a obscuring of a discussion on the interaction between smuggling and state structures. 9 As a result, terminologies have also diverged. We use the term "smuggling" to describe all forms of cross-border trade that violate the formal law applicable at the respective border. We differentiate between licit and illicit goods, with licit goods referring to those for which a legal trade corridor exists that is not subject to additional security clearance. Examples of illicit goods include narcotics, expired medicine, firearms, endangered animals, or archaeological artifacts. 10 More recent literature on smuggling has highlighted the existence of stable and systematic interactions between smugglers and state and nonstate regulatory actors. For example, in her work on Somalia 11 and Afghanistan, 12 Aisha Ahmad has pointed to the predictability of rules around the taxation and regulation of the flows of illicit goods as key factors fostering mutually beneficial relationships between both nonstate armed groups and protostates. One of us has noted elsewhere that smuggling in North Africa is commonly structured through impersonal institutions that are negotiated between state institutions and smuggling networks. 13 Similarly, the literature on war economies points to more structured engagements between smugglers and state structures. 14 These findings overlap with an expansive anthropological literature, particularly on African borderlands, that has chronicled how interactions between state and nonstate actors in borderlands have generated forms of order not sufficiently described by evasion or corruption. 15 The insights of these studies have not been appropriately reflected in security studies and its thinking on borders. In addition, while the anthropological literature contains significant work on the relationships between smugglers and state officials at borders, efforts to create structured typologies of these relationships 16 and theorize their effects on state structures, the effects of border fortifications, and smuggling routes are lacking. This paper aims to address this gap.

Methodology
Empirical work on smuggling and informal and illicit forms of regulation provides unique methodological challenges. 17 Understanding practices that are frequently hidden, misrepresented, or conducted at the very periphery of the state requires deep contextual knowledge and access to local networks. As a result, the study of the illicit suits itself particularly well to ethnographic methodologies and has been dominated by single case studies.
We follow this practice in the field by also drawing extensively on ethnographic methods: the data collected for this paper is the result of political ethnography, interviews, and historical institutional analyses. This has allowed us to trace practices that are commonly unwritten and embedded in everyday social practice, triangulate information through different forms of data, and build relationships that allow both conducting of interviews on such sensitive topics and the contextualization of their content. In this we follow a recent trend in political science that has advocated for the use of political ethnographies, especially in the study of the illicit. 18 We do not, however, limit our analysis to a single case study. This paper draws on data collected in different borderlands across two continents, focusing on Southeast Asia and Africa. In this, we aim to contribute to the use of comparative methods in the study of smuggling and highlight that the patterns discussed in this paper are not limited to a particular geographic, cultural, or political context. The borderlands considered here have been selected following two broad considerations. One was the ease of access and the ability to conduct sensitive research in these environments, which has been facilitated by our research experience in these areas. The other was a desire to include a set of borderlands that is sufficiently diverse to be able to examine the dynamics discussed here in varying environments. The borderlands presented here vary in their geographic position and geological makeup, and include what Gregor Dobler referred to as "green," "grey," and "blue" borders. 19 They include urban and rural borderlands, the existence and absence of formal checkpoints, and varying positions within wider global licit and illicit trade corridors.
The analysis conducted here draws on author-collected data through over two years of field research conducted between 2014 and 2018 in nine borderlands: Bangladesh-Myanmar; Myanmar-China; Myanmar-Thailand; Thailand-Malaysia; Malaysia-Indonesia; Indonesia-Philippines; Tunisia-Libya; Morocco-Algeria; and Morocco-Spain. In these borderlands, we have conducted over 350 qualitative interviews with smugglers, bureaucrats, local security forces, civil society activists, politicians, and journalists. This has been supported through participant observations at border crossings and in informal markets across the borderlands. 20 Naturally, conducting qualitative and ethnographic fieldwork on smuggling networks and in borderlands presents a wide set of methodological challenges. As a full review of these challenges goes beyond the scope of  19 Dobler, "The Green, the Grey and the Blue." 20 As smuggling typically is a dynamic activity, it is worth noting that the dynamics described here may have shifted since their observation in the context of this project. For example, by the time of writing, large parts of the smuggling activities between Morocco and both Algeria and the Spanish enclaves of Ceuta and Melilla have collapsed. However, as the purpose of this paper is to point to the diversity of arrangements, we do not believe this detracts from the arguments made here. this paper, we highlight three central aspects here and note how they have affected the data collection and analysis of this project. 21 First, in working with borderland communities and smugglers and security forces in these borderlands, gaining access and building trust are common challenges for researchers. In the context of our research, these challenges have primarily been mitigated through building relationships over time, through repeated and extensive engagement with interlocutors, as well as the mediation of local intermediaries. Long-term and repeated research in the regions, as well as a combination of lengthy interviews and participant observation, has also helped us better understand how to frame questions about smuggling and illegal activities without suggesting judgment or disapproval, which has further helped build trusting relationships. Here, the existence of structured relationships between smugglers and security forces in the borderlands that we describe below has helped the data collection somewhat, as some activities have been normalized and tolerated enough that participants did not fear detection or exposure by talking with us about their activities.
Second, however, given the illegality of many of the activities involved, protecting the security and anonymity of participants is crucial in conducting qualitative fieldwork on smuggling. Common mitigation strategies in this project have included focusing in interviews on questions around structures rather than individuals to avoid pressing participants on personal data (for example, by asking not for the name of a supplier but whether someone typically uses the same supplier).
Third, despite all mitigation strategies, from long-standing relationships to careful question development questions and an effort to speak to diverse informants and groups, we recognize that misreporting and gaps in our access are still substantive challenges in research on this issue and our own data collection. Although we have sought to extensively triangulate the observations on which we base our argument, data verification was a key challenge due to the nature of our research subject. However, we are confident, due to the wider focus and inductive approach of this article, that the broader patterns we describe are robust to the empirical challenges in the more particular case observations.
As is fitting for its purpose and subject matter, this paper is built on an inductive research design. We begin by mapping six patterns of relationships between smuggling networks and states and then draw on these patterns to analyze deeper drivers and implications.

Ideal types
Conducting extensive field research in different borderlands across the world, and drawing on ethnographic case studies of smuggling from a variety of fields, we repeatedly found the same structuring patterns of state-smuggler interactions. These patterns, presented below, can be seen as "ideal types" of state-smuggler interactions. Drawing on Max Weber's understanding, ideal types constitute methodological utopia-models that help us categorize and systematize empirical observations. 22 However, going beyond Weber, the suggested categories also aim to explain why state-smuggler interactions are structured in a particular way, constituting patterns of explanations (Erkl€ arungsmuster). It is important to note that Erkl€ arungsmuster can overlap or exist in parallel in an empirical setting. For instance, different state authorities may engage differently with different smugglers in the same border region, or the same state authority may engage differently with different smugglers.
Before turning to the typology, some caveats are in order. We outline patterns of interaction between smugglers and states as snapshots of a wider interaction, which may change over time. Our intent is neither to claim that these relationships are stable, nor provide a complete explanation of the underlying factors and institutional histories that let one form or relationship emerge in one place and a different one in another. Our intention is instead to highlight the existence of diverse but patterned relationships and their relevance for both smugglers and states. As we present ideal types, we also do not wish to claim that the elements of the relationship between smugglers and state structures we highlight here capture the empirical entirety. Various forms of local formal and informal institutions, histories, shared social ties and understandings, moral conceptions, and ambitions shape the relationships between smugglers and state institutions. We highlight a fraction of them, as we see them pointing to patterns that explain larger processes, but we do not lay any claim to providing an exhaustive picture of this relationship.
From these caveats and the approach through which the ideal types have been generated emerge some limitations for the structures presented here. Developed based on both our field research and interactions described in the wider literature on smuggling, the ideal types outlined here are 22 What we characterize as ideal types here could, in line with Weber, also be described as Gattungsbegriffe (class or generic concepts) that we developed inductively from our empirical findings. According to Weber, there is a gradual transition from Gattungsbegriffe, used to classify empirical phenomena, to ideal types, which consider "complicated historical patterns." While empirical findings underpin our typology, it nonetheless presents analytical utopia that overemphasize characteristics in line with Weber's understanding of ideal types. While we therefore decided to talk about ideal types, due to the gradual difference between Gattungsbegriff and ideal type, either label applies for our typology. primarily inductive. Consequently, they are intended to be representative but not necessarily exhaustive. This does not limit the main claims of the paper: that these forms of interaction present an important alternative to simplistic antagonistic views of the relationship between smugglers and states, and that they are central to how smuggling networks choose routes. Room exists for the development of further ideal types or reordering of these ideal types without detracting from these wider points.
We describe these ideal types as we see them on one side of the border-either on the exporting or importing side with respect to a particular set of goods. This is not to imply that they are necessarily limited to one side of the border, or that they are necessarily mirrored across the border. Similar ideal types may appear on two sides of a border, or they may differ dramatically upon entering another territory.

Genuine enforcement
The most common conceptualization of state-smuggler interaction is one where state authorities genuinely try to enforce the law, and attempt to detect and eliminate smuggling, while smugglers attempt to evade law enforcement. Hence, smugglers avoid interaction with state authorities. If a high level of state capacity accompanies Genuine Enforcement it makes smuggling costly. For instance, it is costly to avoid state detection and to find and hire people willing to take the risk. This creates high barriers of entry to the smuggling business. Genuine Enforcement is the normative ideal of what we usually expect states to do, and this ideal type dominates much of the academic literature and policy discourse.
Genuine Enforcement may, for instance, be found at Singapore Changi Airport, as the Singaporean state has the capacity and the interest to attempt to detect drug smuggling. Similarly, gasoline smugglers in the Moroccan cities of Oujda and Berkane told us that although customs agents at the border were highly corruptible, their experience suggested those agents controlling the roads further inland were following a different regime-cars caught transporting larger quantities of contraband gasoline would typically be confiscated alongside their wares, and the drivers arrested. The difference in these interactions could be observed to have shaped both the drivers and their cars. Aiming to avoid all interactions with the state, gasoline transporters moving contraband gasoline within Morocco would primarily drive without lights at night, at high speed, and at tremendous risk to their health and those of other drivers. In the borderlands of northern Morocco, this has earned them the local nickname "muqatila"-fighter.

Toleration
A state-smuggler interaction may also be characterized by high degrees of Toleration. In this case, state authorities tolerate the smuggling of a certain amount of goods and/or of particular types of goods. Hence, there is no enforcement of formal rules that prohibit smuggling, and smugglers do not need to pay bribes to state authorities. This is an informal, but regulated and institutionalized, practice, which does not depend on individual officers being on duty. This type of interaction may be driven by a lack of state capacity to deal with all cases of smuggling. Alternatively, it may be based on a conscious decision to enable survivalist activities and to not disrupt the activities of small businesses and other "little fish," such as the owners of small shops, who smuggle daily consumption goods such as fruits, rice, and cooking oil. At times, there may be an upper limit to the goods that people are allowed to smuggle. Nonetheless, all people benefit from this Toleration in the same way, as they can smuggle the same amount or kind of goods. This ensures a high level of predictability and enables anyone to enter the smuggling business without facing any barriers.
We observed this type of state-smuggler interaction in the case of the Indonesia-Malaysia border region, in the Indonesian city of Nunukan in North Kalimantan. The city is only a short boat ride from Tawau in Malaysia. There are various small harbors in Nunukan, which we watched being used for the legal transportation of goods and people during the day and for the illegal transportation of goods, mainly ordinary consumptions goods such as gas cylinders, and migrant workers, hoping to find a job on a palm oil plantation in Malaysia, at night. In some cases, small police offices exist in the harbors. However, these usually only operate during daytime and are closed at night. Hence, even if not stated officially, local Indonesian police tolerate the irregular border crossings into Malaysia.
The border between Morocco and Algeria provides an interesting example of this form of Toleration as well. Although the land border between the two countries has officially been closed since 1994, smuggling of consumer goods and the movement of people remained highly prevalent across this border until the construction of additional fortifications in 2014. The smuggling of gasoline and consumer goods, such as textiles and cosmetics, was largely conducted through makeshift border crossings: doors in the border fence, which soldiers from the Moroccan military manned. Smugglers across these gates who we spoke to recounted that from early 2011 onward, soldiers stopped taking bribes from small-scale traders while still demanding them from wholesalers and larger traders. Smugglers themselves were not always in agreement about the origins of the shift, at times speculating that it represented a response to concerns about social instability in the region, at times interpreting it as a pro-poor policy of the Moroccan king. Although we could not conclusively trace the precise origin of the shift, it is worth noting that it introduced, for some years, a complete Toleration of survivalist smuggling across the border.

Flat rate
A third-way state authorities and smugglers interact can be described as a Flat Rate model. In it, state authorities tolerate smuggling operations in exchange for a fee that is fixed and paid on a regular basis-per crossing, per car, or per week or month, for example-regardless of the amount of goods smuggled (like a mobile phone provider that allows customers to make as many phone calls as they want for a certain fixed price). The payment is not continuously negotiated; it represents an institutionalized and regulated interaction. In a smuggling context, the payment of the fee can either happen at the point of a border crossing or it can be arranged beforehand. For smugglers, this model ensures a high level of predictability regarding the costs of the smuggling operations and a low level of risk. Meanwhile, state authorities benefit from a stable level of (additional) income and a low level of work and enforcement activities. Contrary to the Toleration ideal type, a required payment represents a barrier of entryeven though the costs per item smuggled are likely to still be below the costs of a formal trade route.
Until 2019, the Barrio Chino border crossing between Morocco and the Spanish enclave of Melilla offered a particularly clear example of a Flat Rate model. 23 Every day, hundreds of individuals, commonly referred to by the press as "portadores" or "human mules," could be observed carrying large bundles of consumer goods into Morocco: clothes, foodstuffs, cleaning products. Their trade is illegal-the transporters do not pay any formal tax or tariff to the Moroccan customs officers. But they also do not pay a bribe. Most of the traders work as subcontractors for a small number of large-scale smugglers that operate through the border crossing. As some of the transporters pointed out to us, their wholesalers make a regularalthough illegal-payment to high-level officials within the Moroccan customs service to smuggle their goods through the crossing-a Flat Rate payment. They then mark their goods with numbers and symbols that identify them as theirs-around the border crossing, stacks of numbered bundles can be observed being unloaded from trucks and divided up between waiting transporters. In some cases, wholesalers even hand out "tickets" to transporters that they can show to customs officers to identify their goods as a particular wholesaler's property. 24 Similarly, in Narathiwat Province, in the "deep south" of Thailand at the Malaysian border, the Flat Rate is a prominent model. As mentioned in the introduction, large warehouses line the Golok River, which defines the border between the two countries, and small boats transport daily consumption goods across it. For instance, we observed T-shirts being taken from Thailand to Malaysia, while petrol that is considerably cheaper in Malaysia is shipped to Thailand. This large-scale smuggling of goods often happens near bridges, which constitute formal border crossings with customs terminals and immigration counters. According to interviewed owners of such warehouses, they often have Flat Rate deals with officials in both countries, allowing them to transport as much as they want for a fixed fee that is shared between the various authorities that control the border, including customs, border police, and immigration.

State as smuggler
A fourth option is high-ranking state authorities being directly involved in the smuggling business. In this case, state authorities fully control the smuggling business in a particular sector or are, at the very least, the dominant actor. This model is particularly prominent in cases of smuggling illicit high-value and high-risk goods, such as drugs and guns. State authorities smuggle the goods themselves, they hire smugglers, or they partner with existing smuggling networks. Detection and enforcement are only practiced to deal with rival smugglers or smuggling networks, which refuse to cooperate with state institutions. This imposes high costs on competitors and can help to further monopolize the smuggling business. Here, in particular, it is worth noting that this ideal type describes the relationship on one side of the border-which could be either the importing or exporting side with respect to a particular good-and is not necessarily mirrored on the other side. The smuggling of arms to a separatist movement by government officials of a foreign country sympathetic to their cause, for example, could see a State as Smuggler model on one side of the border and a Genuine Enforcement environment on the other.
For instance, Cox's Bazar in Bangladesh gained widespread international attention in 2017 after the arrival of more than 700,000 Rohingya from Myanmar following state-driven violence against them. Cox's Bazar is near the Myanmar border and has long been a major transit hub for not only refugees from Myanmar but also for smuggled licit and illicit goods. Apart 24 Gallien, "Informal Institutions and the Regulation of Smuggling in North Africa." from licit goods such as instant coffee, a frequently smuggled illicit good is yaba, a methamphetamine produced in Myanmar popular in Bangladesh. It is an open secret in Cox's Bazar that a local former member of parliament and member of the ruling party and his family are key players in control of yaba smuggling from Myanmar in the area, using their state authority to protect their smuggling business and monopolize it. 25 Even interviewed Bangladeshi army officers were afraid and stated that they did not have the ability to act against such influential actors.

Cat and mouse
A more complex type of interaction between smugglers and state authorities may be called Cat and Mouse. In this case, a repeated "game" is played, in which law enforcement officers try to catch smugglers, who, in turn, attempt to remain undetected. If the law enforcement officers succeed, they are rewarded with a payment by the smuggler, often a defined amount. As in the case of Genuine Enforcement, law enforcement officers attempt to detect smuggling while smugglers attempt to evade law enforcement. However, in contrast to that ideal type, detection does not necessarily result in the full application of the law but may result in a bribe payment. Hence, detection is costly for the smuggler. State representatives' role is not understood as intended to eliminate the smuggling activity altogether, and the interaction contains an equilibrium state in which smuggling is prevalent.
As with the Toleration ideal type, this does not describe a case of Genuine Enforcement, and there is an acceptance on an aggregate level by state agents that smuggling occurs and that smugglers are not pursued outside of the "game." However, in contrast to the Toleration ideal type, this model imposes a cost on smugglers through a nonzero probability of capture/payment and repeat interaction. In this aspect, it is similar to the Flat Rate ideal type. But the interaction has less day-to-day predictability and hence presents a barrier of entry not only based on cost but also on risk aversion. Nonetheless, this type of smuggler-state interaction ensures a high level of predictability for smugglers on a macro level, as the rules are clearly defined and smugglers can include the costs for a certain number of detections in their monthly transportation costs. Furthermore, Flat Rate necessitates a certain knowledge of the actors to be able to identify those who have already paid or are covered by a specific payment. Conversely, the Cat and Mouse ideal type does not require such knowledge.
The border between Tunisia and Libya provides a particularly fitting example of this kind of model. Outside the few formal points of entry, the border runs through hundreds of miles of difficult terrain, across deserts and mountains. Smugglers hailing from small villages near the border cross it by night in powerful 4x4s, stocking up on tea, cigarettes, or gasoline in Libya and then driving back into Tunisia. The reason for the strong cars is not just the difficulty of the terrain, but a nightly game of cat and mouse with the local customs officers, who are waiting in their own powerful jeeps along the border. If a smuggler manages to avoid them, they are home free. If they get caught, they face paying a stiff bribe, but, crucially, they are not arrested. The next night, the game begins anew. The actors know each other, and one of the most popular smuggler cafes sits just across from the local customs office; the powerful cars that chase across the desert at night are parked side by side during the day. There is no systematic effort on the side of the customs authorities to eliminate the trade, and there are few arrests or raids on storage facilities. Custom officials who do not participate in the game are swiftly convinced to participate or transferred. And while the nightly trip through the border's difficult terrain bears real risks for the young men who drive the cars, the costs of an occasional bribe have been calculated into the profit margins of the larger wholesale smugglers who employ the drivers.

Petty corruption
Finally, state-smuggler interactions may be characterized by what is commonly called Petty Corruption. It describes an interaction where some state representatives tolerate some smuggling activity in exchange for a bribe. Hence, just like Genuine Enforcement, Petty Corruption closely resembles the state-smuggler interaction as it is imagined in much of the mainstream literature. However, in contrast to the Flat Rate ideal type, this practice is not regulated. It results in a low level of predictability for smugglers, as the relevant state official, the willingness of that official to forgo enforcement in exchange for a bribe, and the form and level of that bribe may all vary. As structured repeated interactions are not always ensured, corrupt state officials may have an incentive to set bribes at predatory levels. But contrary to the Cat and Mouse ideal type, no clear expectation exists on both sides of what happens in the case of detection. The barrier to entry is structured around the ability to bear the costs of bribes, not around access. However, the inability to calculate the costs makes the Petty Corruption model unattractive for professional smugglers.
For instance, police checkpoints that smugglers of licit goods in Tunisia pass as they transport their goods further into the country commonly fit this type. Corruption is widespread-most smugglers pass through the checkpoints with their goods after paying the officers. The payments are relatively unpredictable, however, because personal connections to individual members of the security services, or changes in the rotation, can make a crucial difference. Local smugglers commonly complain about this lack of predictability-an observation shared by Ahmad's account of smuggling in Afghanistan in the pre-Taliban period. 26 Table 1 below summarizes the interactions described in this section by highlighting their differences with reference to three aspects. The first is how predictable the interaction is for both smuggler and state agents; the second is whether bribes are necessary and whether their amount can be anticipated by smugglers beforehand; and the third refers to whether smugglers try to evade detection as a part of this interaction.

Choosing routes
A clear set of questions emerge from this variety of interactions: What shapes these interactions? Who drives their emergence? And why do smugglers-and states-engage with them? As we are building here from the empirical observation of these patterned interactions, we do not aim to provide a complete model of their origins, which would require further research, as suggested in the conclusion. Alongside the borderlands' extended geography, local histories of border-making, historical institutional path dependencies, and local normative conceptions of smuggling, the preferences of both smugglers and state representatives over different interactions clearly play a role. To support our argument that states shape smuggling routes, we focus here on outlining three related observations. First, the type of interaction available will affect the routes that smuggling networks choose. Second, the type of interaction networks choose will differ among networks depending on their features. And third, smuggling networks do not exclusively drive the types of interaction available-they are affected by states. Naturally, smuggling networks' motivations and practices are diverse and cannot be reduced to their interactions with states alone. However, clearly the different interactions between state structures and smugglers sketched above have heterogeneous consequences for the day-to-day business operations of smuggling networks. They differ with respect to the predictability of their interaction with state agents, the costs of these interactions, and the access to these interactions. It follows naturally that depending on their resources, goods, business models, connections, and capital, smuggling networks have different preferences over these features.
The dynamics outlined above, for example, suggest that the costliness of the interaction with state structures provides diverging incentives for smuggling networks depending on their profit margins and access to capital. Small-scale survivalist activities based in the borderlands will likely prefer routes that approximate the Toleration ideal type. Due to their embeddedness in local social networks, they may be able to operate in environments that approximate the Petty Corruption ideal type, whereas they would avoid more expensive environments such as the Flat Rate. More affluent networks, however, may prefer more expensive interactions if they bring additional benefits, such as the use of state-provided trade infrastructure, predictability, or the elimination of competition.
The responsiveness of different networks to the varying costs of different interactions will also depend on the costs of evasion-or the difficulty with which they can avoid detection. Networks that can avoid detection at relatively low costs, such as networks smuggling narcotics in the body of transporters, may prefer a Genuine Enforcement or a Cat and Mouse context, which would be substantially less attractive to networks that face much higher costs of evasion, such as those smuggling heavy artillery.
The predictability of their interaction with state structures also affects networks heterogeneously. Networks trading in very high-value goods but with a lower frequency of border crossings, such as cocaine trafficking networks, would have a significantly lower tolerance for uncertainty than networks trading lower-value goods, such as gasoline, in higher frequency. It is important, however, to conceptualize predictability wider than merely the likelihood of unexpected confiscation or the likelihood of paying a higher cost than originally anticipated. Unpredictability in the context of smuggling can also have a wider influence on people's lives and well-being. Predictability of interactions with state structures commonly correlates with the social normalization and legitimization of illegal economic activity in the borderlands. This can affect traders' social standing in local communities, their own career choices and preferences, and hence preferred cargo and routes.
Different forms of interaction between state structures and smuggling networks also have heterogeneous effects on the barriers to entry to the interaction, and hence affect the market concentration within the specific smuggling segment. The complete toleration of specific trades limits barriers to entry, and even actively counteracts a concentration of market power if it also includes maximum quantities that can be smuggled in a given trip. In contrast, Flat Rate interactions raise barriers to entry relative to the costs of the interaction, whereas State as Smuggler interactions typically include a complete monopolization of the trade. Naturally, the preferences of networks regarding barriers to entry are highly differentiated. They differ depending on the likelihood of a specific network being able to overcome the barrier and the additional profit that can be generated out of a higher market concentration. For a network trading in illegal narcotics, a Toleration interaction is less attractive than a Genuine Enforcement interaction if profit margins are primarily related to market concentration and the cost of evasion from rivals. The costs of interactions with state structures, the predictability of these interactions, and the market concentration all have direct and significant influence on smuggling enterprises.
We have focused primarily on smuggling networks' preferences over different interactions as a consequence of their business model and cost-risk trade-offs because these point to some of the clearest variation between networks outside of their local contexts. However, these are likely not the only factors: local moral economies, normative conceptions both of smugglers and the communities they live in, traditions of mobility, the presence of other nonstate regulatory actors, or political ties could all similarly shape networks' preferences. For example, normative perceptions about the permissibility of bribery could drive smugglers' preferences to opt for Toleration structures. At the same time, networks that see the state as an illegitimate actor that must not be collaborated with could purposefully avoid Toleration interactions. Similarly, perceptions about the impermissibility of trading certain goods could drive some networks to choose a different form of interaction than neighboring networks.
What follows from the above discussion is also that smuggling networks' preferences over what forms of interaction to engage with correlates with the types of goods they smuggle. Some illicit goods, such as narcotics or arms, for example, typically face substantially higher penalties in a Genuine Enforcement environment, which likely makes this less attractive. Depending on the environment, goods also differ in the costs and difficulty of evasion depending on their size and detectability-gold is much more easily stashed in the wheels of a car than a goat would be. Perishable goods such as meat, fruits, and vegetables also put time pressures on networks. That said, we suspect the relationship between goods traded and the preference for particular forms of interaction will remain somewhat imprecise. This is primarily due to the diversity of networks trading in many goods. Some gasoline smuggling networks, for example, are small scale and survivalist in scope; others are highly capitalized wholesale operations, which we would expect to have rather different preferences.
None of the discussions above seek to imply that networks always choose routes-that they always have choices available, that they always make this choice, or that there are no (literal) path dependencies. But it seeks to highlight that in contexts where choices are being made, the interaction between smugglers and state structures is a crucial aspect. To argue that states play a role in shaping smuggling routes, it is worth noting that these interactions are not purely created or shaped by smugglers, that state actors play a role in shaping these relationships. Naturally, this points to a question of what motivates state interactions with smugglers. Although a full discussion on the topic goes beyond the scope of this paper, recent work both in political science and borderland studies, as well as some of the literature on "state criminality," have offered a variety of starting points to explore these questions.
More traditional explanations for state agent involvement in illegal activities have foregrounded financial gain as a motive, and thus frequently tend to assume structures that mirror our Petty Corruption ideal type. 27 Other scholarship has noted how distributional politics and a need to provide incomes for otherwise neglected regions can lead states to tolerate smuggling, or how political acquiescence can lead to the toleration of other illegal economic activities. 28 Assuring domestic supply of goods on which trade restrictions exist can also motivate states to get more directly involved in smuggling: scholarship on smuggling between a divided Germany, for example, showed the involvement of high levels of the German Democratic Republic security services in importing rare goods. 29 It is also worth noting that states are not always most productively analyzed as unitary actors but may reflect different strategies and preferences with respect to smuggling within different parts of the state apparatus. Naturally, a dynamic relationship also exists between the preferences smuggling networks may have for certain interactions and the preferences of state structures. However, a more systematic discussion of this would require a separate study. From the considerations suggested in this section, we can offer an explanation for the puzzle presented at the beginning of the paper. In all borderlands studied in this project, most smuggling activity has been conducted through points of entry, where smugglers' ability to avoid detection is significantly reduced in comparison to "green border" territories. The relationships outlined above suggest smugglers choose to operate through border crossings. Border crossings not only provide advantages in terms of infrastructure; they can also offer advantages in terms of costs, predictability, social respectability, and market concentration. Which forms of networks operate through border crossings and which work around them, then, is crucially determined by the form of state-smuggling interaction at the border crossing.

Choosing routes: Two case studies
To illustrate some of the dynamics described above, we sketch out the geographies of interaction between smugglers and states along two borders and highlight how different networks choose their routes.

The Tunisia-Libya border
The northern section of Tunisia's border with Libya offers two different terrains for smugglers. One is the main point of entry, the Ras Jedir border crossing. The other is the terrain south of the border, a wild landscape of deserts, dry lakes, and mountains, which locals navigate with powerful 4x4s. We find significant smuggling activities across both terrains. We argue that different networks' taken route is critically influenced by the different interactions with state structures. Here, it is important to highlight, as we have noted above, that different interactions can overlap across the same terrain. 30 First, there is a layer of Genuine Enforcement that spans across both routes but is limited to specific goods. Anyone trying to bring cocaine or weapons across either route would typically expect to encounter a Genuine Enforcement environment, albeit individual instances of Petty Corruption may occur. On top of this, we can identify at least two more interactions at the border crossing. We find complete toleration for traders bringing in most licit goods as long as they remain under a certain value threshold per crossing-typically a full tank of gasoline, a few carpets, and a few home appliances. For larger quantities of licit goods, a Flat Rate agreement is in place, where traders pay a fixed fee per crossing, which is then divided between customs officers on the Tunisian side of the crossing and customs officers and local security forces on the Libyan side. Outside of the border crossing, we find the Cat and Mouse game described earlier, where traders crossing the desert in 4x4s by night often manage to avoid security forces but must pay a hefty bribe on the nights when they do get caught. Interspersed across this space are smaller elements of Petty Corruption. Bribes at police checkpoints are common but not standardized, and a common practice among customs officials at the border crossings involves the artificial creation of traffic jams that smugglers can bypass for a fee.
As smugglers of different goods operate through both routes, the different interactions offer key insights into why they do so. For example, the smuggling of illicit goods, such as drugs and arms, primarily occurs outside the border crossing. This is intuitive, as the same interaction-Genuine Enforcement-applies for illicit goods on both routes, but evasion is easier outside the border crossing. Among those trading licit goods, route choices are more diverse. Small-scale traders of licit goods with low levels of capital typically operate through the border crossing, as the Toleration agreement there provides predictability and low costs. Its profit margins, however, are also limited. As a result, traders in licit goods with higher capital endowments, less risk averseness, or higher profit margins often operate outside the border crossings. Although some high-value licit goods, such as phones, are typically smuggled outside the crossing, some commodities appear on both routes. For example, small quantities of gasoline are smuggled through the border crossing under the Toleration agreement, but traders with access to powerful 4x4s can bring across thousands of liters at a time.

The Thailand-Malaysia border
In Narathiwat Province, at the Thailand-Malaysia border, our research shows smugglers have three main options for crossing the border. They use one of the bridges over the border river Golok and pass the official checkpoints, they cross the river with a small irregular boat, or they cross the mountains further inland. Our analysis suggests that territory and the type of smuggler-state interaction determine the smugglers' choice.
The official border crossings, positioned at bridges across the river, tend to be only used for small-scale smuggling of licit goods. Our interviewees explained this by pointing to the high likelihood of being checked, as Genuine Enforcement happens here, and the availability of easier alternatives. However, according to residents on both sides of the border, the smuggling of small amounts of goods for personal consumption or to supply small shops is tolerated. For instance, people from Thailand do extensive shopping in Malaysia and refill petrol for cheaper prices there before driving back home.
Similarly, the mountain crossing is not a particularly popular route, as smugglers explained in our interviews. While there is little enforcement in the mountains, the area is difficult to cross due to a lack of infrastructure, including roads. In addition, there is an ongoing violent conflict in the deep south of Thailand, and the mountains often serve as hideouts for insurgents. Such routes, however, have been used to traffic and exploit people. In Songkhla Province, further west along the Thailand-Malaysia border, mass graves were found in 2015. Refugees and migrant workers from Myanmar, who wanted to reach Malaysia, were held hostage and tortured there for ransom. 31 Those who died were buried locally. In the months following the discovery of the mass graves, more than a hundred people from Bangladesh, Myanmar, and Thailand were arrested, including one Thai army general and fifteen other Thai state officials. For groups like this, who are looking for seclusion rather than a quick transit route, the mountain route is the preferred option. Meanwhile, however, the network also benefited from being close to the state, enabling it to conduct such a high-risk and high-value business.
Finally, however, the most used smuggling route is via small boats that take people and goods across Golok River, near the official border crossings. Here, as we observed, large quantities of licit goods, such as cooking oil, cigarettes, and T-shirts, as well as illicit goods, such as drugs, are smuggled across the border with small boats. The advantage of this route is good transportation links, as the river can be easily accessed via roads on both sides of the border, which connect the area with other parts of Thailand and Malaysia, as a drug smuggler explained. Different options for smuggler-state interactions exist here, making it the best available route for most smugglers. As outlined before, big businesses have Flat Rate deals with state officials, enabling them to operate with fixed costs and a high level of predictability. Even though many businesses operate on the same basis, making market concentration low, the smuggling is profitable because of the large amounts and the different types of goods that can be transported. At the same time, state authorities usually tolerate individuals crossing the border and conduct no passport checks. However, some informal practices have been established. For instance, Thai authorities often collect the ID cards of Malaysians who come to Thailand in the evenings to drink, as the Malaysian province bordering Narathiwat is dry, and only return them when those crossing take the boat back to Malaysia.

Implications
The analysis and considerations detailed above have several wider implications for discussions on smuggling and the state. Here, we discuss three of these implications in more detail, connecting them to existing discussions in the literature.

Active conflict zones are not attractive transit routes for smugglers
Much of the contemporary policy literature on smuggling is driven by the idea that conflict zones present unique opportunities for smuggling networks. The literature on "war economies" illustrates how today's violent conflicts can often be described as mutual enterprises in which the involved parties have little interest to stop fighting, as they economically benefit from ongoing violence. 32 War economies rely heavily on the smuggling of goods across international borders and borders demarcating zones controlled by different armed actors. For instance, conflict zones are the destination of many licit smuggled goods, as people in such areas often are cut off from supplies or have little access to goods. Hence, many people rely on smuggled goods for survival, creating a profitable business for smugglers. 33 Furthermore, conflict zones are the source of many illicit smuggled goods, such as guns and, in some cases, drugs. 34 Even though a high level of smuggling activities certainly exists in conflict zones, financing armed groups and possibly even ensuring the survival of affected populations, our analysis suggests such areas are unpopular transit routes for smugglers. 35 This finding matches that of scholars such as Justin V. Hastings,36 illustrating that this analysis is likely to have validity beyond the contexts explored in our research. 37 We suggest that conflict zones are only attractive to smuggling groups if they can provide interactions with local security providers that are more attractive than alternative routes. However, conflict zones are less likely to provide these interactions, as force is often not monopolized and smugglers must negotiate with multiple authorities and security forces. This analysis is not limited to state structures; it can be extended to nonstate groups that possess some form of territorial control. Each interaction we identify can therefore be replicated with rebel groups, insurgents, or other militias that control sections of a border. However, a group's ability to provide an attractive interaction for smugglers in terms of cost, predictability, and the ability to sustain a certain form of market concentration decreases as their territorial control becomes more contested. Hence, we suggest there are good reasons to believe established professional smuggling networks tend to prefer routes in which they have to interact and negotiate with as few authorities as possible, avoiding areas of violent conflict.
The effectiveness of changes in state capacity depends on the interaction type Driven by the discussed dominant understanding that states' inability to prevent smuggling is the result of limited state capacity, 38 contemporary approaches toward combating smuggling across the globe heavily rely on programs designed to increase state enforcement capacity. 39 These typically involve the training of border agents or the installation of additional security infrastructure, thereby either increasing border guards' ability to detect smuggled goods at points of entry or making it more costly for smugglers to cross "green borders" outside points of entry. President Donald Trump's advocacy for a wall at the US-Mexico border probably is the most prominent example of such initiatives. We suggest that these policy interventions are largely based on the assumption that the dominant interaction type between smugglers and state officials is Genuine Enforcement or low-level Petty Corruption, and that smuggling is primarily conducted outside of points of entry, or through the avoidance of detection within points of entry. We suggest that both assumptions are empirically untrue. Based on 37 For a further discussion of smuggling in conflict zones in Southeast Asia, see Weigand the analysis outlined above we can make three statements about state capacity-building measures. First, anti-corruption training for lower-level border security officials, if in fact genuinely effective in decreasing the willingness of selected officials to engage in corruption, is only going to be successful in reducing overall smuggling quantities conditional upon the wider environment of state-smuggling interactions along the border. In areas where State as Smuggler dominates, these measures likely are of negligible effect. In areas where Petty Corruption dominates, this may either decrease volumes or incentivize smugglers to move toward a different form of interaction, such as by approaching higher state officials to institute a Flat Rate interaction. In all areas, the effect of these measures on the cost of smuggling may lead to a reorientation of smuggling routes.
Second, improvements in surveillance infrastructure at points of entry such as scanners or better-trained personnel only have an effect on smuggling volumes if avoidance of detection is a significant part of the current interaction. Third, border infrastructure that makes crossing "green" borders more expensive for smugglers, such as walls and fences, only have a significant effect on smuggling volumes if trade is primarily conducted through points of entry. By raising the costs of transport outside points of entry, the construction of walls also incentivizes actors to shift routes toward points of entry and raises the costs actors are willing to pay to either engage in certain forms of interactions at border crossings (such as Petty Corruption or Flat Rate) or to transform the dominant interaction at ports of entry into one of these forms. As a result, we suggest that the construction of border infrastructure along "green" borders likely increases corruption at border crossings.

Violence is primarily driven by changes in the interaction
One of the most noteworthy aspects of the interactions we outline above, and of the practice of smuggling we have observed in borderlands across North Africa and Southeast Asia, is that they involve remarkably little amounts of violence in their day-to-day operation. Most interactions discussed here are built on stable expectations on both sides, as well as repeat interaction, and do not necessitate violence. Based on our analysis, we suggest that violence is primarily observed in the context of smuggling at moments of transformation from one form of interaction to another, or as one or multiple actors attempt to force a change in the interaction. We also posit that the likelihood of violence is particularly high in changes that affect market concentrations.
Tunisia's border with Algeria presents an example here. While the smuggling operations at that border were largely controlled through networks connected to the former dictator Zine El Abidine Ben Ali, these collapsed when his regime fell in 2011. Widespread occurrences of violence across this border after 2011 have commonly been connected to a lack of clarity among the actors about the new power structures and interactions. 40 Here, our observation largely follows similar conclusions in recent research on the relationship between smuggling and violence. 41 Bringing the state back in By examining some of smugglers' considerations when choosing routes for crossing borders, the article has illustrated that their relationship with state actors is a key factor. Here, the costs and conditions of evasion or side payments, the predictability and respectability of their interaction with the state, the availability of trade infrastructure-all are typically more significant factors for determining costs than geographical considerations. Hence, in contrast to a common perception, smuggling does not just happen in remote areas. To the contrary, many types of smuggling activities are much more prominent at official border crossings and easy-to-access stretches of borders.
Building on our empirical research, we developed six ideal types that describe smuggler-state relations. We have also noted that each has specific implications for smuggling networks and often benefits certain types of smugglers. For instance, small-scale occasional smugglers prefer Petty Corruption, but this is unattractive for professional smugglers who prefer a higher level of predictability. Meanwhile, Flat Rate models with higher barriers to entry can exclude small-scale smugglers and benefit professional smugglers. These considerations have implications for several related and more applied debates.
First, they contribute to disentangling the notion of "corruption." Corruption often features as a dominant analytical lens in contemporary writing on smuggling, especially in security studies. Conversely, this paper suggests that the concept of corruption in a smuggling context is of limited analytical power. "Corrupt" state-smuggler interaction has various forms, each having specific implications for smugglers, the state, and routes. Furthermore, our analysis shows that the type of corruption most associated with smuggling in the common imaginary-Petty Corruption-is a type of interaction many smugglers consciously avoid.
Second, the findings illustrate that common policy interventions aimed at addressing smuggling and transnational crime fail to respond to the underlying problems. For instance, capacity building is a popular development policy that aims at enhancing state capacity. However, our findings show smuggling often happens in border areas where state capacity is highest. Hence, enhancing state capacity runs the risk of simply increasing the costs for smugglers when interacting with the state.
Third, this paper systematizes a variety of different conceptualizations of the state-smuggling relationship, highlighting not only their relevance for the choice of smuggling routes but also their heterogeneous effects on market structures among smuggling networks. State-smuggler interactions not only shape smuggling routes but also influence what kind of networks can be successful within a certain environment. Consequently, they likely also affect dynamics between smuggling networks, organizational structures, and the relationship of networks with local populations.
Our findings show that the politics of smuggling and the interaction between smugglers and states should take a more prominent place in the study of both smuggling and border fortifications. Here, there is significant room for future research, not just for scholars of illicit economies but also for those interested in international security. In particular, further research could help develop a more nuanced understanding of the motivations and considerations that drive both smugglers and state actors. While many of our discussions here have focused on the cost-risk calculations of smuggling networks, more context-dependent factors, such as local normative perceptions, local histories of border-making, and state-society relationships all deserve further analysis both in the context of smuggling networks and ideal types. In addition, it would be worth exploring whether additional ideal types exist that may describe state-smuggler interactions. As we have focused here on these interactions from the perspectives of smugglers, additional research on the perspective of state institutions toward these interactions, as well as unpacking the notion of state capacity in this context, could be particularly fruitful.
Finally, while we have focused here on an analysis of different interactions and their effects on how smuggling networks choose routes, significant scope remains for additional work on how these interactions evolve over time. As we have outlined, there are diverse sets of motivations both for different smuggling networks and different state actors to prefer or tolerate different forms of interactions. While these are likely critical aspects of generating and maintaining such interactions, they need to be considered alongside the historical and institutional context in which they operate. Crucially, there is a question of which institutional and contextual factors shape the relative power of different actors-both between states and smugglers and between different smuggling networks-to influence these interactions and set the terms for how smuggling can be conducted. This also connects more directly to questions around the wider political economy of smuggling. In particular, the observations presented here note that the interests between smaller survivalist networks and highly capitalized, politically connected networks frequently need to be differentiated and that the role of other actors who maintain an interest in the global formal trade infrastructure and the global bordering industry should also not be neglected. This is sharpened by the fact that these interactions are likely dynamic over time. As we describe here, certain types of interactions set the context for the emergence of more monopolistic structures among smuggling networks, which in turn may then have a different capacity to influence their interactions with state agents than previous networks. In the long term, then, states shape not only smuggling routes, but also smuggling networks-which may end up shaping states themselves.