Posner (2004): Based on a natural experimental setting in Africa, Posner was able to find that different ethnic groups will fight each other if they are large enough for politicians to mobilize and exploit. Otherwise, because ethnicity is an ever-changing definition, it shouldn't really matter.
Habyarimana (I forgot the year
): It is common knowledge that ethnically diverse areas tend to have less resources devoted to public benefits. Using a set of experiments done in Uganda, Habyarimana found that a subset of the population will, when offered $1000 and the chance to anonymously distribute the money between them and a member of the same ethnicity and a member of a different ethnicity, would keep all the money to themselves. However, in the same setting without anonymity, they will share some of the money, but only with their co-ethnics. They will do the bare minimum to prevent social punishment. People of the same ethnicity will more likely have connected social networks and such to avoid punishment, these "greedy" people avoid punishment but discriminate against those who cannot punish them. Not everyone is like this, but the presence of a group of people like this can still hinder the distribution of public goods.
Fearon and Laitin (2003): Two common theories for the onset of civil war are a grievance-based theory and economic opportunity theory. The former suggests that people will be motivated to fight when discrimination or something of the same vein occurs. The latter contests that when rebelling against the federal government becomes economically beneficial, people will do so. This study used the available civil war data of nations to find what factors make fighting the government feasible. From their studies, they concluded that GDP/capita is by far the strongest predictor of civil war onset.
Terrain also plays a role; mountainous terrain also corresponds to greater likelihood of civil war onset, as does political instability, high population and the presence of substantial oil exports. However they did not find grievances to be relevant in this case.
Cederman (2011): The previous study examined how these factors influence people on the individual level. Cederman's work examined what leads to civil war at the group level. Previous efforts looked at universal political grievances, such as political terror or repression, and measures of the potential for ethnic conflict, measured by ethno-linguistic fractionalization, and balances of group size. It doesn’t capture the conditions that groups, rather than individuals, experience. Looking at horizontal inequalities in politics and economics, Cederman concluded that at the group level, grievances do matter. However, this does not mean that individual factors do not matter. It just means that we have to consider more factors when trying to predict/understand civil war.
Walter (1994): Civil wars tend to last a long time. This is due to the inability to easily reconcile the two opposing sides. This is the "credible commitment" problem. Once a bargain is reached, the power to enforce that bargain goes to one of the sides involved in combat, who has no (or weak) incentives to comply, and strong incentives to destroy their former enemy. Walter concludes that a non-partial third party can guarantee a successful conclusion to the war. However, in real life, it is hard to find an unbiased group willing to make negotiations.
Ross (I don't remember the year again
): This study looked at the "resource curse". Areas with abundant natural resources are more likely to be engaged in civil war. Resource-rich states tend to have two severe (and interrelated) problems: Lower rates of economic growth compared to resource-poor states and a greater proportion of their population living in poverty. That is because the economy depends on exporting the resource and less on taxing the population and such. So the government does not rely on the people as much, so they don't care as much for the well-being of the public. In other words, they have little to no incentive to invest in their human capital. In specific regions of the country where there are many resources, secession in those areas are more likely. And where natural resources are concentrated in the territory of a given ethnic/cultural/religious group, there may be both economic (split the dividends with fewer people) and grievance-based motivations for secession for another group. Thus, the presence of natural resources can be considered indicators of governmental negligence and poverty, which in turn lead to civil war.