PROGRAM ON POLITICAL AND ECONOMIC CHANGE

INSTITUTE OF BEHAVIORAL SCIENCE

UNIVERSITY OF COLORADO

BOULDER, COLORADO, USA

80309-0487

Prepared for the Annual Conference of the International Studies Association, Toronto, Canada, March 18th-23rd, 1997. ©1996 Michael D. Ward. This research is supported by a grant from the National Science Foundation, SBR 9511577. We are grateful to Corey L. Lofdahl, Jordin S Cohen, and Michael Shin for their continuing support and help in this project.. The authors may be reached at mdw@isere.colorado.edu and at ksg@isere.colorado.edu.

This project's URL is adder.colorado.edu/~ibsgad/spacetime.html

DEMOCRATIZING FOR PEACE

MICHAEL D. WARD UNIVERSITY OF COLORADO

KRISTIAN S. GLEDITSCH UNIVERSITY OF MARYLAND

Version 3.01

08.11.96 06:24:50

DEMOCRATIZING FOR PEACE


MICHAEL D. WARD & KRISTIAN GLEDITSCH

8 November, 1996

ABSTRACT


We show that as contemporary polities become more democratic the probability that they will be involved in interstate war is substantially reduced. Moving from a weak to a strong democratic authority structure reduces the risk of war involvement by approximately 75%. Polities that move from a highly authoritarian structure to a highly democratic one reduce their risk of war by approximately 50%. To reach these conclusions, we developed and applied a logistic model linking authority characteristics and war involvement using data from the recently revised Polity III and from the Correlates of War data sets. The findings demonstrate that democratizing polities, even those going through mild reversals, are substantially less war prone than previously argued (e.g., Mansfield & Snyder, 1995a,b). In short, the risks of war are reduced by democratization, both in the long term and while societies undergo democratic change as well. We also show that reversals of democratization are associated with a slight increase in the risk of war.

INTRODUCTION


Do polities become more peaceful as they democratize? Or, is political change towards greater political democracy associated with increased likelihood of war? Research following Babst's

(1964) initial observation of an apparent absence of a war between democracies has produced a

considerable volume of empirical evidence for the proposition that democracies rarely if ever fight each other (Ma'oz and Abdolali 1989; Raknerud and Hegre 1997; Russett and Ma'oz 1992; Ma'oz and Russett 1993; Bremer 1992,1993). Despite substantial variation and healthy debate over the specific causal mechanism linking joint democracy and peace, there appears to be an emerging, if heady, consensus at least on the empirical existence of a "democratic peace" or an absence of war between democracies. In April of 1995 (p. 18) The Economist summarized this stylized fact by noting that "It takes two not to tango." In the words of Margaret Thatcher (1990 quoted in The Economist 1995), "…democracies don't go to war with one another." U.S. leaders from Bush to Clinton have echoed these sentiments.

Research on the democratic peace has focused almost exclusively on this specific dyadic proposition,

relating the behavior of pairs of countries to summed or overall aspects of their relationship. The units of analysis in these studies are dyads consisting of two countries, typically classified as jointly democratic, mixed or jointly autocratic. Some studies have also used measures of the mean or average level of political democracy in a dyad. However, when the unit of analysis is shifted to the individual state or the monadic level most studies have failed to find strong or significant differences between the war participation of democracies and that of non-democracies. While this conclusion has been questioned in some writings (Rummel 1983, 1985, & 1995; Ray 1995; Hewitt and Wilkenfeld 1996), the conventional wisdom remains that while democracies do not fight each other, they appear to be at least as war-prone as other political systems (see e.g. Singer and Small 1976; Russett and Ma'oz 1992, 1993; Morgan and Campbell 1991; Morgan and Schwebach 1992; Dixon 1993, 1994; although see also Gates et alia 1996). Why there is such a disjuncture between the national and bilateral versions of the idea that democracy leads to more peaceful foreign policies continues to be puzzling. One early study suggested that ethnocentrism was at part to blame (Chan 1984). Another (Spiro 1994) suggested that this finding is simply a result of having a very rare set of occurrences to study: unusual (i.e., rare) events involving rare structures (democracies) are especially rare. The issue of whether it is useful to aggregate initiators and targets into a single dyad is also viewed as problematic by some. The probability of a given country participating in a war in a given year may for example depend upon the number of politically relevant dyads (Ma'oz and Russett 1993; Lemke 1995), as well as the regime characteristics of the other countries in these relevant dyads. Relevant dyads are those where there exists an actual opportunity for war, typically gauged by geographical proximity and/or the density of political interactions.

The conundrum of why democracies apparently do not wage wars against one another but are not in general more pacific has failed to spawn a great deal of research. Thompson and Tucker (1996 forthcoming) note that in the past ten years there have been over one hundred empirical papers targeted at the peaceful joint democracy proposition. On the other hand, a much smaller number of scholars have looked into why democracies are not considered sui generis to be more pacific. Small and Singer (1976) initially suggested no difference exists between democracies and autocracies in the baseline probability of war involvement. Chan's (1984) research more systematically demonstrated this empirical regularity. Both of the two major explanations of the democratic peace proposition, encompassing both norms and institutions, would seem to have broader implications for the putative peaceful nature of the modern nation-state. Why this conundrum is so important is not only that it calls into some considerable question the theoretical underpinnings of the democratic peace hypothesis, but also that it undercuts the idealist policy prescription that is shared by such diverse political leaders as William Jefferson Clinton and Margaret Thatcher; namely, that widespread democratization will lead to widespread peace.

In this light, recent evidence has emerged that seeks - in part - to reconcile the differences between the bilateral and state-centric versions of Kant's long-standing thesis. Mansfield and Snyder (1995a,b) suggested that while democracies may eventually become more peaceful than autocratic societies, the transition phase in which a state is newly democratic is especially dangerous. Nascent democracies are thought to be quite likely to become involved in foreign conflict. This is a consequence of the fragile position of the new elite which is forced to appeal to nationalistic forces in attempts to solidify its tenacious positions. Accordingly, this leads to an increased propensity for war involvement. On the one hand, this argument is a plausible explanation for the divergent empirical findings, for if democracies are truly more peaceful, except during their very early stages of democratic institutionalization, the non-pacific nature of democracies in general might be explained away by war proneness in young democracies balanced by the pacific behavior patterns of mature democracies. On the other hand, the notion that democratizing societies are war prone and conflict laden does call into question whether democracies do in general have norms and institutions that prohibit conflict with other (similar) societies, and whether in an era of widespread proliferation of weapons of mass destruction, democratization will always be easily adjudged a good thing. The absence of strong monadic relationship between democracy and peace would not necessarily disprove the dyadic proposition that democracies rarely if ever fight each other: nor, would its presence put to rest all questions about the validity of the democratic peace. Yet, it would get us closer to the answers to these questions. These ambiguities between different levels of analysis suggest that existing research on the democratic peace proposition may have left unexplored many possible implications of domestic political structure for conflict behavior.

In this research, we examine in some detail the monadic aspects of the democratic peace proposition by revisiting the question of whether democratizing states are less war prone than autocratic states. We probe whether this may help to explain why democracies do not appear to be very likely to wage war with other democracies. In focusing on transitions, we follow the suggestion of Mansfield and Snyder (1995a,b) that non-mature democracies undergoing rapid change in institutional arrangements may be more prone to war involvement. We do not examine simultaneous factors that might resolve the differing relationships between political structure and conflict behavior found on the various levels of analysis. Nor, do we address the issue of the actual timing of changes of political or institutional arrangements and structures, something unexplored in the extant data on regime characteristics. Instead, we examine another set of neglected features of domestic political structure with possible implications for conflict behavior, namely effects of political transitions and regime changes.

In short, we re-examine whether states that are democratizing are more war prone. Complementarily we examine whether reversals of the democratization process, i.e., democratic states on a trajectory toward greater autocracy, will also by this logic prove to be less pacific.

THE DANGER OF DEMOCRATIZATION


Many studies of the democratic peace examine the impact of current domestic political structures on war participation. We focus primarily on the degree of change in these domestic political configurations. Three views can be identified of the possible consequences of democratization on war. Democratization might on the one hand involve changes towards greater extent or magnitudes of the properties found among democracies considered to prevent conflict from escalating into war. Typically, explanations of the democratic peace focus either on the existence of certain norms that facilitate resolution of conflicts before escalation to higher levels of violence-and ultimately to war-m y occur (Dixon 1993, 1994; Raymond 1994; Hewitt and Wilkenfeld 1996) or the role of various structural features more salient in democracies that make decision-makers more cautious in the use of force (Morgan and Schwebach 1992; Morgan and Campbell 1993; Bueno de Mesquita and Siverson 1995). These latter properties might include both formal institutional constraints on the executive decision to declare war such as the need for approval from a legislature or other independent governmental agency, as well as other, non-constitutional mechanisms that enhance the dependence of executive leaders on popular approval. A democratic political system with freely contested elections increases the executive leader's sensitivity to political risks since it may call into question the tenure of the ruling authorities. The fear that failure may lead to loss of political power in subsequent elections will ensure more restraints in the use of force. The distinction between the two views is commonly drawn as a difference between so-called structural and normative explanations of the democratic peace. The conflict between Hungary and Rumania over the ethnic Hungarian population in Transylvania is a recent example in which democratization may have prevented escalation to war. Accordingly, the likelihood of war should to be the greatest at low levels of democracy and diminish gradually as countries democratize.

Many who accept the notion that democracies do not fight each other still consider democratization and political change something of a double-edged sword. While stable, well-established democracies may not fight each other, a possibly rocky process of democratization or transition towards a fragile democracy need not necessarily imply that countries become more peaceful in the immediate run or as they democratize. The degree of democratic commitment or time since democratization may be important for explanations emphasizing the role of norms. It is quite plausible that some time may have to pass before democratic norms or informal institutions become sufficiently well established to have the effect of inhibiting conflicts. Some have suggested employing a turnover or two consecutive elections rule as classification criteria for democracy (Ray 1995a,b; Huntington 1991; and Przeworski et alia 1996). This perspective calls into question whether one should expect to find a decreasing likelihood of war attendant to democratic transitions.

Yet, much stronger and more pessimistic claims have been made with respect to the impact of democratization on the likelihood of conflict and war. Political instability and change in general is often considered to be associated with increased likelihood of conflict and possible subsequent escalation and war involvement. Rapid democratization might bring about weak regimes unable to establish effective control and political order. While these new regimes may be less repressive and permit greater political freedom than their precursors they would also be prone to instability and attempts by challenging groups on seizing political power. Political instability and disorder might in turn encourage attacks from other countries. These factors suggest that stable, highly autocratic regimes in fact might be less prone to conflict and escalation to war than regimes in transition. Similarly, while political leaders might become more dependent on popular support in democratizing regimes and young democracies, absent fully established institutions and civic traditions, these are often also considered to be prone populist policies emphasizing short term maximization of political support. Democratization is often considered to go together with nationalism, where in some cases certain groups, typically ethnic minorities, may become targeted as second-class citizens and made scapegoats for all existing problems. The transition towards democracy in the former Communist Countries in Eastern Europe has according to some observers left the population "free to hate". Similarly, war and aggression in the international realm often appears to enjoy widespread support amongst the population. The invasion of the Malvinas/Falkland Islands initially appeared to provide a boost to the popularity of the Argentinean military regime, for example. Populist politicians in young democracies might subsequently come to employ external aggression as a way of enhancing political support. This perspective might easily draw support from the example of the republics of former Yugoslavia, where increased political freedom and external aggression seem to have gone hand in hand.

The above discussion suggest two competing hypothesis on the effect of democratization on the likelihood of war. Fortunately, these hypotheses are non-nested and evidence can be collected on each of them at the same time. However, our belief is that the dangers of democratization are less likely to be borne out in the analysis. Why? First and foremost, we think that there is a logical, if not necessarily empirical, inconsistency in accounting for the evidence on the democratic peace proposition while at the same time maintaining that democratizing states are violence prone. It seems likely that if the notion of a dangerous transition were true it would mean that newly institutionalizing states would come to rely more on violence in their foreign policies precisely because they will have relied on it during an important evolutionary period of their development. That reliance will have served their internal political purposes, both in terms of retention of power but also in terms of how the institutions and norms were initially shaped. Thus, if violence comes to be evolutionarily dominant and successful in states undergoing the process of democratization, it would seem logical that the institutions and norms would emerge in a way to reflect this. From a logical perspective, evidence of a general (i.e., non dyadic) trend of pacific relations in democracies would help to explain why democracies don't appear to wage war with one another. If, however, democracies - even nascent ones - are more belligerent this raises some considerable doubt about the logical status of a democratic peace. This explains perhaps the great interest in the recent work of Mansfield and Snyder (1995a,b) by scholars working in the area of democratic peace.

Second, the first studies of this notion are, it must be remembered first efforts, and as yet do not have the accumulated, supporting replications that would benefit such law like generalizations. Thompson and Tucker (1996 forthcoming) detail a number of potential weaknesses in the Mansfield and Snyder (1995a,b) research design, and in one replication come to the conclusion that "democratization and war are independent" (p. 19 in draft). Similarly, Enterline (1995) found in another replication that democratization has a "negative impact on the likelihood of a state being on the initiating side of a dispute" (p. 197). All of these studies used slightly different case selection, variable definition, and statistical methods in coming to the three differing conclusions. What is clear is that more studies are needed and that the current state of the art findings on this question seem especially fragile. Our model seeks to gain more information about this question by directly posing it as a probability problem: are states undergoing transitions, especially democracies, more likely to be involved in interstate wars? We choose to model this problem as logistic probability equation. Basically, we want to know the probability of a polity being involved in an interstate war at a given point in time based on its authority characteristics and how they have changed over the prior ten year period. Thus, we have a logistic equation: , where represents the authority characteristics (i.e., {Democracy, Democracy minus Autocracy, Executive Constraints} and superscripts {d,m,v}refer respectively to the direction, magnitude, and variance of change in the values over the prior ten year period. Thus, the probability of war for a given observation has the following logistic form: (1)

This simple model is quite different in some respects from most other efforts. First, our choice of independent variables focuses entirely on the transition process. Many other studies have only changes exceeding some specific magnitude or have used dichotomous categories (i.e., change versus no-change) over some aggregate time period. These categorizations of political change risk being fairly arbitrary. The mere existence of change, however, is not by itself an adequate measure of democratization. Unlike those who conceive of democracy as an "all-or-nothing" affair we think of democratization and political change as movements (though not necessarily smooth movements) along a continuum and examine the direction and first two moments of the distribution of change. Following the notion of executive constraints as conducive to peace greater changes towards democracy should in our view be associated with decreasing likelihood of war while minor steps towards greater political democracy and cosmetic changes should exercise less effect on the avoidance of war. We explicitly take the magnitude of change into account. Political change or processes of democratization can however take many different forms and need not proceed in a unidirectional or linear fashion. Much of the literature on democratization has claimed that large or rapid change towards democracy often tends to be associated with greater instability and higher risk of democratic reversals (e.g. Huntington 1984, 1991). These views seem to imply that gradual changes involve a smaller likelihood of war than rapid democratization. Furthermore, greater variability of political change within a country's political system would be associated with higher likelihood of war. Accordingly, we focus on the moments of the distribution of change in authority characteristics, including the direction, magnitude, and variance of change, along with the actual values, and an interaction between the direction and magnitude of change (i.e., the actual change).

Such a specification is novel in this literature, but allows us to glean more details about the democratization process and how, if at all, it relates to the probability of being involved in a war. Unstable democratic transitions would, for example, exhibit a high degree of variance in their democratization scores. Countries moving quickly would exhibit a large value for the magnitude of change, but slow democratizing societies would exhibit a small magnitude of change, with low variance, over time. The next section describes the data, methods, and results from an empirical examination of the issue of whether democratization increases the risk of war.

EMPIRICAL EVIDENCE


For assessing authority characteristics, we employ the 5.1996 version of the Polity III data set. A more complete description of these data may be found in author's work (1996), Gurr (1974), Gurr, Jaggers, and Moore (1990), Gurr, Jaggers, and Moore (1989), and Jaggers and Gurr. (1995). These data are more up-to-date that those available previously, having information on polities through the 1994 period. In addition, a number of important revisions have been undertaken in this version of the data. These are detailed in the web site that supports the data (URL:xxx). Most important among these new codings is the corrected democracy scores of 19th Century Monarchies and the inclusion of transition codes. Table 1 shows in tabular format the coding rules for assigning democracy and autocracy scale values in the Polity III framework.Table 1. Authority Characteristics and Democracy and Autocracy Scale Values in Polity III, adapted from Jaggers and Gurr (1995)

Authority Dimensions
Democracy "Points"
Autocracy "Points"
Competitiveness of Political Participation [1,7] (PARCOMP)
Competitive
3
0
Transitional
2
0
Factional
1
0
Restricted
0
1
Suppressed
0
2
Not Applicable
0
0
Regulation of Political Participation [1,5]
Regulated
0
0
Factional or Transitional
0
0
Factional/Restricted
0
1
Restricted
0
2
Unregulated
0
0
Competitiveness of Executive Recruitment [1,4]
Election
2
0
Transitional
1
0
Selection
0
2
Openness of Executive Recruitment [1,7]
Election
1
0
Dual: Hereditary&Election
1
0
Dual: Hereditary&Designation
0
1
Closed
0
1
Constraints on Chief Executive [1,7] (C)
Parity or Subordination [7]
4
0
Intermediate 1: between & [6]
3
0
Substantial Limitations [5]
2
0
Intermediate 2: between & [4]
1
0
Slight to Moderate Limitations [3]
0
1
Intermediate 3: between & [2]
0
2
Unlimited Power [1]
0
3
Total Scores
Democracy "points"
Autocracy "points"

From these data we are able to calculate democracy scale values {D}; democracy scale values minus autocracy scale values {D-A); the extent of executive constraints on the executive decision maker {C}; and the direction-coded positively for movements toward democracy-and magnitude of change as well as the variance of over the prior ten years in the above three variables, represented respectively by d, m, and v superscripts.

In addition we have used the most recent version of the Correlates of War data set to ascertain whether each polity was involved in an interstate or extra-systemic war during each year from 1815-1992. We call this variable war involvement, denoted {W}. The Correlates of War project defines an interstate war as violent conflict between two or more members of the international system involving at least 1,000 annual battle deaths. The extra-systemic wars are conflicts involving territories or groups outside the state system, a category largely comprised by colonial wars. We have coded countries as being at war in years in which they were classified by the Correlates of War data set as parties to one or more interstate or extra-systemic wars. We obtained these data from the Inter-University Consortium for Political and Social Research, housed at the University of Michigan. Thus, data from the Polity III (May 1996 version) and the Correlates of War data sets were merged to provide annual observations on each independent polity over the period from 1815-1992. Thus, with these data we estimate the model given in equation (1).

Table 2 portrays the results of three logit estimations of the basic equation. These results are quite strong in suggesting that both the direction and the magnitude of change are important in determining the log-odds of a polity being at war, but that their effects are of opposite sign. Thus, the direction and magnitude of change toward democracy are important predictors of the log-odds ratio of being at war. When the magnitude of democratic change increases, the log-odds falls. When the direction of democratic transition is positive, toward democracy, the log-odds ratio rises.

Table 2. Three Sets of Empirical Estimates of Logistic Models of the Probability of War based on Authority Characteristics

Democracy

Scale

Democracy minus

Autocracy

executive

Constraints
Independent

vARiABLES
Estimated

Coefficient
Standard

Error
estimated

Coefficient
Standard

Error
estimated

Coefficient
Standard

Error

Intercept

-2.659

.05

-2.644

.044

-2.870

.076

Current

value


.0167

.01

.0003

.005

.053

.016

direction of

Change

.353

.132

.208

.097

.289

.140

magnitude

of Change

-.111

.034

-.029

.014

-.118

.047

variance of

Change

.032

.019

.0018

.005

.006

.039

Direction

Magnitude


-.070


.033


.006


.016


.075


.043

Likelihood

ratio 2

15.9

p = .007

5.9

p = .320

23.1

p = .0003
Number of Cases

10,700

10,700

10,700

What does this really mean about the probability of being at war? Log-odds ratios are not especially easy to write about in substantive terms, so we prefer to interpret our results in terms of probabilities. Thus, a somewhat more intuitive way of looking at these results is to calculate the probabilities for the dependent variable as a function of some scenario involving the independent variables. We look at the process of democratization and examine scenarios in which there is a transition of the authority characteristics over a ten year period. We look at both democratization, as well as "reversal" in which a democratic state becomes less democratic and/or more authoritarian. We show these results graphically. Figure 1. Estimates of the Declining Risk of War Attendant to Democratization

Figure 1 teases out the estimates of the probability of war involvement for various combinations of the independent variable in cases of democratization. This yields the marginal changes in probabilities for changes in the underlying process. A smooth transition has unit variance while a rocky transition is given a variance score of three. As this figure shows in the left column of graphics, democratization, whether toward mild or strong democratic degrees of government, is accompanied by reduction, not increase, in the risk of war. On the other side of the figure in the right column, reversals and retrenchments are shown to be accompanied by increased risks of war involvement. However, these are increases are much more modest than the benefits of further democratization. It must be remembered Figure 1 only deals with the case of societies that have some degree (i.e., at least a democracy score of 1) of democratization. On the whole, these results provide strong evidence of monadic effect of the process of democratization: it reduces the probability that a country will be involved in a war.

Figure 2 shows the decreasing probability associated with moving from a prototypic autocracy to a prototypic democracy within a ten year period. Although probability of war involvement does not decrease linearly, it does decrease monotonically so that over the entire range of democracy minus autocracy values, there is about a fifty percent reduction. Thus, during the democratic transition, at every point along the way, as well as at the end points, there is an attendant reduction in the probability of a polity being at war.Figure 2. The Probability of War as Polities Democratize

It has been argued that institutional constraints are theoretically important in translating the effect of democracy into foreign policy (Bueno de Mesquita, Siverson, and Woller 1992; Siverson 1995). If we break the idea of democracy into its major component parts, it has been shown that the degree of executive constraints dominate the democracy and autocracy scales (Gleditsch and Ward, 1996). Accordingly, we demonstrate that moving toward stronger executive constraints also yields a visible reduction in the risk of war. Figure 3 illustrates that if a polity moves from having no constraints on the executive decision maker (i.e., a value of 1.0 on the x-axis in the above graphic) to a position of parity or subordination with the legislature (a value of 7.0), the probability of being involved in an interstate war is reduced from about 9% to about 3%. This sheds light on what exactly it is about democratization that may reduce the probability of war: shared power with popularly elected officials. To the extent that changes toward democracy bring with them constraints on the executive branch of government, the attendant reduction in the risk of war appears quite robust.
Figure 3. The Probability of War as Executive Constraints are Established


CONCLUSION


Our results show that the process of democratization is accompanied by a decrease in the probability of being involved in a war, either as a target or initiator. These results were obtained with a more up-to-date (and corrected) data set than has been used in earlier studies of this question. They also focused more clearly on the process of transition. In comparison to earlier studies that looked only at the magnitude of change in authority characteristics, we looked at the direction and the smoothness of the transition process.

However, it is necessary to temper these results. First, we have focused on aspects of the transition process to the exclusion of other important aspects of domestic and international politics that may affect the probability of being at war. We do not suggest to have presented a complete model of this process. Although our model passes conventional statistical muster, it should not be used as a predictive model of war. We can correctly predict over 90 percent of the correct outcomes, but all of our correct predictions are for the absence of war. Cases in which there was a war are rare, of course, and the predicted probabilities associated with these cases are higher than in other cases, but do not approach the 0.50 level that would be a minimum for predicting the presence of a war.

Moreover, we understand that having focused on the monadic aspects of changes in regime characteristics ignores the strongest findings of the democratic peace literature. As a result we have ignored the potentially powerful hypotheses that expose the effects of joint democracy, common interests and alliances, the balance of capabilities, and of economic interdependence and international commerce. Nor do we investigate the spatial aspects of these relationships to ascertain whether contiguity and common spatial relationships are important. However, by focusing on the transition argument alone, we lose any traction that might be gained from these notions as well. In the same way, by focusing on democratization processes, we avoid generalizing about the causes of war. All of these aspects are undoubtedly important, and each of them is receiving considerable attention in the literature. Our purpose here is not to bring any evidence to bear on these ideas, instead focusing on the issue of whether democratizing states are more war prone than other states. In that limited way, we find no evidence of a heightened propensity for war involvement, but rather a diminished likelihood of being in a war during the transition.

There remain, however, strong images that support the notion of a dangerous democratization. One such image at the time of this writing is that of an incapacitated Russian leader faced with a welter of nationalist competition inside Russia. The small bandage placed over the gaping wound in Grozny may be perceived as a opportunity that motivates a democratic consolidation. The stylized fact of the dire conflict in the Balkans is often offered as evidence that democratization is not only fragile but dangerous. Recent events in Eastern and Central Europe after the breakup of the Yugoslavia and the Soviet Empire have spawned considerable speculation about the dangers of democratization (e.g., Mansfield and Snyder, 1995a,b). Typical examples offered are Serbia (Yugoslavia), Croatia, Armenia and Azerbaijan. While these regimes may enjoy considerable popular support, it seems dubious whether they might be accurately characterized as democracies. These countries did not exist as countries prior to 1991, and it thus hard to say how democratization-even if it is underway-can be separated from the process of state-building and institutionalization. It is appropriate to question whether these examples are really typical at all. Many conflicts in the area have been predicted that have failed as yet to materialize (e.g., Hungary versus Romania, Russia versus the Ukraine, Poland versus Lithuania, Russia versus the Baltics). Thus, even the stylized facts are quite complicated and do not necessarily lead to the conclusion that newly emergent democracies are war prone. Not all political change is toward greater institutionalization of democratic norms and practices; not even all post Cold War change in Europe.

The democratic peace hypothesis has maintained that democracies don't go to war against other democracies. The dangerous democratization hypothesis has suggested that emergent democracies may be quite prone to international violence, largely as a result of "deformed" institutional forces. Our research does not rule out such an outcome, but it does demonstrate that on the whole the process of democratization proceeds in a way to reduce the risk of war. It also shows that there is little statistical evidence to suggest that the transition of a state toward more democratic practices makes it more dangerous as a threat to international peace.

REFERENCES


Babst, Dean V. 1964. "Elective Governments-A Force for Peace." The Wisconsin Sociologist 3:9-14.

Bremer, Stuart A. 1992. "Dangerous Dyads." Journal of Conflict Resolution 36 (June):309-41.

Bremer, Stuart A. 1993. "Democracy and Militarized Interstate Conflict, 1816-1965." International Interactions 18:231-49.

Bueno de Mesquita, Bruce and Randolph Siverson. 1995. "War and the Survival of Political Leaders: A Comparative Analysis of Regime Type and Accountability." American Political Science Review 89 (December):841-55.

Bueno de Mesquita, Bruce, Randolph Siverson and Gary Woller. 1992. "War and the Fate of Regimes." American Political Science Review 86 (September):638-46.

Chan, Steve 1984. "Mirror, Mirror, on the Wall…Are the Freer Countries More Pacific?" Journal of Conflict Resolution 28 (December): 617-48.

Dixon, William J. 1993. "Democracy and the Management of International Conflict." Journal of Conflict Resolution 37 (March):42-68.

Dixon, William J. 1994. "Democracy and the Peaceful Settlement of International Conflict." American Political Science Review 88 (March):14-32.

Enterline, Andrew J. 1995. "Correspondence on 'Driving While Democratizing'." International Security 20 (Summer): 183-196.

Gagnon, V. P., Jr. 1994/5. "Ethnic Nationalism and International Conflict: the Case of Serbia." International Security (Winter):130-166.

Gates, Scott, Torbjørn L. Knutsen, and Jonathon W. Moses. 1996. "Democracy and Peace: A More Skeptical View." Journal of Peace Research 33 (February):1-10.

Gleditsch, Kristian S. and Michael D. Ward. 1996. "Double Take: a Re-examination of Democracy and Autocracy in Modern Polities." Presented to the annual meetings of the Peace Science Society (International), Houston, Texas, USA.

Gurr, Ted Robert. 1974. "Persistence and Change in Political Systems, 1800-1971." American Political Science Review 74 (December):1482-1504.

Gurr, Ted Robert, Keith Jaggers, and Will H. Moore. 1990. "The Transformation of the Western State: The Growth of Democracy, Autocracy, and State Power since 1800.'' Studies in Comparative International Development 25 (Spring):73-108.

Gurr, Ted Robert, Keith Jaggers, and Will H. Moore. 1989. "Polity II Codebook." Unpublished manuscript, University of Colorado.

Hewitt, J. Joseph and Jonathan Wilkenfeld. 1996 forthcoming. "Democracies in International Crisis." International Interactions (in press).

Huntington, Samuel P. 1984. "Will More Countries Become Democratic?" Political Science Quarterly 99 (Summer):195-8.

Huntington, Samuel P. 1991. The Third Wave: Democratization in the Late Twentieth Century. Oklahoma: Norman, OK.

Jaggers, Keith and Ted Robert Gurr. 1995. "Tracing Democracy's Third Wave with the Polity III Data." Journal of Peace Research 32 (November):469-482.

Lake, David. 1992. "Powerful Pacifists." American Political Science Review 86 (March):24-37.

Lemke, Douglas. 1995. "The Tyranny of Distance: Redefining Relevant Dyads." International Interactions 21(no. 1):23-8.

Mansfield, Edward D. and Jack Snyder. 1995a. "Democratization and the Danger of War." International Security 20 (Summer): 5-38.

Mansfield, Edward D. and Jack Snyder. 1995b. "Democratization and War." Foreign Affairs 74 (May/June): 79-97.

Ma'oz, Ze'ev. 1989. "Joining the Club of Nations: Political Development and International Conflict." International Studies Quarterly 33 (June):199-231.

Ma'oz, Ze'ev and Nasrin Abdolali. 1989. "Regime Types and International Conflict, 1816-1976." Journal of Conflict Resolution 33 (March):3-35.

Ma'oz, Ze'ev and Bruce Russett. 1993. "Alliance, Contiguity, Wealth and Political Stability: Is the Lack of Conflict between Democracies a Statistical Artifact?" International Interactions 17 (no. 3):245-67.

Morgan, T. Clifton and Sally Campbell. 1991. "Domestic Structure, Decisional Constraints, and War." Journal of Conflict Resolution 35 (June):187-211.

Morgan, T. Clifton and Valerie Schwebach. 1992. "Take Two Democracies and Call Me in the Morning: A Prescription for Peace?" International Interactions 19 (no. 4):305-20.

Oneal, John R. and James L. Ray. 1996. "New Tests of the Democratic Peace Controlling for Economic Interdependence, 1950-85," unpublished manuscript, University of Alabama, USA.

Raknerud, Arvid and Håvard Hegre. 1997 forthcoming. "The Hazard of War: Reassessing Evidence of the Democratic Peace." Journal of Peace Research.

Ray, James L. 1995a. "Global Trends, State-specific Factors and Regime Transitions, 1825-1993." Journal of Peace Research 32 (February):49-63.

Ray, James L. 1995b. Democracy and International Politics. Columbia: University of South Carolina Press.

Raymond, Gregory. 1994. "Democracies, Disputes, and Third-Party Intermediaries." Journal of Conflict Resolution 38 (March):24-42.

Rousseau, David L., Christopher Gelpi, Dan Reiter, and Paul K. Huth. 1996. "Assessing the Dyadic Nature of the Democratic Peace, 1918-88." American Political Science Review 90 (September): 512-33.

Rummel, Rudolph J. 1983. "Libertarianism and International Violence." Journal of Conflict Resolution 27 (March):27-72.

Rummel, Rudolph J. 1985. "Libertarian Propositions on Violence within and between Nations." Journal of Conflict Resolution 29 (December):419-55.

Rummel, Rudolph J. 1995. "Democracies ARE Less Warlike than Other Regimes." European Journal of International Relations 1 (December):457-79.

Russett, Bruce and Ze'ev Ma'oz. 1993. "Normative and Structural Causes of Democratic Peace, 1946-1986." American Political Science Review 87 (September):624-38.

Siverson, Randolph J. 1995. "Democracies and War Participation: In Defense of the Institutional Constraints Argument." European Journal of International Relations 1 (December):481-9.

Small, Melvin and J. David Singer. 1976. "The War Proneness of Democratic Regimes." Jerusalem Journal of International Relations 1(no. 1):50-69.

Spiro, David E. 1994. "The Insignificance of the Democratic Peace." International Security 19 (Fall):50-86.

The Economist. April 1st 1995. "The Politics of Peace," pp. 17-18.

Thompson, William R. and Richard M. Tucker. 1996. "A Tale of Two Democratic Peace Critiques." Journal of Conflict Resolution (in press).