DOMESTIC STRUCTURE, LEARNING, AND THE DEMOCRATIC PEACE: AN AGENT-BASED COMPUTATIONAL SIMULATION A. Maurits van der Veen University of Georgia maurits@uga.edu David Rousseau University of Pennsylvania rousseau@sas.upenn.edu August 24, 2004 Draft, not for citation Abstract This paper uses agent-based modeling to study the impact of domestic political structure on the evolution of a democratic peace. We show that democratic peace can emerge even with a very limited set of basic assumptions about the relationship between levels of domestic opposition and the costs of initiating conflict. In addition, we find that learning among democracies and autocracies alike reduces both the incidence of international conflict and the rate at which the international system consolidates into fewer states. Finally, we show that introducing even a fairly weak mechanism for the punishment of ‘pariah’ states (autocracies that attack democracies) suffices to eliminate any semblance of a democratic peace: democracies become more likely to attack not just autocracies but also other democracies. To be presented at the annual conference of the American Political Science Association, Chicago, IL, 5 Sept. 2004 Domestic structure, learning, and the democratic peace DOMESTIC STRUCTURE, LEARNING, AND THE DEMOCRATIC PEACE: AN AGENT-BASED COMPUTATIONAL SIMULATION “Democracies don’t attack each other” — Bill Clinton, State of the Union, 1994 "We have no desire to dominate, no ambitions of empire. Our aim is a democratic peace" — George W. Bush, State of the Union, 2004 Introduction Although fifteen years has elapsed since Levy argued that democratic peace is “the closest thing we have to an empirical law” in international relations (Levy, 1989: 88), the causal mechanisms behind the observed pattern remain elusive. However, despite our lack of understanding of the empirical patterns, the democratic peace has become the cornerstone of American foreign policy in the post-Cold War era. This increases the urgency and importance of investigating the causal mechanisms that may explain the democratic peace. In this paper, we present one approach to doing so, using the tool of agent-based computational simulation. We present a computational model of international conflict built on Cederman’s GeoSim model (Cederman, 2003, 2001a; Cederman & Gleditsch, 2002), into which we introduce important roles for domestic structure and for learning. Although the analysis in this paper is largely preliminary, early runs of the model produce a number of interesting findings. First, basic assumptions about the influence of domestic opposition on a state’s ability (or willingness) to initiate conflict suffice to produce a pattern resembling the democratic peace. Second, in mixed conflict dyads, democracies are more 1 Domestic structure, learning, and the democratic peace likely to escalate to war than are autocracies if they are the challenging state which initiated the conflict, but they are less likely to escalate if they are the target state. Third, autocracies and democracies alike learn to prefer attacking weaker states. Fourth, learning helps reduce the incidence of international conflict as well as slow down the rate of consolidation of states. Finally, and surprisingly, introducing a mechanism by which autocracies that attack democracies are punished not only dramatically increases the incidence of international conflict, but also completely eliminates the democratic peace, generating a system in which democracies are noticeably more likely to be at war, regardless of the regime type of their adversary. This finding should give pause to those who advocate trying to create a democratic peace through preventative war against autocracies: doing so may not simply be ineffective; it may indeed erode the existing democratic peace. Explaining the democratic peace Rousseau’s extensive empirical study (forthcoming) reveals a complex causal process linking domestic politics to international behavior, and finds both monadic and dyadic causal factors informing the democratic peace. Specifically, Rousseau shows that democratic states are constrained at the initiation phase by the presence of domestic political opposition that can punish a chief executive for using military force. However, the use of force by an international opponent reduces domestic opposition to the escalation of conflict once democracies are engaged in militarized crises. The chief 2 Domestic structure, learning, and the democratic peace exception to this process is the dyadic democratic peace: even when they enter a rare crisis, democracies are less likely to escalate against other democracies. The importance of domestic opposition emerges in different forms also from other recent work on the democratic peace (Fearon, 1994; Bueno de Mesquita, Smith, Siverson, & Morrow, 2003; Gelpi & Griesdorf, 2001). It remains somewhat unclear, however, whether domestic opposition factors by themselves are sufficient to create a democratic peace such as we find it in the empirical data. One possibility is that they suffice to sustain a democratic peace but may not generate one by themselves. This raises the issue of the origins of the democratic peace. Cederman (2001a) has shown that the democratic peace may have evolved along the lines originally suggested by Immanuel Kant, by states learning to cooperate peacefully. Interestingly, however, he finds that such learning is not limited to purely democratic dyads: mixed and purely autocratic dyads also appear to learn to cooperate more peacefully. The present paper investigates these issues systematically by testing the implications of different causal mechanisms for the evolution of a democratic peace. In particular, we examine (1) the empirical patterns that result from introducing recent theoretical insights about the role domestic opposition plays in democracies as well as autocracies; (2) the impact of different rates of learning from neighboring states on empirical outcomes regarding international conflict; and (3) the possible contribution of a strategy of aggressively punishing autocracies that violate peaceful coexistence to the creation of a democratic peace. 3 Domestic structure, learning, and the democratic peace Modeling War and Peace: DomGeoSim A recurring problem in the democratic peace literature is the limited number of cases available to us. We have only one ‘run’ for our world, making it very difficult to test the myriad counterfactuals that arise when theorizing different causes for the democratic peace. One way around this restriction is to generate additional ‘runs’, by studying the evolution of an artificial world in which states interact, fight, and conquer. We apply this approach to the study of the democratic peace, by building an agent-based model of interstate conflict which we can run as often as necessary, while subjecting it to fine-grained changes in the parameters that govern its world. Additional ‘world histories’ thus generated cannot, of course, tell us anything about how the real world works, but they can tell us a lot about the validity of our theories for explaining real world patterns. For example, if a theory posits a certain causal mechanism as the driver behind an empirical pattern, we can program a simulation in which we can vary that causal mechanism, keeping all other aspects of the world constant. If the output remains the same nevertheless — and, importantly, if the other components of the model correctly incorporate any additional assumptions or specifications of the theory — this casts serious doubt on the causal mechanism in question. As with all methods of investigation, computer simulations have strengths and weaknesses.1 On the positive side of the ledger, five strengths stand out. First, as with 1 For a more extensive discussion of strengths and weakness of agent-based modeling, see (Axtell, 2000; Johnson, 1999; Rousseau, 2004). 4 Domestic structure, learning, and the democratic peace formal mathematical models, simulations compel the researcher to be very explicit about assumptions and decision rules. Second, simulations allow us to explore extremely complex systems that often have no analytical solution. Third, simulations resemble controlled experiments in that the researcher can precisely vary a single independent variable (or isolate a particular interaction between two or more variables). Fourth, as suggested above, while other methods of inquiry primarily focus on outcomes (e.g., do democratic dyads engage in war?), simulations allow us to explore the processes underlying the broader causal claim (e.g., how does joint democracy decrease the likelihood of war?). Fifth, simulations provide a nice balance between induction and deduction. While the developer must construct a logically consistent model based on theory and history, the output of the model is explored inductively by assessing the impact of varying assumptions and decision rules. On the negative side of the ledger, two important weaknesses stand out. First, simulations have been criticized because they often employ arbitrary assumptions and decision rules (Johnson 1999, 1512). In part, this situation stems from the need to explicitly operationalize each assumption and decision rule. However, it is also due to the reluctance of many simulation modelers to empirically test assumptions using alternative methods of inquiry. In our model, we address this problem by using assumptions and interaction rules based on the empirical findings in Rousseau (forthcoming). Second, critics often question the external validity of computer simulations. While one of the strengths of the method is its internal consistency, it is often unclear if the simulation captures enough of the external world to allow us to generalize from the artificial system we have created to the real world we inhabit. However, this shortcoming is hardly limited 5 Domestic structure, learning, and the democratic peace to agent-based modeling: all models, even the most thickly descriptive ones, abstract from the real world. The more relevant question is whether the elements essential to a particular theory have been incorporated. As often as not, criticism that a model is missing some crucial feature indicates that the theory it attempts to test has been incompletely specified. Our model builds on Lars-Erik Cederman’s GeoSim model (Cederman, 2003), whose code he generously made available to us. Like his, our model is programmed in Java, using the Repast simulation toolkit (see http://repast.sourceforge.net). Although the internal workings of the model have been restructured to allow us, among others, to introduce domestic political structure and learning process, much of the set-up remains the same. Cederman has used his model to explore many different aspects of interstate conflict (e.g. Cederman, 1997). Indeed, he has previously used it to investigate the democratic peace (Cederman & Gleditsch, 2002; Cederman, 2001b). Examining the implications of strategic tagging, regime influenced alliance formation, and collective security for the emergence of a peaceful liberal world, he found that these three causal mechanisms, first proposed by Kant over two centuries ago, could collectively increase the probability of the emergence of a liberal world. While Cederman’s innovative research makes an important contribution to the literature, for our purposes it has certain important limitations. For example, while Cederman’s model of the democratic peace illustrates conditions under which a stable democratic peace can emerge, he assumes that the dyadic democratic peace exists (Cederman, 2001b: 480). In his simulation, democratic states cannot attack other democratic states by definition. In contrast, in our model democracies can (and do) fight 6 Domestic structure, learning, and the democratic peace each other; the question explored is whether over time democracies might learn to stop fighting each other. In order to maximize the flexibility of our adaptation of GeoSim, we have reprogrammed the internal structure of the model, so that configurations other than a straightforward rectangular grid can be modeled (Cederman himself is moving in this direction too). In addition, we have turned many features of the model that were hardwired in the original code into parameters that can be changed by the user. The obvious risk here is that the number of parameters can become bewildering to the user. On the other hand, however, it dramatically increases our ability to perform robustness checks by testing how dependent our findings are on different, apparently unrelated, parameters. To help keep the parameters manageable, we have produced a parameter dictionary, which is attached as an appendix. In order to reflect its close relationship to GeoSim, we will refer to our model below as DomGeoSim.2 Conflict in the DomGeoSim world The model world consists of a population of state agents that interact on a square 50x50 lattice which does not wrap around. Each state agent is composed of one or more of the 2500 territory squares, and possesses certain attributes that are modified through interaction with other agents in the landscape. In particular, each state has a certain wealth, a domestic structure (autocratic vs. democratic, domestic opposition levels), and a 2 For a detailed description of GeoSim, see (Cederman, 2003, 2001b). Our model was programmed by Maurits van der Veen. While DomGeoSim can produce results very similar to those of GeoSim with the appropriate parameter settings, the results will not be identical, due to the correction of a few minor problems which do not alter Cederman’s substantive findings. We would like to thank Lars-Erik Cederman for generously providing the original code to us. 7 Domestic structure, learning, and the democratic peace set of behavioral rules. The individual territory squares are considered ‘provinces’ and international conflict centers around disputes over these provinces. Thus, the model is in reality a network in which provinces are connected to a state capital and states that have adjoining provinces can interact with one another. The initial number of states is a model parameter, and was set to 100 for all simulations reported here. There are three types of states: autocracies, democracies, and pariah states. Pariah states are autocracies that have initiated a dispute with a democracy. Pariah status wears off over time, but while it lasts it has implications for the likelihood that a pariah state will become enmeshed in a conflict. In other words, pariah states behave identically to autocracies, but democracies may behave differently towards them. An early ‘state of the world’ snapshot is shown in figure 1. Democracies at peace are light blue and autocracies at peace are light yellow. When they go to war, their color becomes darker. Pariah states are orange, and become darker red when they go to war. Two contiguous states at war have a bright red border drawn between them. Normal (notat-war) borders are drawn in black. [Figure 1 about here] States can also ally with other states that feel threatened by the same (larger) states. Allies need not be contiguous with one another — merely contiguous with the threat they are allying against. This rule permits two-front conflicts that are common in the history of international relations (e.g., the Polish-French alliance versus Germany prior to World War II). For example, the small single-province state near the top of figure 8 Domestic structure, learning, and the democratic peace 1 could seek to ally against its warring autocratic neighbor to the south with the larger democratic state to its right which is already fighting that autocratic neighbor.3 A parameter governs whether an alliance will have operational implications — i.e. whether anyone comes to the aid of an ally at war. In our simulations here, there is a 50% chance that a state will join in a war being fought by an ally against the state they are jointly allied against (i.e. buck-passing does occur).4 Each simulation is run for 5000 iterations or rounds. If we think of each iteration as a period of time on the order of a calendar month, then each run of the simulation models the rise and spread of democracy across 400 years of human history. In each iteration of the simulation, agents must complete four tasks: 1) tax their provinces (at a rate of 2.5% of resources available in the default simulation); 2) allocate a portion of the tax revenue to battle fronts along the borders (with moveable resources limited to 50% of tax revenue in the default simulation); 3) update the alliance portfolio by adding or subtracting alliance partners; and 4) decide whether to enter into any new disputes with neighbors and how to handle ongoing disputes. After a dispute reaches the level of war, war damages are subtracted from resources available at the battle front. If the balance of power shifts decisively on the front, the province in dispute falls to the attacker. Victory is probabilistic once the attacker achieves a 3:1 advantage on the front (another parameter setting). 3 Although several neighborhood types are available in the model (e.g., von Neumann (only the 4 neighbors located North, South, East, and West), hexagonal (6 of the 8 possible neighbors in an alternating pattern from row to row), and Moore (all 8 neighbors including diagonals)), all the results reported are based on the von Neumann neighborhood used in the GeoSim model. 4 Waltz (1979) argues that buckpassing and chainganging help make the multipolar world more conflictual than a bipolar world. Christensen and Snyder (1990) use the offense-defense balance to explain when each phenomenon is likely to occur. Specifically, they argue that chainganging is more likely in a offense dominant world and buckpassing is more likely in a defense dominant world. The hypotheses can be explored in the simulation by varying the probability that allies aid the state (i.e., P_aidAllies), the margin of power needed for victory (i.e., VictoryRatio), and the costs of war (i.e., F_warCost). 9 Domestic structure, learning, and the democratic peace Cederman’s models address certain issues of domestic politics. For example, as states grow they add provinces composed of conquered territory. The provinces are taxed and border provinces receive additional resources to help defend the state. However, Cederman largely black boxes domestic politics because his theoretical interests have lain elsewhere. In DomGeoSim, we incorporate domestic politics into the model in three ways: 1) the conflict is divided into phases to allow (but not require) domestic politics to affect each phase differently; 2) state behavior is a function of traits that can evolve over time; and 3) domestic political opposition can influence the decision to use engage in interstate conflict. [Figure 2 about here] In DomGeoSim, interactions between a challenger and a target state are structured according to a fairly simple bargaining decision tree used extensively in the formal modeling literature.5 As displayed in Figure 2, the bargaining game has four phases: 1) peace or status quo; 2) dispute; 3) crisis; and 4) war. Peace is the baseline condition in which agents face no threats or violence. A dispute phase begins when a challenging state stakes a claim on a province of the target state. The dispute ends when the target state either rejects the demand, concedes to it, or the dispute diffuses peacefully. A target state can also do nothing, in which case the dispute remains ongoing. A crisis phase begins when either the target rejects the demand of the challenger or a challenger becomes impatient with the target’s delaying tactics. The crisis ends when the 5 Kinsella and Russett (2002: 1046) argue that empirical models are beginning to test the stages of conflicts employed in formals models. Thus, introducing stages into agent-based simulation, as we do here, may facilitate comparative analysis of formal models, large N quantitative studies, and computer simulations. 10 Domestic structure, learning, and the democratic peace challenger either escalates the conflict to war, concedes, or the crisis diffuses peacefully. In addition, a challenging state can do nothing, in which case the crisis remains on-going. War can proceed for many rounds and ends in victory, defeat, or a draw. Once war ends, the dyad returns to the peace phase, assuming loss of the territory did not have implications for the ability of the losing state to survive as a territorially contiguous state. State behavior and learning DomGeoSim permits a wide variety of decision strategies and social learning. Each agent has a set of traits that evolve by learning from more successful neighbors, as well as through a certain amount of random experimentation. The evolution of state behavior is inspired by the literature on genetic algorithms.6 The behavior of agents is determined by the eleven traits summarized in table 1. The first three traits determine the role of power in the decision making process. Trait #0 determines whether an agent considers the dyadic balance of power when deciding whether to initiate a challenge. If the attribute on Trait #0 is “0”, then that agent initiates demands against all states regardless of the balance of power. If the attribute on Trait #0 is “1”, then the agent only initiates demands against weaker agents. Realist theory predicts that over time successful agents would acquire attribute “1” and other agents would emulate these more successful agents (Waltz 1979:118). Bueno de Mesquita et al. predict that democracies will be particularly sensitive to the balance of 6 Many authors prefer to restrict the use of the term “gene” to situations involving death and reproduction. For them, the learning model proposed here would be more appropriately labeled a “meme” structure (Dawkins, 1976). We have chosen to use the terms “traits” and “attributes” in order both to sidestep the debate and to reduce confusion. 11 Domestic structure, learning, and the democratic peace power (2003). Traits #1 and #2 determine whether the balance of power influences decisions to “reject” and “escalate,” respectively. Creating distinct traits for each phase of the conflict allows variation in the power of variables across decision stages. Reed (2000: 88), for example, finds that estimated coefficients can vary significantly between the onset and escalation stages of conflict. Traits #3, #4, and #5 govern the role of alliances in decisions to use force. An attribute value of “1” implies that the agent will not initiate (or reject or escalate) against current allies. Alliances are typically formed in the face of a common threat and therefore can indicate a degree of shared interest (Bueno de Mesquita, 1981). Numerous scholars predict that states will be less likely to initiate against allies than non-allies (e.g. Gowa, 1999; Bennett & Stam, 2004). In the literature overall, the alliance hypothesis has received mixed support because neighbors are both more likely to ally and more likely to fight. Traits #6, #7, and #8 allow the regime type of the opponent to influence decisions to use force. For example, if the attribute on Trait #6 is “0”, then the agent ignores regime type in decisions to initiate conflict. If the attribute is “1”, the agent will only initiate against autocratic opponents. Finally, if the attribute is “2”, the agent will only initiates demands against democratic opponents. Traits #7 and #8 operate in an analogous fashion for decisions to “reject” and “escalate,” respectively. Trait #9 governs the regime type of the agent. If the attribute is a “0”, the agent is an autocratic polity. Conversely, if the attribute is a “1”, the agent is a democratic polity. As discussed below, democratic and autocratic polities differ with respect to their ability to repress domestic political opposition. Finally, Trait #10 governs the satisfaction of the 12 Domestic structure, learning, and the democratic peace agent. If the attribute is “0”, the agent is satisfied with the status quo. If the attribute is “1”, the agent is a revisionist state. Status quo and revisionist states differ in two ways. First, revisionist states ignore domestic opposition when calculating whether or not to initiate conflict. Therefore, revisionist states are more likely to initiate conflict than status quo states, ceteris paribus. Second, revisionist states opportunistically attack neighbors that are already under attack. This opportunistic rule implies that revisionist states are likely to gang up on states under threat. The initial percentage of revisionist states and the frequency of opportunistic behavior are parameters that are set at 0.20 and 0.25, respectively, in our simulations here. Revisionist states suffering a regime change transform into a status quo state. For example, the war-induced regime changes in Germany after World Wars I and II transformed the state into a status quo power, temporarily in the former case and permanently in the latter. The eleven-trait string consists of eight dichotomous and three trichotomous genes. This implies that there are 6912 possible strategies for maximizing growth and security in the anarchic environment (i.e., 2*2*2*2*2*2*3*3*3*2*2). Agents search among these possible strings through mutation and learning. It is important to remember that the fitness of strings is often a function of the current environment. This implies that there may be no movement toward a global optimum over the course of the simulation. For example, a strategy that aids an agent in rapid growth in the short run may be undermined by the adoption of the same strategy by other agents in the neighborhood. In the real world, states constantly shift strategies as new politicians and bureaucrats take office. Good ideas are both forgotten and stumbled upon in the process. In the simulation, this experimentation process is captured by random mutation. If the 13 Domestic structure, learning, and the democratic peace mutation parameter is set at 0.01, there is a 1% chance that the attribute for each trait is switched during an iteration. Given that there are eleven traits in the string, there is roughly a 11% chance of a single attribute changing during each iteration because the probability of mutation for each trait is an independent event. Although mutation allows agents to stumble upon good strategies which might not be available in the immediate neighborhood, it can also be lethal by making the agent unsuited for survival in a competitive environment. The default value for the mutation parameter is 0.001. Learning is a more directed form of change. In the real world, states that are performing poorly often study the strategies of their neighbors in the hope of identifying and adopting a more successful strategy. In the simulation, agents update their strategies using one of three decision rules. The "Look to the most successful" rule implies that agents copy from the most successful agent in the neighborhood, defined as the agent with the most wealth. This rule leads to the rapid diffusion of traits. The “If below mean look above the mean" rule implies that agents first determine if their wealth is below the average in the neighborhood. If so, the agents copy from any agent with wealth above the average of the neighborhood. This rule, which is employed in the default simulation, slows the evolutionary process because only about half the agents learn in each iteration and agents do not always learn from the most successful agent in the neighborhood. Finally, the "If the worst, look to anyone else" rules implies that agents look to see if they are the most unsuccessful in the neighborhood in terms of total power. If so, the agent copies from any other agent in the neighborhood. This rule results in relatively slow learning because few agents learn in each iteration and agents often copy from other pretty unsuccessful agents. 14 Domestic structure, learning, and the democratic peace Agents do not consider changing strategies every round. This reflects the fact that in the real world it often takes some time for a consensus to emerge that a problem exists. For this reason, the parameter P_updateType sets the probability an agent updates in a given round (for all three decision rules). The default value of this parameter is 0.10, implying that states have a 10% chance of updating any one trait if they meet the criterion of the UpdateRule. If both the mutation and the learning parameters are set to 0, agents will never change their behavior, (although they may still change regime type as a result of exogenous coups or democratizations). Incorporating domestic political opposition The domestic politics component of the model is based on three core assumptions. These assumptions, which have extensive theoretical and empirical support, are similar to those discussed in (Rousseau, forthcoming). The power of the model stems from the fact that even a very simple institutional structure can have a profound impact on foreign policy behavior. Assumption #1: All states, whether autocratic or democratic, have domestic political opposition (Bueno de Mesquita et al., 2003). While the extent of opposition can vary from state to state, it exists to some degree in all states. Assumption #2: Although there is great variance in the repressive power of autocratic states, on average autocratic states can repress domestic political opposition more than democratic states. Assumption #3: Domestic political opposition reduces the probability of initiation and escalation for status quo states. In contrast, revisionist states ignore domestic political opposition. 15 Domestic structure, learning, and the democratic peace At the initialization of the simulation, each agent is randomly assigned a level of domestic opposition drawn from a uniform distribution bounded by a minimum and a maximum (0.20 and 0.60, respectively, in the default simulation). Democracies and autocracies do not differ with respect to the initial levels of opposition. Domestic opposition then rises and falls over the course of the simulation between the bounds of 0 and 1.0 based on four factors: 1) rate of economic growth; 2) level of repression; 3) severity of military conflict; and 4) the “rally around the flag” effect.7 First, domestic political opposition changes in proportion to changes in economic growth. For example, if the economy grows by 2.5% in a year, domestic political opposition declines by 5%. Economic growth is a function of the growth rate minus the rate of consumption and the costs of war. During each round of the simulation, a growth rate is randomly selected from a normal distribution with a mean and a standard deviation (set at .0025 and .005, respectively, in the default simulation).8 The economic growth factor captures the fact that random factors outside of military conflict can influence domestic politics. Second, political repression reduces domestic opposition. The extension of economic and political civil liberties in democratic polities coupled with respect for rule of law implies that domestic political opposition is less likely to be silenced by censorship and coercion. The ability to repress in autocracies is a function of their “repressive power endowment” and regime stability. Repressive power endowment is 7 Although not addressed in this paper, each of these factors is parameterized in the model, allowing the user to conduct sensitivity analysis by selectively zeroing out individual factors. 8 If each iteration is analogous to a month, then the growth rate is about 3% per year. In future reversion of the model we hope to incorporate temporal correlation into the model in order to model the impact of economic cycles. The large standard deviation relative to the mean implies recessions take place in the model. 16 Domestic structure, learning, and the democratic peace randomly assigned for each agent at the initialization of the simulation by drawing from a normal distribution with a mean (.02) and a standard deviation (.05). Regime stability is a function of how long the democracy has been a democracy (or the autocracy has been an autocracy). In the simulation, this is operationalized by creating a Stability variable that is equal to 1 divided by the number of years since the last regime change. On average, the longer the regime has existed, the more it is able to repress the political opposition. Therefore, the ability to repress is equal to repressive power endowment minus stability. For example, in an autocratic state, a repressive power endowment of .02 will reduce domestic opposition by 1% in the first year of existence and 1.99% during the 100th year of existence. Third, domestic political opposition grows during military conflicts and due to military defeats (Stein, 1980). During military conflicts, domestic political opposition rises in proportion to the cost of war. For example, if the cost of war in a particular iteration is 1.25% of GNP, then domestic opposition rises by this amount. In the military defeat, the domestic opposition rises in proportion to the amount of GNP lost by the defeated state. Finally, if the state concedes, domestic opposition rises in proportion to the loss in power resulting from the loss of the province. Fourth, domestic political opposition declines due to a “rally around the flag” effect. Numerous studies of public opinion have shown that the popularity of chief executives rises during times of military conflict whether the state is the aggressor or the target (Mueller 1973, 1994; Cotton 1987; Page and Shapiro 1992). In the simulation, domestic political opposition declines at the start of a dispute, crisis, or war by a random number drawn from a uniform distribution bounded by a minimum and maximum 17 Domestic structure, learning, and the democratic peace defined by the user. In the default simulation, the rally minimum and the rally maximum are set at 0 and 5%, respectively. Regime change occurs when the value on Trait #9 shifts from 0 to 1 (democratization) or 1 to 0 (autocratization). Regime change is a function of four factors: 1) rising domestic political opposition; 2) regime stability; 3) random shocks (e.g., coup) in the form of mutation of the regime type trait; and 4) learning from successful agents in the neighborhood. Having addressed mutation and learning above, all that needs to be specified is the impact of rising domestic political opposition and regime stability. Opposition and stability are combined into a single function because of their countervailing properties. In general, the higher the level of domestic political opposition, the greater the probability of regime change. However, the longer an agent has been a democracy (or autocracy), the less likely it is to experience a regime change holding all else constant (such as domestic opposition). Therefore, a long lived democracy such as the United States is more likely to weather a period of high political opposition than a young democracy such as Panama or South Korea. In the simulation, the probability of regime change is equal to (((DomesticOpposition/2)+Stability)/2)2. The division of domestic opposition by two implies that the maximum is .50; this ensures that stability and domestic opposition are equally weighted in the function. It also guarantees that the sum of the two factors never exceeds one. The averaging of the two factors implies that stability can offset domestic opposition and vice versa. Finally, the squaring of the term reduces the probability of change and implies that there is non-linear relationship. For example, if a regime is in its second year and the domestic opposition is 18 Domestic structure, learning, and the democratic peace .75, there is a 19% chance of a regime change. However, if the regime has been in place for 100 years, the probability falls to 4% despite the same level of opposition. While there are obviously an infinite number of ways to formulate the impact of opposition and stability on regime change, the proposed function addresses the tension between stability and opposition in a simple manner. Moreover, the fact that the same rule is applied to all types of regimes reduces the probability that the developer surreptitiously building a model that automatically produces desired results. In fact, the only difference between democracies and autocracies is the presence of repressive capabilities within autocracies. This single difference produces important differences in behavior. Simulation results It should be obvious by this point that ours is a rather complex model, which undeniably violates the KISS principle.9 However, the complexity is necessitated by two factors. First we aim to test the theoretical insights from the literature as precisely as possible. Second, we have decided to parametrize a large number of the hard-coded assumptions and values in GeoSim, facilitating the robustness testing of our findings. Nevertheless, with so many parameters to change, much groundwork is necessary in order to establish the baseline performance of the model, and the stability of outcome patterns as various parameters change. As a result, the findings presented here must be considered preliminary. This makes them no less interesting, but it means we need to be wary of placing too much emphasis on them. 9 KISS = Keep it simple, stupid. Axelrod (1997: 5) strongly advocates following this principle because it reduces the likelihood that results are affected by insufficiently understood interaction effects between a large number of parameters. 19 Domestic structure, learning, and the democratic peace As noted earlier, for the purposes of the present paper we varied just two parameters central to the investigation of the evolutionary process of a democratic peace: the rate of learning and the tendency of democracies to punish pariah states. All other parameters are held constant at their default values, as listed in the appendix. Baseline results All of the results presented below are averages over 100 runs with different randomly generated initial configurations of the world, subject to the parameter values specified. Each run lasted 5000 rounds. The baseline configuration featured no pariah states, an initial configuration of 25% democracies, and the intermediate-speed learning rule (imitate someone better if your resources are below average). Runs started with 100 states and ended, on average, with 57 states, of which 36.5% were democracies. In other words, the fraction of democracies increased slightly over the course of each run, but at the same time the international system was consolidating, so that the average number of democratic states actually shrank. The states at the end of the run were configured such as to produce an average of 208 dyads. In terms of revisionism, at the end of each run on average 14.15% of the democratic states were revisionist, compared to 14.92% of the autocratic states, a difference with no statistical significance. On the other hand, figures for domestic opposition were notably different, with opposition levels at the end of each run in democracies 27% on average, compared to 15.23% for autocracies. It is worth recalling that average opposition levels at startup were not different between the two regime types, 20 Domestic structure, learning, and the democratic peace so these differences are a result of the evolution of the system over the course of a run. Given the role opposition levels play in determining state decisions regarding conflicts, it will not come as a surprise that these difference are reflected in the conflict data. On average, over the course of each run, 6.66% of the dyads were at the dispute level, 2.52% were at the crisis level, and 1.07% were at war. Democracies initiated disputes roughly 1.03% of the time, and autocracies did so 1.11% of the time. This may seem like a small difference, but it is highly statistically significant: a two-tailed t-test gives a probability of 10-132. Differences between autocracies and democracies continue at higher levels of international conflict, when we look at different types of dyads. On average 2.38% of the democratic dyads are in crisis, 2.48% of the mixed dyads, and 2.59% of the autocratic dyads. The parallel figures for dyads at war are 0.96%, 1.11%, and 1.22% respectively. Again, the differences between these values are strongly significant statistically. Finally, it is interesting to look briefly at the tendency of democracies and autocracies to escalate to war, depending on the type of dyad and the nature of the state that initiated the conflict. We find that challenger states are more likely to escalate to war against states of the opposing regime type. Thus, for conflicts in which a democracy is the original challenger, it is more likely to escalate to war if the target state is an autocracy, and vice versa. The same pattern holds for the much smaller fraction of conflicts escalated to the war level by the defending state. Interestingly, the conflict-type most likely to be escalated to war overall (whether by the challenger or the defender) is that of a democracy challenging an autocracy. More broadly, once a democracy has initiated a conflict, it tends to be more likely also to be willing to escalate than an 21 Domestic structure, learning, and the democratic peace autocracy,10 whereas, conversely, autocracies that have been targeted are more willing to escalate than are democracies that have been targeted. It is worth investigating these findings further, in reference also to arguments in the democratic peace literature about the resolve of democratic states once involved in crises (e.g. Lake, 1992).11 Learning The next issue to examine is the degree to which learning takes place over the course of a run. Table 2 shows the results here. The amount of learning is not tremendously great, but some learning clearly occurs. Moreover, it occurs where one might expect it to, especially regarding the power ratio (traits 0-2). Among democracies and autocracies alike more states prefer to take on only weaker states than do not. Alliance status (traits 3-5), on the other hand, has no clearly discernible impact. The reason for this is likely that alliances are too fleeting and too few and far between for there to be many states who have experience with the option of attacking an ally. Moreover, it is common for allies not to share a common border, making intra-alliance conflict impossible. These findings are in line with the weak findings for the alliance variable in the quantitative literature (e.g. Rousseau, forthcoming). Regime type (traits 6-8), on the other hand, does matter, albeit not as strongly as does the power ratio. Interestingly, there is a tendency in the system to learn to become specific in terms of what types of regimes one attacks (trait value 0 is least common). Interestingly, however, whereas autocracies learn to prefer to stay away from fellow 10 11 However, this finding is not statistically significant. The finding is not vulnerable to changes in the rate of learning. 22 Domestic structure, learning, and the democratic peace autocracies (more of them only escalate against democracies), democracies are roughly evenly split between preferring to attack autocracies and democracies. In some ways, this is the opposite of what one might expect from the democratic peace literature.12 A final finding of interest here is that, apart from the regime type findings, differences between autocracies and democracies in either what they learn, or how well they learn it are essentially non-existent. It might be interesting to investigate this issue further, for example to compare our monadic findings to Cederman’s (2001a) finding that learning takes place not just among democratic dyads, but also among mixed and autocratic dyads. [Table 2 about here] Rate of learning As indicated earlier in the paper, we tested the impact of having different rates of learning. The results presented so far represent those for the intermediate rate of learning. We repeated the same set-up with fast and learning as well, to see what differences emerge. Not surprisingly, with the slowest update rule, noticeably less learning takes place in terms of the different traits of states. The patterns from table 2 are attenuated or even disappear altogether. On the other hand, with the fastest update rule, the patterns are not all that different from those for the intermediate rule. This tendency is reflected as well when one looks at outcomes in terms of international conflict. 12 One interesting alternative way to operationalize the issue would be to have trait values representing a preference for attacking one’s own regime type versus the opposite regime type. This might produce results more in line with findings in the literature. 23 Domestic structure, learning, and the democratic peace The first finding that emerges is that rapid learning helps slow down the rate of consolidation of the state system. It does so, the findings suggest, because even the relatively minor learning evident in the results from table 2 are sufficient to reduce the incidence of international conflict in a meaningful fashion. Table 3 shows how a number of the variables already discussed vary as the rate of learning changes. [Table 3 about here] Throughout the table, it is clear that the key differences arise between medium and slow learning, while the difference between fast and medium learning tends to be smaller. Nevertheless, even the latter difference is usually statistically significant (the exception being the fraction of democracies at the end). The table does not present all the variables we discussed earlier, but the trends for the omitted variables are all the same. Faster learning reduces the tendency of states to initiate disputes, it reduces the fraction of dyads at all stages of conflict, and it reduces the tendency of states to be (or remain) revisionist. Given the relatively small differences from an even distribution across trait values we saw in table 2, some of the differences in table 3 are quite striking. For example, the number of dyad wars increases by 50% as the rate of learning slows down from fast to slow. 24 Domestic structure, learning, and the democratic peace Punishing pariah states The final experiment we conducted was to introduce a punishment strategy for socalled pariah states: autocracies that initiate conflicts with democracies. As noted earlier, one possible mechanism for the creation of the democratic peace is that autocracies that ‘invade’ a democratic peaceful region by attacking a democracy will be attacked in turn by other, surrounding democracies. Along these lines, it has been suggested in the literature that democracies tend to come to the aid of other democracies (e.g. Cederman, 2001b). More generally, the notion that punishment of defectors may facilitate the evolution of cooperation is widely accepted in the cooperation literature. Our experiment tests the generalizability of this notion to the case of the democratic peace. We repeated our baseline experiment with each of our two pariah-state parameters set to 0.25. This means that on average, a pariah state remains such for a period of 4 rounds (not very long, in other words), and that there is a 25% chance that a democracy bordering on a pariah state will target it for a conflict. Such explicit targeting of autocracies that have violated the democratic peace has been suggested in the literature as one possible mechanism for producing a stable democratic peace. One might assume that in our model, since such states are likely to do very poorly, they will tend to learn from their neighbors not to attack democracies in the future, which might indeed produce a more peaceful system overall. However, the opposite turns out to be the case. Autocracies attack democracies with sufficient frequency that more often than not the system contains one or more pariah states. As a 25 Domestic structure, learning, and the democratic peace result, the number of wars increases too. This has a number of important effects. Table 4 shows the same variables as table 3, but now showing the effect of introducing pariahs. [Table 4 about here] First, there is a far greater degree of consolidation in the system. The average number of states at the end is reduced by more than 1/3. The average ratio of democracies is slightly greater, but the difference is not significant at the 0.05 level. The same goes for differences in the rate of revisionism observed at the end of the run. More interesting, however, are the striking differences in the incidence of conflict. As table 4 shows, disputes, crises, and wars all become more common, but the relative increase is by far the greatest for the incidence of wars, which occur two and a half times as often. We also see that democratic states become more than twice as likely to initiate disputes as autocracies (as well as more than twice as likely as they were in the no-pariah condition). Moreover, the impact is felt not just by autocracies targeted as pariahs — indeed, fully democratic dyads experience wars three times as often as they did in the no pariah condition! What might account for these dramatic differences? An important contributor to the effect is undoubtedly the opportunism present in our model. Revisionist states may opportunistically attack another state that is already at war, hoping to benefit by forcing the target state to divide its war-fighting resources. When making this decision, revisionists do not take into account the regime type of their target. As democracies get involved in punishing autocracies, therefore, they also make themselves vulnerable to attacks from other, revisionist, democracies. 26 Domestic structure, learning, and the democratic peace Discussion Our simulation results have a number of implications. First, and most obviously, they provide support for the monadic hypothesis as an explanation for the democratic peace. In other words, as a causal mechanism, introducing a role for domestic opposition suffices to support a democratic peace — no special dyadic features are required. Indeed, the empirical finding that democratic dyads are less likely to engage in war than either mixed or autocratic dyads also emerges in our model. On the other hand, it is worth noting that the monadic mechanism does not appear sufficiently strong to lead inexorably to an outcome where all states are democratic. Instead, the fraction of democracies in the world seems to stabilize around 40%. To create a democratic world in the long run, in other words, some additional factors appear to be necessary. Second, although the number of initiations per opportunity in the simulations may seem low for both types of states, this simply reflects the rarity of war and the large number of dyads in the system. The mean fraction of conflict initiations on the part of democracies is 1.03%, versus 1.11% for autocracies. However, with the number of dyads declining slowly over time from about 350 to about 125, this still means that over three new conflicts will be initiated on average in the early phases, declining to just over one new conflict per round by the end. If anything, this seems like a large number of conflicts. However, the Militarized Interstate Dispute (MID) data set (version 3.01) reveals that in just under 200 years the United States has been involved in 344 disputes and the United Kingdom has been involved in 263 militarized disputes. 27 Domestic structure, learning, and the democratic peace One might argue that the development of so many democracy-democracy wars calls the validity of our model into question. After all, there has never been a case of large-scale violence between two democratic states, and as Levy (1989) points out (and the opening quotations illustrate), the finding that democracies do not fight each other is often understood as a law. To some degree, the number of wars in our model is a function of its simplicity: both regime type and war are dichotomous variables. If we were to create a range of regime types and if we restricted the term “war” to cases of major losses on the battle front, the number of wars would drop drastically. Moreover, we want to underscore that we do not intended to (nor can we!) make point predictions about particular historical eras or the future. The purpose is to determine if a handful of simple assumptions can produce patterns congruent with those we observe in the real world around us. Therefore, the important point to take from the model is not that democracies do sometimes fight one another, but rather that democratic dyads are involved in significantly fewer wars than other types of dyads.13 Our findings also give support to suggestions in the literature that a learning process may inform the gradual evolution of a democratic peace, as well as of other empirical patterns in the incidence of international conflict. For example, states learn to prefer attacking weaker rather than stronger states. On the other hand, our results also suggest that such learning is not restricted to democracies alone, a finding which also emerged from Rousseau’s empirical data (forthcoming) and Cederman’s statistical 13 In fact, if we introduced additional assumptions about the likely initial opposition levels in autocracies and democracies, we could probably easily create a world in which democracies virtually never fight one another. The value of our finding is that we do not need to make such assumptions to produce the emergence of a democratic peace. 28 Domestic structure, learning, and the democratic peace analysis of learning and the democratic peace (2001a), but which is at odds with the argument made by Bueno de Mesquita et al. In their recent work (2003). Finally, it is worth noting that, not altogether unexpectedly, democratic states in the simulation runs tend to cluster spatially. We have not yet developed a good quantitative measure of this tendency to cluster, but visual inspection of virtually any run of the simulation shows that the vast majority of democracies tend to cluster in one or at most two groups. Two mechanisms are likely shape this pattern. First, learning of traits from contiguous states tends to lead to clustering. If there are several democracies in the neighborhood and they are doing very well, an autocratic state is more likely to decide to democratize as part of its adaptation.14 Second, the spatial clustering appears to be a function of lower frequency of democracy-democracy war. If an agent clusters with states that are less likely to attack it, it can both avoid the costs of war and allocate forces to fronts that are more likely to become engaged. Thus, democracies, which are less likely to initiate in general and less likely to become involved in democracy-democracy wars, are likely to cluster together in the anarchical environment. This result is consistent with Deutsch’s explanation of the emergence of security communities (1957). Conclusions Using just a handful of basic assumptions related to domestic opposition and the costs of initiating conflict when such opposition is high, a form of democratic peace emerges in the simulations discussed above. Our results show both monadic and dyadic 14 On the spatial diffusion of democracy see (Cederman & Gleditsch, 2002; Gleditsch, 2002; Gleditsch & Ward, 2000). 29 Domestic structure, learning, and the democratic peace features of democratic peace, with democracies less likely to initiate conflicts overall, and democratic-democratic dyads less likely to become embroiled in conflicts than democratic-autocratic dyads, which in turn are more peaceful than autocratic-autocratic dyads. This basic result is particularly powerful because it did not require the introduction of many of the additional assumptions often introduced in the democratic peace literature, such as: - democracies suddenly dropping their peaceful norms of conflict resolution when facing an autocratic opponent (Maoz & Russett, 1993) - political opposition existing only in democratic polities. - democratic polities behaving differently because they expect their democratic opponents to behave differently (Bueno de Mesquita & Lalman, 1992) - autocratic leaders being more likely to focus on private benefits than their democratic counter parts (Bueno de Mesquita et al., 2003) - democracies not fighting other democracies (Cederman, 2001b) - democracies being more likely to settle disputes through mediation and arbitration (Dixon, 1993, 1994; Raymond, 1994) - democracies only allying with other democracies While many of these assumptions or causal mechanisms may in fact reinforce the monadic and dyadic peace, the simulation demonstrates that they are not necessary for its emergence. One needs only to make three relatively uncontroversial assumptions: domestic political opposition exists in all states, autocratic states have greater powers of repression, and domestic political opposition inhibits the use of force. 30 Domestic structure, learning, and the democratic peace It is also worth pointing out that the dynamic nature of the model means that it is not as obvious as it might seem after the fact that these three assumptions would produce the monadic and dyadic democratic peace. Instead, it was quite possible that realists such as Waltz would have been correct: If domestic politics limit a state’s ability to balance threats, then the state will be eliminated from the gene pool in the long run. However, the simulation revealed the exact opposite: agents with greater domestic interference thrived within the anarchic system. Indeed, as noted earlier, the proportion of democratic states rose, on average, from 25% to 40% over the course of 5000 rounds. Beyond this central finding, our experiments also uncovered some additional interesting patterns, each of which is well worth further investigation. Two are worth mentioning here. First, in mixed conflict dyads, democracies are more likely to escalate to war than are autocracies if they are the challenging state which initiated the conflict, but they are less likely to escalate if they are the target state. This finding is congruent with claims in the literature regarding the resolve of democracies, but is particularly interesting because, again, we did not make any particular assumptions regarding resolve in setting up the model. Instead, the pattern emerged as a result of the interaction of other parameters. Second, introducing a mechanism by which autocracies that attack democracies are punished not only dramatically increases the incidence of international conflict, but also completely eliminates the democratic peace. Indeed, even when autocracies are not labeled as pariahs for a particularly long period and when punishment is far from automatic, the result is a system in which democracies are noticeably more likely to be at war, regardless of the regime type of their adversary. Given the current popularity among 31 Domestic structure, learning, and the democratic peace foreign policy experts of the notion that a democratic peace can be created in a pro-active fashion by working to democratize the worst autocracies, this finding raises some potentially troubling questions. In closing, we should emphasize that we are still in the early stages of investigating the impact of different parameters in the model. The results presented here, therefore, are preliminary in nature (although a number of basic robustness tests were performed which did not substantively affect the observed patterns). In addition, additional empirical research, both qualitative and quantitative, may suggest changes to be made to the model’s assumptions, decision rules, and overall structure. After all, simulations are useful tools, but they must be used in conjunction with alternative methods of inquiry to ensure a comprehensive analysis 32 Domestic structure, learning, and the democratic peace References Axelrod, R. (1997). The Complexity of Cooperation: Agent-Based Models of Competition and Collaboration. Princeton, NJ: Princeton University Press. Axtell, R. (2000). Why agents? On the varied motivations for agent computing in the Social Sciences (No. Working Paper No. 17). Washington, DC: Brookings Institution, Center on Social and Economic Dynamics. Bennett, D. S., & Stam, A. C. (2004). The Behavioral Origins of War. Ann Arbor, MI: University of Michigan Press. Bueno de Mesquita, B. (1981). The War Trap. New Haven, CT: Yale University Press. Bueno de Mesquita, B., & Lalman, D. (1992). War and Reason. New Haven, CT: Yale University Press. Bueno de Mesquita, B., Smith, A., Siverson, R. M., & Morrow, J. D. (2003). The Logic of Political Survival. Cambridge, MA: MIT Press. Cederman, L.-E. (1997). Emergent actors in world politics : how states and nations develop and dissolve. Princeton, N.J.: Princeton University Press. Cederman, L.-E. (2001a). Back to Kant: Reinterpreting the Democratic Peace as a Macrohistorical Learning Process. American Political Science Review, 95(1), 1531. Cederman, L.-E. (2001b). Modeling the Democratic Peace as a Kantian Selection Process. Journal of Conflict Resolution, 45(4), 470-502. Cederman, L.-E. (2003). Modeling the Size of Wars: From Billiard Balls to Sandpiles. American Political Science Review, 97(1), 135-150. Cederman, L.-E., & Gleditsch, K. (2002). Conquest and Regime Change: An Evolutionary Model of the Democratic Peace. Paper presented at the Annual Meeting of the American Political Science Association, Boston, MA. Christensen, T., & Snyder, J. (1990). Chain Gangs and Passed Bucks: Predicting Alliance Patterns in Multipolarity. International Organization, 44(2), 137-168. Dawkins, r. (1976). The Selfish Gene. New York: Oxford University Press. Deutsch, K. (1957). Political Community and the North Atlantic Area. Princeton: Princeton University Press. Dixon, W. J. (1993). Democracy and the Management of Internaitonal Conflict. Journal of Conflict Resolution, 37(1), 42-68. Dixon, W. J. (1994). Democracy and the Peaceful Settlement of International Conflict. American Political Science Review, 88(1), 14-32. Fearon, J. (1994). Domestic Political Audiences and the Escalation of International Disputes. American Political Science Review, 88(3), 577-592. Gelpi, C. F., & Griesdorf, M. (2001). Winners or Losers? Democracies in International Crisis, 1918-94. American Political Science Review, 95(3), 633-647. Gleditsch, K. (2002). All International Politics is Local: The Diffusion of Conflict, Integration, and Democratization. Ann Arbor, MI: University of Michigan Press. Gleditsch, K., & Ward, M. D. (2000). War and Peace in Space and Time: The Role of Democratization. International Studies Quarterly, 44(1), 1-29. 33 Domestic structure, learning, and the democratic peace Gowa, J. (1999). Ballots and Bullets: The Elusive Democratic Peace. Princeton, NJ: Princeton University Press. Johnson, P. E. (1999). Simulation Modeling in Political Science. American Behavioral Scientist, 42(10), 1509-1530. Kinsella, D., & Russett, B. (2002). Conflict Emergence and Escalation in Interactive Internaitonal Dyads. Journal of Politics, 64(4), 1045-1068. Lake, D. (1992). Powerful Pacifists: Democratic States and War. American Political Science Review, 86(1), 24-38. Levy, J. S. (1989). Domestic Politics and War. In R. Rotberg & T. Rabb (Eds.), The Origin and Prevention of Major Wars (pp. 79-100). New York: Cambridge University Press. Maoz, Z., & Russett, B. (1993). Normative and Structural Causes of Democratic Peace, 1946-1986. American Political Science Review, 87(3), 624-638. Raymond, G. A. (1994). Democracies, Disputes, and Third-Party Intermediaries. Journal of Conflict Resolution, 38(1), 24-42. Reed, W. (2000). A Unified Statistical Model of Conflict Onset and Escalation. American Journal of Political Science, 44(1), 84-93. Rousseau, D. L. (2004). Identifying Threats and Threatening Identities: Constructivism in International Relations.Unpublished manuscript, Philadelphia, PA. Rousseau, D. L. (forthcoming). Democracy and War: Institutions, Norms, and the Evolution of International Conflict. Stanford, CA: Stanford University Press. Stein, A. A. (1980). The Nation at War. Baltimore, MD: Johns Hopkins Press. Waltz, K. N. (1979). Theory of international politics. Reading: Addison-Wesley. 34 Domestic structure, learning, and the democratic peace Figure 1. Snapshot of the DomGeoSim world. 35 Domestic structure, learning, and the democratic peace 36 Win War Draw Escalate Lose Challenging State Reject Demand Target State Status Quo Delay Dow n Delay Bac k Dow n Challenging State Not Bac k Diffused Status Quo Figure 2. The international conflict game tree. Target Concedes Diffused Status Quo Challenger Concedes Domestic structure, learning, and the democratic peace Nr. 0 Phase Peace Name Power 1 Dispute Power 2 Crisis Power 3 Peace Alliance 4 Dispute Alliance 5 Crisis Alliance 6 Peace Regime type 7 Dispute Regime type 8 Crisis Regime type 9 n.a. Regime type 10 n.a. Satisfaction Value 0 1 0 1 0 1 0 1 0 1 0 1 0 1 2 0 1 2 0 1 2 0 1 0 1 Table 1. Trait Structure of the Model Description Ignore power ratio Only escalate against weaker states Ignore power ratio Only escalate against weaker states Ignore power ratio Only escalate against weaker states Ignore alliance status Only escalate against non-allies Ignore alliance status Only escalate against non-allies Ignore alliance status Only escalate against non-allies Ignore regime type Only escalate against autocracies Only escalate against democracies Ignore regime type Only escalate against autocracies Only escalate against democracies Ignore regime type Only escalate against autocracies Only escalate against democracies Autocracy Democracy Revisionist Status quo 37 Domestic structure, learning, and the democratic peace Trait 0 1 2 3 4 5 6 7 8 Regime Democracy Autocracy Democracy Autocracy Democracy Autocracy Democracy Autocracy Democracy Autocracy Democracy Autocracy Democracy Autocracy Democracy Autocracy Democracy Autocracy 0 43.84 44.05 45.76 46.34 45.49 43.96 48.90 49.66 49.03 49.71 48.33 48.90 30.83 32.42 32.69 31.71 29.69 30.48 1 55.33 55.09 53.41 52.79 53.74 55.15 50.30 49.44 50.16 49.44 50.88 50.26 33.99 31.37 33.98 31.81 34.00 32.39 2 34.08 34.97 32.20 35.16 35.14 35.81 38 T-test 0 0 0.002 0.001 0 0 0.56 0.91 0.61 0.88 0.20 0.47 0.08 0.03 0.45 0.04 0.006 0.006 Table 2. Learning: strategy choices for various traits at end of run, by regime type. T-test values shown for traits 6-8 are for the difference between the most and least popular trait values. (Note: percentages do not add to 100 due to small changes in number of states in the system that occur between the counting of different strategies and of the number of states.) Domestic structure, learning, and the democratic peace Variable Fast Medium Nr. states at end 59.57 57.34 Fraction of democracies at end 0.38 0.37 Democratic revisionism at end 12.58 14.15 Autocratic revisionism at end 13.59 14.92 Disputes / dyads 6.57 6.66 Crises / dyads 2.39 2.52 Wars / dyads 0.99 1.07 Democratic initiation 1.01 1.03 Autocratic initiation 1.06 1.11 DD crises / DD dyads 2.28 2.38 DD wars / DD dyads 0.86 0.96 Table 3. Implications of the rate of adaptation. 39 Slow 48.36 0.41 19.20 18.66 7.35 3.08 1.45 1.20 1.32 2.68 1.17 Variable No pariahs Pariahs Nr. states at end 57.34 34.24 Fraction of democracies at end 0.37 0.40 Democratic revisionism at end 14.15 13.68 Autocratic revisionism at end 14.92 15.73 Disputes / dyads 6.66 7.58 Crises / dyads 2.52 4.09 Wars / dyads 1.07 2.69 Democratic initiation 1.03 2.57 Autocratic initiation 1.11 1.21 DD crises / DD dyads 2.38 3.91 DD wars / DD dyads 0.96 3.13 Table 4. Implications of introducing pariah states. Domestic structure, learning, and the democratic peace 40 Appendix: DomGeoSim Parameter Dictionary 1. Simulation Set-up Name WorldXSize Description Horizontal dimension of the world grid Default 50 Cederman 50 50 50 Whether the world wraps around left-to-right Type Integer (1…100) Integer, (1…100) true/false WorldYSize Vertical dimension of the world grid WrapHorizontal false WrapVertical Whether the world wraps around top-to-bottom true/false false MaxRounds StartStationary Number of rounds (steps, ticks) to run the system Round in which to start monitoring system for output, and also possibly to restructure the system by turning some states into democracies Whether to run an approximation of Lars-Erik Cederman’s Geosim2 model Whether to keep track of wars and their size (warCounting in Geosim2) Whether to keep track of governance type (democracies vs. autocracies) (democracy in Geosim2) Integer (1…) Integer, less than MaxRounds true/false 5000 20 false (hardcoded) false (hardcoded) 10500 500 true/false true n.a. (but true) True true/false true False Cederman CountWars DemocracyMatters false 2. Initialization Specs Name InitSystem InitPolarity P_hegemon P_revisionist S_neighborhoodType Description Whether to reduce the number of states in the system at the start (if not, every grid location is a state at the start) Number of states desired at the start Probability that a state will receive 10 times the standard quantity of resources at initialization time. (Note that this may not matter much if InitSystem is true, since the amalgamation of states will make resource disparities at the individual-territory level rather less noticeable) The probability of becoming a revisionist state at startup The connectivity structure of the world: • 0 – von Neumann neighborhood (only the 4 straight-line neighbors) • 1 – hexagonal neighborhood (6 of the 8 possible neighbors, in an alternating pattern from row to row, to mimic a hexagonal structure) • 2 – Moore neighborhood (all 8 neighbors, including 4 diagonal ones) Type true/false Default true Cederman True Integer, (1…WorldXSize* WorldYSize) Fraction (0…1) 50 200 0.2 0.2 Fraction (0…1) 0.2 n.a. 0, 1, or 2 0 0 Domestic structure, learning, and the democratic peace F_democracies InitDemsAtStart InitDemocracyBias Fraction of states to turn into democracies at the start (propDem in Geosim2) Whether to turn states into democracies at the start, or (if set to false) at the start of the stationary period) Whether to make democracies stronger at the start. If set to true, pick half of the democracies at random from among the 5% most powerful (richest in resources) states, and the other half at random from among the other 95% of states. Will go wrong if F_democracies exceeds 10%. 41 Fraction (0…1) 0.5 0.1 true/false true False true/false false False 3. Agent Specs Name S_updateRule P_updateType P_changeType Description How to learn from neighboring states: 0 – unless richer than all neighbors, learn from richest neighbor 1 – unless richer than average neighbor, learn from a neighbor whose wealth is above average 2 – if poorer than all neighbors, learn from a randomlyselected neighbor Probability of an attempt to learn from neigbhouring states Probability of a random change in strategy (i.e. exploration / mutation) Type 0, 1, or 2 Default 1 Cederman n.a. Fraction (0…1) Fraction (0…1) 0.1 n.a. 0.001 n.a. Type true/false Default True Cederman false Fraction (0…1) Fraction (0…1) Fraction (0…1) Positive real (0…) 0.025 1.0 0.0025 n.a. 0.005 n.a. 10.0 Positive real (0…) 5.0 True/false False 100.0 at start, 1.0 thereafter 50.0 at start, 5.0 thereafter true Fraction 0.0022 0.99 4. Resource Settings Name CumulativeResources TaxRate M_growth SD_growth M_harvest SD_harvest ConsumeFixed Consumption Description Whether resource gathering is cumulative from round to round. Variable used only for Cederman’s model (WarAndPeace model is always cumulative; Geosim uses the complementary parameter terrRes, true if resources non-cumulative). Fraction of a territory’s resources a capital can extract Mean growth rate of a territory’s resources, per round Standard deviation of the growth rate Mean harvest size for set-ups with cumulative resources (Geosim uses mRes and mHarvest here). Standard deviation of the harvest size (Geosim uses sRes and sHarvest here). Whether to consume a fixed share of resources each round Fraction of resources fixed from one round to the Domestic structure, learning, and the democratic peace 42 next. Used only for Cederman’s model when (0…1) resources non-cumulative (Geosim uses a complementary parameter, called resChange, representing the fraction that changes from one round to the next). ProDemocracyBias Resource extraction bias modifier for Positive real 1.0 1.0 democracies. A states’ total extraction of (0…) resources is multiplied by this value (i.e. a value below 1 means democracies extract fewer resources; above 1 means they extract more than autocracies). Used only if Cederman is true (demBias in Geosim2). Note: double-check that ConsumeFixed and Consumption are described correctly. More generally, need to double check once more all the resource functions, since war chests end up being very small, while average GDP per province gradually increases. 5. Distance Settings Name S_distanceCosts DistanceGradient T_distance SD_distance DistanceSlope DistanceOffset Description The way in which a state’s ability to extract resources from distant provinces as well as its ability to mobilize forces in those provinces to face opponents there are affected by distance (comparable parameter in Geosim2 is distRes). • 0 – no costs associated with distance • 1 – costs follow a geometric pattern (distance^gradient). With the default settings, the fraction of resources extractable/mobilizable at successive integer distances is: 1 – 1, 2 – 0.5, 3 – 0.33, 4 – 0.25, 5 – 0.2, 6 – 0.17. • 2 – costs (apart from offset) follow a logistic pattern: 1/(1+e^(slope*ln(distance/threshold))). With the default settings, the fraction of resources extractable/mobilizable at successive integer distances is: 1 – 0.90, 2 – 0.55, 3 – 0.31, 4 – 0.2, 5 – 0.15, 6 – 0.13, 7 – 0.12. Used when S_distanceCosts = 1. Extractive and mobilizing ability are calculated as (distance from province to capital)^DistanceGradient. So for the default value of -1, this is 1/(distance from province to capital) Threshold of distance at which the inflection point of the logistic curve occurs (distDrop in Geosim2). Standard deviation of the initial distance threshold, relevant for distance-loss functions that are statespecific, i.e. as a result of shocks (distDropSD in Geosim2) Slope of logistic distance-loss curve The fraction of resources unaffected by the logistic distance loss function. 6. Conflict Initiation Type 0, 1, or 2 Default 1 Cederman 2 Positive real (0…) -1.0 0.0 Positive real (0…) Positive real (0…) 2.0 2.0 0.0 0.0 Positive real (0…) Fraction (0…1) 3.0 3.0 0.1 0.1 Domestic structure, learning, and the democratic peace Name P_unprovokedAttack Description The probability of initiating an unprovoked conflict P_selectWeakest The probability, when initiating a conflict, of selecting the weakest possible target, as opposed to a randomly-selected possible target. The probability that a revisionist state will behave opportunistically and attack a neighboring state already embroiled in a conflict. Whether to allow campaigns. If true, states will keep fighting against opponents even after a territory changes hands. The ‘unit’ of a war is a singleterritory conflict. If true, states will go on attacking the target of their campaign as long as they share a border (governed, of course by the value of P_dropCampaign). The probability of dropping a campaign against a state The probability a state will initiate another conflict (i.e. without provocation) if already involved in a conflict Whether to go on the alert when a neighboring state is currently embroiled in a conflict (activeNeighs in Geosim2) Once on the alert (for example because a neighbor is in a conflict), the probability that alert status will be dropped (dropActive in Geosim2) P_revOpportunism Campaigns P_dropCampaign P_twoFrontWar MonitorNeighbors P_dropAlert 43 Type Fraction (0…1) Fraction (0…1) Default 0.05 Cederman 0.01 0 0 Fraction (0…1) 0.25 n.a. true/false true true Fraction (0…1) Fraction (0…1) 0.2 0.2 0 0 true/false true true Fraction (0…1) 0.1 0.1 Type Fraction (0…1) Default 0.5 Cederman 0.5 Fraction (0…1) true/false 0.33 0.1 true true true/false false true Positive real (0…) 3.0 3.0 7. Conflict Implementation Name F_mobile F_warCost NoisyWar BalanceFront SuperiorityRatio Description The fraction of resources in a state’s war chest that is mobile and can be allocated according to the strength of opposing forces. The remaining fraction is evenly divided across all fronts (mobil in Geosim2) Cost of a war to one’s opponent, expressed as a fraction of the resources one has mobilized at the front. Whether the outcome of a battle (win/loss) is deterministic or stochastic (in Geosim2, not a separate parameter, but true unless VictorySlope = 0) Whether to consider only the resources a state has mobilized at our mutual front, or instead against the sum total of the resources it has in its war chest (not a variable in Geosim2 — always balance against front only; in WarAndPeace balancing against a front is not meaningful since a front only exists if a war is being fought) The superiority ratio used to decide whether to attack an opponent. (sup in Geosim2). Depending on the presence of alliances and the value of parameter BalanceFront, if the relevant resources of a state are X times greater than those that can be mustered by the target, a state will Domestic structure, learning, and the democratic peace SuperioritySlope SD_superiority VictoryRatio VictorySlope consider the target attackable. The slope of the logistic function used to decide whether to attack an opponent (if 0, then simply use the SuperiorityRatio value) (supC in Geosim2). The logistic function is analogous to that for distance loss, apart from the offset: 1/(1+e^(slope*ln(SuperiorityRatio/actualPowerRatio))). With the default values, this means that the probability a state will find a target attackable, for various power ratios is: 1 – 3E-10, 2 – 0.0003, 3 – 0.5, 4 – 0.997, 5 – 0.99997. Standard deviation of the superiority ratio, relevant for ratios (and thus attack decisions) that are state-specific, i.e. as a result of shocks (supSD in Geosim2) Functions analogously to SuperiorityRatio above, but now for deciding whether a state will win (vict in Geosim2). When shocks apply, and thus state-specific ratios, a state’s victory ratio is always kept equal to its superiority ratio Functions analogously to SuperioritySlope above, but now for deciding whether a state will win (victC in Geosim2) 44 Positive real (0…) 20.0 20.0 Positive real (0…) 0 0 Positive real (0…) 3.0 3.0 Positive real (0…) 20.0 20.0 Type Fraction (0…1) Integer Default 0.2 Cederman 0 100 n.a. Integer 2 20 true/false False true true/false true true Type true/false Integer Default true 5 Cederman false 2.8 8. Conflict Resolution Name P_peace Patience MaxWarMemory DisintegrateCutoffs KeepConnected Description The probability that a given war will end suddenly in any given round, with a peace treaty The number of rounds a state will allow its opponent in a dispute or crisis simply to delay taking further action The number of rounds after a war is over that a state will remember it was in a war (called maxShadow in Geosim2) When a capital is conquered, or when a section of a state is cut off from the rest, whether to keep cut-off or head-less provinces together as one or more new states, or instead to atomize them all into singleprovince states Whether to keep all territories in a state contiguous 9. Alliance Settings Name AllowAlliances T_invokeAlliance Description Whether to allow alliances The resource ratio beyond which one is willing to join an alliance against an opposing state. For example, if 5, then we join an alliance against an opponent if that opponent has 5 times as many resources as we do in our war chest or at our mutual front (depending on the value of BalanceFront) (in Geosim2, this parameter is called minThreat and is negative, but otherwise with Domestic structure, learning, and the democratic peace P_aidAllies AllyContribution the same implications) The probability that one will come to the aid of one’s allies (i.e. defect against an alliance target) (prOblig in Geosim2) The proportion of the forces of states allied with a target that a state takes into account when deciding whether or not to attack that target (contrib in Geosim2) 45 Fraction (0…1) 0.5 0.5 Fraction (0…1) 0.5 0.5 Type true/false Default true Cederman false Fraction (0…1) 0.8 n.a. Fraction (0…1) 0.002 0.002 Fraction (0…1) 0.001 0.001 Fraction (0…1) 0.25 0 Fraction (0…1) 0.25 1 (?) 10. Democracy Settings Name Democratize Description Whether states are subject to coups and democratizations (democratization in Geosim2) T_regimeChange The threshold in terms of wealth above beyond which a state may change regime. For a value of 0.8, if wealth per territory (a proxy for GDP per capita) is in the top 20% of states and a state is an autocracy, it may turn into a democracy, whereas if wealth per territory is in the bottom 20%, a democracy may turn into an autocracy, as determined by P_democratize and P_coup, respectively. P_democratize The probability that an autocracy will turn into a democracy in any given round. In Geosim2, this probability is modified by a complicated ad hoc function involving a logistic calculation based on the degree to which a state is surrounded by democracies. The parameter value is multiplied by the resulting value, which, for different fractions of surrounding democracy, is: 0.1 – 0.56, 0.2 – 0.62, 0.5 – 0.77, 0.9 – 0.89, 1.0 – 0.91. In WarAndPeace, the probability only comes into effect if the regime change wealth threshold (T_regimeChange) is met. Moreover, in WarAndPeace, regime change may also take place based on domestic opposition, instability, or random exploration (see P_changeType above) P_coup The probability that a democracy will turn into an autocracy in any given round. In Geosim2, modified in the same way as P_democratize, except now we count the fraction of surrounding states that are already autocracies. Comments for WarAndPeace are the same as under P_democratize. P_losePariahStatus The probability that a state marked as a pariah will lose its pariah status. If 0, states will never be marked as pariahs (collSec in Geosim2) P_targetPariah The probability that a pariah state will be targeted for an unprovoked escalation by a democracy it borders Double-check implicit value of P_targetPariah in GeoSim Domestic structure, learning, and the democratic peace 46 11. Domestic Opposition Name F_domOpp_DemMin Description Type Default Min. level of domestic opposition for a new Fraction 0.2 democratic state (0…1) F_domOpp_DemMax Max. level of domestic opposition for a new Fraction 0.6 democratic state (0…1) F_domOpp_AutMin Min. level of domestic opposition for a new Fraction 0.2 autocratic state (0…1) F_domOpp_AutMax Max. level of domestic opposition for a new Fraction 0.6 autocratic state (0…1) SD_opposition Standard deviation of the normal function from Fraction 0.05 which the next round’s initial opposition level is (0…1) drawn (with mean equal to the current round’s opposition level) RepressOpposition Reduction in the opposition level that autocracies Fraction 0.02 can achieve in a given round (will be modified by (0…1) their length of tenure as an autocracy), expressed as a fraction of the maximum possible reduction. RallyMin Minimum level by which domestic opposition will Fraction 0 fall if a state is attacked (expressed as a fraction of (0…1) the maximum possible drop). RallyMax Maximum level by which domestic opposition will Fraction 0.05 fall if a state is attacked. (0…1) Note: Currently rallying takes place every time a conflict shifts into a higher phase, and both attacker and attacked have the rally effect. Cederman n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 12. Shock Settings Name P_shock TaxShock TechnoShock ProDemBias_Shocks Description Probability that a state will be subjected to a shock in any given round, starting from StartStationary (see above) Size of the extractive shock. The shock affects the distance threshold (i.e. the inflection point of the logistic distance loss function). A shocked state’s distance threshold is set to the default model threshold plus the fraction of the shock size that corresponds to the fraction of the period between StartStationary and MaxRounds that has elapsed. In other words, the threshold will gradually increase over time, meaning extraction ability improves over time. If true, a shocked state’s victory and superiority ratios will be reset to the result of a random draw from a normal distribution with mean SuperiorityRatio and standard deviation SuperiorityRatio*SD_superiority Degree to which democracies are more or less likely to be shocked. Multiplied by the shock probability, so that a value below 1 means democracies are less likely to be shocked, and a value above 1 means they are more likely to be. Type Fraction (0…1) Default 0.0001 Cederman 0.0001 Real 20.0 20.0 true/false false false Positive real (0…) 1.0 1.0 Domestic structure, learning, and the democratic peace 47 13. Output Choices Name S_WarSizeMeasure ReportInterval Description The measure used to gauge the size of wars: 0 – the cost of the war 1 – the duration of the war, in rounds 2 – the total number of states involved in the war 3 – the duration * the total number of states 4 – the number of state-rounds (less than 3, since not every state will be involved in every round the war is ongoing). All of these values are written to the output, but only the selected one is displayed, if the war-size chart is displayed during the run The interval at which data gets written to a file. Data about wars is written out for each war, but tracking data about the number of states, ongoing conflicts, etc. is only written out every ReportInterval rounds (as well as prior to round 1). Type 0, 1, 2, 3, or 4 Default 4 Cederman 0 Integer 20 n.a.