Theorising state–narco relations in Bolivia’s nascent democracy (1982–1993): governance, order and political transition

Abstract Conventional policy and academic discourses have generally held illicit drug economies in Latin America to be synergistic with violence and instability. The case of post-transition Bolivia (1982–1993) confounds such assumptions. Applying a political economy approach, this article moves beyond mainstream analyses to examine how the Bolivian drug trade became interwoven with informal forms of governance, order and political transition. I argue that state–narco networks – a hangover from Bolivia’s authoritarian era – played an important role in these complex processes. In tracing the evolution of these interactions, the article advances a more nuanced theorisation of the relationship between the state and the drug trade in an understudied case.

political order. Following democratic transition in 1982, these state-narco networks continued to function. Reflecting the diffusion of political power, though, once-centralised networks now became more atomised: running through the military, the police and the political parties. These took on new meaning. Mutual toleration of drug links was now bound to the maintenance of fragile political equilibrium between these actors during an uncertain process of transition. Tracing the evolution of state-narco networks during a period of political flux, the article reveals the significance of these interactions to regime consolidation and an uneven democratisation.
This analysis addresses the undertheorisation of the relationship between the state and drug trade in Latin America (and beyond) within conventional policy and academic discourses of drugs. These discourses are informed by a liberal, Weberian conception of the state, where the drug trade is viewed as necessarily violent, destabilising and synonymous with 'weak' , 'fragile' or 'failed' states. 1 However, this perspective tends to misrepresent the function of the state in postcolonial settings, and hence fails to capture the distinct forms of interaction that may emerge between the state and the drug trade in different contexts. This is an important variable in the relationship between violence, order and illicit drugs. Analysis of state-narco relations must, therefore, take the 'empirical reality' 2 of the state as a point of departure: engaging with 'how states actually function, rather than how they ought to' . 3 Building on a political economy framework, the article explores the role of drug rents 'as a key source and vector of power' 4 in Bolivia's historic political development. As such, it unpacks the top-down assumptions of the mainstream policy and academic discourse to bring greater nuance to theorisations of state-narco relations.
While much of the political economy of illicit drugs literature focuses on authoritarian and (post-)conflict societies, this article considers how informal forms of governance connected to the drug trade partly underpinned Bolivia's uneven democracy. The term 'uneven democracy' is drawn from O'Donnell, referring to a common phenomenon seen across Latin America's Third Wave. 5 The region's new democracies adopted the formal rules and procedures of democracy, such as regular elections, while maintaining elements of the authoritarian past. This included clientelistic practices and unaccountable institutions. State actors operated according to informal rules and structures. This informalism -in part a consequence of factionalised states and intra-elite competition -is crucial to understanding the function of state-narco networks during the post-transition period. Governmental tacit acceptance of state-narco networks and patron-client bonds, and reluctance to implement democratising institutional reforms of the security forces, formed part of an unspoken post-transition settlement between the main actors. The article contends that within the context of the fragmented Bolivian state, the control of drug rents was enmeshed with the path of democratisation.
I employ an intensive case study research design and qualitative methodology to trace the evolution of state-narco relations. This approach includes the use of secondary and documentary sources to construct a historical baseline account of the case, outlining the interplay of the changing political context and the development of Bolivia's coca-cocaine economy. These sources are supplemented by 38 interviews conducted largely with highlevel political actors from the post-transition period. Interviewees include US Embassy personnel previously based in La Paz, and former Bolivian government ministers and high-ranking state officials. Participants are afforded anonymity where requested. This was addressed as part of the process of informed consent, applied in accordance with the conditions of institutional ethical approval for research. 6 While the article considers the effects of US policy preferences in Bolivia on the development of state-narco networks, the analysis primarily centres on dynamics between domestic actors. This approach allows for a historical perspective of the development of state-narco relations in Bolivia.
Individuals were selected for interview due to the key role they played during this period in Bolivian politics, policy implementation and counterdrug efforts. The positionality of interlocutors, as political elites situated within or closely linked with the Bolivian state, gave them insight into the relationship between state agents and the drug trade. Their historical accounts reveal distinct understandings and interpretations of state-narco interactions, how they related to wider political pressures and the groundedness of these networks in local contextual factors. Interviews thus offer novel insights on the function -and significance -of state-narco relations.

Theorising state-narco relations in Latin America
In this section, I unpack the conventional discourses of illicit drug economies in Latin America. Policy and academic discourses are considered separately. The latter has tended to be critical of the former: challenging the effectiveness of counterdrug responses and highlighting their often-damaging consequences. Both discourses, though, share similar underlying assumptions around the nature of the drug trade. Crucially, this leads to an undertheorisation of state-narco relations.

Policy discourses
Within policy discourses, the illicit drug economy in Latin America is commonly associated with insecurity, instability and ineffective governance; the 'war on drugs' prescribes 'more state' and securitised policies in response. Elements of this paradigm may be traced back to the inception of the modern US drug war in the 1980s, as US counterdrug policy adopted the language and logic of securitisation. Two facets of the Andean drug trade as 'threat' were advanced: the societal harms caused by drug use in the US, 7 and the destabilising effects of the coca-cocaine economy in 'America's backyard' . 8 In terms of the latter, this included the violence and corruption of powerful drug-trafficking organisations, real and imagined collusion with 'narco-guerrillas' , and attached social, economic and political instability. 9 These were used as justification for the extension of US counter-supply efforts abroad. Defeat of the drug trade through a militarised response was held as a catch-all cure for the interacting 'security threats' of drugs. The tenets of this drug war approach have had an enduring influence not only on US counterdrug policy, but on other key actors of international drug control. The United Nations Office on Drugs and Crime (UNODC), for example, has warned of the onset of 'narco-states' , overrun by the power and influence of illicit drugs. 10 In certain cases, such warnings are linked to the greed/economic model of post-Cold War civil conflict: associating 'lootable' resources, such as drugs, to insurgency and potential state breakdown. 11 Wider Western security paradigms of 'fragile' , 'weak' or 'failed/failing' states are evident within these policy narratives. 12 Underpinned by 'an idealised notion of the state' , the presence of the drug trade is demonstrative of 'state fragility' and lack of capacity to 'centralise the means of coercion' . 13 'Fragile states' are 'perceived as providing breeding grounds and safe havens for transnational terrorism, weapons proliferation, and organised crime' . 14 These 'ungoverned spaces' thus represent a challenge to Western global order, spawning cross-border security threats and regional instability. 15 Establishing effective state institutions within such cases is viewed as a fundamental prerequisite to sustainable development. 16 Liberal state-building notions of establishing 'effective sovereignty' , reforming government institutions, bolstering security forces and creating alternative rural economies have bled into securitised counterdrug responses as a result.

Academic discourses
The academic literature on drugs in Latin America is diverse, but critical voices of international drug control predominate. Securitisation of drugs is largely rejected, with authors such as Bagley, Buxton and Loveman, for example, highlighting the role of the US 'war on drugs' in deepening drug-related violence and instability in Latin America. 17 This includes human rights abuses carried out by heavily armed counterdrug units. Such policy interventions may also interact with the market dynamics of prohibition. 'Successful' counterdrug policies -eg the capture of a 'drug kingpin' -disrupt markets, creating scarcity and raising prices, thereby increasing incentives and intensifying violent competition between rival groups. 18 The 'war on drugs' , therefore, is an important explanatory factor of violence and instability in Latin America.
Areas of this literature, though, also draw on many of the same premises of the policy discourse, including the assumption of 'drugs, violence and instability' and a tendency to view the drug trade through the lens of state fragility. Youngers and Rosin, for example, highlight the role of 'the drug trade itself' in generating violence, and fuelling armed conflict across Latin America, as well as the weakness of state institutions to resist the power of drug trafficking organisations. 19 The Mexican state's failure to fulfil its Weberian function grounds Bunker and Sullivan's analysis of the challenge posed by Mexican Cartel violence, 20 while Rosen and Kassab draw links between low state capacity and the proliferation of drug trafficking, violence and insecurity throughout the region. 21 In addition to destructive and inadequate policy responses, the damaging presence of the illicit drug trade in Latin America is thus perceived as a problem of ineffective governance and 'weak states' .

Political economy of illicit drugs
This perspective has limitations in accounting for the development of state-narco relations. It is certainly the case that major drug producing and transit nations in Latin America have frequently fallen short of the Weberian ideal type state. For example, fragmented and poorly institutionalised states create opportunities for organised crime to pursue its business interests. This may entail, for example, subversion of state institutions through narco-corruption, mafia de facto control of areas of the national territory and inter-group violence. 22 State institutions lack the capacity (or willingness) to extend public goods evenly across society in such contexts, including security and the rule of law. But in maintaining a 'dogmatic adherence' to a Weberian conceptualisation of the state, this framework fails to engage with the 'distinctive forms of empirical statehood' 23 observable across Latin America. As Morton argues, this represents 'a pathological view of conditions in postcolonial states as one of deviance, aberration and breakdown of the norms of Western statehood' . 24 Throughout Latin America, distinct and enduring forms of statehood have emerged from specific historical trajectories. Forms of political authority and modes of state power have frequently diverged from the principles of the Weberian nation-state. 25 In some contexts, the drug trade may be intimately bound up in these divergent practices. While the lens of 'fragile states' may demonstrate that major production and trafficking of illicit drugs represents a failure of the liberal-Weberian state with attached negative connotations for democratic development, it also obscures important contextual factors that determine state-narco relations and their wider social, political and economic effects.
My application of a political economy approach to illicit drugs incorporates this insight and seeks to unpack the underlying assumptions of the mainstream drugs discourse: problematising the relationship between the state and the drug trade, and tracing the implications of these interactions beyond the general expectation of 'drugs, violence and instability' . This framework necessitates contextualisation of the Bolivian state, and a focus 'on the distribution of power and wealth between different groups and individuals, and the processes that create, sustain and transform these relationships over time' , 26 including the distribution of drug rents.
Snyder's work on 'lootable wealth' is useful in explaining such dynamics, while helping to unpick the relationship between 'drug, violence and instability' . 27 Whether the drug trade is associated with political order, or violence and instability, is in part dependent on the political economy of extraction. In some cases, for example, societal actors may independently control resources, such as illicit drugs, accumulating wealth that allows them to challenge central state authority, thereby fuelling political disorder. Alternatively, the establishment of 'institutions of joint-wealth extraction' -'informal bargains rooted in personal ties among rulers, their cronies, and private elites' 28 -may solidify common interests between state and societal actors. Here, the state protects and fosters illicit enterprise in return for rents. Through joint wealth-extraction, a level of order and stability may be achieved. Potentially hostile rivals are absorbed into existing political institutions, and drug rents provide political leaders with the tools to govern. This incorporates the maintenance of patrimonial state systems, and the management of competition between different actors in the illicit economy. The construction of state authority in these contexts occurs in negotiation with societal actors and is linked to the establishment of shifting alliances between elite factions, within which the control of drug rents becomes a factor. In such cases, top-down narratives of drugs fail to account for the 'structured variation' of political, social and economic conditions in source/transit nations, 29 the distinct forms of interaction that arise between the state and the illicit economy in different contexts, and how these influence the wider effects of the drug trade.

Bolivia's authoritarian era state-narco networks (1964-1982)
Different theories have been advanced for the relative lack of drug-related violence in Bolivia during the post-transition period. Thoumi, for example, argues that Bolivia was vertically integrated into cocaine transit routes in the 1980s, controlled by the Colombian Cartels. 30 Violent competition for dominance occurred predominantly further up the commodity chain, rather than in Bolivia itself. Additionally, Bolivia was not subject to the same intensity of militarised counterdrug policies as its regional neighbours. All-out war was declared against Pablo Escobar and the Medellin Cartel, for example; less-confrontational policies were pursued against Bolivian organised crime. This meant that the Bolivian drug market remained more stable.
At this time, Bolivia's regional role in the Andean cocaine trade was primarily that of coca producer (see Figure 1). Coca provided small-scale producers -organised into sindicatos or unions -with a relatively modest but vital source of income. These producers defended their ancestral right to cultivate coca, given its domestic licit use and cultural significance to indigenous populations. Major local organised crime, though, also operated in Bolivia during this period. These groups grew wealthy from processing coca into low-grade products, before transporting them to their partners in Colombia. Estimates of the illicit trade and assertions around the division of its profits should be treated with caution, but Figure 2 indicates that Bolivia's trafficking clans headed an industry estimated to be worth 15% of gross domestic product (GDP) by 1988. While it is certainly the case that these organisations were dwarfed by their Colombian counterparts, they were significant within the Bolivian context. This view of Bolivian organised crime clashes with local narratives around the country's role in the drug trade. Bolivian politicians argued against the onset of the US 'war on drugs' , claiming Bolivia as a humble producer of coca, linked to underdevelopment. 31 According to this narrative, power and wealth were accumulated by external criminal groups, who were responsible for the most deleterious effects of the drug trade. In-depth analysis of US-Bolivian relations and counterdrug responses is outside the scope of this article. 32 Control of coca, however, largely formed the focus of the US drug strategy in Bolivia -through a mix of enforcement and development efforts. This created friction with coca-growing communities, but avoided the intensive drug war violence witnessed in Colombia.
The regional structure of the drug trade and policy responses are, therefore, important explanatory factors for the nature of Bolivia's illicit economy during this period. Analysis of state-narco networks adds an important element to these perspectives, accounting for the role of informal forms of governance around the drug trade in helping to mediate violence and engender stability in the Bolivian leg of the Andean cocaine trade. These networks were first established during the military authoritarian government of General Rene Barrientos (1964Barrientos ( -1969, then extended and deepened under the regimes of General Hugo Banzer (1971)(1972)(1973)(1974)(1975)(1976)(1977)(1978) and General Luis García Meza (1980)(1981). I use the term 'state-narco networks' to refer to stable, regular modes of exchange between agents of the state and high-level actors of the drug trade. 33 The coca unions, while important actors in the Bolivian case generally, are excluded from my analysis. Focus is instead given to highlevel organised crime, as the formation of state-narco-type systems based around drug crops, such as coca, is unlikely due to the low barriers to entry at this level of the trade and its diffuse nature. 34 Modes of exchange, therefore, typically involve the extraction of drug rents from high-level actors for non-enforcement of the law and/or the official protection of state agents. Such networks formed one facet of the clientelistic mode of governance that enabled these military regimes to co-opt political support and consolidate power, while managing competition in the drug trade.
This reflected the Bolivian state's patrimonial and fragmented nature. 35 Intra-elite competition for capture of the state characterised the period, as loose coalitions, formed of different sectors, looked to control the state apparatus in pursuit of their particular interests. Both 'strong' and 'weak' elements may be identified within the Bolivian state, with attached variations in capacity (and willingness) to exercise authority in different domains. Here, the factionalised Bolivian military was preeminent, not only in its control of executive power but through the military officers assuming political posts and managerial roles in state enterprises, as well as the institution's ties 'to peasant groups and crucial political elites' . 36 Military leaders looked to the patronage of the state to build alliances within the institution, while also binding additional state and societal actors to their regime. Drug rents constituted an  (1986)(1987)(1988)(1989)(1990). data sources: estimates of the size of the Bolivian coca-cocaine economy plotted against Bolivian gross domestic product (GdP) data. Franks, "la economía de la Coca", 20; World Bank, "Bolivia: GdP data".
important part of this system of control. With national coca production used as a proxy measure, Figure 3 indicates the exponential growth of the coca-cocaine economy over this period. 37 Military authoritarian regimes sought to colonise rents from the illicit trade with political allies, including military officers and 'crucial political elites' . In this sense, drug rents were one factor in the 'shifting alliances of power between elites and social actors' that determined the uneven and 'discontinuous construction of state authority' in Bolivia. 38 The personalistic regime of Barrientos, for example, used rents from the emerging Bolivian drug trade to forge patron-client bonds through the state. 39 Barrientos encouraged supporters within the military and the police to facilitate, and profit from, cocaine production in the departments of the Beni and Santa Cruz. As well as buying support, this form of patronage also gave the regime leverage over potential rivals, eg by threatening to expose drug corruption. 40 Lands were also granted to backers of the government in Bolivia's eastern departments, which were fast becoming a key strategic point in the Andean drug trade. Although Barrientos had courted the peasant sector with promises of advancing the redistributive principles of the 1952 revolution, the regime instead used land reform laws to award 'larger plots to fewer people in a more discrete exploitation of the statute for political ends' . 41 As well as benefitting new agricultural elites, parcels of land in the Beni and Santa Cruz would be used by military officers to set up coca processing labs and hidden landing strips. These practices helped to create linkages between the state and the incipient drug trade.
Banzer also used drug rents as a form of patronage. The cocaine boom of the 1970s and the mid-decade collapse of international cotton prices, though, meant that coca-cocaine took on greater importance. Banzer cultivated societal support in the eastern lowlands of Bolivia, as Santa Cruz and its surrounding agri-business became an engine of the national economy. 42 High prices for cotton in particular had encouraged large investments in the crop. But when world market prices fell dramatically in the mid-1970s, the lucrative returns promised by coca-cocaine ventures offered a solution to potential bankruptcy. 43 In the second half of the 1970s, the drug economy would grow significantly in Santa Cruz and the neighbouring Beni department as agri-business elites turned their attention to illicit enterprise. Reports suggest that state support played a role in this shift, with soft loans provided by the state-backed Banco Agrícola de Bolivia. 44 Accelerated land redistribution to military officers in the Beni, meanwhile, also deepened state-narco networks. Such officers were able to extract rents for use of their land as a logistical stop-off point for trafficker flights. During this period, therefore, 'the illegal industry grew and developed a tight nexus with Santa Cruz's elite and the military' . 45 This nexus was facilitated by personal links between state actors and traffickers within Bolivia's small elite, as well as the fluidity of the 'licit' and 'illicit' spheres. Most notably, Roberto Suárez Gómez -'the King of Cocaine' -was well connected to Bolivian elite circles as a large land-owner with 'aristocratic' family ties to the military. 46 This included, for example, his first cousin, Colonel Luís Arce Gómez, who was later dubbed García Meza's 'Minister of Cocaine' . Agri-business elites moved between legitimate enterprise and the drug trade; officers straddled their military careers and ventures in coca-cocaine. This dynamic blurred lines between these spheres and helped to establish a network of patron-client relations. Reflecting profit maximisation motives and Bolivia's clientelistic culture, traffickers viewed themselves as 'functional economic actors' , and, working through elite networks, pursued 'a strategy of peaceful coexistence with the state' . 47 These relations cemented generally cooperative dealings between Bolivia's main trafficking groups, acting to manage competition. According to journalist reports, this was also aided by the targeting of potential competitors to statelinked traffickers by the security forces. 48 Snyder and Duran-Martinez note that this kind of consolidation -also encouraged by the consolidation of trafficking routes by the Colombian cartels 49 -eases systems of rent extraction by reducing the number of actors and creating more certain transactions. 50 As such, the colonisation of the drug trade by the state and the consolidation of its structure helped to create more centralised, stable state-narco networks.
While drug links were downplayed during the governments of Barrientos and Banzer, state-narco networks came to define the García Meza regime. Against the backdrop of the Cold War, the US had turned a blind eye to the indiscretions of their anti-communist allies in the Bolivian military. The presidency of Jimmy Carter (1977)(1978)(1979)(1980)(1981), though, shifted US foreign policy in the Americas. The Banzer regime was pressured to implement democratising reforms and, after a series of inconclusive elections, the leftist Unidad Democrática y Popular (UDP) 51 eventually came to power, in 1980. García Meza's 'cocaine coup' of July that year brought to a halt these first faltering steps towards democracy. The regime had gambled on gaining US backing for its 'anti-communist' credentials from the incoming Reagan administration. Instead, the Drug Enforcement Administration (DEA) and local journalists fully exposed the regime's close links to drug trafficking, 52 forcing the Reagan administration to condemn the coup and withhold US support. The case demonstrated the rise of the US 'war on drugs' in the Andes.
In building internal support for his coup, García Meza looked to exploit his links to the drug trade, distributing patronage in the form of drug rents. 53 Drug trafficking groupsfearful of the potential business implications of open, democratic government -were more than willing to put up such funds. Suggestions that the UDP was consulting with the DEA to increase counterdrug cooperation with the US seemed to confirm their suspicions. 54 It was reported that Roberto Suárez convened Bolivia's top drug traffickers several months before the coup to pool their resources in support of García Meza. 55 In this sense, the interests of traffickers entered into the calculations of would-be regime leaders, such as García Meza, because drug rents could be diverted to political rivals. The new government thus called on its personal ties to these traffickers and co-opted their support with the promise of protection.
The García Meza regime lasted little more than a year due to its lack of a wide base of societal support, failure to address a growing national debt crisis and pressure to relinquish power from both the US and elements within the military. This demonstrates the limits of regime consolidation based solely on state-narco networks. In the case of Barrientos, for example, his pact with the rural sector was crucial to sustaining his rule, while Banzer's delivery of economic growth largely formed the basis of his government's 'legitimacy' . Left-Right divisions also engendered support for these authoritarian regimes from military and police factions, right-wing political parties, business, the middle classes and the country's international patron, the US. In this sense, the role of Bolivian state-narco networks in reinforcing -or indeed weakening -regime authority must be considered in relation to these wider factors.
Nevertheless, it is important to note that regime leaders of this period used drug rents to build factional coalitions with both state and societal actors, including wealthy trafficking groups. These networks managed competition between drug trafficking clans, and built on the kinship ties that ran through Bolivia's small elite. Distribution of economic benefits discouraged violence and disorder, binding the interests of these actors to the regime. In this sense, drug rents were a tool of governance: political actors exploited the illicit economy to consolidate power. According to liberal-state building perspectives, the growth of the drug trade evidences the state as 'weak' . However, this perspective has limited utility in accounting for the Bolivian state's distinct historic development and the drug trade's interaction with these processes, including the reinforcement of the Bolivian state's patrimonial and fragmented nature.

Post-transition state-narco networks (1982-1993)
Democratic transition in 1982 represented a point of fracture for these centralised statenarco networks. First, the end of the authoritarian era led to the diffusion of political power, with the parties and police becoming more prominent. Second, shifting US priorities in Bolivia placed the governments of the post-transition period under increasing pressure to apply securitised counterdrug policies. Third, a new generation of drug traffickers rose to prominence, changing the structure of Bolivia's organised crime and severing previous links to the main state actors of the authoritarian period. Gamarra argues that the process of democratisation thus 'had the paradoxical effect of democratising the structure of organised crime' . 56 These developments prompted the break-up of centralised systems and the establishment of more atomised state-narco networks in their place. Actors within the drug trade sought to forge new links and build on 'existing modes of illegal practice in the armed forces, police and leading political parties in a relatively "unpartisan" fashion' . 57 As before, drug trafficking groups were able to work through state structures. 58 These adapted networks took on new meaning in the post-transition period. Beyond building factional coalitions and reinforcing the political order as had been the case previously, state-narco interactions were now interwoven with a fragile equilibrium between the main political actors. The path of Bolivia's political transition is an important contextual factor here. The opening of the political system under Banzer had led to a period of political instability, revealing the deeply factionalised nature of the Bolivian state and deep fissures in society. 59 The collapse of García Meza's regime had ensured the restoration of the UDP government, as the disgraced military withdrew from power. Deep divisions and fragmented state power, though, still remained. This was 'democratisation by default' , with the main 'praetorian' actors 'discredited and marginalised rather than persuaded of the benefits of cooperation' . 60 The transition was not rooted in convergence between different sectoral interests, but rather in the exhaustion of other routes to power. Hence, there was uncertainty over each actor's commitment to democratic transition. Challenging the interests of the military or other key political actors was viewed to place the transition at risk. In this game of political elites, destabilising actions and a possible spike in drug-related violence were to be avoided. 'Democratic governability' was prioritised. 61 This incorporated political pacts and the trading of patronage, the autonomy of the military and police, and a lack of transparency and accountability in Bolivia's uneven democracy.

The main political parties: ADN, the MIR and the MNR
A series of drug scandals exposed each of the main political parties' links to actors in the illicit trade. For example, the 1988 'Narcovideo' scandal showed officials from both the Movimiento Nacionalista Revolucionario (MNR) and Acción Democrática Nacionalista (ADN) meeting personally with Roberto Suarez to discuss contributions to the parties. Furthermore, President Jaime Paz Zamora's (1989Zamora's ( -1993 Movimiento de Izquierda Revolucionaria (MIR) was accused of taking campaign contributions from prominent drug trafficker -and former army captain -Isaac 'Oso' Chavarría. It was believed that the parties typically accepted 'kick-backs' in exchange for promises of protection. 62 The US Embassy suspected the MIR, in particular, of having a close relationship with drug trafficking following the appointment of drugtainted officials to prominent counterdrug postings. 63 The extent of these links, and the influence of traffickers on the governments of the day, is open to debate. However, many believed drug money lubricated patron-client bonds within the parties. 64 While politicians would occasionally throw accusations of corruption at each other for political gain, inquiries into the depth of drug links in the political class rarely uncovered anything of substance. Critics branded periodic congressional investigations of drug scandals as whitewashes, with convenient scapegoats identified and 'bad apples' blamed for systemic corruption. 65 The parties, though, maintained a collective 'vow of silence' over the issue. 66 Elite Bolivian political actors prioritised the political stability of the transition. 67 Siles Suazo's UDP administration (1982)(1983)(1984)(1985), for example, had been marked by economic crisis and legislative deadlock. The parties subsequently turned to political pacts 68 as a way of establishing stable government. Politicians of the period were shaped by their experiences of historic political instability and decades of authoritarianism. 69 The Pacto por la Democracia under Paz Estensorro (1985Estensorro ( -1989 and the subsequent Acuerdo Patriótico of Paz Zamora's presidency (1989Zamora's presidency ( -1993 created alignment between the president and congress, breaking deadlock and allowing the executive's legislative agenda to pass. The political parties hailed this spirit of compromise, effective governance and the coming of age of the self-proclaimed 'democracy generation' . 70 In the case of the Acuerdo Patriótico, for example, Paz Zamora's leftist MIR party entered into government with Hugo Banzer's right-wing ADN, despite having been subjected to repression under the former dictator's regime. Once-enemies now claimed to be putting democracy, political stability and the national interest before a painful history. 71 Within this context, exposing drug links threatened to bring down the edifice of the 'democracy generation' and collapse political compromise. The issue was downplayed, and the maintenance of political stability prioritised.
Of course, mutual interests in the illicit economy may also have played a part in this dynamic. No party could claim a spotless record over the issue, and so each had much to lose over the disclosure of drug corruption among the political class. In addition to this, some highlighted the clientelistic nature of the political pacts of the period. For example, James Cason, the US Embassy's Political Counsellor, argued that such pacts were part of a cynical elite game, played by the political parties to 'steal as much money as they could for themselves' . 72 In forging these agreements, political patronage was dished out amongst the various parties involved. This included the addition of three new ministries, 16 new vice-ministerial posts and 20,000 new public-sector employees during Paz Zamora's presidency. 73 Pacted democracy allowed political elites to share power and divvy up the patronage of the state. From this more critical view, the politics of the clientelistic and authoritarian past were still in full effect. There were material costs in pursuing the issue of state-narco networks running through the political parties, and this led to the collective 'vow of silence' . These linkages thus formed part of the political dynamic between the parties during the uncertainty of the post-transition period, and indicated the enduring patrimonial nature of the Bolivian state.

The military
Equally, there remained concerns that the military's commitment to democracy was contingent. The civilian political leaders of this era had lived with a politicised military for decades, and so continued 'to fear its political leverage and a possible return to military rule' . 74 Reports of an abortive coup attempt in early 1984 planned by recently discharged officers connected to the 'cocaine coup' supported such suspicions. Indeed, the suggestion that the plot had been hatched in the wake of attempts to investigate the military for their involvement in drug corruption confirmed the anti-democratic tendencies of certain elements. The now dominant 'institutionalist' faction had pragmatically lent support for democracy and carried out a 'token purge' of the drug-liked officers, 75 given the changing preferences of the military's North American patron, culpability for economic crisis and low-standing with the Bolivian public. But the factional nature of the institution made predictions of its future direction difficult. While the tide appeared to be turning towards democracy, Bolivian civilmilitary relations were still delicately balanced. As such, politicians practised a policy of non-interference in the internal matters of the military, delaying institutional reform in return for the military's withdrawal from politics. 76 Indicative of the fragmented Bolivian state, the military enjoyed autonomy from civilian governmental control.
The extraordinary rendition of Arce Gomez to the US in December 1989 demonstrated the finely balanced nature of civil-military relations during this period. After years on the run, the former 'minister of cocaine' was located by DEA agents, living openly in in Santa Cruz. For some in the US Embassy, the brazenness of Arce Gomez hinted at protection from high levels of the Bolivian state. 77 This lack of trust was also reflected in the secrecy of the operation, with suspected corrupt elements within the Bolivian government excluded from the planning and implementation of the operation. Still apparently wary of provoking the military by attempting prosecution within Bolivia, President Paz Zamora agreed that Arce Gomez should face trial in the US on drug trafficking offences. Despite the absence of an extradition treaty, the former colonel was flown out of the country by the DEA and arrested in Miami. The Bolivian government's reluctance to a public trial of such a high-profile 'military man' indicated the fear civilian politicians still had of provoking a coup d'état.
Military institutional autonomy had implications for internal drug corruption. The military's retreat from power meant that the centralised systems of rent extraction of the authoritarian era no longer functioned. 78 In key areas of the coca-cocaine trade, though, military regional commands deepened links to drug traffickers. Pay-offs to secure profitable posts in the Chapare and the Beni then filtered through the institution. 79 These formed part of a hierarchical system of patron-client bonds. The US government shared the perception that state-narco interactions remained embedded in the military. For example, the DEA claimed that 'all elements of the military are involved in drug trafficking to some extent' , 80 while a US Congress report stated that 'most United States and Bolivian officials' admit corruption is 'widespread and generally accepted within the Bolivian military' . 81 However, within an uncertain process of transition, toleration of autonomy -and by extension state-narco links -was linked to keeping the military in its barracks.

The police
The Bolivian police had benefitted from the democratisation process. Where previously the institution had been subordinated to the military, police funding and prerogatives were now enhanced. Post-transition governments looked to the police to act as an institutional counterweight to the military, in return granting a high degree of autonomy from civilian control. 82 Institutional reform was placed on hold, as the political parties instead used patron-client bonds to achieve their aims. 83 This included power over appointments to command positions, and disbursement of discretionary government funds. Particularism continued to define relations between different state actors. Post-transition governments feared that encroachments on the institution risked its support. Furthermore, as political leaders were reluctant to push for prosecutions of officers suspected of corruption or abuses, impunity was common. Between 1982 and 2002, for example, 20 national police commanders were linked with significant cases of corruption -many of these related to drugs -but not one was prosecuted. 84 Taken together, these factors demonstrated the persistence of the patrimonial and fragmented Bolivian state despite transition. Lack of reform meant that the authoritarian tendencies of the police remained unchallenged, limiting transparency, accountability and oversight. This facilitated police links to the illicit sphere.
Linkages between the police and the drug trade thus formed a third facet of the constellation of post-transition state-narco interactions. The institution's natural proximity to illicit activity created numerous opportunities for the extraction of bribes. Poorly paid officials in the Chapare, for example, were able to supplement their wages with ad hoc pay-offs, allowing contraband to flow in and out of the region. 85 Reports suggested, though, that police drug corruption was also more systematic. The police had been incorporated into state-narco networks during the authoritarian era and, in the post-transition period, such practices continued. This reportedly included collusion with traffickers to ensure non-enforcement around transportation of large shipments. 86 Patron-client interactions within the unreformed police institution were in turn solidified with such drug rents. For example, pyramid structures, in which subordinates would pay superiors for ranking positions, ensured drug money flowed through the institution. 87 Indicative of uneven democratic institutionalisation, formal codes and institutional arrangements ran parallel to irregular practices, such as hierarchical systems of corruption. 88

Uneven democracy
Across these different political actors, therefore, tacit acceptance of state-narco networks was associated with maintaining political stability. Political elites of the period, in general, did not consider the drug trade to be a threat to national security. 89 There was ambivalence towards it, given its relatively peaceful nature and its importance to the national economy. For example, Bolivia was not threatened by powerful drug cartels as in Colombia; and, furthermore, coca-cocaine had acted to stabilise the national economy following economic crisis and the implementation of neoliberal structural reforms. On this latter point, coca cultivation provided inward investment and employment for a significant sector of the population, while narco dollars had stabilised the banking system following the liberalisation of capital controls. 90 For the 'democracy generation' , drugs were not the priority, but rather stability, economic growth and democratic consolidation. 91 This included the 'compromises' of state-narco networks.
Such compromises were indicative of Bolivia's uneven democratisation. The Bolivian state's 'deviation' from the liberal ideal type had implications for the process of democratisation. Different state actors pursued their own particular interests. Institutionalisation of Westernliberal democratic norms was uneven across the different arms of the state. As such, 'democracy had triumphed but was known to have feet of clay' 92 : the 'old reflexes and assumptions, and vested interests created by the previous history' stood alongside democratic institutions. 93 This included the continuation of clientelistic politics and the autonomy of the security forces. State-narco networks were embedded within these surviving remnants of the authoritarian past. Bolivia's enduring form of 'empirical statehood' and the drug trade's interactions with it thus continued to shape the country's 'democratic' future.
The legacies of authoritarianism -and the state-narco networks contained within -were tolerated within an uncertain process of political transition, as support for democracy from the main political actors was viewed as provisional. As such, non-enforcement of the norms of accountability and transparency limited the potential resistance of 'informal veto players' to (formal) democracy, inducing 'powerful actors to accept rules they would otherwise seek to overturn' . 94 In this sense, drug rents were not used to build factional coalitions behind a central authority, as was the case during military authoritarian governments. Rather, different political actors pursued drug rents independently (to different degrees). Reflecting the fragmentation of the Bolivian state, mutual toleration of these practices between different actors was tied to the maintenance of a tacit post-transition political settlement. As a result, agents of the state formed alliances with private actors to exploit drug revenues, eroding democratic notions of an even and universal rule of law. 95

Conclusion
The case of post-transition Bolivia confounds the typical expectations of the mainstream policy and academic discourse of drugs. Moving beyond the blanket assumption of 'drugs, violence and instability' , this article elucidates the role of state-narco networks in managing and mediating violence, and the relationship between the drug trade and Bolivia's political transition. This includes the informal forms of governance and political order that arose as part of state-narco interactions. I argue that the framework of 'fragile states' is insufficient to explain these interactions and their wider significance. The application of a political economy approach reveals the role of drug rents in Bolivia's wider political development: the continuation of clientelistic and authoritarian practices in Bolivia's uneven democracy, and the interconnectedness of these phenomena, state-narco networks and a delicately balanced post-transition political settlement. In this way, the article brings greater nuance to theorisations of the relationship between the Bolivian state and the drug trade.
Re-thinking these linkages has important implications for drug policy, state-building and democratic development. Across Latin America, counterdrug efforts aimed at defeating the drug trade and establishing security have frequently been associated with heightened violence. Drug war 'wins' destabilise the drug trade, and lead to the emergence of new actors and (violent) competition to establish supremacy. In addition to this, they may disrupt the informal forms of governance -ie state-narco networks -that act to mediate violence in some cases. Within such contexts, the assumptions of a Weberian liberal-state perspective that underpin drug war policies misrepresent the actual functioning of state power. In Bolivia, state-narco links were viewed by local actors to be interwoven with political order. This is not to say that state-narco networks are desirable or sustainable in the long term. These relations were part of Bolivia's uneven democracy, contributing to limited transparency, accountability and political competition. Stability and the maintenance of transition may have been used to justify such practices, but these also represented the accumulation of wealth and power among elite actors. As Gutierrez notes, 'this kind of stability is short-term and may come at the expense of building more legitimate and accountable governance institutions' . 96 Hence, this analysis raises important questions of how such particularistic and exclusionary systems may be decoupled from the state, while avoiding heightened violence and institutional decay.

Disclosure statement
No potential conflict of interest was reported by the authors. of illicit drugs, the 'war on drugs' in Latin America, US-Bolivian relations, and the interstices of illicit economies and development.