Essays on Conflict, Institutions, and Ethnic Diversity

Abstract: The thesis consists of five self-contained papers. Paper 1: Social Capital vs Institutions in the Growth Process Is social capital a substitute or a complement to formal institutions for achieving economic growth? Research on the impacts of social capital and formal institutions on economic development have so far mainly emerged as two distinct …elds. In the social capital literature, trust, networks, social norms, and associational activity are believed to be central aspects of successful economies. Although micro studies suggest that social capital has a larger e¤ect on economic performance when formal institutions are weak, this has not been con…rmed at the macro level. In the institutional literature, it is emphasized how formal institutions such as those regulating the strength of property rights, the constraints against the executive, and the power of courts, are fundamental determinants of long-run growth (North 1990, Acemoglu et al. 2001). These studies, however, never attempt to quantify the e¤ect of informal institutions such as inter- personal trust. Based on the micro evidence, we outline an investment game between a producer and a lender in an incomplete-contracts setting. The key insight from the model is that social capital may have its greatest positive impact on the total monetary surplus from the game (economic growth) at lower levels of institutional development, and that the positive impact eventually vanishes if institutions become strong enough. This basic prediction about substitution …nds support in a cross-country growth regression the marginal impact of our proxy for social capital (interpersonal trust) decreases with the quality of formal institutions. This implies that attempts at building social capital create, if successful, a pro-growth potential for countries with bad institutions. Paper 2: The Roots of Ethnic Diversity The level of ethnic diversity is believed to have consequences for economic and political development. Accepting this observation naturally leads to the question: Why are some countries more ethnically fractionalized than others? For instance, why is the probability that two randomly chosen individuals belong to di¤erent ethnic groups roughly 93 percent in Uganda but only 0.2 percent in South Korea? In the paper, we explore the two main hypotheses regarding the formation of ethnic identities. The constructivist”view is that ethnic identi…cations are socially constructed phenomenon appearing during modernity (Gellner 1983, Tilly 1992). The evolutionary view contends that ethnic divisions have deep roots in history and ecology and should be analyzed in an evolutionary framework. Ethnic identification is here regarded as a x natural and evolutionarily successful behavior that has existed throughout history. The process of evolution is tied to the geographical context, and in the paper we discuss the implications for ethnic diversity from a number of stylized ecological facts. We develop a formal model where ethnic groups endogenously emerge among periph- eral populations in response to an insu¢ cient supply of public goods. A key prediction is that the current level of ethnic fractionalization in a given area should be positively correlated with the antiquity of human settlement. Our empirical analysis introduces the historical duration of human settlements for all countries in the world. The dating is based on research in genetics, archeology, climatology and on fossils, as synthesized by primarily Oppenheimer (2003). The theoretical prediction of a positive e¤ect of the historical duration of human settlements on ethnic diversity receives strong empirical support, and there are clear indications that ethnic diversity is higher where geographical conditions have favoured isolation, and lower where early civilization proved more successful, and where the state was stronger during the modern nation-state era. Hence, a genuine understanding of ethnic diversity requires a synthesis of evolutionary and constructivist arguments. Our results have implications for how social scientists investigate the e¤ects of eth- nic diversity. An often employed method for assessing the e¤ect of ethnic diversity on economic and political development has been to treat ethnic diversity as an exogenous determinant. Since a stronger state in the nation-state era is associated with having less ethnic diversity, and there is a positive correlation between indicators of this state strength and several indicators of economic and political development, the negative coe¢ cient on ethnicity obtained in these regressions could re‡ect an omitted variable bias. Paper 3: The Causal E¤ects of Ethnic Diversity: An Instrumental Variables Approach High levels of ethnic diversity have been linked to various poor economic and political outcomes, e.g., lower income levels, poor economic growth, more corruption, and a lower provision of public goods (Easterly and Levine 1997, La Porta et al. 1999). The standard approach in this literature has been to treat ethnic diversity as if it were exogenous to economic development, but that is a misspeci…cation. The historical literature has documented how populations in more developed countries have become more homogenous over time, and increasingly so during the last couple of centuries, through a combination of deliberate homogenizing e¤orts and endogenous processes (Gellner 1983, Tilly 1992). Recent research has found that ethnic diversity is determined both by historical forces and by geographical factors: Ethnic diversity is higher in countries with a longer duration of human settlement, and in countries that have a naturally fragmented geography, that lie closer to the equator, and that have had low levels of territorial state capacity during the modern era. xi In the paper we discuss how previous studies on ethnic diversity and long-run devel- opment may have obtained biased estimates due to omitted variables, simultaneity, or measurement error, but also that the use of instrumental variables allows us to deal with exactly these problems. Our main instruments capture the historical duration of human settlements, the degree of geographical fragmentation, and the number of years since the date of independence. With these at hand, we …nd that high levels ethnic diversity is associated with lower income levels, poor economic growth, more corruption, and poor provision of public goods, and that results obtained in OLS may underestimate the true e¤ects. While previous studies have shown signi…cant partial correlations between ethnic diversity and economic outcomes, the present paper demonstrates that there indeed are causal e¤ects of ethnic diversity. We also …nd that the e¤ects of ethnic diversity and property rights institutions on economic development among former European colonies can be separated from each other. This suggests that countries that have problems due to high levels of ethnic diversity could alleviate these problems by improving the quality of their formal institutions. On a more general level, the results presented in the paper promise that an acceptance of the endogenous nature of ethnic diversity does not preclude meaningful empirical analyses of the long-run e¤ects of ethnic diversity. Paper 4: Nationalism and Government E¤ectiveness Nation-building, which generally refers to a process of unifying the population in a country by constructing a national unity, is believed to have positive e¤ects on aggregate performance, and has been proposed as a possible remedy against problems associated with high levels of ethnic fractionalization (Miguel 2004). However, systematic empirical evidence that the creation of a national unity is a worthwhile policy is still largely absent, and nationalism, an indicator of successful nation-building, has been empirically linked to protectionism and intolerance, which suggests that dismal performance is a more likely outcome. Furthermore, there is an obvious problem with the idea that the unity of a country’s population can be enhanced by encouraging nationalism –a national identity is created in relation to other national identities, and for there to be an “us”there has to be a “them.” The paper investigates whether nationalism a¤ects the ability of governments to e¤ec- tively formulate and implement good policies, i.e., government e¤ectiveness, and whether it mitigates the negative e¤ects of ethnic fractionalization, or is associated with less trade openness. We discuss how nationalism may have a positive e¤ects, as it can increase in- group altruism, trustworthiness, and state authority, and how it may have negative e¤ects, as it can breed prejudice, out-group animosity, and skepticism of new ideas, implementa- xii tion techniques, and goods, if these are not of national origin or are not considered to be in line with national traditions. We hypothesize that the positive e¤ects will dominate at low levels of nationalism but that the negative e¤ects will dominate at higher levels of nationalism. The empirical analysis con…rms that nationalism has an inverted U-shaped e¤ect on government e¤ectiveness, and also shows that this e¤ect does not capture the in‡uence of factors such as income, economic growth, democracy, and income inequality. Comfort- ingly, the qualitative result is the same also when we instrument for nationalism, with instruments that represent a number of historical and cultural circumstances. Further- more, nationalism can mitigate the negative e¤ects of ethnic fractionalization in former colonies, but has no clear e¤ect on trade openness. Taken seriously, the results suggest that most countries already have too nationalistic populations, and probably would function better if these sentiments where downplayed. Paper 5: Earthquakes and Civil War There are two diametrically opposing views in the literature on natural disasters and violent con‡ict. According to the …rst view, natural disasters can contribute to de-escalate con‡ict, as previous disagreements seem relatively unimportant. This is view shared by several relief organizations and policy makers (WBGU 2008, Brancati 2007). The second view is that natural disasters make violent con‡ict more likely, and this view is supported by most systematic empirical studies. Nel and Righarts (2008) …nd that natural disasters in general increase the probability of onset of civil war, and Brancati (2007) …nds that earthquakes are positively associated with the incidence of civil war, and argues that earthquakes of higher magnitude have a stronger e¤ect. Consider the e¤ects of the great tsunami in South-East Asia in 2004. The tsunami is believed to have exacerbated the con‡ict in Sri Lanka, but in Aceh, Indonesia, the …ghting came to an end. It appears that the con‡ict de-escalated where the tsunami had its most severe e¤ects and escalated where the e¤ects were less severe. In the paper we ask whether more destructive natural disasters are associated with a higher or a lower risk of violent con‡ict. We take the argument that natural disasters can de-escalate existing con‡icts seriously, and investigate the e¤ects not only on the incidence or onset of con‡ict, but also on the termination of con‡ict. We develop an economic model of rebellion in the wake of a natural disaster. It shows how the moderate destruction caused by moderate natural disasters can make rebellion feasible, by lowering the opportunity cost of potential recruits. More intensive destruction means that the material payo¤ in the event of a victorious rebellion is lower. Taken together, the model predicts that violent con‡ict may be more likely after moderate disasters and less likely after very strong disasters. xiii The empirical results are well in line with the theoretical predictions. We employ an exhaustive dataset on earthquakes from 1947 to 2001, and develop a new set of exogenous indicators of the size of earthquakes, e.g., the seismic energy released by earthquakes. We …nd that earthquakes a¤ect both the onset and termination of con‡icts and, as a result, the incidence of con‡ict. The size of an earthquake is of fundamental importance, as are the social conditions in the area surrounding the epicenter. The association between earthquakes and the incidence of civil war can be explained by three e¤ects: (i) Moderate earthquakes increase the risk that new con‡icts are started, (ii) strong earthquakes make it less likely that new con‡icts are started, and (iii) strong earthquakes make the termination of existing con‡icts more likely. As such, the paper is the …rst systematic study to establish that a natural disaster can give the impetus needed to end existing con‡icts and prevent new violent con‡icts from emerging.