ABSTRACT
Over the past decades, extractive booms have been more of a curse than a blessing for fragile, resource-rich states such as the Democratic Republic of the Congo, Equatorial Guinea, or Nigeria. The main dynamics at play involve Dutch disease, rentier states, rentseeking, and armed conflict (e.g. Mehlum et al., 2006; Carbonnier, 2007, 2011; Collier and Goderis, 2008). Despite a rich literature, the relationship between resource extraction and development remains a controversial issue, and there is little consensus on the transmission channels from the exploitation of non-renewable natural resources to sustainable development outcomes. Most of the empirical literature on the resource curse follows Sachs and Warner by
assessing the impact of resource export dependence on development outcomes in terms of economic growth (Sachs and Warner, 1997). The exploitation of fossil fuels and minerals translates into immediate economic growth but also into the depletion of the natural capital base of a country. The outcome can be regarded as sustainable only if the extractive revenues are reinvested in other forms of capital in a way that increases, or at least maintains, the overall capital stock. In the case of exhaustible resources in low-income countries, this is obviously all the more important as the extractive rent will die off at some point, depend-and the extraction pace (Stevens, 2011). This paper captures
DEFENSE SPENDING, NATURAL RESOURCES, AND CONFLICT
the linkages between resource extraction and sustainability by incorporating natural capital in calculating national income and wealth (Hamilton, 2004, p. 31). Genuine savings serve as a sustainability indicator (Hartwick, 1977; Perman et al., 2003, p. 89; World Bank, 2006), whereby national wealth is based on a comprehensive list of assets which includes produced, natural and intangible capital, i.e. human and social capital (Goodwin, 2007). Resource extraction corresponds to an instant decumulation of genuine savings as it reduces the natural capital stock. Genuine savings can nonetheless be maintained or even grow with compensatory capital accumulation in physical and human capital. But since the capital approach rests on the unrealistic assumption that natural capital can be fully substituted by produced and human capital, it is referred to as an indicator of ‘weak sustainability’ (Everett and Wilks, 1999; Thiry and Cassiers, 2010). The World Bank has calculated genuine savings or ANS for 209 countries from 1970
onwards using the following formula (Bolt et al., 2002):
ANSit ¼ ðGSit � DEPCitÞ|fflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflffl} Net Savings
þEEit � RRDit � CDit; (1)
where ANSit is the adjusted net savings of country i at time t. It is composed of gross national savings GSit net of the depreciation of produced capital DEPCit, augmented by (non-fixed capital) expenditure on education EEit and reduced by the rents from depletion of natural capital1 and damages from carbon dioxide emissions CDit. In words, adjusted net savings measure the true rate of savings after taking into account investments in human capital, depletion of natural resources, and damage caused by pollution. Thus, increases in total wealth are measured by positive adjusted net savings rates indicating sustainability in the economy. Therefore, ANS is a suitable policy indicator as it underpins the need for domestic savings, and at the same time includes environment and resource management in a framework that finance and development planning ministries can understand. Moreover, growth at the cost of resource and environmental depletion is captured by decreased adjusted net savings rates. ANS data have recently been used more often in the empirical literature on the resource
curse (e.g. Stoever, 2011; Barbier, 2010; Van der Ploeg and Poelhekke, 2010; Van der Ploeg 2010; Dietz et al., 2007). Existing studies tend to consider ANS as an outcome variable expressed as a percentage of national income. This raises an endogeneity issue, especially when national income appears as an explanatory variable in the empirical model. Instead, we consider ANS per capita. We further follow Barbier (2010) in excluding CO2 damage from the ANS calculation since it suffers from important methodological flaws. Our empirical model presented in section 4 seeks to assess the impact of resource
dependence, armed conflict, and violence on ANS per capita, with a specific focus on human capital. The conceptual framework draws on a range of theoretical considerations and empirical studies. The negative impact of resource dependence on economic growth in fragile states draws on several prominent explanations including in particular rentier-state and rent-seeking dynamics together with macroeconomic issues associated with Dutch disease (see e.g. Sachs and Warner (1997), Ross (1999)). Commodity price volatility and the ensuing voracity effect have further been singled out as additional variables explaining the so-called resource curse. While these transmission channels have been tested with respect to GDP growth, there is yet little evidence concerning the impact on genuine savings. We seek
DEFENSE SPENDING, NATURAL RESOURCES, AND CONFLICT
to address this gap and focus on a period of rising commodity prices, which is expected to increase natural capital depletion, i.e. to reduce genuine savings, ceteris paribus. This can be (partly) offset when extractive revenues are reinvested in physical and human capital. A series of studies from the mid-1990s onward show that armed conflict associated with
resource dependence are important factors in explaining poor economic growth in resourcerich and yet weak states. Often based on extraction theory, they posit that primary commodities can be either taxed or looted. Rulers and rebels both seek to extract maximal revenue under a set of constraints including transaction costs. Rebels seek to capture the state or secede from it in an attempt to rule over the portion of territory where the resources are located. The incentive to rebel diminishes with the state’s capacity to defend itself and the opportunity cost of joining the rebellion. It increases with the probability of successful rebellion and the prize associated with seizing power. Taking GDP per capita, GDP growth, and primary commodity dependence as proxies, the link between natural resources, and civil war in developing countries has notably been established by Collier (1999) and Collier and Hoeffler (2002, 2004). Following Collier, we consider four channels through which armed conflict may negatively affect ANS to an even greater extent than GDP. Civil war causes the destruction of physical and human capital such as infrastructure and the labor force through killing and displacement. It further disrupts social capital and causes political instability that results in less private investment in physical capital. Civil war may also divert public expenditures away from education and investment in fixed capital. Moreover, dissaving further destroys the capital stock, for instance when the elite put their capital in a safer place outside the country. There are thus convincing arguments to establish a negative link between armed conflict and genuine savings. The linkage between armed conflict and resource dependence remains more contested 15
years after Collier and Hoeffler’s initial work on the economic causes of civil war. For Brunnschweiler and Bulte (2008), there is no evidence of a causal relationship between resource dependence and armed conflict. Resource abundance per se is even associated with a weaker probability of civil war. Van der Ploeg and Rohner (2012) draw another conclusion by treating resource extraction as endogenous on the basis that fighting affects the extraction pace. Other authors find that resource dependence tends to prolong rather than cause armed conflicts (Di John, 2007), and underline that the location and type of resources matter a great deal (see e.g. Lujala, 2009). There is general agreement that countries endowed with strong institutions and good governance face a lower risk of armed conflict associated with resource dependence than weak states, notwithstanding issues of circular causality (Carbonnier et al., 2011). This is echoed by Teorell (2009) who finds evidence supporting the hypothesis that civil war is associated with the lack of impartial institutions. Relative deprivation remains an important driver of armed conflict associated with resource dependence, pointing to the link between resource dependence, exclusion, and conflict (Billon, 2001). We posit ANS per capita to be affected not solely by natural capital depletion but also by a lack of investment in human and physical capital associated with armed conflict. We consequently explore the impact of resource dependence and armed conflict on education expenditure to assess to what extent the overall negative impact on ANS can be explained by human capital. Finally, recent research on the local dynamics of resource extraction in Latin America
has highlighted the prevalence of so-called ‘socio-environmental’ conflicts pitching host communities against extractive firms. Disputes often revolve over land property issues, water quality and quantity, waste water treatment, and job opportunities. Such disputes tend to turn violent in institutional settings where appropriate mechanisms for peaceful conflict (Vasquez, 2010; Arellano-Yanguas, 2008). Yet, these conflicts
