Immigration and Civil War - Should You Be Worried?

By Alex Nowrasteh

Some modern immigration restrictionists are arguing that immigration will cause a civil war in the United States or other countries unless it is curtailed or radically altered. Reihan Salam’s recent book Melting Pot or Civil War? is the most glaring example. In addition to the title, he points to immigration being a problem in and of itself that also increases the severity of other issues dividing American society. The result could be a civil war. Salam writes that “[t]he divisions that define this moment in American history are not yet as worrisome as those that led to the Civil War or the bloody battles putting workers against industrialists at the dawn of the last century … [n]evertheless, it is hard to shake the feeling that our luck might soon run out.” 

David Frum hints at the possibility of a racialized civil war by arguing that young white voters are also worried about immigrants bringing demographic changes, leading to the rise (again) of nationalist political parties in Europe that are running on platforms to restrict immigration. The problems of immigration are apparently so well-known that even countries that are not the destinations for immigrants, such as Hungary and Poland, are turning to nationalist politicians – sometimes. The notion of a racialized civil war caused by immigrants or as a reaction to them has even entered popular culture. Michel Houellebecq’s novel Submission, which is about a Muslim political party winning a presidential election in France in the near future and pursuing policies to turn that nation into an Islamic theocracy, includes conversations between characters about a civil war in France. 

But could immigration cause a civil war? Since it’s been raised so often, I think it’s an important issue to address. Much of my research over the last several years is about how immigrants affect the economic and political institutions of the countries where they settle, finding positive or null effects. But if immigrants did cause civil wars, which are usually the deadliest types of wars, then that would be a very large cost that we’d need to consider. 

Fortunately, there is a large set of peer-reviewed literature on the causes of civil wars that should diminish the fears of immigration restrictionists who think the United States or other Western countries could sink into racialized civil wars due to immigration. According to a wonderful review in the Journal of Economic Literature by Christopher Blattman and Edward Miguel, civil wars are more likely to occur in countries that are poor, are subject to negative income shocks, have weak state institutions, have sparsely populated peripheral regions, and possess mountains. Modern developed countries do not possess most of those features. 

Civil wars likely have causes on both the micro level and on the macro level. On the micro level, a theoretical multiplayer game model developed by Joan Esteban and Debraj Ray where each player has imperfect information about the costs of conflict shows that Pareto-improving social decision making becomes impossible and conflict is certain to ensue with four or more players. Based on additional research by Ray (cited here), conflict may be unavoidable even with enforceable contracts between coalitions. Thus, in a situation where society divides along multiple lines – by geography, religion, race, ethnicity, or economic class – it may be impossible for the government to arrange a set of transfers or policies that prevent conflicts among all divisions simultaneously. If immigration increases the number of divisions in society, then it is theoretically possible that it would increase the chance of civil war according to these models.

On the macro level, a country’s degree of ethnic fractionalization reduces the chance of civil war, income inequality has no effect, and democratic government is not a significant predictor of conflict risk conditional on the existence of poverty, negative income shocks, weak state institutions, sparsely populated peripheral regions, or mountains. The finding that more ethnic fractionalization does not lead to civil war seems counter-intuitive, but that’s due to observer bias.  Economist Paul Collier observed that “[c]onflicts in ethnically diverse countries may be ethnically patterned without being ethnically caused. International media coverage of civil wars often focuses on history and ethnicity because rebel leaders adopt this sort of discourse.  Grievances are to a rebel organization what image is to a business. The rebel group needs to stimulate a sense of collective grievance to build cohesion in its army and to attract funding from its diaspora living in rich countries.” 

Furthermore, republican institutions reduce the chance of civil war as they help to enforce intertemporal commitments and lower transaction costs. As a result, immigrants would be more likely to cause a civil war if they weakened republican political institutions, but there is no evidence of that, no evidence that democratic political institutions attract immigrants (independent of other factors), and plenty of evidence that immigrants move between countries with similar levels of democracy.

The exception to this is that refugee flows increase the chance of civil wars under very specific circumstances that do not exist in developed countries. From 1951 through 2001, Salehyan and Gleditsch found that the baseline chance of a country fighting a civil war if there were no refugees present and no civil war in a neighboring country was about 3.5 percent per year. That percentage rose to 4.5 percent per year if the ratio of refugees to the population goes up to the global average. A similar increase in refugees combined with a civil war in a neighboring country further increased the annual chance of having a civil war to 6.2 percent. From zero refugees and no neighboring civil war to an average number of refugees and a neighboring civil war, the chance of having a civil war increased by 2.7 percentage points or 77 percent. Stronger democratic governments and more interregional trade diminish the chance of civil war even in the presence of civil war in a neighboring country and refugee flows who are members of cross-border ethnic groups. All of the civil wars during the 1951 through 2001 period that Salehyan and Gledistch considered occurred in poor countries with weak governing institutions. 

Since a civil war has never been caused by immigrants in a developed country and refugee flows only increase the chance of civil war under very specific circumstances in developing countries from 1951 through 2001, immigration restrictionists should feel relieved. On the other hand, the extreme downside risk of electing nationalist governments in the developed world is very high. The extreme downside risk of civil war caused by refugee or immigrant inflows into a developed country has historically been zero. Both or either of these findings could change in the future and there is a possibility that immigration could lead to the election of nationalists, but nationalism is the far greater threat for those concerned about civil war or other radical shifts that could damage our civilization.  In either case, managing the nationalist reaction or immigration seems easier and more likely to succeed than acceding to their policy demands before they are elected.

The above originally appeared at Cato.org