Deterrence and Restraint: Do Joint Military Exercises Escalate Conflict?

• Published date: 01 January 2022
• Date: 01 January 2022
• Author: Raymond Kuo,Brian Dylan Blankenship
• DOI: 10.1177/00220027211023147
• Subject Matter: Articles

Deterrence and

Restraint: Do Joint

Military Exercises

Escalate Conflict?

Raymond Kuo

1

and Brian Dylan Blankenship

2

Abstract

Multinational military exercises are among the most notable demonstrations of

military cooperation and intent. On average, one is initiated every 8.9 days. But it has

often been argued that joint military exercises (JMEs) increase the risk of war. Using

a relational contracting approach, we claim that formal military alliances mediate the

effect of JMEs. Exercises and alliances serve complementary functions: The former

allows targeted responses to military provocations by adversaries, while the latter

provides institutional constraints on partners and establishes a partnership’s overall

strategic limitations. In combination, alliances dampen the conflict escalation effects

of exercises, deterring adversaries while simultaneously restraining partners. We

test this theory using a two-stage model on directed dyadic data of JMEs from 1973

through 2003. We find that JMEs in general do not escalate conflict, and that JMEs

conducted with allies in particular reduce the probability of conflict escalation.

Keywords

military alliance, restraint, relational contracting, moral haza rd, deterrence, joint

military exercises, signaling

1

Minneapolis, MN, USA

2

Department of Political Science, University of Miami, Coral Gables, FL, USA

Corresponding Author:

Brian Dylan Blankenship, University of Miami, 1300 Campo Sano Ave., Office 140E, Coral Gables, FL

33124, USA.

Email: bdb2125@columbia.edu

Journal of Conflict Resolution

ªThe Author(s) 2021

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/00220027211023147

journals.sagepub.com/home/jcr

Article

2022, Vol. 66(1) 3 –31

4 Journal of Conflict Resolution 66(1)

Introduction

In the wake of his 2018 summit with No rth Korean leader Kim Jong Un, U.S.

President Donald Trump called for a halt to what he described as U.S.-South Korean

“war games.” Military exercises between the two countries were “very provocative”

and “very expensive.”

1

The Kim regime had previously stated that Vigilant Ace, a

large joint military exercise intended to signal allied resolve to denuclearize the

Korean peninsula, would bring the parties “to the brink of war.” In response,

Washington and Seoul postponed a subsequent, annual exercise—Foal Eagle—until

after the 2018 Winter Olympics.

Concerns thatmultinational military exercises would lead to open militaryengage-

ments are not without precedent or foundation. In another case ten years prior, Geor-

gia’s participation in Sea Breeze 2008, an exercise with sixteen country participants,

including five NATO members, was followed one month later by the 2008 Russia-

Georgia War. In combination with other signals, Georgian President Saakashvili felt

he had Western support to take increasingly assertive steps to counteract Russian

support for the breakaway provinces of South Ossetia and Abkhazia. This culminated

in military incursions against the capital of the self-proclaimed Republic of South

Ossetia on August8. Russian forces thenofficially intervened, leadingto open warfare

between the two sides (Driscoll and Maliniak 2016; Lanoszka 2018).

Yet in many cases, exercises have not escalated conflict. NATO’s annual Exer-

cise REFORGER never produced direct engagements between NATO and the War-

saw Pact throughout the 1970s and 1980s (Blackwill and Legro 1989).

Consequently, in this paper, we ask: Under what conditions do joint military exer-

cises (JMEs) escalate conflict?

2

We use a relational contracting approach to argue

that the effect of JMEs depends on whether they are conducted between states that

have a military alliance. Outside of an alliance context, JMEs carry the potential to

increase the likelihood of war and interstate fighting by both participants and targets.

Emboldened by new capabilities and outside support, exercise participants escalate

conflict. Following the spiral model (Jervis 1976, ch. 3), their targets respond simi-

larly. However, JMEs conducted within an alliance reduce escalation by both parti-

cipants and adversaries. Alliances and exercises complement one another. States use

the former to establish the parameters of their security relationship and constrain

adventurism. They use the latter to more flexibly respond to short-term shifts in the

military balance. In combination, partners can enhance their ability to restrain one

another while maintaining a specific level of deterrence.

We use statistical analysis to test this theory. Following Braithwaite and Lemke

(2011), we use two-stage models to, first, partial out the factors leading states to

initiate conflict with one another, then run a second stage modeling the effect of

JMEs on conflict escalation. We show that when conducted with alliance partners,

these exercises reduce escalation by both their participants and their targets. The

results are subjected to a battery of robustness checks presented here and in the

Online Appendix.

2Journal of Conflict Resolution XX(X)