THOU SHALT NOT KILL A FELLOW LIBERAL: PATTERNS OF MILITARIZED DISPUTES BETWEEN LIBERAL STATES, 1816-1992

Abstract

In this paper, I provide fresh reasoning and evidence about how liberal societies avoid war against each other. While such regimes often use force against each other in militarized inter-state disputes, they have shed blood in such encounters only five times in almost two full centuries of record keeping. When blood has been shed, the states involved are just recently liberal, and almost always have a long and still active history of enduring rivalry against one another. Kant, writing in 1795, was familiar with liberal republics (but not with modern democracy) and more recent authors such as Doyle (1986) and Rummel (1983) have followed Kant's lead by focusing on, and documenting, a virtual absence of inter-state war between liberal or free states. To throw new light on this question, I examine the militarized disputes (MIDs) that have occurred between liberal societies, to see how close their disputes have come to war. In this frame of reference, war is defined as sustained combat resulting in substantial fatalities (at least 1,000 killed). MIDs are armed conflict more broadly conceived -- with 96% of them less violent than war itself -- involving the explicit threat to use force or the display of force as well as the use of force. There have been 128 militarized disputes between liberal societies as defined and measured by Doyle (1986; 1997), with the United States and the United Kingdom the most frequent protagonists. I scrutinize these cases using the combination of scientific method and case study found in Ray (1995), who examines cases that some may have considered wars between democracies. I expand this method to inter-liberal MIDs, first examining severe cases. Wars are discussed (e.g., Israel vs. Lebanon 1948, India vs. Pakistan 1999), but they do not qualify (e.g., Israel was not independent, Pakistan was not free). With no actual wars between liberal states, their most severe MID is a declaration of war, by Britain on Finland, Dec. 6, 1941, when both were democratic and liberal. Thanks to the case-study emphasis, I also have discovered corollary declarations of war not in the MID data set: by Australia and Canada on Finland. But one must question the true severity of these incidents, since none of these three Commonwealth nations actually did go on to engage in combat against Finland.

Moving from wars themselves to the MIDs short of war, it is unusual to have such a large number of militarized disputes with so little risk of a war. To see why war is so often avoided, I examine the context of the dispute, as well as the highest severity of action reached by each protagonist in the dispute. Severity is assessed using not only the threat, display, or use of force distinction, but also the more specific underlying action codes, such as blockade, seizure of territory, and mobilization. The context and severity (Maoz 1982) of the actions in the disputes allows an assessment of how close liberal societies have come to waging war against each other. Some incidents involve seizure of fishing boats, or protests about fighter plane incursions into airspace, and hence do not seem to bear the seeds of war. More serious incidents do occur, some in the context of World War I or II, as, e.g., one liberal state seizes another's territory just ahead of an invasion by the Axis powers. In some ways inter-liberal MIDs look as serious as other MIDs. Liberal MIDs often involve the use of force. Liberal MIDs are multilateral (24% of the time) more often than other MIDs (15%). Liberal MIDs are less often reciprocated (44% of the time), but this is not significantly different than the rate for other MIDs (50%). Liberal MIDs tend to be tit-for-tat incidents (i.e., rarely escalate vertically). Most distinctively, liberal MIDs rarely involve bloodshed and, when leading to use of force, tend toward seizures rather than clashes. The results of my hypothesis tests are consistent with the explanation that liberal societies avoid war with each other for two related reasons: (1) consistent with costly signaling theories, liberal targets tend not to respond militarily to liberal states who initiate MIDs against them, but this difference is not strong enough to achieve statistical significance; (2) a vivid norm of avoiding bloodshed characterizes most of their interactions, and provides a powerful pacifying effect. The main peril of inter-liberal MIDs comes from the weakness of another important norm, against ganging up.

This study is of value for a number of reasons. First, and straight-forwardly, the findings help us understand what type of situations breed an armed clash between two liberal societies, and consequently how close these clashes have come to warfare. Second, my findings provide new evidence that has relevance to the democratic peace debate. Several authors have emphasized the existence of a "democratic" peace (Ray 1995; Senese 1997). Studies such as mine, on MIDs in the "liberal" peace, can shed light on the democratic peace debate. While the absence of war between two "democracies" has been hailed as the closest to a law-like finding in world politics, and has stimulated enormous research efforts over the last two decades, much of this research has been based on the Polity data on democracy, which in turn seems to be heavily dependent on a measure of constraints on executive authority (Gurr et al. 1989; Gleditsch and Ward 1997). As an indication of the close connections between the liberal and the democratic peace literature, Russett and Oneal (1997), in perhaps the classic empirical study of the peace in question, title their study, "The Classic Liberals Were Right: Democracy, Interdependence, and Conflict (emphasis added)." Much of the research on the inter-democratic peace is inductive and empirical; to gain a large number of cases for multivariate statistics, the analysts typically use the Correlates of War Project's militarized dispute data rather than its war data (Oneal and Russett 1997; Henderson 2002). Using the outbreak of a MID as the dependent variable in these multivariate study of the democratic peace, however, implicitly assumes that inter-liberal MIDs are just as severe as other MIDs. As I show, that assumption is not warranted. Examining the types of militarized clashes that have occurred between liberal societies allows the reader to ponder the meaning and face validity of the many extant empirical tests of the inter-democratic peace. Although my case studies lend a bit of new support to critics of the democratic peace by pointing out declarations of war that even several authorities on the subject are not aware of, the main weight of my findings is strongly supportive of the democratic peace literature. This is because my statistical patterns, especially on battle deaths, reinforce the democratic peace proposition. Because inter-liberal MIDs (which are mostly also inter-democratic MIDs) are much less likely to lead to mass killing than other MIDs, the analysts like Russett who said "the classic liberals are right" got the correct message out, but somewhat underestimated the full degree to which the classic liberals were right.

With most cases of inter-liberal MIDs surprisingly low in severity and bloodshed, but with some controversial cases (Lebanon-Israel, India-Pakistan, Britain-Finland, Australia-Finland) of war or declaration of war, the debate over the democratic peace is likely to continue.