Geography, natural resources, and civil war : synthesizing ‘Greed and Grievances’ : case studies of Sierra Leone and the Sudan

(1)

1

Geography, Natural Resources, and Civil War: Synthesizing

‘Greed and Grievances’

Case Studies of Sierra Leone and The Sudan

T.O. Kropholler (Bsc) Master's Thesis

Master's Programme: International Relations Graduate School of Social Sciences

Student number: 6176216

Supervisor: Dr. Polly Pallister-Wilkins Second Reader: Dr. Ursula Daxecker Date: 24-05-2016

(2)

2

Abstract

The aim of this thesis is to get more in depth knowledge about how geography and natural resources relate to civil war. It also attempts to synthesize the narratives of ‘greed’ and ‘grievances’ which are important in the debate concerning causes of civil war. I found that (1) in each case the location of armed conflict and natural resource do not precisely overlap, but fighting over control of resource rich areas does take place in between the capital cities and the resource rich areas. (2) The

hypotheses of Le Billon concerning the relationship between the geography of natural resources, and the type of armed conflict are only partly correct. (3) When the characteristics of terrain and

‘grievance’ factors are included, the hypotheses of Le Billon increase in validity, thus beginning to synthesize the ‘Greed vs Grievance’ narratives.

These results were realized by performing two case studies (1) The Sierra Leone Civil War, (2) The Second Sudanese Civil War. In both of these case studies, I tested the hypotheses of Le Billon (2001) which guides most of the debate concerning geography, natural resources and civil war, by

overlapping diamond and oil deposits with the location of armed conflict. I backed these ‘mappings of natural resources and armed conflict’ up with a literature study on the course of each conflict.

(3)

3

1. Introduction

Civil war is one of the most complex phenomenon to study in social sciences, there are always numerous actors, causes, interests and feelings involved which are almost never clear to the observer, but are obscure. Recent revolutions and civil wars in North Africa and

Europe acknowledge the fact that civil wars are still contemporary, in fact if we look at the world today it is overwhelmed with civil war. Civil war is also one of the greatest destroyer of peoples’ lives. Not only the people who die directly in battle, but also the people who die in starvation, sickness and ethnic cleansing. Civil war also is one of the greatest causes of forced migration, which inevitably leads to more pain and misery with refugees. It is also one of the greatest destroyers of wealth for societies and those who manage to survive. All these factors make civil war an interesting and relevant area of research for political scientist. In recent years scholars have done a lot of research on the causes of civil wars. The assumption is ‘if we know the cause, we know the remedy’. These researches are often quantitative large N studies which manage to produce strong correlations between the presence of natural resources, and the onset of civil war in countries. These results juxtapose

conventional wisdom that civil wars are fought because of resentment among ethnic groups in countries. The debate that followed is called the ‘Greed vs Grievance’ debate, in which the greed camp is currently the winning narrative. This is largely because the ‘greed’ narrative is better at quantifying indicators. The ‘grievance’ camp however seems to be fighting its way back, which is something to be glad of, because in social sciences reality is never

dichotomous, black and white, but always a grey soup of synthesis.

Both however do already seem to agree on one thing: that natural resources and spatiality, or geography, matters. I therefore take this agreement between both narratives as a starting point for my thesis in an attempt to synthesize both narratives, which of course will

hopefully lead to further and better research in order to fully understand and eradicate civil war.

(5)

5

The research questions that will guide me on this endeavor are formulated as follows:

 Did the geographic location of diamonds and oil affect which type of armed conflict

took place in The Sierra Leone Civil War and The Second Sudanese Civil War?

 How did the geography of diamonds and oil affect the course of the Sierra Leone Civil

War and The Second Sudanese Civil War?

 To what extent is a synthesis of the ‘greed’ and ‘grievances’ narratives possible in the way The Sierra Leone Civil War and The Second Sudanese Civil War were conducted? In order to answer the first question I will test the following hypotheses:

Hypotheses:

H1: The more proximate a point resource to a country's capital, the more likely a coup

d'état or state control will take place.

H2: The more distant a point resource to a country's capital, the more likely secession will take place.

H3: The more proximate a diffuse resource to an country's capital, the more likely

rebellion or looting will take place.

H4: The more distant a diffuse resource to a country's capital, the more likely warlordism

will take place.

These hypotheses are derived from Le Billon’s 2001 article “The Political Ecology of War” which deals with the relation of the geography of natural resources and civil war, and takes up an important position in the academic debate.

I will test these hypotheses by (1) analyzing the location and concentration of diamonds, oil, and natural gas in relation to the location of actual fighting; (2) doing a literature study of the actual fighting: "who fights who, and to what objective?"; (3): Where the evidence does not concur with above mentioned hypotheses, I will provide possible explanations for this difference, which should result in more a more comprehensive understanding of armed conflict in relation to natural resources and geography.

(6)

6

This is the first chapter, the introduction. In the second chapter of this thesis I will present the methodology. I will (1) argue why I have chosen Sierra Leone and Nigeria as my case studies; I will (2) present the data with which I will analyze the case studies and why it is so compelling; (3) present the limitations of the methodological problems and possible limitations. In the third chapter of this paper I will present the theoretical framework. I will present the two main narratives: ‘greed’ and ‘grievance’ as the causes for civil war. In the second part of the theoretical framework I will present the theory behind the hypotheses of Le Billon. In the fourth chapter of this thesis, I will present the case studies. The first parts of each subsection are the descriptive parts, and then the analysis follows. The fifth chapter concludes and reviews.

(7)

7

2. Methodology

Case Selection

The Sierra Leone Civil War is a classic case of a ‘resource war’ the term ‘blood diamond’ was coined in the aftermath of this conflict. After and even during the war, scholar reopened the academic debate concerning geography, natural resources and civil war. The result of this is that there is a lot of published work available on the Sierra Leone Civil War, which is

advantageous to my research. Furthermore, the Sierra Leone Civil War is often depicted as purely a war fought over its diamond riches, which leaves room for the ‘grievance’ narrative to add explanatory power.

The Second Sudanese Civil War was one of the longest running civil wars, and one of the most lethal. It is often depicted as both a ‘resource war’, and an ‘ethnic’ war. It is also much publicized, and the war still resonates in South Sudan and in Darfur, which gives it an up-to-date literature.

Both cases are similar in the fact that they are both complex and well documented. They are also both former British Colonies, which limits the feasibility of any historical critiques concerning these civil wars. They also both have large Muslim and Christian populations, which decreases any religious claim on a cause for civil war. The most important difference between the two cases is state size. I have deliberately chosen two cases which differ in size, because theory concerning geography, natural resources, and civil war suggests that

distance, and therefore size does matter.

Data Selection

In order to map and overlap armed conflict and natural resources, I will use three PRIO datasets: (1) The PRIO Conflict Site dataset, which is a geo-referenced dataset on armed conflict (Halberg, 2012); (2) the PRIO Diamond Resources Dataset (Gilmore et al, 2005); and (3) the PRIO Petroleum Dataset (Lujala et al, 2007).

The Conflict Site dataset includes all armed conflicts from 1989 until 2008. It basically provides me with two things that I need: (1) The coordinates of the core of armed conflict each year; (2) The radius of those armed conflicts. It defines armed conflict as a conflict

(8)

8

between at a minimum of 25 battle deaths per year per given geo-referenced armed conflict. This might sound as a small number of battle deaths, and it is, but considering the thousands of battle deaths in both wars I do not expect any problems regarding validity of this

measurement.

The diamond dataset contains all the known diamond deposits in the world up to 2005. It provides me with three things I need: (1) Coordinates of the diamond deposits in Sierra Leone; (2) What type of diamond is found there; (3) when the deposit was discovered and first mined.

The petroleum dataset contains all the known oil and gas fields in the world up until 2007, both onshore and offshore. It provides me with two things I need: (1) Coordinates of the on-shore oilfields in Sudan; (2) When they were discovered, and when production started. In the case of Sierra Leone, I take the coordinates of the armed conflict sites and pinpoint them on a map. I do the same with the coordinates of the diamond deposits, and overlap them with the pinpoints of the armed conflict sites. In the case of Sudan I take the

coordinates of armed conflict sites, and the coordinates of the petroleum sites and overlap them as well.

I chose to analyze only these two natural resources: oil and diamonds, as they can generate significant amounts of money, and are relatively easily produced and/or looted, which is relevant in relation to civil war according to theory.

Methodological Problems & Limitations

An important limitation to this research is that the PRIO dataset on armed conflict still is not disaggregated enough, which gives a distorted image of the situation on the ground.

Furthermore, more information about the size and scope of the natural resources dataset is needed for the same reason as with the armed conflict dataset. Another important

limitation is that true motive of individuals and groups for some course of action are difficult to infer from a literature study, although it does give more insight than large N quantitative studies which have dominated the discourse in recent decade.

(9)

9

3. Theoretical Approach to Civil War

3.1 Causes of Civil War: Greed & Grievances

Even though civil wars are not new to mankind and the academic debate, the reason why civil wars occur is still heavily debated. Of recent years the two most important two narratives which deal with the causes of civil wars are known as the ‘greed vs grievance’ narratives. On the one hand, those who favor the ‘grievance’ narrative claim that identity related issues cause civil war. On the other hand, those who favor the ‘greed’ narrative claim that economic related issues cause civil wars. Economist Herschel Grossman speaks of a rebel industry in which “insurgents are indistinguishable from bandits or pirates” (Grossman, 1999: 269). More recent supporters of the greed narrative however, speak more in terms of ‘opportunity’ than of an industry of greed in civil war (Fearon & Laitin, 2003); (Collier & Hoeffler, 2004); (Collier, 2006); (Collier et al, 2008).

Perhaps the most adamant supporter of the ‘greed’ narrative is economist Paul Collier, who has produced several important publications on the matter. Together with Anke Hoeffler, he published the first research on the economic causes for the onset of civil war. They find that four variables are significant determinants of the onset of civil war: (1) The higher is the per capita income, the lower the risk of civil war; (2) increased natural resources increase the risk of war; (3) the larger the population, the greater the risk of war; (4) highly fractionalizes societies are have no higher conflict risk than highly homogeneous societies. Polarized societies do have a higher risk of conflict (Collier & Hoeffler, 1998).

The work of Collier and Hoeffler has spurred more research on the economic causes of civil war. Indeed Fearon and Laitin find that conditions that favor insurgency such as poverty, a large population, and instability, are better predictors of the onset of civil war than

indicators of ethnic and religious diversity, economic inequality, lack of democracy, and discrimination against minorities (Fearon & Laitin, 2003).

In 2004 Collier and Anke Hoeffler revise their publication of 1998, and build on the work of Fearon and Laitin. In their 2004 publication they propose a theory which juxtaposes

opportunities for rebellion against the costs, or constraints of rebellion. They find three key factors which influence the opportunity for rebellion: (1) the availability of finance; primary

(10)

10

commodity exports substantially increases conflict risks because of the opportunity of extortion these commodities provides; (2) the cost of rebellion; male secondary education enrollment, national income per capita, and GDP growth rate reduce conflict risk; (3) military advantage; a dispersed population increases the risk of conflict, and they find weaker

evidence that mountainous terrain might advantage rebels. They did not find any significant correlation between the onset of civil war and ‘grievance’ proxies such as: inequality, political rights, ethnic polarization, and religious polarization. Of the grievance proxies they measured, only ‘ethnic domination’ had a significant correlation with the onset of civil war. Furthermore, they find that conflict risk is proportional to a country’s population (Collier & Hoeffler, 2004).

In 2005 James Fearon counters the 2004 Collier and Hoeffler research by arguing that (1) even when using Collier and Hoeffler’s civil war codings and model specifications, the empirical association between primary commodity exports and civil war outbreak is neither strong nor robust; (2) oil exports predict higher civil war risk because high oil exports indicate a weaker state, not because oil provides better financing opportunities for rebels (Fearon, 2005).

Apparently, even though there appears to be some disagreement in how economic factors influence the onset of civil war, there seems little doubt that the ‘greed’ camp is winning the debate on causes for civil war by not finding any significant correlation between grievance indicators and the onset of civil war. The problem with greed as the sole cause for civil war is however (even if it is mauled in an ‘opportunity’ framework) that it, again, depicts rebels or insurgents as mere bandits or pirates. It also quite heavily presupposes a condition of complete ‘rationality’ on the side of the rebels, especially in the work of Collier.

Depicting actors in social sciences as completely rational actors, significantly narrows the ability of the observer to come to a complete description of reality. In the end, political science studies the interaction of humans, and mankind is simply not a completely rational being. This is what makes political science such a complex study. Because actors are not completely rational, ‘greed’ cannot be the only explanation for civil war. Grievances,

although difficult to measure, should not be overlooked in any research on such a primal act as war, and especially civil war. As with most debates in political science, there is hardly ever

(11)

11

one truth, and therefore a synthesis is needed in the ‘Greed vs Grievances debate’. Luckily, the critiques of ‘greed’, and research on ‘grievance’ are solid.

Grievance as a cause for civil war is not new but stems from a long history of research on the subject of revolution. In 1962 Davies argued that revolutions are motivated by an increasing gap between the unfulfilled personal aspirations on the one hand, and actual economic status on the other hand (Davies, 1962). Gurr acknowledged this and linked relative

deprivation theory to conflict by arguing that violence is a reaction to frustrations stemming from unfulfilled aspirations related to personal well-being (Gurr, 1970). Later this correlation between grievance and violence, was linked to ethnicity. First Horrowitz found, and later Gurr confirmed, that ethnic grievance contributed indirectly to collective violence through ethnic mobilization (Horrowitz, 1985); (Gurr, 2000).

Cederman, Weidman, and Gleditsch in 2011 further developed the ethnic grievance and conflict idea into a quantitative research that would bring the ‘grievance’ narrative back in the debate. By using geocoded data on ethnic groups’ settlement combined with spatial wealth estimates, they find that ethnic groups both above and below the country average in terms of per capita income are overrepresented in civil conflict. Their empirical analysis also shows that political inequality between regions influence civil conflict, based on access to central power (Cederman et al, 2011).

This landmark study on ethnic inequalities and violence shows that grievance is still relevant in the debate concerning the onset of civil war, and paves the way for further research in this area. In this thesis I will use their findings in order to get to a synthesis of the greed and grievances debate in the case of Sierra Leone and Sudan.

(12)

12

3.2 Feasibility of Civil War: Natural Resources and Geography

The truce in the ‘greed vs grievance’ debate leaves us with the question: if there is not a clear cause of civil war, why is it then that civil war is so omnipresent? This is where feasibility of civil war through natural resources and geography comes into play.

In an attempt to leave the ‘greed vs grievances’ debate, Collier and Hoeffler published in 2008 their most comprehensive study of civil war, although it is largely re-do of their earlier work. This time they conclude that the results of their earlier work still stands, but also introduce some new insights. In this article they present the ‘feasibility hypothesis’ which proposes that where rebellion is feasible it will occur. They conclude that motivation for rebellion is indeterminate, as motivation is “supplied by whatever agenda happens to be adopted by the first social entrepreneur to occupy the viable niche, or itself endogenous to the opportunities thereby opened for illegal income” (Collier et al, 2008:26).

Earlier, Ross has focused more on natural resources as an economic cause for civil war. He finds that resource wealth is indeed causally linked to civil conflict. Resource wealth appears not only to increase the probability of conflict, but Ross also finds evidence that resource wealth makes conflict last longer, and produce more casualties when it does occur. Ross also finds that not all types of natural resources influence conflict. Oil, nonfuel minerals, and illicit drugs do appear to influence conflict, while agricultural commodities are unrelated to civil war. More strikingly, Ross finds that the proposed causal mechanisms through which natural resources affect conflict –looting and grievance- do not appear to be valid. Lastly, Ross does find that resources play a different role in separatist conflicts than in non-separatist conflicts (Ross, 2004).

So, although Collier finds that rebellion has a precondition of feasibility, Ross finds that looting those natural resources does not affect civil war onset, which is seems inconsistent at the very least. Also interesting is that Ross finds that resources play a different role in different conflicts, for instance separatist conflicts and non-separatist conflicts. This relationship between natural resources and the type of conflict which takes place comes from Le Billon’s 2002 article “The Political Ecology of War”. In this article Le Billon describes the relation between the nature/geography of a resource and the type of conflict that

(13)

13

follows. He identifies distance and concentration and the mode of production of a resource as determinates for which type of armed conflict will take place (Le Billon, 2001).

Table 11

Table 1 presents relation between the nature/geography of a resource and type of conflict, as proposed by Le Billon. The proximate/distant category describes the distance of a resource to a countries center of power such as a national or regional capital. The point/diffuse category describes the concentration of a resource.

According to Le Billon, a point resource is more easily monopolized than a diffuse resource, but its desirability often makes it heavily contested, and often depends on international recognition for attracting investors. Therefore, point resources such as oil, are likely to ignite secession or coup d’état wars depending on the distance to the center of power. In the case of proximate point resources, it would be beneficial to control the whole of the land as resource rents would be the prize for controlling the state. In the case of distant point resources, the prize does not require to control the whole state, but only the part in which the resources are found. This bid for secession should be especially likely “when ‘outsiders’

1

Le Billon, P. (2002), “The Political Ecology of War: Natural Resources and Armed Conflicts”, Political

(14)

14

are perceived to extract ‘local’ resources without sharing the wealth, and when local populations are displaced by the extractive industry or suffer from its environmental costs” Le Billon, 2002: 574). Le Billon defines a ‘warlord’, as a “strongmen controlling an area through their ability to wage war and who do not obey higher (central) authorities.” (Le Billon, 2002: 575). Warlords are thus often dispersed, and heavily rely on the natural resources they find in their occupied territories.

Evidently, geographic features such as distance and concentration are important in Le Billon’s hypotheses. There are however geographic features which he does not include in his hypotheses. This in my view limits the applicability of his hypotheses. Other geographic features such as mountainous areas, forests, swamps, deserts should also matter in relation to natural resources and armed conflict. These geographic feature should not only affect the onset of armed conflict, but also the course armed conflict takes, and who wins.

(15)

15

4. Case Studies

In this chapter I will present the two case studies of Sierra Leone and Sudan. First, I will provide a descriptive section of the case studies. This is done by (1) presenting maps

containing conflict sites and diamond or oil deposits in Sierra Leone and Sudan. I constructed these maps using data from three different PRIO datasets. The PRIO Conflict Site dataset, which is a geo-referenced dataset on armed conflict (Halberg, 2012); the PRIO Diamond Resources Dataset (Gilmore et al, 2005); and the PRIO Petroleum Dataset (Lujala et al, 2007). (2) In addition to these maps, I provide an overview of both conflicts in which I describe the actors involved, what their objectives are, and the general conduct of both wars. This is done by using existing academic literature on both wars, and by using NGO reports. Then, I will present my analysis in which I will link the empirical findings back to the theory and research question presented earlier in this thesis.

4.1 Sierra Leone Civil War (1991 – 2002)

4.1.1 Mapping Conflict and Diamonds

The variously colored pinpoints in Map 1 show all the known diamond deposits as of 2005. The various colors represent the geological structure of the diamond occurrence. As shown in the legend of Map 1, these geological occurrences are classified as: Alluvial, Kimberlite, Marine, Unkown, and both Alluvial/Kimberlite. These classifications are important because they say something about the ‘lootability’, of the diamonds, or put differently: the ease of which the diamonds can be mined from the soil, and thus the ease of which the diamonds can be looted by the one who controls the territory.

Out of the three forms of diamonds which occur in Sierra Leone, the alluvial deposits are the most accessible and also the most common. These alluvial deposits can be mined by hand and requires digging pits in riverbeds and/or pan for diamonds like one would pan for gold. Consequently, apart from physical labor, the cost of mining this kind of diamonds is rather low. Kimberlite diamond deposits lie deeper in the ground and are therefore less easily accessible. These deposits are not mined by hand, but requires heavy machinery and is in

(16)

16

need of a more organized operation. Therefore, mining kimberlite deposits is also more costly than mining the alluvial form of diamonds (Silberfein, 2004).

The third form of diamonds in Sierra Leone, marine deposits, can be mined in three ways. (1) by dragging submerged rich diamonds sands onto shore; (2) by removing diamonds from 20m deep water through suction pipes; and (3) by deep sea marine vessels which deploy underwater excavators to extract the diamonds. The marine diamond deposits are as of 2005 still under exploration, and were not being produced during the civil war in Sierra Leone (Gilmore et al 2005). Therefore, the marine deposits could not be looted and do not seem to have played a role in the war.

As Map 1 also shows, the diamond deposits in Sierra Leone are predominantly found in the southern and especially eastern part of the country. There is only one diamond deposit located in the north of the country, in the town of Kamakwie. There is little information known about this deposit and does not seem to have played any substantial role in the civil war.

Even though the diamond deposits are located in the eastern part of the country and the capital city of Freetown lies in the most western part of the country, the diamond deposits are still relatively proximate to the capital. The most distant deposit is the Gori Hills alluvial diamond deposit located near the Guinean border to the north of Pendembu at

(17)

17

The numerical pinpoints in Map 1 show the core of the conflict zone in each given year. The conflict zone in 1991 had a radius of 100 kilometers and from 1992 until 2000 each conflict zone had a radius of 150 km (Hallberg, 2012). Because Sierra Leone is a relatively small country (radius of 170km), one can assume -taking the conflict radii into account- that fighting took place virtually in the whole of the country. Having a conflict zone with a radius of 150km in a country with a radius of 170km, means that 88.24% of the country is in effect a conflict zone. The majority of the fighting, however, took place in the core of each conflict zone. Therefore, we can see in Map 1 how the war started, and how the war developed in Sierra Leone.

As Map 1 shows, the war in Sierra Leone clearly started in the south-eastern part of the country near the border with Liberia in 1991. For as long as two years the core of the conflict zones remained within an area between the regional city of Kenema and the border of Liberia, at roughly 20km of each core and 250km of Freetown, the capital city of Sierra Leone . In 1994 there is a remarkable shift of roughly 120km towards the center of the

(18)

18

country, which will later be explained by providing additional information about the course of the war. For the remaining six years of the civil war in Sierra Leone, the core of the conflict zones stayed within roughly 30km of each other and 150km of Freetown.

Map 2 is the same as Map 1 except that the core of the conflict zones and diamond deposits are here shown as clusters, or groups. This shows that the diamond deposits are relatively diffused within the eastern region of Sierra Leone. Another remarkable thing is that the larger cluster of core conflict zones lies precisely between Freetown and the largest cluster of diamond deposits.

In map 3 the most important diamond deposits are highlighted. It shows that three out of the four important diamond deposits are located in the largest eastern cluster of diamond deposits. The most northern of these three is the alluvial Yengema diamond field. The other two are the Koidu diamond field and the Tongo diamond field, both fields contain alluvial and kimberlite diamond deposits. To the south near the Liberian border lies another important diamond field in the town of Zimmi which is rich in alluvial diamonds.

(19)

19

Map 4 is a zoomed-in version of Map 3 in which the two marine diamond deposits and the diamond field of Kamakwie in the north of Sierra Leone are excluded, but all the other diamond fields and conflict cores remain clearly mapped, and the most important diamond fields are again highlighted. More importantly, Map 4 also shows that the two clusters of conflict cores lie next to two forest reserves. The first cluster of conflict cores ranging from number 1 to 3 lie next to the Golo forests at the border of Liberia and which are outliers of the larger Lofa-Mano National Park in Liberia. The second cluster of conflict cores ranging from number 4 to number 10 lies next to the 85.73 square km Kangari Hills Forest Reserve.

(20)

(21)

21 4.1.2 The Course of the Conflict

On the 23rd of March 1991 the RUF under the leadership of Foday Sankoh, and supported by

rebels belonging to Charles Taylor’s National Patriotic Front of Liberia, launched their insurgency into Sierra Leone from their safe haven in Liberia. The RUF attacked areas in the eastern and southern provinces of Sierra Leone as part of a pincer movement, with the aim of coming together further in the interior (Silberfein, 2004). Initially the RUF stated that their aim is to overthrow the APC regime of President Momoh in Freetown. Although the RUF was numerically inferior to the Sierra Leone Army (SLA), they succeeded in controlling two thirds of the country by the end of 1991. Partly because the Momoh government was taken completely by surprise by the RUF attack (Davies, 2000), and partly because the Momoh government initially did not regard the RUF as a serious threat, which gave the RUF

momentum in their insurgency (Richards, 2003). In October 1991 the RUF managed to take control of the Kono diamond district. Reports came out that government forces were themselves digging for diamonds instead of maintaining a defense against the RUF.

Government forces who fought for their personal gain posed a problem for the SLA. These soldiers came to be known as ‘sobels’, soldiers by day rebels by night. These sobels partook in looting of villages and cooperating with the RUF in dividing up the spoils of war (Silberfein, 2004).

The neglect of the Momoh government in handling the RUF led to a coup by dissatified

young army officers on the 29th of march 1992, and Captain Valentine Strasser, who headed

the coup became the new leader of Sierra Leone (Davies, 2000). After this coup, the SLA was far better equipped than before, and started to gain ground on the RUF, which were driven back into the border region with Liberia. The situation for the RUF became so dire that they chose to abandon their heavy weapons and retreat into the Gola Forrest, and fight from the bush (Richards, 2003).

In December 1993, the government of Valentine Strasser announced a unilateral cease-fire, as the rebels were virtually routed from all affected war areas (Davies, 2000). The RUF by now became a “forest survivalist movement” (Richards, 2003: 15), and were able to reconsolidate their position thanks to the cease-fire (Davies, 2000). By 1994 the RUF regained the initiative, as they had secured their forest camps, and the SLA was not ably trained and equipped for jungle warfare. Moreover, the RUF had learned the geography of

(22)

22

the bush paths which interconnected with all of Sierra Leone’s rural areas. In 1995 the RUF established bush camps in all of Sierra Leone and attacked the SLA from those positions, and cut them off from their supply lines to the east. The RUF increasingly made use of hit-and-run tactics and raiding villages which were intended to spread fear among the civilians instead of holding terrain against the government of Sierra Leone (Richards, 2003).

The advance of the RUF seemed unstoppable, and the government of Strasser hired South African private military contractor Executive Outcomes (EO) to train the SLA in jungle warfare and to halt the RUF’s advance on Freetown (ibid). In March 1995, EO prevented the RUF from taking Freetown by battling the rebels thirty miles from the capital and driving them back into the bush (Davies, 2000).

In January 1996, two months before presidential elections Valentine Strasser is replaced by his deputy general Maada Bio in a coup. In the years leading up to the coup, the regime of Strasser became increasingly unable to battle the rebels, and members of the regime of Strasser were being accused of smuggling diamonds. In March 1996 Ahmed Tejan Kabba was declared winner of the Presidential election (Zack-Williams, 1999). President Kabba initiated talks with the RUF to end the war in Sierra Leone. These talks culminated in the Abidjan Peace Accord of 1996, which was signed by President Kabba and RUF leader Foday Sankoh. Unfortunately due to bad implementation the Abidjan Peace Accord failed and fighting resumed (Davies, 2000).

In March 1997 RUF leader Foday Sankoh was arrested and taken prisoner in Nigeria while brokering a deal trading arms for diamonds. In the absence of Foday Sankoh the RUF started to split in groups under the leadership of Sankohs former lieutenants (citation needed). Two months later, a military coup against President Kabba was staged, which resulted in the rule of the Armed Forces Revolutionary Council (AFRC) in cooperation with the RUF and the fleeing of Kabba to Guinea. Because the AFRC had no chance in getting international recognition, they had no other option than to cooperate with the RUF (ibid).

The AFRC junta was short-lived, under pressure of the Economic Community of West African States (ECOWAS) the AFRC signed the Conakry Accord which stipulated that all hostilities would be ceased and that Kabba would be reinstated as President of Sierra Leone. When the AFRC failed to meet the demands of the Conakry Accord, ECOMOG (the military apparatus of

(23)

23

ECOWAS) intervened and drove the AFRC and the RUF out of the capital of Freetown. In February 1998, Kabba would be reinstated as President of Sierra Leone. (ibid)

Nigeria subsequently handed over Foday Sankoh to the government in Sierra Leone where he was put to trial and sentenced to death. This angered the RUF officers and started their invasion of Freetow, which was eventually repulsed by ECOMOG. In july 1999 Sankoh and Kabba signed the Lomé Peace Accord. This accord would grant amnesty for Sankoh and all combatants in the Sierra Leone Civil War, and would make Sankoh Vice President of Sierra Leone (ibid).

The aftermath of the peace accord saw an overwhelming number of rebels returning from the bush into Freetown, which gave the Sankoh and the RUF the opportunity to try and get full control of the capital in May 2000. In the same month, British paratroopers were deployed and cleared Freetown of the RUF, which was eventually crushed. By the end of 2001, apart from renegade rebel groups in the bush, only UN peacekeeping forces and 8,000 troops of the new SLA were bearing arms, and the war was eventually officially ended in February 2002. (ibid) (see Table 1 for a summary of the events)

(24)

24

Table 22

2

Davies, V.A.B. (2000), “Sierra Leone: Ironic Tragedy”, Journal of African Economies, 9(3): 349-369.

March 1991 Onset of the rebel war. Sierra Leonean fighters of the RUF,

supported by rebels belonging to Charles Taylor’s National Patriotic Front of Liberia, enter Sierra Leone from Liberia and attack areas in the Eastern and Southern Provinces.

29 april 1992 Young, disgruntled, junior ranking soldiers protesting

neglect at the war front overthrow President Momoh’s APC, setting up the NPRC, headed by Captain Valentine Strasser.

December 1993 Declaration of unilateral cease-fire by NPRC. Rebels virtually

routed from all war-affected areas, reconsolidate.

May 1995 ‘Executive Outcomes’ mercenaries repulse a rebel advance

on Freetown

January 1996 Palace coup: Captain Strasser is replaced by his deputy,

General Maada Bio.

February 1996 General and presidential elections, subsequently won by

Tejan Kabbah’s Sierra Leone’s People Party.

March 1996 Inauguration of President Kabbah

November 1996 Signing of abortive Abidjan Peace Accord between

government and the RUF.

March 1997 Arrest of Foday Sankoh in Nigeria ostensibly for carrying

arms.

May 1997 Overthrow of Tejan Kabba by the army and establishment

of the Armed Forces Revolutionary Council in coalition with the RUF.

October 1997 Signing of abortive Conakry Peace Plan calling for

restoration of Tejan Kabbah in six months.

February 1998 Restoration of Tejan Kabba government by ECOMOG.

January 1999 Invasion of Freetown by rebels, repulsed within two to

three weeks by ECOMOG.

July 1999 Signing of Lomé peace accord.

April 2000 Replacement of ECOMOG by UN peacekeeping force.

May 2000 Over 500UN peacekeepers are taken hostage and the rebels

advance on Freetown. Twenty people killed in march on Foday Sankoh’s residence in Freetown. Sankoh escapes from residence and is recaptured a week later. Rebel advance repulsed with help of 1,000 British paratroopers.

June 2000 Withdrawal of British paratroopers.

July 2000 UN resolution calls for prohibition of diamond imports from

(25)

25 4.1.3 Analysis of the Conflict

Looking at the empirics on Sierra Leone presented earlier, I argue that Sierra Leone falls within the distant/diffuse category of Le Billon’s framework. Although Sierra Leone is a very small country, especially for African states, ‘distant’ here must be seen as relative to state size. As seen earlier the capital of Freetown, lies in the westernmost point of Sierra Leone, and the bulk of the diamond deposits lie in the eastern and southernmost points of the country. This distance between the capital and resources is problematic for Sierra Leone, as a states military capability is strongest at the capital and decreases over distance. Although the diamond deposits are concentrated in the eastern and southern parts of the country, they are diffused within those regions, taking up virtually all of the eastern and western parts of Sierra Leone.

Le Billon’s hypothesis expects a distant/diffused geography of a resource to correlate with a ‘warlord’ type of armed conflict. In reality however we do not see ‘warlordism’ taking place at the onset of the war, but fighting for control of the state, or coup d’état. This cannot be inferred from the mappings of conflict and diamonds, but this is supported by the evidence presented in the literature review on the course of the conflict in which the RUF under the leadership of Foday Sankoh intended to overthrow the government in Freetown. This

discrepancy is of course not in compliance with the hypotheses of Le Billon and shall be dealt with further on.

Another precondition for the acceptance of Le Billon hypotheses is that the actual fighting should take place at the location of the resource in question. The assumption that armed conflict in resource wars should take place at the location of the resource seems at first hand, not to be consistent with what actually happened in Sierra Leone. Looking at the maps

presented earlier in this chapter, the location of actual fighting and the location of the diamond deposits do not seem to overlap. The core of fighting in the initial three years seems to take place in the one spot in which there are no diamond deposits in Sierra Leone (see Map 1). This would imply that the ‘resource war’ in Sierra Leone is not actually fought in the area where the resources are located, and this would not be in compliance with Le Billon’s hypotheses. The literature review on the course of the conflict, however, shows that the RUF did control two thirds of the country, including the important diamond deposits in

(26)

26

the Kono district. This shows that although the location of the bulk of the fighting does not overlap with the location of the natural resources, rebel control of territory does overlap with the location of natural resources.

This distinction between location of actual fighting and control over territory seems spurious as one first need to fight over territory in order to control territory. In actuality this is need not always have to be the case. The western cluster of armed conflict cores highlighted in Map 2 also does not exactly overlap with any of the locations of diamond deposits to the east or south. The cluster does however, lie directly between Freetown in the west and the Diamonds in the east. The position of this cluster between the capital and the resources could mean that government forces were cut off from the diamond rich areas. This is also supported from the literature study on the course of the civil war. This shows that the majority of the fighting need not be in the exact same area of the natural resources in order to gain control over those resources.

Although initially the Sierra Leone Civil war was fought between the RUF and the

government over state control, we later see a situation of ‘warlordism’ developing in which each warlord fights over control of the diamond deposits. Warlords formed not only from the ranks of the SLA, the so called ‘sobels’, soldiers which both fought and cooperated with the RUF for the spoils of war, but the situation of ‘warlordism’ also developed because of fractionalization of the RUF in later stages of the civil war. Abiodun Alao (2007: 125) describes the situation as:

“While the leader of the only armed faction against the government, Foday Sankoh, remained the key warlord, the nature of the conflict was such that a number of key supporters of Sankoh became major actors on his behalf while he was incarcerated in Nigeria. Hence, an individual such as the late Sam Bockarie could be effectively described as a warlord. Even the members of the armed forces who later left the force to join with the rebel movement, such as the former leader Johnny Paul Koroma, represent, to a large extent, warlords.”

Although Le Billon’s hypothesis suggests that warlordism should take place at the onset of the conflict, the case of Sierra Leone shows that warlordism is a later development. I argue this is because of the terrain of Sierra Leone, and especially the terrain of the parts in which the diamonds are located. This part of Sierra Leone is characterized by mountainous, and dense forest terrain (as shown in Map 1, and Map 4). This kind of terrain makes warlordism

(27)

27

more feasible, even at short distances, because controlling mountainous and forest terrain is relatively easily controlled, which increases competition for control in those areas.

This shows that terrain had a significant impact on how the civil war in Sierra Leone developed. This is also true for the survival of the RUF. As reported in the literature study, the RUF was nearly demolished by government forces in 1993. The RUF could however retreat into the bush of eastern and southern Sierra Leone, regroup, regain strength, and attack the SLA from their strongholds in the bush. This further supports my point that terrain was instrumental in the way the war in Sierra Leone developed.

So terrain explains why ‘warlordism’ developed in the Sierra Leone Civil War, but it does not explain why the RUF fought with state control as their primary objective. This can only explained by other factors than the ‘greed’ factors which are often used to explain the war in Sierra Leone. Multiple scholars have depicted the RUF as mere bandits looking to gain

financially from looting of diamonds (Hirsch, 2001); (Vehnamaki, 2002); (Silberfein, 2004). But these scholars completely ignore the fact that the Sierra Leone central government neglected the peripheral regions for decades leading up to the Civil War (Richards, 2000). Whether RUF leader Foday Sankoh himself fought for personal gain, or for the neglected people of Sierra Leone does not really matter, because it was neglect through which he mobilized the support for his cause.

Several things are to be concluded from analyzing the Sierra Leone Civil War in relation to Le Billon’s hypotheses, geography and the ‘greed vs grievance’ debate. (1) Although the

location of armed conflict and location of diamonds do not completely overlap, we do see that fighting over control of diamond rich areas took place. (2) Le Billon’s hypothesis which predicted ‘warlordism’ at the onset of the war does not concur with the analysis in which fighting over state control was the initial objective of the war. (3) The conflict over state control can be explained by a ‘grievance’ narrative. (4) The development of warlords can be explained by a ‘greed’ narrative, and by the characteristics of terrain in diamond rich areas.

(28)

28

4.2 The Second Sudanese Civil War (1989 – 2005)

Although the Second Sudanese Civil War took place between 1983 and 2005, I only have geo-referenced armed conflict data of Sudan available between 1989 and 2008. Although 1983 is the actual year of the onset of the war, 1989 can be regarded as another important starting point, as in 1989 an important armistice was broken, and Field Marshall Omar al-Bashir seized power in Khartoum. Furthermore, the war was officially ended in 2005, but another civil war in western Sudan (War in Darfur) broke out in 2004, which is considered a spill-over from the Second Sudanese Civil War. Therefore, I chose to include those conflict years in the maps presented below.

At the time of writing this thesis Sudan is not the same country as it was during the years of the Second Sudanese Civil war. As of 2011 southern Sudan seceded from Sudan to form a new country; South Sudan. The remaining part of Sudan simply continued to be called Sudan. Whenever I use the term Sudan in this thesis, I refer to Sudan before 2011, thus including southern Sudan. If I use the term southern Sudan I refer to the geological southern region in Sudan. When I use the term South Sudan I refer to the country South Sudan as of 2011, if Sudan is mentioned in context with South Sudan I refer to Sudan the country as of 2011. In the maps presented below, I have included a dotted line which marks the present-day border between South Sudan and Sudan. This is done because all of the known oil fields as of 2007 in Sudan are located along this border.

4.2.1 Mapping Conflict and Oil

The variously colored pinpoints in Map 6 represent the three oilfields in Sudan which were discovered up to 2007 in Sudan. The core of the Melut oilfield lies in present day South Sudan and Sudan next to the Ethiopian border. The core of oilfield Muglad 1 lies precisely on the border of present day South Sudan and Sudan, and the core of the oilfield Muglad 2 lies in western Sudan on the border with provinces Southern Darfur and Southern Kordofan (both within present day Sudan). The Melut oilfield was discovered in 1982 and the production of this oilfield started in 1997. The Muglad 1 oilfield was discovered in 1980 an

(29)

29

production started in 1996; Muglad 2 was discovered in 1979 and it is unclear when the production of this oilfield started.

The numerical pinpoints in Map 6 show the core of the conflict zone in each given year. The conflict zones between 1989 and 1995 had radii ranging between 450 and 550 kilometers. In 1996 the radius of the conflict zone expanded to 850km, and further increased in 1997 and 1998 to 950 km. In 1999 the radius of the conflict zone decreased to 900km, and from 2000 until 2004 the conflict zone radii ranged between 750km in 2000 and 600km in 2004. It rapidly decreased in conflict years 2005 and 2006 to 200km and 350km, and subsequently expanded back up to 500km in 2007 and 2008 (Hallberg 2012).

Although the abovementioned radii are large distances, they must be regarded in relation with the size of Sudan, one of the largest countries in Africa. Sudan has an approximated radius of 1,000 km. Having a conflict zone radius of 950km, and a country radius of 1,000km would mean that 95% of the country is effectively a conflict zone. However, as in the case with Sierra Leone, again the majority of fighting took place in the core of the conflict zones, thus justifying this depiction of the conflict.

Map 6 clearly shows that fighting began in southern Sudan in 1989 and remained located within a radius of 200km in southern Sudan for 6 years. Conflict year 1994 and 1995 had

(30)

30

exactly the same core of conflict zones, and are therefore both marked by pinpoint number 7. In 1996 the core of the conflict zone shifted in north-eastern direction, marked by

pinpoint 8 which is hidden behind pinpoint 14 in this map. From 1996 to 2002 the core of the conflict zones remained within roughly 300km of each other in eastern Sudan. And as of 2003 the core of the conflict zone shifted in western direction towards the region Darfur, and stayed there in subsequent years up to 2008.

In Map 7 these cores of conflict zones are again represented as groups, and a clear divide in conflict area is seen. Furthermore, it is clear that the three grouped conflict zones surround the three grouped oil fields. The upper two conflict zone clusters in western and eastern Sudan lie directly in between the three oil fields and the capital of Khartoum. The core of the southernmost grouped conflict zone is roughly 860km away from Khartoum. In the case of the westernmost grouped conflict zones, this is roughly 320km; and the easternmost grouped conflict zones are distanced at roughly 680km from Khartoum. The center of the grouped oil fields is roughly 600km away from Khartoum.

(31)

31

Map 8 is a zoomed-in version of Map 6 and shows a clearer image of the different cores of the conflict zones.

Map 9 is a further zoomed-in version of Map 6, focused on southern Sudan. It still includes the Melut and Muglad 1 oilfields, as well as the first 7 cores of the conflict zones from 1989 until 1995. As mentioned earlier, the conflict zones of 1994 and 1995 have exactly the same core, and are therefore represented by pinpoint number 7.

(32)

32

Map 10 is also a zoomed-in version of Map 6, with a focus on the conflict zones in eastern Sudan. What is clear from this map is that in the years 2000 and 2002, the core of the conflict zones shifted across the border into Ethiopia. Furthermore, the core of conflict zone number 13 lies in the Nuba Mountains of Southern Kordofan, which as will later be clear played an important role in the war.

(33)

33

Lastly, Map 11 is again a zoomed-in version of Map 6, now with a focus on western Sudan. Included is the Muglad 2 oilfield, right on the border of Southern Darfur and Southern Kordofan. Pinpoint number 17 represents the core of the final conflict zone in the Second Sudanese Civil War in 2005. Pinpoints 18, 19, and 20 exclusively represent the cores of conflict zones in the War in Darfur.

(34)

34 4.2.2 The Course of the Conflict

In May 1983 President Ja’afar Numayrie announced without warning the Republican Order Number One, and declared Sudan an Islamic State under Sharia Law, including the

predominantly non-Islamic southern Sudan. This declaration effectively ended the Addis Ababa Agreement of 1972, which ended the First Sudanese Civil War and established the Southern Sudan Autonomous Region. With the ending of the Addis Ababa Agreement, The

Second Sudanese Civil War began(Collins, 2007).

On July 31st 1983, former Sudanese Army Colonel John Garang, founder and leader of the

months earlier established Sudanese People’s Liberation Movement/Army (SPLM/A), released the Manifesto, South Sudanese People’s Movement. In this manifest Garang announced that the civil war was indeed a revolution. However, the aim of the SPLM/A’s revolution was not to form a separate South Sudan, but to build a ‘ New Sudan’, which would be a federation with a central government tasked to fight racism and tribalism in Sudan (ibid).

What followed is six years of heavy fighting, resulting important SPLM/A victories in southern Sudan including villages in the strategic Upper Blue Nile River region near the Ethiopian border, and the Nuba mountains at the present-day South Sudan / Sudan border. Due to these important losses, the morale of the Sudanese Army was at an all-time low. The SPLM/A on the other hand was well organized, fought for a purpose, and seemed invincible (ibid).

On June 30th 1989, a group of army officers under the leadership of Omar al-Bashir

overthrew the civilian government in a coup d’état. The purpose of Bashir’s revolution was to spread Arab culture in Sudan, and impose a Salafist Islamist theology on all Sudanese. In the years to come, these juxtaposed positions of Bashir and Garang would further

deteriorate the situation in Sudan. One of the first steps of Bashir, however, was to negotiate a ceasefire with the SPLM/A (ibid).

Garang used this ceasefire to gain support for his cause internationally with succes. Although the SPLM/A under Garang was considered successful and well organized, some senior

(35)

35

tolerate dissent within his ranks. In 1991 this dissatisfaction resulted in the splitting up of the SPLM/A into the SPLA under the leadership of Garang and the SPLA-Nasir under the

leadership of Lam Akol and Riek Machar (Jok & Hutchinson, 1999).

The split between Garang and Akol/Machar was not only because of the dictatorial style of Garang, but also because there was disagreement about the objective of the revolution of the SPLM/A. Garang, as posed earlier, strived for an united ‘New Sudan’ including southern Sudan. Machar and Akol on the other hand wanted to fight for an independent South Sudan (ibid). What followed was a struggle for power in southern Sudan between the two SPLA factions until the reunion of the SPLM/A in 2002.

Important to note is that this power struggle was basically another civil war fought along a divide in ethnicity. Garang belonged to the Dinka ethnic group, and Machar to the Nuer ethnic group. Both leaders mobilized support for their cause from their respective ethnic groups, basically drawing the ‘race card’. What followed were rifts of ethnic cleansing on both sides (ibid).

In need of material and financial support, Machar brokered a deal with Bashir which enabled the SPLA-Nasir to continue its battle against Garang. Bashir on the other hand would benefit from further ‘infighting’ in southern Sudan, as it would weaken both SPLA factions. If their power struggle would be settled, Bashir hoped to either defeat the winner, or make peace. Either way, Bashir would succeed in his objective of controlling the whole of Sudan (ibid). Eventually, Machar’s SPLA-Nasir disintegrated because Machar’s claim to fight for an independent southern Sudan contradicted their dependence on the GoS for arms and money. The disintegration of the SPLA-Nasir led Machar to form the South Sudan

Independence Movement/Army (SSIM/A) in 1994, with which he continued to fight Garang for control of the South and work with the GoS simultaneously (Collins, 2007)

The SSIM/A however, again was not a success. By the end of 1995 nearly all of the leaders in southern Sudan had distanced themselves from Machar, which left him isolated in his struggle. On April 4th 1996 Machar capitulated to Bashir, again in return for arms and cash. Machar now was exclusively fighting Garang’s SPLM/A, which legitimized his dwindling authority with the Nuer people, and more or less defaulted on his secession ambition (ibid).

(36)

36

The SPLA on the other hand, was on the offensive and won battle over battle against the GoS. From 1994 to 1997 the SPLA retook towns and cities in Equatoria (southernmost part of Sudan), the Upper Blue Nile (border with Ethiopia), and the Nuba Mountains (north of the South Sudan border). During one battle in 1997 in Equatoria the SPLA inflicted heavy losses on the GoS, killing 8,000 soldiers and capturing 1,700 prisoners (ibid).In 1998 however, the GoS gained military and political control over the Western Upper Blue Nile region, including its oil fields (Jok & Hutchinson, 1999). The GoS continued to gain territory on the SPLA in 1999 and 2000 in the Nuba Mountains, which almost saw the SPLA completely driven out of that area if it were not for the cease-fire brokered by the US (Collins, 2007).

In the years leading up to 2002, Riek Machar became increasingly marginalized as an actor in the civil war. On January 6th 2002 he therefore signed the ‘Nairobi Declaration’ in which he acknowledged Garang as the supreme authority in southern Sudan, reunited his SSIM/A with the SPLA, and became vice president of the SPLM/A. This move of course also meant a truce between the Dinka and Nuer peoples (ibid).

In May 1999 the first barrels of oil pumped up from the fields in the Upper Blue Nile region were exported to China. These oilfields were operated by the Chinese National Petroleum Corporation, and the revenue generated from the oil exports was predominantly allocated to the defense budget of Sudan. By 2002 the defense budget had increased to $665mln, which was half of the governments total budget, and 5% of the national GDP (ibid). Naturally, the people of southern Sudan were very much opposed to the idea that foreign companies could exploit southern Sudan’s oilfields to the benefit of the central government of Sudan. The drilling of oil wells in southern Sudan by the Chinese and GoS, which resulted in the displacement of local people in southern Sudan, and therefore inflicted heavy fighting between the SPLA and the GoS. The issue of ‘who gets what’ from oil revenues was

subsequently heavily debated in the peace agreements between GoS and the SPLM in 2005 (ibid).

In October 2002 the SPLA recaptured the town of Torit in eastern Equatoria near the border with Uganda. The defeat of the GoS forces in Tori led to the start of the peace process between the SPLM/A and the GoS. In July 2003 Bashir and Garang met for the first time, and pledged their commitment to bringing peace to the country. Both leaders held true to their

(37)

37

words, as 18 months later the Comprehensive Peace Agreement (CPA) was signed between Garang and Sudan’s Vice President Taha on January 9th 2005, bringing an end to 22 years of civil war in Africa’s largest country (ibid). (see Table 2 for a summary of the events).

The Comprehensive Peace Agreement stipulated that after six years of the signing, the GoS and the SPLM/A would organize a referendum in which the southern Sudanese people were to decide to stay within Sudan, or to secede from Sudan. It also states that power must be divided between the ruling party of Sudan, and southern Sudan. The President of the Government of South Sudan, would also be First Vice President of Sudan. Furthermore the CPA stated that the wealth generated from Sudanese oilfields had to be divided. It stipulates that 50% of the net revenue from oilfields in South Sudan would flow to the Government of South Sudan, and 50% of the net revenue would go to the Government of Sudan (CPA, 2005).

1989 Military coup led by Field Marshall Omar Bashir.

SPLA in control of large areas of Equitoria, Bahr al Ghazal, and Upper Nile provinces. Government in control of cities of Juba, Wau, and Malakal.

An informal cease-fire broke down in October

1991 SPLA splits into Torit and Nasir groups

Groups, and SPLA start fighting each other.

1992 Abuja I – Nigeria peace conference to unite southern rebels

1993 SPLA-United formed,

1994 SPLA-United re-named SSIM/A, continues fighting SPLA.

1997 Fighting in the east begins

2002 SSIM/A reunites with SPLA

2003 New rebel group in Darfur, SLM/A is launched

Government of Sudan retaliates and Darfur crisis begins

2005 Comprehensive Peace Agreement is signed, paving the way

for a independence referendum in southern Sudan. Garang of the SPLA becomes first Vice President of Sudan. Table 23

3_{Raftopoulos, B. & Alexander, K. (2006), “Peace in the Balance: The Crisis in Sudan”, Cape Town, ZA: Institute}

(38)

38 4.2.3 Analysis of the Conflict

Looking at the empirics on Sudan presented earlier, I argue that Sudan falls within the distant/point category of Le Billon’s framework. As seen earlier Sudan’s capital Khartoum, lies in the north of Sudan, while the oilfields are located at roughly 700km south to

Khartoum, on the present-day border with South Sudan. This distance between the capital and resources is problematic for Sudan, as a states military capability is strongest at the capital and decreases over distance.

Le Billon predicts that distant point resources evoke secessionist wars, because in order to benefit from the ‘prize’ of the resource, one merely needs to control the part of the country in which the resource is located. In the case of Sudan however, we do not see a secessionist war concerning southern Sudan taking place. Although Sudan is often framed as resource war where the aim of the southern Sudanese was to secede from Sudan, it was at the onset of the Second Sudanese Civil War actually a war fought with the objective of state control. In the case of Sudan, I identify three different types of armed conflict / political outcomes: (1) war of state control; (2) warlordism; (3) secession.

Another precondition for the acceptance of Le Billon’s hypotheses is that the actual fighting should take place at the location of the resource in question. In the case of Sudan this seems not to be the case initially, as in the first 7 conflict years the cores of the armed conflict do not overlap. It is only from the 8th conflict year onwards that conflict and resource pinpoints start to overlap (see Map 6).

The spatial development of the civil war in Sudan seems to follow the same path as the war in Sierra Leone. It starts farthest away as possible from the capital of the central government, and then spreads out in the direction of the capital, and settles in an area between the

capital and the resources. Again, this is probably because the projected power of the central government is weaker as distance increases.

(39)

39

In the case of Sudan, the characteristics of terrain also played a significant role. Collins points out that:

“Although the government had the advantage of more men, more armor, and more firepower, SPAF and the PDF were at a distinct disadvantage having to conduct their campaigns in swamps, grasslands, bush, and deep forests, terrain alien and unknown to them.” (Collins, 2007: 1791).

This is consistent with what maps 9 and 10 show. Map 9 shows that fighting took place in the area of the Ez Zeraf Game Reserve, which mainly consists of swamps and deep forests. Map 10 shows that fighting also took place in the riverine area of the Blue Nile

The year 1991 is an important year within the Second South Sudanese Civil War as it saw the splitting up of the SPLM/A into two warring factions, and further complicating the war. The motivation of this split is heavily debated. Some claim that Riek Machar and John Garang disagreed over the objective of the war in which Garang would support his cause for a

unified ‘New Sudan’, and Machar wanted to pursue an objective of secession from Sudan instead. Others claim that Machar pursued personal political objectives, by which he needed to oust Garang from the SPLM/A. Whatever the actual motivation for the split of the SPLM/A was does not really matter as both warlords mobilized support for their respective camps on the basis of old ethnic grievance between the Dinka and Nuer peoples.

As the literature study on the course of the war shows, secession from Sudan only became an option when oil revenues from Chinese exploited oilfields in southern Sudan began to flow to the central government in Sudan, and local people had to be moved in order to drill for new oil wells. This is in accordance to what Le Billon (2002: 574) states concerning natural resources and the probability of secession:

“The likelihood of political secession increases when ‘outsiders’ are perceived to extract ‘local’ resources withouth sharing the wealth, and when local populations are displaced by the extractive industry or suffer from its environmental costs”(Le Billon 2002: 574)

This is supported by Cederman and his colleagues who say that:

“As opposed to objective conditions such as horizontal inequalities, grievances are intersubjective perceived phenomena” (Cederman et al 2011:481)

Although this shift from conflict over state control towards secession seems to be based in a ‘greed’ argument, namely an economic one, it is in actuality a ‘grievance’ based argument. It is not about the oil rents itself that flow from south Sudan to the central government. It is

(40)

40

the perception of identity, namely ‘outsider’ ‘insider’ in relation to economic inequality that shifted the conflict from ‘state control’ towards ‘secession’, and makes it a ‘grievance’ based motivation.

Several things are to be concluded from analyzing The Second Sudanese Civil War in relation to Le Billon’s hypotheses, geography and the ‘greed vs grievance’ debate. (1) As in the case of Sierra Leone, the location of armed conflict and location of oil do not completely overlap, I that the spatial development of the Second Sudanese War took a similar path, and that fighting over control of oil rich areas took place. (2) Le Billon’s hypothesis which predicted ‘secession at the onset of the war does not concur with the analysis in which fighting over state control was the initial objective of the war. (3) The subsequent ‘warlord’ and

secessionist conflict can explained by ‘grievance’ narratives. (4) As in the case of Sierra Leone, terrain played an important role, as it increased the relative capabilities of the SPLM/A

(41)

41

5. Conclusion

The aim of this thesis was to get more in depth knowledge about how geography and natural resources relate to civil war. It also attempted to synthesize the narratives of ‘greed’ and ‘grievances’ which are important in the debate concerning causes of civil war. These aims were realized by performing two case studies (1) Sierra Leone Civil War, (2) The Second Sudanese Civil War. In both of these case studies, I tested the hypotheses of Le Billon (2001) which guides most of the debate concerning geography, natural resources and civil war, by overlapping diamond and oil deposits with the location of armed conflict. I backed these ‘mappings of natural resources and armed conflict’ up with a literature study on the course of each conflict.

I found that (1) in each case the location of armed conflict and natural resource do not precisely overlap, but fighting over control of resource rich areas takes place in between the capital cities and the resource rich areas. (2) The hypotheses of Le Billon concerning the relationship between the geography of natural resources, and the type of armed conflict are only partly correct. (3) When characteristics of terrain and ‘grievance’ factors are included, the hypotheses of Le Billon increase in validity, thus beginning to synthesize the ‘Greed vs Grievance’ narratives.

An important limitation to this research is that the PRIO dataset on armed conflict still is not disaggregated enough, which gives a distorted image of the situation on the ground.

Furthermore, more information about the size and scope of the natural resources dataset is needed for the same reason as with the armed conflict dataset. Another important

limitation is that true motive of individuals and groups for some course of action are difficult to infer from a literature study, although it does give more insight than large N quantitative studies which have dominated the discourse in recent decade.

Further research can make use of the ACLED dataset, also from PRIO, which contains more disaggregated data on armed conflict, but was not used for practical reasons in this thesis. Also, a field study in countries currently involved in so called ‘resource wars’ might be better equipped to explain the immensely complex nature of civil wars.

(42)

42

6. Bibliography

Alao, A. (2007), “Natural Resources and Conflict in Africa: The Tragedy of Endowment”, Rochester, NY: University of Rochester Press.

Ballentine, K. & Sherman, J. (2003), “The Political Economy of Armed Conflict: Beyond Greed and Grievance”, Boulder: Lynne Rienner Publishers.

Cederman, L.E., Weidmann, N.B. & Gleditsch, K.S. (2011), “Horizontal Inequalities and Ethnonationalist Civil War: A Global Comparison”, The American Political Science Review, 105(3): 478-495.

Collier, P. & Hoeffler, A. (1998), “On economic causes of civil war”, Oxford Economic Papers, 50: 563-573.

Collier, P. & Hoeffler, A. (2004), “Greed and Grievance in Civil War”, Oxford Economic Papers, 56(4): 563-595.

Collier, P., Hoeffler, A. & Rohner, D. (2009), “Beyond greed and grievance: feasibility and civil war”, Oxford Economic Papers, 61(1): 1-27.

Collins, R. (2007), “Civil Wars in the Sudan”, History Compass, 5(6): 1778-1805.

The Comprehensive Peace Agreement Between The Government of The Republic of Sudan and The Sudan People’s Liberation Movement/Army (2005), opened for signature 9-01-2005,

entered into force 11-10-2007

Davies, J.C. (1962), “Toward a Theory of Revolution”, American Sociological Review, 6(1): 5-19.

Davies, V.A.B. (2000), “Sierra Leone: Ironic Tragedy”, Journal of African Economies, 9(3): 349-369.

Fearon, J.D. & Laitin, D.D. (2003), “Ethnicity, Insurgency, and Civil War”, The American

Political Science Review, 97(1): 75-90.

Fearon, J.D. (2005), “Primary Commodity Exports and Civil War”, Journal of Conflict

Resolution, 49(4): 483-507.

Gberie, L. (2005), “A Dirty War in West Africa: The RUF and the Destruction of Sierra Leone”, Bloomington, IN: Indiana University Press.

Gilmore, E., Gleditsch, N.P., Lujala, P. & Rød, J.K. (2005), “Conflict Diamonds: A New Dataset”,

Conflict Management and Peace Science, 22(3): 257-292.

Geography, natural resources, and civil war : synthesizing ‘Greed and Grievances’ : case studies of Sierra Leone and the Sudan