Danger Beyond Dyads: Third-Party Participants in Militarized Interstate Disputes

Abstract

Stuart Bremer often reminded us that third parties—directly or indirectly—affect the initiation, evolution, and termination of conflict. He encouraged scholars to research the phenomenon of joining behavior further and personally investigated it. Questions about joining behavior are indeed deeply intertwined with a variety of theories of conflict. However, existing records on third-party interventions are limited to states' military involvement in conflict. The limitations imposed by the data can lead researchers to biased or incomplete conclusions about many international phenomena. We heed Bremer's encouragement and present here the results of an effort to collect new evidence on nonneutral (partisan) interventions in militarized interstate disputes for the 1946—2001 period. The data we present differ from existing records in that: (1) they provide information on both third parties' military and nonmilitary activities; (2) they broaden the notion of what constitutes a third party by including coalitions of states, intergovernmental organizations (IGOs), and nongovernmental organizations (NGOs); (3) they expand the investigation framework by recording interventions that occur before and after a militarized dispute. We test the usefulness of the data by exploring the issue of major powers' interventions in conflicts, as Yamamoto and Bremer did in their 1980 “Wider Wars and Restless Nights” article. We offer strong support for Yamamoto and Bremer's finding that major powers drag one another into ongoing conflicts and show how the data may help us raise and answer new and more complex hypotheses about third parties and the dynamics of joining behavior.

Determinants of Third Parties’ Intervention and Alignment Choices in Ongoing Conflicts, 1946–2001

“…However, unlike previous studies of joining behavior, both militarized and nonmilitarized interventions are considered, first together, then separately. The data that allow for this coding come from Corbetta and Dixon (2005) and are described more fully below 11…”

“…The data track military and low‐intensity, nonmilitary (economic and diplomatic) interventions over the course of a dyad’s conflictual history for the 1946–2001 period 14 . In Corbetta and Dixon’s (2005) data set, military interventions range from the offer and provision of military assistance to actual fighting by a third party on the side of one of the combatants. Nonmilitary interventions, instead, include: statements of support or opposition toward one side in the conflict; appeals/demands for cease‐fire or troop withdrawal; promises of economic assistance; promises of military assistance; direct diplomatic assistance; diplomatic sanctions; offer and provision of economic assistance; economic sanctions 15…”

“…The data on third‐party interventions from Corbetta and Dixon (2005) were integrated with dyadic MID data (Maoz 2005) to produce a data structure in which each observational unit represents a decision by a potential joiner to enter the conflict and align itself with either disputant during each year of a dispute. With the exception of micro‐states, all members of the international system are considered as potential third parties during each dispute‐year.…”

“…Rather, we can think of intervention as a continuous variable ranging from indifference to the conflict and abstention to full‐scale military involvement. Corbetta and Dixon (2005), for instance, identify a series of joining techniques of increasing intensity ranging from verbal statements by a third party state’s public officials to full military involvement. When only extreme values of all of the possible realizations of the dependent variable are used, researchers are in effect dealing with a truncated sample.…”

“…This paper proposes a model of joining behavior that includes both security considerations and homophilous network ties as determinants of third parties’ intervention and alignment choices. The model is tested on military and nonmilitary interventions data from Corbetta and Dixon (2005) for the 1946–2001 period. The results from a boolean logit estimation indicate that (1) the intervention decision is the result of both security‐driven, pragmatic considerations and homophily between joiners and disputants; and (2) opposition to a party in a conflict is as important as a third party’s social proximity to the side being supported.…”