Cockroach states ; what makes for a successful drug trafficking country

F

M

G

Afdeling Politicologie

Civil Wars and State Building

2016-2017, Semester II

Research Paper

Cockroach States: What makes for a successful drug trafficking

country.

Words: 21156

Supervised by Dr. Abbey Steele

Read by Dr. Jana Krause

23/06/2017

Final Project for Political Science: International Relations Toby Murray – 11267275 – toby.murray@student.uva.nl

Cockroach States:

What makes for a successful drug trafficking country.

By

Toby Murray

Dr. Abbey Steele

Acknowledgements

First, my gratitude to the never-ending patience of my supervisor Dr Abbey Steele. For

supporting me through numerous changes of heart, across a number of interests, and for

convincing me that “water” does not constitute a thesis.

My thanks also to Annabel Flinn, for your unwavering support and faith through this project and

every other. For convincing me to come here, and for making sure I made it through.

As with any 24-year-old postgraduate student, thanks to my parents. I quite literally couldn’t

have done it with you.

To my friends Benoît Siberdt and Henry Stennett for their proof-reading and help. Any mistakes

are on your heads.

Abstract

Drug transit countries are becoming an increasingly widespread problem as globalised drug consumption increases. This thesis presents research regarding the question: what factors explain why some countries become drug transit countries and others do not? This thesis sets out three aims to achieve this. First it produces an appropriate measure of the relative level of drug trafficking in a drug transit country. Second, through statistical analysis, it determines what the generalizable factors that explain drug transit countries are, and third, it investigates the causal mechanisms that operate behind these factors. By employing a nested analysis to achieve these three aims, this thesis finds that institutional corruption is the key contributing factor to the settlement of drug trafficking, and that a weak judicial system acts to propagate drug trafficking further. It also tentatively suggests that high levels of poverty, emigration and low levels of state penetration also increase the level of drug trafficking.

Table of Contents

Acknowledgements ... 3

Abstract ... 4

CHAPTER 1: INTRODUCTION ... 7

CHAPTER 2: LITERATURE REVIEW ... 10

Corruption ... 10

State Penetration ... 11

Internal Conflict and Insurgency ... 12

GDP Per Capita ... 13 Poverty ... 14 Unemployment ... 15 Police Force ... 15 Judicial Strength... 16 Ethnic Cleavages ... 17 Migration ... 18 Inequality ... 18 Language ... 19 Conclusion ... 20

CHAPTER 3: METHODOLOGY ... 21

Heroin and Cocaine ... 22

Phase I: Quantitative Study ... 22

Dependent Variable ... 23

Data Collection ... 24

Production Data ... 24

Transit and Consumption Data ... 25

Transit_Seizure and Consumer_Seizure ... 26

Variable I: T-Value ... 28

Variables II & III: Ranked_CIA & Binary_CIA ... 30

Independent Variables ... 31

Corruption ... 31

State Penetration ... 31

Civil War and Insurgency ... 32

GDP Per Capita... 32 Poverty ... 32 Unemployment ... 33 Police Force ... 33 Judicial Strength ... 33 Ethnic Fractionalisation ... 34 Migration ... 34 Inequality... 34 Language ... 35

CHAPTER 4: RESULTS & DISCUSSION ... 36

CHAPTER 5: CASE STUDY - GUINEA BISSAU ... 46

Case Selection ... 46

Case Examination ... 47

Pre- drug trafficking history (1974 – 1998) ... 48

Emergence of drug trafficking (1998 – 2005) ... 49

Proliferation of drug trafficking (2005 – present) ... 50

Analysis ... 51

Theory Building ... 53

CHAPTER SIX: CONCLUSION ... 55

APPENDIX 1: CASE COUNTRIES ... 57

APPENDIX 2: DEPEDENT VARIABLE VALUES ... 58

CHAPTER 1:

INTRODUCTION

“They are terrorists, criminals, drug dealers … They take advantage of the open

borders, free markets and technological advances that bring so many benefits to the world’s people. They thrive in countries with weak laws and institutions and they show no scruple about resorting to intimidation or violence. Their ruthlessness is the very antithesis of all we regard as civil.”

- Kofi Annan

“I imagine someone who is fumigating his house” Claudia Mendez Arriaza, a Guatemalan journalist is once reported to have said, “Where do the cockroaches go? To the house of a neighbour who is trying to recover from past problems. That describes the situation in Guatemala” (Bridges, 2010). In the world surrounding the illicit trafficking of drugs, finding cockroaches in your house also means finding corruption, violence and addiction. Guatemala is not the only country to find itself afflicted. So too are its neighbours, Honduras and Mexico. So too are Tajikistan, Iran, Guinea-Bissau and Ghana. These countries form a subset of key locations in the global drug trade known as transit countries - countries through which drug traffickers, desperate to avoid “fumigation”, move in order to carry on their illicit trade in more auspicious conditions. Over the last 25 years, the rise of globalisation, in combination with the U.S.-led War on Drugs, has made drug trafficking an ever more essential and convoluted part of getting drugs to consumers. As technology made travel cheaper, and as stricter enforcement measures made travel more necessary, drugs have trekked through increasingly circuitous routes from production to consumption. Despite huge cost, the War on Drugs has had little success combatting global drug consumption. Instead, in its attempts to stem the flow from countries of production, it has ended up driving traffic through a greater number of countries, increasing the number of global players. (Bagley, 2012). This is the process of drug trafficking, and is described by Naylor (2003) as the multilateral exchange of inherently illegal goods between producers, distributors and consumers in a market-like context. Ngwube builds on this, adding that in the movement from source to user, there is any given number of intermediaries that are essential for the movement of the illegal commodity. These intermediaries, which are typically occupied by the distributors in this supply chain, are what are more commonly known as transit countries (Ngwube, 2014). Extending this, I will define a drug transit country as ‘a country where drugs are imported with the explicit intention of later exporting those same drugs’. Under this definition, an idealised model of a transit country should at all times have 0 net drug traffic, i.e. all drugs that come in should leave again. Whether a drug trafficking country must have its own trafficking infrastructure and organisations, or whether it can simply play host to a foreign drug trafficking operation is not important to this definition and both situations have been observed in a variety of transit countries (Desroches, 2007).

As drug trafficking methods have evolved, so too has the behaviour of the criminal organisations and networks involved. While drug trafficking used to operate primarily along geographical lines, traveling the path of least distance, increasingly inexpensive transport has forced this move to a market-like context. Literature has traced the parallel change that has seen trafficking operated by primarily large vertically integrated organisations like the Medellin and Cali cartels, that oversaw the end-to-end process from production to consumption, to trafficking through networks of organisations (Williams, 1998). With this move, drug trafficking organisations (DTOs) have changed their behaviour to match operating in a transnational environment. Chee and Benson (2014) discuss in a study that applies prospect theory, an economic theory, to the process of drug trafficking, that DTOs, much like firms, will operate to minimise loss and maximise profit. They conclude that as anti-trafficking law enforcement efforts in any one country increases, DTOs that are sufficiently good at reading the market will ‘offshore’ their operations to avoid costly risks, whether through forming new network partnerships, or moving operations overseas. Wainwright (2016) applies this logic to suggest that in a similar manner to certain countries offering more ‘business-friendly’ environments to companies and firms, other countries may offer more business-friendly environments to drug traffickers. However, he concludes, unsurprisingly, they are not likely to be the same conditions, but provides no further investigation into what these factors may be.

This is a common thread throughout the literature concerning the global drug trade. Much of the writing in both academic studies and organisational reports will appeal to country-specific factors to explain presence of drug trafficking in any one country. However, there is very little that constitutes a generalizable theory of what factors form a drug-trafficking friendly country. Factors appealed to range from socio-economic reasons, such as inequality and unemployment, to cultural factors such as transnational linguistic links. However, there are many poor countries, many countries speaking the same language and many countries with high levels of unemployment, but still only a select number of these host drug trafficking operations. Producing a general theory to explain why some countries become drug transit countries will allow the isolation of a key factor, or combination of factors, that are central to this settlement. Understanding this will allow better understanding of the causes of drug trafficking, not the symptoms. With countries that play host to drug trafficking seeing their economic growth suffer by 2.4%, understanding the systematic reasons behind this can help move towards addressing this problem (UNODC, 2016).

The question that this thesis will aim to answer then, is what makes a country a transit country? What are the factors that constitute a drug trafficking friendly country? Why do some countries become transit countries, and why do some see more trafficking than others? This question will take the form:

To answer this question, this thesis will have three aims. First is to determine an appropriate proxy to represent the relative level of drug trafficking within a country. Measuring drug trafficking is, by its nature, difficult to do accurately, but in order to understand which countries are used to traffic drugs, it is essential to produce a method of understanding the flow of drugs. By employing seizure data from the United Nations Office on Drugs and Crime (UNODC) and intelligence data from the Central Intelligence Agency, this thesis will produce three unique measures of drug trafficking. Second, it will identify general factors that explain why some countries becomes drug transit countries. This is the first step in building a theory that will answer the research question posed above. Third, and finally, it will qualitatively examine the causal mechanisms that operate between these factors and drug trafficking to determine the causation behind the relevant factors.

This thesis will begin with a review of the current literature on drug trafficking. For this, I will select 22 case-countries from a list of “prominent” drug trafficking countries provided by Rexton Kan (2016). Using case studies that describe the drug trafficking situation in each of these countries, I will extract the reasons given for the settlement and proliferation of drug trafficking domestically. These factors will form the hypotheses that are tested later. Following this, I discuss the methodological approach to answering the research question and the approaches used to produce three proxy measures of drug trafficking. This will take the form of a mixed method analysis. In this, I will first undertake a statistical analysis to isolate the generalizable factors that affect in-country drug trafficking. Then, I will investigate the causal mechanisms behind these factors with a case study examining Guinea-Bissau, a country that has seen a rapid increase in drug trafficking over the last 15 years, and now sees the country processing approximately 300 tonnes of cocaine a year. I conclude that institutional corruption and a weak judiciary are the key factors that explain why countries become drug transit countries, while tentatively suggesting that higher levels of poverty, migration and low state penetration are also contributing factors.

CHAPTER 2:

LITERATURE REVIEW

In this chapter, I assess the current literature on drug transit countries and collect a group of covariate factors that explain why specific countries have become drug transit countries. There is a lack of literature that discusses, holistically, the phenomenon of transit countries. The majority of literature concerning drug trafficking - and drug transit countries specifically - is instead made up of case studies. These studies often do not explicitly discuss drug transit countries, but drug trafficking within a country. This body of text will inform my analysis. Using a cross section of countries that were mentioned by Rexton Kan (2016) as significant transit hubs (found in Appendix 1), I will assess case studies of each country and collect a number of variables that have been suggested as causative factors. As my analysis will be an exercise in theory building, I also suggest hypotheses to be tested after the discussion of each factor.

Corruption

Corruption can come in many forms and as a result is difficult to define. For the purposes of this thesis I will use Nye’s definition, that corruption entails “any conduct that strays from the normal duties inherent to the public welfare as a result of private interests, be they family, clan, or friendship, to obtain personal benefit in money or social status” (Nye, 1967: 3). Public power in this context can mean any public institution or role, whether state government, local government, the police or judiciary.

Corruption is one of the most persistent reasons provided for a country’s position as a drug-transit country, with 17/22 case studies making some reference to the importance of corruption as a factor. For example, Duffy discusses how “trafficking in Belize is the direct result of complicity in the institutions that are intended to prevent smuggling” (Duffy, 2000: 556). While Duffy does not discuss how corruption may have led to the emergence of drug trafficking in Belize, she does say that “the ability of the government of Belize to combat trafficking was severely undermined by deeply entrenched corruption which reached into senior levels of government” (ibid). This is supported by Dudley who suggests that Panama’s eminence as a drug transit country began when “General Manuel Noriega [a senior politician and military officer at the time] also let Medellin Cartel traffickers use Panama as a safe-haven, bank and launching pad for drug shipments through the 1980s” (Dudley, 2010: 71).

However, Soberon discusses how corruption is a method employed by DTOs operating within a country to avert conflict with the state. Corruption, he says, is not a prerequisite for drug trafficking, but is a result of DTOs attempting to minimise their costs by offsetting violence. However, his examples for this behaviour are limited to countries responsible for drug production – Peru and Colombia – not countries considered to be exclusively trafficking countries. This is significant because in countries where cocaine

would have already been present, illicit activity must have necessarily existed regardless of government corruption (Soberon, 1997).

Though this questions the direction of causation, Paoli et al in their study of the drug trade in Tajikistan, state that “Corruption is not only an almost inevitable by-product of wholesale drug trafficking, but also one of its most powerful breeding grounds. In Tajikistan, it existed before the development of the illicit drug industry” (Paoli et al, 2007: 957). This explicitly states that corruption was required for the emergence of drug trafficking, and within their study of Tajikistan, the authors assert that “Such ‘Eastern’ corruption powerfully enhances Tajikistan’s advantage in the world opiate industry. Coupled with the growing means invested in corruption by drug traffickers, it also undermines the consolidation of democratic and accountable state structures” (2007: 957). Paoli et al. suggest that corruption facilitates the emergence of drug trafficking operations as senior officials neglect their democratic duty in pursuit of personal profit, which is later clarified as purely financial: “The bribe taker, who has often bought his own post, is required to pass a share of his earning to the boss who helped him” (2007: 957). That said, as the authors discuss in the first quote, corruption does not necessarily operate in one direction. Hence, if corruption can be both by-product and requirement, determining the causal relationship between drug trafficking and consumption is clearly non-trivial. This is corroborated by Dorn et al. (2015) who suggest that corruption and drug trafficking enjoy an endogenous relationship, where an increase in one will typically see an increase in the other.

H1: Countries with higher levels of corruption will see higher levels of drug trafficking

State Penetration

State penetration is second only to corruption as a systematic reason appealed to by the case studies, with 14 out of 22 studies referencing the importance of weak state penetration to the settlement of drug trafficking. State penetration as I understand it will be in line with Soifer’s territorial reach where “the territorial reach of the state defines the geographic area within which its policies can be enforced” (Soifer, 2008: 243). In his paper, Soifer uses the example of taxation to illustrate that the territorial incidence of tax (or the amount of tax that is collected per capita regionally) shows the territorial reach of the state. For example, Monaco – a country with high territorial reach – should collect a consistently high level of tax from all regions, whereas the Democratic Republic of Congo – with a lower territorial reach – may see a much larger variation when comparing urban to rural areas. For the purposes of this thesis, we will understand “territorial reach of the state” as “state penetration”.

In the drug trafficking literature, state penetration is either referred to directly, or in terms of “porous borders”, “unpatrolled coastlines”, etc. For example, in Duffy’s study of Belize, she refers to Belize’s “large tracts of forested land, unprotected coastline, numerous Cayes and inland waterways, and unpopulated

rainforest and coastal area [which] present the opportunity for significant transhipment of illicit narcotics” (Duffy, 2000: 555). This is echoed by Farah who suggests that the drug trade was able to penetrate into El Salvador because the country has “more than 300 unmonitored points of entry” (Farah, 2011: 10). The theory then, is that a lack of state presence allows organised crime to enter and settle under the radar of the government. This may occur in a number of ways. First, it is plausible that governments have little idea a smuggling operation is settling within their territory if there is no government presence where this is happening. Second, areas without government presence afford a degree of plausible deniability to corrupt governments, allowing them to hide trafficking operations from public eyes. Finally, this could refer to a situation where there are inhabited areas of a country that are not under state control.

This final scenario is illustrated by Chandran in his examination of Pakistan, where he discusses how “with the process of nation building yet to be completed, there are various zones that are improperly coordinated with the mainland. The Sardar and Malik systems in Pakistan emphasise less on the authority of central law over the majority of areas in the NWFP [North-West Frontier Province] […] in Pakistan, the Malik system is considered to be the greatest obstacle in fighting drug trafficking” (Chandran, 1998: 908). This emphasises that if there are areas that are not ‘unmonitored’, but are outside of government remit – perhaps having set up independent institutions to make up for a lack of infrastructure provided by the government – then this lack of state penetration offers opportunity for DTOs.

H2: Countries with lower levels of state penetration will see higher levels of drug trafficking

Internal Conflict and Insurgency

The literature suggests that conflict may impact on the origin of drug trafficking in two ways. The primary mechanism concerns ongoing civil war or insurgency within a country. This effect is discussed within civil war literature, where it is generally accepted that the presence of drugs in a conflict acts to lengthen the duration of the civil war (Ross, 2004). This is particularly relevant in the case of insurgency, where rebels are able to fund campaigns effectively through the trafficking of narcotics. Due to international prohibition and pressure, this mostly acts to the benefit of rebel groups (Ross, 2004). Cornell focuses on the Islamic Movement of Uzbekistan (IMU), an insurgent group operating on the Afghan border of Uzbekistan and describes how “in the complex political climate of Afghanistan and Central Asia at the time, the IMU was in a singularly well-placed position to control the drug trade from Afghanistan to Central Asia” (Cornell, 2007: 630). Cornell describes how, in order to fund campaigns and attacks, the IMU would turn to the trafficking of Afghan opium. Further, he explains how following the termination of funding to the IMU from Al-Qaeda the IMO increased trafficking to fill the gap in lost income. This same mechanism can be observed in Lebanon. Marshall states that “illicit drugs became one of the prime sources of funds for militia rearmament, operations and enrichment” and explains how the drug trade went on to “inflame the civil war” (Marshall, 2005: 75). Further, he proposes that the mechanism for this was trade with Western

European DTOs, where drugs were smuggled through the Lebanon and swapped for guns. The interaction in both cases is that in existing conflict, insurgent groups with little external funding are likely to resort to drug trafficking and, once settled, this trade is likely to stay entrenched.

However, the UNODC suggests that “while the drug trade provides some funds for the conflict, more significant is the cover the conflict provides for the drug trade. Those who profit most from heroin trafficking are professional criminals” (UNODC, 2010: 232). Rather then, DTOs may target conflict-afflicted states as a means to an end as opposed to trafficking being driven by combatants.

A second mechanism is proposed by Farah. He describes how “after the conflict officially ended in 1992, small groups on both sides did not disarm, but kept their weapons, safe houses and logistical hubs […] these groups, often taking over the leadership of local gangs that were forming at the time, immediately provided a new level of sophistication and brutality to several types of existing criminal activities” (Farah, 2011: 8). Farah suggests then, that even if not engaged in drug trafficking during the war, the armed groups that are left without purpose, and often with extensive means to execute organised crime will turn their hand to drug trafficking post conflict. Both mechanisms highlight that a history of civil war may therefore have a positive correlation with that country being a transit country.

H3: Countries with a history of civil war or currently experiencing civil war will see higher levels of drug trafficking

GDP Per Capita

Of the 22 countries used from the state department list, 21 are found in the bottom half of the World Bank’s ranking of countries by GDP per Capita. This measure gives a proportional representation of the wealth of each country, and doesn’t discriminate against smaller countries in terms of wealth, where GDP is “the sum of value added by all resident producers plus any product taxes” (World Bank).

GDP per Capita, in contrast to measuring poverty, can be used as a measure of a country’s economic strength. This is appealed to by a report by the US State Department as a cause of transit countries – “Factors that contribute to drug trafficking include… inadequate educational and employment opportunities for at-risk youth who engage in crime; and a struggling economy” (US State Department, 2014: 194). However, no reason is given for why we should expect a struggling economy in and of itself to contribute to a country becoming a drug transit country. This is echoed by Vigh, who describes how in Guinea Bissau “[A] combination of economic decline, diminishing resources and withering livelihood…[means] we are not looking just at specific groups of people who see possibilities in trafficking cocaine, but rather a whole population grasping for even the slightest or most risky possibility of gaining a better life” (Vigh, 2012: 149). This does not provide a specific reason for why a struggling economy may turn a country into a drug trafficking hub, but does suggest that the detrimental effects of the dire state of

the economy coerces civilians to turn to drug trafficking to earn a living. This means that drug trafficking settlement may be a bottom up mechanism, although it is unclear whether this mechanism is caused by a struggling economy or by a resultant factor, i.e. poverty, unemployment.

Akyeampong makes a separate point when he describes how “the Ghanaian economy was in recession in the 1990s […] free trade had opened the economy to cheap imports […] manufacturing had collapsed as Ghanaian goods had become uncompetitive […] small business owners – and some big ones – turned their attention to drug trafficking” (Akyeampong, 2005: 429). This suggests, rather than the economy working through individuals i.e. poor civilians, a weak economy turns legitimate businesses to illegitimate means of income. Businesses with employees and methods of distribution make for good trafficking partners to DTOs. However, it remains to be investigated how this would facilitate the settlement of drug trafficking, whether licit businesses seek out drug trafficking as a method of income, or whether drug trafficking must already be present.

H4: Countries with lower GDP per Capita will see higher levels of drug trafficking

Poverty

The world’s biggest drug king-pin Joaquin ‘El Chapo’ Guzmán Loera is quoted as saying his childhood poverty was the defining factor in his choice to become a drug trafficker (Penn, 2016). However, the academic or institutional literature surrounding the effects of poverty on drug trafficking finds limited evidence of any link. Gberie suggests “drug traffickers have exploited the widespread poverty and corruption to co-opt government officials, military and law enforcement officers, and political and traditional leaders into an opportunistic network that underpins a very profitable criminal enterprise” (Gberie, 2016: 2). However, no explanation is offered for how ‘widespread poverty’ has contributed to the status of both Guinea and Mali as drug trafficking countries. Indeed, the inclusion of corruption in this analysis undermines Gberie’s implication of the relevance of poverty. In each case study reviewed, while poverty is mentioned (as a number of countries in this subsection suffer from high levels of relative poverty), few appeal to poverty as a causal factor in a country being a drug trafficking hub. Outside of the case studies used, there is a dearth of evidence in political science literature in general that there should be any link between poverty and drug trafficking.

However, a study by Dunlap et al. into the link between poverty and domestic drug dealing in the US states that there is “a clear connection between poverty and entrance into the drug market” after controlling for relevant factors such as race and social class (Dunlap et al, 2010: 123). In their study, they found that a majority of respondents would appeal to poverty as their primary reason for pursuing a career in drug dealing, as it could provide for their family when a legitimate primary income was otherwise unattainable. However, the application of this study is limited because it only investigates small scale drug dealing in the

US, and makes no suggestion of the applicability of this link at larger scales of drug dealing in other environments. However, enough suggestion has been made that poverty should exist as a causal factor, therefore:

H5: Countries with higher levels of poverty per Capita will see higher levels of drug trafficking

Unemployment

In the 2012 World Drug Report, the UNODC reported that “Unemployment appears to be another key socio-economic driver of drug trafficking and illicit drug use. Among young males, in particular, unemployment increases the likelihood of participation in the illicit drug trade and illicit drug use” (UNODC, 2012: 88). However, as with poverty, there is little evidence in our case studies of a causal relationship between unemployment and drug trafficking. There is a brief discussion by Beittel that in Mexico “Unemployment […] provide[s] ready recruits for gangs who are hired by the DTOs to fight as their proxies”, proposing that unemployment may not directly aid drug trafficking, but may facilitate DTOs by providing them with disposable assets that can defend their interests (Beittel, 2011: 20). There is no further discussion of this point though. In a separate paper the UNODC suggests that unemployment generally drives retail level drug dealing, not drug trafficking, as this is the labour-intensive element of the trafficking process. None of the studies appeal exclusively to unemployment as a causal factor, and in all cases, it is supported by another causal factor. Interestingly, most attempts to justify unemployment as a factor largely fall to appeals to poverty. Regardless, testing this factor will be key to building a general theory.

H6: Countries with higher levels of unemployment will see higher levels of drug trafficking

Of note, the converse relationship between the effects of drug trafficking on the rate of unemployment may be more significant. A report by the UNODC proposes that drug trafficking will often lead to an increase in drug use locally, and further to this, drugs like cocaine and heroin can often lead to an increase in the levels of unemployment, as addiction begins to affect productivity in the workforce (UNODC, 2014: 7). This can result in a vicious cycle, where boredom from unemployment increases the level of drug abuse and as a result increases unemployment further. Therefore, should there be a correlation between drug trafficking and unemployment it should be treated carefully.

Police Force

There is a distinction drawn in this thesis between a weak police force and a weak judiciary. A weak police force is underfunded, understaffed and undertrained, whereas a weak judiciary is one with low efficiency,

low independence and weak sentencing power (WEF, 2016: 64). The literature surrounding judicial strength will be discussed in the next section.

In his examination of drug trafficking in Haiti, Griffiths claims that “Haiti has several factors that have facilitated trafficking […] poorly monitored coasts, mountainous interior, about twenty unpatrolled airstrips, inadequate law-enforcement resources and corruption” (Griffiths, 2010: 74). Although there is no further exploration of these reasons, we have discussed all but inadequate law-enforcement resources already in this review. However, Griffiths provides no suggestion of how poor policing resources will influence drug trafficking likelihood. The same is true of Rosenburg who explains that “conditions in Honduras are right for this role [as a cocaine transhipment point…] its Caribbean coast is underpopulated, the adjacent bay islands are under-policed; and there are an estimated 100 clandestine airstrips throughout the country” (Rosenburg, 1988: 148). Unfortunately, he too provides no evidence for why this is the case. Paoli et al. propose that an underpowered police force may result in drug trafficking because under-provision means that “the Tajik state agencies were unable to protect the country borders or to prevent any type of smuggling […] they lacked sufficient human and technical means to effectively patrol the long Tajik-Afghan border” (Paoli et al., 2007: 956). This suggests that an underfunded, understaffed police force is unable to keep up with a DTO that is capable of highly technological methods of smuggling across large borders. This is supported by a report by the International Crisis Group which proposes that one reason drug smuggling was able to permeate Guatemala was an under-policed border with Honduras. The provide the example of “a police officer at El Florido […] [who] said his substation had only one working car that often lacked gasoline” (ICG, 2014: 1). In this case, a DTO does not even need to be particularly savvy or tactically advanced to outperform the police force and enter the country.

H7: Countries with weaker police forces will see higher levels of drug trafficking

Judicial Strength

As discussed earlier in my thesis, judicial strength will refer to three factors as defined by the World Economic Forum. First, the independence of the judiciary, or how far removed it is from the influence of government or powerful individuals. Second, the efficiency of the judiciary, or how quickly the law can be enforced and effective sentencing can occur. Finally, the power of the judiciary, or whether the judiciary has the legal power to enforce its will or has the legal instruments to prosecute drug traffickers where the judiciary refers to the judges of each country collectively (WEF, 2017).

Emmers’ investigation into transnational crime in Southeast Asia reveals that judicial strength is a factor in two countries positions as drug transit states: “The rule of law is only partially applied in Myanmar and Cambodia; two failed states where corruption and domestic struggles for control undermine administrative

and judiciary systems. They therefore provide ideal environments for organised crime” (Emmers, 2003: 10). This is echoed by a Council of the European Union report into Jamaica, which proposes that “despite a relatively high interdiction rate, traffickers persist given the low conviction rates and weak judicial systems” (Council of the European Union, 2017: 2). A number of other case studies of countries also made reference to the importance of a weak judiciary for fostering drug trafficking, including Mexico, Thailand and Ecuador (State Department, 2014; Bonner, 2010). This suggests that even if traffickers are caught in a country, they are still able to persist in the same location because a weak judiciary will let them onto the streets again to continue smuggling. The mechanism through which this would operate is unclear, but common sense would tell us that if a DTO does not think it will lose important members to interdiction, they would continue to operate out of a location with this safeguard.

H8: Countries with weaker judiciaries will see higher levels of drug trafficking

Ethnic Cleavages

Chandran observes in his study of trafficking in Asia that “Groups, mostly of tribal or ethnic origin, also come to be used by powerful drug mafias… The “Shan United Army”, the most powerful drug network in Myanmar, utilises its ethnic resentment as a screen to effectively restrain any governmental action toward controlling drug trafficking” (Chandran, 1998: 907). The Shan are an ethnic minority in Myanmar, making up just 6% of the country, and in this deeply divided society, drug trafficking operating through an ethnic minority provides cover because “measures to combat drug trafficking become extremely difficult, as at stake is the delicate ethnic peace” (Ibid.).

Exploitation of ethnic cleavages seems to occur in a number of the case countries, including Mali, where Tinti et al. highlight how “drug trafficking networks in Mali were often split along ethnic lines, especially involving ethnic Tuareg Rebels who considered themselves alienated from the flow of central state resources” (Tinti et al., 2013: 2). This follows a slightly different mechanism, though still operating from ethnic resentment, where ethnic groups that suffer state discrimination are able to obtain resources from an alternate source such as drug trafficking. This is reflected in examples from Tajikistan and Thailand as well (Paoli et al, 2007; Windle, 2015). Ellis suggests that this is not unusual, this process also being found in Nigeria, where he also observes that different ethnic groups also dominate other illegal trades out of Nigeria, including “the trade in prostitutes from Nigeria to Europe, which is overwhelmingly in the hands of the people from Edo State.” (Ellis, 2009: 189).

Migration

The importance of emigration to the success of drug trafficking is discussed by Akyeampong, who suggests there exist “transnational networks that have emerged to promote drug trafficking” (Akyeampong, 2005: 430). From a Ghanaian perspective, he discusses how transnational networks have formed as a result of migration patterns and diasporas. Akyeampong uses the example of a heroin ring that transported drugs from Thailand, through Ghana to the USA through a series of Ghanaian nationals settled in each country. He suggests that increased ability to travel globally and establish communities in foreign nations as a result of globalisation has increased the efficiency of the drug trade, and will often lead back to one’s original country of origin as a base. This is supported by Paoli et al. who observe that “Between 1992 and 1999 more than 600,000 people left Tajikistan. At least 60,000 of those immigrants fled into Afghanistan, a flow that not only revived common ethnic and clan ties, but also created the basis for future transactions in opiates” (Paoli et al, 2007: 958). Further, they argue that a similar situation on the other side of the border (a large Tajik immigrant population in Russia) has created similar linkages and effects for the traffic of heroin to Russia. The proposal then is that countries that have a large number of emigrants may see transnational networks emerge between those remaining with their country of origin and drug production regions/drug consumption regions. The networks may be vulnerable to exploitation by DTOs, and may create risk of becoming a transit country for the country of origin.

A separate effect to be aware of with respect to migration is discussed by Schloenhardt, who highlights how “strong connections have been found between drug and migrant trafficking activities” (Schloenhardt, 1999: 14). He explains that a growing industry of migrant trafficking is often operated by organised crime elements that also may be central to the drug trade. Criminal organisations that will be successful at migrant trafficking will tend to be organisations with “well-tested trafficking routes and with personnel located in different countries along these routes” (Schloenhardt, 1999: 14). While he does not provide specific examples for this, Schloenhardt suggests that this is particularly prevalent in Asia. We can therefore infer that countries which also offer a lucrative market in migrant trafficking will be an attractive location for DTOs to operate.

H10: Countries with higher levels of emigration will see higher levels of drug trafficking.

Inequality

Inequality is referenced a number of times by the UNODC as a reason that countries become hosts to drug trafficking (UNODC, 2014; UNODC, 2016). In addition, crime literature has found significant links between inequality and crime (Wilkinson et al, 1998). However, no reference was made explicitly to the effects of inequality in any of the studied case countries. There is a brief suggestion in Marshall’s investigation of the Lebanese drug trade that “the long-standing traffic in drugs supported rather than

challenged existing political and economic power relationships as wealthy landowners […] reaped the bulk of the profits” (Marshall, 2005: 5). However, it is unclear whether this is a commentary on the inequality that affects the Lebanon, or if it is a reference to the effects of corruption. Nevertheless, the UNODC indicates from a holistic position that “societies characterized by high income inequality tend to be more prone to crime, and in some extremely unequal societies, members of marginalized groups may view involvement in criminal activities such as drug trafficking as the only feasible strategy for upward social mobility” (UNODC, 2016: 79).

H11: Countries with higher levels of inequality will see higher levels of drug trafficking

Language

Finally, in their book on drugs and failed states, Inkster and Comolli discuss a potential link that explains the spread of drug operations from South America to West Africa based on linguistic links between countries: “Linguistic links, especially in Portuguese-speaking countries […] drew Latin American traffickers to the region” (Inkster and Comolli, 2012: 105). Inkster and Comolli suggest that the reason drug operations are so prevalent in Guinea-Bissau (and Cape Verde) is the Lusophone population. While Brazil may be the only major South American Portuguese speaking population, the economic dominance of Brazil on the continent, combined with Brazilian immigration, has led to Portuguese being spoken around the Southern Cone as a working language. The relevance of linguistic links is not discussed much in the case studies, but the importance of linguistic links to trade is discussed extensively in economic literature. A study by Havrylyshyn and Pritchett uses a gravity model to compare cultural links between bilateral trade partners. Their results show that sharing a language has a greater effect on increasing trade between two countries than distance (Havrylyshyn and Pritchett, 1991: 6). A similar study conducted by Frankel et al. finds that a similarity in linguistic links can result in a 55 percent increase in trade between those two countries than would otherwise have occurred (holding distance and trade barriers constant). This is an empirical study and a number of reasons are provided for this being the case, including similar languages often also overcoming cultural barriers (Frankel et al, 1997: 74).

While the trade modelling is obviously limited to the licit economy, the basic principal could be seen to apply to trade in illicit goods as well. We can imagine that when investigating alternative countries to operate through, a Portuguese-speaking DTO would choose to operate across the Atlantic Ocean in a country where both partners can speak their primary language, as opposed to settling in nearby Arabic-speaking Mauritania.

H12: Countries that share common languages with major drug producing countries or major drug consuming countries will see higher levels of drug trafficking.

Conclusion

In conclusion, the indicators drawn from the literature surrounding drug trafficking countries seems to follow three mechanistic themes

i) Corruption and GDP per capita suggest the importance of top-down settlement and proliferation of drug trafficking.

ii) Poverty, unemployment and another understanding of the effects of GDP per capita suggest the importance of having a readily available, ‘on-the-ground’ workforce, and emphasise bottom-up settlement and propagation

iii) Civil war, police force, judicial strength, ethnic fractionalisation, migration, inequality and language emphasise the importance of environmental factors that facilitate trafficking. These are factors that do not require interaction on the ground in a transit country, but help create a ‘drug trafficking’ friendly environment conducive to the settlement and proliferation of drug trafficking.

Having established a set of covariates and the hypotheses that will test these covariates, the next stage will be to establish a methodology and set of proxy measures that will represent both the independent variables (the covariates) and the dependent variables (trafficking of drugs).

CHAPTER 3:

METHODOLOGY

The three aims of this thesis are to: 1) determine an appropriate proxy to represent the dependent variable (the relative level of drug trafficking within a country), 2) identify relationships between relevant factors that affect the relative level of trafficking within a country, and 3) examine the causal mechanisms that operate between the dependent and relevant independent variables. To accomplish this, I employ a mixed-method nested analysis. This is a two-stage process, combining both quantitative and qualitative research methods. First, by employing a large-N cross-national study, I assess the statistical significance and correlative strength of each of the indicators found in the literature from the previous chapter. This provides an overview of the general relationship between each covariate and the dependent variable, indicating which factors can be used to describe the global narcotics trade and which are country-specific factors. The process of this phase will address aims 1 and 2.

Following this, a qualitative case study will examine the origins of drug trafficking in Guinea-Bissau. Conducting a case study complements the previous quantitative study by testing the measurement validity of my statistical study – examining whether proxies that have provided statistical significance are correctly measuring the relevance of a covariate. It also allows for process tracing to examine exactly how covariates link to the dependent variable and provide a theoretical framework explaining which factors contribute to countries see higher levels of drug trafficking. This phase of analysis will address aim 3 and is discussed in chapter five.

A nested analysis is an appropriate choice for this study because it allows for both an examination of the general relationships that explain which countries are afflicted by drug trafficking and for understanding the mechanisms behind causation. As discussed by Lieberman, “The strategy of combining the two approaches aims to improve the quality of conceptualisation and measurement, analysis of rival explanations, and overall confidence in the central findings of a study” (Lieberman, 2005: 436). This captures the essential nature of a nested analysis, especially when applied to a subject that is so laden with potential measurement error as drug trafficking. Evidence of correlation provided by quantitative study does not equal causation, but combination with a qualitative study allows evaluate the relevance of each factor with greater certainty. Additionally, data concerning the drug trade is limited in its availability and quality, meaning there is a high potential of measurement error associated with our dependent variable. Measurement error can result in a systematic error and subsequent misinterpretation of the degree of correlation. Measurement validity can therefore be tested qualitatively to assess systematic error.

Heroin and Cocaine

For the purposes of this thesis, research is limited to investigating transit hubs for heroin and cocaine. While this may exclude certain important transit countries from the research, or decrease the magnitude of drug trafficking ascribed to a country, it will also decrease the associated measurement error for a number of reasons:

First, the production of cocaine and heroin, unlike other drugs, occurs in a limited number of countries. Cocaine is limited to the Andean region of South America (which provides the only sustainable environment for the growth of the coca plant) and heroin, while not limited to any one country from necessity, can trace 98% of its global production back to three Asian countries (UNODC, 2016). Other major drugs such as marijuana and amphetamines are less limited geographically in where they are synthesised or grown globally. Production data is largely collected by international institutions, and maintaining accurate records in a limited number of countries allows for greater reliability and accuracy than global data.

Second, while the UN Single Convention on Narcotic Drugs 1961 introduced international prohibition of narcotic drugs, this is only true globally of heroin and cocaine (Wainwright, 2016). While a drug like marijuana is legal in relatively few countries, these often offer important ports for drug trafficking, and may behave anomalously in global comparisons This is more apparent in the trade of synthetic drugs, which often inhabit a grey area of legality globally as often they are produced as ‘legal highs’ (Reuter and Greenfield, 2001).

Finally, when compared to synthetic narcotics, cocaine and heroin are better suited to study because they come in a limited number of forms. While the UNODC database on narcotic seizures has only 4 different forms of cocaine-based narcotic, and 5 different forms of opiate – each of which has a well-defined purity associated with it, there are over 50 different types of synthetic drug. As tracking narcotics trafficking relies on tracking quantities of pure substance, excluding these from our study will help to maintain accuracy of seizure measurements (UNODC, 2017).

Phase I: Quantitative Study

This section will discuss the methods employed to produce an appropriate dependent variable to measure the relative amount of drug trafficking with a country. This will be followed by a discussion of the proxies used to represent the independent covariates that will be used in the quantitative analysis. For the dependent variable, three separate measures were produced. These were analysed using either a linear regression analysis or a logit analysis. The choice of analytical technique differed based on the dependent variable. Though the research is designed to only utilise a single measure, the data that this design is based on is

largely incomplete. Therefore, over the course of the development the measure, two robustness checks were produced, totalling three dependent variable measures: one continuous, one discrete and one dummy variable. Linear regression was applied to the first two measures, and a binary logit analysis was applied to the third. The unit of analysis for all variables discussed will be at country-level and the list of countries used is taken from the UNODC Database (2015), which contains 194 countries.

Dependent Variable

This section will address aim 1 – producing an appropriate proxy to represent the dependent variable. It is conceivable that using raw seizure data concerning the amount of drugs captured by law enforcement within a country would be a satisfactory proxy measurement. However, as I am interested in understanding the amount of trafficked drugs specifically, this would not provide a sufficient distinction between the purpose of drugs in each country. 500 one kilogram seizures caught at retail, where it is intended for consumption, would count the same as a 500-kilogram seizure in a trafficking operation. It is the distinction between these two kinds of seizure that is key to producing an appropriate measure.

Theoretically, the best way to produce a measure of the relative amount of drug trafficking of a country would require three sets of data on quantities of narcotics: the amount of narcotics produced by each country, the amount of narcotics consumed by each country, and the amount of narcotics seized in each country. Using these three sets of data, the following equation can be devised from a standardised linear relationship:

𝑻 =

𝑺+𝑷+𝑪 (Eq. 1)

where T is the relative level of drug trafficking in a country, S is the weight of narcotics seized within each country, P is the weight produced and C is the weight consumed. Each value must represent the weight of active ingredient to account for cutting and dilution over the course of the trafficking process (UNODC, 2016). The resulting value of T is a unit-less ratio where a negative value indicates a country as either a ‘producer’ or a ‘consumer’, and a ‘trafficker’ is described by a positive value of T. While this method means that consumer and producer states are indistinguishable by this measure, this is justified as we are only interested in the amount of drugs that are trafficked through each country. If necessary, the equation can be restructured to obtain the value of either a producer or consumer. The equation is standardised to the range 1, giving the relative measure of drug trafficking in each country. This will obscure measures of magnitude, but the aim of this study is to understand countries that are primarily used for trafficking, not the countries that see the largest quantities of drug trafficked.

The value of T is therefore a measure of a country’s relative incidence of narcotics, where those narcotics are best understood not as a result of production or for consumption by the populace, but are imported with the intention of further transport to a new location. If net drug traffic is calculated by the equation:

𝑵𝒆𝒕 = 𝑰𝒏 − 𝑶𝒖𝒕 (Eq. 2)

then an idealised transit hub should have a net value of 0.

Data Collection

The difficulties of collecting data concerning illicit trades are well-documented (Reuter and Greenfield, 2001). A key issue is the quality of data that can be found concerning the drug. As per Dorn et al. highlight, data on illicit trades is “generally either lacking (in respect to the reality of uncaptured organised crime), of uncertain quality, ambiguous, or somewhat theoretical” (Dorn et al, 2005: 2). Each of these factors applies to every stage of the drug trade. During the construction of this measure, attempts were made to mitigate each of these shortcomings.

Production Data

Two primary sources of data are used throughout the literature to estimate the level of production in the international narcotics trade: the UNODC and the United State Office of National Drug Control Policy (ONDCP). Inkster and Comolli provide a comparison of the two sources. Data from the UNODC is, the suggest, the more methodologically rigorous of the two, offering a combination of both national government statistics and independently conducted, on the ground investigations. In contrast, the ONDCP data relies on satellite surveys of drug-producing areas, but this approach is limited to surveying only known production sites – often underestimating production numbers (Inkster and Comolli, 2012). While the UNODC also routinely underestimates production data, its methodology is rigorously applied to all regions. While this will produce a systematic error, which will result in underestimation of the dependent variable, it will apply similarly globally - avoiding bias when compared to data from the ONDCP (King et al., 1994: 156).

Additionally, the UNODC does not rely on directly measuring the amount of finished product, or actual drugs, from each country. Instead, measurements were made from production yields of each crop. By multiplying production yield against the measured quantities of each crop an estimate of drug production can be made. A weakness at this stage is again underestimation as yields produced by new technology are unmeasured in calculations, which rely on known - and likely outdated - methods, (Inkster and Comolli, 2012). Again, this is a systematic error across all data and calculation, and as discussed above, should not produce a bias in the dependent variable. Non-surveyed countries or regions also cause similar

underestimation, which unfortunately is not systematic, but will be mitigated by focussing on opiates and cocaine, which, can only be found in certain regions. This thesis will therefore utilise UNODC production data to produce the dependent variable.

The production data used will be taken as a per year average mean from the 2010-2015 annual reports by the UNODC. Taking an average over five years will mitigate any missing data that may occur on an annual basis or anomalous years that deviated from the norm of production because of external factors. The period of five years was chosen because this matches the available seizure data which will be used for transit and consumption data.

Transit and Consumption Data

The most widely used seizure data, and the data used in this thesis, is collected by the UNODC by collating reports from domestic police forces and regional drug policy and enforcement bodies (UNODC, 2017). This is collected as a detailed record of the seizures reported by each body with details of exactly what region in which country the seizure was made, including details of what drug and what quantity was captured. As discussed below, it also contains data on the country of production, country of last origin (or the country the drug was received from) and destination country, though each of these measures is markedly less available. Additionally, these values can be inaccurate, as the information that travels with each trafficking shipment is purposefully limited. A single DTO will frequently not oversee the end-to-end supply of each shipment (Williams, 1998). A change in trafficker at any point in the supply chain can obscure data, protecting upstream organisations. Additionally, seizures normally result in the arrest of participants who are not aware of the bureaucracy surrounding each shipment further than their specific journey (Kenney, 2007). Consequently, a sum of each measure will not give the desired size of production (origin), consumption (destination) or transit (seizure or last point). There is little alternative data available, so method of data manipulation is proposed later that can be applied to mitigate these shortcomings. Consumption data also has a number of innate weaknesses. Foremost amongst these is measurement validity - once a drug is consumed, it is often impossible for its quantity to be measured accurately in retrospect. Most measures of consumption rely on testimony and survey data from drug users, but these are vulnerable to subjective reporting. As Inkster and Comolli highlight, only 0.05% of respondents in Japan admit to consuming illicit narcotics when surveyed, despite Japan having a significant level of retail seizures (Inkster and Comolli, 2012: 17). As a result, drug consumption data, especially outside of richer developed countries where surveys are not conducted as regularly or comprehensively, is particularly sparse. Furthermore, when consumption data can be found, it is often reported as percentages of the population who have consumed drugs in the past year (UNODC, 2017). This does not give a useable quantity of drugs consumed, and there is no reliable measure of the average amount consumed, as this value varies hugely by

both country and individual. A similar method to address the inadequacy of drug consumption data will also be discussed beneath.

Transit_Seizure and Consumer_Seizure

This section will discuss a method to utilise the UNODC seizure data on origin, last origin and destination to represent transit and consumption data. Using this method, eq. 1 will be corrected to use these values produced from the raw data. There is a problem with seizure data that relates to its method of data capture – that seizures often do not relate to which countries are most affected by the drug trade (Dorn et al, 2005). Instead, seizure data shows which country is best at capturing drugs that flow through their country. While often magnitude and quantity of seizures can be used to show a correlation, the magnitude is often greatly skewed towards richer developed countries, which are able to afford expensive interdiction and seizure efforts that are required to capture drugs within their territory. As Csete and Sánchez discuss, “Drug seizures are a reflection of the level of activity of law enforcement officials. Judging year-to-year trafficking trends from seizure data assumes that interdiction activity of the police is constant or at least similar [in every country]” (Csete and Sánchez, 2012: 6).

As well as giving values for transit and consumption quantities of narcotics, the method employed in this section will mitigate this issue by disregarding the importance of the country of seizure and instead focussing on the data that can be garnered from that drug seizure, meaning that even if a seizure is made in Spain, but has come from Nigeria, the seizure will count for both as transit countries (or transit and destination). While a separate issue with seizure data double counting, where a seizure may be counted by both countries if seized in transit between the two, this will not be an issue for this analysis, as if a kilogram of heroin passes through 4 countries, this should contribute to each of those four countries’ T-values. It also means that a seizure is only counted once for the destination country as this will be a unique value for each seizure. We will use the following model to determine the values of transit and consumption for each country1_.

where SCn is the seizure amount for the country in which the seizure is made (where n refers to the country),

“Origin” is the same seizure amount attributed to the reported origin point of the seizure recorded in UNODC seizure dataset. “Last” is the same seizure amount but attributed to the country of last origin reported in the UNODC seizure dataset and “Destination” is the same seizure amount attributed to the country of destination as reported in the UNODC seizure dataset. X is the theoretical producer country, T1 &T2 are theoretical transit countries and Y is the theoretical consumer country. This model characterises

a theoretical “simple supply network” consisting of four nodes. For clarity, I have used two transit countries to model for at least one change in each value at every stage, but this can be reduced to the simplest configuration (Producer  Consumer), or expanded to include as many nodes as the chain requires. Where a seizure is made, if the seizure country is the same as the reported origin then we are able to say with confidence that this seizure can be attributed to that country as a Producer_Seizure (meaning we count the seizure amount as an amount of drugs produced by that country). Where a seizure is made and the seizure country is not the same as the reported origin, the reported last transit, or the destination then we can say with confidence that this seizure can be attributed to that country as a Transit_Seizure (where we count the seizure amount as an amount of drugs trafficked through that country). Finally, where a seizure is made and the seizure country is the same as the destination country then we can say with confidence that this seizure can be attributed to that country as a Consumer_Seizure (where we count the seizure amount as a quantity of narcotics that will be consumed within that country). We can see from the model above that these are the only rules that hold true across all seizures and reported values. Any other inferences cannot be made with certainty: for example, the reported origin country of a seizure made in country Y

could be one of three values, so to count all origin values as a Producer_Seizure does not work. Following these rules can correctly attribute the drug seizures to each country's appropriate status in the drug trade. The rules are summarised below.

Where S is the country of seizure, O is the reported origin, L is the reported last origin and D is the reported destination.

Variable I: T-Value

While this system derives a value to represent the level of production per country, I will use the production data from the UNODC instead. This measure has greater validity than the proposed Producer_Seizure measure, as it more accurately corresponds to the real-world value for cocaine and heroin produced by each country. While this value will be significantly larger than the values for Transit_Seizure and Consumer_Seizure, its magnitude will not affect the standardised, relative measure of the dependent variable. A large producer value will still reduce to a value of -1, indicating a producer state. While the principal behind Transit_Seizure and Consumer_Seizure is abstract and does not accurately represent the actual level of drug transit or consumption within a country, it still has a high level of measurement validity. While the quantities of seizures represented in these equations will be significantly lower than their real-world levels, the use of a relative, standardised measure will negate the need for magnitude validity. Instead it will give an accurate representation of the balance of drug trafficking and consumption within a country which is what aim 1 requires. An obvious weakness is that should a country find itself consistently misreported by traffickers, this will give an incorrect value of T. However, as the dataset used has over 200000 applicable data points, the probability of this significantly affecting the values of all countries remains low.

Transit_Seizure and Consumption_Seizure data will be taken as a per-year average over the period 2010-2015, as this is the data available.

To accommodate for these new values, equation (1) becomes:

𝑻 = 𝑻𝒓𝒂𝒏𝒔𝒊𝒕_𝑺𝒆𝒊𝒛𝒖𝒓𝒆 − (𝑷𝒓𝒐𝒅𝒖𝒄𝒕𝒊𝒐𝒏 𝑫𝒂𝒕𝒂 + 𝑪𝒐𝒏𝒔𝒖𝒎𝒑𝒕𝒊𝒐𝒏_𝑺𝒆𝒊𝒛𝒖𝒓𝒆)

𝑻𝒓𝒂𝒏𝒔𝒊𝒕_𝑺𝒆𝒊𝒛𝒖𝒓𝒆 + 𝑷𝒓𝒐𝒅𝒖𝒄𝒕𝒊𝒐𝒏 𝑫𝒂𝒕𝒂 + 𝑪𝒐𝒏𝒔𝒖𝒎𝒑𝒕𝒊𝒐𝒏_𝑺𝒆𝒊𝒛𝒖𝒓𝒆 (Eq. 3)

Using the data from the UNODC seizures database in this equation gives the following structure of countries:

Transit Countries Producer and Consumer Countries

Albania, Algeria, Angola, Armenia, Azerbaijan, The Bahamas, Bangladesh, Belarus, Belize, Benin, Bermuda, Bosnia and Herzegovina, Bulgaria, Burundi, Cambodia, Cape Verde, Chile, Costa Rica, Cote d'Ivoire, Cuba, Dominica, Dominican Republic, Ecuador, El Salvador, Ethiopia, The Gambia, Ghana, Georgia, Greece, Guatemala, Guinea, Guinea-Bissau, Haiti, Honduras, Hong Kong, India, Iran, Jamaica, Kazakhstan, Kenya, Kyrgyzstan, Laos, Latvia, Lebanon, Liberia, Lithuania, Macedonia, Madagascar, Mali, Mauritius, Mexico, Moldova, Morocco, Nepal, Nicaragua, Nigeria, Pakistan, Panama, Paraguay, Philippines, Romania, Saint Vincent's & Grenadines, Sao Tome and Príncipe, Senegal, Serbia, Suriname, Syria, Taiwan, Tajikistan, Tanzania, Thailand, Togo, Trinidad and Tobago, Turkey, Turkmenistan, Ukraine, United Arab Emirates, Uruguay, Uzbekistan, Venezuela, Zambia, Zimbabwe

Afghanistan, Andorra, Anguilla, Antigua and Barbuda, Argentina, Australia, Austria, Bahrain, Barbados, Belgium, Bhutan, Bolivia, Botswana, Brazil, Brunei, Burkina Faso, Canada, Cayman Islands, Cameroon, Central African Republic, Chad, China, Colombia, Comoros, Congo Republic of, Croatia, Cyprus, Czech Republic, Democratic Republic of the Congo, Denmark, Djibouti, Egypt, Equatorial Guinea, Eritrea, Estonia, Fiji, Finland, France, Gabon, Germany, Grenada, Guyana, Hungary, Iceland, Indonesia, Iraq, Ireland, Israel, Italy, Japan, Jordan, Kosovo, Kuwait, Lesotho, Liechtenstein, Libya, Luxembourg, Macau, Malawi, Malaysia, Maldives, Malta, Mauritania, Monaco, Mongolia, Montenegro, Montserrat, Mozambique, Myanmar, Namibia, Netherlands, New Zealand, Niger, North Korea, Norway, Oman, Papua New Guinea, Peru, Poland, Portugal, Puerto Rico, Qatar, Russia, Rwanda, Saint Kitts and Nevis, Saint Lucia, Samoa, Sao Tome and Príncipe, Saudi Arabia, Seychelles, Sierra Leone, Singapore, Slovakia, Slovenia, Solomon Islands, Somalia, South Africa, South Korea, South Sudan, Spain, Sri Lanka, Sudan, Swaziland, Sweden, Switzerland, Timor-Leste, Tunisia, Turks & Caicos Islands, Uganda, United Kingdom, United States of America, Vietnam, Yemen

Individual values for each country can be found in Appendix 2.

Table 1: A list of countries and their status and transit, producer or consumer, as determined by the T-Value model

Variables II & III: Ranked_CIA & Binary_CIA

As a robustness check, two other drug trafficking indicators are produced to provide a comparative measure based on data from the CIA Factbook. The CIA Factbook produces an annual description of all countries that are important to the global drug trade, one that offers the most in depth analysis of all countries in one place from the same point in time. There are two other sources that report on the state of narcotics trafficking in a consistent manner. The US State Department produces an annual International Narcotics Control Strategy Report that comments on the state of countries involved in the drug trade. However, each year the report focuses exclusively on countries that “directly affect the United States” (State Department, 2017). The second is the UNODC, which produces a number of in depth country reports. However, these are not comprehensive in scale of breadth, covering only a few countries. Additionally, these reports are often spread across decades, meaning comparisons may be misleading (UNODC, 2017b_{). The CIA}

Factbook produces a shallow contextual description of each country’s relation to the global drug trade. Whilst providing little analysis, the scope and breadth it provides makes it an ideal source for producing a descriptive measure of transit countries. The Factbook is considered highly reliable and produces academically rigorous data because of its application of US intelligence, though this means its methodology and sources are unverifiable (CIA, 2015). The data used comes from the CIA World Factbooks 2010 - 2015 so as to match the period of data used for the UNODC-derived T-value. The values determined at each level are averaged over this period to produce a single value. The measures are then rounded to the nearest integer.

Using the descriptions that are used in the Factbook I produced two measures:

Binary_CIA: A dummy variable that categorises a country as either a transit country “1” if mentioned by the CIA as such; or non-transit “0” if there is no mention or if it is considered a key producer or consumer. The variable is named binary as it can only produce two values, 1 or 0.

Ranked_CIA: A ranked variable will also be used, where the following categorical definitions will be employed to differentiate between the data. The following classifications provide the boundaries that define the ranked variable:

0 No Trafficking/Producer/Consumer

1 Minor Trafficker. "Minor; small amounts; limited" 2 Standard Trafficking. "Transit Nation"

3 Significant. "Significant; Increasingly Active" 4 Major Trafficker. "Major, "Important", "Key"