<![if !supportEmptyParas]>�<![endif]>

<![if !supportEmptyParas]>�<![endif]>

<![if !supportEmptyParas]>�<![endif]>

INCIDENCE OF MILITARIZED DISPUTES BETWEEN LIBERAL STATES,

<![if !supportEmptyParas]>�<![endif]>

1816-1992

<![if !supportEmptyParas]>�<![endif]>

<![if !supportEmptyParas]>�<![endif]>

Frank W. Wayman

<![if !supportEmptyParas]>�<![endif]>

Univ. of Michigan-Dearborn

<![if !supportEmptyParas]>�<![endif]>

A paper presented at the annual meeting of the International Studies Association, New Orleans, La., Mar. 23-27, 2002.

<![if !supportEmptyParas]>�<![endif]>

Manuscript of April 6, 2002, 08:35, c:\war\ISA02V4.doc

<![if !supportEmptyParas]>�<![endif]>

ABSTRACT:� Although liberal states have not fought a war against each other, they have occasionally been on opposing sides in the militarized interstate disputes that can escalate to war.� Among some 2,000 militarized interstate disputes (MIDs) since 1816, I have identified sixty-two that have been between liberal states.� What is the implication of this for the democratic peace literature?� This paper examines the severity of these sixty-two clashes to assess what degree of hostility they represent.� While often involving the use of force, liberal MIDs are almost always confined to two parties (i.e., do not escalate horizontally), and are usually unreciprocated or tit-for-tat incidents (i.e., rarely escalate vertically).� While inter-state war between liberal societies may occur in the future, not many of the MIDs that have occurred between them in the past two centuries can be seen as cause for alarm.��

INCIDENCE OF MILITARIZED DISPUTES BETWEEN LIBERAL STATES,

1816-1992

<![if !supportEmptyParas]>�<![endif]>

Frank W. Wayman

Univ. of Michigan-Dearborn

<![if !supportEmptyParas]>�<![endif]>

�� The inter-democratic peace sometimes seems to be the major focus of those seeking to understand the causes of war (e.g., Oneal and Russett 2001), and to prevent war in our own time (e.g., the Clinton administration effort to foster democracy).� But most of the empirical research on the inter-democratic peace examines militarized disputes (MIDs) rather than war.� Concerning democracies clashing with other democracies in inter-state wars (i.e., sustained combat involving more than a thousand battle deaths), it is hard to say more than this:� there have been fewer than a hundred such wars since the birth of modern democracy in the nineteenth century; and there are no cases of inter-state wars between states that are clearly liberal democracies (Rummel 1994: 2).� Those who believe strongly that democracies do not fight other democracies say flatly: this means there are no wars between democracies.� Indeed, I join them on one narrow point.� I believe (modifying Jack Levy's aphorism) that this is as close to a counter-intuitive law-like generalization that we have in the social sciences.� It is something surprising for which there are no exceptions.�

<![if !supportEmptyParas]>�<![endif]>

�� Critics respond that maybe clashes such as Athens fighting Syracuse or the American Civil War or Syria fighting Israel in 1948 are counter-examples.� Ultimately, most of these seem a little forced, since typically one of these so-called democracies is drawn from a rogue's assemblage of slave-owning societies and states prone to chronic military rule.�

<![if !supportEmptyParas]>�<![endif]>

�� The better argument by critics of the democratic peace is that there have been very few inter-state wars and very few democracies, so the lack of war between democracies is not so surprising (Small and Singer 1976).� And it is especially unsurprising given that many of the democracies that have existed have been allies in NATO, driven into each other's arms by their sense of threat from the Soviet Union during the Cold War era.� Other factors also may have driven them together, such as shared trade and prosperity, membership in common in international organizations, and the global drift toward less inter-state war nowadays.� I agree with these critics, insofar as I would say that we need more democracies with the potential to fight each other, and who despite this did not fight, before we rush to the conclusion that this law has gained our confidence.� I am confident in Boyle's law (the volume of a gas is proportional to its temperature in degrees Kelvin) to an extent I will never be about the democratic peace law.� I have gained more confidence in the democratic peace as two peaceful decades have passed since Rummel first announced its discovery.� I have also gained confidence in it because of Russett and Maoz's observation that India and Pakistan, which have fought three or four (Oneal and Russett 2001: 48) wars with each other since 1946, have not fought a war against each other when both were democracies.� But, if we rely solely on whether there has been an inter-democratic war, it is going to take many more decades of peace to build our confidence in the stability of the democratic peace.

<![if !supportEmptyParas]>�<![endif]>

�� The objection that there haven't been enough modern wars to feel confident that democracies don't fight each other, especially when combined with the further objection that there might be a need to control for third variables (such as level of economic development, trade, joint alliance membership), has led to an explosion of research on MIDs between democracies.� Crucial to such research has been the existence of the Correlates of War Project militarized inter-state dispute data set, containing about 2,000 militarized inter-state disputes (MIDs) since 1816.� MIDs are cases involving the use of force, display of force, or explicit threat to use force.� This large number of cases of armed conflict between states allows one to get around the problem that there is a small number of cases (N) of modern wars.� With MIDs data, one can test for whether there is a statistically significant absence of armed conflict between democracies, while controlling for a host of� other variables.� This has generated a huge and valuable literature beyond one's power to comprehensively review in a short paper, but useful literature citations are in Maoz (1997), Oneal and Russett (2001), and Henderson (2002).�

<![if !supportEmptyParas]>�<![endif]>

CLASSIFYING STATES AND ARMED CONFLICTS

<![if !supportEmptyParas]>�<![endif]>

�� The contribution I seek to make in this paper is to examine a portion of the MIDs data on which this vast literature rests.� While the scholars studying the democratic peace spend a great deal of attention on measurement (e.g., what is a democracy, conceptually and operationally), much of the measurement work has focused on how one should measure democracy (as well as regime similarity, and other control variables), whereas I want to switch the attention to the kinds of armed conflicts the states get into.�

<![if !supportEmptyParas]>�<![endif]>

CATEGORIES OF ARMED CONFLICT

<![if !supportEmptyParas]>�<![endif]>

�� I switch attention to the somewhat less-examined militarized dispute side, to see what mililitarized disputes these liberal societies have had with each other.� This is important for two reasons.� One, the empirical literature just discussed tends to treat MIDs in a dichotomous fashion (there is a MID or there isn't), and debate has been mostly limited to whether one should code an ongoing MID as well as the onset of a new MID as an instance of armed conflict.� Second, regardless of this empirical literature on whether joint democracy affects MID incidence while controlling for other variables, there remains an important substantive question:� have the MIDs between democracies usually gotten close to the brink of war, or are they minor technical clashes with little risk to international peace and security?

<![if !supportEmptyParas]>�<![endif]>

�� In other words, my purpose in doing this is to examine some of the implications of switching from the original focus (do liberal societies go to war against each other) to the statistical focus (do liberal societies get into militarized disputes with each other).�� My purpose is not to quarrel with the fine scholarship that has been done, but rather to complement that work and to help everyone in that debate get a richer sense of the evidence base on which their arguments rest.� On the one hand, a MID can be a severe event like a declaration of war.� On the other hand, a militarized dispute between liberal societies may just be argument between tuna fishermen, in which the coast guard seizes an illegal fishing boat by force.� If I drive from Dearborn, Michigan, across the border to Windsor, Ontario, and get stopped for speeding and escorted to the courthouse by a police car to make sure I pay the fine, this is not coded as an international incident, but if a fishing boat has an analagous encounter and is taken into port under armed guard, this is a MID; do we from this have evidence that democracies are on the brink of war?� I will argue that we should be very hesitant to use all the MIDs data to reach inferences about the inter-democratic peace.� There is evidence in the MIDs data for both sides.� Some inter-democratic MIDs are not going to alarm anyone except the NATO pilot who finds the Swiss angry because he's not in Germany anymore.� But on the other hand the MIDs data do contain surprising cases of intense armed belligerence between two liberal states.� One inter-democratic MID is so severe it surprises even some of the researchers who have published statistical studies of the democratic peace.�

<![if !supportEmptyParas]>�<![endif]>

�� There have been some studies similar to what I am doing, and the most recent and most similar work is that of Senese (1997).� Senese conducts a multivariate test to see if, once a MID has started and before it expands to war, joint democracy is a pacifying condition that reduces its chances of escalating.� This is a commendable study, but each individual study has its limits.� One limit of Senese's work is that the MID data set variables do not allow one to measure escalation in the usual sense of the word, namely, the process by which a less severe act by one side is responded to by a more severe act by the other side.� This is not possible to examine because the MID data only record the highest action taken by each side, and do not tell us the other actions or the sequence in which the highest actions occurred.� A second limitation of Senese's study is that it is only a multivariate statistical study of a large number of cases, and so I my purpose is to complement it by looking more carefully at a subset of important cases, to see exactly how severe each one was.

<![if !supportEmptyParas]>�<![endif]>

CATEGORIZING SOVEREIGN STATES

<![if !supportEmptyParas]>�<![endif]>

�� One needs to decide what sorts of states are in the pacific union.� Should one single out countries with "elective governments" (Babst 1964) -- i.e., elections in which the bulk of the population selects their rulers?� Or should one look at countries that are free (as emphasized by Kant, Rummel, and Doyle 1986), and if so should one focus on a free economy, a judicial and legislative branch free of executive branch dominance (as emphasized by the Polity data sets), or whether the people are have civil and personal freedom (as emphasized by Freedom House)?

<![if !supportEmptyParas]>�<![endif]>

�� In this paper, I conceive of these different definitions of free states, from Doyle to Polity to Freedom House, as all being slightly varying measures of the basic underlying dimension of being a liberal state.� They are all highly correlated with each other, and are usually in agreement in classifying particular governments.� I use Doyle's because it goes back the farthest (I need to get back to 1816), and, when Doyle stops (in 1982) I update his ratings with Freedom House rankings, which go more up to the present than any other pertinent measure.� In the decade on which they overlap (1972 to 1982), Freedom House and Doyle's (1986) ratings are highly consistent with each other.�

�� An important case illustrating state ratings is the Indo-Pakistani war (or near-war) of 1999.� Over three months in the second quarter of 1999, India and Pakistan fought daily, with a total killed in combat of perhaps 1200.� This seems just over the COW threshold of 1,000 battle deaths, the operational indicator of war.� At that time, both India and Pakistan had elected governments (the Musharraf military government not having yet seized power in Pakistan).� However, Doyle and Freedom House do not count Pakistan as free.� Doyle counts India as liberal in every year except the emergency year toward the end of Indira Gandhi's rule.� But to Doyle, Pakistan is never a liberal state, and so India-Pakistan is never a liberal dyad.� Freedom House has a similar view.� They rank states as "free" if they score below 3 on a scale, ranging from 1 for fully free to 7 for totally not free.� Freedom House assesses this freedom scale for both political rights and civil liberties, and a state with a score of "2, 2" would be rated as free in both political rights and civil liberties.� So India, with a "2, 3" Freedom House rating in 1998-1999 and in 1999-2000, is classified as free.� Pakistan, with a "4, 5" in 1998-1999, is rated "partially free."� Later, after the coup, Pakistan in 1999-2000 had a score of "7, 5" in 1999-2000, which is "not free."� (Freedom House 2002)� In short, Doyle and Freedom House agree that the apparent war between India and Pakistan was not a war between free or liberal states, even though the Pakistani as well as the Indian government had been popularly elected.�� And so by using the Doyle (1986) ratings of a country, updated with Freedom House ratings of whether the country remained "free," I examine each MID up to 1992.

<![if !supportEmptyParas]>�<![endif]>

�� This method allowed me to identify the sixty-two MIDs in which a liberal state was on each side (table 1).� This list, culled from the Correlates of War Project's (COW) version 2, MIDs B list of some 2,000 MIDs in the entire globe since 1816, is presented in table 1 in chronological order.� The table displays the MID's code number, the year the MID started, the initiator, the target, the highest level reached in the belligerent acts by the initiator, and the highest level of such acts by the target (Gochman and Maoz 1984).� These levels of action are on a twenty-two point scale of action developed by the COW project:

<![if !supportEmptyParas]>�<![endif]>

�� It is obviously of central importance to this investigation to get a sense of what might be called the severity or importance of these militarized disputes, so we can see how close liberal societies have come to the brink of war.� After all, no war has occurred without a prior MID, and if wars occur, it has been because MIDs escalate.� What is the ecalatory ladder, and is its code embedded in the twenty-two categories of hostility level?� A first answer is that the index itself is organized in a prima facie way from high severity (war, level 22) to low severity (no militarized action, level one).� This is a useful first step, and I will occasionally rely on that ordering in this paper.� A second answer is provided by the hostility level codes in the two right-hand columns of table 1.� These are the codes used by most of the research in the field.� Hostility level five (corresponding to action code 22) is inter-state war itself.� Hostility level four (corresponding to action codes 15 to 21) is use of force short of war.� Hostility level three (corresponding to action codes 7 to 14) is display of force.� Hostility level two (corresponding to action codes 2 to 6) is explicit threat to use of force.� Hostility level one (corresponding to action codes 1 and -9) is no detectable or detected militarized action.

<![if !supportEmptyParas]>�<![endif]>

�� A deeper answer is provided by Maoz (1984), who created the first scale of MID severity.� Maoz has the first scientifically validated measure of severity of MID actions, and also the first measure that can give exact numerical scores to the severity of each action.� Maoz ranks the MID actions on a scale of zero to 100, with 100 representing war.� Unfortunately, Maoz's original scale (which he is now updating) does not scale all 22 levels of hostility.� His original scaling (Maoz 1984: 224) is as follows:

<![if !supportEmptyParas]>�<![endif]>

100.� War (Most Severe)

99.� Declaration of War

96.� Blockade

84.� Occupation of Territory

82.� Mobilization

81.� Seizure of Material or Personnel

75.� Clash

65.� Other Use of Military Force

49.� Show of Force

30.� Threat to Use Force

29.� Threat to Declare War

20.� Alert

8.� Threat to Occupy Territory

1.� Threat to Blockade (Least Severe)

<![if !supportEmptyParas]>�<![endif]>

While Maoz's list follows approximately the same ranking as the official COW codes, it differs in rank-ordering in two important particulars:� alert, a display of force, is according to Maoz much less severe.� He rates alert as about as severe as an average display of force.� On the other hand, mobilization, another display of force, is rated by Maoz as as severe as an average use of force.� So the Maoz scaling technique suggests that the COW action and hostility codes are imperfectly correlated with a severity scale.� Since each of these indicators has something to be said for it, I use both the Maoz severity codes and the official COW codes in this paper.

<![if !supportEmptyParas]>�<![endif]>

�� A updated Maoz severity scale would have more value as a research tool.� As one of the preliminary steps to updating Maoz's scale, I have interpolated, based on my own judgment, starred scores that I hypothesize would complete Maoz's scale.� In each case I have been sure to give a unique score that will avoid ties:

<![if !supportEmptyParas]>�<![endif]>

100.� War (Most Severe)

99.5� Use of Chemical, Bacteriological, and Radiological (CBR)�� Weapons*

99.� Declaration of War

96.� Blockade

84.� Occupation of Territory

82.� Mobilization

81.� Seizure of Material or Personnel

75.� Clash

65.� Other Use of Military Force

60.� Border Violation*

50.� Show of Troops*

49.� Show of Ships*

48.� Show of Planes*

42.� Nuclear Alert*

40.� Threat to Use Nuclear Weapons*

30.� Threat to Use Force

29.� Threat to Declare War

20.� Alert

8.� Threat to Occupy Territory

1.� Threat to Blockade (Least Severe)

<![if !supportEmptyParas]>�<![endif]>

�� I have taken the Doyle list of liberal states (Doyle 1986) and coded which MIDs have occurred between these states.� Although so far I have identified 62 MIDs between liberal states,� the work is preliminary, and there are probably up to 90 such inter-liberal MIDs.� The sixty-two MIDs that all my coders agree on can be ranked by the highest hostility level each reached.� When classified by their most severe incident (Table 3) fifty of the MIDs have a severity level between 50 and 100, whereas only twelve have severity between 1 and 49.� Thus, the modified Maoz severity scale indicates that these inter-liberal MIDs are mostly characterized by a peak incident that is indeed severe.

<![if !supportEmptyParas]>�<![endif]>

�� By the more traditional classification, of threat, display, or use of force (Table 4), we also have signs that these inter-liberal MIDs often peak in a serious event.� There are forty uses of force by the initiating side.� Of these, most often the initiator's highest act is a seizure (17 cases) or "other use of force" (16 cases).� There are also four clashes, two occupations of territory, and one declaration of war.�

<![if !supportEmptyParas]>�<![endif]>

�� In addition to these cases of use of force by the initiator, sometimes the target is the only side to use force.� In my 62 cases, this unilateral use of force by a target occurs three times.� In all, this makes 43 cases in which either the initiator or the target used force.� This means that 69% (43 out of 62) of the inter-liberal MIDs went to use of force.� This is exactly the average for the MID data set as a whole (Jones, Bremer, and Singer 1996: 197).

<![if !supportEmptyParas]>�<![endif]>

THE CASE OF THE U.K. DECLARATION OF WAR ON FINLAND:� DEC. 6, 1941

<![if !supportEmptyParas]>�<![endif]>

�� By far the most severe of these events was the declaration of war by Britain on Finland during World War II.� This occurred on Dec. 6, 1941.� A declaration of war, rating 99 on Maoz's severity scale, is obviously serious.� Finland had regularly scheduled elections and a parliamentary government throughout this period, and it goes without saying that the U.K. was a liberal democracy under Winston Churchill.� On the other hand, there are several mitigating circumstances, starting with the oddity that this declaration of war did not lead to an actual war.� Finland had been invaded by the Soviet Union a couple of years before, and at the conclusion of that war Finland was forced to cede 10% of its territory.� When Hitler invaded the Soviet Union, Finland did, too, not as a Nazi ally but as a co-belligerent.� "The Western Allies had ambivalent feelings, torn between their residual good will for Finland and the need to support their vital ally, the Soviet Union.� As a result, Britain declared war against Finland, but the United States did not; there were no hostilities between these countries and Finland." (www.wikipedia.com/wiki/Continuation_War)� Churchill apparently had tried to convince Stalin it would be counter-productive to declare war on Finland, and, failing to be convincing, declared war (Churchill, 1950: 528).� Under such circumstances, it is perhaps understandable that Britain did not attack Finland, and that Finland did not attack Britain.

�� However, in researching this case, I found evidence that there has been at least one other declaration of war by a liberal regime, and I suspect there may have been a third.� Australia, following Britain's lead, apparently also declared war on Finland (Australian Constitutional Commission 1988).� The Dominion of Canada may have done so also.� Neither an Australian nor a Canadian declaration of war on Finland appear in the MIDs data set, so the data set may understate the number of severe MIDs between liberal states.

<![if !supportEmptyParas]>�<![endif]>

CONTAINMENT VERSUS ESCALATION OF INCIDENTS:� THE PACIFIC SIGNATAURE OF INTER-LIBERAL MIDS.��

<![if !supportEmptyParas]>�<![endif]>

�� Mansbach and Vasquez (1981) argue that hostile events occurring between friends tend to be interpreted as not as serious as the same events occurring between enemies, and that consequently the friend attacked will tend to respond in a less hostile manner.� This pattern is suggested by these declarations of war on Finland, to which the Finns do not violently respond.� The pattern is characteristic of many of the inter-liberal MIDs (see Table 5 and Table 6).� In table 5, the principal diagonal shows the tit-for-tat pattern of the target responding with the same action level as the initiator (14% of the cases); the upper right portion of table 5 (above the principal diagonal) shows the cases in which the target response was a less severe action, indicating de-escalation (76% of the cases).�� There are only six cases in which the target's response is a higher action than the initiator took (representing 10% of the cases).�

<![if !supportEmptyParas]>�<![endif]>

�� Simplifying this pattern into use, display, or threat of force, we see (table 6), that liberal states usually do not respond to liberal initiators, even up to declarations of war.� In a majority of cases, there is no response from the target, even when there has been use of force against it.� In 65% of the cases of initiator use of force, there is no response from the target; in 59% of the threats or displays of force, there is no response from the target.� When the target does respond in an inter-liberal MID, the response is usually proportional to the provocation--what might be called tit-for-tat.� In table 7 we see that the these patterns are significantly more pacific than the patterns for the control group of MIDS that are not inter-liberal.

<![if !supportEmptyParas]>�<![endif]>

�� Sometimes, this is because of the context of the inter-liberal MID.� There are cases of Swiss protest of NATO overflights of Swiss airspace.� While the Swiss have to protect their neutrality and territorial integrity, they do not have to worry about deterring a NATO invasion.� Sometimes the MID action may even be seen as having protective intent.� There is a case of Britain moving troops into Scandinavia--but the intent is clearly to preclude a non-liberal state's hostile occupation of the territory.� This is to say, some of the inter-liberal MIDs occur in, and are being interpreted in, the context of a wider war or wars.� Liberal states are trying to cope with NAZI and Stalinist aggression, and end up seizing each other's territory or even declaring war on each other.� In these cases, considering the more dangerous context, no wonder the target's response to the use of force by a fellow-liberal state is a muted response.� Hence, even in the more provocative cases of use of force, we have no response a greater percentage of the time in inter-liberal MIDs than in the rest of the MID data set.� Also, a substantial portion of MIDs involve more than two states, but, astonishingly, of the sixty-two inter-liberal MIDs thus far identified, only one goes beyond the bilateral state, and that MID involves only three countries (see Table 1).� The chi-squared test indicates that the inter-liberal MIDs are significantly more bilateral than the control group of all other MIDs (table 8).�� In sum, it seems important that inter-liberal MIDs are usually self-limiting in the sense that they don't escalate horizontally or vertically.

<![if !supportEmptyParas]>�<![endif]>

�� Moreover, in support of this line of argument, inter-liberal MIDs rarely involve bloodshed.� Only three have any recorded battle deaths (table 9).� Oddly, all three of these cases involve, of all places, Ecuador (twice against Peru,once against the U.S.).� None involve more than 100 battle deaths.� Battle deaths are significantly less common in inter-liberal MIDs than in the control group (i.e., all other MIDs).

CONCLUSIONS

<![if !supportEmptyParas]>�<![endif]>

�� The scientific study of the democratic peace has drawn attention because of its potential ramifications for international relations theory (are realists their liberal-idealist opponents more correct about the state of the intenational system? [Wayman and Diehl 1994]), and for foreign policy.� Oneal and Russett (2001) represent the predominant view of the empirical-statistical studies, that when one controls for all pertinent alternative causal explanations of armed conflict, one finds that democracies don't engage in armed conflict with each other.� Oneal and Russett's findings give a flavor for what the dominant school this literature can produce:� they conclude that Immanuel Kant's original conception was correct, namely, that free societies (what I have been calling democracies) will not fight against each other, especially if they trade with each other and are embedded in a free confederation of states.� Henderson (2002) is a good example of the contrarian view, namely, the view that, when one controls for the correct additional variables, one finds that Russett is wrong.� Henderson carefully uses Oneal and Russett's measures and procedures, and reports that, when one uses the correct set of control variables, joint democracy (two states both being democracies) does not contribute to the reduction of armed conflict.� My conclusion from this debate is (1) both Henderson and Oneal/Russett have done superb statistical analyses; (2) a reader's conclusion about whether joint democracy leads to a reduction in MIDs involvement depends on whether the reader finds Henderson's or Oneal/Russett's work more convincing; (3) if Oneal and Russett are right, then a blow has been dealt to political realism (Mearsheimer 2001) and support has been provided for the Clinton administration efforts to spread democracy; but if Henderson is right the realists are defended and Clinton's efforts were brought into question; (4) neither Russett nor Henderson is addressing the original finding of Rummel (and predecessors) that there have been no wars between democracies.� Even if Henderson's analyses are correct, it remains true that there have been no inter-state wars between a clear-cut liberal democracy and another such state.� All this empirical literature hangs on the examination of MIDs rather than inter-state wars, so it is important in assessing this literature to see what the occurrence of a MID between liberal democracies has meant, in terms of severity of armed conflict.�

<![if !supportEmptyParas]>�<![endif]>

�� The examination of the democratic peace question in this paper in fact lends support to both sides.� On the one hand, a liberal democracy has (once in the MID data set, and perhaps in two other instances not recorded in the data set) declared war on another liberal democracy.� More generally, inter-liberal MIDs involve the use of force more often, and have greater severity, than one would expect.� On the other hand, these inter-democratic MIDs often are limited in horizontal and vertical spread, and have been little noted or long remembered except by diligent Correlates of War Project coders who (admirably) live and die by sticking to their coding rules.� This paper, however, provides some evidence that just examining what the inter-liberal MIDs were like can be a useful complement to the important statistical tests of Oneal/Russett, et al.� Perhaps this paper will inspire more attention to this subject.� As the case study of declaration of war indicates, we may need more sensitivity to the limitations of the MIDs data set; this data set, like any human creation, may have some errors, at least errors of omission.� We need further work on event severity, better understanding of what sort of incidents and MIDs are likely to escalate to war.� We certainly could use an incident-level data set, like Leng's BCOW data, so that we do not have to infer such things as tit-for-tat behavior from the bare data of the highest coded hostility reached by the target and by the initiator.� We need a better sense of what the distinction is between war and the lesser MID incidents (i.e., threat, display, and use of force short of war).� The MID data set has sustained a flourishing industry on the empirical study of democracy and war.� While this industry has provided us with an improved understanding of the inter-democratic peace, those seeking to understand this literature should be made aware of the issues of validity raised in this paper.��

<![if !supportEmptyParas]>�<![endif]>

�� The many large N statistical studies of the democratic peace usually use the occurrence of a MID as the dependent variable.� This implicitly assumes that inter-democratic MIDs are on average at the same severity level as the other MIDs.� I have shown that in three ways (bilateralism, low rates of response, low levels of fatalities), inter-liberal MIDs are less severe than other MIDs.� The inter-liberal MIDs are less likely to involve third parties, less likely to be reciprocated, and less likely to involve the loss of life (compared with other MIDs).� Hence, the inter-liberal peace is more powerful than one would have concluded from the extant large N, statistical literature.� The inter-liberal peace proposition is also, for the same reasons, more powerful than one would have concluded by simply counting the inter-liberal MIDs or calculating what proportion of them involved use of force.� The 185-year peace between free societies may end tomorrow, but it has been, on net, slightly more stable than we'd reckoned.

Table 1.� MIDS between Liberal Countries

<![if !supportEmptyParas]>�<![endif]>

MID

Year

Initiator

Target

Hi Act

Host Level

Fatalities

<![if !supportEmptyParas]>�<![endif]>

Number

Name

Number

Name

Initiator

Target

Initiator

Target

<![if !supportEmptyParas]>�<![endif]>

301

1835

220

FRN

USA

373

1846

200

USA

<![if !supportEmptyParas]>�<![endif]>

207

1858

200

USA

1512

1891

USA

155

CHL

<![if !supportEmptyParas]>�<![endif]>

2076

1891

155

CHL

160

ARG

2300

1893

220

FRN

200

1513

1897

155

CHL

160

ARG

1898

200

220

FRN

1093

1900

160

ARG

155

CHL

-9

1902

USA

200

<![if !supportEmptyParas]>�<![endif]>

2077

1905

160

ARG

155

CHL

<![if !supportEmptyParas]>�<![endif]>

2078

1909

155

CHL

160

ARG

<![if !supportEmptyParas]>�<![endif]>

1772

1911

325

ITA

100

<![if !supportEmptyParas]>�<![endif]>

3330

1911

325

ITA

220

FRN

<![if !supportEmptyParas]>�<![endif]>

316

1912

325

ITA

220

FRN

1654

1914

200

210

NTH

<![if !supportEmptyParas]>�<![endif]>

3350

1914

385

NOR

200

1732

1915

200

380

SWD

<![if !supportEmptyParas]>�<![endif]>

1726

1916

220

FRN

160

ARG

<![if !supportEmptyParas]>�<![endif]>

1660*

1917

200, 2

UK, USA

380

SWD

17, 17

<![if !supportEmptyParas]>�<![endif]>

4, 4

<![if !supportEmptyParas]>�<![endif]>

1776

1917

200

120

NTH

<![if !supportEmptyParas]>�<![endif]>

1273

1921

290

POL

255

GMY

-9

3700

1939

200

210

NTH

<![if !supportEmptyParas]>�<![endif]>

3704

1940

200

385

NOR

<![if !supportEmptyParas]>�<![endif]>

3707

1940

200

390

DEN

<![if !supportEmptyParas]>�<![endif]>

3800

1940

200

225

SWZ

<![if !supportEmptyParas]>�<![endif]>

3802

1940

200

380

SWD

<![if !supportEmptyParas]>�<![endif]>

1786

1941

200

375

FIN

<![if !supportEmptyParas]>�<![endif]>

1702

1952

130

ECU

USA

<![if !supportEmptyParas]>�<![endif]>

3209

1954

USA

225

SWZ

1705

1954

130

ECU

USA

<![if !supportEmptyParas]>�<![endif]>

1-25

3242

1955

130

ECU

USA

<![if !supportEmptyParas]>�<![endif]>

3243

1955

135

PER

USA

<![if !supportEmptyParas]>�<![endif]>

3222

1956

MEX

USA

<![if !supportEmptyParas]>�<![endif]>

2845

1957

155

CHL

USA

2875

1960

395

ICE

200

2876

1960

USA

305

AUS

<![if !supportEmptyParas]>�<![endif]>

2877

1960

305

AUS

325

ITA

<![if !supportEmptyParas]>�<![endif]>

1900

1961

390

DEN

200

<![if !supportEmptyParas]>�<![endif]>

1187

1962

145

BOL

155

CHL

624

1963

220

FRN

140

BRA

2952

1974

USA

CAN

<![if !supportEmptyParas]>�<![endif]>

2953

1975

USA

CAN

<![if !supportEmptyParas]>�<![endif]>

619

1975

395

ICE

200

2335

1976

666

ISR

USA

3177

1976

666

ISR

350

GRC

<![if !supportEmptyParas]>�<![endif]>

MID

Year

Initiator

Target

Hi Act

Host Level

Fatalities

<![if !supportEmptyParas]>�<![endif]>

Number

Name

Number

Name

Initiator

Target

Initiator

Target

<![if !supportEmptyParas]>�<![endif]>

2968

1979

CAN

USA

<![if !supportEmptyParas]>�<![endif]>

3105

1980

130

ECU

USA

<![if !supportEmptyParas]>�<![endif]>

2119

1981

135

PER

130

ECU

26-100

2970

1981

385

NOR

390

DEN

<![if !supportEmptyParas]>�<![endif]>

2323

1982

101

VEN

100

COL

2118

1984

135

PER

130

ECU

1-25

2541

1984

220

FRN

230

SPN

<![if !supportEmptyParas]>�<![endif]>

2542

1984

205

IRE

230

SPN

<![if !supportEmptyParas]>�<![endif]>

2543

1985

205

IRE

230

SPN

<![if !supportEmptyParas]>�<![endif]>

2598

1986

230

SPN

200

<![if !supportEmptyParas]>�<![endif]>

2737

1987

CAN

220

FRN

<![if !supportEmptyParas]>�<![endif]>

2812

1987

100

COL

101

VEN

2768

1988

100

COL

101

VEN

-9

3900

1989

CAN

USA

<![if !supportEmptyParas]>�<![endif]>

3972

1991

USA

CAN

<![if !supportEmptyParas]>�<![endif]>

3573

1992

820

MAL

830

SIN

<![if !supportEmptyParas]>�<![endif]>

*200 and 2 are allies in this dispute.� Both reached the same hostility level.

Minus 9 on fatalities indicates missing data.
TABLE 2.� HOSTILITY LEVELS OF MILITARIZED INTERSTATE DISPUTES (MIDs).

<![if !supportEmptyParas]>�<![endif]>

22. Inter-state war

21. Use of chemical, bacteriological, or radiological weapons

20. declaration of war

19. other use of force

18. clash

17. seizure

16. occupation of territory

15. blockade

14. border violation

13. fortify border

12. mobilization

11. nuclear alert

10. alert

�9. show of planes

�8. show of ships

�7. show of troops

�6. threat to use nuclear weapons

�5. threat to declare war

�4. threat to occupy territory

�3. threat to blockade

�2. threat to use force

�1. no militarized action

-9. missing data (which in these cases means no codable use of force, display of force, or explicit threat to use force by target side has been detected, despite extensive search; later, for visual clarity of presentation, I will sometimes render the -9 fields as blank).

TABLE 3.� Liberal MIDs Ranked by Most Severe Act

EVENT AND SEVERITY�� NUMBER OF CASES

<![if !supportEmptyParas]>�<![endif]>

100.� War (Most Severe)�� 0��

99.5� Use of CBRadiological Weapons*�� 0

99.� Declaration of War�� 1

96.� Blockade�� 0

84.� Occupation of Territory�� 0

82.� Mobilization�� 0

81.� Seizure of Material or Personnel�� 21

75.� Clash�� 6

65.� Other Use of Military Force�� 15

60.� Border Violation*�� 5

50.� Show of Troops*�� 2

49.� Show of Ships*�� 6

48.� Show of Planes*�� 0

42.� Nuclear Alert*�� 0

40.� Threat to Use Nuclear Weapons*�� 0

36.� Fortify Border*�� 1

30.� Threat to Use Force�� 3

29.� Threat to Declare War�� 0

20.� Alert�� 2

8.� Threat to Occupy Territory�� 0

1.� Threat to Blockade (Least Severe)�� 0

Table 4.� Inter-Liberal MIDs Classified by Highest Hostility Level Taken by Initiator.

<![if !supportEmptyParas]>�<![endif]>

Hostility Level�� N

War�� 0

Use of Force�� 40

Display of Force�� 18

Threat to Use Force�� 4

Total�� 62

Table 5.� Initiator Actions and Target Actions in Liberal MIDs.

<![if !supportEmptyParas]>�<![endif]>

<![if !vml]>

<![endif]><![if !mso]>

<![endif]>

22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1

<![if !mso]>

<![endif]><![if !mso & !vml]>�<![endif]><![if !vml]>

<![endif]>Initiator Action

Target Action

Total

USE

DISPLAY

THREAT

NONE

<![if !supportEmptyParas]>�<![endif]>

WAR

<![if !supportEmptyParas]>�<![endif]>

CBR

<![if !supportEmptyParas]>�<![endif]>

DECL

<![if !supportEmptyParas]>�<![endif]>

OTHRUSE

<![if !supportEmptyParas]>�<![endif]>

CLASH

<![if !supportEmptyParas]>�<![endif]>

SEIZE

<![if !supportEmptyParas]>�<![endif]>

OCCUP

<![if !supportEmptyParas]>�<![endif]>

BLOCK

<![if !supportEmptyParas]>�<![endif]>

VIOL

<![if !supportEmptyParas]>�<![endif]>

FORT

<![if !supportEmptyParas]>�<![endif]>

MOBIL

<![if !supportEmptyParas]>�<![endif]>

NUCALR

<![if !supportEmptyParas]>�<![endif]>

ALERT

<![if !supportEmptyParas]>�<![endif]>

SPLAN

<![if !supportEmptyParas]>�<![endif]>

SSHIP

<![if !supportEmptyParas]>�<![endif]>

STROO

<![if !supportEmptyParas]>�<![endif]>

THREAT

THRNUK

<![if !supportEmptyParas]>�<![endif]>

THRWAR

<![if !supportEmptyParas]>�<![endif]>

THROCC

<![if !supportEmptyParas]>�<![endif]>

THRBLK

<![if !supportEmptyParas]>�<![endif]>

THRFOR

<![if !supportEmptyParas]>�<![endif]>

NONE

NOMIL

<![if !supportEmptyParas]>�<![endif]>

Total

<![if !supportEmptyParas]>�<![endif]>

Key to Action Abbreviations:

<![if !supportEmptyParas]>�<![endif]>

USE

DISPLAY

WAR (22) = Interstate War

VIOL (14) = Border Violation

CBR (21) = Use of CBR Weapons

FORT (13) = Fortify Border

DECL (20) = Declaration of War

MOBIL (12) = Mobilization

OTHRUSE (19) = Other Use of Force

NUCALR (11) = Nuclear Alert

CLASH (18) = Clash

ALERT (10) = Alert

SEIZE (17) = Seizure

SPLAN (9) = Show of Planes

OCCUP (16) = Occupation of Territory

SSHIP (8) - Show of Ships

BLOCK (15) = Blockade

STROO (7) = Show of Troops

<![if !supportEmptyParas]>�<![endif]>

THREAT

NONE

THRNUK (6) = Threat to Use Nuclear Weapons

NOMIL (1) = No Militarized Action

THRWAR (5) = Threat to Declare War

<![if !supportEmptyParas]>�<![endif]>

THROCC (4) = Threat to Occupy Territory

<![if !supportEmptyParas]>�<![endif]>

THRBLK (3) = Threat to Blockade

<![if !supportEmptyParas]>�<![endif]>

THRFOR (2) = Threat to Use Force

<![if !supportEmptyParas]>�<![endif]>

Table 6.� Initiator and Target Hostility Levels in Inter-Liberal MIDs.

<![if !supportEmptyParas]>�<![endif]>

�� Initiator Action

<![if !supportEmptyParas]>�<![endif]>

Target�� Use of�� Display of�� Threat to��

Action�� Force�� Force�� Use Force��

Use of Force�� 11�� 3�� 0

�� (28%)�� (17%)�� (0%)

<![if !supportEmptyParas]>�<![endif]>

Display of Force�� 1�� 5�� 1

�� (3%)�� (28%)�� (25%)

��

Threat of Force�� 2�� 0�� 0

�� (5%)�� (0%)�� (0%)

<![if !supportEmptyParas]>�<![endif]>

No Response�� 26�� 10�� 3

�� (65%)�� (56%)�� (75%)

Table 7.� Reciprocation (Target Response to Initiator) in Inter-Liberal and Other MIDs.

<![if !supportEmptyParas]>�<![endif]>

�� MID type, classified by Participants:

�� Inter-Liberal�� Other�� All

<![if !supportEmptyParas]>�<![endif]>

Response�� 23�� 990�� 1,013

<![if !supportEmptyParas]>�<![endif]>

No Response�� 39�� 990�� 1,029

<![if !supportEmptyParas]>�<![endif]>

Total:�� 62�� 1,980�� 2,042

<![if !supportEmptyParas]>�<![endif]>

Chi Square 4.02 (1 d.f.), significant at .05 level for 2-tailed test.

Table 8.� Bilateralism and Inter-Liberal MIDs.

<![if !supportEmptyParas]>�<![endif]>

�� MID Type, Classified by Participants:

�� Inter-Liberal�� Other�� All

<![if !supportEmptyParas]>�<![endif]>

Bilateral�� 61�� 1,651�� 1,712

<![if !supportEmptyParas]>�<![endif]>

Multilateral�� 1�� 329�� 330

<![if !supportEmptyParas]>�<![endif]>

Total:�� 62�� 1,980�� 2,042

<![if !supportEmptyParas]>�<![endif]>

Chi Square 9.99 (1 d.f.), significant at .01 level for 2-tailed test.

Table 9.� Battle Deaths and Inter-Liberal MIDs.

<![if !supportEmptyParas]>�<![endif]>

�� MID Type, Classified by Participants:

�� Inter-Liberal�� Other�� All

<![if !supportEmptyParas]>�<![endif]>

At least one death�� 3�� 347�� 350

<![if !supportEmptyParas]>�<![endif]>

No Reported Fatalities�� 59�� 1,633�� 1,692

<![if !supportEmptyParas]>�<![endif]>

Total:�� 62�� 1,980�� 2,042

<![if !supportEmptyParas]>�<![endif]>

Chi Square 6.82 (1 d.f.), significant at .01 level for 2-tailed test.

ACKNOWLEGDMENTS:

I would like to thank Phil Schafer for discussing the sixty-two cases of inter-liberal MIDs with me, Errol Henderson for discussions of the democratic peace, and Bram Wayman for creating the tables.

<![if !supportEmptyParas]>�<![endif]>

REFERENCES:

Australian Constitutional Commission (1988) Final Report of the�� Constitutional Commission, Commonwealth of Australia,�� reproduced in part at http://www.statusquo.org/indaus.htm.

BABST, DEAN (1964) "Elective Governments a Force for Peace." The�� Wisconsin Sociologist 3, 1: 9-14.

---� (1972) "A Force for Peace."� Industrial Research (April):�� 55-58.

CHURCHILL, WINSTON (1950) The Grand Alliance.� Boston:� Houghton�� Mifflin.

DOYLE, MICHAEL (1986) "Liberalism and World Politics."� American�� Political Science Review 80:� 1151-61.

Freedom House (2002) "Freedom in the World Country Ratings 1972-�� 73 to 2000-2001."� http://www.freedomhouse.org.

GOCHMAN, CHARLES, and ZEEV MAOZ (1984) "Militarized Interstate�� Disputes, 1816-1976: Procedures, Patterns, and Insights."� ��Journal of Conflict Resolution 28: 585a-616.

HENDERSON, ERROL (2002) Democracy and War.� Boulder, Colo.:�� Lynne Reiner.

JONES, DANIEL, STUART BREMER, and J. DAVID SINGER (1996)�� "Militarized Interstate Disputes 1816-1992:� Rationale,�� Coding Rules, and Empirical Patterns."� Conflict Management�� and Peace Science 15 (No. 2): 163-213.��

MANSBACH, RICHARD, and JOHN VASQUEZ (1981) In Search of Theory:�� A New Paradigm for Global Politics.� N.Y.:� Columbia Univ.�� Press.

MAOZ, ZEEV (1982) Paths to Conflict: International Dispute�� Initiation, 1816-1976.� Boulder, Colo.:� Westview Press.

---� (1997) "The Controversy over the Democratic Peace," ��International Security, Vol. 22, No. 1 (Summer): 162-197.

MEARSHEIMER, JOHN (2001) The Tragedy of Great Power Politics.�� N.Y.:� W.W. Norton.

ONEAL, JOHN, and BRUCE RUSSETT (1997) "The Classic Liberals Were�� Right:� Democracy, Interdependence, and Conflict: 1950-1985."� ��International Studies Quarterly 41, No. 2 (June): 267-294.

---� (2001) Triangulating Peace.� N.Y.:� W.W. Norton.

RAY, JAMES LEE (1995)� Democracy and International Conflict:� An�� Evaluation of the Democratic Peace Proposition.� Columbia,�� S.C.:� Univ. of South Carolina Press.��

RUMMEL, R.J. (1994) Death by Government.� New Brunswick, N.J.:�� Transaction Publishers.��

SENESE, PAUL (1997) "Between Dispute and War:� The Effect of�� Joint Democracy on Interstate Conflict Escalation."� Journal�� of Politics 59, No. 1 (Feb.): 1-27.

SMALL, MELVIN, and J. DAVID SINGER (1976) "The War Proneness of�� Democratic Regimes."� Jerusalem Journal of International�� Relations 1 (Summer): 50-69.

VASQUEZ, JOHN (1998) The Power of Power Politics.� Cambridge, UK:�� Cambridge University Press.

WAYMAN, FRANK, and PAUL DIEHL (1994) "Realism Reconsidered:� The�� Realpolitik Framework and Its Basic Propositions," pp. 3-28�� in Frank Wayman and Paul Diehl, eds., Reconstructing�� Realpolitik.� Ann Arbor, Mich.:� Univ. of Michigan Press.

Feb	APR	Jun
	05
2007	2008	2011