Skip to Main Content
AdvertisementNavbar Search Filter Mobile Microsite Search Term Search
Close
search filter search input Search
Advanced Search
Search Menu
Skip Nav Destination


Article Navigation
Volume 48
Issue 3
September 2004
Article Contents
Democratic Synergy and Victory in War, 1816–1992
Ajin Choi
Ajin Choi
Yonsei University
Search for other works by this author on:
Oxford Academic
Google Scholar
International Studies Quarterly, Volume 48, Issue 3, September 2004, Pages 663–682, https://doi.org/10.1111/j.0020-8833.2004.00319.x
Published:
29 July 2004Navbar Search Filter Mobile Microsite Search Term SearchClose
search filter search input Search
Advanced Search
Search Menu
Abstract
This study investigates the question of why democracies are more likely to win wars than non-democracies. I argue that due to the transparency of the polities, and the stability of their preferences, once determined, democracies are better able to cooperate with their partners in the conduct of wars, and thereby are more likely to win wars. In support of my argument, the main findings in this study show that, other things being equal, the larger the number of democratic partners a state has, the more likely it is to win; moreover, democratic states are more likely to have democratic partners during wars. These results are in contrast with those in current literature about the high likelihood of prevailing by democracies in wars, which emphasize, on the one hand, the superior capacity of democratic states to strengthen military capabilities and, on the other hand, to select wars in which they have a high chance of winning.
The influence of democracy on world politics and foreign policy choices is one of the most intensively studied questions in the field of International Relations. Traditionally, democracies have been regarded as inefficient actors in the international arena due to the incompetence of their mass public and the decentralized nature of their institutions (Tocqueville, [1835] 1945; Morgenthau, 1960, 1985; Kennan, 1977).
In contrast to these expectations, however, democracies have proven themselves to be powerful and successful actors in the international system. This has been demonstrated by the accumulation of tremendously important political and military victories that were won by the U.S. and its democratic allies during the last century. In the first half of that century, the advanced democratic states defeated military challenges by such aggressive central powers as Germany and Austria–Hungary in World War I, and they won military victories against such potent expansionist enemies as Germany and Japan in World War II. In the second half of the 20th century, the U.S., with support of a number of advanced democracies, won the Cold War by outlasting and effectively defeating the Soviet Union and its satellites.
A systematic positive relationship between the democratic political system and state war performance has also been recently reported by one line of what below is termed the democratic efficacy literature (Lake, 1992; Reiter and Stam, 1998, 2002; Bueno de Mesquita, Morrow, Siverson, and Smith, 1999; Gelpi and Griesdorf, 2001). Based on a variety of statistical tests, and controlling for the impact of a number of other potentially important factors, this line of literature has demonstrated that, compared to non-democracies, democracies enjoy a disproportionately high likelihood of prevailing in interstate crises and wars.1
Although these recent empirical and theoretical findings agree that democratic states are formidable and competitive actors in military affairs, the arguments in this literature have only focused on the internal efficacy of individual democratic states in explaining their victories in military conflict. In a path-breaking article, David Lake argues that democratic states win wars because they earn fewer domestic rents, tend to create fewer domestic economic distortions, and therefore possess greater national wealth and devote more resources to security. Furthermore, democracies are able to enjoy greater societal support for their policies, and therefore have a greater capacity to extract resources from society to wage a war. This argument emphasizes the capacity of democratic states to strengthen their own military capabilities in the conduct of wars (Lake, 1992). In a competing argument, Dan Reiter and Allan Stam suggest that democracies are more likely to achieve victories in wars because they initiate wars in which they have high chances of winning. Democratic leaders are more likely to lose power after defeat since the electorate punishes their leaders' policy failure at the polls. This mechanism provides an incentive for democratic leaders to make careful decisions about what adversaries to pick a fight with when they start wars (Reiter and Stam, 1998, 2002; Bueno de Mesquita et al., 1999; Gelpi and Grisdorf, 2001).
Indeed, these explanations capture important features of democratic political systems and cannot be dismissed cavalierly. Nonetheless, they cannot be considered as conclusive. The current explanations have paid scant attention to the dynamic developments that occur inside the conflicts, and therefore to the possibility of superior diplomatic efficacy on the part of democratic states during those conflicts (Clausewitz, [1832] 1976; Blainly, 1988; Goemans, 2000; Wagner, 2000; Werner, 2000; Powell, 2001). Moreover, these two explanations have overlooked a key empirical pattern: democratic states tend to be the targets of military challenges and to fight in multilateral wars (Lake, 1992; Rasler and Thompson, 1999; Gelpi and Grieco, 2001; Grieco, 2001).
In this study, I shift the theoretical and empirical focus on democracies' diplomatic and military behavior during war, and argue that it is their superior performance in wartime cooperation that gives democratic states a significant advantage in wars. In a nutshell, due to the transparency of their political systems, and the stability of their preferences, democratic states are better able to cooperate with their partners in the conduct of wars, and thereby are more likely to win wars. In support of my argument, the main findings in this study show that, all else being equal, the larger the number of democratic partners a state has, the more likely it is to win a war; moreover, democratic states are more likely than non-democratic states to have democratic partners during wars. These findings not only challenge the two current explanations, but also suggest different foreign policy implications and broaden the research program of the democratic efficacy school by integrating the impact of democracy on military conflict with its impact on international cooperation.2
The analysis below proceeds in three steps. First, I elaborate the theoretical framework for this study. In particular, I introduce three main hypotheses elaborating the causal mechanisms of my own argument about democracies as cooperative partners and thereby winners. Then, I present two alternative arguments about democracies' victory, namely that of Lake and that of Reiter and Stam, and discuss their weaknesses. Second, I report the results of my empirical analysis and assess the relative explanatory power of the three competing arguments about democracies' victory. Third, and finally, I summarize the key findings and discuss their implications for international relations theory.
Determinants of Democratic Victory
Democracy has been identified as one of the major determinants of victory in war. However, the key question of locating the precise mechanism through which democratic states are particularly advantaged in the prosecution of wars remains inconclusively answered.
In what follows, I draw three new hypotheses to explain the causes of democracies' victory connecting their stable preferences and transparent political system to their cooperative behavior with partners, and test them against the two current alternative explanations: one connects democracies' strong military capabilities to their domestic support; and the other connects democracies' prudent initiation of war to their leaders' electoral accountability.
Main Hypotheses: Democracy and International Partnerships
States form wartime coalitions to increase their capabilities in the face of external threats because it is often unlikely that one state alone can make the “great and sudden shifts” in capabilities necessary to prevail in a war effort. In this sense the decision of third parties to join or not to join a war coalition can lead to different war outcomes from those the war participants had initially anticipated (Gartner and Siverson, 1996; Smith, 1996; Werner, 2000).
However, there can be serious uncertainty about the behavior of allies and the reliability of their ties with one another in the conduct of wars (Olson and Zeckhauser, 1966; Stam, 1996).
That is to say that there can be variation in the levels of cooperation among wartime allies and that having a partner that is reliable as well as powerful is a key element in winning a war. Even if a state obtains the same amount of military capabilities from its partners during a war, war outcomes could be different depending upon the amount and depth of cooperation among wartime allies. In other words, I argue that democratic states are better able to cooperate with their partners in the conduct of military conflicts and therefore more likely to win in those conflicts.
Why could it be the case that democratic states cooperate more effectively with their wartime partners and thereby lead the war to win? The causal process of the relationship among democracy, international partnerships and victory is based on two dimensions of wartime cooperation: (1) unity of cobelligerents and (2) coordination in the war effort.
First, a strong unity of cobelligerent states is important to lead them to win the war. As Sun (1971) points out, one of the main strategies pursued by adversaries is, therefore, to make the other side disunite. Even though all partners in a wartime coalition share the fundamental common interest of winning the war, each state has its own conflicting or divergent interests because it is tempted to minimize its own costs and risks. Therefore, a state may give up its commitment to its partners by leaving the war before it ends and thus may lead the coalition to dissolve and increase the risk of losing the war. Historically, attempts for separate peace have been ubiquitous. During World War I, for example, as the stalemate of the war continued into 1917, most of the war-participant states began to consider the possibility of separate peace negotiations. Germany proposed separate peace negotiations to Britain at least three times, and Austria–Hungary tried to undertake a separate peace with the Entente Powers at the expense of its alliance with Germany. Furthermore, Russia withdrew from the war, and Romania left, reentered, and left again in the midst of World War I (Stevenson, 1988).
Democratic states are better able than non-democratic states to maintain international commitments during wars because democratic institutions, in particular, are comprised of effective veto players in decision-making processes and, after a time, produce highly stable domestic preferences (Tsebelis, 1995, 1999; Martin, 1999). It may be harder or slower for democratic states to enter into international commitments because of the competing preferences of diverse interests. It may be even harder to gain support if international commitments involve the use of force. For example, just before the breakout of World War II, Britain had signaled its potential allies and adversaries that it did not want to be entangled in bloody continental war once again. As this war developed, U.S. President Franklin Roosevelt began to recognize the need to participate in this war. U.S. participation was, however, delayed until the country was actually attacked because the president was constrained by the isolationist public and political groups (Kissinger, 1994).
However, once democratic states have made decisions and these decisions have been ratified through democratic institutional mechanisms, then it is also harder to reverse them because of the effective institutional constraints that operate in democracies on preference changes of individual branches or leaders. Therefore, if they decide to make an international commitment and do so, it is unlikely that democratic states will change course (Fearon, 1994; Gaubatz, 1996; Schultz, 1998).3
In contrast, when non-democratic states make decisions, they are less constrained by either home publics or domestic institutions, and their decisions are usually made by a single leader or based on a smaller number of group preferences. Therefore, non-democratic states may respond quickly to allies' requests, but for the same reason, their ability to maintain commitments may be rather unreliable and unpredictable. In other words, their promises are “not enforceable by an independent judiciary or any other independent source of power…. Because of this and the obvious possibility that any dictator could, because of an insecure hold on power or the absence of an heir, take a short-term view, the promises of an autocrat are never completely credible” (Olson, 1993:571).
Specifically, even if the top leader or another branch of government in a democracy wanted to change the status quo, they might see their efforts effectively blocked by other members in the government. In contrast, veto players are absent and ineffective in a non-democratic political system, and thus the preferences of the top leader are less likely to be constrained by other members. For example, when Germany proposed peace negotiation to Britain in September 1917, David Lloyd George, the British Prime Minister, expressed his interest in this peace proposal despite the costs of some of his cobelligerents' interests, but his preference was faced with strong opposition from the War Cabinet members. The Prime Minister had to take their voice into account to guarantee the survival of his government and continued to fight until Germany was defeated (Woodward, 1971). In contrast, Austria–Hungary's Emperor Karl could initiate a separate peace with the Entente Powers despite that most of the major members in the government strongly opposed the break up of the alliance with Germany during World War I (Shanafelt, 1985).
The second key factor in the conduct of a coalition war is the ability of member states to coordinate with each other to allocate their efforts and resources on the critical place and time to undertake a military operation effectively. The importance of such concentration of war efforts in the major theaters of operations was emphasized by Carl von Clausewitz:
There is no higher and simpler law of strategy than that of keeping one's force concentrated…. Since obviously greater force is more likely to lead to success, it naturally follows that we can never use too great a force, and further, that all available force must be used simultaneously. The simultaneous use of all means intended for a given action appears as an elementary law of war ([1832] 1976:194–207 emphasis added).
The lack of coordination for the concentration of efforts and resources among the Axis Powers during World War II, and among Arab countries during prolonged conflicts against Israel, are two well-known historical counterexamples. In contrast, although it was initially attacked by Japan, the U.S. chose the “Germany First” strategy and successfully concentrated its major efforts and resources with those of its wartime partners in the European theaters in World War II.4
Openness and transparency in a democratic political system can reinforce cooperation with other states because an effective way to achieve cooperation among states is to increase “the quantity and quality of communication” so as to reduce uncertainty and find it relatively easy to monitor and, if necessary, to punish one another's derogations (Keohane, 1983; Cowhey, 1993). A state needs “not merely information about the other government's resources and formal negotiating positions, but rather knowledge of their internal evaluations of the situation, their intentions, the intensity of their preferences, and their willingness to adhere to an agreement even in adverse future circumstances.” As a result, “‘closed’ governments will be viewed with more skepticism by potential partners, who will anticipate more serious problems of bounded rationality in relations with these closed governments than toward their more open counterparts” (Keohane, 1983:162–163). This suggests that successful cooperation among states depends not only on interests and power, but also on the level of shared information and expectations.
By encouraging communication and reducing uncertainty, an open and transparent political system in democratic states can facilitate cooperation among coalition member states to concentrate their efforts and resources in the conduct of wars. In contrast, due to the lack of openness and transparency, non-democratic states cannot provide a mechanism to increase confidence that they will remain within the terms of and operate based on a mutual agreement.
For example, although the Grand Alliance partners agreed to cooperate on the employment of air power and the exchange of related military technologies against Germany during World War II, the lack of openness and transparency in the Soviet political system contributed to limited cooperation, thus reducing the effectiveness in the conduct of war. On the one hand, Russia's fear of political contamination, which it believed might come about by allowing its allies to establish air bases in Russian territory prevented its allies from implementing “Velvet,” a joint air force operation in 1942 to destroy Germany's oil route in the eastern front. On the other hand, since Britain and the U.S. were not confident that Russia was doing what it said it would do, and at the same time, they were not sure that Russia would not do what it should not do, the Grand Alliance partners failed to conclude the agreement on the exchange of military technologies with Russia during that war (Lukas, 1970; Ulam, 1974; Beardsley, 1977).
In sum, I argue that these distinctive theoretical and empirical elements of democracies can make significant differences in wartime cooperation and lead their coalition to win in war.5 Furthermore, due to these elements, democratic states are likely to have more democratic partners. Therefore, I derive the three following hypotheses:
Alternative Explanations
Democracy and Military Capabilities: “War-Fighting” Model. It is far from surprising that military capabilities have a major impact on war outcomes.
Organski and Kugler (1980:4) argue that, “power is the most important determinant of whether a war will be won or lost.”Waltz (1979:118 and 168) also points out that the internal effort to increase their own capability is a “more reliable and precise” way to survive in the international system. A number of empirical studies on this subject consistently have shown that it is difficult to overcome an opponent's advantage in military capabilities during wars (Wayman, Singer, and Goertz, 1983; Rosen, 1992).
What is more important, however, is that different political systems have different abilities to extract or mobilize resources from society for waging a war (Tilly, 1975; Organski and Kugler, 1980; Rasler and Thompson, 1985; Kugler and Arbetman, 1997). Recent studies have suggested that authoritarian regimes are more interested in extracting rents, and thus are more likely to distort their economies, while democratic regimes are constrained in seeking rents and thus are more likely to produce stronger national economies that are able to generate more resources needed for national security. Lake (1992) especially identifies this factor as the most important determinant to explain the superior performance of democracies in war.
Moreover, since democratic states “respect and effectively represent morally autonomous individuals,” they are less repressive and more legitimate (Doyle, 1983:325). Therefore, democratic governments can attract greater social support and mobilize more resources from society to wage their wars. Thus, democratic states are better able to strengthen their military capabilities, therefore more likely to win wars.
Democracy and War Initiation: “Selection-Effect” Model. There is a persistent finding that states that initiate wars are more likely to be the victors at the end. Bueno de Mesquita (1981:21–22) notes that “the initiators of most interstate wars during the past 160 years do appear to have enjoyed an advantage…the nations initiating combat won forty two times [among 58 interstate wars].” This finding is based on the idea that the decision to go to war is prudently calculated by the leader, and therefore is “not accidental or unintentional.”Gartner and Siverson (1996) support this finding by showing that the states that initiate wars are inclined to choose as the targets those states that they believe they can defeat by themselves and that will not be joined by war-time coalition partners.6 There are also findings that initiators are advantaged in wars since they can display a high level of motivation and use effective strategies regardless of whether they are militarily stronger or weaker than the targets (Maoz, 1983; Mearsheimer, 1983).
All states, then, can benefit from the motivational and logistical advantage that fighting as initiators provides. Retier and Stam, however, argue that this is even more so for democracies because democratic states are more cautious in making foreign policy choices of going to war. Two main mechanisms underlie their argument: electoral punishment and information.7
First, democratic states may be more cautious than non-democratic states in choosing or avoiding adversaries because they know they must compete with opposition rivals who want to take office through open and competitive elections, and they know they will be evaluated at the polls by the people who bear most of the costs of their policy. Bueno de Mesquita, Siverson, and Woller (1992) have shown that war involvement and its outcomes induce violent leadership changes by punishing the failure of its foreign policy. In a later study, Bueno de Mesquita and Siverson (1995) have found that defeats in wars are more likely to have such effects on a democratic regime, thus more readily replacing its leaders with a new one as a result of their policy failure. These findings imply that democratic leaders must be more sensitive to outcomes, and therefore, more careful when evaluating a decision to go to war. In other words, if democratic states do go to war, and especially if they initiate a war, such states are more likely to have a high chance of winning (Reiter and Stam, 1998, 2002; Bueno de Mesquita et al., 1999).
The second mechanism points to the fact that democratic states may have better information than non-democratic states. It is because democracy as a form of government encourages openness of information and competition among ideas (Feaver, 1996; Schultz, 1998; Reiter and Stam, 2002). Without competition and openness of ideas, a policy, which only reflects the narrow interests of a certain group would prevail. Snyder (1984, 1991) particularly points out that the German elite's bias for an offensive strategy prior to World War I was generated from the parochial organizational interests of their professional military.
Limitations of the Alternative Explanations. It can hardly be denied that the two dominant arguments about democracies' war effectiveness identify important explanatory elements. Nonetheless, they can only be considered as partial answers to the empirical puzzle of democracies' ability to prevail in military conflicts. Their most serious flaw is that both explanations underestimate the dynamic interactions among war participants that take place during wars. What happens during war is indeed crucial for the war outcome. As Clausewitz put it, “war is simply a continuation of political intercourse, with the addition of other means… in itself does not suspend political intercourse” ([1832] 1976: 605). This indicated that political bargaining among states does not end even when fighting begins, and war ends because the bargaining while fighting reveals information to reach negotiated settlement (Goemans, 2000; Wagner, 2000; Powell, 2001). Therefore, the neglect of the processes inside the war can be problematic: as Wagner (2000:497) puts it, “this way of posing the problem can only lead to misleading conclusions.”
The war-fighting model, however, ignores the importance of developments that take place inside the war. Not only does political bargaining continue while fighting, but also a third party's incentive to join or leave the war largely depends on what happens battle after battle (Sun, 1971; Blainly, 1988; Werner, 2000). Therefore, war outcomes can be determined by states' diplomatic skills and cohesiveness with their cobelligerents during wars as well as their own capacity to build up military capabilities. Morgenthau (1985:141) points out that, “a competent diplomacy can increase the power of a nation what one would expect it to be in view of all the other factors combined.” Given the fact that a third party's decision inside the war significantly matters in determining war outcomes, the war fighting argument, which solely emphasizes the capacity of democratic states to strengthen their own capabilities in the face of external threats, can only provide an incomplete answer for democracies' victory.
The selection-effect model, on the other hand, attempts to explain democracies' victory by assuming that war outcomes can be entirely determined by the choices made by leaders before fighting actually begins. This argument is problematic insofar as it does not fully account for the uncertainties that policy makers face on the eve of war. There are, at least, three kinds of uncertainty which could affect leaders' calculations: uncertainty about the feasibility of their own capabilities, uncertainty about the adversary's capabilities and intentions, and uncertainty about the reactions from third parties. Moreover, as Fearon (1995:390–401) points out, states have a strategic incentive to misrepresent their intentions and capabilities in the bargaining stage before fighting starts. Given the fact that both uncertainties and strategic misrepresentations are inevitable, the probability of winning calculated by even “a very cautious” democratic leader will be updated, or entirely defined, as a war progresses.
Moreover, the two current arguments do not take into account the nature of wars where democracies go to fight when they explain the causes of democracies' victory in wars. First, democratic states have disproportionately participated and emerged as winners in multilateral wars: democratic states participated in eight bilateral wars, while they participated in 16 multilateral wars during the period from 1816 to 1992; and democratic states emerged as winners five times from the eight bilateral wars, while 12 times in 16 multilateral wars. Another feature of democratic states in war is that they are more likely to have been the target of military challenges. During the period from 1816 to 1992, only 25 percent of democratic war participants are initiators, and about 33 percent of democratic war participants fight on the initiator side. More than 67 percent of democratic war participants fight on the defender side, and about 64 percent of democratic winners are defenders. These limitations in the two current explanations allow this study to undertake a new inquiry on the causes of democracies' victory.
Analysis
Since the dependent variable, win and lose, is dichotomous, I use a probit model with robust standard errors to test my hypotheses about the determinants of democracies' victory; key results are presented in Table 1.8 While Model I presents the results of testing whether democratic states are more likely to win wars than non-democratic states, Model II presents the results of testing the hypotheses on why democratic states are more likely to win wars than non-democratic states.9 In addition, Model III presents the results of testing one of the two competing arguments, which connects democracies' victory with their superior capacity to strengthen their military capabilities.
1
Probit Models of State War Performance
 Model I Model II Model III 
Independent variables 
Democracy .92*** 33% 0.39 14% .26 10% 
 (.24)  (.39)  (.37)  
Number of democratic partners — — .55*** 63% .57*** 65% 
   (.11)  (.11)  
Number of non-democratic partners — — −.16** −47% −.18** −50% 
   (.06)  (.06)  
Democracy*initiation — — −.02 −1% .31 10% 
   (.59)  (.59)  
Democracy*number of democratic partners — — .10 13% .20 24% 
   (.16)  (.17)  
 Control variables 
Military capabilities 2.04*** 66% 1.95*** 62% — — 
 (.43)  (.42)    
Initiation .41* 16% .50** 19% 0.70*** 26% 
 (.20)  (.20)  (0.19)  
Assisted military capabilities 1.93*** 63% 1.64*** 54% 0.78* 28% 
 (.35)  (.48)  (.43)  
Constant −1.31 — −1.18 — −.40 — 
 (.22)  (.22)  (.15)  
Number of observations 266  266  266  
Observations correctly predicted 168  204  162  
Percentage correctly predicted 63%  77%  61%  
Wald χ2 (degrees of freedom) 52.16 (4)***  98.35 (8)***  72.73 (7)***  
Log-likelihood −149.86  −133.34  −154.29  
 Model I Model II Model III 
Independent variables 
Democracy .92*** 33% 0.39 14% .26 10% 
 (.24)  (.39)  (.37)  
Number of democratic partners — — .55*** 63% .57*** 65% 
   (.11)  (.11)  
Number of non-democratic partners — — −.16** −47% −.18** −50% 
   (.06)  (.06)  
Democracy*initiation — — −.02 −1% .31 10% 
   (.59)  (.59)  
Democracy*number of democratic partners — — .10 13% .20 24% 
   (.16)  (.17)  
 Control variables 
Military capabilities 2.04*** 66% 1.95*** 62% — — 
 (.43)  (.42)    
Initiation .41* 16% .50** 19% 0.70*** 26% 
 (.20)  (.20)  (0.19)  
Assisted military capabilities 1.93*** 63% 1.64*** 54% 0.78* 28% 
 (.35)  (.48)  (.43)  
Constant −1.31 — −1.18 — −.40 — 
 (.22)  (.22)  (.15)  
Number of observations 266  266  266  
Observations correctly predicted 168  204  162  
Percentage correctly predicted 63%  77%  61%  
Wald χ2 (degrees of freedom) 52.16 (4)***  98.35 (8)***  72.73 (7)***  
Log-likelihood −149.86  −133.34  −154.29  
Note: The coefficients of explanatory variables are computed by Stata 7.0, and their marginal impacts are computed based on King et al. (2000). CLARIFY: Software for Interpreting and Presenting Statistical Results. Version 2.0, Cambridge, MA: Harvard University, June 1 〈​http://gking.harvard.edu​〉.
*
p<.1;
**
p<.01;
***
p<.001 based on two-tailed test.
Open in new tab
1
Probit Models of State War Performance
 Model I Model II Model III 
Independent variables 
Democracy .92*** 33% 0.39 14% .26 10% 
 (.24)  (.39)  (.37)  
Number of democratic partners — — .55*** 63% .57*** 65% 
   (.11)  (.11)  
Number of non-democratic partners — — −.16** −47% −.18** −50% 
   (.06)  (.06)  
Democracy*initiation — — −.02 −1% .31 10% 
   (.59)  (.59)  
Democracy*number of democratic partners — — .10 13% .20 24% 
   (.16)  (.17)  
 Control variables 
Military capabilities 2.04*** 66% 1.95*** 62% — — 
 (.43)  (.42)    
Initiation .41* 16% .50** 19% 0.70*** 26% 
 (.20)  (.20)  (0.19)  
Assisted military capabilities 1.93*** 63% 1.64*** 54% 0.78* 28% 
 (.35)  (.48)  (.43)  
Constant −1.31 — −1.18 — −.40 — 
 (.22)  (.22)  (.15)  
Number of observations 266  266  266  
Observations correctly predicted 168  204  162  
Percentage correctly predicted 63%  77%  61%  
Wald χ2 (degrees of freedom) 52.16 (4)***  98.35 (8)***  72.73 (7)***  
Log-likelihood −149.86  −133.34  −154.29  
 Model I Model II Model III 
Independent variables 
Democracy .92*** 33% 0.39 14% .26 10% 
 (.24)  (.39)  (.37)  
Number of democratic partners — — .55*** 63% .57*** 65% 
   (.11)  (.11)  
Number of non-democratic partners — — −.16** −47% −.18** −50% 
   (.06)  (.06)  
Democracy*initiation — — −.02 −1% .31 10% 
   (.59)  (.59)  
Democracy*number of democratic partners — — .10 13% .20 24% 
   (.16)  (.17)  
 Control variables 
Military capabilities 2.04*** 66% 1.95*** 62% — — 
 (.43)  (.42)    
Initiation .41* 16% .50** 19% 0.70*** 26% 
 (.20)  (.20)  (0.19)  
Assisted military capabilities 1.93*** 63% 1.64*** 54% 0.78* 28% 
 (.35)  (.48)  (.43)  
Constant −1.31 — −1.18 — −.40 — 
 (.22)  (.22)  (.15)  
Number of observations 266  266  266  
Observations correctly predicted 168  204  162  
Percentage correctly predicted 63%  77%  61%  
Wald χ2 (degrees of freedom) 52.16 (4)***  98.35 (8)***  72.73 (7)***  
Log-likelihood −149.86  −133.34  −154.29  
Note: The coefficients of explanatory variables are computed by Stata 7.0, and their marginal impacts are computed based on King et al. (2000). CLARIFY: Software for Interpreting and Presenting Statistical Results. Version 2.0, Cambridge, MA: Harvard University, June 1 〈​http://gking.harvard.edu​〉.
*
p<.1;
**
p<.01;
***
p<.001 based on two-tailed test.
Open in new tab
First, the basic model of war performance, presented on the left side in Table 1, includes the democracy variable and three control variables—military capabilities, initiation, and assisted military capabilities. The coefficients of all these variables are strongly positive and statistically significant. Most importantly, the coefficient on the democracy variable is strongly positive and statistically significant (b=.92, p=.00). This finding indicates that democratic states are more likely to win wars than non-democratic states controlling for other important factors explaining war outcomes.
The coefficients in the probit model do not, however, directly illuminate the magnitude of impact of the explanatory variables on the dependent variable. In order to evaluate and compare how much each of the explanatory variables substantively affects the probability of winning a war all else being equal, I present the results of a marginal impact analysis in the second column of Model I.10 Four key points emerge from that analysis. First, the predicted probability of winning a war increases by 66 percentage points on average, plus or minus 10 percentage points, as the relative military capability variable changes from its minimum to its maximum. Second, the probability of winning a war increases by 63 percentage points on average, plus or minus 8 percentage points, as the assisted military capability variable changes from its minimum to its maximum. Third, the probability of winning a war increases by 16 percentage points on average, plus or minus 8 percentage points, as the initiation variable changes from 0 to1. Finally, the probability of winning a war increases 33 percentage points on average, plus minus 8 percentage points, if the democracy variable is changed from 0 to 1.11
Overall, the results presented in Model I suggest that, most importantly, the democracy variable significantly affects war outcomes. With regard to the control variables, all of them significantly affect war outcomes. Among these three variables, a state's own military capabilities and assisted military capabilities have a greater impact on war outcomes than does initiation. This finding suggests that war outcomes are more likely to be a function of internal and external balancing of capabilities rather than a function of caution, motivation, and strategy in which the initiator side is usually advantaged. Furthermore, another interesting result in this model is that the impact of assisted military capabilities is almost as great as that of a state's own military capabilities, and therefore, as crucial as the latter for explaining war outcomes.12
In order to investigate the mechanism through which democratic states are particularly advantaged in the conduct of wars, Model II estimates the effects on the probability of winning a war of the number of democratic and non-democratic partner variables, two interaction-term variables between democracy and initiation and between democracy and the number of democratic partners, and the four basic variables employed in Model I.13
Confirming the results of Model I, we can observe in column 3 in respect to Model II that the coefficients for the military capability, initiation, and assisted capability variables continue to be strongly positive and statistically significant. According to the results of a marginal impact analysis associated with the second model and presented in column 4 of the table, as each variable changes from its minimum to its maximum and all else being equal, a state's own military capabilities increase the probability of winning by 62 percentage points on average, plus or minus 9 percentage points. A state that initiates a war increases the chance of winning by 19 percentage points on average, plus or minus 7 percentage points. Assisted capabilities from cobelligerents also increase the probability of winning by 54 percentage points on average, plus or minus 12 percentage points.
Most importantly, I include the number of democratic and non-democratic partners variables in Model II since I expect that, depending upon their domestic political regime type, the behavior of allies can be different, and therefore affect war outcomes to different degrees even if the same amount of military capabilities are provided by allies. The coefficients for both democratic and non-democratic partners variables are statistically significant, and the direction of those coefficients is opposite of one another; in a manner consistent with the expectation discussed in the main hypothesis section, the number of democratic partners variable has a positive impact on war outcomes, while the number of non-democratic partners variable has a negative impact on them.14
Figure 1 shows that the degree to which the number of democratic partners and non-democratic partners respectively affect the probability of wining a war, all other factors being equal. According to a marginal impact analysis summarized in Figure 1 as well as in Table 1, the number of democratic partners variable increases the probability of winning a war by 63 percentage points on average, plus or minus 6 percentage points, as the variable moves from 0 to 5 democratic partners. In contrast, the number of non-democratic partners variable decreases the probability of winning by 47 percentage points on average, plus or minus 14 percentage points, as the variable changes from 0 to 9 non-democratic partners (see the dotted lines with squares in the figure). The results of this marginal impact analysis suggest that the total impact of the number of democratic partners variable is greater than that of the assisted military capability variable, that even small numbers of democratic partners can generate a tremendous impact on the probability of winning, and that the probability of winning substantially continues to increase as each democratic partner joins wartime coalitions. These findings imply that any powerful state alone cannot be the contributor for democratic war coalitions, and cooperation among democracies is crucial for victory.
1
Open in new tabDownload slide
The Mariginal Effects of the Number of Partners on the Probability of Winning a War
1
Open in new tabDownload slide
The Mariginal Effects of the Number of Partners on the Probability of Winning a War
Furthermore, the difference between the number of democratic partners and non-democratic partners in terms of their respective impacts on war outcomes can be shown clearly when each change in the numbers of partners simultaneously occurs with the maximum change in assisted military capabilities from allies. Even if a state receives the maximum observable amount of military capabilities from allies during a war, the impacts of these capabilities on the probability of winning vary depending on the domestic regime type of its allies. For example, if a state obtains the maximum observable amount of military capabilities from three democratic partners, its probability of winning increases by 77 percentage points on average, plus or minus 7 percentage points. In contrast, if a state obtains the observable maximum amount of military capabilities from three non-democratic partners during a war, its probability of winning increases only by 39 percentage points on average, plus or minus 11 percentage points. Considering that the assisted military capability variable alone increases the probability of winning by 54 percentage points as shown in Model II, if a state obtains those capabilities only from democratic partners, it can radically enhance its probability of winning. In contrast, if a state obtains those capabilities only from non-democratic partners, its probability of winning is below or smaller than change in the probability of winning that the assisted capability variable alone generates (the solid lines with triangles in the figure). In sum, even if a state receives the maximum possible capabilities from its partners, the impacts of these capabilities on the probability of winning are different depending upon the domestic regime type of its partners, and those differences become even greater as the number of each type of partners increases.
From these findings, we can infer that if a state attracts democratic partners, it can increase its probability of winning by obtaining military capabilities from as many allies as possible. In contrast, if a state only attracts non-democratic partners, it would be better off if it obtained those military capabilities from as small a number of allies as possible. That is to say that, while a collective action problem is operating in coalitions among non-democratic states, a synergistic effect is operating in coalitions among democratic states.
Moving to focus on the democracy variable, very interestingly and importantly, after including the variable measuring the number of democratic partners in Model II, the coefficient for the democracy variable is no longer statistically significant. Moreover, the marginal impact of democracy on war outcomes drastically drops from 33 percentage points to 14 percentage points. This finding suggests that the effect of the democracy variable by itself is a weak predictor of war outcomes once the effects of the number of democratic partners are taken into account. That is, the victories of democratic states cannot be explained directly by the advantages solely located within their national political structure and should be explained by the dynamics in wartime cooperation among coalition member states.
Another result reported in Model II, further challenges the argument that emphasizes the internal efficacy of individual democratic states. As noted above, the predominant argument about democratic war efficacy centers around the view that democracies are superior in choosing targets against whom they initiate wars. This line of inquiry is addressed in the second model. We find that the interaction-term variable linking initiation and democracy is negative and statistically insignificant (b=−.02, p=.97). This finding suggests that we cannot accept this relationship as the strongest determinant of democracies' victories, and, at the same time, we also cannot be strongly confident that initiation is negatively associated with war performance of democratic states because there is a rather high chance that there might be no relationship.
The large standard errors in multivariate statistical models, which usually result in statistical insignificance, are in part due to a small number of cases. Given that, cross tabulation distributions could be an alternative and sometimes less demanding approach to analyze the difference in the relationships of two variables between two groups (Rasler and Thompson, 1999:428). Therefore, I further compare two cross-tabulations to measure the level of association of initiation with war outcomes between democratic and non-democratic war fighters; the results of this analysis are presented in Table 2.
2
Democracy, War Initiation, and State War Performance
 Lose Win Total 
1. Non-democratic states 
Did not initiate 79 (59%) 55 (41%) 135 
Did initiate 30 (39%) 47 (61%) 76 
Total 109 (52%) 102 (48%) 211 
 Fisher's exact=.006 
2. Democratic States 
Did not initiate 9 (22%) 32 (78%) 41 
Did initiate 2 (14%) 12 (86%) 14 
Total 11 (20%) 44 (80%) 55 
 Fisher's exact=.709 
 Lose Win Total 
1. Non-democratic states 
Did not initiate 79 (59%) 55 (41%) 135 
Did initiate 30 (39%) 47 (61%) 76 
Total 109 (52%) 102 (48%) 211 
 Fisher's exact=.006 
2. Democratic States 
Did not initiate 9 (22%) 32 (78%) 41 
Did initiate 2 (14%) 12 (86%) 14 
Total 11 (20%) 44 (80%) 55 
 Fisher's exact=.709 
Note: One of the requirements of χ2 test, especially in 2 × 2 tables, is that all expected counts in each cell should be 5 or greater. Fisher's exact test can be used when any counts in each cell have 5 or fewer. See Rasler and Thompson (1999).
Open in new tab
2
Democracy, War Initiation, and State War Performance
 Lose Win Total 
1. Non-democratic states 
Did not initiate 79 (59%) 55 (41%) 135 
Did initiate 30 (39%) 47 (61%) 76 
Total 109 (52%) 102 (48%) 211 
 Fisher's exact=.006 
2. Democratic States 
Did not initiate 9 (22%) 32 (78%) 41 
Did initiate 2 (14%) 12 (86%) 14 
Total 11 (20%) 44 (80%) 55 
 Fisher's exact=.709 
 Lose Win Total 
1. Non-democratic states 
Did not initiate 79 (59%) 55 (41%) 135 
Did initiate 30 (39%) 47 (61%) 76 
Total 109 (52%) 102 (48%) 211 
 Fisher's exact=.006 
2. Democratic States 
Did not initiate 9 (22%) 32 (78%) 41 
Did initiate 2 (14%) 12 (86%) 14 
Total 11 (20%) 44 (80%) 55 
 Fisher's exact=.709 
Note: One of the requirements of χ2 test, especially in 2 × 2 tables, is that all expected counts in each cell should be 5 or greater. Fisher's exact test can be used when any counts in each cell have 5 or fewer. See Rasler and Thompson (1999).
Open in new tab
Table 2 shows that 61 percent of non-democratic initiators win while almost the same percent of non-democratic non-initiators are defeated. On the other hand, 86 percent of democratic initiators win, but at the same time, 78 percent of democratic non-initiators also win. In other words, initiation is positively and statistically significantly associated with war outcomes in the case of non-democratic states (p=.01), while initiation is not statistically significantly associated with outcomes in the case of democratic states (p=.71). Theses findings in Table 2 clearly present that it is hard to accept the selection effect argument that democratic states are more likely to win a war because they can carefully pick the war.15
Next, in order to test more directly the war-fighting argument, another explanation about democracies' victory based on their internal efficacy, I construct MODEL III simply by removing the military capability variable from the second war performance model. The results for Model III are reported in column 5 of Table 1. Gelpi and Griesdorf (2001:641) point out that military capabilities may be an intervening variable between democracy and war outcomes in Lake's argument, and suggest that if this is the case, when the democracy and military capability variables are included in the model together, there is a chance that the effect of the democracy variable is likely to be smaller. In other words, if it were found that the effect of the democracy variable has become stronger after the military capability variable is removed from Model II, this could be evidence in support of Lake's argument about “powerful” democratic states.
According to the results for Model III, however, the coefficient for the democracy variable has become smaller rather than larger, and without much change in its standard errors. This result suggests that it is hard to connect democracies' victory with their military capabilities. Indeed, a marginal impact analysis reported in column 6 in Table 1, showing the smaller impact of the democracy variable on the probability of winning in Model III compared to its impact in Model II, consistently suggests that the military capability variable may be less important for democratic war fighters than non-democratic counterparts. Rather, when moving from Model II to Model III, a great shift in the impacts of the explanatory variables on war outcomes occurs in that of the initiation variable. The initiation variable increases the probability of winning by an additional 7 percentage points in Model III with a smaller standard error compared to Model II. It appears that initiation and military capabilities are positively related with each other, in a manner consistent with conventional wisdom, a state which has stronger military capabilities compared to its adversaries is likely to initiate a war and to achieve a victory.
Finally, in order to test whether democratic states benefit more from democratic partners for winning wars, I include an interaction-term variable between democracy and the number of democratic partners. The coefficient for this variable is positive, but is not statistically significant (b=.10, p=.52). That is, although the positive sign of the coefficient shows that the probability of winning is enhanced when democratic states join other democratic states, the large standard error of the coefficient casts doubts on this relationship. Therefore, up to this point, we can only say that both democratic and non-democratic states can increase their probability of winning by attracting democratic partners as discussed in Hypothesis 1.
Nonetheless, it is plausible that the large standard error of the interaction term variable between democracy and the number of democratic partners is attributable to the relatively small number of observations. Moreover, the coding rule in mixed war coalitions between democratic and non-democratic states allows non-democratic states to have more democratic partners than democratic states, thus weakening the effect of this interaction term variable. It is, therefore, very important to examine which states are more likely to have democratic partners during wars. I test this question by regressing the military capability, role in the war, and domestic regime type of war participant states on the number of democratic partners in Table 3.16
3
Poisson Regression Model of the Number of Democratic Partners
Independent Variables Coefficients SE 
Relative military capabilities −3.42* .93 
Defender 1.12* .32 
Democracy 1.20* .23 
Cons −1.01 .33 
Number of observations 266  
Dependent variable Mean=0.68, min=0 max=5 
Wald χ2 (degree of freedom) 148.35(3)*  
Independent Variables Coefficients SE 
Relative military capabilities −3.42* .93 
Defender 1.12* .32 
Democracy 1.20* .23 
Cons −1.01 .33 
Number of observations 266  
Dependent variable Mean=0.68, min=0 max=5 
Wald χ2 (degree of freedom) 148.35(3)*  
*
p<.000 based on two-tailed test.
Open in new tab
3
Poisson Regression Model of the Number of Democratic Partners
Independent Variables Coefficients SE 
Relative military capabilities −3.42* .93 
Defender 1.12* .32 
Democracy 1.20* .23 
Cons −1.01 .33 
Number of observations 266  
Dependent variable Mean=0.68, min=0 max=5 
Wald χ2 (degree of freedom) 148.35(3)*  
Independent Variables Coefficients SE 
Relative military capabilities −3.42* .93 
Defender 1.12* .32 
Democracy 1.20* .23 
Cons −1.01 .33 
Number of observations 266  
Dependent variable Mean=0.68, min=0 max=5 
Wald χ2 (degree of freedom) 148.35(3)*  
*
p<.000 based on two-tailed test.
Open in new tab
After estimating the poisson regression model with robust standard errors presented in Table 3, we observe, in the first place, the smaller a state's military capabilities, the more likely it is to have democratic partners. This finding shows that democratic partners appear to be balancing rather than bandwagoning by joining the militarily weaker side, and suggests that it is hard to support one of the alternative arguments based on selection-effect mechanism, that democratic partners simply join the side which they think is more likely to be victorious. Second, the defender side is more likely to have democratic partners. If this finding is combined with the impact of the number of democratic partners on war outcomes, it can in part explain how democratic states are more likely than non-democratic states to appear as a target in international conflicts, and democratic targets are more likely to win wars (Rasler and Thompson, 1999; Gelpi and Grieco, 2001). Finally and most importantly, as expected in the discussion of Hypothesis 3, democratic states are more likely than non-democratic states to have democratic partners; in particular, democratic states are likely to have 0.96 more democratic partners than non-democratic states on average, plus or minus .3, all else being equal. All these findings are statistically significant (p<.00).17
Finally, summarizing the relationships regarding the impacts of democratic and non-democratic partners on the war performance of democratic and non-democratic states, in the first place, both democratic and non-democratic states can increase their probability of winning by attracting partners due to the capabilities they bring into the battlefields. Once the impacts of these assisted capabilities are controlled, both democratic and non-democratic states can increase their probability of winning by obtaining support from only democratic partners while neither democratic nor non-democratic states increase the probability of winning by obtaining that support from non-democratic partners. Furthermore, as the number of democratic partners increases, both democratic states and non-democratic states can enhance their probability of winning. In contrast, as the number of non-democratic partners increases, both democratic and non-democratic states decrease their probability of winning. Most importantly, it is democratic states that are more likely to attract democratic allies as discussed above in Table 3, and therefore more likely to enjoy a higher probability of winning than non-democratic states.
Conclusion
Students of international relations have rightly observed that democratic states are formidable in war. This study has sought to contribute to our understanding of the mechanism by which democracies enjoy their superior performance in war, and has suggested that it in part may concern their superior capacity to cooperate with their partners, especially by attracting democratic partners during wars. Due to the stability of their preferences and the transparency of the polities, once determined, democracies are better able to cooperate with their partners in the conduct of war, and thereby are more likely to win wars.
These findings have important implications for international relations theory. Most importantly, the key finding, linking wartime cooperation among coalition member states to democratic war success, challenges the current literature that tends to emphasize exclusively the internally generated efficacy of democratic systems as they use force abroad. By overlooking the dynamic features of war and by neglecting the nature of wars where democratic states fight, those two current arguments reveal serious limitations in explaining democracies' victory. Identifying a weak association between the war performance of democratic states and initiation on the one hand, and military capabilities on the other, the results in this study suggest that democratic states are more likely to win wars because democratic states are cooperative partners not because they are selective initiators or they are effective resource generators in the prosecution of wars.
More broadly, the findings above challenge realist arguments about world affairs. In the first place, in contrast to traditional realist perspective, my findings suggest that democratic states are more effective and reliable actors with whom to work. Moreover, in contrast to neorealist arguments, domestic political regime type matters greatly in explaining world politics. This study demonstrates that the military capabilities from wartime allies largely affect war outcomes, but depending upon their domestic political regime type, there are significant variations in how these capabilities are effectively employed and operated in the conduct of wars. This finding implies that securing and supporting democratic allies in the face of external threats and military challenges should be a high priority for the prudent democratic leaders.
Finally, beyond challenging the present literature, these findings expand the research program on the role of democracy in foreign policy and world politics. They connect the impact of democracy in military conflict with the impact of democracy in international cooperation. Moreover, they suggest directions for future research by asking to test cooperation mechanisms more thoroughly and systematically and posing the question of whether the superior capacity of democracy to cooperate also occurs in peacetime international interactions.
Appendix
Domain of Cases
The domain of cases for my study is drawn for the most part from the Correlates of War (COW) project's data set of Interstate Wars during the period from 1816 to1992. This data set includes 75 interstate wars. I made four changes in the COW Interstate War data set for the purposes of my analysis.18 First, I excluded 26 cases of draws between states in six wars during this period. I did this to have observations that are consistent with my dependent variable, winning and losing in wars.19 Next, I divide World War I and World War II into several different wars on the basis of the major theaters of operation. World War I is divided into two distinct wars. The “Eastern Front War” includes Austria–Hungary, Bulgaria, Germany, and Turkey versus Britain, France, Greece, Italy, Romania, Russia, and Yugoslavia. The “Western Front War” includes Austria–Hungary and Germany versus Belgium, Britain, France, Italy, and the U.S. I also treat World War II as twelve separate wars. There are four dyadic wars between Germany and Belgium, Norway, Netherlands, and Poland, respectively, and one dyadic war between Russia and Poland. There are seven multilateral wars between the Axis Powers and the Allied Powers.20 In this process, the countries that did not fight in the main battles are excluded. In addition, Belgium, Norway, Netherlands, France, Italy, and several Eastern European countries are excluded from the subsequent wars if they had fought and surrendered in earlier main battles. Dividing these two world wars into a series of separate wars helps us understand clearly which participants fought, what they fought for, and when they surrendered or achieved victory. Finally, I include Greece as a participant during the Turko-Cyprus war of 1974–1975. This war is a part of a protracted conflict between Greece and Turkey over Cyprus. The crisis was triggered by a military coup in Cyprus supported by the Greek government. Greece mobilized its forces during the war, and later it participated in the peace negotiations (Brecher and Wilkenfeld, 1997:366–371). In sum, I examine in the analysis below a total of 266 state-participants in 80 wars during the period from 1816 to 1992.
In order to examine whether my coding rules might affect the results of the statistical analysis, I conducted sensitivity tests as follows: I tested the main state war performance model (Model II) removing all the U.S. cases from the data set (b=.54, p=.00); second, I removed Australia, Canada, and New Zealand (b=.55, p=.00); third, in order to avoid that large coalition wars determine the results, I removed all World War I (b=.46, p=.00) and World War II (b=.59, p=.00) cases; and finally, I included only the periods after 1900 (b=.85, p=.00), 1918 (b=.75, p=.00) and 1945 (b=1.67, p=.00) in the data set. Overall, as shown by the coefficients and p-values in the parentheses, although the size of the coefficients for the number of democratic partners variable changes across these models, the coefficients are positive and statistically significant across all these models.
Dependent Variable: Win and Lose
To measure the dependent variable, wining or losing in a war, I use the war outcome variable in the COW Interstate War data set. The war outcome variable is 1 if a state won, and 0 if the state lost in an interstate war. In this coding process, I modified its coding for the outcome variable in order to correspond to the disaggregation of the two world wars. For example, I coded Germany as a winner in a series of bilateral wars against Poland, the Netherlands, Belgium, and Norway in 1939 and 1940, a winner in multilateral wars against France and Britain in 1940, Greece in 1940–1941, and Yugoslavia in 1941, and a loser in a series of multilateral wars against the Allied Powers, Australia, Britain, Canada, New Zealand, Russia, and the U.S. between 1941 and 1945.
Independent Variables
Democracy. To measure the democracy score of each war participant state, I use the Polity IV data set (Marshall and Jaggers, 2003). Based on this data set I first construct a score for each war participant by following the formula: domestic regime type=democracy score−autocracy score+10. This domestic regime-type score ranges from 0 to 20, and the higher the score ascribed to a country the more fully democratic that country is assumed to be. In my analysis, I use a dummy democracy variable by coding the war participant as 1 if its score is 17 and greater, and 0 otherwise.21
Number of Democratic Partners. The number of democratic partners variable is based on how many democratic partners a war participant state has during the war it fought.
Number of Non-Democratic Partners. The number of non-democratic partners variable is based on how many non-democratic partners a war participant state has during the war it fought.
Democracy and the Number of Democratic Partners. In order to examine which states are more likely to benefit from democratic partners for victory, I construct an interaction term variable between democracy and the number of democratic partners by multiplying these two variables.
Democracy and Military Capabilities. I treat this variable as an intervening variable, and estimate its impact by constructing State War Performance Model III in the analysis section.
Democracy and Initiation. I construct an interaction-term variable between democracy and initiation by multiplying the democracy variable with the initiation variable whose measurement is discussed below.
Control Variables
Military Capabilities. To measure the military capabilities of war participants, I mostly use the COW National Material Capabilities of States data set constructed by Singer and Small (1994,1999). Based on this data set, I calculate each state's military capability relative to all participants in a war as follows.22 First, I convert three elements, number of soldiers, amount of military expenditure, and military expenditure per soldier, to a percentage of the world total of each element. Next, each value is discounted based on the distance between the actor and the location of the war (Fitspatrick, 1986). The measurement of this distance-discounted capability was developed by Bueno de Mesquita (1981:103). Third, I convert each value to a percentage of those of all participants in each war and average the values of the three elements.
Initiation. I follow the coding of the initiation variable in the COW Interstate War data set. I code the initiation variable as 1 if a state initiated a war, and 0 otherwise.
Assisted Military Capabilities. I measure the assisted military capability variable by subtracting a state's own military capabilities from the sum of military capabilities on its side during a war.
Footnotes
Author's note: I would like to thank Bear Braumoeller, Giacomo Chiozza, Peter Feaver, Christopher Gelpi, Joseph Grieco, Paul Gronke, Dan Reiter, Allan Stam, and anonymous reviewers of International Studies Quarterly for their suggestions. I also would like to thank John M. Olin Institute for Strategic Studies at Harvard University for its support of this project. I am alone responsible for any remaining errors. Data are available upon request.
1
Desch (2002) recently raises doubts on this proposition.
2
I thank an anonymous reviewer for suggesting to add this point here. If my argument is correct, different kinds of cooperative behaviors among democracies should be observed; for example, compared to non-democracies, democracies should be less likely to abandon their partners or more likely to adopt joint military commands to concentrate their efforts in the conduct of war. For the empirical tests regarding the reliability and effectiveness of democracies as wartime allies, see Choi (2001, 2003).
3
There are only two historical cases that democratic states failed to maintain commitments to their partners during wars. These are Australia and the U.S. during the Vietnam War. However, it took about 8 years for these two states to reverse their commitment to South Vietnam and withdraw from the war, see Choi (2001, 2003).
4
For the success and failure of cooperative efforts among the Grand Alliance partners during World War II, see Lukas (1970), Kitchen (1986), Reynolds et al. (1994), Lane and Temperley (1995) and Reynolds (1995).
5
Therefore, I expect mixed wartime coalitions between democratic and non-democratic states to be better able than wartime coalitions among non-democratic states to cooperate with each other to win wars. This expectation contrasts with the argument that mixed dyads between democracy and non-democracy are less likely than jointly autocratic dyads to cooperate with each other. See Leeds (1999).
6
See also Wang and Ray (1994).
7
Reiter and Stam (2002) have expanded the analytical purview of their argument to include the effectiveness of military leadership and the high level of solders' initiatives on the battlefield as additional explanatory factors for democracies' victory in war. Although the battlefield effectiveness argument emphasizes the importance of wartime behavior, in Reiter and Stam's (2002) formulation, it still does not capture the dynamics of allies' choices and behaviors during war, and thus may be subject to the same line of criticism I elaborated with respect to the war-fighting model.
8
There may be a possibility of selection bias in my analysis. For example, relating to my main argument, states may be more likely to go to war when they believe that they have a high probability of support by their democratic partners. Therefore, the reliability of democratic partners may affect the decision of states to initiate or to retaliate if attacked. This could mean that many of the cases in my war outcome equation appear only after the relevant actors have made estimates as to whether they have reliable democratic partners. As a result, the states that participated in war are not necessarily a representative sample of the pool of all states. These effects could be non-random, and therefore they could affect my findings. I solve this possible problem by showing how selection bias may affect my findings. In the censored sample, the coefficient of the number of democratic partners variable in the war outcome equation is likely to be underestimated, regardless of its true impact. This point suggests that although my findings might be biased, it is at least not the case that this bias yields statistical results favorable toward my argument. See Achen (1986).
9
One of the important requirements in scientific research is to report the limitations of research. I would like to report that there is the possibility of non-independence among observations in this study (King, Keohane, and Verba, 1994). In order to examine whether this non-independence among observations seriously distorts my findings, I used several different tools since there is no decisive solution for this issue in the field of quantitative international relations. First, I ran the models with robust standard errors clustering on the war number. The size and direction of the coefficients for explanatory variables do not substantially change, and they are statistically significant although their standard errors become somewhat larger. For example, in regard to the number of democratic partners variable, its coefficient does not change (from .55 to .55) and its p-value only changes from .00 to .01.
Second, I randomly selected 75 cases from all of the cases and I also randomly selected only one state per war following the suggestion presented in Stam's book (1996). The results were indistinct in the first place, so I reran the model taking out interaction term variables due to the limited number of the total observations, and including the period of post-1900 due to the infrequency of democratic war participants in the 19th century. The standard errors are somewhat larger, but the overall results are consistent with the main results presented in this study (b=.43, p=.02).
Third, to address the non-independence problem relating to the effects of the main variables in this study, the number of democratic and non-democratic partners, in a manner suggested by Bear Braumoeller, I included the number of democratic and non-democratic partners variables on the other side in the equation, and conducted the F-test of joint significance of the coefficients for the number of democratic partners variable and the number of democratic partners on the other side variable, and of the coefficients for the number of non-democratic partners variable and the number of non-democratic partners on the other side variable. It turned out that each of these two sets of variables is statistically the same and jointly significant (p=.00).
10
To calculate the marginal impacts of the explanatory variables, the democracy and initiation variables, and the interaction term variable between democracy and initiation, are set at their modal category while the number of democratic and non-democratic partners, relative military capability, and assisted military capability variables, and the interaction term variable between democracy and the number of democratic partners are set at their mean value.
11
The interpretation of the marginal effects of the explanatory variables on the predicted probability of winning a war is based on King, Tomz, and Wittenberg (2000).
12
This finding is different from the expectation in the recent literature which argues that the impact of alliances on war outcomes should be significantly smaller than that of a state's own capabilities. See Stam (1996).
13
By adding the number of democratic and non-democratic partners variables and the two interaction term variables in addition to the four explanatory variables in Model I, the explanatory power of Model II is significantly improved according to a log-likelihood ratio test on the two war performance models (p<.00).
14
I, first, replicated Reiter and Stam's (1998) main model, Model 5 in Table 2, and then ran their model again including the two new variables, the number of democratic and non-democratic partners. According to the results of this new test, although their interpretation of several multilateral wars is different from mine, these two variables have significant impacts on war outcome: the number of democratic partners variable is positively associated with the probability of winning (b=1.75, p=.01) whereas the number of non-democratic partners variable is negatively associated with it (b=−.34, p=.04) as predicted in this study. Before I ran this test, I had to change one coding in their data set; without it, the democratic partner variable completely predicts war outcomes and those observations are excluded from analysis. I assigned one democratic partner to the Norway case during World War II based on their suggestion.
15
Explaining the advantages of initiators in war, Rasler and Thompson have also not found any support for the argument that democratic initiators are more selective than non-democratic initiators. They rather have found that defender democratic states disproportionately tend to win wars. See Rasler and Thompson (1999).
16
Since the number of democratic partners is a count variable in a fixed domain with no certain upper bound, I use a poisson regression model instead of an ordinary least squares model. See King (1988). To calculate the marginal effects of the democracy variable using CLARIFY, the military capability variable is set at its mean value and the defender variable is set at its modal value. I also test the question of the number of democratic partners based on a negative binomial regression model by being aware of the possibility that having one democratic partner is likely to attract other democratic partners. See King (1998). According to the results of this analysis, democratic states are likely to have 1.6 more democratic partners than non-democratic states during wars.
17
The tendency of democratic states to have more democratic partners during wars can be related to the type of wars democracies fight. To test this conjecture, I exclude major wars from the model reported in Table 3. I find that when excluding World War I (b=1.31, p=.00), World War II (b=1.02, p=.00), and both World War I and II (b=1.07, p=.00) cases, respectively, the coefficient on the democracy variable remains positive and highly significant.
18
Given the limited number of cases, the coding rules about cases and variables may generate different findings between this study and Reiter and Stam's study. To compare the coding rules in these two studies, see Choi (2001:223–228) and Reiter and Stam (2002:52–57).
19
This coding decision is consistent with those of Lake (1992) and Reiter and Stam (1998, 2001). In order to examine whether the exclusion of these draw cases might affect the statistical results, I re-estimated the main state war performance model (Model II) in Table 1 using an ordered probit model in which the dependent variable is coded as 0 for defeats, as 1 for draws, and as 2 for victories. The results from the ordered probit model confirm that the number of democratic partners variable has a positive and statistically significant impact on war outcome (b=.16 p=.00).
20
For this disaggregation of the two world wars, I mostly follow Wright (1964), Dupuy and Dupuy (1993) Clodfelter (1992), Ellis (1993), Dear and Foot (1995), and Reiter and Stam (1998, 2002).
21
The formula for this democracy variable was developed by Rousseau, Gelpi, Reiter, and Huth (1996).
22
This formula was developed by Rousseau et al. (1996).
References
Achen
C.
(
1986
)
The Statistical Analysis of Quasi-Experiments
.
Berkeley
:
University of California Press
.
Google Scholar
OpenURL Placeholder Text
 
Beardsley
E.
(
1977
)
Secret between Friends
.
Social Studies of Science
7
:
447

473
.
Google Scholar
Crossref
Search ADS
 
Blainly
G.
(
1988
)
The Causes of War
.
New York
:
The Free Press
.
Google Scholar
OpenURL Placeholder Text
 
Brecher
M.
Wilkenfeld
J.
(
1997
)
A Study of Crisis
.
Ann Arbor
:
University of Michigan Press
.
Google Scholar
OpenURL Placeholder Text
 
Bueno de Mesquita
B.
(
1981
)
The War Trap
.
New Haven, CT
:
Yale University Press
.
Google Scholar
OpenURL Placeholder Text
 
Bueno de Mesquita
B.
Morrow
J.
Siverson
R.
Smith
A.
(
1999
)
An Institutional Explanation of the Democratic Peace
.
American Political Science Review
93
:
791

807
.
Google Scholar
Crossref
Search ADS
 
Bueno de Mesquita
B.
Siverson
R.
(
1995
)
War and the Survival of Political Leaders
.
American Political Science Review
89
:
841

855
.
Google Scholar
Crossref
Search ADS
 
Bueno de Mesquita
B.
Siverson
R.
Woller
G.
(
1992
)
War and Fate of Regimes
.
American Political Science Review
86
:
638

646
.
Google Scholar
Crossref
Search ADS
 
Choi
A.
(
2001
) “Cooperation for Victory: Democracy, International Partnerships and State War Performance, 1816–1992.” Ph.D. Dissertation, Duke University.
Choi
A.
(
2003
) “Friends and Allies: The Abandonment Behavior of States during War, 1816–1992.” Presented in the Annual Meeting of American Political Science Association in Philadelphia.
Clausewitz
C.
(
[1832] 1976
)
On War
.
Princeton, NJ
:
Princeton University Press
.
Google Scholar
OpenURL Placeholder Text
 
Clodfelter
M.
(
1992
)
Warfare and Armed Conflicts
.
Jefferson
:
McFarland
.
Google Scholar
OpenURL Placeholder Text
 
Cowhey
P.
(
1993
)
Domestic Institutions and Credibility of International Commitments
.
International Organization
47
:
299

326
.
Google Scholar
Crossref
Search ADS
 
Dear
I. C. B.
Foot
M. R. D.
, eds. (
1995
)
The Oxford Companion to World War II
.
Oxford
:
Oxford University Press
.
Google Scholar
OpenURL Placeholder Text
 
Desch
M.
(
2002
)
Democracy and Victory
.
International Security
27
:
4

47
.
Google Scholar
Crossref
Search ADS
 
Dolye
M.
(
1983
)
Kant, Liberal Legacies and Foreign Affairs, Part I and II
.
Philosophy and Public Affairs
12
:
205

235
and 323–353.
Google Scholar
OpenURL Placeholder Text
 
Dupuy
E.
Dupuy
T.
(
1993
)
The Encyclopedia of Military History from 3500 BC to the Present
.
New York
:
Harper & Row Publishers
.
Google Scholar
OpenURL Placeholder Text
 
Ellis
J.
(
1993
)
World War II
.
New York
:
Facts on File
.
Google Scholar
OpenURL Placeholder Text
 
Fearon
J.
(
1994
)
Democratic Political Audiences and the Escalation of International Disputes
.
American Political Science Review
88
:
577

592
.
Google Scholar
Crossref
Search ADS
 
Fearon
J.
(
1995
)
Rational Explanations for War
.
International Organization
49
:
379

414
.
Google Scholar
Crossref
Search ADS
 
Feaver
P.
(
1996
)
The Civil–Military Problematique
.
Armed Force and Society
23
:
149

179
.
Google Scholar
Crossref
Search ADS
 
Fitspatrick
G.
(
1986
)
Direct-line Distances
.
Metuchen
:
Scarecrow Press
.
Google Scholar
OpenURL Placeholder Text
 
Gartner
S.
Siverson
R.
(
1996
)
War Expansion and War Outcome
.
Journal of Conflict Resolution
40
:
4

15
.
Google Scholar
Crossref
Search ADS
 
Gaubatz
K.
(
1996
)
Democratic States and Commitment in International Relations
.
International Organization
50
:
109

40
.
Google Scholar
Crossref
Search ADS
 
Gelpi
C.
Grieco
J.
(
2001
)
Attracting Trouble
.
Journal of Conflict Resolution
45
:
794

817
.
Google Scholar
Crossref
Search ADS
 
Gelpi
C.
Griesdorf
M.
(
2001
)
Winner or Loser? Democracies in International Crisis, 1918–1988
.
American Political Science Review
95
:
633

48
.
Google Scholar
Crossref
Search ADS
 
Goemans
H.
(
2000
)
War and Punishment
.
Princeton, NJ
:
Princeton University Press
.
Google Scholar
OpenURL Placeholder Text
 
Grieco
J.
(
2001
)
Repetitive Military Challenges and Recurrent International Conflicts, 1918–1944
.
International Studies Quarterly
45
:
295

316
.
Google Scholar
Crossref
Search ADS
 
Kennan
G.
(
1977
)
The Cloud of Danger
.
Boston
:
Little Brown
.
Google Scholar
OpenURL Placeholder Text
 
Keohane
R.
(
1983
) “
The Demand for International Regimes
.”In
International Regime
, edited by
Krasner
S.
, pp.
141

172
.
Ithaca
:
Cornell University Press
.
Google Scholar
OpenURL Placeholder Text
 
King
G.
(
1988
)
Statistical Models for Political Science Event Counts
.
American Journal of Political Science
32
:
838

863
.
Google Scholar
Crossref
Search ADS
 
King
G.
(
1998
)
Unifying Political Methodology
.
Ann Arbor
:
University of Michigan Press
.
Google Scholar
OpenURL Placeholder Text
 
King
G.
Keohane
R.
Verba
S.
(
1994
)
Designing Social Inquiry
.
Princeton, NJ
:
Princeton University Press
.
Google Scholar
OpenURL Placeholder Text
 
King
G.
Tomz
M.
Wittenberg
J.
(
2000
)
Making the Most of Statistical Analyses
.
American Journal of Political Science
44
:
347

361
.
Google Scholar
Crossref
Search ADS
 
Kissinger
H.
(
1994
)
Diplomacy
.
New York
:
Simon & Schuster
.
Google Scholar
OpenURL Placeholder Text
 
Kitchen
M.
(
1986
)
British Policy towards the Soviet Union during the Second World War
.
New York
:
St. Martin's Press
.
Google Scholar
OpenURL Placeholder Text
 
Kugler
J.
Arbetman
M.
, eds. (
1997
)
Political Capacity and Economic Behavior
.
Boulder, CO
:
Westview Press
.
Google Scholar
OpenURL Placeholder Text
 
Lane
A.
Temperley
H.
, eds. (
1995
)
The Rise and Fall of the Grand Alliance, 1941–1945
.
New York
:
St. Martin's Press
.
Google Scholar
OpenURL Placeholder Text
 
Lake
D.
(
1992
)
Powerful Pacifists
: Democratic States and War.
American Political Science Review
86
:
24

37
.
Google Scholar
Crossref
Search ADS
 
Leeds
A.
(
1999
)
Domestic Political Institutions, Credible Commitments, and International Cooperation
.
American Journal of Political Science
43
:
979

1002
.
Google Scholar
Crossref
Search ADS
 
Lukas
R.
(
1970
)
Eagles East
.
Tallahassee, FL
:
Florida State University Press
.
Google Scholar
OpenURL Placeholder Text
 
Maoz
Z.
(
1983
)
Resolve, Capabilities, and the Outcomes of Interstate Disputes, 1816–1976
.
Journal of Conflict Resolution
27
:
195

229
.
Google Scholar
Crossref
Search ADS
 
Martin
L.
(
1999
)
Democratic Commitments
.
Princeton, NJ
:
Princeton University Press
.
Google Scholar
OpenURL Placeholder Text
 
Mearsheimer
J.
(
1983
)
Conventional Deterrence
.
Ithaca
:
Cornell University Press
.
Google Scholar
OpenURL Placeholder Text
 
Marshall
M.
Jaggers
K.
(
2003
)
Polity IV Project
.
College Park
:
Integrated Network for Societal Conflict Research and Center for International Development and Conflict Management, University of Maryland
.
Google Scholar
OpenURL Placeholder Text
 
Morgenthau
H.
(
1960
)
The Purpose of American Politics
.
New York
:
Alfred Knopf
.
Google Scholar
OpenURL Placeholder Text
 
Morgenthau
H.
(
1985
)
Politics Among Nations
.
New York
:
Alfred Knopf
.
Google Scholar
OpenURL Placeholder Text
 
Olson
M.
(
1993
)
Dictatorship, Democracy, and Development
.
American Political Science Review
37
:
567

576
.
Google Scholar
Crossref
Search ADS
 
Olson
M.
Zeckhauser
R.
(
1966
)
An Economic Theory of Alliance
.
Review of Economics and Statistics
48
:
266

279
.
Google Scholar
Crossref
Search ADS
 
Organski
A. F. K.
Kugler
J.
(
1980
)
The War Ledger
.
Chicago
:
University of Chicago Press
.
Google Scholar
OpenURL Placeholder Text
 
Powell
R.
(
2001
) “Bargaining While Fighting.” Unpublished Paper.
Rasler
K.
Thompson
W.
(
1985
)
War Making and State Making
.
American Political Science Review
79
:
491

507
.
Google Scholar
Crossref
Search ADS
 
Rasler
K.
Thompson
W.
(
1999
)
Predatory Initiators and Changing Landscapes for Warfare
.
Journal of Conflict Resolution
43
:
411

433
.
Google Scholar
Crossref
Search ADS
 
Reiter
D.
Stam
A.
(
1998
)
Democracy, War Initiation, and Victory
.
American Political Science Review
92
:
377

389
.
Google Scholar
Crossref
Search ADS
 
Reiter
D.
Stam
A.
(
2002
)
Democracies at War
.
Princeton, NJ
:
Princeton University Press
.
Google Scholar
OpenURL Placeholder Text
 
Reynolds
D.
(
1995
)
Rich Relations
.
London
:
HarperCollins Publishers
.
Google Scholar
OpenURL Placeholder Text
 
Reynolds
D.
Kimball
W. F.
Chubarian
A. O.
(
1994
)
Allies at War
.
New York
:
St. Martin's Press
.
Google Scholar
OpenURL Placeholder Text
 
Rosen
S.
(
1992
) “
War Power and the Willingness to Suffer
.”In
The Scientific Study of Peace and War
, edited by J. Vasquez and H. Marie, pp.
255

273
.
New York
:
Lexington Books
.
Google Scholar
OpenURL Placeholder Text
 
Rousseau
D.
Gelpi
C.
Reiter
D.
Huth
P.
(
1996
)
Assessing the Dyadic Nature of the Democratic Peace, 1919–88
.
American Political Science Review
90
:
512

533
.
Google Scholar
Crossref
Search ADS
 
Schultz
K.
(
1998
)
Domestic Opposition and Signaling in International Crises
.
American Political Science Review
90
:
829

844
.
Google Scholar
Crossref
Search ADS
 
Shanafelt
G.
(
1985
)
The Secret Enemy
.
New York
:
Columbia University Press
.
Google Scholar
OpenURL Placeholder Text
 
Singer
D.
Small
M.
(
1994
)
Correlates of War Project: Interstate War Data, 1816–1992
.
Ann Arbor
:
Inter-university Consortium for Political and Social Research (Study No. 9905)
.
Google Scholar
OpenURL Placeholder Text
 
Singer
D.
Small
M.
(
1999
)
Updated National Material Capabilities Data, 1816–1993
.
Ann Arbor
:
Inter-university Consortium for Political and Social Research (Study No. 9903)
.
Google Scholar
OpenURL Placeholder Text
 
Smith
A.
(
1996
)
To Intervene or Not to Intervene
.
Journal of Conflict Resolution
40
:
16

40
.
Google Scholar
Crossref
Search ADS
 
Snyder
J.
(
1984
)
Civil Military Relations and the Cult of the Offensive, 1914–1984
.
International Security
9
:
108

146
.
Google Scholar
Crossref
Search ADS
 
Snyder
J.
(
1991
)
Myths of Empire
.
Ithaca
:
Cornell University Press
.
Google Scholar
OpenURL Placeholder Text
 
Stam
A.
(
1996
)
Win, Lose or Draw
.
Ann Arbor
:
University of Michigan Press
.
Google Scholar
OpenURL Placeholder Text
 
Stevenson
D.
(
1988
)
The First World War and International Politics
.
New York
:
Oxford University Press
.
Google Scholar
OpenURL Placeholder Text
 
Sun
T.
(
1971
)
The Art of War
.
New York
:
Oxford University Press
.
Google Scholar
OpenURL Placeholder Text
 
Tilly
C.
(
1975
)
The Formation of National States in Western Europe
.
Princeton, NJ
:
Princeton University Press
.
Google Scholar
OpenURL Placeholder Text
 
Tocqueville
A.
(
[1835] 1945
)
Democracy in America
.
New York
:
Alfred Knopf
.
Google Scholar
OpenURL Placeholder Text
 
Tsebelis
G.
(
1995
)
Decision Making in Political Systems
.
British Journal of Political Science
25
:
289

326
.
Google Scholar
Crossref
Search ADS
 
Tsebelis
G.
(
1999
)
Veto Players and Law Production in Parliamentary Democracies
.
American Political Science Review
93
:
591

608
.
Google Scholar
Crossref
Search ADS
 
Ulam
A.
(
1974
)
Expansion and Coexistence
.
New York
:
Praeger Publishers
.
Google Scholar
OpenURL Placeholder Text
 
Wagner
H.
(
2000
)
Bargaining and War
.
American Journal of Politics
44
:
469

484
.
Google Scholar
Crossref
Search ADS
 
Waltz
K.
(
1979
)
Theory of International Politics
.
Reading, MA
:
Addison-Wesley
.
Google Scholar
OpenURL Placeholder Text
 
Wang
K.
Ray
J. L.
(
1994
)
Beginners and Winners
.
International Studies Quarterly
38
:
139

154
.
Google Scholar
Crossref
Search ADS
 
Wayman
F.
Singer
D.
Goertz
G.
(
1983
)
Capabilities, Allocations, and Success in Militarized Disputes and Wars, 1816–1976
.
International Studies Quarterly
27
:
497

514
.
Google Scholar
Crossref
Search ADS
 
Werner
S.
(
2000
)
Deterring Intervention
.
American Journal of Political Science
44
:
720

732
.
Google Scholar
Crossref
Search ADS
 
Woodward
D.
(
1971
)
David Llyod George, a Negotiated Peace with Germany and the Kuhlman Peace Kite of September 1971
.
Canadian Journal of History
6
:
75

93
.
Google Scholar
OpenURL Placeholder Text
 
Wright
Q.
(
1964
)
A Study of War
.
Chicago
:
The University of Chicago Press
.
Google Scholar
OpenURL Placeholder Text
 
© 2004 International Studies Association

Issue Section:
Original Articles
Download all slides
Advertisement
111 Views
27 Citations
View Metrics
×
Email alerts
Article activity alert
Advance article alerts
New issue alert
Receive exclusive offers and updates from Oxford Academic
Citing articles via
Web of Science (27)
Google Scholar
Crossref
Why They Fight: How Perceived Motivations for Military Service Shape Support for the Use of Force
Economic Decline, Social Identity, and Authoritarian Values in the United States
Rebel Fragmentation and the Recruitment of Child Soldiers
Economic Statecraft by Other Means: The Use and Abuse of Anti-Bribery Prosecution
Paradigms and Practice
Advertisement
Advertisement
Journals Career Network
Connect
Resources
Explore
Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide
Close
This Feature Is Available To Subscribers Only
Sign In or Create an Account
Close
This PDF is available to Subscribers Only
View Article Abstract & Purchase Options
For full access to this pdf, sign in to an existing account, or purchase an annual subscription.
Close
IssuesMore ContentAdvance articlesISQ Data ArchiveSubmitAuthor GuidelinesSubmission SiteOpen AccessPurchaseAlertsAboutAbout International Studies QuarterlyAbout the International Studies AssociationEditorial TeamEditorial BoardAdvertising and Corporate ServicesJournals Career NetworkSelf-Archiving PolicyDispatch DatesISA Portal