Chad Looney

Fluctuations in global petroleum prices can explain part of the variation in political violence in three African petrol states: Angola, Nigeria, and Algeria. As petrol states, all these countries have depended on imports of food and other basic commodities purchased with oil export receipts. Thus, when petrol prices and the economic performance of these countries fall, the states decrease imports to their citizenry. In turn, disaffected populations become open to recruitment by rebel leaders or political entrepreneurs who oppose the highly centralized and exclusive patrimonial African petrol states. Wealthy petrol states usually choose to meet violent opposition with violence because acceding to rebel demands only incites others to take up arms against the state. The process of violent rebellion and state countermeasures is shown using two regression models of violence during civil war periods and non-civil war periods. VIOLENCE IN AFRICAN PETROL STATES One can understand why violent opposition movements are not co-opted in economically weak less developed countries, but why do they continue to manifest themselves in various forms throughout resource-rich countries? Specifically, what factors cause violence in African petrol states to arise, and why do resource-rich governments choose to engage these movements with violence rather than simply co-opt them with state wealth? Using the dependent variable violence, this paper uses two main regression models—a civil war model and a non-civil war model—to assess causes of the variation in politically related violence in resource-rich petrol states. This paper finds that, other things being equal, changes in African petrol state violence are partially explained by variations in global petroleum prices. Specific evidence from three states—Angola, Nigeria and Algeria—demonstrates that when petrol states depend upon food commodities imported with oil monies, the African petrol state makes up revenue shortages by decreasing the importation of goods to its citizenry. This generally leads to the recruiting of disaffected and impoverished citizenry by political entrepreneurs who pursue rival political and economic goals. After reviewing petrol state literature and statistical/econometric-based civil war analyses, I offer theoretical explanations and statistical findings. Then, I support them with qualitative evidence and conclude that variations in global petroleum prices explain variations in violence in African petrol states. AFRICAN PETROL STATES: BACKGROUND ANALYSIS How does one recognize a petrol state? First, its export structure is predominantly based on oil. Three of the states on the panel described below have export structures ranging from 75% to 98% oil over the 26 years analyzed. Correlatively, gross domestic products (GDP) are linked directly to the global demand for oil, which was especially evident in the 1970s when petrol states’ GDPs were rapidly increasing (Karl, 1999, 6-7). Petrol states generally suffer from “Dutch Disease,” which occurs when oil revenues overvalue the exchange rate and make other sectors uncompetitive in global markets. As a result, the non-oil economy declines, leading to increasing reliance on oil. In addition, petrol states generally become rentier states receiving monies from outside of the country, which makes their populations unnecessary since they do not need to rely C. LOONEY: AFRICAN PETROL STATES -2on taxes as a revenue source (Karl, 1999, 5-6). Thus, the state becomes an isolated entity, completely autonomous from any pressures that do not threaten its vital input: petrol dollars. Because there are no democratic avenues for the populace to make grievances known, violence is usually the only means to draw attention to grievances; however, while petrol states may be autonomous from their populations, their structure and governing behaviors hinder their ability to quell violence. State structures are weak. In particular, African state structures are weak because generally, they are remnants from colonial times. European powers created state structures within their colonial territories in order to promote their interests while claiming they were “civilizing” barbaric peoples. However, the African states that Europeans established lacked essential democratic attributes such as checks, balances, and transparency (Berman, 1998, 314). Furthermore, in many instances after African states gained independence, local elites merely took over the role of the colonial power. Why would they change the state structures that were designed so efficaciously to accommodate corrupt governance? Since oil monies are poorly accounted for in these countries and oil companies do not disclose how much states receive, corruption abounds, most obviously manifested in the clandestine pocketing of oil revenues, but more fully demonstrated in government policies (Karl, 2003, 26). For instance, domestic investments occur, but are generally directed at lavish projects that predominantly benefit those who conceive them. This altering of government policy to benefit individuals, or rent-seeking, becomes the political norm and the number of rent-seekers, or patrons, increases during booms along with new grievances about current distributions (Karl, 2003, 19, 21, 24). By contrast, busts obviate petrol dollars while patrons of the governments still expect revenues to continue, which can drastically affect the stability of the state (Karl, 2003, 19, 21, 24). Distributive policies keep these states stable. Money flows into the pockets of regional leaders and central government coteries that are essential to keeping dictators in power. As a result, decreases in oil prices affect the state before the general population. Thus, when a state’s networks become agitated, it may resort to brazen maneuvers such as denying its citizens basic goods. African petrol states do, however, have International Financial Institutions (IFI) to fall back on in a crisis. IFIs such as the World Bank Group, and Export Credit Agencies (ECA) play a crucial role in contributing to short-term regime stability in African petrol states by promoting the principle of comparative advantage. ECAs are specifically designed to increase exports from lesserdeveloped countries to the developed world (Karl, 2003, 7). ECAs finance projects in high risk localities and are not, like the World Bank, restricted by socio-environmental regulations (Karl, 2003, 15-16). For example, while U.S. President Ronald Reagan was supporting UNITA rebels and trying to oust the MPLA government during the Angolan Civil War, U.S. ECAs (ExportImport Bank and the Overseas Private Investment Corporation) were contributing to MPLA’s hold on Luanda (Karl, 2003, 17). It is obvious that without Angola’s resource wealth, financial institutions would not have considered venturing into a country so ridden with civil war. THE EVOLUTION OF STATISTICAL STUDIES Statistical variables need to be derived and interpreted through some sort of causal template. In 1998, Collier and Hoeffler ran two regression models on the determinants of occurrence and duration of civil wars in all countries from 1960-1992. They derived their variables from a “rebel utility function,” specifying the theoretical cost-benefit analysis guiding a political leader’s decision to rebel. They interpreted their regression results through the analytical framework that the rebel utility function provided. While an exact presentation of this function is beyond the scope of this paper, what follows is a simplified explanation derived from their landmark work, “On Economic Causes of Civil War.” Their results not only exemplify how statistical results pertaining to violence can be interpreted, but also demonstrate the link between primary commodity exports (e.g., oil) and civil war. TCNJ JOURNAL OF STUDENT SCHOLARSHIP VOLUME X APRIL, 2008 -3“The objective of rebellion is either to capture the state or secede from it. In general, the incentive for rebellion is the product of the probability of victory, which depends upon the capacity of the government to defend itself” (Collier and Hoeffler, 1998, 564). If the taxable capacity within a state increases, military capacity also increases. This decreases the probability of conflict. Per capita income serves as a proxy for taxable capacity. If secession is the motivation, then natural resources may be in the calculus to initiate violence; primary commodity exports as percent of GDP proxy natural resources. Ethno-lingual divisions are also included in the models. From the Singer and Small data on civil wars from 1960-1992, Collier and Hoeffler used a probit model to assess what variables increased the probability of civil war and a tobit model to assess the variables that explain the duration of civil wars. Their tobit model incorporates more variables, so it is better at elucidating significance, but unable to show precise effect (Collier and Hoeffler, 1998, 568). The probit model tells us that, with other things equal, for each dollar increase in per capita income, the probability of a civil war decreases by 0.1 %. The Ethno-lingual fractionalization index was insignificant in explaining the occurrence of civil war, and the primary commodity exports effect is nonmonotonic, which means that “until extremely high levels of primary commodity export dependence is reached (after about 26%), the risk of civil war declines” (Collier and Hoeffler, 2002, 16-17). Collier and Hoeffler’s study provides empirical evidence that violent opposition towards government is related to economic factors, which corroborates this paper’s thesis. Low per capita income in petrol states is directly correlated to the Dutch Disease and distributive/rent-seeking policies. It should be noted, however, that Angola, Nigeria, and Algeria all have primary commodity export structures in the upper 90 percent, which suggests that petrol states are not as susceptible to civil war as lesser-developed countries. In “On the Incidence of Civil War in Africa” (2002), Collier and Hoeffler focused their analysis to isolate Africa from the rest of the developing world. The study was initiated following the Stockholm International Peace Research Institute’s (SIPRI) assessment in 1999 that “Africa is the most conflict ridden continent of the world and the only region in which the number of armed conflicts is on the increase” (Collier and Hoeffler, 2002, 13). In this study, their global panel set was expanded from their 1998 study by incorporating the years 1992 to 1999. The panel is composed of “161 countries for eight five-year periods: 1960-1964, 1965-1969,...,1995-1999.” Collier and Hoeffler focused on five-year periods leading up to civil wars and had 750 observations of which “46 were characterized by civil wars.” The ratio of primary commodity exports to GDP is only slightly higher with the average being 19% in Africa and 17% in the rest of the developing world (Collier and Hoeffler, 2002, 21). However, per capita income is nearly double in the rest of the developing world compared to Africa and GDP growth disparities are even greater (Collier and Hoeffler, 2002, 24-25). This finding corroborates this paper’s argument by illustrating that disenfranchisement of citizenry leads to violence. What greater disenfranchisement can there be than denying citizenry food? This paper focuses on more than civil war, which is not the only serious violence that occurs in Africa. Collier and Hoeffler drew on Singer and Small’s definition of civil war, according to which at least 1,000 battle deaths must occur, with the stronger forces sustaining a minimum of 5% of the casualties (Collier and Hoeffler, 1998, 567-568). Using this definition as a basis for data collection obviously excludes many acts of political violence. Collier and Hoeffler’s studies do not focus on occasional incidents of violence, and their dependent variables are based only on whether or not a war started and how many years it continued. Because violence is an amorphous phenomenon, this study focuses on a wider range of acts of recorded violence dependent upon different conditions. “Greed and Grievance” literature explores whether conflict arises for economic (greed) or political reasons (grievances) while Collier and Hoeffler’s work chiefly focuses on economic incentives. For instance, their regression model is interpreted through a cost-benefit analytical framework (a rebel utility function). Recent research on this topic by others suggests that civil C. LOONEY: AFRICAN PETROL STATES -4wars may start for political reasons then turn economic (Cater, 2003, 28-29). Although costbenefit frameworks can provide a basis for statistical analysis on the economics of civil war, quantifying grievances is extremely difficult. Even Collier and Hoeffler had to rely heavily on proxies and then stretch their inferences in order to interpret their regression through an economic analytical framework. The following regression analyses, however, focus specifically on African petrol states and all annually recorded political violence, including acts during and not during civil wars from 1980 to 2006. Focusing on all recorded acts of political violence may more satisfactorily explain variations in violence. Because what causes changes in severity during the course of a quarter century conflict such as the Angolan Civil War? Where and how do UNITA, MEND, and others recruit their followers and what makes the citizenry vulnerable to recruitment? What variables are most conducive to violence in an African petrol state? By treating each violent act as its own entity, this study hopes to offer a new quantitative perspective on violence in African petrol states.