External Intervention, Identity, and Civil War

• Date: 01 December 2020
• Published date: 01 December 2020
• DOI: 10.1177/0010414020912279
• Author: Stergios Skaperdas,William Wohlforth,Nicholas Sambanis
• Subject Matter: Articles

https://doi.org/10.1177/0010414020912279

Comparative Political Studies

2020, Vol. 53(14) 2155 –2182

© The Author(s) 2020

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/0010414020912279

journals.sagepub.com/home/cps

Article

External Intervention,

Identity, and Civil War

Nicholas Sambanis1* ,

Stergios Skaperdas2*,

and William Wohlforth3*

Abstract

We examine how external intervention interacts with ethnic polarization

to induce rebellion and civil war. Previous literature views polarization as

internally produced—the result of demographic characteristics or intergroup

differences made salient by ethnic entrepreneurs. We complement these

approaches by showing that polarization is also affected by international

politics. We model intervention’s effect on civil war via the pathway of ethnic

identification—a mechanism not previously identified in the literature. In

our model, local actors representing different groups are emboldened by

foreign patrons to pursue their objectives violently. This, in turn, makes

ethnic identity salient and induces polarization. Without the specter of

intervention, polarization is often insufficient to induce war and, in turn, in

the absence of polarization, intervention is insufficient to induce war. We

illustrate the model with case evidence from Ukraine.

Keywords

civil war, polarization, identity, intervention

1University of Pennsylvania, Philadelphia, PA, USA

2University of California, Irvine, CA, USA

3Dartmouth College, Hanover, NH, USA

*All authors contributed equally to this work; their names are listed alphabetically.

Corresponding Author:

Nicholas Sambanis, University of Pennsylvania, 133 S. 36th Street, Philadelphia, PA 19104-

6243, USA.

Email: sambanis@upenn.edu

912279CPSXXX10.1177/0010414020912279Comparative Political StudiesSambanis et al.

research-article2020

2156 Comparative Political Studies 53(14)

Introduction

What is the relationship between ethnic heterogeneity and civil war? Despite

decades of scholarly effort, research on this question remains inconclusive. In

this article, we show that while the literature has looked for domestic expla-

nations linking conflict to ethnic fractionalization or polarization, the rela-

tionship between ethnic differences and conflict is also affected by

international politics. We model the international system’s effect on civil war

via the pathway of ethnic identification—a mechanism that has heretofore

not figured prominently in civil war research. In our model, intervention by

foreign patrons emboldens local actors to pursue their objectives violently.

Mobilization for violence along ethnic lines, in turn, makes ethnic identity

more salient and induces polarization. Without the specter of intervention,

ethnic polarization is often insufficient to induce war and, in turn, in the

absence of polarization intervention is insufficient to induce war. The model

serves as a necessary bridge between international relations and comparative

political-economy approaches to internal armed conflict and helps identify

promising avenues for further study.

Our theory builds on three previous waves of research on ethnicity and

intrastate armed conflict (civil war). In the first wave of empirical studies, the

most common approach was to look for a direct, usually linear, association

between indices of ethno-linguistic fractionalization (ELF) and civil war

onset (Fearon & Laitin, 2003). That wave produced a puzzle: while most

theories of violent conflict suggest a positive correlation between ethnic frag-

mentation and civil war, most empirical studies found no evidence in support

of that hypothesis (Hegre & Sambanis, 2006). This could be due to a mis-

match between theory and the empirical proxies (Sambanis, 2004), as all

first-wave studies used static measures of countries’ ethnic makeup, which

cannot capture the depth or the political salience of ethnic cleavages (Chandra

& Wilkinson, 2008).

A second wave of studies innovated by replacing the commonly used

index of ELF with different indices measuring ethnic polarization (a few

large groups rather than many small ones) or ethnic dominance (size of the

largest group). These analyses established a positive and statistically signifi-

cant correlation between ethnic polarization and civil war (Montalvo &

Reynal-Querol, 2005). Yet these new measures also suffered from the same

key limitations of first-wave studies. Commonly used ethnic polarization and

fractionalization indices failed to capture the changing salience of ethnic

identity over time. The static property of these measures runs afoul of con-

vincing scholarship arguing that ethnic differences per se cannot explain vio-

lent conflict (Chandra, 2006); rather, other variables such as shared culture or