Ethnic Conflict: Questions and Answers (coauthored w/ Dr. Patrick James)
Nov. 26 2016 Ethnopolitics - download link (gated)

This article implements a diagrammatic exposition based on systemism to identify what is known so far about ethnic conflict, along with the most and least useful developments.  On the positive side, it is possible to point out greater awareness of the nexus of international and domestic conflict.  Overreaction against primordialism and in favor of constructivism stands out on the negative side.  Assessment of the field guided by systemism argues in favor of a research agenda focusing on causal mechanisms involving ethnic diversity, local armed conflict and diffusion of ethnic conflict, with investigation based on mixed methods.

Recently cited at the Washington Post's Monkey Cage blog.


Why do some civil war torn countries produce more refugees relative to their internally displaced population and others displace more of their population internally than across borders? Surprisingly, the relationship between internally displaced persons and conflict has been woefully underexplored. The aim of this article is to fill this gap in the literature. Using a panel dataset of civil conflicts by country-year from 1993-2010 and a two-step Heckman selection model, I show that civil wars fought along ethnic lines produce greater refugee flows relative to IDP flows than non-ethnic civil wars. I account for this finding by relying on insights drawn previous work on civilian victimization. Specifically, I argue that in conflicts where combatants are recruited along ethnic lines, ethnic markers allow for less costly and more discriminate targeting of rival civilian populations, which in turn increases the share of forced migrants who seek refugee across borders relative to those displaced internally. 


Recent quantitative work on civil wars has identified a link between refugee movements and the spread of conflict across borders. One commonly proposed mech- anisms that accounts for this finding identifies refugee flows as a form of population pressure, which increases violence between host populations and incoming refugees. Another commonly proposed mechanism suggests refugees with ethnic ties to groups within the host population increase violence when their presence disrupts the ethnic balance of power between rival groups. Spatial regression results using a novel geo-coded dataset of substate violence in Lebanon between March 2013 - April 2015 reveal support for the latter mechanism. 


In this article, I explore the ethnic balance of power theory in the context of two different forms of conflict -ethnic armed rebellion and one-sided violence -using global panel data on intrastate violence at the group and country levels. Although, the analysis fails to identify a significant association between the presence of coethnic refugees and ethnic armed conflict, coethnic refugees do significantly increase the probability of rebel led one-sided violence. I account for this by introducing a novel theory, "the logic of population control", which contends that rebels have an incentive to target civilians loyal to rival groups as a means of undercutting the support these rivals enjoy. However, targeting civilians is costly if rebels cannot properly discriminate between their own supporters and civilians supporters of the government or opposing groups. In ethnically polarized states, rebels can take advantage of the salience of ethnicity by mobilizing along ethnic lines and using ethnic markers to aid in the identification of potentially disloyal civilians. This logic, also applies to coethnic refugees. Rebels view refugees that share ethnic ties with the civilian populations represented by rival armed groups (government or otherwise) as a threat and respond to this threat with violence directed towards these unarmed civilians. Therefore, I conclude that while coethnic refugees do not increase the likelihood of conflict onset, they do increase the chances that civilians will be targeted, refugee or otherwise. 

Drone Wars: The Terrorists Strike Back 
​Working Paper (coauthored with Fouad Pervez & Yu-Ming Liou)

As the use of drones has increased dramatically over the past decade, so too has the debate over its effectiveness as a counterinsurgency strategy. Some argue that drones are efficient ways to target terrorists with minimal loss of life. Others claim that anger over civilian casualties from drone strikes, coupled with concerns over violations of state sovereignty, serve to radicalize populations and motivate those targeted to retaliate in kind. One perspective predicts violence to decrease in response to drone strikes, while the other expects it to increase. We test these two propositions using a newly released geocoded dataset of drone strikes in Pakistan. The results of our analysis suggests that drone strikes actually significantly increase the probability of localized terrorist related violence in Pakistan at the substate level, particularly soon after the strike takes place. 


Previous studies of regime type and coup risk have borrowed wholesale from the broader literature on regime durability and leave us with sets of predictors unable to discern between coups, civil wars, or even popular protests. This article aims to address this gap by introducing a theory of regime durability tailored specifically to coup risk. I argue that single-party dictatorships where executive power is constrained by other state institutions are the least coup prone regime type for two reasons: 1) executive constraint provides dissatisfied insiders with meaningful institutional mechanisms to overturn the executive that are not typically found in other autocracies and 2) single-party regimes are less attractive to would-be putschists because other state institutions reduce the flexibility of post-coup executives to drastically alter policies in ways that would benefit his/her coup plotting patrons. This theory is evaluated using a model trained on a global sample from 1961-2013 and tested against out-of-sample data using 10-fold cross-validation. 

For replication data and scripts see my github found here.

© 2017 by Cyrus Mohammadian

  • apple-touch-icon_2
  • LinkedIn Social Icon
  • Octocat