Alliance Dynamics:


lb xg


t]ymptoms of an Unstable World


International Relations


11/14/1998


Common sense tells us that humans survive better collecti vely than individually, workers have more success winning contract disputes when in labor unions, and alliances among states increase the likelihood of peace in an anarchic international arena. However, in the case of the latter, conventional wisdom has been challenged; indeed some have asserted the polar opposite, that war is caused by alliances. Stuart A. Bremer, in his quantitative analysis of indicators of war, first showed that alliance increases the probability of war, and then, that the absence of alliance increases the likelihood of war. Conversely, Glenn Snyder offers a theoretical model of alliance dynamics, largely based on the prisonerís dilemma which makes for a dismal outlook and the prediction that alliances will breed conflict and insecurity among states.


The key difference between the two articles is that Bremer studies dyads; his approach only examines the occurrence of war between allied states. Snyder tackles alliances and the potential for war with all other states, and he does not really touch on the potential for war between allies. Although Bremer ultimately concludes that the absence of alliance between two states predicts the potential for war, and Snyder asserts that the existence of alliances create war, Bremerís and Snyderís articles are not necessarily in opposition. Both recognize that alliances vary in commitment and sincerity. It is the degree of commitment one ally expresses to another that ultimately determines whether it will war with its ally or any other state.


Bremerís article divides alliances into three categories (derived from the Correlates of War data set): defense pacts, neutrality agreements and ententes1. He found that, independent of the presence of any other variables associated with war, neutrality agreements and defense pacts have the highest correlation with the occurrence of war within dyads (probability 3.09 X 103 and 1.79 X 103, respectively), and ententes the lowest (probability 0.28 X 103); the absence of any alliance had a slightly higher probability of producing war than an entente (probability 0.33 X 103).2 His finding that a lack of alliance, independent of any other variable, is not highly associated with the potential for war, is not very surprising, since the number of states in existence is so large and a lack of alliance does not necessarily indicate a presence of animosity; some states may have no interest in each other either way.


The discrepancy in probabilities between the two former forms of alliances and the latter is not surprising either, since an entente represents a strong commitment of alliance between two states. The results indicate that a strong commitment between two states makes it highly unlikely that they will ever war with each other, but says nothing of their likelihood of going to war with states outside the alliance. Hence, this finding does not contradict Snyderís model. Neu trality agreements and defense pacts however represent weaker alliances, or alliances of convenience. Bremer might say that they represent alliances in the traditional view, ď. . . [which] sees them growing out of expediency and reflecting nothing deeper than a temporary need of two or more states to coordinate their actions against one or more other states. . . . alliances are not seen as contracts but rather as bargains, wherein it is understood by all parties that each has the right to withdraw quickly should a better deal come along.Ē3 These weaker alliances might work even better under Snyderís model, since they represent strategic alliances, negotiated under specific circumstances for security purposes. They do not represent natural (predisposed)4, alliances, or those negotiated on ideological or ethnic grounds.


Snyder presents a dilemma model, a no-win situation. According to him, if a state expresses a great deal of commitment to its ally, the ally will feel more secure in pursuing its conflicts with its adversary. The ally perhaps feels bigger than it actually is; it is confident initiating battles or pursuing its interest with its adversary knowing it has the backing of its partner, the original state. This leaves the original state in a situation of entrapment.5 By expressing a high level of commitment to its ally, it has a heightened the potential to become ensnared in its