Notes - Chapter 4: Political Violence

A prince ought to have two fears, one from within, on account of his subjects, the other from without, on account of external powers.

- Niccolo Machiavelli, 1513.

Niccolo Machiavelli. Wikimedia Commons

Classic theoretical models of conflict, such as those in Grossman (1991) and Skaperdas (1992), have been applied to understanding civil war. In common with the model developed in this chapter, these authors see conflict as the outcome of an equilibrium process in which the incentives of the various parties are modeled explicitly. Those incentives arise from the technology of conflict, the pto_be_inserteds of the protagonists, and the underlying economic constraints. Grossman (1992) makes the point that foreign aid may increase the propensity of civil conflict for the same reason as the one stressed in this chapter. For excellent reviews of the theoretical literature, see Blattman and Miguel (2009) on general issues and Garfinkel and Skaperdas (2007) on the research that uses contest functions - Dixit (1987) and Skaperdas (1996) survey the use of contest functions more generally. Aslaksen and Torvik (2006), Azam (2002), Caselli (2006) and Chassang and Padro i Miquel (2009) are more recent theoretical contributions which take somewhat different approaches. Azam (2005) discusses state types in Africa where the distinction between one-sided and two-sided political violence is important. Parker (1988) is a classic on the role of military technology and its historical impact on the organization of the state.

Much progress has been made in this theoretical research, but most of it has been pursued separately from the empirical literature and the models have not generally been formulated with empirical testing in mind. Fearon (2008) is an exception, but he follows a rather different modeling approach to the one adopted in this chapter. Another exception is the literature on polarization and conflict, where the theory by Esteban and Ray (1999) and has both been informing and informed by empirical work such as Reynal-Querol (2002) and Montalvo and Reynal-Querol (2005).

In recent years, a large empirical literature has emerged, which looks at civil war its determinants. This literature is reviewed by Elbadawi and Sambanis (2002) and Blattman and Miguel (2009). A robust finding in this literature is that poor countries are disproportionately involved in civil war, even though the direction of causation may be difficult to establish. But the interpretation of this correlation is open to debate. Fearon and Laitin (2003) see it as reflecting limited state capacity to put down rebellions, while Collier and Hoeffler (2004) see it as a reflection of the lower opportunity cost of fighting when incomes are low.

There is also considerable debate about other prospective drivers of civil war, such as ethnic divisions and political institutions. When it comes to natural resources, results diverge as well. While some researchers have found natural resources to significantly raise the probability of onset and/or duration of civil war, other researchers have failed to find such an effect (see Ross, 2004 for a review of the research on this topic). However, most empirical researchers would probably agree that there is a significant cross-sectional correlation between oil dependence and civil war. A few papers, like Drury and Olson (1998) and Nel and Righarts (2008) look at the empirical links between natural disasters and civil conflict, while very recent studies by de Ree and Nillesen (2009) and Nunn and Qian (2010) explore the links between aid and civil war.

A smaller set of studies focus on variation in conflict within countries. For example, Deininger (2003) uses community-level data from Uganda finding that scarcity of economic opportunities (proxied by infrastructure) and the presence of cash crops are correlated with the civil strife. Another within country study by Dube and Vargas (2008) builds on the theoretical framework developed by dal Bó and dal Bó (2006) to explain the incidence of conflict within Colombian municipalities, exploiting the time-series variation in coffee and oil prices.

There is also a considerable literature in political science, which aims at finding economic, political and social variables that explain government repression and violations of human rights. This literature, surveyed in Davenport (2007), is almost exclusively empirical. The literatures on civil war and government repression have largely been developed in parallel, without too much contact with each other, even though civil war sometimes appear as a "right-hand-side" variable in empirical studies of repression. Azam and Hoeffler (2002) show how violence against civilians (repression) can be effective at deterring rebellion and applies this to data on refugees.