International Studies Review (2021) 0,1–31
FORUM: CONFLICT DELEGATION IN
Swedish Defence University, Stockholm, Sweden
University of Reading, Reading, UK
University of North Texas, Denton, USA
University of Nottingham, Nottingham, UK
Koç University, Istanbul, Turkey
New America, Washington, USA
Military Academy at ETH Zurich, Zurich, Switzerland
Reichman University, Herzliya, Israel
ETH Zurich, Zurich, Switzerland
University of St Andrews, St Andrews, UK
ERIN K. JENNE
Central European University, Vienna, Austria
Leiden University, Leiden, The Netherlands
Authors’ note: Names are listed in order of the authors’ contribution to the forum. The conveners of this collec-
tive discussion are Niklas Karlén and Vladimir Rauta, with all contributors sharing otherwise equal authorship. The
corresponding author is Vladimir Rauta, email@example.com.
Karlén, Niklas et al. (2021) FORUM: CONFLICT DELEGATION IN CIVIL WARS. International Studies Review,
© The Author(s) (2021). Published by Oxford University Press on behalf of the International Studies Association. This is an
Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License
(https://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in
any medium, provided the original work is properly cited. For commercial re-use, please contact firstname.lastname@example.org
Downloaded from https://academic.oup.com/isr/advance-article/doi/10.1093/isr/viab053/6419778 by guest on 03 November 2021
2Conﬂict Delegation in Civil Wars
DAVI D S. SIROKY
Arizona State University, Tempe, USA
University of Oxford, Oxford, UK
Dartmouth College, Hanover, USA
University College London, London, UK
University of Oxford, Oxford, UK
This forum provides an outlet for an assessment of research on the dele-
gation of war to non-state armed groups in civil wars. Given the signiﬁcant
growth of studies concerned with this phenomenon over the last decade,
this forum critically engages with the present state of the ﬁeld. First, we
canvass some of the most important theoretical developments to demon-
strate the heterogeneity of the debate. Second, we expand on the theme of
complexity and investigate its multiple facets as a window into pushing the
debate forward. Third, we draw the contours of a future research agenda
by highlighting some contemporary problems, puzzles, and challenges to
empirical data collection. In essence, we seek to connect two main litera-
tures that have been talking past each other: external support in civil wars
and proxy warfare. The forum bridges this gap at a critical juncture in this
new and emerging scholarship by offering space for scholarly dialogue
across conceptual labels.
Este foro proporciona una salida a la evaluación de la investigación sobre
la delegación de la guerra a los grupos armados no estatales en las guer-
ras civiles. Dado el importante crecimiento de los estudios relacionados
con este fenómeno en la última década, este foro aborda de forma crítica
el estado actual del tema. En primer lugar, repasamos algunos de los de-
sarrollos teóricos más importantes para demostrar la heterogeneidad del
debate. En segundo lugar, ampliamos el tema de la complejidad e inves-
tigamos sus múltiples facetas como ventana para impulsar el debate. En
tercer lugar, trazamos los contornos de una futura agenda de investigación
destacando algunos problemas, enigmas y desafíos contemporáneos para
la recopilación de datos empíricos. En esencia, tratamos de conectar
dos literaturas principales que han sido incapaces de comunicarse entre
sí: el apoyo externo en las guerras civiles y la guerra indirecta. El foro
cubre este vacío en una coyuntura crítica de este nuevo y emergente es-
tudio ofreciendo un espacio para el diálogo académico entre las etiquetas
Cette tribune offre un média pour une évaluation des recherches sur la
délégation de la guerre à des groupes armés non étatiques lors des guerres
civiles. Compte tenu de l’augmentation signiﬁcative du nombre d’études
consacrées à ce phénomène au cours de cette dernière décennie, cette
tribune se penche sur l’état actuel du domaine. Nous commençons par
passer en revue certains des développements théoriques les plus impor-
tants aﬁn de démontrer l’hétérogénéité du débat. Ensuite, nous dévelop-
pons le thème de la complexité et étudions ses multiples facettes aﬁn
Downloaded from https://academic.oup.com/isr/advance-article/doi/10.1093/isr/viab053/6419778 by guest on 03 November 2021
NIKLAS KARLÉN ET AL.3
d’offrir une ouverture par laquelle faire avancer le débat. Enﬁn, nous
traçons les contours d’un futur programme de recherche en soulignant
certains problèmes, certaines énigmes et certains déﬁs contemporains
de la collecte de données empiriques. En substance, nous cherchons à
rétablir une relation entre deux grandes littératures entre lesquelles il
existait autrefois un dialogue: celle sur le soutien extérieur lors des guer-
res civiles et celle sur les guerres par factions interposées. Cette tribune
comble cette lacune à un moment critique de ces nouvelles recherches
émergentes en offrant un espace pour un dialogue intellectuel allant au-
delà des étiquettes conceptuelles.
Keywords: conﬂict delegation, external support, proxy warfare,
civil war, principal–agent theory
Palabras clave: delegación de conﬂictos, apoyo externo, guerra
indirecta, guerra civil, teoría del principal–agente
Mots clés: délégation des conﬂits, soutien extérieur, guerres par
factions interposées, guerre civile, théorie du principal–agent
Swedish Defence University
University of Reading
Conﬂict delegation has been studied across several subﬁelds—international re-
lations, conﬂict research, and international security—under a myriad of labels.
External support in civil wars, proxy war/warfare, state sponsorship of insur-
gency/terrorism, rebel patronage, indirect intervention, informal/transnational al-
liances, paramilitary operations, internationalized conﬂict, indirect interstate con-
ﬂict, security assistance, subversion, substitution, and military aid are just some of
the concepts employed in the ever-expanding semantic ﬁeld of sponsorship of non-
state armed groups. Current research has developed within separate debates inter-
ested in largely the same phenomenon. Although signiﬁcant analytical differences
worth emphasizing exist,1there is ample space to think creatively and pluralistically
about the present and future research on the topic.
In this forum, we offer a discussion by a collective of scholars situated at the inter-
section of two main literatures: external support in civil wars and proxy warfare. The
two literatures straddle an intellectual boundary reinforced by the usual dynamics of
knowledge production: employment of different concepts, methodological prefer-
ences, and theoretical traditions. To simplify: while a plethora of theoretical labels
have been used within both literatures, research on external support tends to be
more empirical, quantitative, and outcome-oriented, while that on proxy wars has
been more conceptual, qualitative, and process-oriented. Notwithstanding the im-
portance of such markers to the formation and consolidation of disciplinary iden-
tities, what distinguishes these two strands of research is, however, narrower than
most often assumed. Instead of fragmentation, we observe compartmentalization,
1For instance, scholars differ in terms of explanatory and geographical focal points, the agency of various actors,
the formalization of the relationship, the type of targets, the degree of interdependence between external actors, the
type of support, and the use of principal-agent theory.
Downloaded from https://academic.oup.com/isr/advance-article/doi/10.1093/isr/viab053/6419778 by guest on 03 November 2021
4Conﬂict Delegation in Civil Wars
visible most often in the different bibliographies accompanying studies of external
support and proxy warfare. In short, what we see are efforts to pursue two spe-
cialized research enterprises interested in answering questions concerning a key
security challenge in the twenty-ﬁrst century: conﬂict delegation.
Conﬂict delegation refers to a strategy in which a foreign2government commits
material resources or military expertise to a non-state armed group to target a per-
ceived adversary. Delegation requires some degree of control—that is, state spon-
sors are likely to inﬂuence the aims, strategies, and tactics of rebel groups (Salehyan
2010, 501). This forum scrutinizes the development of research on conﬂict delega-
tion at a key moment in time. Interest in the topic has grown in recent years because
of the considerable external involvement in conﬂicts around the globe, along with
its signiﬁcant media exposure and the attention it has received by policymakers.
However, the delegation of war to non-state armed groups is not a new phe-
nomenon, representing “a perennial strand in the history of warfare” (Mumford
2013, 1). Neither is it rare (Salehyan 2010, 497). San-Akca has estimated that al-
most two-thirds of all rebel groups active since World War II have beneﬁted from re-
sources from outside states (San-Akca 2016, 1) and Grauer and Tierney have shown
that the likelihood of rebels receiving aid has increased considerably (Grauer and
Tierney 2018). Examples are abundant. Saudi Arabia, Qatar, Turkey, and the United
States have channeled arms, equipment, and ﬁnancial resources to various oppo-
sition groups in Syria (Baylouny and Mullins 2018;Rauta 2020, 2–3). Russia has
provided vital support to rebels in Eastern Ukraine (Platonova 2021) while at the
same time smuggling arms to the Taliban in Afghanistan (Byman 2020,6).Iran’s
proxy relationship with Hezbollah has demonstrated signiﬁcant persistence, with
a wide range of assistance being provided for almost four decades (Borghard and
Rapp-Hooper 2013, 85–86). Sierra Leone, Guinea, and Ivory Coast have backed
rebel groups in neighboring Liberia (Hazen 2013, 107–37). Pakistan has aided sev-
eral militant groups in Kashmir (Byman 2020, 6), Malaysia has supported groups
in the Philippines, and Thailand has aided the Khmer Rouge in Cambodia (Lee
2020, 136–37, 157–59). More than a dozen different states have provided support
to various Palestinian groups in Israel over the years (San-Akca 2016, 70–79). This is
still just a snapshot of some relationships of an otherwise global phenomenon, and
by evaluating the progress and prospects in conﬂict delegation research we tap into
the urgency proxy wars impress on policy makers (Benowitz and Ceccamese 2020;
Plana 2020;Stark 2020a).
How do approaches to conﬂict delegation differ conceptually, analytically, and
empirically? How do we bridge the theoretical and methodological traditions com-
monly associated with the terms? What are the strengths and weaknesses of the
inﬂuential principal–agent theory? What are the best practices of translating the
recent conceptual work behind proxy wars into the ongoing efforts toward clear
operationalization and robust measures of external support? How do we compare
and integrate the empirical insights from case study work on speciﬁc proxy wars with
the patterns and trends in the provision of external support world-wide? Finally, how
do we move forward by harnessing the potential of an integrated, future-oriented
The contributions to this forum tackle these questions in an exchange that cov-
ers concepts, theory, data, and policy. We open the forum by canvassing some of
the most important theoretical developments to demonstrate the heterogeneity of
the debate. The ﬁrst three contributions highlight different approaches to under-
standing conﬂict delegation. Idean Salehyan begins to reﬂect on how the ﬁeld has
evolved during the last decade since the publication of his article, “The Delegation
of War to Rebel Organizations.” We have come to know a lot about the causes and
consequences of conﬂict delegation, and, in the process, principal–agent theory
2This sets this literature apart from domestic delegation to progovernment militias within a country (Cohen and
Nordås 2015;Eck 2015;Carey and Mitchell 2017;Biberman 2018;Fisk 2021).
NIKLAS KARLÉN ET AL.5
has become the dominant theoretical framework. This is followed by a discussion
by Andrew Mumford advocating for closer conceptual integration between the lit-
erature on proxy warfare and civil wars. The starting point here is Mumford’s Proxy
Warfare monograph, which recast the study of proxy wars in the wider research on
contemporary warfare, framing its relevance away from Cold War accounts and to-
ward its many contemporary facets. Taken together, Salehyan and Mumford offer
two visions of progress and integration that map onto two distinct directions of
knowledge production: consolidation of research on external support in civil wars
and the emergence of proxy war studies. Drawing on arguments made in States in
Disguise: Causes of State Support for Rebel Groups, Belgin San-Akca elaborates on the se-
lection model. San-Akca’s contribution provides an assessment of the strengths and
weaknesses of the existing theoretical accounts. San-Akca then makes the case for
reﬁning the selection model as a theoretical alternative, itself a broader invitation
to think differently about the delegation of war to non-state armed groups.
The following three contributions expand on the theme of complexity and in-
vestigate its multiple facets as a window into pushing the debate toward solving em-
pirical, policy-relevant problems, while removing conceptual and theoretical hur-
dles. Niklas Karlén and Vladimir Rauta propose some extensions to the standard
principal–agent model by discussing delegation in an integrated typology of alter-
native actor arrangements. They highlight that research on delegation in other dis-
ciplines has offered a much more comprehensive understanding of delegatory re-
lationships by extending the standard principal–agent framework, which is equally
relevant but entirely missing in research on conﬂict delegation. Alexandra Stark
adds empirical weight to the complexity of delegation by focusing on the roles re-
gional actors play in waging wars by proxy in the Middle East. Michel Wyss and Assaf
Moghadam highlight the often overlooked ability of non-state armed actors to as-
sume the role of patrons themselves in an argument that breaks with the orthodoxy
of state sponsorship. This discussion of complexity then makes way for the next
three contributions, which draw the contours of a future research agenda by high-
lighting a number of contemporary problems, puzzles, and challenges to empirical
Drawing on research in the context of political violence in Africa, Allard Du-
ursma and Henning Tamm discuss mutual interventions whereby neighboring states
respond in kind to rivals’ provision of external support. Next, Erin Jenne, Milos
Popovic, and David Siroky outline a robust argument linking great power politics
and proxy warfare with a focus on security hierarchies. Vanessa Meier then highlights
the challenges inherent to quantitative data collection efforts, in a contribution
that relates conceptual and operationalization choices to the difﬁculties of organiz-
ing complex and, often messy, empirics. Finally, ways forward are outlined. Empha-
sizing theoretical plurality, methodological diversity, and a commitment to rigor,
Alexandra Chinchilla, Kit Rickard, and Giuseppe Spatafora conclude by highlight-
ing avenues for future research. In essence, they stress the need to further theorize
proxy agency, to modify principal–agent theory, and to re-evaluate the international
conditions surrounding the provision of support.
A Decade of Delegation
University of North Texas
In 2010, I wrote “The Delegation of War to Rebel Organizations,” which appeared
in the Journal of Conﬂict Resolution (Salehyan 2010). This was not the ﬁrst article
to examine the issue of foreign support for rebel organizations or proxy warfare.
Indeed, during the Cold War, observers noted that many armed conﬂicts were
6Conﬂict Delegation in Civil Wars
fueled by superpower support for combatants (Deutsch in Eckstein 1964;Dunér
1981;Bar-Siman Tov 1984). Rather, in that article I sought to bridge the divide be-
tween scholars of international and civil conﬂict as well as to introduce delegation
theory (or principal–agent theory) as a way to think about this phenomenon.
In a nutshell, scholars of international war have vastly understated the level of
armed conﬂict in the international system by ignoring indirect conﬂict strategies.
Rather than open hostilities between the armed forces of states, a large number of
international conﬂicts involve the provision of resources to rebel organizations. In
other words, states provide support to domestic combatants within their rivals as
a substitute for direct hostilities. Yet, in large-N studies, these are normally coded
as civil wars with external involvement and do not appear in datasets on interna-
tional disputes. Russian involvement in Ukraine, Iranian and Saudi proxy warfare
in Yemen, and the wide array of state actors backing militants in Syria are but a
few recent examples in which governments have empowered insurgents instead of
ﬁghting one another directly.
As I argued in “The Delegation of War”, one theoretical framework that promises
to shed light upon these relationships is principal–agent theory. Accordingly, a for-
eign government (the principal) delegates armed conﬂict against its rival by enlist-
ing the support of a rebel organization (the agent). In doing so, the principal for-
goes the costs of direct warfare, including lives lost, military resources, and interna-
tional or domestic condemnation and gains a partner that may possess greater local
information and popular legitimacy. Yet, the principal also loses some control over
its foreign policy, particularly if it cannot adequately screen agents for reliability and
competence or sanction them for bad behavior. For the insurgents, they must weigh
the beneﬁts of augmented resources against the potential for lost autonomy.
In “The Delegation of War”, I pointed to the important distinction between del-
egation and intervention. However, delegation should also be theoretically distinct
from the related concept of alliances (Hughes 2012, 11–14; Tamm 2016a, 151–52;
Rauta 2021b, 15). For example, Tamm provides a broader usage of the term al-
liance, while my emphasis is on the distinction between arrangements that come
with hierarchical control over the group versus those in which both parties retain
autonomy. With delegation, the foreign state expects, to varying degrees, control
over the rebel’s agenda. It is a hierarchical relationship in which the principal pro-
vides funding and support, while the rebel group adapts its goals, strategies, and
tactics to suit the needs of its patron. The rebels are dependent upon foreign re-
sources and are asked to do their patron’s bidding. Alliances do not involve the
same hierarchical relationship. Rebels and external states may cooperate on the
battleﬁeld and resources may be offered, but with little control over the rebel’s
agenda. Empirically, while it is easier to ascertain whether or not a foreign govern-
ment provides resources, it is much harder to measure the degree of hierarchy that
exists in the relationship or how dependent upon external support the rebels are.
While the line between a horizontal alliance and hierarchical delegation is often
unclear in practice, they are not interchangeable concepts.
Recently, Abbott et al. (2020) have offered an expanded typology of indirect gov-
ernance relationships, on which Heinkelmann-Wild and Mehrl (2021) expand in
a comprehensive discussion of variation in control mechanisms. Central to this
conceptualization, and my distinction between delegation and alliance, is the de-
gree of hierarchy expected in the partnership between the sponsor and the rebel
(Abbott et al. 2020, 14). Hezbollah is largely an Iranian creation and answers to
Tehran; the National Union for the Total Independence of Angola in Angola was
formed independently but was co-opted by the United States and South Africa and
depended upon these governments for resources and direction. Non-hierarchical
relationships afford the insurgents considerable freedom to use foreign resources
to meet their own needs, even if the patron shares common goals.
Notwithstanding conceptual differences, several studies have sought to un-
cover the causes of foreign support for rebel organizations, building upon “The
NIKLAS KARLÉN ET AL.7
Delegation of War”. Salehyan, Gleditsch, and Cunningham (2011) ﬁnd that insur-
gents ﬁghting states with an international rival and those embedded in transna-
tional communities are more likely to receive foreign support. Moreover, while weak
rebels (relative to the state) are less likely to receive support as they are not viable,
the strongest groups also tend not to receive support as they would prefer to raise
their own revenue, without foreign constraints. Maoz and San-Akca (2012) delve
deeper into how international rivalries shape rebel support, ﬁnding that the weaker
state in a rivalry is more likely to use indirect conﬂict strategies. As discussed below,
San-Akca (2016) provides a rich theoretical and empirical analysis of state support
for rebels, focusing on international interests, ideological/cultural similarity, and
domestic incentives to empower rebels. Finally, Tamm (2016a) argues that support
for rebels in the two Congo Wars was often driven by external leaders’ desire
to insulate themselves from a coup at home, including by securing resources to
distribute to supporters.
Other studies have focused on the consequences of foreign backing of rebels. For
instance, Tamm (2016b) ﬁnds that external resources can promote rebel cohesion
or splintering depending on how funds are allocated among rebel elites. Salehyan,
Siroky, and Wood (2014) ﬁnd that insurgents supported by external states are more
likely to kill civilians, although democratic sponsors tend to restrain the groups they
fund. Sawyer, Cunningham, and Reed (2017) show that rebels that receive highly
fungible support, such as money and guns, are less likely to see conﬂict termination
than rebels that do not. Karlén (2017) demonstrates that the presence of external
support during a civil war increases the risk of conﬂict recurrence in the short term.
Popovic (2017) examines the causes of rebel defection against their sponsors and
ﬁnds that decentralized groups are less accountable to foreign patrons. We also
now know that differences in the provision of external support are associated with
the adoption of violent/nonviolent tactics (Petrova 2019), the likelihood of civil
war negotiations (Karlén 2020), that gender framing impacts rebel efforts to secure
support (Manekin and Wood 2020), and that foreign sponsorship affects groups’
incentives to engage in rebel governance (Huang and Sullivan 2021). Finally, Stein
and Cantin (2021) show that rebels sponsored by foreign states are more likely to
participate in high-intensity inter-rebel conﬂicts than rebels receiving no support
from external states.
Many of the articles cited above (including my own) infer delegation from for-
eign support. On the one hand, foreign funding may matter for conﬂict dynamics
precisely because it is foreign, namely not derived from domestic sources such as
civilians. Funding from wealthy foreigners or members of the diaspora may have
similar effects, as it makes rebels less accountable to local populations. On the other
hand, external resources given as a donation versus those given with the expecta-
tion of obedience may be vital for other matters. What is clear is that conﬂict del-
egation continues to be a promising research topic, and we should pursue these
research questions and puzzles with the intellectual creativity, diversity, pluralism,
and rigor outlined in the conclusion to this forum. In shaping the outlines of the
future research agenda, principal–agent theory will surely inform scholarship on
the international dynamics of civil wars for years to come. I look forward to seeing
what the next ten years of research will yield.
In Search of Proxy War Studies
University of Nottingham
This contribution attempts to bridge the study of proxy wars with the broader civil
war literature, by focusing on the evolution of the proxy war scholarship since the
publication of my book Proxy Warfare (Mumford 2013). I hope to encourage greater
8Conﬂict Delegation in Civil Wars
terminological clarity and theoretical integration between the strands of literature
introduced by Karlén and Rauta at the start of the forum. This move is premised
on the remarkable expansion of proxy war literature into something that can in the
future become the subﬁeld of proxy war studies (Rauta 2020).3On the one hand,
this allows to take stock of new research into proxy wars and great power conﬂict
(de Soysa 2017; Jenne, Popovic, and Siroky in this forum), counterterrorism
(Larsdotter 2014;Cragin 2015,2020), cyber war (Borghard and Lonergan 2016),
and conﬂict mediation and management (Hellmüller 2021;Irrera 2021). On the
other hand, it presents a genuine opportunity to answer de Soysa’s question of “how
might we incorporate the idea of proxy war in the study of civil war to broaden the
general understanding of how civil wars occur and end?” (de Soysa 2017).
I observed in my book, Proxy Warfare, that the phenomenon had “not been an
adept cross-disciplinary traveller” (Mumford 2013, 2). Research in subsequent years
has done little to persuade me to revise that argument. The lack of cross- (or even
inter- or multi-) disciplinarity in the study of proxy wars has led to the creation
of methodological silos, which in turn has created alternative terminological dis-
courses across different subﬁelds (Rauta 2018). This terminological déjà vu is the
product of major disciplinary and methodological challenges facing the study of
proxy war. As a phenomenon, proxy wars are empirically ubiquitous yet still thought
of as conceptually obscure. A recent article evaluated the analytical merits of proxy
war starting from the paradoxical rejection of the concept in the absence of proper
conceptual analysis (Rauta 2021b). What remains to be asked is: when is a proxy
war not a proxy war? Conceptually, proxy wars seem to be sharing a fate with the
famous American judicial interpretation of pornography, namely, we know what it
is when we see it, but there is a wide spectrum of judgment involved in assessing its
meaning. This resonates with Salehyan’s remarks above about the differences be-
tween intervention, delegation, and alliance and Meier’s observations on clarity in
operationalizing external support.
There has been some excellent recent research on the role of third parties in
intrastate conﬂicts, some of which have already been highlighted by Salehyan. This
builds upon a wider body of civil war literature from the last two decades that has
posited explanations of third-party interference in the conduct and cessation of
such wars (Regan 2002;Regan and Aydin 2006). Yet much of this literature has fo-
cused predominantly on direct state intervention and not indirect means of exter-
nal support channeled through proxies, creating a situation in which civil war stud-
ies and proxy war studies have been talking past each other. Some civil wars scholars
do not classify indirect intervention as an act of proxy war. This reluctance to con-
ceptually separate proxy wars (an indirect intervention) from overt state military
interference (a direct intervention) lumps together two different forms of conﬂict
intervention by third parties. Furthermore, it overlooks their very different strategic
motives (plausible deniability versus show of strength) and their general method of
undertaking (covert versus overt).
External support is, although widely used, perhaps not the most appropriate term
to describe third-party interference as it does not adequately account for the dynam-
ics of the relationship behind such support. The phrase external support is abound
in many modern studies of third-party intervention in civil wars and is indeed inte-
gral to the conclusion of some works, but such interventions are rarely cast as being
constitutive of a proxy–benefactor relationship. Sawyer, Cunningham, and Reed, for
example, ﬁnd that “highly fungible external support is as likely, if not more likely,
than a shift in military power to prolong conﬂict” (Sawyer, Cunningham, and Reed
2017, 1176). Karlén concludes that “external support to rebel movements increases
the probability of conﬂict recurrence in the short term” (Karlén 2017, 500). These
3Whereas other subﬁelds of the study of conﬂict, such as civil wars, have their own journals and dedicated academic
fora, the idea of proxy war studies is tangibly in its infancy galvanized around the forthcoming Routledge Handbook of
Proxy Wars edited by some of the contributors to this forum.
NIKLAS KARLÉN ET AL.9
ﬁndings have signiﬁcant implications for assessments of a third party’s strategic op-
tions, especially if the aim is to prolong a conﬂict so that a relatively weaker proxy
can reach a stalemate that would not otherwise have been possible or to destabilize
a particular country or region through encouraging conﬂict recurrence. Future re-
search should integrate more closely Regan and Aydin’s observation according to
which third-party efforts to compel victory or engender stalemate “would be con-
sistent with some of the cold war rhetoric about proxy wars” (Regan and Aydin
2006, 743). We need not just see the connection but join the dots. This forum in
general, and my contribution, in particular, is a rallying call to scholars across the
civil war/proxy war studies divide to continue to see connections such as this and
join the dots between the two related phenomena. It picks up on Salehyan, Gled-
itsch, and Cunningham’s call for research on conﬂicts to “look closely at circum-
stances where the lines between civil and international war are blurred” (Salehyan,
Gleditsch, and Cunningham 2011, 734–35). Proxy wars operate in this blurry con-
ﬂict zone where internal wars are made more complex and intractable because of
external intervention. Analysis of proxy wars needs to be cleaved apart from generic
assessments of external intervention and given a consistent, independent platform
of critical evaluation. To this end, thinking about, contributing to, and building on
the idea of proxy war studies is key.
The Role of Agency in the Formation of
This contribution provides a critique of dominant approaches and propose an al-
ternative theoretical framework to understand the relationship the relationship be-
tween external states and non-state armed groups (NAGs): selection theory. Recent
conﬂicts have been complicated not only by the high number of armed groups
populating the battleﬁeld, but also by third-party states competing to inﬂuence the
process and the outcome of these civil wars by allying with armed groups. Multiple
states frequently found themselves competing to support the same NAG while NAGs
were free to pick and choose their partners in the marketplace of supporters.
Neither the term proxy warfare nor conﬂict delegation is sufﬁcient to capture
this new nature of state–NAG relations. My approach is to examine these relations
as alliances or partnerships since both states and NAGs have agency to accept or
reject the formation of these partnerships. Since the end of World War II, almost
107 states fought a NAG, which pursued governmental and/or territorial objectives
(San-Akca 2016). In almost 77 percent of these internal conﬂicts, there was a third-
party state supporting the NAG side in the conﬂict by providing safe havens, training
camps, guns, weapons, and/or funds. Out of 537 NAGs that emerged between 1946
and 2019, 52 percent were able to receive such third-party state support.4
Building on a central theme of this forum, there is little scholarly consensus on
how to conceptualize the complex relations between states and NAGs. Most of the
existing research examines these relations as part of states’ foreign policies. The new
form and nature of relations between states and armed groups that have emerged
in the past three decades require us to develop more comprehensive conceptual,
analytical, and empirical tools. Although there is a prominent body of research
4These ﬁgures are calculated using the latest updated version of the NAGs Dataset (ver. Dec 2020), which includes
information on 537 armed groups and their state supporters for the period between 1946 and 2019. The data is available
10 Conﬂict Delegation in Civil Wars
examining external support of armed groups in detail, it still falls short in captur-
ing the various dimensions of state–NAG relations. While the conﬂict delegation
literature often limits the agency of NAGs, which can take action without the
consent of their patrons, delegation theory starts telling the story if partnerships
between the two fail. States delegate to armed groups to realize their foreign policy
objectives, which might lead to a principal–agent problem. As Salehyan remarks
in this forum, states settle for less-competent groups knowing that delegation to
strong groups is risky since groups might abandon the supporter goals after they
get necessary resources from their supporters. From the armed group’s perspective,
the delegation logic attributes agency to armed groups when the latter makes a
choice between accepting external support or relying on its own resources, but
it does not tell us much about the process through which armed groups actively
search for state allies and choose among many.
Invaluable insights have been offered by framing state–rebel relations as germane
to problems, such as moral hazard and/or agency slack and adverse selection. Di-
vergent preferences, together with asymmetric information, create the possibility
that agents will not perform as intended. The principal has no way of knowing ex
ante whether it is selecting an agent that is competent and reliable; hence, it runs
the risk of adverse selection. Agency slack, or moral hazard, refers to the risk that
the agent can, once the relationship has been established, shirk its responsibilities
by failing to devote optimal effort or that it can take actions that run counter to the
preferences of the principal (Salehyan 2010, 502). The logic of delegation implicitly
suggests that states just delegate rebels to conduct their foreign policy business and
rebels tacitly agree to the tasks delegated to them. This perspective, though it helps
capture when the relations between states and rebels are hierarchical, misses the
details about the state–NAG interactions in the form of alliances or partnerships.
We do not learn much about the conditions under which states select some NAGs
while not others. By the same token, it is not sufﬁcient to understand the complex
mechanisms through which NAGs select their external supporters.
The selection theory offered in States in Disguise—Causes of State Support for Rebel
Groups goes beyond the logic of proxy warfare and delegation (San-Akca 2016).
NAGs are treated as potential allies or partners of states and the formation of part-
nership or alliances between the two is shown to occur as a result of a selection
process on the side of both states and NAGs. In other words, NAGs are attributed
agency in accepting or rejecting the offers of states. I proposed selection theory
to emphasize a process of selection involving both the supporters and the groups.
While principal-agent theory explains the complications after the alliance or co-
operation between states and armed groups are established, I focus on the process
prior to the decision to initiate a partnership through either delegation, alliance, or
intervention by states and acknowledge that armed groups can and do also initiate
such cooperation with states. I examine the role of interstate rivalry, ideational ties
between supporters and target states as well as between supporters and NAGs, and
the domestic political setting that motivates cooperation between states and armed
groups. I build on conventional theories of international relations as informed by
realism, liberalism, and constructivism to hypothesize about the onset and level of
such cooperative arrangements between states and NAGs.
Moreover, the logic of selection theory helps us move beyond the recent schol-
arly focus on external state support of rebels, which is only one type of alliance
between states and NAGs. NAGs are capable of choosing their supporters as well
as going into the territories of other states to raise funds, recruit individuals, and
spread propaganda. Through their own initiative, NAGs establish a form of de facto
alliance with these states, even though the states have not directly created chan-
nels to support these groups (San-Akca 2014,2016). The delegation logic does not
speak to such cases since it is no longer possible to talk about a patron or a prin-
cipal intentionally delegating tasks to the agent in question. Indeed, in the period
between 1946 and 2019, 43 percent of all groups were able to raise funds, ﬁnd safe
NIKLAS KARLÉN ET AL.11
havens, establish training camps, and acquire weapons within the borders of states
independently. In other words, states do not always deliberately help NAGs, but
NAGs nonetheless manage to acquire resources within other states’ borders. I re-
fer to these cases as de facto or tacit alliances between states and NAGs. NAGs do
not only make a choice about whether to accept or reject a state’s offer, but also
actively seek outside sources of support in an effort to systematically diversify their
My goal is not to state that the delegation logic does not apply at all. Rather, it
focuses on the subsequent process after a state–NAG alliance begins. Selection the-
ory emphasizes the process at the onset of such an alliance and offers theoretical
insights into how states and NAGs engage with each other through several forms
of cooperation. The delegation logic falls short speciﬁcally when it comes to cases
in which NAGs are able to acquire support without the backing of an external pa-
tron and no strings are attached by their outside patrons. We need more work to
predict the patterns of behavior by such NAGs. Future research needs to engage in
sophisticated analysis in order to enrich our understanding of why multiple states
support the same NAG and why a NAG seeks support from multiple states. Both
of these are intriguing since both states and NAGs end up competing over their
potential allies. In addition, we should pay more attention to how many state–NAG
alliances are formed as a result of the coincidental convergence of interests or de-
liberate choices on the part of both states and NAGs. The detailed analysis should
include the role these alliances play in conﬂict instigation and resolution. Given that
we are in the era of nuclear weapons, alliances between states and NAGs will con-
tinue to be one of the most popular instruments used by states that want to avoid
the material and reputational costs involved in direct war-making.
Complex Conﬂict Delegation in Civil Wars
Swedish Defence University
University of Reading
In the last decade, principal–agent theory has almost completely dominated the-
orizing on conﬂict delegation (Byman and Kreps 2010;Salehyan 2010;Salehyan,
Gleditsch, and Cunningham 2011;Salehyan, Siroky, and Wood 2014;Popovic 2017).
Even scholars using alternative labels, such as beneﬁciary–proxy (Mumford 2013)
or sponsor–insurgent (Tamm 2016a), adhere to many of its core assumptions. Al-
though not without criticism (Borghard 2014, 25–31), it is fair to say that principal–
agent theory has become the dominant framework through which we have come to
study external support to non-state armed groups. This is not surprising as it speaks
directly to the fundamental process of delegation, providing us with valuable in-
sights into conﬂict delegation in civil war, as mentioned by the previous contribu-
However, despite the widespread use of this theoretical framework we have still
not utilized it to its fullest potential. The current standard application of principal–
agent theory has focused entirely on one direct relationship: that between princi-
pal and agent. This has been the model almost exclusively employed in the con-
ﬂict delegation literature. Although suitable to address some research questions, it
might cloud our vision in relation to other important issues. As San-Akca has al-
ready noted, one consequence of employing this theoretical perspective is that it
limits the agency of non-state armed groups. Another issue is that we may overlook
the central role of other actors in the delegation process as highlighted next by
12 Conﬂict Delegation in Civil Wars
Figure 1. Complex delegation patterns
Stark. This may lead to an oversimpliﬁcation as to how conﬂict delegation in civil
wars works in practice.
To this end, we believe that we should engage more with complexity in relation
to conﬂict delegation. This does not have to lead to a complete rebuttal of the
current framework, as principal–agent theory offers a multitude of extensions that
we so far have underutilized. More complex delegation chains have been exten-
sively discussed in other literatures, such as in the study of international organi-
zations (Nielson and Tierney 2003;Michaelowa, Reinsberg, and Schneider 2018),
economics (Laffont and Martimort 1998), and bureaucratic politics (Bennedsen
and Schultz 2011).
Adapting such insights to conﬂict delegation may yield important theoretical in-
sights. In this contribution, we argue that by extending the complexity of conﬂict
delegation beyond the usual two actors, principal and agent, we shift the focus onto
more complex delegation patterns with the potential to open up a range of hith-
erto unexplored questions. In ﬁgure 1, we highlight a selection of ﬁve delegation
patterns that extend standard principal–agent theory.5These extensions of the stan-
dard model offer other conﬁgurations as to the type of principal(s), the agent(s),
and the nature of their relationships. Of these extensions, only multiple delegation,
5Given the extent to which this issue has been ignored in the literature on conﬂict delegation, the labels we propose
are tentative and provide a ﬁrst-cut mapping of complexity. Even in other sub-ﬁelds employing principal–agent theory,
there is little conceptual agreement on exact labeling.
NIKLAS KARLÉN ET AL.13
that is, separate principals providing support to the same agent, has received at-
tention in the literature on conﬂict delegation. Tamm (2016b) has looked at how
multiple principals affect rebel cohesion and splintering, while Popovic (2018) has
explored how this impacts inter-rebel alliances. Although other forms of delegation
patterns have been less systematically explored, alternative arrangements are visible
in empirical cases.
We identify four other types: collective delegation,specialized delegation,dual delega-
tion,andsimultaneous delegation. To begin with, collective delegation accounts for a
situation in which several principals coordinate delegation efforts and jointly del-
egate authority. A contemporary example of this would be state sponsors working
together to channel support to the Free Syrian Army in Syria through “military op-
eration rooms” in Turkey and Jordan (Lister 2016,24).Specialized delegation refers
to tasks subsequently delegated along a chain of increasingly specialized agents.
Dual delegation describes a process in which the delegation of authority is delegated
to one agent who then further delegates this to another agent. In this situation,
the principal–agent relationship is mediated by a middleman. This could, for in-
stance, be great powers routing support through regional or local allies, with ex-
amples including United States support to the Mujahedin ﬁghters in Afghanistan
delivered through Pakistan (Hughes 2012). Lastly, double delegation is a simultaneous
delegation to more than one agent. Syria has provided support to a range of Pales-
tinian groups, such as Hamas and the Palestinian Islamic Jihad (San-Akca 2016, 75),
while the United States provided parallel support to various rebel groups in Syria, at
times with unexpected consequences. In 2016, the CIA-armed militia Fursan al Haq,
or Knights of Righteousness, clashed with the Pentagon-backed Syrian Democratic
Forces (Bulos, Hennigan, and Bennett 2016).
Empirically, we can observe different types of delegation patterns as these ex-
amples demonstrate. Nonetheless, we have so far restricted most analyses to the
standard model. By acknowledging complexity and its variation, we open several
potential research avenues. First, why do principals employ middlemen? Second,
do proxies assume different roles for different principals? Third, what are the ef-
fects of shorter and longer delegation chains? Future research needs to look deeper
into how state sponsors coordinate and jointly channel support and how this affects
rebel groups. Moreover, discussing complex conﬂict delegation presents a unique
opportunity to evaluate the issue of control in the delegation of war to non-state
armed actors, a key topic the debate has yet to address systematically.
Complicating the Proxy War Model
As the contributions to the forum already identify, the proxy war literature models
delegation relationships as an ideal–typical connection between an external state
sponsor and a local non-state group proxy, where the external sponsor provides in-
direct support to local actors (such as ﬁnancing, weapons, training, intelligence,
and diplomatic support) in exchange for a degree of command over the proxy
group (Hughes 2012;Mumford 2013;Rauta 2018;Groh 2019). However, Karlén
and Rauta’s discussion on complex delegation patterns invites a broader theoreti-
cal reconsideration of this simple sponsor–proxy model. While the standard model
can be a useful heuristic, it fails to capture the much broader array of proxy rela-
tionships and types of interventions that we see in civil wars today. Speciﬁcally, there
are two advances that the proxy wars literature must make to capture these dynam-
ics: conceptualizing the role of regional actors in relation to both great powers and
14 Conﬂict Delegation in Civil Wars
local actors, and understanding intervention as a menu of options rather than a
The proxy war literature has not conceptualized the role of regional state actors
in relation to both great powers and local actors, and it has not fully reckoned with
the types of support that external actors can provide. While regional actors do act
in the traditional proxy role by providing indirect support to local actors, they in-
creasingly also act as both proxies and sponsors, receiving support from great pow-
ers while sponsoring local actors themselves. Regional states’ interventions in some
conﬂicts may be better thought of as one node in a chain of proxy relationships,
where a regional state may both receive support from a great power sponsor and
provide support to local proxies.
Regional actors also intervene in civil wars in ways that blend forms of indirect
and direct support, providing funding, weapons, and/or intelligence support to
local actors, for example, while at the same time deploying Special Forces on the
ground to train and sometimes ﬁght alongside local actors or conducting drone
strikes or other types of airstrikes. Additionally, aspects of this kind of interven-
tion are often covert or defy straight-forward attribution, making these relationships
even more analytically complex than the straightforward sponsor–proxy heuristic.
While intervention is sometimes conceptualized along a ladder or spectrum, rang-
ing from indirect to direct forms of intervention, regional intervention in today’s
proxy wars may be better thought of as a menu of options from which policymakers
The Saudi-led coalition’s intervention in Yemen beginning in 2015 is a useful
theory-building case that illustrates these dynamics in action. The coalition of nine
Arab states, led by Saudi Arabia and the United Arab Emirates (UAE), intervened
militarily in Yemen’s civil war beginning on March 26, 2015, at the invitation of the
internationally recognized government, after it had been driven out of the capital
Sana’a by the Houthi insurgent group. A de facto division of labor had the UAE
lead efforts to ﬁght the ground offensive in the south and train an array of local
militant groups there, while Saudi Arabia led an air campaign in the north. The
coalition also blockaded air and sea routes into Yemen.
From the start of the intervention, the United States provided logistical support
to the operation, including aerial refueling for aircraft engaged in airstrikes as well
as targeting assistance and other forms of intelligence-sharing (the United King-
dom and France have also provided support). A joint US–Saudi planning cell in
Riyadh coordinated support for the air campaign. In addition to these more direct
forms of support, the United States also continued to provide Saudi Arabia and the
UAE arms sales packages, military training, and other forms of support throughout
the intervention. While this support is part of long-standing bilateral security part-
nerships, it both sustained these states’ military capacities to carry out such inter-
ventions and demonstrated implicit US support for the intervention (Stark 2020b).
While the UAE began drawing down its forces from Yemen in 2019, the coalition’s
air campaign continued. The United States has also intermittently engaged with
the Houthis and more systematically in a counterterrorism campaign in Southern
Yemen in partnership with the UAE. Since mid-2016, US Special Forces from Joint
Special Operations Command have worked with Emirati Special Forces in Southern
Yemen in what they say is an advisory capacity to counterterrorism operations and
intelligence gathering against Al-Qaeda in the Arabian Peninsula (AQAP). About a
dozen US Green berets have reportedly also been stationed on the Saudi–Yemeni
border since late 2017 to locate and destroy Houthi ballistic missiles and launch
sites (Cooper, Gibbons-Neff, and Schmitt 2018). US forces also conducted missile
strikes on radar facilities in Houthi-controlled territory in October 2016 in response
to anti-ship cruise missiles launched at US Navy warships in the Red Sea (Stewart
First, throughout the Saudi-led coalition’s intervention in Yemen, Saudi Arabia
and the UAE have acted as nodes in a complex sponsor–proxy relationship that
NIKLAS KARLÉN ET AL.15
links the United States indirectly with local actors. Each actor in this chain had
their own interests behind intervening in Yemen, with some overlap in interests as
well as considerable areas of divergence. The United States provided support to
the coalition not because it had a strong interest in the outcome of the local con-
ﬂict and who ultimately governed Yemen but because the Obama Administration
(and later the Trump Administration) felt compelled to support their Gulf security
partners in exchange for their acquiescence to the Joint Comprehensive Plan of
Action (JCPOA) among other issues. The United States’ support for the interven-
tion has meant, however, that it has substantial leverage over the coalition. At times,
this has been used to affect the coalition’s behavior: for example, United States’
pressure reportedly pushed Saudi Arabia to sign the Stockholm Agreement in late
2018 and contributed to the UAE’s decision to begin drawing down in 2019 (Stark
2020c). It is not possible to fully understand the motivations of sponsors or their in-
tervention behavior in Yemen without locating Saudi Arabia and the UAE as nodes
between the United States, on the one hand, and local actors, on the other. Second,
Saudi Arabia and the UAE did not select options along a spectrum of intervention:
rather, they chose to intervene both directly and indirectly in the conﬂict, deploying
their forces not only to ﬁght but also to train and provide other forms of indirect
support to local actors. Similarly, it is difﬁcult to place the United States’ interven-
tion in Yemen in such a two-dimensional spectrum: while most US assistance was
indirect, it did engage sporadically with the Houthis and more systematically in a
counterterrorism campaign in Yemen, working alongside security partners on the
Future scholarship should build on more complex conceptualizations that ex-
pand our fairly narrow heuristics for understanding the relationship between spon-
sors and proxies as also suggested by Karlén and Rauta. As the Yemen case illustrates,
these relationships and intervention options affect the decision to intervene and the
conduct of the intervention itself. Scholarship needs to understand the degree to
which global and regional actors are able to affect the behavior of proxies or, even
better, the conditions under which they can affect this behavior. It also needs to en-
gage with intervention decision-making so that we can build a better understanding
of when and under what conditions an intervention or proxy sponsorship is likely
to occur as well as how and why proxy wars end. Direct intervention and indirect
proxy sponsorship are not mutually exclusive or even necessarily distinct options
that exist on a two-dimensional spectrum of intervention in civil wars; rather, they
often occur in tandem.
Conﬂict Delegation by Non-State Actors
Military Academy at ETH Zurich
A crucial aspect of conﬂict delegation in international affairs is the relationship
between the sponsor and the proxy. However, a key puzzle remains unsolved: what
types of actors can occupy these two roles? Complimentary to Stark’s observations
on the complexity of the sponsor–proxy relationship, this question has hitherto
received an orthodox reply: the existing literature on proxy wars and conﬂict del-
egation suggests that the role of the sponsor is ﬁlled by states, while proxy roles are
16 Conﬂict Delegation in Civil Wars
reserved for non-state actors (NSAs; Hughes 2012,8–11;Groh 2019, 27–29).6Some
scholars (Mumford 2013, 45) have challenged this conventional wisdom, noting
that NSAs as varied as the National Patriotic Front of Liberia, Hezbollah, Al Qaeda
in the Arabian Peninsula, or the People’s Protection Units (YPG) have functioned
as non-state sponsors that have utilized other NSAs as proxies (Moghadam and
Non-state sponsorship involves cooperation between at least two NSAs, which
raises the question of how to distinguish this particular sponsor–proxy relation-
ship from cooperation among militant actors more generally. Greater conceptual
clarity about the similarities and differences of these obviously related phenom-
ena would advance our understanding of an important aspect of contemporary
conﬂict. Cooperation can be deﬁned as formal or informal collaborative arrange-
ments designed to pursue common interests between two or more parties. The
main characteristic of cooperation is rather generic, centering on the notion that
parties design or participate in a common endeavor from which they perceive to
derive a mutual beneﬁt. Without further attributes, however, the concept speci-
ﬁes neither the objects/domains of the arrangement nor the substance and na-
ture of the relationship between the cooperating parties. The same applies to co-
operation among militant actors, which, absent additional speciﬁcations, can take
on a multitude of forms. The speciﬁc collaborative arrangement between militant
groups can vary in terms of its expected duration (i.e., from short- to long-term)
or with respect to the quantity, quality, and variety of activities that the agree-
ment governs. Furthermore, the generic label of cooperation reveals little about
such variables as the degree of interdependence between the cooperating militant
groups, their relative power, their ideological compatibility, or their mutual level of
How, then, do sponsor–proxy relations involving a set of NSAs relate to coopera-
tion among militant groups writ large? We argue that the former is best understood
as a subset of the latter. The use of proxies by non-state sponsors is a form of co-
operation between militant actors in which two additional characteristics typically
apply. The ﬁrst is that the capabilities of the sponsor and proxyare asymmetric, with
the sponsor enjoying privileged status. The second characteristic derives from the
ﬁrst: the sponsor’s political and military objectives take clear precedence over those
of the proxy. Looking at proxy relationships in general, sponsors typically provide a
combination of ﬁnancial and material assistance, as well as training and operational
planning. Proxy assistance usually consists of some combination of ﬁghting a com-
mon adversary, collecting intelligence, patrolling and holding rear areas, and/or
exerting governance on behalf of the sponsor. Since the ﬁghting capabilities of non-
state sponsors typically outweigh those of their proxies, however, non-state sponsors
are less concerned about using their proxies on the frontlines and instead have
them carry out secondary security tasks. Non-state sponsors prize their proxies more
for their political than their military value. Their proxies help them in gaining ac-
cess to or liaising with local communities that typically share a constituency with the
proxy but might be mistrustful of the non-state sponsor’s intentions. In short, non-
state sponsorship is mostly concerned with enhancing and maintaining the spon-
sor’s legitimacy vis-à-vis domestic and/or external stakeholders (Moghadam and
How might we identify a sponsor–proxy relationship involving NSAs in practice?
There are at least four factors that serve as good indicators that a given coopera-
tive relationship between militant actors involves a sponsor–proxy relationship, al-
though not all have to be present. In illustrating these four factors, we will contrast
the sponsor–proxy relationship to what is arguably the most similar collaborative
6Although the contributions by both Stark, Jenne and Popovic and Siroky in this forum also mention governments
as potential agents.
NIKLAS KARLÉN ET AL.17
partnership, namely the relationship between a militant or terrorist alliance hub
(such as Al Qaeda or the Islamic State) that delegates armed conﬂict to its local
afﬁliates (Bacon 2018).
Non-state sponsors will often seek to assert their dominance over their surro-
gates. For example, in the case of the Syrian Democratic Forces, the Kurdish YPG
prevented its proxies from becoming too powerful by fragmenting them, spreading
discord, and detaining key leaders (Wilkofsky and Fatah 2017). Militant hubs, in
contrast, will seek to enforce their strategic outlook on their afﬁliates but often lack
the physical means to impose their will on their partners.
The ability to coerce is to a large degree contingent upon the distance between
the non-state sponsor and its proxy. Physical proximity to their proxies is more
critical for non-state sponsors than for state sponsors, because NSAs typically lack
the military capability or ﬁnancial leverage required to inﬂuence actors from afar.
Practically speaking, non-state sponsors and their proxies will typically be colocated
in the same theater. This distinguishes non-state sponsorship relations from some
related—but nevertheless distinct—forms of cooperation, such as the relationship
between Al Qaeda and most of its afﬁliates and associates. In the latter relation-
ships, strategic headquarters, such as Al Qaeda’s central leadership, are often hard-
pressed to rein in their afﬁliates, who are out of their physical reach. Al Qaeda’s
Pakistan-based leadership, for example, never managed to control its Iraq-based
afﬁliate (Fishman 2016).
Material and/or Financial Preponderance
Non-state sponsors typically command more resources than their proxies, thereby
ensuring the proxy’s dependence on their benefactors. Typical cooperative relation-
ships between militant groups that do not amount to sponsor–proxy relationships
do not always feature such a stark material or ﬁnancial asymmetry. At times, the sup-
posed junior partner in an alliance hub may even be in a better ﬁnancial shape than
the overall headquarters. In the case of the relationship between Al Qaeda and Al
Qaeda in Iraq, for example, the Iraqi afﬁliate eventually surpassed Al Qaeda Central
in ﬁnancial strength, even providing ﬁnancial support to the central leadership in
Pakistan (Byman 2017).
In many instances, the non-state sponsor’s superior resources will result from ex-
ternal support that it is enjoying. The Lebanese Hezbollah, for example, was able
to establish its own proxy militia, the Lebanese Resistance Brigades, in large part
thanks to decade-long Iranian ﬁnancial and material support to the Party of God,
in addition to donations from the Lebanese diaspora (Levitt 2013).
To date, scholars have presented only a handful of case studies of non-state spon-
sorship. For this reason, the list of factors discussed above is far from exhaustive.
With scholarly interest in non-state sponsorship growing, additional factors are
likely to be identiﬁed. We expect these contributions to offer additional insights
into sufﬁcient and necessary conditions of non-state sponsorships, thereby enabling
a clearer drawing of conceptual boundaries between non-state sponsorship and re-
18 Conﬂict Delegation in Civil Wars
Why Peace Rarely Trickles Down: The
Settlement of Mutual Interventions and the
Persistence of Civil Wars in Africa
University of St Andrews
Amutual intervention involves two states that simultaneously intervene in each
other’s intrastate conﬂicts by supporting rebel groups (Cliffe 1999,90).Inprevious
research, we identiﬁed twenty-three mutual interventions in Africa between 1960
and 2010 (ﬁgure 2). Nine of them were reciprocal proxy wars in Mumford’s sense
of involving solely indirect interventions by both states. The other fourteen cases
featured troop intervention by one state, similar to the blend of direct and indirect
support Stark highlights in the Middle East. We also showed that more than half of
Africa’s mutual interventions terminated via negotiated settlements in which both
states committed to ending their support for rebels (Duursma and Tamm 2021).
Here, we reﬂect on why these peaceful interstate settlements rarely led to the termi-
nation of the associated intrastate conﬂicts.
Peace trickled down from the international to the national level in only three of
the conﬂicts linked to mutual interventions. Perhaps most prominently, the Tripar-
tite Accord that resolved the mutual intervention between Angola and South Africa
in 1988 also laid the basis for the resolution of the civil war in Namibia. Crocker,
the US diplomat who successfully mediated the Tripartite Accord, explained that
the resolution of the conﬂict in Namibia was a result of “the right alignment of
local, regional, and international events—like planets lining up for some rare astro-
nomical happening” (quoted in Zartman 1989,234).
The other two intrastate conﬂicts that can be said to have ended as a result of the
prior peaceful settlement of mutual interventions occurred in Sudan. Following the
resolution of the mutual intervention between Sudan and Ethiopia in March 1971,
Ethiopia not only ended its support to the Southern Sudan Liberation Movement
(SSLM) but also began to push for a mediated solution to end the war between the
SSLM and the Sudanese government. This led to the ﬁrst round of secret prelim-
inary peace talks in Addis Ababa in November 1971 (Alier 1990, 80). Subsequent
ofﬁcial negotiations mediated by Emperor Haile Selassie led to the conclusion of
the Addis Ababa Agreement in February 1972 (Beshir 1975, 83–85). Another case
is the resolution of Sudan’s second civil war, which partly resulted from the negoti-
ated settlements of the mutual interventions between Sudan, on the one hand, and
Eritrea, Ethiopia, and Uganda, on the other, in the late 1990s and early 2000s. Dur-
ing these mutual interventions, Eritrea, Ethiopia, and Uganda all worked toward
regime change in Sudan. Yet, following the termination of the mutual interven-
tions, the three neighbors played a constructive role in the peace process that led
to the resolution of the conﬂict between the Sudanese government and the Sudan
People’s Liberation Movement/Army in 2005 (Brosché and Duursma 2018,565).
So why do most intrastate conﬂicts persist after the resolution of a mutual inter-
vention? One could expect that such a resolution makes the resolution of the as-
sociated intrastate conﬂicts more likely because the loss of external support should
shift the balance of power in favor of the government side. This can not only lead
to the defeat of the rebel side, but also widen the bargaining space because a lower
NIKLAS KARLÉN ET AL.19
Figure 2. Mutual interventions in Africa
20 Conﬂict Delegation in Civil Wars
likelihood of winning the war can force the rebel side to accept a negotiated set-
tlement that it would otherwise have refused (Kydd 2010). However, rebel groups
often rely on multiple sources of funding or even manage to ﬁnd new sources of
support after a patron ends its support. For example, the Justice and Equality Move-
ment (JEM) turned to South Sudan for support after Chad resolved the mutual
intervention with Sudan.
A second reason why one could expect that the resolution of a mutual interven-
tion makes the resolution of an intrastate conﬂict more likely is that the former
mutual intervention adversary can act as a mediator in the intrastate conﬂict, as il-
lustrated in the mediation efforts to end Sudan’s ﬁrst and second civil wars. Yet, with
the exception of these two cases in Sudan, this type of mediation effort has gener-
ally failed. South Africa did not succeed in its mediation attempt between Mozam-
bique’s government and the Mozambican National Resistance, following the ter-
mination of the mutual intervention through the Nkomati Accord in March 1984.
Sudan’s mediation effort between the Lord’s Resistance Army and the Ugandan
government between 2006 and 2008 led to several partial peace agreements, but
eventually also failed. Chad mediated between the Sudanese government and JEM
following the termination of the mutual intervention with Sudan in 2010. While this
led to the conclusion of a peace agreement, JEM continued to ﬁght the Sudanese
government. These failed mediation efforts are not that surprising when consider-
ing quantitative research on biased mediation. Biased third parties that previously
supported the rebel side are signiﬁcantly less successful than unbiased mediators or
mediators biased toward the government side (Svensson 2007). Biased mediation is
also unlikely to lead to durable settlements (Duursma 2020).
To conclude, a mutual intervention is a form of interstate conﬂict that by deﬁ-
nition is linked to two intrastate conﬂicts, which raises the question of how mutual
interventions inﬂuence these conﬂicts. We previously found that intrastate conﬂict
dyads in Africa that were at some point involved in a mutual intervention expe-
rienced on average nearly six times as many battle-related deaths as other dyads
(Duursma and Tamm 2021). However, mutual interventions that were resolved via
a negotiated settlement rarely led to the resolution of the associated intrastate con-
ﬂicts. Peace typically fails to trickle down because rebel groups rely on multiple
sources of funding and because biased third parties are ineffective mediators. Over-
all, this suggests that mutual interventions have a greater impact on the severity of
intrastate conﬂicts than on their termination.
Great Power Rivalry and Proxy Wars
ERIN K. JENNE
Central European University
DAVI D S. SIROKY
Arizona State University
The term proxy war has acquired at least two distinct meanings. The ﬁrst denotes
all wars fought by a local proxy on behalf of a third-party sponsor (Mumford 2013;
Salehyan, Siroky, and Wood 2014;San-Akca 2016;Karlén 2017,2019;Rauta 2018,
2020,2021b). In the context of civil wars, this covers indirect armed interventions
by foreign sponsors on behalf of non-state armed proxy forces as part of offensive
interventions against the government. This deﬁnition can be expanded to cover
NIKLAS KARLÉN ET AL.21
a foreign sponsor’s indirect armed support for the embattled government against
the armed opposition as part of defensive interventions (in this forum, Stark like-
wise contends that states, not just non-state armed groups, can serve as proxies in
sponsor–proxy relationships). This second meaning of proxy war, which we focus
on in this brief commentary, covers the subset of proxy wars in which two or more
rival states—often great powers—face off on opposite sides of a civil conﬂict in a
third country (Deutsch 1964;Dunér 1981;Bar-Siman Tov 1984).
One of the earliest post-World War II examples of what we call symmetrical proxy
wars (similar to what Duursma and Tamm call mutual interventions but between
rival powers) was the conﬂict in North Yemen (1948 and 1962) in which the United
Kingdom supported the royalists, along with Jordan, Israel, and Saudi Arabia, while
the Soviet Union along with Egypt championed the republicans. Similarly, the
United States and the USSR squared off in Nicaragua (1981–1989), El Salvador
(1979–1991), and Angola (1975–1991). In both Vietnam and Thailand, meanwhile,
the United States and China assisted opposing sides of the conﬂict (Carson 2016).
While it is true that revanchist states and regional powers have been proliﬁc inter-
veners in the neighborhood (Jenne 2004,2007,2015;Popovic 2017), great powers7
remain some of the most frequent and inﬂuential proxy warriors around the world,
due to their global interests, power-projection capabilities, and status as the world’s
leading arms purveyors.8
In our book-in-progress, we analyze great power side-taking in ongoing civil wars
from 1975 to 2015, and by examining how great powers “take sides” in the conﬂict,
we follow Stark’s admonition in this forum to examine both direct intervention and
indirect foreign sponsorship. We show that great powers have provided weapons,
logistical support, money, air support, and sometimes troops to their favored side
in approximately 80 percent of all civil conﬂicts. Furthermore, by far, the most com-
mon side-taking conﬁguration is defensive, where great powers support governments
against armed opposition groups (ﬁgure 3). Offensive interventions, by contrast,
make up only 8 percent, while symmetrical proxy interventions comprise only 11
percent of great power interventions in ongoing conﬂicts.
The book presents a theory of security hierarchies to account for these patterns in
side-taking. Security hierarchies are composed of arms transfers and security com-
mitments that ﬂow from the great power to their allies and client states. Great powers
preside at the top of these networks, and allied states—with whom great powers share
close economic, political, and ideological ties—stand just below, playing auxiliary or
support roles (such as collective delegation, using Karlén and Rauta’s terminology)
in military engagements. Finally, there are client states, which exist in a strictly sub-
sidiary relationship to their great power patrons (this parallels Salehyan’s insight
in this forum that allies are not in the same hierarchical relationship with sponsor
states as clients/proxies). Security hierarchies are arranged as a “series of paral-
lel power hierarchies, each of which functions similarly to the others and to the
overall international power hierarchy” (Lemke 2002, 48). Security hierarchies may
change over time: new client states may be acquired, either voluntarily or involuntar-
ily, through war, while others may be abandoned (Sylvan and Majeski 2009,33).
How do these prior security relationships inﬂuence great power side-taking?
Drawing on a patron–client logic, we argue that great powers intervene militarily
primarily to support embattled governments that are already embedded in their or
their allies’ security hierarchies. Arms transfers establish a close security relationship
between patron and client in which the client purchases security at the expense of
some of its autonomy; subsequent servicing and support for advanced weapon sys-
tems may lock the seller into providing military support to the government in the
7We operationalize great powers as the permanent ﬁve members (P5) of the United Nations Security Council
8Great powers controlled nearly 90 percent of the global arms market in 2008–2016. The US share amounted to 40
percent (Theohary 2016, 21).
22 Conﬂict Delegation in Civil Wars
Figure 3. Great power side-taking conﬁgurations (1975–2015)
NIKLAS KARLÉN ET AL.23
event of a domestic armed challenge (Kinsella 1998, 9). For their part, great pow-
ers use their arms procurement to inﬂuence the foreign policies of their clients
(Menon 1986, 214) or balance against a rival great power’s inﬂuence (Krause 1991,
Our analysis establishes a pattern of mostly defensive interventions in regime con-
ﬂicts to support besieged governments with which they already have a security rela-
tionship. In marked contrast to most existing research on conﬂict delegation, this
suggests the need to look more at governments rather than non-state armed groups
as proxies. For democratic and nondemocratic powers alike, these ties impact side-
taking choices more than gross human rights violations, regime similarity with the
government, and several other commonly advanced explanations. Great powers are
also more likely to support the embattled governments of former colonies. What
this adds up to is that great power interventions are mostly aimed at defending
client governments within their respective security hierarchies, the boundaries of
which are marked by their and their allies’ arms transfers. One central policy impli-
cation is that great powers ought to be less fearful of rival power interventions, since
defensive interventions rarely elicit corresponding offensive ones from rival great
powers that lead to proxy wars. Even during the Cold War, great powers assumed a
generally defensive rather than offensive military posture when responding to civil
wars around the world.
Making the Clandestine Public: Challenges in
Collecting Data on External Support and
University of Oxford
Research on external support has been equally driven and limited by data availabil-
ity. The type and quality of the data we have does not only affect how rigorously we
can put our theories about external support to the test but also prescribe the very
questions we can feasibly ask, at least in quantitative research designs. As such, the
collection of data is never just a stock-taking exercise. Rather, it frames the academic
debate and sketches possible paths for future research. This is why the Third-Party
Intervention (TPI), Non-State Actor (NSA) and, more recently, Non-State Armed Groups
(NAG) and Uppsala Conﬂict Data Program External Support (ESD) datasets have been
so inﬂuential in shaping academic debate—both in the specialized study of exter-
nal support and proxy wars, and the larger literature on armed conﬂict. Having
worked extensively on the ESD, I believe there are a number of conceptual and
methodological challenges that all data collection efforts on external support face.
Despite much academic interest, as Mumford remarks above, external support
remains a vague concept with many of its key components poorly deﬁned. In fact, a
key theme of the forum is conceptual limitation, and external support shares this with
the related notions of proxy war,delegation,andalliances. This begins with questions
as elementary as which actors should be considered external to a conﬂict: For state
supporters, this is usually a question of where the line between supporting a warring
party and becoming a primary actor oneself lies. The Democratic Republic of the
Congo, for example, ﬁghts Rwandan rebels based predominantly on its territory,
that only contest Rwandan and not Congolese rule, however. Similar problems arise
when states support coalition efforts without offering direct support to a warring
party as we see with United States’ support to the Saudi-led coalition in Yemen
as discussed by Stark. With non-state armed groups operating transnationally,
24 Conﬂict Delegation in Civil Wars
splintering, and switching allegiances, it can be even more difﬁcult to discern
when support originates from the same group in a different location, a different
group altogether, or an independent fraction of a group (e.g., foreign ﬁghters).
While rebel alliances within a conﬂict are not uncommon, they do not introduce
fundamentally new resources to the balance of power and are rightly not included
as external support.
Another key conceptual challenge arises from the need to distinguish external in-
terference in armed conﬂict from the regular workings of the globalized economic
and political order. Financial motives, for example, drive large parts of the sales
of military equipment worldwide and sellers often show little interest in how their
equipment is used. Including all weapon transfers would then risk overcounting
cases of external support. To guard against this, the ESD requires the intent to assist
in a conﬂict as an inclusion criterion. As such, the United States is considered an
external supporter to the Pakistani government in its ﬁght against the Pakistani Tal-
iban, whereas China—its number one source of weapons—is not. This approach,
however, risks undercounting the foreign-provided resources on the ground. Yet, it
is inevitable to choose either a purpose-based or resource-based approach to the
collection of data on external support.
Coding observations on civil war dynamics from media-based, open-source ac-
counts introduces a number of potential reporting biases that have been discussed
in detail elsewhere (Weidmann 2016). When coding external support, there is the
additional challenge of clandestineness and desired deniability that can make it
hard to ﬁnd reliable information. With an increased interest in external support
and the ampliﬁcation of voices from conﬂict zones via social media and other
web-based communication technologies, reporting on it has become more fre-
quent and detailed over the last decade. Hence, data collection efforts are more
likely to pick up on more recent cases of external support. At the same time, of-
ﬁcial government information is only released to the public long after the fact.
Ideally, these countervailing forces balance each other out with time. In prac-
tice, it requires constant updating of existing data and great scrutiny when com-
paring the development of external support over time to distinguish variation
in external support from variation in the availability of information on external
Although a plethora of data on external involvement in civil wars exist, a num-
ber of data “blind spots” remain. These include systematic, reliable, and consis-
tent information on support before and after, and during inactive periods of civil
wars, the termination of support (and reasons for the decision), the quantity and
quality of support (measurements thereof still to be developed), and coalition
contributions—all for both state and non-state recipients. Studies using network
approaches to explore questions related to proxy relationships not only offer a par-
ticularly fruitful way forward but also create a demand for more relational data
(Jackson, San-Akca, and Maoz 2020). The plurality of sources that contain informa-
tion on these topics and the necessity to apply criteria, such as the intent of the
supporter to distinguish them from closely related phenomena, suggest that auto-
mated coding remains, if anything, a distant possibility.
Addressing these data limitations offers numerous ways of moving forward, and
the data collection efforts highlighted here have taken the ﬁrst steps to accomplish
these tasks. Yet, the discrepancy in coding criteria, scope, and units of analysis means
that it is still nearly impossible to draw on the information in one uniﬁed frame-
work.9Advances in data management that allow for the automated integration of
conﬂict event data from multiple sources could facilitate the merging of external
9The NAG includes observations on inactive conﬂict-years and support termination, but only for rebel recipients.
Extensions of the TPI data provide information for years with high risk of political instability, but no conﬂict (Regan
and Meachum 2014). The ESD includes information on coalitions of state supporters, but not states supporting the
coalition effort without providing direct support to the recipient.
NIKLAS KARLÉN ET AL.25
support datasets where operationalizations of concepts permit it. At a minimum,
this would allow to cross-validate ﬁndings by running the same analyses on different
datasets in a much-simpliﬁed manner. Ideally, it would expand to combine data on
external support with data on other forms of foreign interference.
To tackle issues around reliability, it is further paramount to ﬁnd ways to express
the uncertainty that forms part of open source-based data, especially in politically
fragile contexts. We need to develop systematic ways to express this uncertainty, for
example, by training coders to assign uncertainty scores based on their knowledge
of the conﬂict and the sources, by incorporating expert surveys into review pro-
cesses, or by reporting divergent accounts in the data (Weidmann and Rød 2015).
Tools from latent variable models may also be used to formally estimate measures of
uncertainty using various sources of information. Scholars can then model the in-
herent uncertainty, rendering estimations more robust. Still, the work of qualitative
researchers to put these conjectures to the test in case work remains key and greater
collaboration on mixed-methods research designs could help alleviate some of the
data-quality concerns that persist even when the discussed challenges are addressed.
Ways Forward: A Research Agenda on
University College London
University of Oxford
As the authors of this forum have shown, conﬂict delegation is a rich and diverse
ﬁeld of inquiry, be it conceptualized as proxy wars, external support, interventions,
or alliances. However, the forum also shows that there is still a lot that is undertheo-
rized that future research should consider. Mapping out the likely evolution of the
ﬁeld can help answer several questions: What should we know about conﬂict dele-
gation that we do not know yet? What has received little attention? How can new
research challenge or conﬁrm current understandings? In the concluding piece of
this forum, we propose new avenues for research on conﬂict delegation in civil war.
Existing research overwhelmingly conceptualizes the delegation of war as a fun-
damentally dyadic relationship between a principal/sponsor and an agent/proxy,
which is nested within dyadic international rivalries between states (usually, the
sponsor and the target). However, as several of the contributions in this forum
allude to, both the sponsor and the proxy are embedded in overlapping systems
that are signiﬁcantly undertheorized. Further, the preferences and strategies of the
principal typically receive the most attention in the existing literature. This draws
attention to three important points. First, future work should focus more on the
agency of the recipient of external patronage, as existing theories are predomi-
nantly developed from the point of view of the sponsor. Second, not only should
future work determine when and how delegation takes place, but it should also fo-
cus on the changing interaction over time and the tools employed by both actors to
maximize their goals. Finally, the opportunities and risks of providing support from
the point of view of the principal can only be understood by developing theories
incorporating the international dynamic that inﬂuences whether and in what form
intervention takes place.
26 Conﬂict Delegation in Civil Wars
Theorizing Proxy Agency
Some of the most acclaimed studies in proxy warfare explore the ways in which ex-
ternal states accept, address, or circumvent the problem of agency slack (Hughes
2012;Mumford 2013;Groh 2019). Agency slack refers to the risk that the agent
can, once the relationship has been established, shirk its responsibilities by failing
to devote optimal effort or that it can take actions that run counter to the prefer-
ences of the principal (Salehyan 2010, 502). However, we lack the same degree of
attention to the proxy. While scholars readily acknowledge that local actors have
agency, they usually characterize them through an external lens: as insubordinate
actors that need to be controlled (Rauta 2020;Heinkelmann-Wild and Mehrl 2021).
Hardly ever do we see a more direct focus on the recipients of support, their goals
in seeking sponsorship, and their incentives to shirk or go against their patrons’ in-
terests. Proxies exhibit signiﬁcant agency to choose their sponsors as well as decide
how much they will accede to pressure from their sponsors to change their policies
or behavior. San-Akca’s contribution in this forum speaks directly to this.
Future research must pay more attention to the role of the proxy actors as the
main object of research. Speciﬁcally, we need to know more about what proxies
hope to gain out of their relationship with a sponsor. It may not always be mate-
rial factors, such as aid and arms, but rather political support against other political
elites, advice, prestige, or protection from other external actors. Proxies may also
employ multiple strategies to manage their sponsor(s), such as threatening to seek
another sponsor or threatening to collapse, altering their rhetoric or ideology, or
strategically choosing to comply with some demands of the sponsor. Future work
should develop a holistic picture of what the proxy and the various military and po-
litical factions within it want out of the sponsor—and from one another—and the
reciprocal demands they make of the sponsor. Ultimately, proxies are not merely
the pawns of external powers. They have means and strategies to reassert their
Revising principal–agent theory
Once we recognize the agency of proxies, we must also reconsider when sponsors
have more control as well as expand our conceptualization of the tools available to
sponsors to reign in their proxies. Part of this argument is made here by Karlén and
Rauta, starting from the common assumption that the sponsor is thought of as the
principal retaining the ultimate say over support (Salehyan 2010). However, in re-
ality, the proxy has the ability to accept or reject sponsorship (San-Akca 2016); the
sponsor, on the other hand, often lacks a market of available rebel groups to choose
between and will have to content itself with the actors at hand. In some cases, the
proxy will do the job of ﬁghting with or without the principal, who eager to gain
inﬂuence in a conﬂict often must work with the available proxy on terms largely
set by the proxy. In addition, unlike classical principal–agent theory, information
asymmetries are not especially important in determining when sponsors support,
and continue to support, suboptimal proxies. Sponsors often spend substantial time
learning about a prospective proxy before supporting it and invest heavily in mon-
itoring to observe the proxy’s effort and performance after initiating support. As
a result, supporters will often sponsor proxies with eyes wide open about their in-
compatible interests and continue to support proxies despite ample evidence of
disagreements. In other instances, proxies and sponsors begin their relationship
closer together but then drift apart, which sometimes leads to abandonment of the
proxy and sometimes escalation of the sponsor’s involvement. Future scholarship
should then investigate the ongoing and dynamic relationship between sponsors
and proxies. It should also investigate the conditions under which sponsors have
more control over proxies, which may not always be the conditions of complete
information and enforceable contracts like principal–agent theory suggests.
NIKLAS KARLÉN ET AL.27
Future work should also expand our conceptualization of the tools available to
the sponsor to exert greater control over the proxy. How can sponsorsmanage their
relationship with the proxy and induce it to pursue the sponsor’s priorities? Some
work has begun to address when and what kind of incentives are used, and when
they are successful (Elias 2020), but more is needed in this vein. A better alterna-
tive to conceptualizing aid conditionality as the main means of management is to
recognize that different types of support provide different sets of opportunities and
constraints. Sponsors may rely on additional tools besides aid conditionality, such as
energetic diplomacy, civilian and military advisors, and different kinds of weapons
or aid. In short, the study of the reasons behind the form that external support
takes and variation in the effects of assistance on proxy–sponsor cooperation rep-
resents an untapped opportunity for research, which can signiﬁcantly advance our
understanding of the sponsor–recipient relationship.
The international community
Appreciating that there are different relationships between sponsors and proxies
draws attention to our ﬁnal point. If certain forms of support ensure greater de-
grees of control and compliance, then why do states not always opt for these op-
tions? Key in this are constraints and opportunities provided by the international
system, which should urge future work to focus on understanding the risks associ-
ated with the provision of different forms of support. Mitchell (1970) claimed that
in order to understand why and how external states become involved in civil wars,
analysts must explore four sets of factors: (1) within the conﬂict state, (2) within the
external state, (3) the links between the previous two, and (4) the international sys-
tem. While the ﬁrst three factors have received the bulk of attention since, too little
is known about the fourth. Future work should seek to understand how changes in
the international system account for the complex patterns of external involvement
in civil wars “in nonobvious, yet decisive ways” (Kalyvas and Balcells 2010, 427).
Existing scholarship has focused predominantly on the use of external support
or proxy conﬂict to target international rivals. While this is important, the targets
are often not directly the international rivals themselves. For instance, in 2020 the
Libyan conﬂict hosted regional competitive dynamics that had little to do with ri-
valries between external states and the Libyan government, but between Turkey,
Qatar, and Sudan, versus the UAE, Saudi Arabia, and Egypt. Communities of states
coordinated their support to counter other communities of states. Yet they are more
than dyadic relations—they are networks. Future work has much to gain by employ-
ing network theory to understand changing structures of the international network
of states. The decision to provide external support to a rebel group targeting an-
other state is interdependent and strategic, that is, it is conditioned on the actions
and expected actions of other states, whether they occur, are expected to occur,
or, even, do not occur. A network understanding of intervention and external sup-
port can shed light not only on why and how states become involved in foreign
civil wars, but also on the sometimes confusing patterns of external intervention. In
the future, scholars should develop theories to further understand the mechanisms
driving relations between the international system and the process through which
wars become internationalized.
To conclude, the work on conﬂict delegation that this forum surveys and to which
it adds has increased our understanding of why states support non-state actors as
well as begun to address the problems and management of disagreement between
the external supporter and the proxy. However, future scholarship still has to tackle
many important questions on delegated war. We suggest three avenues: shift focus to
the proxy as an agent whose goals and strategies can signiﬁcantly alter the sponsor–
proxy relationship, reexamine the conditions when sponsors have more control
over their proxies and disaggregate the tools that sponsors can use to achieve their
28 Conﬂict Delegation in Civil Wars
sponsorship goals, and re-evaluate the international conditions under which sup-
port happens. Finally, all of these avenues need to be empowered by comprehensive
and systematic data collection efforts that tackle the empirical blind spots identiﬁed
while at the same time building and improving on existing scholarship.
ABBOTT,KENNETH,PHILIPP GENSCHEL,DUNCAN SNIDAL,AND BERNHARD ZANGL. 2020. The Governor’s Dilemma:
Indirect Governance beyond Principals and Agents. Oxford: Oxford University Press.
ALIER,ABEL. 1990. Southern Sudan: Too Many Agreements Dishonoured. Exeter: Ithaca Press.
BACON,TRICIA. 2018. Why Terrorist Groups Form International Alliances. Philadelphia, PA: University of Penn-
BAR-SIMAN-TOV,YAACOV. 1984. “The Strategy of War by Proxy.” Cooperation and Conﬂict 19 (4): 263–73.
BAYLOUNY,ANNE MARIE,AND CREIGHTON A. MULLINS. 2018. “Cash is King: Financial Sponsorship and Chang-
ing Priorities in the Syrian Civil War.” Studies in Conﬂict & Terrorism 41 (12): 990–1010.
BENNEDSEN,MORTEN,AND CHRISTINA SCHULTZ. 2011. “Arm’s Length Delegation of Public Services.” Journal
of Public Economics 95 (7-8): 543–52.
BENOWITZ,BRITTANY,AND ALICIA CECCANESE. 2020. “How the U.N. Can Help Prevent the Spread of Proxy
Conﬂicts.” Just Security. May 27, 2020.
BESHIR,MOHAMED OMER. 1975. The Southern Sudan: From Conﬂict to Peace. London: Barnes & Noble Books.
BIBERMAN,YELENA. 2018. “Self-Defence Militias, Death Squads, and State Outsourcing of Violence in India
and Turkey.” Journal of Strategic Studies 41 (5): 751–81.
BORGHARD,ERICA. 2014. “Friends with Beneﬁts? Power and Inﬂuence in Proxy Warfare.” Ph.D. disserta-
tion, Columbia University.
BORGHARD,ERICA,AND MIRA RAPP-HOOPER. 2013. “Hezbollah and the Iranian Nuclear Programme.” Survival
BORGHARD,ERICA,AND SHAWN W. LONERGAN. 2016. “Can States Calculate the Risks of Using Cyber Proxies?”
Orbis 60 (3): 395–416.
BULOS,NABIH,W.J.HENNIGAN,AND BRIAN BENNETT. 2016. “In Syria, Militias Armed by the Pentagon Fight
those Armed by the CIA.” LA Times, March 27.
BROSCHÉ,JOHAN,AND ALLARD DUURSMA. 2018. “Hurdles to Peace: A Level-of-Analysis Approach to Resolving
Sudan’s Civil Wars.” Third World Quarterly 39 (3): 560–76.
BYMAN,DANIEL. 2017. “Judging Al Qaeda’s Record, Part I: Is the Organization in Decline?” Lawfare,
———. 2020. “Understanding and Misunderstanding, State Sponsorship of Terrorism” Studies in Conﬂict
BYMAN,DANIEL,AND SARAH E. KREPS. 2010. “Agents of Destruction? Applying Principal-Agent Analysis to
State-Sponsored Terrorism.” International Studies Perspectives 11 (1): 1–18.
CAREY,SABINE C., AND NEILL J. MITCHELL. 2017. “Progovernment militias.” Annual Review of Political Science
CARSON,AUSTIN. 2016. “Facing off and Saving Face: Covert Intervention and Escalation Management in
the Korean War.” International Organization 70 (1): 103–31.
CLIFFE,LIONEL. 1999. “Regional Dimensions of Conﬂict in the Horn of Africa.” Third World Quarterly 20
COHEN,DARA KAY,AND RAGNILD NORDÅS. 2015. “Do States Delegate Shameful Violence to Militias? Patterns
of Sexual Violence in Recent Armed Conﬂicts.” Journal of Conﬂict Resolution 59 (4): 1–22.
COOPER,HELENE,THOMAS GIBBONS-NEFF,AND ERIC SCHMITT. 2018. “Army Special Forces Secretly Help Saudis
Combat Threat from Yemen Rebels,” New York Times,May3.
CRAGIN,KIM R. 2015. “Semi-Proxy Wars and US Counterterrorism Strategy” Studies in Conﬂict & Terrorism
38 (5): 311–27.
———. 2020. “Tactical Partnerships for Strategic Effects: Recent Experiences of US Forces Working by,
with, and through Surrogates in Syria and Libya” Defence Studies 20 (4): 318–35.
DESOYSA,INDRA. 2017. “Proxy Wars: Implications of Great-Power Rivalry for the Onset and Duration of
Civil War.” Oxford Research Encyclopedia of Politics. November 4, 2017.
DEUTSCH,KARL W. 1964. “External Involvement in Internal War.” In Internal War, edited by Harry Eckstein,
100–10. New York: Free Press of Glencoe.
DUNÉR,BERTIL. 1981. “Proxy Intervention in Civil Wars.” Journal of Peace Research 18 (4): 353–61.
DUURSMA,ALLARD. 2020. “African Solutions to African Challenges: The Role of Legitimacy in Mediating
Civil Wars in Africa.” International Organization 74 (2): 295–330.
NIKLAS KARLÉN ET AL.29
DUURSMAALLARD,AND TAMM,HENNING. 2021. “Mutual Interventions in Africa.” International Studies Quarterly
Online First: 1–10.
ECK,KRISTINE. 2015. “Repression by Proxy: How Military Purges and Insurgency Impact the Delegation of
Coercion.” Journal of Conﬂict Resolution 59(5): 924–46.
ELIAS,BARBARA. 2020. Why Allies Rebel: Deﬁant Local Partners in Counterinsurgency Wars. Cambridge: Cam-
bridge University Press.
FISHMAN,BRIAN. 2016. The Master Plan: ISIS, Al-Qaeda, and the Jihadi Strategy for Final Victor y. New Haven,
CT: Yale University Press.
FISK,KERSTIN. 2021. “Plausible Deniability? An Investigation of Government and Government-Outsourced
Violence in Refugee Hosting Areas.” International Interactions 47 (2): 191–220.
GRAUER,RYAN ,AND DOMINIC TIERNEY. 2018. “The Arsenal of Insurrection: Explaining Rising Support for
Rebels.” Security Studies 27(2): 263–95.
GROH,TYRONE. 2019. Proxy War: The Least Bad Option. Stanford, CA: Stanford University Press.
HAZEN,JENNIFER. 2013. What Rebels Want: Resources and Supply Networks in Wartime. Ithaca, NY: Cornell
HEINKELMANN-WILDTIM,AND MEHRL,MARIUS. 2021. “Indirect Governance at War: Delegation and Orches-
tration in Rebel Support.” Journal of Conﬂict Resolution Online First: 1–29.
HELLMÜLLER,SARA. 2021. “The Challenge of Forging Consent to UN Mediation in Internationalized Civil
Wars: The Case of Syria.” International Negotiation 26: 1–28.
HUANG,REYKO,AND PATRICIA L. SULLIVAN. 2021. “Arms for Education? External Support and Rebel Social
Services.” Journal of Peace Research 58 (4): 794–808.
HUGHES,GERAINT. 2012. My Enemy’s Enemy: Proxy Warfare in International Politics. Brighton: Sussex Academic
IRRERA,DANIELA. 2021. “Non-state Actors and Conﬂict Management in Proxy Wars.” Oxford Research
Encyclopedia of International Studies.
JACKSON,JAMIE A., BELGIN SAN-AKCA,AND ZEEV MAOZ. 2020. “International Support Networks and the Calcu-
lus of Uprising.” Journal of Peace Research 57 (7): 632–47.
JENNE,ERIN K. 2004. “A Bargaining Theory of Minority Demands: Explaining the Dog that Did not Bite
in 1990s Yugoslavia.” International Studies Quarterly 48 (4): 729–54.
———. 2007. Ethnic Bargaining: The Paradox of Minority Empowerment. Ithaca, NY: Cornell University Press.
———. 2015. Nested Security: Lessons in Conﬂict Management from the League of Nations and the European
Union. Ithaca, NY: Cornell University Press.
KALYVAS,STATHIS N., AND LAIA BALCELLS. 2010. “International System and Technologies of Rebellion: How
the End of the Cold War Shaped Internal Conﬂict.” American Political Science Review 104 (3): 415–29.
KARLÉN,NIKLAS. 2017. “The Legacy of Foreign Patrons: External State Support and Conﬂict Recurrence.”
Journal of Peace Research 54 (4): 499–512.
———. 2019. “Turning off the Taps: The Termination of State Sponsorship.” Terrorism and Political Vio-
lence 31 (4): 733–58.
———. 2020. “Escalate to De-Escalate? External State Support and Government’s Willingness to Negoti-
ate.” Studies in Conﬂict & Terrorism.
KINSELLA,DAV ID . 1998. “Arms Transfer Dependence and Foreign Policy.” Journal of Peace Research 35 (1):
KRAUSE,KEITH. 1991. “Military Statecraft: Power and Inﬂuence in Soviet and American Arms Transfer
Relationships.” International Studies Quarterly 35 (3): 313–36.
KYDD,ANDREW H. 2010. “Rationalist Approaches to Conﬂict Prevention and Resolution.” Annual Review of
Political Science 13: 101–21.
LAFFONT,JEAN-JACQUES,AND DAV ID MARTIMORT. 1998. “Collusion and Delegation.” The RAND Journal of Eco-
nomics 29 (2): 280–305.
LARSDOTTER,KERSTI. 2014. “Regional Support for Afghan Insurgents: Challenges for Counterinsurgency
Theory and Doctrine.” Journal of Strategic Studies 37 (1): 135–62.
LEE,MELISSA. 2020. Crippling Leviathan. How Foreign Subversion Weakens the State. Ithaca, NY: Cornell Uni-
LEMKE,DOUGLAS. 2002. Regions of War and Peace. Cambridge: Cambridge University Press.
LEVITT,MATTHEW. 2013. Hezbollah: The Global Footprint of Lebanon’s Party of God. Washington, DC: George-
town University Press.
LISTER,CHARLES. 2016. “The Free Syrian Army: A Decentralized Insurgent Brand.” The Brookings Project
on U.S. Relations with the Islamic World, Analysis Paper No. 26.
MANEKIN,DEVORAH S., AND REED MWOOD. 2020. “Framing the Narrative: Female Fighters, External Audi-
ence Attitudes, and Transnational Support for Armed Rebellions.”Journal of Conﬂict Resolution 64 (9):
30 Conﬂict Delegation in Civil Wars
MAOZ,ZEEV,AND BELGIN SAN-AKCA. 2012. “Rivalry and State Support of Non-State Armed Groups (NAGs),
1946-2001.” International Studies Quarterly 56 (4): 720–34.
MENON,RAJAN. 1986. “Soviet Arms Transfers to the Third World: Characteristics and Consequences.”
Journal of International Affairs 40 (1): 59–76.
MICHAELOWA,KATHARINA,BERNHARD REINSBERG,AND CHRISTINA J. SCHNEIDER. 2018. “The Politics of Double
Delegation in the European Union.” International Studies Quarterly 62 (4): 821–33.
MITCHELL,CHRISTOPHER. 1970. “Civil Strife and the Involvement of External Parties.” International Studies
Quarterly 14 (2): 166–94.
MOGHADAM,ASSAF,AND MICHEL WYSS. 2020. “The Political Power of Proxies: Why Nonstate Actors Use Local
Surrogates.” International Security 44 (4): 119–57.
MUMFORD,ANDREW. 2013. Proxy Warfare. Cambridge: Polity.
NIELSON,DANIEL L., AND MICHEAL J. TIERNEY. 2003. “Delegation to International Organizations: Agency
Theory and World Bank Environmental Reform.” International Organization 57 (2): 241–76.
PETROVA,MARINA G. 2019. “What Matters Is Who Supports You: Diaspora and Foreign States as Ex-
ternal Supporters and Militants’ Adoption of Nonviolence.” Journal of Conﬂict Resolution 63 (9):
PLANA,SARA. 2020. “Seven Bad Options to Counter State Sponsorship of Proxies.” Lawfare. September 13,
PLATONOVA,DARIA. 2021. The Donbas Conﬂict in Ukraine. Elites, Protest and Partition. London: Routledge.
POPOVIC,MILOS. 2017. “Fragile Proxies: Explaining Rebel Defection against Their State Sponsors.” Terro r-
ism and Political Violence 29 (5): 922–42.
———. 2018. “Inter-Rebel Alliances in the Shadow of Foreign Sponsors.” International Interactions 44 (4):
RAUTA,VLADIMIR. 2018. “A Structural-Relational Analysis of Party Dynamics in Proxy Wars.” International
Relations 32 (4): 449–67.
———. 2020. “Proxy Warfare and the Future of Conﬂict: Take Two.” The RUSI Journal 165 (2): 1–10.
———. 2021a. “Framers, Founders, and Reformers: Three Generations of Proxy War Research.” Contem-
porary Security Policy 42 (1): 113–34.
———. 2021b. “‘Proxy War’: A Reconceptualization.” Civil Wars 23 (1): 1–24.
REGAN,PATRICK M. 2002. “Third Party Interventions and the Duration of Intrastate Conﬂicts.” Journal of
Conﬂict Resolution 46 (1): 55–73.
REGAN,PATRICK M., AND AYSEGUL AYDIN. 2006. “Diplomacy and Other Forms of Intervention in Civil Wars.”
Journal of Conﬂict Resolution 50 (5): 736–56.
REGAN,PATRICK M., AND M. SCOTT MEACHUM. 2014. “Data on Interventions during Periods of Political In-
stability.” Journal of Peace Research 51 (1): 127–35.
SALEHYAN,IDEAN. 2010. “The Delegation of War to Rebel Organizations.” Journal of Conﬂict Resolution 54
SALEHYAN,IDEAN,KRISTIAN GLEDITSCH,AND DAV ID A. CUNNINGHAM. 2011. “Explaining External Support for
Rebel Groups.” International Organization 65 (4): 709–44.
SALEHYAN,IDEAN,DAV ID SIROKY,AND REED WOOD. 2014. “External Rebel Sponsorship and Civilian Abuse: A
Principal-Agent Analysis of Wartime Atrocities.” International Organization 68 (3): 633–61.
SAN-AKCA,BELGIN. 2014. “Democracy and Vulnerability: An Exploitation Theory of Democracies by Ter-
rorists.” Journal of Conﬂict Resolution 58 (7): 1285–310.
———. 2016. States in Disguise: Causes of State Support for Rebel Groups. New York: Oxford University Press.
SAWYER,KATHERINE,KATHLEEN GALLAGHER CUNNINGHAM,AND REED WILLIAM. 2017. “The Role of External Sup-
port in Civil War Termination.” Journal of Conﬂict Resolution 61 (6): 1174–202.
STARK,ALEXANDRA. 2020a. “Give Up on Proxy Wars in the Middle East.” Foreign Policy. August 7, 2020.
———. 2020b. “The Monarchs’ Pawns? Gulf State Proxy Warfare 2011-Today.” New America. June 15,
———. 2020c. “Prospects for Ending External Intervention in Yemen’s War.” In MENA’s Frozen Conﬂicts
(POMEPS Studies 42), edited by Marc Lynch, 74–80. Washington, DC: POMEPS.
STEIN,ART HUR,AND MARC-OLIVIER CANTIN. 2021. “Crowding out the Field: External Support to Insurgents
and the Intensity of Inter-Rebel Fighting in Civil Wars.” International Interactions 47 (4): 662–91.
STEWART,PHIL. 2016. “U.S. Military Strikes Yemen after Missile Attacks on U.S. Navy Ship.” Reuters,
SVENSSON,ISAK. 2007. “Bargaining, Bias and Peace Brokers: How Rebels Commit to Peace.” Journal of Peace
Research 44 (2): 177–94.
SYLVAN,DAVI D,AND STEPHEN MAJESKI. 2009. U.S. Foreign Policy in Perspective: Clients, Enemies and Empire.Lon-
NIKLAS KARLÉN ET AL.31
TAMM,HENNING. 2016a. “The Origins of Transnational Alliances: Rulers, Rebels, and Political Survival in
the Congo Wars.” International Security 41 (1): 147–81.
———. 2016b. “Rebel Leaders, Internal Rivals, and External Resources: How State Sponsors Affect In-
surgent Cohesion.” International Studies Quarterly 60 (4): 559–610.
THEOHARY,CATHARINE A. 2016. Conventional Arms Transfers to Developing Nations, 2008-2015. Washington,
DC: Congressional Research Service.
WEIDMANN,NILS B. 2016. “A Closer Look at Reporting Bias in Conﬂict Event Data.” American Journal of
Political Science 60 (1): 206–18.
WEIDMANN,NILS B., AND ESPEN GEELMUYD RØD. 2015. “Making Uncertainty Explicit: Separating Reports and
Events in the Coding of Violence and Contention.” Journal of Peace Research 52 (1): 125–28.
WILKOFSKY,DANIEL,AND KHALID FATAH. 2017. “Northern Syria’s Anti-Islamic State Coalition Has an Arab
Problem.” War on the Rocks, September 18.
ZARTMAN,WILLIAM I. 1989. Ripe for Resolution: Conﬂict and Inter vention in Africa. Oxford: Oxford University