Signaling, Extended Deterrence, and Regional Conflict Escalation: A Game Theory Model

In my dissertation I seek to determine the conditions of regional conflict escalation. In particular, I pose two questions in this dissertation: first, why do small states decide to fight against relatively more powerful states in spite of very low likelihood of winning?; and second, once regional conflicts occur, why do extra-regional states decide to join one of the regional contestants even when they expect a minimal gain? To understand this theoretical puzzle, I set up a signaling game model which implements an extended deterrence situation. I solved the game for perfect Bayesian equilibrium and got seven equilibria, five of them pooling and two separating. Further examination of these equilibria revealed that only two patterns of equilibrium behavior are substantially interesting: successful extended deterrence and failed extended deterrence. I found that the key difference between the two outcomes lies in the quality and the scope of pre-conflict signals. That is, the deterrence success appears to be highly associated with credibility, intensity, and coherence of pre-conflict signals. To test this finding, I chose to conduct case studies. Particularly, I followed the methods of structured, focused comparison to systematically compare the Korean War of 1950 as a failed deterrence case and Quemoy-Matsu Crisis (or First Taiwan Straits Crisis) of 1954 as a successful deterrence case. The analysis of pre-conflict exchange of signal-response sequences in both cases largely confirms my argument.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Political Science

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• Lee, Sang-Hyun

Contributors dc:contributor

• Zinnes, Dina A.

• Political Science, General

Language dc:language

eng

Identifier

(MiAaPQ)AAI9944919

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/82602