Democratic peace theory: Difference between revisions

Browse history interactively ← Previous edit Next edit →Content deleted Content addedVisual WikitextInline

Revision as of 19:05, 20 November 2006 editDoremítzwr (talk \| contribs)1,063 edits →Young democracies: Corrected grammar (less used with a countable noun)← Previous edit		Revision as of 22:54, 29 November 2006 edit undo82.235.131.83 (talk) added fr wiki linkNext edit →
Line 645:		Line 645:
	]		]
	]		]
			]

Revision as of 22:54, 29 November 2006

The democratic peace theory, democratic peace thesis or simply democratic peace (often DPT) is a theory and related empirical research in international relations, political science, and philosophy which holds that democracies—usually, liberal democracies—never or almost never go to war with one another.

The original theory and research on wars has been followed by many similar theories and related research on the relationship between democracy and peace, including that lesser conflicts than wars are also rare between democracies, and that systematic violence is in general less common within democracies.

History

The democratic peace theory is a relatively new development. One explanation is that democratic governments were scarce before the late 19th century. Although the philosophical idea has circulated since Emmanuel Kant, it was not scientifically evaluated until the 1960s.

File:Kant 2.jpg

Emmanuel Kant

Kant foreshadowed the theory in his essay Perpetual Peace written in 1795, although he thought that constitutional republics was only one of several necessary conditions for a perpetual peace. Kant's theory was that a majority of the people would never vote to go to war, unless in self defense. Therefore, if all nations were republics, it would end war, because there would be no aggressors. Other explanations have been proposed since, but the modern theory is principally the empirical claim that democracies rarely or never fight.

Dean Babst, a criminologist, was the first to do statistical research on this topic. He wrote an academic paper supporting the theory in 1964 in Wisconsin Sociologist; he published a slightly more popularized version, eight years later, in the trade journal Industrial Research. Both versions initially received little attention.

Melvin Small and J. David Singer (1976) eventually responded and denied that democracies were in general less war-like than other nations. However, they found only two marginal cases of democracies fighting each other. This paper was published in a political science journal which finally brought more widespread attention to the theory, as did Michael Doyle's (1983) lengthy discussion of the topic. Rudolph J. Rummel was another early researcher and drew considerable lay attention to the subject in his later works.

Maoz & Abdolali (1989) extended the research to lesser conflicts than wars. Bremer (1992) and Maoz & Russett (1992) found the correlation between democracy and peacefulness remained significant after controlling for many possible confounding variables. This moved the theory into the mainstream of social science. Supporters of Realism in international relations and others responded by raising many new objections. Other researchers attempted more systematic explanations of how democracy might cause peace, and of how democracy might also affect other aspects of foreign relations such as alliances and collaboration.

There have been numerous further studies in the field since these pioneering works. Most studies have found some form of democratic peace exists, although neither methodological disputes nor doubtful cases are entirely resolved (Kinsella 2005).

Influence

The democratic peace theory has been extremely divisive among political scientists. It is rooted in the idealist and classical liberalist traditions and is opposed to the previously dominant theory of realism. However, democratic peace theory has come to be more widely accepted and has in some democracies affected policy change.

Presidents of both the major American parties have expressed support for the theory. Former President Bill Clinton of the Democratic Party: "Ultimately, the best strategy to ensure our security and to build a durable peace is to support the advance of democracy elsewhere. Democracies don't attack each other." Current President George W. Bush of the Republican Party: "And the reason why I'm so strong on democracy is democracies don't go to war with each other. And the reason why is the people of most societies don't like war, and they understand what war means.... I've got great faith in democracies to promote peace. And that's why I'm such a strong believer that the way forward in the Middle East, the broader Middle East, is to promote democracy." The United States Congress has passed the Advance Democracy Act which states: "Wars between or among democratic countries are exceedingly rare, while wars between and among nondemocratic countries are commonplace, with nearly 170,000,000 people having lost their lives because of the policies of totalitarian governments."

Former European Commissioner for External Relations Chris Patten: "Inevitable because the EU was formed partly to protect liberal values, so it is hardly surprising that we should think it appropriate to speak out. But it is also sensible for strategic reasons. Free societies tend not to fight one another or to be bad neighbours." The A Secure Europe in a Better World, European Security Strategy states: "The best protection for our security is a world of well-governed democratic states."

Some fear that the democratic peace theory may be used to justify wars against nondemocracies in order to bring lasting peace, in a democratic crusade (Chan 1997, p. 59). Woodrow Wilson in 1917 asked Congress to declare war against Imperial Germany, citing Germany's sinking of American ships due to unrestricted submarine warfare and the Zimmerman telegram, but also stating that "A steadfast concert for peace can never be maintained except by a partnership of democratic nations" and "The world must be made safe for democracy." Some point out that the democratic peace theory has been used to justify the 2003 Iraq War, others argue that this justification was used only after the war had already started (Russett 2005). Furthermore, Weede (2004) has argued that the justification is extremely weak, because forcibly democratizing a country completely surrounded by non-democracies, most of which full autocracies, as Iraq is, is at least as likely to increase the risk of war as it is to decrease it (some studies show that dyads formed by one democracy and one autocracy are the most warlike, and several find that the risk of war is greatly increased in democratizing countries surrounded by nondemocracies). According to Weede, if the United States and its allies wanted to adopt a rationale strategy of forced democratization based on democratic peace, which he still does not recommend, it would be best to start intervening in countries which border with at least one or two stable democracies, and expand gradually. Also, research shows that attempts to create democracies by using external force has often failed eventually. Gleditsch, Christiansen and Hegre (2004) argue that forced democratization by interventionism may initially have partial success, but often create an unstable democratizing country, which can have dangerous consequences in the long run. Those attempts which had a permanent and stable success, like democratization in occupied Japan after World War II, mostly involved countries which had an advanced economic and social structure already, and implied a drastic change of the whole political culture. Supporting internal democratic movements and using diplomacy may be far more successful and less costly. Thus, the theory and related research, if they were correctly understood, may actually be an argument against a democratic crusade (Weart 1998), (Owen 2005), (Russett 2005).

Definitions

Research on the democratic peace theory has to define "democracy" and "peace" (or, more often, "war").

Democracy

Democracies have been defined differently by different theorists and researchers; this accounts for some of the variations in their findings. Some examples:

Kant (1795) opposed direct democracy since it is "necessarily despotism, as it establishes an executive power contrary to the general will; all being able to decide against one whose opinion may differ, the will of all is therefore not that of all: which is contradictory and opposite to liberty." Instead, Kant favors a constitutional republic where individual liberty is protected from the will of the majority.

Small and Singer (1976) define democracy as a nation that (1) holds periodic elections in which the opposition parties are as free to run as government parties, (2) allows at least 10% of the adult population to vote, and (3) has a parliament that either controls or enjoys parity with the executive branch of the government.

Doyle (1983) requires (1) that "liberal régimes" have market or private property economics, (2) they have polities that are externally sovereign, (3) they have citizens with juridical rights, and (4) they have representative governments. Either 30% of the adult males were able to vote or it was possible for every man to acquire voting rights as by attaining enough property. He allows greater power to hereditary monarchs than other researchers; for example, he counts the rule of Louis-Philippe of France as a liberal régime.

Ray (1995) requires that at least 50% of the adult population is allowed to vote and that there has been at least one peaceful, constitutional transfer of executive power from one independent political party to another by means of an election.

Rummel (1997) states that "By democracy is meant liberal democracy, where those who hold power are elected in competitive elections with a secret ballot and wide franchise (loosely understood as including at least 2/3rds of adult males); where there is freedom of speech, religion, and organization; and a constitutional framework of law to which the government is subordinate and that guarantees equal rights."

Weart (1998). See Never at War.

Non-binary classifications

The above definitions are binary, classifying nations into either democracies or nondemocracies. Many researchers have instead used more finely grained scales. One example is the Polity data series which scores each state on two scales, one for democracy and one for autocracy, for each year since 1800; as well as several others. The use of the Polity Data has varied. Some researchers have done correlations between the democracy scale and belligerence; others have treated it as a binary classification by (as its maker does) calling all states with a high democracy score and a low autocracy score democracies; yet others have used the difference of the two scores, sometimes again making this into a binary classification (Gleditsch 1992).

Young democracies

Several researchers have observed that many of the possible exceptions to the democratic peace have occurred when at least one of the involved democracies was very young. Many of them have therefore added a qualifier, typically stating that the peacefulness apply to democracies older than 3 years (Doyle 1983), (Russett 1993), (Rummel 1997), (Weart 1998). Rummel (1997) argues that this is enough time for "democratic procedures to be accepted, and democratic culture to settle in." Additionally, this may allow for other states to actually come to the recognition of the state as a democracy.

Mansfield and Snyder (2002, 2005), while agreeing that there have been no wars between mature liberal democracies, state that countries in transition to democracy are especially likely to be involved in wars. They find that democratizing countries are even more warlike than stable democracies, stable autocracies or even countries in transition towards autocracy. So, they suggest caution in eliminating these wars from the analysis, because this might hide a negative aspect of the process of democratization. A reanalysis of the earlier study's statistical results (Braumoeller 2004) harv error: no target: CITEREFBraumoeller2004 (help) emphasizes that the above relationship between democratization and war can only be said to hold for those democratizing countries where the executive lacks sufficient power, independence, and institutional strength. A review (Ray 2003) harv error: no target: CITEREFRay2003 (help) cites several other studies finding that the increase in the risk of war in democratizing countries happens only if many or most of the surrounding nations are undemocratic. If wars between young democracies are included in the analysis, several studies and reviews still find enough evidence supporting the stronger claim that all democracies, whether young or established, go unto war with one another less frequently (Ray 1998), (Ray 2003), (Hegre 2004) harv error: no target: CITEREFHegre2004 (help), while some do not (Schwartz & Skinner 2002) harv error: no target: CITEREFSchwartzSkinner2002 (help).

Wars and lesser conflicts

Quantitative research on international wars usually define war as a military conflict with more than 1000 killed in battle. This is the definition used in the Correlates of War Project which has also supplied the data for many studies on war. It turns out that most of the military conflicts in question fall clearly above or below this threshold (Ray 1995, p. 103).

Some researchers have used different definitions. For example, Weart (1998) defines war as more than 200 battle deaths. Russett (1993, p. 50), when looking at Ancient Greece, only requires some real battle engagement, involving on both sides forces under state authorization.

Militarized Interstate Disputes (MIDs), in the Correlates of War Project classification, are lesser conflicts than wars. Such a conflict may be no more than military display of force with no battle deaths. MIDs and wars together are "militarized interstate conflicts" or MICs. MIDs include the conflicts that precede a war; so the difference between MIDs and MICs may be less than it appears.

Statistical analysis and concerns about degrees of freedom are the primary reasons for using MID's instead of actual wars. Wars are relatively rare. An average ratio of 30 MIDs to one war provides a richer statistical environment for analysis.

The monadic peace and the dyadic peace

Most research is regarding the dyadic peace, that democracies do not fight one another. Very few researchers have supported the monadic peace, that democracies are more peaceful in general. There are some recent papers, which find a slight monadic effect. Müller and Wolff (2004), in listing them, agree "that democracies on average might be slightly, but not strongly, less warlike than other states," but general "monadic explanations is neither necessary nor convincing". They note that democracies have varied greatly in their belligerence against non-democracies. The most militant democracies since 1950 have been India, Israel, the United Kingdom and the United States.

Wars

The straightforward argument for the democratic peace is: given the number of wars over the past two centuries, if democracies fought each other as often as any other pair of states, there should have been many wars between democracies. Instead, depending on the study, there have been zero or very few. A review (Ray 1998) harv error: no target: CITEREFRay1998 (help) lists many studies finding that this peacefulness is statistically significant.

Possible exceptions to no wars

Main article: List of possible exceptions to democratic peace theory

Several researchers find no wars between well-established liberal democracies. Jack Levy (1988) made an oft-quoted assertion that the theory is "as close as anything we have to an empirical law in international relations".

Others see one or two exceptions. Some wars commonly suggested as exceptions are the Spanish-American War, the Continuation War and, recently, the Kargil War. Some of those who see exceptions regard them as marginal cases.

Other authors simply describe war between democracies as "rare", "very rare", "rare or non-existent".

The question of no or few wars may be unimportant. Bremer (1992, 1993), who strongly supports the democratic peace, argues that it is impossible to prove a probability of exactly zero wars between democracies; thus is "fruitless to debate the question of whether democracies never or only very rarely fight one another". It is only possible to show a decrease in the probability of war.

However, at least one researcher (Rummel 1983) has argued that one exception will disprove the theory. Most researchers disagree (Gleditsch 1992).

Lesser conflicts

One problem with the research on wars is that, as the Realist Mearsheimer (1990, p. 50) put it, "democracies have been few in number over the past two centuries, and thus there have been few opportunities where democracies were in a position to fight one another". Especially if using a strict definition of democracy, as by those finding no wars. Democracies have been very rare until recently. Even looser definitions of democracy, such as Doyle's, find only a dozen democracies before the late nineteenth century, and many of them short-lived or with limited franchise (Doyle 1983), (Doyle 1997, p. 261). Freedom House finds no independent state with universal suffrage in 1900.

Wayman (1998), a supporter of the theory, states that "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".

Thus, despite the studies mentioned earlier, some argue that there is not enough data to show that the absence of wars between democracies is statistically significant, especially if trying to check for the possible influence of external factors (see also the discussion in the section on criticism and counter-criticism).

Many researchers reacted to this limitation by studying lesser conflicts instead, since they have been far more common. There have been many more MIDs than wars; the Correlates of War Project counts several thousands during the last two centuries. A review (Ray 2003) harv error: no target: CITEREFRay2003 (help) lists many studies that have reported that democratic pairs of states are less likely to be involved in MIDs than other pairs of states.

Another study (Hensel, Goertz & Diehl 2000) harv error: no target: CITEREFHenselGoertzDiehl2000 (help) finds that after both states have become democratic, there is a decreasing probability for MIDs within a year and this decreases almost to zero within five years.

When examining the inter-liberal MIDs in more detail, one study (Wayman 2002) harv error: no target: CITEREFWayman2002 (help) finds that they are less likely to involve third parties, the target of the hostility is less likely reciprocate, if the target reciprocates the response is usually proportional to the provocation, and the disputes are less likely to cause any loss of life. The most common action was "Seizure of Material or Personnel".

Studies find that the probability that disputes between states will be resolved peacefully is positively affected by the degree of democracy exhibited by the least democratic state involved in that dispute. Disputes between democratic states are significantly shorter than disputes involving at least one undemocratic state. Democratic states are more likely to be amenable to third party mediation when they are involved in disputes with each other (Ray 2003) harv error: no target: CITEREFRay2003 (help).

In international crises that include the threat or use of military force, one study finds that if the parties are democracies, then relative military strength has no effect on who wins. This is different from when nondemocracies are involved. These results are the same also if the conflicting parties are formal allies (Gelpi & Griesdorf 2001) harv error: no target: CITEREFGelpiGriesdorf2001 (help). Similarly, a study of the behavior of states that joined ongoing militarized disputes reports that power is important only to autocracies: democracies do not seem to base their alignment on the power of the sides in the dispute (Werner & Lemke 1997) harv error: no target: CITEREFWernerLemke1997 (help).

Conflict initiation

Most studies have looked only at who is involved in the conflicts and ignored the question of who initiated the conflict. In many conflicts both sides argue that the other side was initiator. Several researchers, as described in (Gleditsch, Christiansen & Hegre 2004), have argued that studying conflict initiation is of limited value, because existing data about conflict initiation may be especially unreliable. Even so, several studies have examined this. Reiter and Stam (2003) argue that autocracies initiates conflicts against democracies more frequently than democracies do against autocracies. Quackenbush and Rudy (2006), while confirming Reiter and Stam's results, find that democracies initiate wars against nondemocracies more frequently than nondemocracies do to each other. Several following studies (Peceny & Beer 2003), (Peceny & Butler 2004), (Lai & Slater 2006) have studied how different types of autocracies with different institutions vary regarding conflict initiation. Personalistic and military dictatorships may be particularly prone to conflict initiation, as compared to other types of autocracy such as one party states, but also more likely to be targeted in a war having other initiators.

Internal violence and democide

Most of this article discusses research on relations between states. However, there is also evidence that democracies have less internal systematic violence. For instance, one study finds that the most democratic and the most authoritarian states have few civil wars, and intermediate regimes the most. The probability for a civil war is also increased by political change, regardless whether toward greater democracy or greater autocracy. Intermediate regimes continue to be the most prone to civil war, regardless of the time since the political change. In the long run, since intermediate regimes are less stable than autocracies, which in turn are less stable than democracies, durable democracy is the most probable end-point of the process of democratization (Hegre et al. 2001) harv error: no target: CITEREFHegreEllingtonGatesGleditsch2001 (help). Abadie (2004) study finds that the most democratic nations have the least terrorism. Harff (2003) finds that genocide and politicide are rare in democracies. Rummel (1997) finds that the more democratic a regime, the less its democide. He finds that democide has killed six times as many people as battles.

Davenport and Armstrong (2004) lists several other studies and states: "Repeatedly, democratic political systems have been found to decrease political bans, censorship, torture, disappearances and mass killing, doing so in a linear fashion across diverse measurements, methodologies, time periods, countries, and contexts." It concludes: "Across measures and methodological techniques, it is found that below a certain level, democracy has no impact on human rights violations, but above this level democracy influences repression in a negative and roughly linear manner." Davenport and Armstrong (2003) states that thirty years worth of statistical research has revealed that only two variables decrease human rights violations: political democracy and economic development. Of this democracy is more important and more easily created.

Possible explanations for how democracy may cause peace

These theories have traditionally been categorized into two groups: explanations that focus on democratic norms and explanations that focus on democratic political structures (Gelpi & Griesdorf 2001) harv error: no target: CITEREFGelpiGriesdorf2001 (help), (Braumoeller 1997) harv error: no target: CITEREFBraumoeller1997 (help). Note that they usually are meant to be explanations for little violence between democracies, not for a low level of internal violence in democracies.

Several of these mechanisms may also apply to oligarchies. The book Never at War find evidence for an oligarchic peace. One example is the Polish-Lithuanian Commonwealth, in which the Sejm resisted and vetoed most royal proposals for war, like those of Władysław IV Vasa.

Democratic norms

One example from the first group is that liberal democratic culture may make the leaders accustomed to negotiation and compromise (Weart 1998) harv error: no target: CITEREFWeart1998 (help), (Müller & Wolff 2004). Another that a belief in human rights may make people in democracies reluctant to go to war, especially against other democracies. The decline in colonialism, also by democracies, may be related to a change in perception of non-European peoples and their rights (Ravlo & Gleditsch 2000) harv error: no target: CITEREFRavloGleditsch2000 (help).

Bruce Russett (1993, p. 5-11, 35, 59-62, 73-4) also argues that the democratic culture affect the way leaders resolve conflicts. In addition, he holds that a social norm emerged toward the end of the nineteenth, that democracies should not fight each other, which strengthened when the democratic culture and the degree of democracy increased, for example by widening the franchise. Increasing democratic stability allowed partners in foreign affairs to perceive a nation as reliable democratic. The alliances between democracies during the two World Wars and the Cold War also strengthened the norms. He sees less effective traces of this norm in Greek antiquity.

Mousseau (2000) argues that it is market-oriented development that creates the norms and values that explain the democratic peace. When opportunities to contract in the market are widespread, as in market-oriented developed countries, a culture of contracting emerges that encourages shared respect for individualism, negotiations, compromise, respect for the law, and equality before the law. Constrained by voters, democratically elected leaders in market-oriented developed countries abide by these norms. In contrast, voters in democracies without developed market economies, and the leaders in nondemocracies, have other norms and values that encourage conflict.

(Braumoeller 1997) harv error: no target: CITEREFBraumoeller1997 (help) argues that liberal norms of conflict resolution vary because liberalism takes many forms. By examining survey results from the newly-independent states of the former Soviet Union, the author demonstrates that liberalism in that region bears a stronger resemblance to 19th-century liberal nationalism than to the sort of universalist, Wilsonian liberalism described by democratic peace theorists, and that, as a result, liberals in the region are more, not less, aggressive than non-liberals.

Democratic political structures

The case for institutional constraints goes back to Kant (1795), who wrote :

"f the consent of the citizens is required in order to decide that war should be declared (and in this constitution it cannot but be the case), nothing is more natural than that they would be very cautious in commencing such a poor game, decreeing for themselves all the calamities of war. Among the latter would be: having to fight, having to pay the costs of war from their own resources, having painfully to repair the devastation war leaves behind, and, to fill up the measure of evils, load themselves with a heavy national debt that would embitter peace itself and that can never be liquidated on account of constant wars in the future"

Democracy thus gives influence to those most likely to be killed or wounded in wars, and their relatives and friends (and to those who pay the bulk of the war taxes) Russett (1993, p. 30). This monadic theory must, however, explain why democracies do attack non-democratic states. One explanation is that these democracies were threatened or otherwise were provoked by the non-democratic states. Doyle (1997, p. 272) argued that the absence of a monadic peace is only to be expected: the same ideologies that cause liberal states to be at peace with each other inspire idealistic wars with the illiberal, whether to defend oppressed foreign minorities or avenge countrymen settled abroad. Doyle also notes (p. 292) liberal states do conduct covert operations against each other; the covert nature of the operation, however, prevents the publicity otherwise characteristic of a free state from applying to the question

Studies show that democratic states are more likely than autocratic states to win the wars. One explanation is that democracies, for internal political and economic reasons, have greater resources. This might mean that democratic leaders are unlikely to select other democratic states as targets because they perceive them to be particularly formidable opponents. One study finds that interstate wars have important impacts on the fate of political regimes, and that the probability that a political leader will fall from power in the wake of a lost war is particularly high in democratic states (Ray 1998) harv error: no target: CITEREFRay1998 (help).

As described in (Gelpi & Griesdorf 2001) harv error: no target: CITEREFGelpiGriesdorf2001 (help), several studies have argued that liberal leaders face institutionalized constraints that impede their capacity to mobilize the state’s resources for war without the consent of a broad spectrum of interests. Survey results that compare the attitudes of citizens and elites in the Soviet successor states are consistent with this argument (Braumoeller 1997) harv error: no target: CITEREFBraumoeller1997 (help). Moreover, these constraints are readily apparent to other states and cannot be manipulated by leaders. Thus, democracies send credible signals to other states of an aversion to using force. These signals allow democratic states to avoid conflicts with one another, but they may attract aggression from nondemocratic states. Democracies may be pressured to respond to such aggression—perhaps even preemptively—through the use of force. Also as described in (Gelpi & Griesdorf 2001) harv error: no target: CITEREFGelpiGriesdorf2001 (help), studies have argued that when democratic leaders do choose to escalate international crises, their threats are taken as highly credible, since there must be a relatively large public opinion for these actions. In disputes between liberal states, the credibility of their bargaining signals allows them to negotiate a peaceful settlement before mobilization.

An explanation based on game theory similar to the last two above is that the participation of the public and the open debate send clear and reliable information regarding the intentions of democracies to other states. In contrast, it is difficult to know the intentions of nondemocratic leaders, what effect concessions will have, and if promises will be kept. Thus there will be mistrust and unwillingness to make concessions if at least one of the parties in a dispute is a nondemocracy (Levy & Razin 2004) harv error: no target: CITEREFLevyRazin2004 (help).

Criticisms and counter-criticisms

There are several logically distinguishable classes of criticism. Note that they usually apply to no wars or few MIDs between democracies, not to little systematic violence in established democracies. See List of possible exceptions to democratic peace theory for a discussion of specific historical cases.

The theory may be empirically wrong or statistically not significant

Only one study (Schwartz & Skinner 2002) harv error: no target: CITEREFSchwartzSkinner2002 (help) appears to have argued that there have been as many wars between democracies as one would expect between any other couple of states. However, its authors include wars between young and dubious democracies, and very small wars.

Others (Spiro 1994), (Gowa 1999) harv error: no target: CITEREFGowa1999 (help), (Small & Singer 1976) harv error: no target: CITEREFSmallSinger1976 (help) state that, although there may be some evidence for democratic peace, the data sample or the time span may be too small to assess any definitive conclusions. For example, Gowa finds evidence for democratic peace to be insignificant before 1939, because of the too small number of democracies, and offers an alternate explanation for the following period (see the section on Realist Explanations). Gowa's use of statistics has been criticized, with several other studies and reviews finding different or opposing results (Gelpi & Griesdorf 2001) harv error: no target: CITEREFGelpiGriesdorf2001 (help), (Ray 2003) harv error: no target: CITEREFRay2003 (help). However, this can be seen as the longest-lasting criticism to the theory; as noted earlier, also some supporters (Wayman 1998) harv error: no target: CITEREFWayman1998 (help) agree that the statistical sample for assessing its validity is limited or scarce, at least if only full scale wars are considered.

It can be interesting to consider the question of "how much" significant the evidence is. One study (Ray 2003) harv error: no target: CITEREFRay2003 (help) tries to answer this question in a straightforward way. According to Ray, who uses a rather restrictive definition of democracy and war, there have been no wars between jointly democratic couples of states in the period from 1816 to 1992. Assuming a purely random distribution of wars between states, regardless of their democratic character, the predicted number of conflicts between democracies would be around ten. So, Ray argues that the evidence is statistically significant, but that it is still conceivable that, in the future, even a small number of inter-democratic wars wipes out such evidence.

There may be problems in definitions, methodology, or data

Some authors criticize the definition of democracy by arguing that states continually reinterpret other states' regime types as a consequence of their own objective interests and motives, such as economical and security concerns (Rosato 2003) harv error: no target: CITEREFRosato2003 (help). For example, one study (Oren 1995) harv error: no target: CITEREFOren1995 (help) reports that Germany was considered a democratic state by Western opinion leaders at the end of the 19th century; yet in the years preceding World War I, when its relations with the United States, France and Britain started deteriorating, Germany was gradually reinterpreted as an autocratic state, in absence of any actual regime change. Shimmin (Shimmin 1999) harv error: no target: CITEREFShimmin1999 (help) moves a similar criticism regarding the western perception of Milosevic's Serbia between 1989 and 1999. Rummel (Rummel 1999) harv error: no target: CITEREFRummel1999 (help) replies to the above criticism by stating that, in general, studies on democratic peace do not focus on the "western" perception of democracy; and in the specific case of Serbia, by arguing that the limited credit accorded by western democracies to Milosevic in the early '90s did not amount to a recognition of democracy, but only to the perception that possible alternative leaders could be even worse.

Some democratic peace researchers have been criticized for post hoc reclassifying some specific conflicts as non wars or political systems as non democracies without checking and correcting the whole data set used similarly. Supporters and opponents of the democratic peace agree that this is bad use of statistics, even if a plausible case can be made for the correction (Bremer 1992), (Gleditsch 1995), (Gowa 1999). A military affairs columnist of the newspaper Asia Times has summarized the above criticism in a journalist's fashion describing the theory as subject to the no true Scotsman problem: exceptions are explained away as not being between real democracies or being real wars.

However, most researchers agree that an objective working definition of "democracy" and "war" can be given. Even so, democracy is an evolving concept which has meant different things at different times, but in almost all cases researchers apply the same criteria to all history.

Definitions of democracy that require an actual transfer of power between different political parties sometimes exclude long periods often viewed as democratic. For example, the United States until 1800, India from independence until 1979, and Japan until 1993 (Ray 1995, p. 100).

Some democratic peace researchers require that the executive result from a substantively contested election. This may be a restrictive definition: For example, the National Archives of the United States notes that "For all intents and purposes, George Washington was unopposed for election as President, both in 1789 and 1792". (Under the original provisions for the Electoral College, there was no distinction between votes for President and Vice-President: each elector was required to vote for two distinct candidates, with the runner-up to be Vice-President. Every elector cast one of his votes for Washington, John Adams received a majority of the other votes; there were several other candidates: so the election for Vice President was contested.)

Spiro (1994) made several other criticisms of the statistical methods used. Russett (1995) and a series of papers described by Ray (2003) responded to this, for example with different methodology.

Sometimes the datasets used have also been criticized. For example, some authors have criticized the Correlates of War data for not including civilian deaths in the battle deaths count, especially in civil wars (Sambanis 2001) harv error: no target: CITEREFSambanis2001 (help). Weeks and Cohen (2006) argue that most fishing disputes, which include no deaths and generally very limited threats of violence, should be excluded even from the list of military disputes. Gleditsch (2004) made several criticisms to the Correlates of War data set, and produced a revised set of data. Maoz and Russett (1993) made several criticisms to the Polity I and II data sets, which have mostly been addressed in later versions. These criticisms are generally considered minor issues.

Correlation is not causation: There may be other explanations

If phenomenon A is found to be correlated with phenomenon B, there are in principle several possibilities regarding the origin of such correlation: A may cause B, B may cause A, both A and B may be caused by a third phenomenon C, or they may be caused by two different phenomena which are themselves correlated, and other, more complex, combinations. Many researchers, while accepting the empirical findings of democratic peace, have looked for different or complementary explanations, connections, and statistical variables which may account for such evidence.

Political similarity may be the explanation: the Autocratic Peace

One general criticism motivating research of different explanations is that actually the theory cannot claim that "democracy causes peace", because the evidence for democracies being, in general, more peaceful is very slight or non existent; it only can support the claim that "joint democracy causes peace". According to Rosato (2003), this casts doubts on whether democracy is actually the cause because, if so, a monadic effect would be expected.

Perhaps the simplest explanation to such perceived anomaly (but not the one the Realist Rosato prefers, see the section on Realist explanations below) is that democracies are not peaceful to each other because they are democratic, but rather because they are similar. This line of thought started with several independent observations of an "Autocratic Peace" effect, a reduced probability of war (obviously no author claims its absence) between states which are both non-democratic, or both highly so (Raknerud & Hegre 1997) harv error: no target: CITEREFRaknerudHegre1997 (help), (Beck & Jackman 1998) harv error: no target: CITEREFBeckJackman1998 (help), This has lead to the hypothesis that democratic peace emerges as a particular case when analyzing a subset of states which are, in fact, similar (Werner 2000) harv error: no target: CITEREFWerner2000 (help). Or, that not similarity in general, but only coherence of strong political regimes such as full democracies and stark autocracies, might affect the probability of war.

Autocratic peace and the explanation based on political similarity is a relatively recent development, and opinions about its value are varied. Henderson (2002) builds a model considering political similarity, geographic distance and economic interdependence as its main variables, and concludes that democratic peace is a statistical artifact which disappears when the above variables are taken into account. Werner (2000) finds a conflict reducing effect from political similarity in general, but with democratic dyads being particularly peaceful, and noting some differences in behavior between democratic and autocratic dyads with respect to alliances and power evaluation. Beck, King and Zeng (2004) use neural networks to show two distinct low probability zones, corresponding to high democracy and high autocracy. Petersen (2004) uses a different statistical model and finds that autocratic peace is not statistically significant, and that the effect attributed to similarity is mostly driven by the pacifying effect of joint democracy. Ray (2005) similarly disputes the weight of the argument on logical grounds, claiming that statistical analysis on "political similarity" uses a main variable which is an extension of "joint democracy" by linguistic redefinition, and so it is expected that the war reducing effects are carried on in the new analysis. Bennett (2006) builds a direct statistical model based on a triadic classification of states into "democratic", "autocratic" and "mixed". He finds that autocratic dyads have a 35% reduced chance of going into any type of armed conflict with respect to a reference mixed dyad. Democratic dyads have a 55% reduced chance. This effect gets stronger when looking at more severe conflicts; for wars (more than 1000 battle deaths), he estimates democratic dyads to have an 82% lower risk than autocratic dyads. He concludes that autocratic peace exists, but democratic peace is clearly stronger. However, he finds no relevant pacifying effect of political similarity, except at the extremes of the scale.

To summarize a rather complex picture, there are no less than four possible stances on the value of this criticism:

Political similarity, plus some complementary variables, explains everything. Democratic peace is a statistical artifact. Henderson subscribes to this view.
Political similarity has a pacifying effect, but democracy makes it stronger. Werner would probably subscribe to this view.
Political similarity in general has little or no effect, except at the extremes of the democracy-autocracy scale: a democratic peace and an autocratic peace exist separately, with the first one being stronger, and may have different explanations. Bennett holds this view, and Kinsella mentions this as a possibility
Political similarity has little or no effect and there is no evidence for autocratic peace. Petersen and Ray are among defendants of this view.

Economic factors may be the explanation: wealth, growth, and interdependence

A majority of researchers on the determinants of democracy agree that economic development is a primary factor which allows the formation of a stable and healthy democracy (Hegre, 2003; Weede, 2004). This in itself is not in contradiction with democratic peace theory; it is just a statement about the nature of democracy; however, if a causal link between some economic factor and peace could be found, one could hope to explain the findings of the theory on a purely economical basis.

Mousseau argues that a culture of contracting in advanced market-oriented economies may cause both democracy and peace (2000; 2002; 2003; 2005). Low economic development may hinder development of liberal institutions and values. Hegre (2000) and Souva (2003) reach similar conclusions. These studies indicate that democracy, alone, is an unlikely cause of the democratic peace. Mousseau (2005) finds that democracy is a significant factor only when both democracies have levels of economic development well above the global median. The poorest 21% of the democracies studied, and the poorest 4-5% of current democracies, are significantly more likely to fight each other. Oneal & Russett (2004) list several other studies which find smaller effects from economic development when controlling for one of the Kantian variables, trade. Mousseau, Hegre & Oneal (2003) confirm that if at least one of the democracies involved has a very low level of economic development, democracy is ineffective in preventing war; however, they find that when also controlling for trade, 91% of all the democratic pairs had high enough development for the pacifying effect of democracy to be important during the 1885–1992 period and all in 1992. Hegre (2003) finds that democracy is correlated with civil peace only for developed countries, and for countries with high levels of literacy. Conversely, the risk of civil war decreases with development only for democratic countries. Considering that both World Wars were fought between countries which can be considered highly developed by most criteria casts serious doubts about whether economic development alone can explain democratic peace.

Gartzke (2005) argues that economic freedom has a much stronger effect than democracy. Rummel (2005) criticizes the methodology used and finds the results invalid.

Several studies find that democracy, more trade causing greater economic interdependence, and membership in more intergovernmental organizations reduce the risk of war. This is often called the Kantian peace theory since it is similar to Kant's earlier theory about a perpetual peace; it is often also called "liberal peace" theory, especially when one focuses on the effects of trade and democracy. (The theory that free trade can cause peace is quite old and referred to as Cobdenism.) Many researchers agree that these variables positively affect each other but each has a separate pacifying effect. For example, in countries exchanging a substantial amount of trade, economic interest groups may exist that oppose a reciprocal disruptive war, but in democracy such groups may have more power, and the political leaders be more likely to accept their requests. (Russett & Oneal 2001), (Lagazio & Russett 2004) harv error: no target: CITEREFLagazioRussett2004 (help), (Oneal & Russett 2004) harv error: no target: CITEREFOnealRussett2004 (help). Weede (2004) argues that the pacifying effect of free trade and economic interdependence may be more important than that of democracy, because the former affects peace both directly and indirectly, by producing economic development and ultimately, democracy. Weede also lists some other authors supporting this view. However, some recent studies find no effect from trade but only from democracy (Goenner 2004) harv error: no target: CITEREFGoenner2004 (help), (Kim & Rousseau 2005) harv error: no target: CITEREFKimRousseau2005 (help).

None of the authors listed argues that free trade alone causes peace. Even so, the issue of whether free trade or democracy is more important in maintaining peace may have potentially significant practical consequences, for example on evaluating the effectiveness of applying economic sanctions and restrictions to autocratic countries.

It was Michael Doyle (1983, 1997) who reintroduced Kant's three articles into democratic peace theory. He argued that a pacific union of liberal states has been growing for the past two centuries. He denies that a pair of states will be peaceful simply because they are both liberal democracies; if that were enough, liberal states would not be aggressive towards weak non-liberal states (as the history of American relations with Mexico shows they are). Rather, liberal democracy is a necessary condition for international organization and hospitality (which are Kant's other two articles) — and all three are sufficient to produce peace. Other Kantians have not repeated Doyle's argument that all three in the triad must be present, instead stating that all three reduce the risk of war.

Other direct explanations of the correlation

Many studies, as those discussed in (Ray 1998) harv error: no target: CITEREFRay1998 (help), (Ray 2005) harv error: no target: CITEREFRay2005 (help), (Oneal & Russett 2004) harv error: no target: CITEREFOnealRussett2004 (help), supporting the theory have controlled for many possible alternative causes of the peace. Examples of factors controlled for are geographic distance, geographic contiguity, power status, alliance ties, militarization, economic wealth and economic growth, power ratio, and political stability. These studies have often found very different results depending on methodology and included variables, which has caused criticism. It should be noted that DPT does not state democracy is the only thing affecting the risk of military conflict. Many of the mentioned studies have found that other factors are also important. However, a common thread in most results is an emphasis on the relationship between democracy and peace.

Several studies have also controlled for the possibility of reverse causality from peace to democracy. For example, one study (Reuveny & Li 2003) harv error: no target: CITEREFReuvenyLi2003 (help) supports the theory of simultaneous causation, finding that dyads involved in wars are likely to experience a decrease in joint democracy, which in turn increases the probability of further war. So they argue that disputes between democratizing or democratic states should be resolved externally at a very early stage, in order stabilize the system. Another study (Reiter 2001) harv error: no target: CITEREFReiter2001 (help) finds that peace does not spread democracy, but spreading democracy is likely to spread peace. A different kind of reverse causation lies in the suggestion that impending war could destroy or decrease democracy, because the preparation for war might include political restrictions, which may be the cause for the findings of democratic peace. However, this hypothesis has been statistically tested in a study (Mousseau & Shi 1999) harv error: no target: CITEREFMousseauShi1999 (help) whose authors find, depending on the definition of the pre-war period, no such effect or a very slight one. So, they find this explanation unlikely. Note also that this explanation would predict a monadic effect, although weaker than the dyadic one.

Weart (1998) argues that the peacefulness appears and disappears rapidly when democracy appears and disappears. This in his view makes it unlikely that variables that change more slowly are the explanation. Weart, however, has been criticized for not offering any quantitative analysis supporting his claims (Ray, 2000).

Wars tend very strongly to be between neighboring states. Gleditsch (1995) showed that the average distance between democracies is about 8000 miles, the same as the average distance between all states. He believes that the effect of distance in preventing war, modified by the democratic peace, explains the incidence of war as fully as it can be explained.

Realist explanations

Supporters of realism in international relations in general argue that not democracy or its absence, but considerations and evaluations of power, cause peace or war. Specifically, many realist critics claim that the effect ascribed to democratic, or liberal, peace, is in fact due to alliance ties between democratic states which in turn are caused, one way or another, by realist factors.

For example, Farber and Gowa (1995) find evidence for peace between democracies to be statistically significant only in the period from 1945 on, and consider such peace an artifact of the Cold War, when the threat from the communist states forced democracies to ally with one another. Mearsheimer (1990) offers a similar analysis of the Anglo-American peace before 1945, caused by the German threat. Spiro (1994) finds several instances of wars between democracies, arguing that evidence in favor of the theory might be not so vast as other authors report, and claims that the remaining evidence consists of peace between allied states with shared objectives. He acknowledges that democratic states might have a somewhat greater tendency to ally with one another, and regards this as the only real effect of democratic peace. Rosato (2003) argues that most of the significant evidence for democratic peace has been observed after World War II; and that it has happened within a broad alliance, which can be identified with NATO and its satellite nations, imposed and maintained by American dominance. One of the main points in Rosato's argument is that, although never engaged in open war with another liberal democracy during the Cold War, the United States intervened openly or covertly in the political affairs of democratic states several times, for example in the Chilean coup of 1973, the 1953 coup in Iran and 1954 coup in Guatemala; in Rosato's view, these interventions show the United States' determination to maintain an "imperial peace".

The most direct counter arguments to such criticisms have been studies finding peace between democracies to be significant even when controlling for "common interests" as reflected in alliance ties (Gelpi & Griesdorf 2001) harv error: no target: CITEREFGelpiGriesdorf2001 (help), (Ray 2003) harv error: no target: CITEREFRay2003 (help). Regarding specific issues, Ray (1998) objects that explanations based on the Cold War should predict that the Communist bloc would be at peace within itself also, but exceptions include the Soviet Invasion of Afghanistan, the Cambodian-Vietnamese War, and the Sino-Vietnamese War. Ray also argues that the external threat did not prevent conflicts in the Western bloc when at least one of the involved states was a nondemocracy, such as the Turkish Invasion of Cyprus, the Falklands War, and the Football War. Also, one study (Ravlo & Gleditsch 2000) harv error: no target: CITEREFRavloGleditsch2000 (help) notes that the explanation "goes increasingly stale as the post-Cold War world accumulates an increasing number of peaceful dyad-years between democracies". Rosato's argument about American dominance has also been criticized for not giving supporting statistical evidence (Slantchev, Alexandrova & Gartzke 2005).

Some realist authors also criticize in detail the explanations given by supporters of democratic peace, pointing to supposed inconsistencies or weaknesses.

Rosato (2003) criticizes most explanations to how democracy might cause pace. Arguments based on normative constraints, he argues, are not consistent with the fact that democracies do go to war not less other states, thus violating norms preventing war; for the same reason he refutes arguments based on the importance of public opinion. Regarding explanations based on greater accountability of leaders, he finds that historically autocratic leaders have been removed or punished more often than democratic leaders when they get involved in costly wars. Finally, he also criticizes the arguments that democracies treat each other with trust and respect even during crises; and that democracy might be slow to mobilize its composite and diverse groups and opinions, hindering the start of a war, drawing support from other authors. Another realist, Layne (1994) analyzes the crises and brinkmanship that took place between non-allied democratic great powers, during the relatively brief period when such existed. He finds no evidence either of institutional or cultural constraints against war; indeed, there was popular sentiment in favor of war on both sides. Instead, in all cases, one side concluded that it could not afford to risk that war at that time, and made the necessary concessions.

Rosato's objections have been criticized for claimed logical and methodological errors, and for being contradicted by existing statistical research (Slantchev, Alexandrova & Gartzke 2005), (Kinsella 2005). Russett (1995) replies to Layne by re-examining some of the crises studied in his article, and reaching different conclusions; Russett argues that perceptions of democracy prevented escalation, or played a major role in doing so. Also, a recent study (Gelpi & Griesdorf 2001) finds that, while in general the outcome of international disputes is highly influenced by the contenders' relative military strength, this is not true if both contenders are democratic states; in this case the authors find the outcome of the crisis to be independent of the military capabilities of contenders, which is contrary to realist expectations. Finally, both the realist criticisms here described ignore new possible explanations, like the game-theoretic one discussed below.

A different kind of realist criticism (see (Jervis 2002) harv error: no target: CITEREFJervis2002 (help) for a discussion) is centered around the role of nuclear weapons in maintaining peace. In realist terms, this means that, in the case of disputes between nuclear powers, respective evaluation of power might be irrelevant because of Mutual assured destruction preventing both sides from foreseeing what could be reasonably called a "victory". An obvious rebuttal is that nuclear powers have been too few to account for the evidence in favor of democratic peace, except a very small part of it. The rebuttal remains valid even considering the mitigating argument that some advanced democracies, for example Germany and Japan, would be able to complete a nuclear program in a very brief period of time if a possible nuclear menace arose. The 1999 Kargil War between India and Pakistan has been cited as a counterexample to this argument (Page Fortna, 2004).

Some supporters of the democratic peace do not deny that realist factors are also important (Russett 1995). Research supporting the theory has also shown that factors such as alliance ties and major power status influence interstate conflict behavior (Ray 2003) harv error: no target: CITEREFRay2003 (help).

Marxist explanations

Immanuel Wallerstein has argued that it is the global capitalist system that creates shared interests among the dominant parties, thus inhibiting potentially harmful belligerence.

Negri and Hardt take a similar stance, arguing that the intertwined network of interests in the global capitalism leads to the decline of individual nation states, and the rise of a global Empire which has no outside, and no external enemies. As a result, they write, "The era of imperialist, interimperialist, and anti-imperialist wars is over. (...) we have entered the era of minor and internal conflicts. Every imperial war is a civil war, a police action." (Hardt & Negri 2000).

There may be limited consequences of the peace

The peacefulness may have various limitations and qualifiers and may not actually mean very much in the real-world.

Many democratic peace researchers do not count as wars conflicts which do not kill a thousand on the battlefield; thus they exclude for example the bloodless Cod Wars. However, as noted earlier, research has also found a peacefulness between democracies when looking at lesser conflicts.

Democracies were involved in more colonial and imperialistic wars than other states during the 1816-1945 period. On the other hand, this relation disappears if controlling for factors like power and number of colonies. Liberal democracies have less of these wars than other states after 1945. This might be related to changes in the perception of non-European peoples, as embodied in the Universal Declaration of Human Rights (Ravlo & Glieditsch 2000).

Related to this is the human rights violations committed against native people, sometimes by liberal democracies. One response is that many of the worst crimes were committed by nondemocracies, like in the European colonies before the nineteenth century, in King Leopold II of Belgium's privately owned Congo Free State, and in Stalin's Soviet Union. England abolished slavery in British territory in 1833, immediately after the Reform Act 1832 had significantly enlarged the franchise. (Of course, the abolition of the slave trade had been enacted in 1807; and many DPT supporters would deny that England was a liberal democracy in 1833 when examining interstate wars.)

Hermann and Kegley (1995) argue that interventions between democracies are more likely to happen than projected by an expected model. They further argue (1996) that democracies are more likely to intervene in other liberal states than against countries that are non-democracies. Finally, they argue that these interventions between democracies have been increasing over time and that the world can expect more of these interventions in the future. The methodology used has been criticized and more recent studies have found opposing results (Gleditsch, Christiansen & Hegre 2004).

Rummel argues that the continuing increase in democracy worldwide will soon lead to an end to wars and democide, possibly around or even before the middle of this century. The fall of Communism and the increase in the number of democratic states were accompanied by a sudden and dramatic decline in total warfare, interstate wars, ethnic wars, revolutionary wars, and the number of refugees and displaced persons. One report claims that the two main causes of this decline in warfare are the end of the Cold War itself and decolonization; but also claims that the three Kantian factors have contributed materially.

Academic relevance and derived studies

File:DP CHART V19.JPG

Charts arguing in favor of DPT. showing research by R. J. Rummel and others. High resolution PDF

File:DP BACKSIDE V 16.JPG

High Resolution PDF

Democratic peace theory is a well established research field with more than a hundred authors having published articles about it. Several peer-reviewed studies mention in their introduction that most researchers accept the theory as an empirical fact.

Imre Lakatos suggested that what he called a "progressive research program" is better than a "degenerative" one when it can explain the same phenomena as the "degenerative" one, but is also characterized by growth of its research field and the discovery of important novel facts. In contrast, the supporters of the "degenerative" program do not make important new empirical discoveries, but instead mostly apply adjustments to their theory in order to defend it from competitors. Some researchers argue that democratic peace theory is now the "progressive" program in international relations. According to these authors, the theory can explain the empirical phenomena previously explained by the earlier dominant research program, realism in international relations; in addition, the initial statement that democracies do not, or rarely, wage war on one another, has been followed by a rapidly growing literature on novel empirical regularities. (Ray 2003) harv error: no target: CITEREFRay2003 (help), (Chernoff 2004) harv error: no target: CITEREFChernoff2004 (help), (Harrison 2005) harv error: no target: CITEREFHarrison2005 (help). Many of these derived studies have been mentioned above, for example those examining lesser conflicts and minor incidents.

Other examples are several studies finding that democracies are more likely to ally with one another than with other states, forming alliances which are likely to last longer than alliances involving nondemocracies (Ray 2003) harv error: no target: CITEREFRay2003 (help); several studies including (Weart 1998) harv error: no target: CITEREFWeart1998 (help) showing that democracies conduct diplomacy differently and in a more conciliatory way compared to nondemocracies; one study finding that democracies with proportional representation are in general more peaceful regardless of the nature of the other party involved in a relationship (Leblang & Chan 2003) harv error: no target: CITEREFLeblangChan2003 (help); and another study reporting that proportional representation system and decentralized territorial autonomy is positively associated with lasting peace in postconflict societies (Binningsbø 2005) harv error: no target: CITEREFBinningsbø2005 (help).