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Draft, not for citation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.