Crowding out the field: External Support to Insurgents and the Intensity of Inter-rebel Fighting in Civil Wars

ABSTRACT

How does external support to insurgents influence the likelihood that the latter will get involved in violent clashes against other rebel groups? In this article, we outline a theoretical framework which contends that, in multiparty civil wars, rebels sponsored by foreign states are more likely to participate in high-intensity inter-rebel conflicts than rebels receiving no support from external states. We argue that this is because external support creates strategic incentives for insurgent leaders to target other rebel contenders in order to signal resolve to their sponsors and to crowd out the battlefield ahead of the post-conflict period. External support, moreover, tends to activate potent socio-psychological mechanisms among rank-and-file combatants that may remove restraints on the use of violence against other rebel fighters. Using data on inter-rebel conflicts from 1989 to 2018, we test these hypotheses with a set of large-N regressions and find strong support for our theory. Further analyzes inductively reveal that our statistical results are likely, to some extent, to be driven by the prevalence of religious insurgencies in contemporary conflicts. Religious insurgencies display organizational features that could reinforce vertical strategic incentives and horizontal socio-psychological dynamics, thereby increasing their involvement in inter-rebel fighting. To further probe the ‘meso-foundations’ of inter-rebel fighting following rebel sponsorship, we then provide qualitative evidence on the Syrian Civil War. Our article contributes to scholarship by highlighting the consequences of external support on conflict processes beyond the insurgent-incumbent dyad.

¿De qué manera el apoyo externo a los insurgentes influye en la probabilidad de que se involucren en enfrentamientos violentos contra otros grupos rebeldes? En este artículo, exponemos un marco teórico que sostiene que, en las guerras civiles en las que hay varias partes involucradas, los rebeldes financiados por estados extranjeros tienen más probabilidades de participar en conflictos de gran intensidad entre grupos rebeldes que los que no reciben apoyo de estados externos. Sostenemos que esto se debe a que el apoyo externo crea incentivos estratégicos para que los líderes insurgentes apunten a otros rivales rebeldes con la finalidad de dar una señal de resolución a sus financiadores y atestar el campo de batalla para tener una ventaja en el período posterior al conflicto. Además, el apoyo externo tiende a activar potentes mecanismos sociopsicológicos entre los combatientes de base que pueden eliminar las restricciones sobre el uso de violencia contra otros combatientes rebeldes. Mediante el uso de datos sobre conflictos entre rebeldes de 1989 a 2018, probamos estas hipótesis con un conjunto de regresiones de N grandes y descubrimos que nuestra teoría tiene un gran sustento. Análisis adicionales revelan por inducción que nuestros resultados estadísticos probablemente, en cierta medida, estén impulsados por la prevalencia de insurgencias religiosas en los conflictos contemporáneos. Las insurgencias religiosas muestran rasgos organizativos que podrían reforzar los incentivos estratégicos verticales y las dinámicas sociopsicológicas horizontales y, por consiguiente, aumentar su implicación en peleas entre rebeldes. Para indagar más las “bases a nivel meso” de las peleas entre rebeldes a consecuencia de la financiación, podemos proporcionar pruebas cualitativas de la guerra civil siria. Nuestro artículo contribuye a la investigación mediante el énfasis en las consecuencias del apoyo externo en los procesos de conflicto más allá de la díada formada por los insurgentes.

Comment le soutien extérieur aux insurgés influence-t-il la probabilité que ces derniers s’impliquent dans des affrontements violents contre d’autres groupes rebelles ? Dans cet article, nous présentons un cadre théorique qui affirme que, dans les guerres civiles multipartites, les rebelles soutenus par des États étrangers sont davantage susceptibles de participer à des conflits entre rebelles de forte intensité que les rebelles ne bénéficiant d’aucun soutien d’États extérieurs. Nous affirmons que cela est dû au fait que le soutien extérieur donne lieu à des incitations stratégiques motivant les chefs insurgés à cibler des groupes rebelles concurrents afin de signaler leur détermination aux acteurs qui les soutiennent et d’évincer ces groupes concurrents du champ de bataille avant la période post-conflit. De plus, le soutien extérieur tend à activer des mécanismes socio-psychologiques puissants pouvant éliminer la retenue des combattants de base à recourir à la violence contre d’autres combattants rebelles. Nous nous sommes appuyés sur des données sur les conflits entre rebelles entre 1989 et 2018 pour vérifier ces hypothèses à l’aide d’un ensemble de régressions à grande échelle, et nous avons constaté que notre théorie était solidement étayée. D’autres analyses inductives ont révélé que nos résultats statistiques étaient susceptibles, dans une certaine mesure, d’être déterminés par la prévalence des insurrections religieuses dans les conflits contemporains. Les insurrections religieuses présentent des caractéristiques organisationnelles pouvant potentiellement renforcer les incitations stratégiques verticales et les dynamiques socio-psychologiques horizontales, ce qui peut accroître l’implication des insurgés dans des combats entre rebelles. Nous proposons ensuite des preuves qualitatives issues de la guerre civile syrienne pour explorer encore davantage les « bases au niveau méso » des combats entre rebelles qui interviennent suite au soutien aux rebelles. Notre article contribue aux recherches en soulignant les conséquences du soutien extérieur sur les processus de conflit au-delà de la dyade insurgés-pouvoir en place.

KEYWORDS:

• Civil wars
• rebels
• external support
• inter-rebel conflicts
• syrian civil war

Notes

¹ A distinction is also commonly made between ‘external support’ and ‘proxy warfare.’ The former encompasses all varieties of support and doesn’t demand that the foreign states have other objectives than helping a local ‘friend.’ In contrast, the latter is the delegation of war’s means to conduct ‘subversive operations’ against a designated target (Mumford 2013a, 40). In proxy warfare, the choice of the surrogate ‘is dependent on the target,’ meaning that the rebels’ identity is less important for the sponsor than the target’s one (Rauta 2018, 452). While the distinction is real at the conceptual level, it is difficult to establish at the empirical level and the study indistinctly looks at both types.

Moreover, we do not consider support provided by non-state actors (Moghadam and Wyss 2020) or assistance given to private military companies (PMC) or pro-government militias (PGM).

² Huang and Sullivan (2020), however, find that external support can favor pro-civilian behaviors .

³ Our argument focuses on leaders’ perceptions of their strategic incentives, which we view as being shaped by a combination of both motivations and opportunities (Schon 2020). While these elements are difficult to disaggregate for each group, we argue that rebel sponsorship influences both leaders’ motivations and opportunities. In some cases, rebels have neither motivations nor opportunities to fight rebel contenders prior to receiving external resources, but rebel sponsorship can redefine both elements. In other cases, rebel leaders can be willing to fight other rebels – they hold the motivation – but lack the resources necessary to do so – they lack the corresponding opportunities. In these cases, they must often ‘instrumentally’ collaborate with groups they dislike or perceive as rivals since it is their only way to acquire necessary resources for their fight against the incumbent (Christia 2012). This is no longer the case if they receive resources from foreign sponsors.

⁴ Following Fjelde and Nilsson, this can be the case whether the recipient of support is strong or weak compared to other groups, since both types have strong interests in targeting contenders (2012). Krause, for his parts, argues that ‘hegemonic’ groups are not likely to experience inter-rebel fighting (2014).

⁵ Interestingly, what rebel leaders can deem rational at first glance – strategically attacking rebel contenders to maximize post-conflict rewards – can in fact foster irrational practices. They can entirely dissociate their behaviors on the battlefield from their actual victory perspectives. This likely occurred in Syria, as later shown in our case study.

⁶ Jones (2017) shows that rebels receiving support from great power win half to two-thirds of the time in civil wars.

⁷ As demonstrated in our case study on Syria, this can go against sponsors’ will, who would prefer to see a united rebellion against the incumbent. Still, with few leverages over rebels’ conduct, sponsors often have no other choice than continuing providing resources to their main local ally.

⁸ The presence of compatible political projects does not necessarily entail a complete alignment of interests and subsequent cooperation. In the 1990s, for instance, the two main Kurdish groups in Iraq Kurdistan region, the Patriotic Union of Kurdistan (PUK) and the Kurdistan Democratic Party (PDK), fought each other despite carrying seemingly compatible visions of the future (Lawrence 2008).

⁹ This is especially the case when rebels evolve within sanctuaries in neighboring countries (Salehyan 2007).

¹⁰ A group is active during a calendar year when its opposition to the government results in at least 25 fatalities.

¹¹ We exclude events opposing rebel groups to pro-government militias or to internal factions of the same group, since these events are arguably driven by different dynamics. We also exclude the conflict opposing the Islamic State (IS) to the Syrian Democratic Forces (SDF) in Syria. In fact, this conflict is the primary objective of the external support provided to the SDF starting from 2014, therefore falling outside the scope of our argument regarding the effect of rebel sponsorship on inter-rebel dynamics (Baczko, Dorronsoro, and Quesnay 2018).

Models using an alternative operationalization of the dependant variable – the aggregate numbers of fighters killed in these conflicts instead of the number of events – are provided in the appendix as robustness checks.

¹² Several of the datasets have been published several years ago. We thus decided to conduct extensive manual research for all variables included, to limit the issue of missing data, and the subsequent biases that could be caused by a large number of dropped observations. Procedures for manual coding are outlined in the appendix.

¹³ To increase the validity of the results, we deliberately excluded from the main models ‘alleged’ cases of external support that were not sufficiently documented. This is an acknowledged limitation of our study, since this might exclude cases of highly concealed sponsorships and ultimately bias our estimates. To limit this issue, we run the same models including ‘alleged’ cases of support in the appendix.

¹⁴ Counts are log-transformed to normalize extreme observations.

¹⁵ A set of alternative models are outlined in the appendix.

¹⁶ A depiction of our dependent variable’s distribution is reproduced in the appendix.

¹⁷ Despite being often studied in conjunction due to alleged similarities, religious and ethnic groups do not appear to present the same prospects in terms of inter-rebel fights. This can be explained by the fact that ethnic groups often have secessionist political projects, which demand overcoming divisions among organizations on the short term, at least until a concrete separation from the central state is achieved. In that regard, a strand of research ‘points to the role of ethnicity in making cooperative behavior more likely’ and in decreasing stereotypes and prejudices toward the out-group among members of a common constituency (Ives 2019a, 6).

¹⁸ This has been salient in Bosnia, Afghanistan, and Chechnya, among others, where faith-based support was common (Byman et al. 2001).

¹⁹ Additional models reproduced in the appendix confirm the unclear influence of religiosity in the relation of interest.

²⁰ Interestingly, if inter-rebel fighting occurred in every Syrian governorate, Gade, Hafez, and Gabbay demonstrate that their majority of such clashes took place in the areas of Aleppo, Damascus, Idlib and Dayr al-Zawr, outside of areas where oil and gas resources are concentrated (2019). This goes against theories solely explaining inter-rebel violence by a strong competition around access to lootable resources.

²² Due to traditional opposition to the group, Saudi Arabia refused, for instance, to fund organizations related to the Muslim Brotherhood, finding alternative groups more in line with its specific faith (Baylouny and Mullins 2018).

²³ Baczko, Dorronsoro and Quesnay argue that religious symbolism was present among most groups, even among those labeling themselves as secular (2018).

²⁴ Interestingly, Kapstein and Ribar show ‘non-locally-dependent’ rebel organizations were more likely to target insurgent contenders that ‘locally-dependent rebel organizations’ (2019).