Citation

Zagare FC. J. Peace Res. 1992; 29(4): 435-454.

Copyright

(Copyright © 1992, SAGE Publishing)

DOI

10.1177/0022343392029004006

PMID

unavailable

Abstract

A multi-stage escalation model is developed in which each stage is taken as a psychologically distinct and empirically feasible mode of conflict available to two states in a hostile, deterrent relationship. Each stage is defined by a generic, extensive-form game of mutual deterrence in which the credibility of each player's threat is allowed to vary. In the multi-stage model these games are linked to one another by way of a common outcome, creating a new game with a distinct dynamic. The results of previous research on one- and two-stage games are summarized and the model extended to games with three stages. A three-stage escalation game composed solely of one of the three possible one-stage deterrence games has strategic characteristics identical to those of its component parts: mutual credibility at each stage implies stable deterrence; mutually incredible threats at every stage result in an unstable deterrence relationship, as does an asymmetry of credibility at all three stages of the game. In other ways, however, three-stage games are more like their two-stage than their one-stage counterparts. Three-stage games exhibit the same stability-instability paradox revealed in two-stage structures and continue to support the argument that escalation dominance at any stage of the game is usually destabilizing. And, as in two-stage escalation games, the end game is the most diagnostic of the stability characteristics of a status quo outcome. The implications of the model for NATO strategy and arms control under a variety of conditions are discussed. Assumptions about end game credibility are shown to be critical in determining the wisdom of eliminating either one or two conflict stages.