Democratic peace theory is cherry-picked, and dyadic between two democracies at best.

This finding or ‘law’ has not only been recognized by scholars of international relations, but also found its way outside academia and has influenced foreign policies to promote peace and democracy, most prominently since the 1990s. However, my book will not draw conclusions based on ‘cherry-picking’ of specific studies showing how peaceful democracies are, but on a systematic overview of studies in this field. Therefore, this book relies on my own database with hundreds of different studies, which are relevant for each chapter; the articles had to engage directly with the chapter’s main research question. The next section will provide more detailed information around the selection criteria. This overview includes both highly cited and recent articles which were selected in a systematic way.

Based on analyses of statistical studies around this topic of democracy and war, it will become clear that the overall statistical support for the democratic peace hypothesis is not strong at all. In the rest of the chapter, I will spell out four reasons why democracy does not cause peace, and why the empirical support for the popular idea of democratic peace is quite weak: (1) most studies do not find a strong correlation between democracy and interstate war at the dyadic level, and they show that there are other—more powerful—explanations for war and peace, or even that the impact of democracy is a spurious one, (2) the theoretical foundation of the democratic peace hypothesis is weak, and the causal mechanisms are unclear, (3) democracies are not necessarily more peaceful in general, and the evidence for the democratic peace hypothesis at the monadic level is inconclusive, and (4) the process of democratization is dangerous and living in a democratizing country means living in a less peaceful country.

In my view, it is difficult—if not impossible—to support the democratic peace hypothesis without any reservations. The key caveats should not be ignored and certainly deserve more attention before we can confidently argue that democracies are more peaceful than other types of political systems. Please notice that I can already reveal that the assumed link between democracy and intrastate war is problematic as well, but this topic will be at the core of the next chapter (Chapter 4).

Selection of Articles: Democracy and War

The instrumental value of democracy cannot convincingly be found in democracy’s expected bond with peace. I have come to this conclusion on the basis of an analysis of statistical studies, which will be discussed in the rest of this chapter. So how did I select the articles for my database?4

For this chapter and Chapter 4, I selected the articles that focused on war and democracy. Using the online database Web of Science (formerly known as Web of Knowledge), I identified a total of almost 8000 articles published in the sampled journals until the end of 2015. I identifed them by entering ‘democr*’ in the basic search field; this asterisk (*)-based ‘wildcard’ allows searching for terms including ‘democratic’, ‘democracy’, and ‘democratization’ (in both British and American spellings) simultaneously, in the title, abstract and/or the keywords. As a next step, I excluded articles in which ‘democracy’ is used as synonym for state (e.g. analysis of the relationship between immigration policies and unemployment in European democracies) or a specific political party or movement (e.g. the ‘Democrats’ in the USA, or Uganda’s People’s Democratic Army) or a specifc country (e.g. the Democratic Republic of Congo). In addition, I identified them by entering ‘war*’ in the basic search feld. The words democracy (democr*) and war (war*) need to be mentioned in title and/or abstract—and I also checked for equivalents of ‘war’ like ‘conflict’ and ‘dispute’ and ‘no peace’.5

As it is not feasible to analyse thousands of articles, it is necessary to take a next step in the selection process. I decided to select these articles, which will be part of the database for the third and fourth chapter, in three different ways. The first method is to choose the articles with the most citations. So, for example, in Chapter 3, the articles which are cited more than a hundred times are included in this first list. The article by Beck et al. (1998) has been cited more than 951 times, and as a consequence, this article is part of the database. But also articles with a much lower number of citations (such as Barbieri 1996, with 179 citations) are included in my analyses.

The second method is simply to include the most recent articles published in the past five years, so since beginning 2011 until end 2015. The most recent articles can easily be overlooked by applying the first method of most quoted articles. In my view, however, they still need to be included as they present the most recent findings and engage with the recent and innovative debates, which cannot be ignored in this book. For example, recent studies on democracy and interstate war (Chapter 3) have paid more attention to the mechanisms (see, e.g., Zeigler et al. 2014), and there is a growing attention for the impact of political institutions in recent studies on democracy and intrastate war (Chapter 4; see, e.g., Walter 2015). Those recent findings cannot be ignored in any systematic analysis of statistical studies on this theme.

The third method is the most subjective approach of selecting articles, as it is based on the ‘snowballing method’. So it includes articles which have not been selected by the first and second methods, but which have been quoted extensively and regularly by the previously selected articles. For example, the article by Bethany Lacina (2006) cannot be selected based on having high citations (the first method) and it cannot be included based on being a recent publication (the second method), but it has been mentioned by key studies and hence surfaces via the snowballing method (a third method). This article is important as it clearly distinguishes the determinants of conflict severity from those for conflict onset, and those determinants seem to be quite different, which is crucial information for Chapter 4.

In this way, my study presents and assesses the findings based on a big pool of statistical studies in the published literature. Based on this assessment, I will be able to draw clearer conclusions concerning the significance of the effects of democracy on interstate war (this chapter) and intrastate war (the next chapter). The Appendix shows more detailed information of the selected articles.

The Democratic Peace Hypothesis, Its Roots and Supporters

The democratic peace hypothesis6 states that democracies never or seldom go to war with one another. Where is this powerful idea of ‘democratic peace’ coming from? Before discussing the main findings of the statistical articles and before describing the four caveats of the ‘democratic peace paradigm’, we need to know a bit more around the background and the roots of this idea.

Immanuel Kant’s 1795 essay Perpetual Peace has often been mentioned as the foundation for this hypothesis. Kant believed peace was difficult to achieve, since ‘the natural state is one of war’ (Kant 1795: 10). A state of peace must therefore be established for—in his view—it is certain that hostilities will be committed and people need to be protected from each other. In such a world, each may treat his neighbour, from whom he demands security, as an enemy. In a dictatorship where ‘the subjects are not citizens, a declaration of war is the easiest thing in the world to decide upon, because war does not require of the ruler, who is the proprietor and not a member of the state, the least sacrifice of the pleasures of his table, the chase, his country houses, his court functions, and the like. He may, therefore, resolve on war as on a pleasure party for the most trivial reasons, and with perfect indifference leave the justification which decency requires to the diplomatic corps who are ever ready to provide it’ (Kant 1795: 13).

In contrast, the situation is different in constitutional republics, according to Kant. He argued that the majority of the people in republics would never vote to go to war, except for pure self-defence. Therefore, a world with only republics would be peaceful, since there would be no aggressors. The republican constitution, which requires the consent of the citizens to start a war, gives the positive prospect of perpetual peace.

It is important to note that the ideas of Kant on the one hand and the modern democratic peace scholars on the other hand are not completely similar. For example, Kant talked about republics instead of democratic states as the ideal states to achieve peace. He defined republican states as states with representative governments, in which the legislature is separated from the executive. Not surprisingly—considering the epoch in which he lived—Kant did not include universal suffrage in his definition, which is now seen as an essential dimension of democracy, even of the most minimalist types of democracy (Dahl 1971; see also Chapter 2). Moreover, Kant argued that republics will be at peace in general, which means that such political systems are expected to be not only in peace with each other, but also with other non-republican systems. Nowadays, only few scholars would support this approach of a ‘monadic democratic peace’. As will become clear at the end of this chapter, there is not much evidence for the idea that democracies are more peaceful in general.

Since the 1960s, most statistical studies have not focused on the ‘monadic democratic peace hypothesis’ but on testing the ‘dyadic democratic peace hypothesis’. This dyadic hypothesis states that it is less likely that democracies fight with each other, compared to other ‘dyads’ or other pairs of different types of political systems. The sociologist Dean Babst was the first scholar who started to build on Kant’s old idea in the ‘dyadic’ way, and decided to test it in statistical studies (Babst 1964, 1972). He concluded that ‘no wars have been fought between independent nations with elective governments between 1789 and 1941’ (Babst 1972: 55). His study was not published in one of the journals in the field of international relations, but in a sociological journal and later in Industrial Research. Therefore, it was not read by international relations scholars, and initially, it did not get the attention it deserved in the field of international politics.

Babst’s work was, for example, not cited by Melvin Small and J. David Singer (1976), and their fndings seemed to contradict Babst’s study. However, Small and Singer did not compare the rates of war proneness for democracies and dictatorships, but instead they focused on the question whether wars involving democratic states have historically been significantly different in length or in degree of violence compared to wars involving only dictatorships. For length and degree of violence during the wars, they did not find a difference between democracies and dictatorships, so they concluded that types of political systems did not matter.7 A few years later, Rudolph J. Rummel did cite Babst’s work and replicated Babst’s idea in statistical tests, which were described in the fourth book of his five-volume Understanding Confict and War (1975– 1981). He found clear support for his eleventh (of the 33) propositions about causes and conditions of conflict, which stated that ‘Libertarian systems mutually preclude violence’ (Rummel 1979: 279).

Eventually, those innovative studies from the 1970s helped to evoke the interest in the democratic peace proposition, and in the expected peaceful nature of relationships among democratic states. Since the 1980s, the number of quantitative studies has increased considerably, accumulating into an impressive field of research in international relations with its own ‘empirical law’ of democratic peace (Levy 1989: 270; see also Ray 1998).

This democratic peace hypothesis has not only received support from political scientists, but also from politicians and policy makers. Particularly since 1993, the idea of a democratic peace has inspired American foreign policies aimed at the promotion of peace and democracy. As the 42nd president of the USA (1993–2001), Bill Clinton was the first politician who explicitly bridged the gap between these findings in international relations on the one hand, and his foreign policy strategy on the other hand, at least rhetorically. Anthony Lake, who was Clinton’s National Security Adviser, stated in 1993 that in order to cope with America’s foreign policy challenges, the expansion of democratic states around the world would be essential because ‘it protects our [U.S.] interests and security’ (see Henderson 2002: 20). In his 1994 State of the Union, Clinton declared that ‘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’.8 Findings from research in the field of international relations seemed to have a direct impact on policy making, and this move of the Clinton administration can be seen as ‘a textbook case of arbitrage between the ivory tower and the real world’ (Gowa 1999: 109).

Clinton’s successor, George W. Bush, went one big step further in his faith that democratic peace holds. He argued that efforts to turn Iraq into a democratic country would have positive effects on Iraqi’s neighbours. The authoritarian regimes in the region would fall as domino stones and follow the Iraqi example. They would start democratizing as soon as they could, which would then result in achieving a peaceful and stable the Middle East. The real motives for attacking Iraq may have been different, but ‘regime change’ was at the heart of Washington’s rhetoric when the USA started to bomb Baghdad in March 2003. The rhetoric of the Bush administration focused on toppling Saddam Hussein’s regime, and replacing the entire underlying dictatorial system with a democracy.

Moreover, George W. Bush used the democratic peace idea to justify the war in Iraq, declaring, ‘The reason why I’m so strong on democracy is democracies don’t go to war with each other…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’.9 In 2004, the 43rd President of the USA said: ‘If you think you can have peace without democracy – again - I think you’ll find that - I can only speak for myself, that I will be extremely doubtful that it will ever happen’.10 In his second inaugural address, he stated that ‘the survival of liberty in our land increasingly depends on the success of liberty in other lands. The best hope for peace in our world is the expansion of freedom in all the world’.

Again, these are just words from speeches and can hence be seen as rhetoric to defend military intervention (cf. Jervis 2003; Kaufmann 2004; Daalder and Lindsay 2005; Owen 2005; Lieberfeld 2005; Schmidt and Williams 2008). Still, in the end, politicians have rationalized their political decisions based on one of the most powerful ideas taken from studies in the field of international relations, clearly showing the influence of this academic idea in political practice.

Hence, the field of international relations seems to have its own law: democracies rarely fight with each other. It cannot be denied that the evidence supporting the democratic peace proposition is quite diverse in character (see Ray 1998): the evidence has been epistemological (Rummel 1975), philosophical (Doyle 1986), formal (Bueno de Mesquita and Lalman 1992), historical (Weart 1994; Ray 1995; Owen 1994), experimental (Mintz and Geva 1993), anthropological (Ember et al. 1992; Crawford 1994), psychological (Kegley and Hermann 1995), economic (Brawley 1993; Weede 1996) and political (Gaubatz 1991). Still, there have been numerous critical studies (see, i.e., Hayes 2011), and the general picture is unclear. We do not yet know much about the overall findings from statistical studies.

So far, it seems as if some quantitative studies—mainly within the field of international relations—have found strong and robust evidence which supports the ‘democratic peace hypothesis’. Those studies show that democracy has had a positive influence on international peace (see, i.e., Rummel 1979; Ray and Russett 1996). In this sense, the idea of a democratic peace seems to be confirmed. Political scientists such as James Lee Ray are passionate supporters: ‘No scientific evidence is entirely definitive’ but ‘based on all the empirical evidence so far’ the more defensible of the two possible definitive answers to the question “Does democracy cause peace?” is “Yes” (Ray 1998: 43). However, based on my own analyses of the empirical studies with statistical evidence, I cannot be as enthusiastic as those scholars; to the contrary, I whole-heartily disagree with them, as a more systematic analysis of the articles shows that there are four important weaknesses, which seriously undermines the idea that peace is one of democracy’s instrumental values.

Caveat 1: It’s Not (Just) Democracy

While analysing the selected articles, the first remarkable finding is that only a relatively small number of studies have actually tested the democratic peace hypothesis. Most of the studies have focused on the mechanisms (see next section, caveat 2), and hence seem to assume that there is a correlation between democracy and war. In this way, the majority of the studies—often unintentionally—reinforce the idea that democratic peace actually exists without testing this proposition. However, none of the studies that directly test the democratic peace hypothesis found strong evidence that democracy is the most important factor when explaining interstate war. All democratic peace studies have controlled for many possible alternative causes of the peace, such as economic development and growth, geographic distance and contiguity, power status, alliance ties, militarization and political stability. The findings show that it is not just democracy which explains war, not at all. Within this group of studies, which explicitly test the democratic peace hypothesis, four different types of findings can be detected. I will discuss those results more in-depth in the rest of this section.

First Result: There Is Correlation, but Other Explanations Are Significant Too

The first subgroup consists of scholars who stress the importance of democratic peace, despite the fact their own analyses have shown that other factors are statistically significant as well (Maoz and Russett 1993; Rousseau et al. 1996; Gleditsch and Hegre 1997; Beck et al. 1998; Ray 2013). For example, some studies (e.g. Rousseau et al. 1996) included alternative independent variables in order to test realist arguments. They tested whether the distribution of power determines decisions to use force, and measures each state’s military capabilities relative to its opponent. A state’s military capability is the average of three elements: number of troops, military expenditures and military expenditures per soldier. They found that this realist variable was strong, positive and statistically significant at the 0.001 level in their analyses (see, e.g., Rousseau et al. 1996: 522, Table 2). However, not only a state’s military capabilities appeared to be an important explanation for peace. In addition, wealth, growth, alliances and contiguity played a crucial role when explaining interstate war (see, e.g., Maoz and Russett 1993: 632, Table 1).11 Moreover, when other factors are included, the impact of democracy on the likelihood of international crises is even spurious (Maoz and Russett 1993: 632; Henderson 2002: 141, see also p. 3).12 Still, scholars in this group keep defending the democratic peace idea, despite the fact that their own analyses showed the significance of alternative explanations.

Second Result: Initially There Is Correlation, but the Impact of Democracy Is Spurious When Other Explanatory Factors Are Included in the Models

The second subgroup of scholars is far more radical. Based on their own analyses, this group concludes that the democratic peace link is a spurious one (Weede 1984, 1996; Barbieri 1996; Mousseau 2013; Gartzke and Weisiger 2014).13 Typically, efforts to demonstrate the spuriousness of the statistical democratic peace pointed to other factors that, when accounted for ‘properly’, eliminated or dramatically reduced the statistical significance of shared democracy. Hence, the studies in this second group did not find strong evidence for the democratic peace hypothesis anymore, once other explanatory factors were included in the models.14

One of the most convincing alternative explanations of peace between countries is that there is no democratic peace, but a capitalist peace instead. The settlement in Germany and Japan succeeded because of the establishment of capitalist peace. Because of economic support by the Americans, who encouraged free trade and offered trade opportunities in practice as well, the poorer economies in Europe and Japan would gain economically, resulting in ‘economic growth, prosperity, and, ultimately, free trade among most of the more technologically advanced economies’ (Rasler and Thompson 2005: 232). By establishing and expanding free trade, the incentives for war would quickly decrease among trading states, according to this approach. To prevent new interstate wars after World War II, the capitalist peace was a far more important factor than the American promotion of democracy and its political institutions.

The capitalist peace, or capitalist peace theory, also states that economic development accounts for both democracy and the peace among democratic nations. Economic development is a key factor to explain democracy (Lipset 1959; see also Hegre 2003; Weede 2004).15 Moreover, economic development also plays a role when explaining peace, and the presence of market-oriented economies in countries have a positive impact on both democracy in those countries and peace between them (Mousseau 2000, 2002, 2003, 2005, 2013; see also Hegre 2014). Democratic peace only exists when both democracies have high levels of economic development, when economic development is well above the global median.

In fact, the poorest 21% of the democracies studied, and the poorest 4–5% of current democracies, are significantly more likely than other kinds of political systems to fight each other (see, e.g., Mousseau 2005). Moreover, if at least one of the democracies involved has a very low level of economic development, then democracy cannot prevent war.16 Still, there is a pacifying effect of free trade and economic interdependence, which is more important than the effect of democracy, because the former affects peace both directly and indirectly, by producing economic development and ultimately, democracy (see Weede 2004).17

Capitalist peace is not the only alternative explanation. Shared interests in general, and political similarities in specific, can also be seen as an important second alternative explanation for war and peace between countries (Farber and Gowa 1995, 1997; Gartzke 2007; Gowa 1999; Henderson 2002). Democracies are not peaceful to each other because they are democratic, but rather because they are similar. So the difference of the scores of both countries also contributes to the conflict proneness of the dyad. If the difference in levels of democracy is big, then the chance of conflict is higher (cf. Oneal and Russett 1997: 281–282).

Many researchers have conflated both the conflict-dampening impact of joined democracy and the confict-exacerbating impact of political distance in the variables focusing on political systems, but as Errol A. Henderson (2002: 32) convincingly argued: ‘Fusing these two contrasting attributes in a single variable makes it difficult to distinguish between the competing processes’. Therefore, it is better to include an additional variable of ‘political dissimilarity’ in the model. Henderson (2002) was one of the first scholars who included this variable and measured it by taking the absolute value of the difference between the two states’ scores. His main variables were not only political similarity, but also geographic distance and economic interdependence, and he concluded that democratic peace is a statistical artefact which disappears when those other variables are taken into account. Political similarity clearly has a pacifying effect18 (see Werner 2000; Henderson 2002; Beck et al. 2004), and it is not democracy per se which is the decisive factor.19

Hence, the benefits of trade and trade interdependence are essential explanations, while democracy is spurious or at least subordinate (see also Rosecrance 1986; Weede 1984, 1996; Hegre 2000, 2014; Jervis 2002; Souva 2003; Rasler and Thompson 2005: 235; Mousseau 2000, 2002, 2003, 2005). Based on those studies, it is safe to conclude that democracy, on its own, is an unlikely cause of the democratic peace.

Third Result: There Is Correlation, but Other Explanations Are Much Stronger

This same point that democracy is just one of the explanations for peace (and not even a very important one) is also at the core of studies in the third subgroup. Scholars of this group keep arguing that there is support for the democratic peace hypothesis, and that the link is not spurious. In this sense, they are less radical than the second group of scholars, as they do not completely reject the value of democracy for peace. On the other hand, their own analyses have clearly shown that alternative factors—hence other factors than democracy or type of political system— are not only statistically significant but also more important when trying to explain interstate war (Bremer 1992; Gelpi 1997; Oneal and Russett 1999a, b; Reiter and Stam 2002; Peterson 2013; Caselli et al. 2015).

Theoretical arguments and empirical evidence suggest that democracy is not the most important factor, while war is more likely to occur between states that are geographically proximate, approximately equal in power, major powers, allied, economically advanced and highly militarized than between those that are not. Bivariate analyses of these factors in relation to the onset of interstate war over all pairs of states in the period from 1816 to 1965 have generally supported these associations. However, multivariate analyses revealed some differences. Stuart Bremer (1992), for example, showed that some factors are far more important than others. The existence of a dangerous, war-prone dyad can be best explained by the presence of contiguity, the absence of an alliance and the absence of more advanced economy. The absence of democratic polity and other factors (absence of overwhelming preponderance, and presence of major power) are less powerful. Overall, these findings suggest that our research priorities may be seriously distorted and that we should not focus too much on the perceived positive impact of democracy, but on other factors (such as alliances and economic factors) instead.

Fourth Result: There Is Correlation, but Only Under Certain Specific Conditions

The final subgroup of scholars argues that we cannot unconditionally accept the idea that democratic peace exists in general, so always and everywhere. Their statistical studies clearly showed that support for this hypothesis heavily depends on other factors. The chance of democratic peace depends not just on the specific historical period (Cold War or not; Gibler and Sarkees 2004; Siverson and Emmons 1991; Weede 1984), but also the stage of the conflict (beginning, duration or severity; see Bremer 1993; Bennett and Stam 1996; Reed 2000), and on the neighbourhood instability (extent of confict in the region; see Gibler and Braithwaite 2013; Gibler and Miller 2013). Despite the differences between the studies, there is one common finding in all studies: when explaining interstate war, we cannot just rely on the impact of democracy, as it is too much dependent on other factors.

Several scholars found strong evidence for the idea that democratic peace exists, but only during some specific historical periods. Based on this evidence, they concluded that democratic peace is simply a statistical artefact of the Cold War. For example, Henry Farber and Joanne Gowa (1995) found statistical support for the idea that peace between democracies is an artefact of the Cold War, when the threat from the communist states forced democracies to ally with one another (see also Mearsheimer 1990). Sebastian Rosato (2003) also argued 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.

Since the Second World War, war has become a very costly affair. Scholars discovered that only a handful of states are ‘capable of engaging in major power warfare. That process of elimination has not yet extinguished the possibility of major power warfare, but it has lowered its probability immensely’ (Rasler and Thompson 2005: 219). The chance to achieve something in a war is low in general, and even lower in a bipolar world with two big power players risking high nuclear war costs (Jervis 2002). While war became more costly, trade became less costly; as a consequence, the war/trade costs increased during the Cold War (Rosecrance 1986; see also Jervis 2002). In such a world, war and conflict have become less attractive, while trade and cooperation have become more appealing (Rasler and Thompson 2005: 219). Hence, more states decided to adopt trading strategies in order to prevent confict and war as much as possible. In the end, democracy was part of the story, but only a very small part with a subordinated role next to the power dynamics during the Cold War, the costs of warfare and the benefits of trade.

Some scholars found evidence that the democratic peace still exists in the post-Cold War period (Park 2013) which weakens this argument. However, most analyses showed that dyadic dispute rates have converged after the Cold War (see, e.g., Gowa 2011). Moreover, jointly democratic dyads are likely to be allied only after 1945 (see Gibler and Sarkees 2004); during the 1816–1944 time period, there is even a negative relationship between democratic dyads and alliance formation.20 These findings cast serious doubts on the idea of a general existence of democratic peace.

Not only the historical period, but also the stage of the conflict is crucial. Some scholars in this group provided evidence that democratic peace is not universal, but that it depends on the stage and whether we focus on the beginning, duration or severity of the conflict. Although joint democracy has some pacifying effects on the onset of conflict, the results suggest that they are unrelated to the escalation of disputes to war (see Reed 2000). Moreover, democratic peace is dependent on the neighbourhood instability. Democracies often have few territorial issues over which to contend, as they tend to be part of a stable region. Democracies only seldom have territorial disputes with their neighbours, and therefore they can more easily choose favourable conflicts to escalate. The type of political system does not predict conflict selection or victory once controls are added for issue salience (Gibler and Miller 2013; see also Park and James 2015). There is an interaction between joint democracy and regional instability, which confirms the idea that the effects of type of political system on continued conflict apply mostly to dyads in peaceful regions (Gibler and Braithwaite 2013; see also Park and James 2015). Very democratic countries might even become more aggressive and faster than other political systems, once the region becomes more hostile (see, e.g., Baliga et al. 2011).

The General Lesson from the Results in a Nutshell (Caveat 1)

In short, regardless of the differences between the statistical studies on democratic peace, all findings have indicated that other explanations are important as well. It is clear that democracy is just one of the explanations, and certainly not the most important one,21 sometimes even spurious and often heavily dependent on other factors. It is not (just) democracy to be preoccupied with, when trying to prevent war between countries (Table 3.1).

Caveat 2: What Are the Causal Mechanisms?

Most of the statistical studies on democratic peace seem to assume that there is a correlation between democracy and war; based on this assumption, they then decide to focus on the mechanisms. This is problematic as none of the democratic peace studies found strong evidence that democracy is the most important factor when explaining interstate war (see the previous section). As a consequence, the next step of looking for mechanisms is quite irrelevant and not necessary in my view, but most studies nevertheless argue that the field lacks strong theoretical foundations and robust empirical evidence that can reveal convincing causal mechanisms.22 Those studies seem to accept the correlation between dyadic democracy and peace, and then start questioning whether democracy really causes peace before investigating potential mechanisms.

Table 3.1 Statistical studies on democracy and interstate war (dyadic level)