When Does War Transpire? Unlikely War Onset Under Uncertainty

Scholars have attempted to clarify whether uncertainty about the outcome of a potential war—victory or defeat—influences state behavior in international crises. Until recently, there was consensus that when two states in a bilateral context have equivalent capabilities, and therefore experience high outcome uncertainty, they are more inclined to undergo conflict. However, this bilateral scheme provides a limited and ultimately misleading understanding of how outcome uncertainty affects the onset of war because bilateral interactions in an international setting do not take place independently of possible third-party involvement. Thus, in a new study, political scientists Muhammet Bas and Robert Schub maintain that estimates of uncertainty about a war’s outcome should incorporate multilateral factors, such as multiple actors’ relative capabilities.

In a contradiction to the conclusions of studies of bilateral interactions, the authors of this study postulate that higher outcome uncertainty decreases the likelihood of conflict. This pacifying effect stems from the fact that potential belligerent states allow for the involvement of a myriad of third-party states, directly or indirectly, at the onset of war. The authors expect that a hierarchical, uneven power distribution clarifies ex ante estimation of the war outcome, and an assured victory increases the likelihood of conflict. From the powerful state’s perspective, no intervention from minor actors will overturn the expected triumph. On the other hand, an even power distribution increases outcome uncertainty because each third-party state’s involvement and support may significantly transform the outcome of war. In the face of such uncertainty, the potential belligerent states exhibit prudence thereby decreasing the likelihood of conflict.

Building upon these theoretical foundations, the authors address how the polarity, the capability distribution, the complex nature of the estimation of third-party response, and the high war costs relate to the degree of outcome uncertainty in multilateral settings. The first two components collectively indicate that an increase in the number of powerful states, which balances the distribution of capabilities, increases outcome uncertainty. This conclusion is straightforward: under broad power parity, any powerful third-party states can join a war and significantly sway its course. This power balance may promote peace because victory is uncertain. The last two components of the theory associate outcome uncertainty with state behavior stipulating that potential conflict initiators respond to higher outcome uncertainty with prudence. This is due to the complexity of estimating third-party reactions to the outbreak of conflict as well as the anticipated high costs of conflict in an uncertain environment.

The authors hypothesize that multipolar power distributions that lean toward perfect power parity among states represent the highest degree of outcome uncertainty, and that higher outcome uncertainty reduces the likelihood of conflict. To test this hypothesis, the authors develop a novel measure that captures outcome uncertainty within a multi-state system. Referred to as ‘system uncertainty,’ it includes all the relevant states for a potential conflict between two belligerent states. Additionally, to capture the fact that more powerful states contribute more to overall uncertainty, the authors give those states a larger weight. By extension, the authors define ‘regional uncertainty’ which uses different states than those used for system uncertainty. The authors divide the world into five geographical clusters: Europe, Asia and Oceania, the Middle East and North Africa, sub-Saharan Africa, and the Americas. All states within the geographical cluster of the conflict-initiating states are included as relevant actors for that conflict. The authors also consider all major powers and any state contiguous to either of the two conflict-initiating states to be a relevant actor.

The authors then empirically test their hypothesis—the effect of outcome uncertainty on conflict—by using both the system and regional uncertainty concepts with two specifications. In the first specification, they use conflicts in their initial year, sourced from the Militarized Interstate Disputes (MID) data set, to examine uncertainty’s impact on the likelihood of the onset of conflict. In the second specification, the authors use the number of conflict onsets per year to examine uncertainty’s impact on the total number of annual conflicts. In order to adjust for factors beyond polarity-based outcome uncertainty, the authors incorporate into their analysis factors including states’ proximity, regime type, rivalry status, preexisting alliances between them, and the presence of a shared border.

The results indicate that greater outcome uncertainty has a statistically significant effect in reducing the likelihood of conflict and the total number of annual conflicts at the systemic level. The extent of this effect varies according to weighted forms and models. Representatively, when increasing system uncertainty from the 10th to the 90th percentile, with all other factors constant, not only does the probability of conflict decrease from 15.6 percent to 9.7 percent, but the number of expected annual conflicts also decreases from around 11 to less than 1. These findings suggest a clear pacifying effect of system-level outcome uncertainty on interstate conflict dynamics. Simply stated, a highly uncertain international setting induces cautious state behavior.

The observed relationship between outcome uncertainty and conflict is also present at the regional level of analysis. Regional outcome uncertainty is negatively correlated with the likelihood of regional conflict and the total number of annual regional conflicts. While the relationship between regional outcome uncertainty and the likelihood of regional conflict in some sub-models is not strong enough to show statistical significance due to suspected misspecification issues, there appears to be an overall negative association between them. Regional uncertainty does have a substantive statistical effect on annual regional conflict count; when increasing regional uncertainty from the 10th to the 90th percentile, all other things being constant, the expected number of regional conflicts per year reduces to near a quarter of the original number.

In summary, the results demonstrate that states will avoid initiating a war when they are uncertain about the chances of their victory, owing to the possibility of other relevant states entering the war and swinging its trajectory and outcome. The authors of this study have highlighted the importance of integrating systemic and regional attributes into the previous bilateral design used in conflict studies. States in the real world are likely to be cognizant of the multilateral nature of interstate conflict. Thus, incorporating third-party actors into the equation of outcome uncertainty is a logical and worthwhile strategy. Through focusing on multilateral settings and uncertainty about the outcome of a potential war, the authors shed light on ambiguity-averse state behavior and establish that states will opt out of a war when the outcomes are uncertain.

Article source: Bas, Muhammet, and Robert Schub. “How Uncertainty about War Outcomes Affects War Onset.” Journal of Conflict Resolution 60, no. 6. (2016): 1099-1128.

Featured photo: cc/(bee32, photo ID: 542078928, from iStock by Getty Images)