Where You Sit Matters: The Power of Brokers in Diplomatic Networks & Interstate Conflict

Abstract

How does a state’s broker position in diplomatic networks influence its tendency to engage in conflict? While the existing scholarship typically characterizes brokers as ideal mediators or facilitators of peace, there is little systematic analysis that examines the impact of broker position on a state’s conflict propensity. I argue that the exclusive emphasis on the mediating role of brokers has elided the fact that broker position serves as strategic assets for states. Building on the existing literature on diplomacy and brokerage, I posit that broker position in diplomatic networks provides a state with leverage that can be used to coerce and co-opt other states into supporting its military operations, thereby increasing its propensity to initiate militarized interstate disputes. I assemble diplomatic networks from the past two centuries to examine the impact of broker position and find that the more a state occupies a broker position, the more likely it is to initiate militarized disputes. These findings challenge the prevalent notions that diplomatic ties are insignificant in shaping state conflict behavior and that broker position is solely used to promote peace.

¿Cómo influye el posicionamiento como intermediario dentro de las redes diplomáticas por parte de un determinado Estado en su tendencia a participar en conflictos? Si bien el mundo académico actual caracteriza, generalmente, a los intermediarios como mediadores ideales o como facilitadores de la paz, existe un número reducido de análisis sistemáticos que examinen el impacto que tiene el hecho de estar posicionado como intermediario sobre la propensión al conflicto de ese Estado. Argumentamos que el énfasis exclusivo en el papel como mediador de los intermediarios ha omitido el hecho de que la posición como intermediario sirve como activo estratégico para los Estados. Postulamos, partiendo de la base que proporciona la bibliografía existente sobre diplomacia e intermediación, que el estar posicionado como intermediario dentro de las redes diplomáticas proporciona a un determinado Estado un grado de influencia que puede usarse para coaccionar y cooptar a otros Estados para que apoyen sus operaciones militares, lo cual aumentaría, de esta forma, su propensión a iniciar disputas interestatales militarizadas. Hemos reunido diversas redes diplomáticas formadas durante los últimos dos siglos con el fin de estudiar el impacto que tiene el posicionamiento como intermediario y llegamos a la conclusión de que cuanto más ocupa un Estado una posición de intermediario, más probable resulta que pueda iniciar disputas militarizadas. Estas conclusiones desafían las nociones prevalentes de que las relaciones diplomáticas resultan insignificantes para la confirmación del comportamiento en materia de conflicto estatal y de que la posición de intermediario se utiliza únicamente con el fin de promover la paz.

Quels sont les effets du rôle de négociateur d’un État au sein de réseaux diplomatiques sur sa propension à prendre part à un conflit ? Bien que la littérature existante définisse généralement les négociateurs comme des médiateurs idéaux ou des facilitateurs de paix, il existe peu d’analyses systématiques qui s’intéressent aux conséquences du rôle de négociateur sur la propension au conflit d’un État. J’affirme qu’en mettant exclusivement l’accent sur le rôle de médiation des négociateurs, nous avons éludé le fait que la position de négociateur constitue un avantage stratégique pour un État. En me fondant sur la littérature existante sur la diplomatie et la négociation, j’émets l’hypothèse qu’une position de négociateur au sein de réseaux diplomatiques fournit à un État un avantage qu’il peut utiliser pour contraindre et coopter d’autres États à soutenir ses opérations militaires. Aussi, sa propension à initier des conflits interétatiques militarisés augmente. Je rassemble des données sur les réseaux diplomatiques des deux derniers siècles pour analyser les effets du rôle de négociateur. J’observe que plus un État occupe un rôle de négociateur, plus il aura de chance d’initier des conflits militarisés. Ces conclusions remettent en question les notions prédominantes selon lesquelles les liens diplomatiques n’ont pas d’incidence sur le comportement d’un État en matière de conflits et que le rôle de négociateur n’est utilisé que pour promouvoir la paix.

Keywords:

• Broker
• diplomatic networks
• militarized interstate disputes
• network analysis

Acknowledgement

I thank anonymous reviewers for thoughtful comments and invaluable feedback. I am also grateful to Douglas Kriner, Dino Christenson, Manjari Miller, and Rosella Cappella Zielinski, and panel participants at the APSA 2020 conference.

Notes

1 States can use the information to convincingly pitch their policy preferences on a particular issue, increase the chance that their policy measures will succeed, and may even persuade others to refrain from opposing their preferences by emphasizing the potential adverse consequences of such opposition (Maller 2010, 67–68).

2 While they do not explicitly focus on diplomatic networks per se, some scholars find that IGO networks significantly shape the politics of war and peace (e.g., Boehmer, Gartzke, and Nordstrom 2004; Dorussen and Ward 2008; Hafner-Burton and Montgomery 2006). According to Hafner-Burton, Von Stein, and Gartzke (2008, 181) “network analysis is useful not only because it provides insight into how states’ overall IO membership portfolio (rather than their participation in one regime alone) shapes behavior but also because it considers how the connections generated by membership affect the mechanisms of cooperation.”

3 Renshon (2017) is a notable exception. He contributes significantly to our understanding of how diplomatic networks influence a state’s conflict propensity by showing that status deficient countries are more likely to initiate interstate conflict. Yet, his focus lies in a state’s status deficit—measured as a discrepancy between material capabilities and structural position within a diplomatic network—rather than the position itself. Thus, my purpose here is to pay explicit attention to a state’s structural position, specifically its broker position in the diplomatic network.

4 Following the existing scholarship, I consider social power as deriving from a particular notion of social capital, which is defined as “the aggregate of the actual or potential resources which are linked to possession of a durable network of more or less institutionalized relationships of mutual acquaintance or recognition” (Bourdieu 2018, 21). Hafner-Burton and Montgomery (2006) argue that the source of social capital has been examined in two respects: the ability to bridge structural holes and the ability to connect many actors. My article focuses on the former.

5 While IR scholarship less explicitly references the concept of leverage, or often uses it interchangeably with influence or as a means to generate influence, I regard leverage as “the creative use of resources and relationships intended to bring about certain effects in the world” (Friman 2015, 203). An actor’s leverage stems from a wide array of resources, including the possession of nuclear weapons, economic resources, or policy instruments. This article considers a state’s broker position as a meaningful source of leverage.

6 The co-optive power corresponds to the second and third faces of power, contributing to one’s soft power (i.e.,“the ability to get preferred outcomes through the co-optive means of agenda setting, persuasion, and attraction” (Nye 2011, 16)).

7 While Henke focuses on degree centrality as an indicator of a state’s capacity to establish military coalitions, I emphasize betweenness centrality because it captures not only a state’s ability to gather information but also its ability to manipulate the information for its own benefit. In addition, the information that can be obtained through bridging structural holes or weak ties can be more heterogeneous and reliable due to its varied sourcing and cross-checked validity (Granovetter 1973).

8 A network consists of a set of nodes linked by a set of edges. It is formally represented as an n x n adjacency matrix, where matrix entry $X_{ij}$ indicates the existence of ties from i to j. In this article, the adjacency matrix shows the presence of ambassador ties from each state to every other state.

9 The dataset records the diplomatic exchange for the years 1817, 1824, 1827, 1832, 1836, 1840, every 5 years between 1844 and 1914, and then 1920, 1925, 1930, 1935, 1940, and then every 5 years between 1950 and 2005. I assembled diplomatic networks accordingly at those years.

10 The military capabilities were obtained from the Correlates of War (COW) dataset (Singer, David, and Stuckey 1972). I used a conventional indicator of the CINC (Composite Index of National Capabilities) to measure a state’s military capability.

11 Although interstate conflict arises from a series of strategic interactions between two states, the initiation of disputes ultimately requires the decision of only one state. Existing studies that examine the factors that influence a state’s opportunity and willingness to initiate MIDs similarly posit monadic arguments (Fordham 2004; Gill-Tiney 2022; Most and Starr 1989; Chiba, Martinez Machain, and Reed 2014). For example, a state’s greater military capabilities and/or major power status increase its opportunity and willingness to engage in aggression, regardless of the characteristics of its opponent. Following this literature, I consider a state’s broker position as a monadic variable and examine whether and to what extent it increases a state’s willingness to initiate MIDs.

12 Note that diplomatic networks are constructed in largely 5-year intervals. Small and Singer (1973), the original collectors of the data of diplomatic representation, recommend researchers who want to use it as an annual variable to either: (1) select the last available data until the next observation, or (2) interpolate the missing years. I chose the latter approach to calculate a country’s betweenness centrality each year.

13 In line with Small and Singer (1982), the following countries are regarded as a major power in the international system: the United Kingdom (1816–present), France (1816–1940, 1945–present), Prussia/Germany (1816–1918, 1925–1945), Austria-Hungary (1816–1918), Russia/the Soviet Union (1816–1917, 1922–present), Italy (1860–1943), Japan (1895–1945), the United States (1899–present), and China (1950–present).

14 The dependent variable was constructed based on the MID dataset (version 3.10) of the Correlates of War (COW) project corrected by Maoz et al. (2019).

15 I also employed cubic polynomials of peace years and used a lagged dependent variable, both of which are alternative ways to address temporal dependence in the dependent variable (Carter and Signorino 2010; Renshon 2017). A peace year is measured as the number of years that a dyad has been at peace since the previous conflict. The results with lagged dependent variables, cubic splines, and cubic polynomials of peace years are largely consistent.

16 I account for temporal dependence by specifying an AR1 lag structure.

17 I believe that betweenness centrality best captures the notion of brokers, but note that a consistent pattern of results can be found with alternative network centrality measures, such as degree centrality, eigenvector centrality, and information centrality.

18 To account for the fact that standard logit regression might underestimate low-probability events, I also employed rare-events logistic regression (King and Zeng 2001) and found no substantive differences in the results. The rare events logistic estimates are available in the online appendix.

19 I appreciate the reviewers for pushing me to develop this point.

20 Heckman selection models need an exclusion restriction to reduce the problem of collinearity between the selection and outcome equations (Heckman 1979). In line with previous studies on conflict escalation (e.g., Peterson and Graham 2011; Hwang 2010), I exclude the cubic polynomials from the second stage equation as the exclusion restriction.

21 The lack of statistical significance in the escalation stage could be attributed to either the increased possibility that an opponent concedes before escalation or that the challenger chooses to back down due to the high costs associated with escalation. Although investigating the differential impact of broker position on dispute initiation and escalation falls outside the scope of this manuscript, it presents an intriguing avenue for future research.

22 The betweenness centralities of China from 1970 to 2005 in 5-year intervals are as follows: 170.63, 121.57, 231.75, 1115.65, 1297.48, 1951.65, 2078.61, and 2807.52. Those of the United States over the same period are as follows: 1763.70, 1290.83, 1305.15, 2627.46, 2801.90, 4544.20, 4262.45, and 3953.95.