The incidence of violent conflict

Balancing the greed-grievance theorem

As armed violence has become increasingly intra-state in nature, a focus has been placed on attempting to understand the sort of environment that can heighten the risk of violent conflict outbreak within state borders.

Much of the traditional literature on armed violence had been focused on the conflict-inducing role of “relative deprivation” (Gurr 1970) and the collective mobilization of groups as a function of political repression (Tilly 1978). But more recently, a predominantly quantitative empirical literature guided by rational choice theory emerged to account for the role that economic opportunities play in influencing the incidence of violent conflict (Grossman 1995; Collier & Hoeffler 1998; De Soysa 2000; Ross 2004).

The seminal papers by Collier & Hoeffler (1998, 2000, 2004) sparked a contentious debate by proposing a parsimonious model of civil war onset that conceptually juxtaposed “two contrasting motivations for armed rebellion” (Collier & Hoeffler 2000:1).

Whilst grievance alludes to the idea that civil war is the result of societal discontent and the desire to redress structural injustices through collective violent action, greed relates to the opportunities for economic predation derived from initiating and participating in armed insurgencies.

Based on data from 1960 to 1999, the Collier & Hoeffler (CH) model employed statistical analysis to quantify the relative explanatory power of these two underlying causal categories. Their econometric results suggested that the greed model outperformed the grievance model in determining the risk of civil war incidence.

Their empirical interpretations essentially downgraded the explanatory validity of the grievance dimension and advocated a shift in the pendulum of causality towards the greed end of the spectrum. As a result, a reductionist approach to civil war onset was proposed — asserting that the financial viability of engaging in armed violence seems to be the predominant determinant of cross-country civil war outbreak (Collier 1999, Collier & Hoeffler 2004; Collier et al. 2009).

Criticisms of this deduction have however arisen. Whilst some have found that grievances arguably have greater explanatory power (Reynal-Querol 2002; Regan & Norton 2005) — others have questioned the methodological validity and interpretation of the statistical proxies used by CH (Nathan 2005; Keen 2008).

Subsequent academic work has stressed the need to apply a more integrative political economy approach to the study of armed violence. The argument is that a more balanced and multifaceted analysis on the causes of violent conflict onset would be beneficial — notably so as to tailor policy and operational responses more fittingly (Ballentine & Sherman 2003; Arnson & Zartman 2005).

The premise of this analysis is to argue that although useful in simplifying the complexity of civil war onset, the greed-grievance theorem is limited by its capacity to account for the intertwined and multifaceted dynamics of violent conflict more generally.

This analysis will attempt to highlight the interpretative biases that have led to the overt endorsement (particularly by the policy community) of the greed dimension. It will attempt to demonstrate how historically-informed qualitative evaluations of conflict drivers can add explanatory validity to the grievance dimension.

This analysis will also seek to shed light on the advantages of adopting a political economy approach to violent conflict and suggest that the (arguably) over-used Good Governance term should be better defined so as to be captured in the equations of causality.

Most importantly, this analysis will employ the broader and more encompassing term of violent conflict to reflect the reality that armed violence has become increasingly asymmetrical. In other words, violent conflict need not necessarily equate only to civil wars opposing identified (i.e. state vs non-state) warring parties. It can also include one-sided violence against civilians, clashes between communal militias and spurious attacks by armed groups against state security forces and public infrastructure, for example.

The reason for this is that there has been a growing trend in the number of low-intensity armed conflicts within countries. When taking the Uppsala Conflict Data Program’s (UCDP) taxonomy of intensity — one that relates to deadliness — the number of armed conflicts with more than 1,000 deaths (referred to as war by UCDP) has in recent decades been trumped by the number of conflicts with less than 1,000 deaths (referred to as minor conflict by UCDP).

If we are to take the UCDP data on armed conflicts as indicative, it sheds light on a new normal in the study of violent conflict. It ostensibly demonstrates that armed violence might be less deadly, but not necessarily less frequent.

For this reason, it is important to depart from the analytical framework restricted to larger-scale armed conflicts like civil wars and towards one that is inclusive of the changing nature of violent conflict. One example of this is the growing number of politically motivated (non-state) armed insurgencies that are spreading across different subregions of Africa and the Middle-East.

Greed versus Grievance

The contemporary literature on armed violence has been dominated by an academic debate over the explanatory salience of two theoretical causal categories: greed and grievance.

Collier & Hoeffler (1999) broadly interpret greed as encompassing loot-seeking behaviour — where the propensity for individuals to participate in armed violence is largely a function of the availability of lootable economic resources. Grievances are broadly interpreted as justice-seeking — whereby collective political violence is driven by a collective endeavour to redress prevailing structural (and sometimes even institutionalized) injustices.

The statistical results emanating from the CH model led the authors to propose a paradigm shift in the analytical approach to the study of civil war onset. Collier & Hoeffler broadly posit that the combination of abundantly lootable primary commodities, low education levels and a high proportion of young unemployed men in declining economies significantly increases the risk of civil war outbreak.

Their econometric results suggest that the traditional consideration of violent conflict as a reaction to socio-economic and political grievances takes the backseat when compared to the assertion that armed violence is more likely to be due to the opportunities for economic plunder by would-be violent groupings.

Indeed, whilst violent conflict has historically been conceptualized as engendering human suffering, infrastructural destruction and economic downturns, it has been proposed that armed violence can also create an “alternative system of profit, power and even protection” (Keen 1998:11) — to the extent that violent conflict can motivate “conflict entrepreneurs” to exploit opportunities for self-enrichment. In other words, some groups might seek to benefit from engaging in armed violence. They may even have an economic interest in the “initiation, perpetuation and renewal of conflict”(Collier 1999:14).

This paradigm shift in the causal interpretation of violent conflict outbreak can arguably be attributed to the employment of economic theory and quantitative methods of analysis. Rational choice theory has been utilised to model “rebel behaviour” such that armed insurgencies are perceived to be more viable under an environment of tangible economic incentives to participate in them.

Overcoming the problem of collective action (Olson 1965) has constituted a fundamental question for conflict analysts — and although economists have acknowledged that society-wide grievances may be essential for rebel cohesion, it arguably suffers from the free-rider problem that dampens the prospects for coordinated violent mobilization.

Because grievance-relief through armed insurgency can be considered a public good, when the rewards to participation are based on mere promises to redress injustices, potential recruits may have insufficient economic incentives to bear the risks and costs associated with engaging in violence — especially when these can be borne by others.

As a result, when the monetary benefits to participating in armed rebellion are not confined to those who participate in it, collective mobilization is likely to be adversely affected. This potentially limits the size and military capacity of grievance-motivated rebel movements — especially when potential recruits are incapable of identifying time-inconsistency problems associated with the economic viability of joining and participating in an insurgency (Collier 1999).

Consequently, although social capital in the form of shared grievances may have the potential to stimulate collective action for a common goal, the absence of resources to finance grievance-motivated rebellions will likely undermine their capacity to be effective in their endeavours, regardless of how strong their political or ideological motivattions might be.

If addressing grievances can only be achieved through engaging in armed violence, it is posited that only insurgent groups with access to economic resources are capable of overcoming the problem of collective action. CH have therefore proposed that the conditions that make rebellions financially viable are ultimately more likely to happen.

As argued by Collier (1999), the grievances emanating from socio-economic and political discontent might merely serve the functional purpose of a self-promoting narrative for the cohesion and recruitment prospects of armed insurgencies. Ultimately, only when the benefits to insurgency participation can be monetized through the provision of selective fiduciary payments may armed rebellions become militarily viable.

The findings from the CH model are indicative of the idea that armed insurgencies may be motivated primarily by economic agendas and that the incidence of civil war outbreak is heightened when rebellions are financially feasible.

With the proposal that commercialized natural resources such as alluvial diamonds, timber and narcotics have played a fundamental role in contemporary civil wars in Angola, Cambodia and Colombia, many of the recent policy and operational implementations have been orientated towards undermining the income-generating opportunities of rebel movements, one example being the curtailment of trade in so-called conflict goods.

The United Nations Security Council (UNSC), for example, implemented trade embargos on the exports of diamonds in Sierra Leone and Liberia in an attempt to dismantle the “conflict diamond” industry. Such actions, as argued by Ballentine & Nitzschke (2003), have been arguably guided by the endorsement of the greed thesis by the international policy community.

Berdal (2005) asserts that this appraisal is due to the policy appeal of viewing civil wars and violent conflict more generally as greed-driven — especially as it deals with quantifiable economic assets like natural resources — as opposed to the more elusively quantifiable and subjective sources of grievances.

Research on how jihadi groups operating in the West African Sahel are tapping on illicit trade markets and engaging in kidnapping activities to sustain themselves financially is a reflection of this. Some of the operational responses to drug trafficking in the West African Sahel have arguably been primarily guided by the intent to curtail the financial resourcefulness of jihadi groups.

But responses to these jihadi armed insurgencies cannot be viewed solely through the economic prism. In the case of jihadi groups, the motivations for engaging in armed violence cannot be restricted to opportunities for self-enrichment. This is especially true when considering that a large share jihadi groups’ activities are not cost-effective. In other words, they are not necessarily or even primarily motivated by generating financial profits but possibly more so by (elusively quantifiable) political and ideological objectives.

This sheds light on the fact that there has been an ostensibly imbalanced policy focus on the economic dimension of violent conflict onset — with a general disregard to the role played by other socio-economic and political factors that equally intervene in the dynamics of causality — but that are usually more challenging to include in econometric models.

Violent conflict in Africa

Violent conflicts in the African continent have often been presented as case-study examples to support the relevance of the greed dimension in accounting for the risks associated with violent conflict onset.

Often characterized by low-income levels, political instability and abundance of natural resources, many African countries seem to fit broadly into the greed paradigm of violent conflict incidence. The endorsement of the greed thesis must however be scrutinized — as it derives from econometric analysis that by defaultmeasure what can be quantified.

But like with all parsimonious models that rely on a small number of variables that can generate statistical predictive power, their omission of things that affect the incidence of violent conflict but that cannot be measured as quantitative data raises some questions for the so-called external validty of such models. The fundamental question rests on whether changing the nature, locality and number of regressors in a model generates consistent outputs.

It is important to note that the statistical significance of the greed model does not fully eliminate the plausibility of grievance-driven effects on violent conflict onset. Rather, ithighlights the need to remain cautious about the appealing — but possibly misleading— interpretations of the drivers of violent conflict in the African continent.

Prior to — and after — the era of colonization, armed conflicts have arguably been a recurrent phenomenon in Africa, albeit to different degrees across countries. About 40% of Sub- Saharan Africa is said to have experienced at least one period of civil war between 1960–1999 (Elbadawi & Sambanis 2000).

Since 1945, the number of armed conflicts in the African continent have been trending upwards, increasing at a base rate far superior to other regions. Together with Asia, it is the region where the large majority of worldwide contemporary armed conflicts have been concentrated.

In the last 20 years, on average, close to 37% of armed conflicts have been located in the African continent, compared to 36% in Asia. Since 2014, the number of armed conflicts in Africa has consistently outnumbered those in other regions, with 2019 marking a high in Africa’s percentage share of global armed conflicts, at 46%.

CH (2002) offer an analytical framework for understanding the economic factors that are believed to influence the incidence of civil war onset in Africa.

They find that average per capita income was $2,000 for Africa compared to $3,625 in other developing countries whilst the average annual growth rate of gross domestic product (GDP) was at 0.5% for Africa relative to 2% for the other regional cohorts from 1965–1999.

The significantly lower average per capita income and rate of macroeconomic growth in Africa are used as proxies of the low costs of organizing armed rebellions: they relate to the bleak income-generating prospects of engaging in other (potentially lesser) economically-rewarding activities.

CH (2002) show that the mean ratio of primary commodity exports to GDP was slightly lower in Africa (17%) than in other developing regions (19%). Although low in relative terms, Elbadawi & Sambanis (2000) argue that since the relation between natural resources and conflict incidence is quadratic, what matters is the dispersion rather than the mean measure of natural resource dependency.

As such, they find that the standard deviation of African countries’ resource dependence is 46% smaller than non-African countries. This is an indication that more African countries are closer to the peak of natural resource dependence (2000:253) — and are therefore at an ostensibly greater risk of violent conflict in an environment of political and societal acrimony.

Whilst we can interpret Africa’s resource dependence through the belief that it is a good proxy for the availability of lootable commodities, it can also be viewed through the prism of the “resource curse” theorem. In other words, grievances can be generated as a result of the unequal distribution of resource benefits by public authorities and the economic distortions that emanate from the rent-seeking behaviour induced by the high economic dependence on natural resource extraction (Sachs & Warner 2001). However, modelling rent-seeking behaviour at all levels of government is difficult, making it unlikely to be included in the equations of econometric models.

Other proposed indicators of civil war onset are the levels of “ethnic fractionalization and dominance” (Collier & Hoeffler 2002). Africa was found to have the highest degree of ethnic fractionalization (61%) and a “relatively low” degree of ethnic dominance at 40% — lower than the averaged estimate for other developing regions, at 57% (Collier & Hoeffler 2002).

A high degree of ethnic fractionalization is indicative of the possibility that it constitutes a hurdle for collective mobilization and thus a hindering factor for rebel coordination. Inversely, ethnic dominance is said to increase the probability of collective political mobilization along ethnic lines — such that the relatively low levels of ethnic dominance in Africa should (in theory) temper the risks of politically-motivated ethnic marginalization.

Collier & Hoeffler (2002) find that Africa had a “slightly higher incidence of conflict relative to other developing regions”. However, whether this higher risk is attributable to greed or grievances is arguably subject to the way in which the statistical proxy variables included in their model are defined and measured.

CH find that the relationships found globally in their model “apply equally to Africa” (2002:20). This leads them to the proposal that Africa’s economic environment matters more in determining the continent’s proneness to violent conflict than its socio-demographic predicaments.

Balancing Greed and Grievances

If we are to follow CH’s interpretation, an interim assessment would be that the variables that are claimed to be related to the economic opportunities for rebellion have a greater statistical explanatory power than the grievance proxies in determining civil war outbreak in Africa.

A crucial caveat however emanates from the fact that this is partially — or at least arguably — the result of relying on statistical analysis to come to such conclusions. CH’s regressions provide probabilistic causal estimates using quantitative data that is limited by its capacity to fully account for all of the (not necessarily quantifiable) transmission mechanisms that link cause with effect in the realm of violent conflict onset.

The CH model is premised on the statistically significant effect of measurable greed proxies against the elusively quantifiable grievances proxies. Statistical biases can therefore emanate from the possibility that grievance-related variables are simply harder to measure. This can translate into the possibility that greed variables have more statistical explanatory power simply because they are easier to quantify.

When acknowledging the shortfalls of econometric models, it is important to take the validity of statistical analysis with caution. The onset (and duration) of violent conflict is arguably influenced by a constellation of evolving triggers and drivers, opening up the idea that econometric models can only tell us so much.

This highlights the need to adopt a more balanced approach to studying the relative weight of causal factors. By complementing quantitative research with historically-informed qualitative analysis — researchers can develop a more representative framework of analysis for evaluating the impact of the sort of variables that are oftentimes missing from econometric models on violent conflict.

The case of Sierra Leone

The conflict in Sierra Leone has become a paradigmatic case of “Africa’s diamond wars”. It has recurrently been used to justify the relevance of economic agendas in driving violent conflict — and how armed rebellions are motivated and sustainted by the ability to plunder profitable economic resources. Charles Cater (2003) explores the case of Sierra Leone and warns against the potentially misleading assumptions and excessive reliance on the “one-dimensional prism of economic opportunism”.

Qualitative analysis can shed light on the equally important role played by grievances in motivating rebellions. In Sierra Leone, government mismanagement is said to have led to the collapse of state institutions and the deterioration of public infrastructure — accompanied by lower economic growth and higher unemployment (Gberie 2005).

This created the sort of dire circumstances that arguably motivated the Revolutionary United Front (RUF) to take up arms against the government. The RUF was allegedly motivated by its resentment towards the political regime. Indeed, rebel leader Foday Sankoh expressed the goal of overthrowing “Sierra Leone’s corrupt rulers and liberate the country’s derelict peasantry and dispossessed” (quoted in Gberie 2005:6).

Sierra Leone somewhat illustrates a case in point where the polarized (as opposed to integrated) greed-grievance approach to causality has limitations. It is arguably a conflict where rebellion was motivated by grievances — notably “anger at lack of good government and educational opportunities” (Keen 2000:35) — but sustained by the economic profits provided by the exploitation of lootable diamonds.

The case of Sierra Leone highlights why it might be warranted to revise the greed-grievance theorem in favour of a more balanced or integrative approach to conflict causality. It also highlights the necessity to focus on the role played by the quality of governance in influencing the incidence of armed violence.

An important limitation with the CH model is related to how its rational choice paradigm places too much focus on rebel-centric approaches to civil war onset — without sufficient attention to how government (in)action can also play a significant role in triggering large scale social unrest, armed insurgencies or military coups.

Repressive marginalization

The interpretation of the CH model is premised on the idea that rebellions are the result of the attraction to opportunities for economic predation. This has generated a one-sided approach whereby conflict is viewed primarily through the analytical lens of rebel economic self-interest.

The case of the Sudans

The case of the Sudans however highlights that armed insurgencies can be the result of a reaction to governments oppressing and marginalizing their denizens. Ylönen argues that the principal motivation for insurgencies in pre-secession Sudan is a consequence of Khartoum’s oppressive policies that “resulted in culturally and regionally imposed political marginalization” (2008:128). He argues that pre-secession Sudan reflects a pattern across many African countries, ones that are primarily characterized by patronage politics — with highly concentrated political, economic and military power in the hands of factional groups.

As such, governmental predation over economic resources and politically-motivated distribution of their rents can shed some explanatory power on the historically recurrent armed insurgencies in the southern regions of Sudan, including modern-day South Sudan. The case of the Sudans sheda light on the interrelated dynamic between greed and grievance — as armed rebellions can be formed out of the grievances generated by the greedy behaviour of governmental elites — and sustained if and when sufficient economic resources are available to finance the operational and human costs of engaging in armed violence.

Most importantly, the history of the Sudans highlights the important role played by inequality in triggering violent conflict, an indicator that does not seem to have rekevance in CH’s econometric results — which suggest that economic inequality has no statistically significant effect on civil war outbreak (see Stewart 2008).

The political economy of violent conflict

The criticism that emerges from the revisionist armed violence literature is that the quantitative research that has supported the greed thesis is limited by its capacity to explain the causal process by which civil wars and violent conflicts more generally are triggered and sustained.

A primary concern is that statistical analysis provides probabilistic estimates of civil war incidence using data that arguably biases the validity of greed variables — given the difficulty to quantify grievance proxies.

Most importantly, statistical significance does not imply causal significance. In other words, econometric regressions using cross-country data do not necessarily capture (in full) the procedural causal mechanisms that are required to gain a sound understanding of violent conflict incidence — one that is necessary to program and implement tailored policy and operational responses accordingly.

As a result, what emanates from the literature on the political economy of violent conflict is an evident rebalancing towards the importance of grievances — and an underlying reverse-causality interaction between greed and grievances. The primary idea is that social capital in the form of, for example, belonging to a particular community or “ethnic” group with shared grievances can make armed insurgencies viable — despite the absence of economic opportunities for predation, as examined in the cases explored by Ballentine & Sherman (2003).

In Kosovo, it has been suggested that the main trigger of collective political violence resulted from the systematic exclusion of ethnic minorities from political power. In Nepal, despite having no access to sources of financing, the Maoist insurgency is claimed to have emerged from the grievances resulting from “extreme levels of landlessness” and the “burdens of the bonded labour system upon the lower castes” (Ballentine & Sherman 2003:261). These cases shed light on the idea that grievances deriving from deliberate exclusion and repression can sometimes suffice in generating collective violent mobilization.

The underlying caveat however relates to the idea that under given instances, attributing causality to either the greed or grievance dimension can be limiting — given that we are dealing with complex anthropogenic phenomena that ultimately require an integrated framework of analysis.

An integrative approach to studying conflict onset can therefore derive its utility from the acceptance that multiple causal factors may be at work in initiating violent conflict — especially those that cannot necessarily be quantified.

In Bougainville, the grievances emanating from the inequitable share of the profits from the Panguna copper mine is said to have been a prominent factor in the formation of the Bougainville Revolutionary Army (Ballentine & Nitschke 2003b). This illustrates how the politics of inequitable distribution of economic resources by a greedy elite can generate the societal grievances that trigger violent conflict — a phenomenon that also serves to partially understand the Aceh conflict in Indonesia or the Boko Haram insurgency in northern Nigeria.

Charles Cater argues that despite the relative (and policy appealing) value of explaining civil war onset through the dichotomized greed-grievance framework, a revamped political economy approach offers a more realistic picture of contemporary armed conflicts by examining the “interrelationship between economic and political causes” (2003:20).

He uses the cases of Sierra Leone and the Democratic Republic of Congo (DRC) to warn against the excessive reliance on reductionist economic accounts and argues that they cannot be “adequately explained by the uniform logic of resource predation” (2005:29).

He asserts that the primary motive for the formation of rebel militias in Sierra Leone and the DRC was the result of the grievances generated from decades of misrule and corruption by a parasitic state elite as well as by the nepotistic exploitation of natural resources. The rebel avidity for the expropriation of diamonds is argued to be a consequence of the decay of the state apparatus and the necessity to finance an armed rebellion to overthrow governments deemed as irresponsive to the needs of its citizenry.

The role of “Good Governance”

The CH model popularized a framework for analysing conflicts through the perspective of “agency” — where the economic motivations for armed rebellion by non-state agents can heighten the risk of violent conflict onset.

However, a crucial criticism that emerges from this framework is the absence of an “institutional component” (De Soysa 2002:399) — one characterized by the role played by a governing institutional structure in the form of a sovereign state and its subnational public authorities.

Good governance is often presented as a conflict-averting phenomenon that can potentially explain why civil war erupted in Angola but not in Botswana. Inversely, poor governance is often seen as playing a conflict-inducing role by generating the sort of structural grievances that motivate repressed groups to take up arms — presumably when all other alternatives have been exhausted.

In Sri Lanka, for example, the repressive and discriminative “Sinhala- only” policies of the central government arguably triggered the political radicalisation of the Liberation Tigers of Tamil Eelam (LTTE) — which capitalized on its diaspora community to finance its militant and “ethno-separatist” agenda.

The underlying idea is that the structural environment created by the actions of the governing authorities matter in the understanding of violent conflict onset. The oft-cited characterization of countries like Somalia, Yemen or Libya as “fragile states” tend to make reference to how governing authorities that are unable or unwilling to protect their citizens from poverty and insecurity usually create the sort of fertile environment for violent conflict to erupt — especially in the absence of any legal or peaceful mechanisms to contest the legitimacy of governing authorities. They also shed light on how repressive states can be the principal perpetrators of grievances — capable of sparking armed insurgencies through the radicalisation of politically or economically disenfranchised groups.

Broadly defined, good governance can refer to the ability of governing authorities to deliver public goods in an equitable manner, to the extent that citizens enjoy a degree of political freedom, security and economic well-being. However, good governance is often utilized as a catch-all term that suffers from ambiguity, making it difficult to tangibly implement policies that align with the multiple components of so-called good governance.

Consequently, while recurrently mentioned as an end-goal, good governance often suffers from being so poorly defined that it is close to impossible to quantify. In a sense, this makes it more difficult to not only implement initiatives that seek to improve governance — it also makes it harder to evaluate whether any form of good governance is having positive externalities.

A matter of research methods

Although heuristic, the quantitative studies that have awarded greater relevance to the greed dimension should be tempered by the fact that the risk of violent conflict onset cannot be confined to one causal category when alternative conflict triggers can also be identified — even if seldom quantified.

Because policymakers often rely on the investigations carried out by academics, it is advisable to ensure that studies delving with the risks of violent conflict outbreak do not skew the causal validity of conflict triggers towards those that are easier to assess via statistical analysis.

—

Any errors author’s own. Analysis to be taken as non-exhaustive.

References

Arnson, C.J. & Zartman, I. W., (2005), Rethinking the Economics of War: The Intersection of Need, Creed and Greed, (Woodrow Wilson Center Press)

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

Ballentine, K. & Nitzschke, H., (2003b) “Policy Lessons from Studies in the Political Economy of Armed Conflict”, International Peace Academy Policy Report

Berdal, M. (2005) “Beyond Greed and Grievance — and not too soon”, Review of International Studies, Vol. 31, pp. 687–698

Cater, C., (2003) “The Political Economy of Conflict and UN Intervention: Rethinking the Critical Case of Africa”, in Karen Ballentine and Jacke Sherman (eds.) The Political Economy of Armed Conflict: Beyond Greed and Grievance, (Lynne Rienner, London)

Collier, P. & Hoeffler, A., (1998) “On Economic Causes of Civil War”, Oxford Economic Papers 50(4): 563–573

Collier, P. & Hoeffler, A., (1999), “Justice-Seeking and Loot-Seeking in Civil War”, World Bank Conflict Research Paper, available at http://www.worldbank.org/research/conflict /papers/justice.html

Collier, P. & Hoeffler, A., (2000) “Greed and Grievance in Civil War”, Policy Research Working Paper, 2355, World Bank Development Research Group

Collier, P., (2000) “Doing Well out of Civil War: An Economic Perspective”, in Mats Berdal & David Malone, eds, Greed & Grievance: Economic Agendas in Civil War (Boulder, CO: Lynne Rienner)

De Soysa, Indra, (2000) “The Resource Curse: Are Civil Wars Driven by Rapacity or Paucity?”, in Mats Berdal & David Malone, eds, Greed & Grievance: Economic Agendas in Civil War (Boulder, CO: Lynne Rienner)

De Soysa, Indra (2002), “Paradise is a Bazaar? Greed, Creed and Governance in Civil War, 1989- 1999”, Journal of Peace Research, Vol. 39, №4

Fearon, J.D. & Laitin D., (2003) “Ethnicity, Insurgency and Civil War”, American Political Science Review, Vol. 97, №1

Grossman, H.I. (1995) “Insurrections”, in K. Hartley and T. Sandier (eds), Handbook of Defence Economics, Vol. I, Elsevier Science B.V., Amsterdam

Gurr, R.T., (1970) Why Men Rebel (Princeton: Princeton University Press)

Keen, D., (1998) “The Economic Functions of Violence in Civil Wars”, Adelphi Paper, 320 (London: International Institute for Strategic Studies)

Olson, M. (1965) The Logic of Collective Action: Public Goods and the Theory of Groups, (Harvard University Press)

Ross, M.L., (2004), “What do we know about Natural Resources and Civil War”, Journal of Peace Research, Vol. 41, №3

Stewart, F. (2008) Horizontal Inequalities and Conflict: Understanding Group Violence in Multiethnic Societies (Palgrave MacMillan)

Tilly, C., (1978) From Mobilization to Revolution (Reading, MA: Addison-Wesley)