Weapons of choice

Abstract

This article investigates the effect of natural resources on whether ethno-political groups choose to pursue their goals with nonviolent as compared to violent means, distinguishing terrorism from insurgencies. It is hypothesized that whether or not the extraction of fossil fuels sparks violence depends both on the group’s characteristics and the state’s reaction. Data are taken from the Minorities at Risk Organizational Behavior (MAROB) project, covering 118 organizations in 13 countries of the Middle East and North Africa over the 1980–2004 period. The multinomial logit models combine group- and country-specific information and show that ethno-political groups are more likely to resort to rebellion rather than using nonviolent means or becoming terrorists when representing regions rich in oil. This effect is enhanced for groups already enjoying regional autonomy or being supported by a foreign state but can be mitigated by power-sharing arrangements. These results are thus in line with the argument that economic considerations, or ‘greed’, dominate over political considerations, or ‘grievances’, with regard to violent conflicts. The opposite appears to hold considering terrorism, as we do not find any evidence for a resource curse here, but find an increasing effect of political discrimination and a decreasing effect of regional autonomy.

Keywords

oil rebellion resource curse terrorism

Introduction

The discovery and exploitation of oil can contribute to a country’s economic growth and prosperity. Resource-abundance can, however, also turn into a threat to stability and peace. While this aspect of the so-called ‘resource-curse’ is widely discussed in the context of civil wars (e.g. Fearon & Laitin, 2003; Collier & Hoeffler, 2004), it has largely been neglected in the literature analyzing the causes of terrorism. This neglect is surprising. In a large number of countries, natural resource abundance has disadvantaged the local population, leading to high regional unemployment and mass immigration (Karl, 2007). It thus seems straightforward that marginalized populations in areas with a wealth of natural resources might resort to terrorism in order to express their grievances. This problem plays a particularly important role in the Middle East and North Africa (MENA) region, which has a large number of oil-rich, fragile states.

Consider Iraq. Political groups such as the Kurdistan Democratic Party or the Patriotic Union of Kurdistan, which represent the Kurdish minority in the north of the country, first fought for more autonomy, then for their own state. During the course of this fighting, they have resorted to violent means, both at a terrorist scale and at a larger battle-sized scale. While the public discourse of the movement focuses on the discrimination of this largest people without their own territory, petroleum reserves are likely to be another important driver of their unrest. Despite their obtaining significant regional autonomy in 1991, the situation has remained tense, with oil revenues being a main cause of conflict both among Kurds (Wimmer, 2002) and between the Kurds and the national government (Chulov, 2009).

In this article, we use a combined framework to investigate whether and to what extent the availability of oil in a region determines whether ethno-political organizations choose to pursue their aims with nonviolent means, resort to terrorism, or start insurgencies. As the previous literature has mostly analyzed the two forms of violence separately, our comparative analysis closes an important gap in the literature. Our focus is on political organizations claiming to represent the interest of specific ethnic populations before their own state – that is, we look at activities within their own country, at the subnational level. We expect that the type of violence applied is a strategic choice, in part based on grievances, in part based on greed, and depending on the organization’s characteristics, the context, and the reaction of the state to its actions. We theorize (in the following section) that groups will weigh risks and benefits of their political actions based on the support they enjoy, their aims, and the strength of the state they face.

We test our hypotheses using data from the Minorities at Risk Organizational Behavior (MAROB) dataset combined with geo-coded data on natural resource reserves. This allows us to match political groups with fossil fuel availability at the regional level, as we explain in more detail in the third section. In the same section we also explain how our multinomial logit panel models combine organization- and country-specific information to test the determinants of an organization’s choice between pursuing their goals with nonviolent means, taking up arms for small-scale terrorist activities, or for a larger-scale rebellion.

A number of studies on the country level predominantly focus either on greed or on the relative importance of greed and grievances (see, inter alia, Collier & Hoeffler, 2004; Collier, Hoeffler & Rohner, 2009; Regan & Norton, 2005). More recently, the focus moved beyond the country level, with Hunziker & Cederman (2012) analyzing the behavior of ethnic groups. We further zoom in on the unit of analysis and look at political organizations, thereby adding an important perspective to the literature. We present our results in the fourth section. They show that insurgencies are more likely with larger resource extraction, both with respect to nonviolent means and terrorism. Both support by foreign states and regional autonomy (and thus a demonstrated will for at least some degree of independence) enhance the escalating impact of oil. The choice to engage in terrorist activities is, however, not affected by resource availability within a group’s territory. This leads us to conclude that economic considerations (or greed) are the main channel through which natural resources affect large-scale violence. While terrorism seems to be driven more by political factors, grievances generated by the extraction of oil are not sufficiently strong to induce terrorist activities among the groups in our sample. We discuss the policy implications of these findings in the final section.

Theory

As Hunziker & Cedermann (2012) point out, the civil war literature widely accepts the existence of a link between petroleum and intrastate conflict. Fearon & Laitin (2003), Humphreys (2005), and de Soysa & Neumayer (2007), among many others, find that countries rich in oil and gas have a higher risk of civil war. This is attributed to a number of factors that can broadly be classified to represent, first, greed or opportunity and, second, grievances. The greed-based hypothesis postulates that resources directly lead to rebellions or coups because controlling an area or state rich in resources is comparably more valuable than one without them. The presence of natural resources has been shown to weaken institutions, as politicians have no incentives to develop them when they do not have to rely on a broad tax base (e.g. Fearon & Laitin, 2003). Furthermore, resource abundance allows rebel groups easy access to finance, making revolutions more feasible (Collier, Hoeffler & Rohner, 2009).

However, the grievance channel to violent behavior cannot be neglected. Karl (2007) stresses the absence of a significant multiplier effect of oil wealth, limited opportunities for technology diffusion, and consequently low living standards for large parts of the population in areas rich in oil. Hunziker & Cederman (2012) point to examples of externalities that include the reorganization of land rights, pollution, disruptions of the labor market due to shifts in demand away from unskilled workers, large-scale in-migration, urbanization, and rapid centralization of state powers. They find that violent reactions of ethnic groups become likely when members of the group feel themselves deprived of their fair share of gains from natural resources and when these resources incur negative externalities on them that are not mitigated by participatory political institutions. They interpret their results as showing the role of grievances to be of equal importance to that of greed in explaining civil war.

In a study on 13 cases, Ross (2004) tests potential causal channels for the resource–conflict relationship. He argues that separatist motives are likely to come into play in cases of grievances over the distribution of benefits from resource extraction or based on the incentive to control these revenues. His results emphasize the importance of the geographical distribution of oil across the country for conflict. He also shows pre-emptive repression of groups by the state out of fear of losing control over resources as well as interventions by foreign states that spark civil wars. This is in line with de Soysa & Binningsbø (2009: 21), who show that natural resource abundance leads to the repression of large parts of the population.

In contrast to the literature on larger-scale civil unrest, natural resources hardly feature in the literature on what determines terrorism. Exceptions to this are Tavares (2004), Bravo & Dias (2006), and Sambanis (2008). Yet, they do not provide specific theories as to why resource wealth should play a role in the occurrence of terrorism, nor do their findings draw a clear picture regarding the relationship between fossil fuels and terror. Gassebner & Luechinger (2011) include the share of a country’s total exports made up of primary goods in their large-scale robustness analysis of what determines terror. Across their models, they do not find a robust relationship between their measure of resource abundance and the number of terrorist attacks against a country’s citizens.

Arguably, the presence of natural resources is important in determining the extent of terrorism as well as insurgencies. The externalities of mineral resource extraction described above all compound into substantial grievances, potentially leading representatives of repressed minorities to resort to terrorist activities.¹ For example, the transformation or modernization of the economy has been found to lead to socio-economic and demographic changes feeding through to terrorism (Ross, 1993). Specifically, the link between political rights and terrorism has been commonly shown in the literature (e.g. Danzell, 2011; Krueger & Malečková, 2003). These rights are often restricted in areas of resource extraction as described above. Furthermore, Piazza (2011) shows that countries featuring minority economic discrimination are more likely to experience domestic terrorist attacks. As discussed above, the exclusion of some groups from the gains of fossil fuels is common. The neglect of natural resources in the literature on terrorism is thus surprising.

So far, our discussion has concerned the choice of violence over nonviolent means, but we have had no hypotheses regarding the likelihood of resorting to terrorism over insurgency or vice versa. Sambanis (2008) argues that terrorism and civil wars are distinct strategic choices – civil wars being driven more by economic factors and terrorist activities rather than by political aspects. He finds that the power (a-)symmetry between the group and the state, as well as the level of public support, determine the degree of mobilization and the type of violence. Regan & Norton (2005) differentiate between the importance of grievances as the backbone of a movement and the importance of resources as the means of paying out selective benefits to group members. They find that, overall, similar factors are related to protest, rebellion, and civil war (namely, income and distributional issues, repressive policies of the state, and access to exploitable resources). Yet, the reaction of the state determines whether violence escalates.

A small and recent literature analyzes groups that apply terrorism during civil war (but not the distinct choice between them), finding that democracies are more vulnerable to civilian casualties and thus terrorism, while groups depending on mass mobilization would not attack civilians (Stanton, 2013). Additionally, terrorism appears to benefit the group’s survival, but not to be effective in reaching its political goals (Fortna, 2015). Findley & Young (2012) describe how terrorism plays different roles before, during, and after a conflict. Similar to our theoretical framework, Polo & Gleditsch (2014) analyze terrorism in civil war as a bargaining process between the group and the state, based on the rebels’ objectives, their available resources, and the expected reaction of the government.

The approach of looking at groups rather than at countries is valuable both from a theoretical and from an empirical point of view. Fine degrees of motivation and strategic policy choices are likely to drive terror as well as larger-scale unrest, which can only be identified by looking at organizations. One possibility to capture these strategic considerations is to assume that a group’s ability for collective violence depends on its members’ expected costs and benefits, taking into account the socio-economic and political context (Conteh-Morgan, 2004). The group does not operate in a vacuum but is affected by its surroundings, especially the state against which it rebels and which reacts to this threat. The institutional environment influences the ability of opposing groups to mobilize, their perceived chances of success, and the political measures at their disposal (Muller & Seligson, 1987). Noticeably, collective action turns violent when those protesting against a certain perceived grievance do not have access to institutions that nonviolently mediate them (Tarrow, 1998).

In the context discussed here, a group’s strategic ‘weapon of choice’ will depend on balancing of costs and benefits of reaching the political aim most efficiently. The extent of mobilization – clearly smaller for terrorist activities than for insurgencies – then depends on both the need or the desired political outcome and the ability, that is, the strength of the state and the number of people willing to join the movement.

Based on theory and evidence described in this section, our empirical analysis is built along the following hypotheses. The mere existence of fossil fuels in a region leads to disturbances which can cause both terrorism and insurgency, while at the same time, revenues can be used to pay selective benefits (e.g. Regan & Norton, 2005).² We thus expect both forms of violence to increase with oil revenues (Hypothesis 1). Following Hunziker & Cederman (2012) this effect on insurgencies should be mitigated by increased citizen participation in the wealth created by the resources and in deciding how to exploit them (Hypothesis 2a). In line with Dreher & Fischer (2010), we expect participation in power to also reduce the extent of terrorism (Hypothesis 2b). Closely related, political discrimination should enhance violence linked to resources (Hypothesis 3a for terrorism and Hypothesis 3b for insurgencies). In line with Sambanis (2008), we consider terrorism to be driven more by political reasons and insurgencies by economic reasons, thus expecting the effect of discrimination to be larger for terrorist activities (Hypothesis 3c).

We interact our oil measure with indicators for access to political power as well as political discrimination to test Hypotheses 2 and 3. In contrast, where separatist motivations exist and a state of autonomy has already been reached, oil revenues might be a motivation to violently strive for complete secession (Ross, 2004). Thus, we interact our oil indicator with a measure of autonomy and expect that it exclusively strengthens the effect of oil on insurgencies (Hypothesis 4). As Karl (2007) points out, oil-induced income inequality is likely to be perceived as more severe compared to similar levels of inequality due to other reasons because the income-generating process is perceived to be unfair. We therefore also interact our measure of resource abundance with economic discrimination and, just as for political discrimination, expect a stronger impact on terrorist activities than on civil wars (giving rise to Hypotheses 5a on terror, 5b on insurgencies, and 5c on the comparative effect). We expect the strength of the group to play a key role. The stronger the state relative to dissenting groups, the higher the probability that such groups will turn to terrorism rather than other forms of violence (Hypothesis 6). In other words, if a group feels strong enough vis-à-vis the state, it will dare to take up arms in a more coordinated fashion (Carter, 2015; Ross, 2004; Regan & Norton, 2005; Sambanis, 2008). We will test this using the variable of whether a group is supported by a foreign state as a proxy for the strength of an organization, as ‘an available source of support external to the arena of conflict can empower organizations to engage in contentious politics in a way inaccessible to those without similar sources’ (Asal et al., 2013: 309f).

Method and data

Our approach follows a number of recent papers focusing on violent organizations relying on multinomial logit regressions (Gaibulloev & Sandler, 2014; Asal, Brown & Schulzke, 2015; Carter, 2012). Closely related to our work, Meierrieks & Krieger (2014) model the choice between terrorism and civil war. The multinomial logit model allows us to determine differential impacts of the variables of interest on the strategic choice of the observed political organizations. This assumes that the process from nonviolence to terrorism to insurgency is not continuous, that is, it is not a process of (de)escalation, but rather represents separate decisions. However, even if the process were ordered, the multinomial specification would still be important to be able to estimate separate coefficients for the explanatory variables for each possible outcome. When organizations engage in terror and larger-scale insurgencies in the same year we code them as insurgencies, as our method of estimation requires the groups to be exclusive.³ Looking at dynamics over time, most groups that are nonviolent remain nonviolent (with a likelihood of 90.6%), but especially with regard to terrorism there is quite some movement both towards escalation and de-escalation.⁴

We implement our specification as a multilevel model, including random intercepts for each organization. This allows us to exploit the panel structure of our dataset and thus variation for the same group over time rather than across organizations. This is a novelty with regard to the other studies using multinomial logit models introduced above.⁵ Our reduced-form empirical model is at the organization-year level:

W E A P O N_{i, t} = α + β R E S O U R C E S_{i, t - 1} + γ X_{i, t - 1} + δ R E S O U R C E S_{i, t - 1} * X_{i, t - 1} + ζ Z_{i, t - 1} + ∊_{i, t},

where WEAPON reflects organization i’s ‘weapon of choice’ in year t, RESOURCES is our indicator of natural resource abundance, where we expect β > 0 (Hypothesis 1). X represents the variables we interact with oil production to test our hypotheses: (i) two indicators for political discrimination and whether the ethnic group shares power with others (Hypotheses 2 and 3); (ii) an indicator for regional autonomy of the ethnic group (Hypothesis 4); (iii) an indicator for the group being economically discriminated against (Hypothesis 5); and (iv) whether a group was supported by a foreign state (Hypothesis 6). We expect δ > 0 in all cases except for power-sharing, where we expect δ < 0. Z contains our control variables (at the country and group level) and ∊ is the error term, which is clustered at the organization level. All our independent variables are lagged by one year in order to minimize the bias due to reversed causality.⁶

Our main variables are taken from Asal, Pate & Wilkenfeld’s (2008) Minorities at Risk Organizational Behavior (MAROB) dataset. The dataset contains an unbalanced panel of 118 political organizations claiming to represent the interests of 22 ethnic groups in 13 countries and territories of the Middle East and North Africa, over the 1980–2004 period.⁷ Our categorical dependent variable measures whether an organization is nonviolent in a given year, whether it carries out any terrorist activity, or whether it is involved in a larger-scale insurgency, thus ranging between 0 and 2.⁸ Distinguishing the two forms of violence is a key challenge to our econometric analysis. We will rely on a combination of action-based (the level of violence) and actor-based (the group’s attributes) approaches (Asal et al., 2012). According to Mickolus et al. (2004: 2) ‘terrorism is the use or threat of use, of anxiety inducing extranormal violence for political purposes by any individual or group, whether acting for or in opposition to established government authority, when such action is intended to influence the attitudes and behavior of a target group wider than the immediate victims’. Criteria for the inclusion of a group in the MAROB database include that it must not be created by the government and that it has to be political in its goals and activities. Following a large number of previous studies, the definition for terrorism applied here is a narrow one, comprising violent attacks on civilians only (including non-security state personnel such as civil service personnel and government representatives), but excluding those on state institutions and the military, which are conceptually different and often termed as guerilla activities (see, inter alia, Abrahms, 2012; Fortna, 2015; Kydd & Walter, 2006). Specifically, any group that attacked civilians on a low scale or forcefully secured their support is deemed to be a terrorist organization.⁹ Large-scale violent events include those targeting security personnel and state institutions and those attacks that attempt to seize control over a town, guerilla activity, and civil wars fought by rebel military units with base areas. Violence arising from groups with control over a specific area with some degree of governance structure is also included in this category.

Asal, Pate & Wilkenfeld’s (2008) data have two main advantages over alternative datasets. First, they are available at the organization level rather than the ethnicity or country level. Compared to data at the country level, this allows using geo-coded data on natural resources to test whether resources in a certain region affect violence related to the same region. More broadly, our data allow the investigation of more differentiated reasons for violence. Rather than attributing violence to ethnicities as a whole, characteristics of groups from the same ethnicity can be distinguished (Asal & Wilkenfeld, 2013). Second, the dataset includes nonviolent as well as violent groups. This is contrary to most previous organizational-level studies that include organizations only once they become violent (Stanton, 2013; Fortna, 2015); these studies are therefore unable to examine the determinants of why organizations choose to be violent per se. However, the data have a number of drawbacks as well. The most important one is the limited regional coverage and the resulting small number of independent observations we can exploit for our regressions. The MENA region is different from other areas in a number of ways, so that we are careful to avoid generalizing our results to other regions of the world. What is more, while Asal, Pate & Wilkenfeld (2008) follow clear guidelines on how to code the actions of organizations, the boundaries between terrorism and insurgencies in particular are sometimes blurred (Sambanis, 2008), and the resulting data are noisy. We have no reason, however, to expect a systematic bias in testing our hypotheses, and we make this distinction as clear as possible by applying the strict definition described above.

We rely on two indicators for natural resource abundance, coded at the regional level. Our main resource indicator follows Hunziker & Cederman (2012), who use data from Horn’s (2010) Giant Oil and Gas Fields of the World database which includes geo-coded information on the location and size of petroleum occurrence across the world (for fields containing at least 500 million barrels of oil or gas equivalents). The data allow us to code the share of a state’s oil reserves that is situated on a specific ethnic group’s territory. The annual value of a country’s oil production (taken from Ross, 2013) is then weighed with these shares to estimate the return to oil production on a group’s territory in a given year in 2009 US$.¹⁰ The resulting resource variable thus varies across groups and time. Given that the variable is highly skewed, we use the log of the variable.¹¹

Our second indicator of resource abundance is a binary indicator based on the geo-coded location of oil and gas fields in PRIO’s Petroleum dataset v.1.2 (Lujala, Rød & Thieme, 2007). It has the advantage of also including small fields. However, these data do not measure the degree of resource abundance. What is more, they hardly vary within groups in the same country and do not vary at all within the same country over time.

We use a number of variables to control for observed heterogeneity at the group level and the country level. At the group level, and also taken from the MAROB database, we control for the goals of a group. Specifically, we include indicator variables for organizations that aim to eliminate political, economic, or cultural discrimination, groups that aim for autonomy or independence, and groups that want to establish an Islamic state.¹² Asal, Pate & Wilkenfeld (2008) coded these variables based on the expressed aims and motivations of the groups as reported in newspapers and other sources. We expect fighting for autonomy or independence, or an Islamic state, to lead groups to take up arms at a larger scale as these are goals that states do not usually give in to. Organizations with ‘other’ goals are the omitted category.

We control for whether organizations receive financial, political, humanitarian or military support from foreign states, as this is likely to fuel violence. We control for negotiations between the state government and the political organization, as members of the group that do not wish to reach an agreement with the state or that expect larger concessions when showing strength could opt for increased violence. In addition, we include whether or not the government uses violence against an organization, that is, if the organization is considered legal or if it faces lethal violence by the state. We also add a variable indicating whether a group provided social services as this requires a certain degree of organization as well as financial means and thus strength.

At the country level, we rely on a number of standard control variables from the terrorism and civil war literature. Due to our very small sample of countries, however, we will not put a huge weight on their estimated coefficients. We control for whether or not the country is a democracy, relying on indicators from Freedom House (2014) for the average levels of civil liberties and political rights, ranging between 1 and 7, with higher values indicating less freedom.¹³ We also include a country’s logged GDP per capita to proxy for its level of development. As Sambanis (2008) points out, the negative correlation between per capita GDP and civil war is widely accepted. GDP per capita, however, is not a robust determinant of terrorism (Gassebner & Luechinger, 2011; Sambanis, 2008). We control for ethno-linguistic fractionalization because of the assumption that a higher degree of fractionalization leads to a higher potential for conflict. However, the empirical evidence regarding the effects of fractionalization is mixed (see Blattman & Miguel, 2010).

In line with the previous literature, we expect greater levels of repression in countries with larger populations, where the chance for conflict is larger (de Soysa & Binningsbø, 2009). Gassebner & Luechinger (2011) find population to be among the few variables that robustly increase terrorism. Collier & Hoeffler (2004) and Collier, Hoeffler & Rohner (2009) find the risk of civil war to increase with population. Following Hunziker & Cederman (2012) we also control for the logged value of oil produced at the national level, which could be related to facets of the resource curse relevant at the country level rather than the group level. We report the sources of all variables, their descriptive statistics, and exact definitions in the Online appendix to this article.

Results

Table I shows the results without interacted variables, with nonviolence being the omitted base category. The coefficients thus allow us to compare the choice of the two forms of violence with respect to nonviolence. We report relative risk ratios (or odds ratios) that can be directly interpreted with respect to the quantitative effect of the variables. These exponentiated multinomial logit coefficients show to what extent the risk of an outcome changes relative to the reference group following a unit change in a variable, for constant values of the other variables in the model.¹⁴ Odds ratios larger than 1 indicate a positive correlation between an explanatory variable and the respective outcome, while odds ratios smaller than 1 indicate negative relationships. By testing whether the difference between the odds ratios for our two violent outcomes is significant, we can also compare them among each other.

We start with including our two oil variables without any control variables (columns 1 and 2 of Table I) before adding group characteristics (columns 3 and 4) and finally estimating the full model (without any interactions, columns 5 and 6). As can be seen, ethno-political groups are more likely to engage in insurgencies the higher the value of the oil resources that were extracted from their territory in the previous year. This effect is robust across the different specifications. The odds ratios in columns 2, 4, and 6 of Table I are significant at the

Table I.

Determinants of terror and insurgency, multinomial logit, 1980–2004, reduced form and full models without interaction effects

	(1)	(2)	(3)	(4)	(5)	(6)
	Terror	Insurgency	Terror	Insurgency	Terror	Insurgency
Log(Group oil production)	1.051 (0.545)	1.316† (0.087)	1.024 (0.734)	1.352** (0.004)	0.986 (0.863)	1.374** (0.000)
Log(National oil production)	0.921 (0.292)	0.802 (0.183)	0.969 (0.636)	0.809* (0.048)	1.011 (0.897)	0.800* (0.040)
Goal: Eliminate discrimination			2.390 (0.373)	3.544 (0.428)	4.638 (0.192)	15.462 (0.125)
Goal: Autonomy, independence			0.293 (0.253)	0.129 (0.213)	0.230 (0.205)	0.740 (0.870)
Goal: Eliminate economic discrimination			0.696 (0.761)	6.526 (0.232)	0.401 (0.430)	5.958 (0.166)
Goal: Eliminate cultural discrimination			0.994 (0.993)	1.105 (0.910)	0.442 (0.263)	0.765 (0.757)
Group supported by foreign state			4.319** (0.003)	3.551** (0.004)	3.455* (0.033)	3.076* (0.017)
Goal: Islamic state			3.820 (0.326)	2.761 (0.631)	7.574 (0.208)	4.513 (0.528)
State uses violence against group			3.192* (0.018)	2.071 (0.175)	9.561** (0.000)	3.653† (0.068)
State negotiated with organization			0.349** (0.038)	0.596 (0.376)	0.314* (0.057)	1.022 (0.969)
Group provides social services			7.893** (0.009)	15.719** (0.003)	15.020** (0.001)	18.315** (0.000)
Log(GDP p.c.)					5.668* (0.014)	2.394 (0.165)
Log(Population)					0.882 (0.691)	1.254 (0.621)
Freedom House					1.943† (0.056)	1.230 (0.558)
Ethno-linguistic fractionalization					0.302 (0.582)	161.61* (0.029)
Number of groups	112		107		105
Number of observations	5,031		4,146		3,360
Log-Likelihood	−842.403		−644.081		−424.210

Odds ratios shown. p-values in parentheses: ^† p < 0.1, *p < 0.05, **p < 0.01. All variables are lagged by one year and standard errors are clustered at the organization level.

10% (column 2) and at the 1% (columns 4 and 6) levels and indicate that the odds of a group being involved in an insurgency rather than in nonviolent activities increase by a factor of 1.37 (full model, column 6) with an increase in the logged value of oil production in the group’s area by 1 (its mean being 7.55, and the baseline odds of insurgency being 0.29).¹⁵ There is no evidence that resource abundance in the group’s territory affects its choice of terrorism versus nonviolence, however (this can be seen from columns 1, 3, and 5 of Table I). Figure 1 shows the predicted probabilities of the three outcomes as a function of oil conditional on all included control variables being held constant at their mean (based on columns 5 and 6 of Table I). As can be seen, the probability of terrorism is rather small in this sample and does not react strongly to an increasing value of regional oil production.¹⁶ However, the probability of insurgency increases notably with resource

Figure 1.

Predicted probabilities of the three outcomes as a function of lagged Log(Group oil production)

availability at the cost of nonviolence which gets increasingly unlikely. These results point to the absence of grievances arising from the extraction of oil strong enough to induce terrorism among the countries and years in our sample. Thus, regarding Hypothesis 1, it appears that the effect of oil on insurgencies as reported in the previous literature finds support at the subnational level, while we do not find a significant relationship with terrorism. When comparing the coefficients for the two violent outcomes, oil makes civil wars significantly more likely than terror (p-value 0.0009).

Regarding the control variables of the models, the results in Table I show that the groups’ official goals do not appear to make a difference regarding their pursuing these aims in a nonviolent or violent manner. In contrast, aiming at eliminating economic discrimination is statistically significantly more related to large-scale violence than to terrorism (p-value 0.008). Having the support of a foreign state makes both forms of violence more likely (columns 3 to 6). There is no significant difference between the two outcomes in this regard (p-value 0.8561). A state using violence against a group robustly increases the likelihood of this group turning to terrorism, while the same is only true for insurgencies in the full model (column 6), but even here the impact is significantly larger for terrorist activities than for insurgencies (p-value 0.04). Negotiations reduce the probability of terrorism (significant at the 5% and 10% levels), while a group providing social services – our proxy for the degree of organization – is more likely to resort to both forms of violence, the effect being significant at the 1% level in all specifications (columns 3 to 6).

As pointed out above, we do not put a lot of weight on the national control variables due to the small number of countries in the sample. Overall, the effect of oil production at the national level is not robustly significant. When adding group characteristics (column 4) and also country-level variables (column 6), however, it seems that extracting fossil fuels somewhere on the national territory decreases the probability of violent outbreaks, possibly due to positive spillovers from these regions in terms of social services or employment. Focusing on the full model (i.e. columns 5 and 6), both richer and less democratic countries are more likely to face terrorist attacks but not to be confronted with larger-scale challenges. While the positive relationship between per capita GDP and terrorism is in line with the literature above, the negative association between democracy and terrorism is not, and might arise both from the specificities of the region under observation or the small sample size. On the other hand, ethno-linguistic fractionalization has a large and positive effect on insurgency (in line with the previous literature) but does not appear to be linked to terror. This is also confirmed by the significant difference between the two odds ratios in columns 5 and 6 (p-value 0.003).

We next turn to Hypotheses 2 and 3, testing whether the effect of oil extraction on the ‘choice of weapons’ depends on political discrimination and possibilities of political participation of the ethnic group. Interpreting the significance of interaction effects in non-linear models such as ours might not be straightforward. However, these difficulties do not pertain to incidence rate ratios, which rely on a multiplicative rather than an additive scale (Buis, 2010) and thus correctly calculate the significance of the incidence ratio. The odds ratio of the interaction then reflects the ratio of the odds ratios of the two interacted variables (which do not depend on the values of the other variables in the model).

An ethnic group is considered to face stronger political discrimination if a group is not only politically underrepresented but if, additionally, there are either no measures taken to remedy the situation or no measures introduced that further restrict the group’s political participation relative to other groups. We take this

Table II.

Determinants of terror and insurgency, multinomial logit, 1980–2004, political participation

	(1)	(2)	(3)	(4)
	Terror	Insurgency	Terror	Insurgency
Log(Group oil production)	0.935 (0.465)	1.300** (0.002)	0.962 (0.660)	1.576** (0.000)
Log(National oil production)	1.124 (0.268)	0.769** (0.022)	1.826* (0.021)	0.779* (0.053)
Interaction term group oil and political discrimination	1.001 (0.961)	1.044 (0.108)
Political discrimination	1.860* (0.038)	0.668 (0.255)
Interaction term group oil and power sharing			1.066 (0.487)	0.744** (0.001)
Ethnic group shares power with others			3.386 (0.347)	14.031* (0.092)
Goal: Eliminate discrimination	1.495 (0.758)	21.714† (0.071)	5.799 (0.228)	32.692* (0.032)
Goal: Autonomy, independence	0.047* (0.014)	0.746 (0.864)	0.043* (0.026)	1.580 (0.769)
Goal: Eliminate economic discrimination	0.552 (0.578)	4.113 (0.192)	0.375 (0.510)	10.153† (0.097)
Goal: Eliminate cultural discrimination	0.451 (0.297)	0.756 (0.757)	0.447 (0.421)	0.451 (0.364)
Group supported by foreign state	3.062† (0.053)	3.125* (0.018)	1.398 (0.608)	3.086† (0.061)
Goal: Islamic state	5.013 (0.333)	3.866 (0.561)	7.108 (0.250)	9.059 (0.304)
State uses violence against group	8.248** (0.000)	3.834† (0.060)	4.547* (0.028)	3.448 (0.144)
State negotiated with organization	0.444 (0.168)	1.086 (0.867)	1.173 (0.811)	0.832 (0.814)
Group provides social services	24.715** (0.000)	17.172** (0.000)	15.369* (0.024)	6.077 (0.143)
Log(GDP p.c.)	4.336* (0.025)	1.858 (0.307)	25.525** (0.001)	1.342 (0.698)
Log(Population)	0.658 (0.211)	1.715 (0.226)	0.006** (0.005)	2.809 (0.186)
Freedom House	1.750† (0.083)	1.130 (0.696)	0.630 (0.460)	0.542 (0.126)
Ethno-linguistic fractionalization	0.244 (0.514)	255.829* (0.015)	4.201 (0.662)	21,042.7** (0.005)
Number of groups	103		88
Number of observations	3,336		2,517
Log-Likelihood	−414.307		−313.042

Odds ratios shown. p-values in parentheses: ^† p < 0.1, *p < 0.05, **p < 0.01. All variables are lagged by one year and standard errors are clustered at the organization level.

variable from the Minorities at Risk (2009) database. In order to measure whether an ethnic group represented by an organization in our sample has a share in central political power, we rely on a variable from the Ethnic Power Relations dataset (Wimmer, Cederman & Min, 2009). As can be seen in Table II, while political grievances per se increase the probability of terror by a factor of 1.9, on average (see column 1), the interaction effect is not significant at conventional levels. The effect of regional fuel extraction for either type of violence does

Table III.

Determinants of terror and insurgency, multinomial logit, 1980–2004, autonomy and economic discrimination

	(1)	(2)	(3)	(4)
	Terror	Insurgency	Terror	Insurgency
Log(Group oil production)	1.017 (0.828)	1.307** (0.001)	0.976 (0.778)	1.322** (0.001)
Log(National oil production)	1.605* (0.050)	0.846 (0.107)	1.108 (0.276)	0.811* (0.036)
Interaction term group oil and autonomy	0.541** (0.000)	1.249* (0.031)
Ethnic group has regional autonomy	0.464 (0.487)	0.156 (0.352)
Interaction term group oil and economic discrimination			0.996 (0.865)	1.014 (0.674)
Economic discrimination			2.070** (0.005)	0.895 (0.734)
Goal: Eliminate discrimination	6.752 (0.183)	5.925 (0.281)	1.545 (0.735)	15.744† (0.100)
Goal: Autonomy, independence	0.106† (0.087)	1.144 (0.934)	0.065* (0.019)	0.779 (0.883)
Goal: Eliminate economic discrimination	0.814 (0.912)	10.723 (0.163)	0.470 (0.476)	4.520 (0.190)
Goal: Eliminate cultural discrimination	0.768 (0.784)	0.650 (0.636)	0.588 (0.446)	0.932 (0.938)
Group supported by foreign state	1.409 (0.608)	2.477† (0.096)	3.087† (0.058)	3.143* (0.017)
Goal: Islamic state	6.482 (0.252)	3.788 (0.502)	5.283 (0.257)	3.919 (0.538)
State uses violence against group	5.208* (0.018)	3.590 (0.125)	9.217** (0.000)	4.012† (0.051)
State negotiated with organization	0.786 (0.750)	0.472 (0.417)	0.451 (0.206)	1.038 (0.944)
Group provides social services	7.529 (0.108)	4.751 (0.261)	23.620** (0.001)	16.872** (0.000)
Log(GDP p.c.)	11.445** (0.002)	3.042 (0.131)	4.528* (0.019)	2.151 (0.195)
Log(Population)	0.016* (0.016)	2.307 (0.278)	0.739 (0.321)	1.301 (0.505)
Freedom House	0.550 (0.302)	0.654 (0.317)	1.614 (0.149)	1.202 (0.569)
Ethno-linguistic fractionalization	21.279 (0.338)	33,481.84** (0.001)	0.405 (0.681)	134.373* (0.030)
Number of groups	88		103
Number of observations	2,517		3,336
Log-Likelihood	−315.478		−415.383

Odds ratios shown. p-values in parentheses: ^† p < 0.1, *p < 0.05, **p < 0.01. All variables are lagged by one year and standard errors are clustered at the organization level.

not depend on grievances, at least according to the proxy we use here (columns 1 and 2). While this is in line with our assumption of terrorism being a more political phenomenon than insurgency, this does not appear to be linked to natural resources. In contrast, the diminishing effect of power-sharing on the relationship between oil and civil war finds support in column 4. In quantitative terms, the odds ratio of 0.74 indicates that the effect of an increase in oil production in the group’s area for groups sharing power is 74% of the effects for groups that do not share power.¹⁷ This result is in line with our hypothesis and Hunziker & Cederman’s (2012) observation that groups that participate in power are less likely to choose violent over nonviolent means.

Columns 1 and 2 of Table III present the results for Hypothesis 4, including an interaction term between our indicator of regional resource abundance and a binary variable indicating whether or not an ethnic group enjoyed regional autonomy. This information is also taken from the EPR database (Wimmer, Cederman & Min, 2009). While regional autonomy appears to make terrorism less likely in combination with oil reserves (column 1), it increases the probability of violent conflict (column 2). The odds ratios show that in autonomous regions the effect of an increase in oil production in the group’s area on the probability of terror (insurgency) compared to nonviolence is 54% (125%) of the effect for groups without regional autonomy.¹⁸ This is in line with Sambanis (2008), showing economic factors to be important for civil wars, while terrorist activities are predominantly driven by political aspects. Political grievances over the appropriation of fuel revenues are largely addressed by the status of regional autonomy while only independence grants full economic control. The finding regarding civil war is also in accordance with Ross’s (2004) result that regions where ethnic groups strive for more autonomy might be driven into secessionist wars where financial incentives from natural resources are available.

Columns 3 and 4 of Table III turn to Hypothesis 5, testing the impact of economic discrimination. The findings resemble those for political discrimination in that economic discrimination per se increases the probability for terrorism twofold (column 3) but does not appear to be related to insurgencies or have a differential effect in oil-extracting areas.

The results for Hypothesis 6 – expecting support by a foreign state to increase the likelihood for insurgency in oil-rich regions due to increased strength of the group relative to the central state and also based on higher stakes for the supporting state – yield some interesting insights: column 1 of Table IV shows that the significant effect of foreign support on terrorism is independent of oil. In contrast, among the groups backed by foreign

Table IV.

Determinants of terror and insurgency, multinomial logit, 1980–2004, support by foreign state

	(1) Terror	(2) Insurgency
Log(Group oil production)	0.971 (0.716)	1.319** (0.004)
Log(National oil production)	1.009 (0.915)	0.796* (0.039)
Interaction term group oil and foreign support	1.079 (0.192)	1.151* (0.037)
Goal: Eliminate discrimination	4.101 (0.233)	12.072 (0.180)
Goal: Autonomy, independence	0.179 (0.151)	0.498 (0.722)
Goal: Eliminate economic discrimination	0.447 (0.505)	6.268 (0.177)
Goal: Eliminate cultural discrimination	0.443 (0.270)	0.717 (0.696)
Group supported by foreign state	2.731† (0.094)	1.677 (0.225)
Goal: Islamic state	7.152 (0.232)	3.950 (0.583)
State uses violence against group	10.234** (0.000)	4.180* (0.048)
State negotiated with organization	0.283* (0.045)	0.830 (0.769)
Group provides social services	16.847** (0.000)	22.143** (0.000)
Log(GDP p.c.)	5.260* (0.021)	2.300 (0.198)
Log(Population)	0.893 (0.714)	1.274 (0.599)
Freedom House	1.876† (0.075)	1.222 (0.580)
Ethno-linguistic Fractionalization	0.346 (0.627)	180.203* (0.027)
Number of groups	105
Number of observations	3,360
Log-Likelihood	−421.664

Odds ratios shown. p-values in parentheses: ^† p < 0.1, *p < 0.05, **p < 0.01. All variables are lagged by one year and standard errors are clustered at the organization level.

states, presence of oil reserves makes insurgency more likely (by a factor of 1.15).¹⁹ This could be explained by Ross’s (2004) conclusion that the high possible income from fossil fuels encourages foreign states to back groups in larger-scale insurgencies (column 2).

In summary, we find evidence in line with the ‘resource curse’ in relation to large-scale violence but not when it comes to terrorism. In our approach to differentiating the two, we find political and economic discrimination to increase terrorism, but independent of oil resources. However, while power-sharing can mitigate the escalating impact of mineral resources, in areas that strive for more independence, the prospect of high revenues appears to induce insurgencies. Similarly, foreign state involvement in regions with fossil fuel reserves is likely to spark civil wars but not terrorism.²⁰

Conclusion

In this article, we investigate what determines ethno-political organizations’ choice between pursuing their goals with nonviolent means and violent action, distinguishing between smaller-scale terrorist activities and larger-scale insurgencies. Combining the two forms of violent behavior within the same framework allows us not only to identify determinants of violence as such but also to distinguish between different forms of violent actions. According to our theory, the extraction of natural resources exerts externalities on ethnic groups populating the regions where resources are extracted, leading to grievances. At the same time, revenues from fossil fuels represent important economic incentives. The consequent risk of both terrorism and rebellion depends on the group’s characteristics as well as the state’s reaction to its actions. Based on this reasoning, we run a multinomial logit model where we include regional and national oil production and then add a number of interaction terms representing these factors. Indeed, our results show that insurgencies are more likely with greater resource extraction, both with respect to nonviolence and with respect to terrorism. The choice to take up arms for terrorist activities is not affected by resource availability, however.

Access to a share in political power reduces the likelihood of insurgencies. While Hunziker & Cederman (2012) interpret this result to indicate the importance of grievances rather than greed, it could also be that ethnic groups participating in power might be able to extract a larger share of the resources in their territory, reducing the chances of greed-based insurgencies. This reading is supported by the finding that a status of regional autonomy helps to diminish the effect of regional oil extraction on terrorist activities but even enlarges the probability of insurgencies. What is more, while economic and political grievances per se increase the risk of terrorism, this relationship appears to be independent from fossil fuel extraction. Additionally, in regions where oil can be extracted, foreign states’ support of the political group also escalates violent processes.

Taken together, our findings indicate that natural resources are an essential factor in the mobilization for civil war – for example, as a means for selective payments – but are less important for terrorist movements, where a smaller degree of mobilization is required. This is in line with the analyses of both Sambanis (2008) and Ross (2004). We also conclude that terrorism is driven by political rather than economic factors. In contrast, greed dominates grievance as a motive for ethno-political organizations to turn to civil wars as a consequence of oil production in the regions of the ethnicities that they represent (in line with the country-level analyses in Collier & Hoeffler, 2004 and Collier, Hoeffler & Rohner, 2009).

In addition, the regional distribution of oil matters as do specific group characteristics and an organization’s standing vis-à-vis the state. The information on all of these aspects is lost when comparing countries, as has been the focus of most of the previous literature. Furthermore, although the geographical coverage of the MAROB data is limited to the MENA region, looking at groups should arguably benefit from a higher external validity. The ‘rationality of actors’, that is, the way humans weigh costs and benefits, should be independent of their country of origin, although the resulting action then follows from the specific context. For example, in a very different environment, the availability of lootable oil resources appears to impact the strategic considerations of maritime pirates in the Gulf of Guinea and rebel leaders in Nigeria (Daxecker & Prins, 2015).

Our results bear important policy implications. In order to reduce the negative consequences of oil resources, central governments need to carefully balance the degree of decentralization and regional autonomy they give to the regions containing these resources. Countries plagued with terrorism (but not larger-scale insurgencies) will do well to give affected regions some autonomy. Countries affected by larger-scale insurgencies should give regional groups a share in political power short of regional autonomy. Returning to the example of Iraq we raised in the introduction, and taking our results at face value, greater regional autonomy has already proved to increase the risk of insurgency while reducing terrorism coming from the country’s Kurdish minority. The larger-scale insurgency of the Islamic State (IS) should thus not result in regional autonomy, but rather with a share in political representation at the central government.

Future research should take up the innovations of this article – notably moving the ‘resource curse’ literature to the subnational level, looking at political groups, and comparing causes of violent behavior per se as well as causes of its different forms – and include further geographical regions and natural resources.

Footnotes

Replication data

The Online appendix, dataset, and do-file for the empirical analyses in this article are available at . We used Stata 13 for all regressions.

Acknowledgements

We thank Dominik Noe for participating in developing the idea and constructing the database for this article. We thank Todd Sandler and other participants of the Terrorism and Policy Conference 2014, as well as the reviewers and editor of this journal for helpful comments, and Jamie Parsons for proofreading.

Funding

Merle Kreibaum gratefully acknowledges funding by the German Research Foundation (DFG).

Notes

References

Abrahms

Max

(2012) The political effectiveness of terrorism revisited. Comparative Political Studies 45(3): 366–393.

Asal

Victor

Wilkenfeld

Jonathan

(2013) Ethnic conflict: An organizational perspective. Penn State Journal of Law & International Affairs 2: 91–182.

Asal

Victor

Brown

Mitchell

Schulzke

Marcus

(2015) Kill them all – Old and young, girls and women and little children: An examination of the organizational choice of targeting civilians. Political Science Research and Methods 3(3): 589–607.

Asal

Victor

Calle

Luis De La

Findley

Michael

Young

Joseph

(2012) Killing civilians or holding territory? How to think about terrorism. International Studies Review 14(3): 475–497.

Asal

Victor

Legault

Richard

Szekely

Ora

Wilkenfeld

Jonathan

(2013) Gender ideologies and forms of contentious mobilization in the Middle East. Journal of Peace Research 50(3): 305–318.

Asal

Victor

Pate

Amy

Wilkenfeld

Jonathan

(2008) Minorities at Risk organizational behavior data and codebook version 9/2008 (http://www.cidcm.umd.edu/mar/data.asp).

Blattman

Christopher

Miguel

Edward

(2010) Civil war. Journal of Economic Literature 48(1): 3–57.

Bravo

Ana Bela Santos

Dias

Carlos Manuel Mendes

(2006) An empirical analysis of terrorism: Deprivation, Islamism and geopolitical factors. Defence and Peace Economics 17(4): 329–341.

Buis

Maarten L

(2010) Stata tip 87: Interpretation of interactions in non-linear models. Stata Journal 10(2): 305–308.

10.

Carter

David B

(2012) A blessing or a curse? State support for terrorist groups. International Organization 66(1): 129–151.

11.

Carter

David B

(2015) When terrorism is evidence of state success: Securing the state against territorial groups. Oxford Economic Papers 67(1): 116–132.

12.

Chenoweth

Erica

Cunningham

Kathleen Gallagher

(2013) Understanding nonviolent resistance: An introduction. Journal of Peace Research 50(3): 271–276.

13.

Chulov

Martin

(2009) Kurds lay claim to oil riches in Iraq as old hatreds flare. Guardian/Observer (http://www.theguardian.com/world/2009/jun/14/kurds-iraq-kirkuk-oil, accessed 9 May 2014).

14.

Collier

Paul

Hoeffler

Anke

(2004) Greed and grievance in civil war. Oxford Economic Papers 56(4): 563–595.

15.

Collier

Paul

Hoeffler

Anke

Rohner

Dominic

(2009) Beyond greed and grievance: Feasibility and civil war. Oxford Economic Papers: 61(1): 1–27.

16.

Conteh-Morgan

Earl

(2004) Collective Political Violence: An Introduction to the Theories and Cases of Violent Conflicts. New York: Routledge.

17.

Danzell

Orlandrew E

(2011) Political parties: When do they turn to terror? Journal of Conflict Resolution 55(1): 85–105.

18.

Daxecker

Ursala

Prins

Brandon

(2015) Financing rebellion through maritime piracy (politicalviolenceataglance.org, accessed 20 August 2015).

19.

de Soysa

Indra

Binningsbø

Helga Malmin

(2009) The devil’s excrement as social cement: Natural resources and political terror, 1980–2002. International Social Science Journal 57(s1): 21–32.

20.

de Soysa

Indra

Neumayer

Eric

(2007) Resource wealth and the risk of civil war onset: Results from a new dataset of natural resource rents, 1970–1999. Conflict Management and Peace Science 24(3): 201–218.

21.

Dreher

Axel

Fischer

Justina AV

(2010) Government decentralization as a disincentive for transnational terror? An empirical analysis. International Economic Review 51(4): 981–1002.

22.

Fearon

James D

Laitin

David D

(2003) Ethnicity, insurgency, and civil war. American Political Science Review 97(1): 75–90.

23.

Findley

Michael G

Young

Joseph K

(2012) Terrorism and civil war: A spatial and temporal approach to a conceptual problem. Perspectives on Politics 10(2): 285–305.

24.

Fortna

Virginia Page

(2015) Do terrorists win? The use of terrorism and civil war outcomes 1989–2009. International Organization 69(3): 519–556.

25.

Freedom House (2014) Freedom in the world: Country and territory ratings and status (http://www.freedomhouse.org/report-types/freedom-world, accessed 1 February 2016).

26.

Gaibulloev

Khusrav

Sandler

Todd

(2014) An empirical analysis of alternative ways that terrorist groups end. Public Choice 106(1): 25–44.

27.

Gassebner

Martin

Luechinger

Simon

(2011) Lock, stock, and barrel: A comprehensive assessment of the determinants of terror. Public Choice 149(3): 235–261.

28.

Horn

Myron K

(2010) Giant oil and gas fields of the world (https://edx.netl.doe.gov/dataset/aapg-datapages-giant-oil-and-gas-fields-of-the-world, accessed 8 May 2016).

29.

Humphreys

Macartan

(2005) Natural resources, conflict, and conflict resolution: Uncovering the mechanisms. Journal of Conflict Resolution 49(4): 508–537.

30.

Hunziker

Philipp M

Cederman

Lars-Erik

(2012) No extraction without representation: Petroleum production and ethnonationalist conflict (http://ssrn.com/abstract=2107176).

31.

Karl

Terry Lynn

(2007) Oil-led development: Social, political, and economic consequences. Encyclopedia of Energy 4: 661–672.

32.

Krueger

Alan B

Malečková

Jitka

(2003) Education, poverty and terrorism: Is there a causal connection? Journal of Economic Perspectives 17(4): 119–144.

33.

Kydd

Andrew H

Walter

Barbara F

(2006) The strategies of terrorism. International Security 31(1): 49–80.

34.

Lujala

Päivi

Rød

Jan Ketil

Thieme

Nadja

(2007) Fighting over oil: Introducing a new dataset. Conflict Management and Peace Science 24(3): 239–256.

35.

Marshall

Monty G

Jaggers

Keith

Gurr

Ted Robert

(2014) Polity IV Project, political regime characteristics and transitions, 1800–2013 (http://www.systemicpeace.org/inscrdata.html, accessed 12 February 2015).

36.

Meierrieks

Daniel

Krieger

Tim

(2014) The roots of Islamist armed struggle. Paper presented at the Terrorism and Policy Conference 2014, UTD (http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2453437).

37.

Mickolus

Edward F

Sandler

Todd

Murdock

Jean M

Flemming

Peter A

(2004) International terrorism: Attributes of terrorist events (ITERATE): 1968 – 2003, data codebook.

38.

Minorities at Risk Project (2009) Minorities at Risk Dataset. Center for International Development and Conflict Management (http://www.cidcm.umd.edu/mar, accessed 6 May 2013).

39.

Muller

Edward N

Seligson

Mitchell A

(1987) Inequality and insurgency. American Political Science Review 81(2): 425–452.

40.

Piazza

James A

(2011) Poverty, minority economic discrimination, and domestic terrorism. Journal of Peace Research 48(3): 339–353.

41.

Polo

Sara

Gleditsch

Kristian Skrede

(2014) Twisting arms and sending messages: Terrorist tactics in civil war. Paper presented at the Peace Science Society Workshop on Disaggregation in Terrorism Studies, University of Pennsylvania, PA.

42.

Rabe-Hesketh

Sophia

Skrondal

Anders

Pickles

Andrew

(2004) Generalized multilevel structural equation modelling. Psychometrika 69(2): 167–190.

43.

Regan

Patrick M

Norton

Daniel

(2005) Greed, grievance, and mobilization in civil wars. Journal of Conflict Resolution 49(3): 319–336.

44.

Ross

Jeffrey I

(1993) Structural causes of oppositional political terrorism: Towards a causal model. Journal of Peace Research 30(3): 317–329.

45.

Ross

Michael

(2004) How do natural resources influence civil war? Evidence from thirteen cases. International Organization 58(1): 35–67.

46.

Ross

Michael

(2013) Oil and gas data, 1932–2011 (https://dataverse.harvard.edu/dataset.xhtml?persistentId=hdl:1902.1/20369, accessed 8 May 2016).

47.

Sambanis

Nicholas

(2008) Terrorism and civil war. In: Keefer

Philip

Loayza

Norman

(eds) Terrorism, Economic Development, and Political Openness. Cambridge: Cambridge University Press.

48.

Sandler

Todd

(1995) On the relationship between democracy and terrorism. Terrorism and Political Violence 7(4): 1–9.

49.

Stanton

Jessica A

(2013) Terrorism in the context of civil war. Journal of Politics 75(4): 1009–1022.

50.

Tarrow

Sidney

(1998) Power in Movement: Social Movements, Collective Action and Politics. Cambridge: Cambridge University Press.

51.

Tavares

José

(2004) The open society assesses its enemies: Shocks, disasters and terrorist attacks. Journal of Monetary Economics 51(5): 1039–1070.

52.

Wimmer

Andreas

(2002) From subject to object of history: The Kurdish movement in northern Iraq since 1991. Kurdische Studien 2(1): 115–129.

53.

Wimmer

Andreas

Cederman

Lars-Erik

Min

Brian

(2009) Ethnic politics and armed conflict: A configurational analysis of a new global dataset. American Sociological Review 74(2): 316–337.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB