Emergent Structure Models: Applications to World Politics Prof. Lars-Erik Cederman Center for Comparative and International Studies (CIS) Seilergraben 49, Room G.2, lcederman@ethz.ch Christa Deiwiks, CIS Room E.3, deiwiks@icr.gess.ethz.ch http://www.icr.ethz.ch/teaching/compmodels Week 12

Applying Geosim to World Politics Configurations Processes Qualitative properties Example 3. Democratic peace Example 4. Emergence of the territorial state Distributional Example 2. State-size distributions Example 1. War-size distributions

Cumulative war-size plot, 1820-1997 Data Source: Correlates of War Project (COW)

Self-organized criticality Per Bak’s sand pile Power-law distributed avalanches in a rice pile

Theory: Self-organized criticality log f f Input Output s-a log s Complex System s Slowly driven systems that fluctuate around state of marginal stability while generating non-linear output according to a power law. Examples: sandpiles, semi-conductors, earthquakes, extinction of species, forest fires, epidemics, traffic jams, city populations, stock market fluctuations, firm size

War clusters in Geosim t = 3,326 t = 10,000

Simulated cumulative war-size plot log P(S > s) (cumulative frequency) log s (severity) log P(S > s) = 1.68 – 0.64 log s N = 218 R2 = 0.991 See “Modeling the Size of Wars” American Political Science Review Feb. 2003

Applying Geosim to world politics Configurations Processes Qualitative properties Example 3. Democratic peace Example 4. Emergence of the territorial state Distributional Example 2. State-size distributions Example 1. War-size distributions

2. Modeling state sizes: Empirical data log Pr (S > s) (cumulative frequency) log S ~ N(5.31, 0.79) MAE = 0.028 1998 log s (state size) Data: Lake et al.

Simulating state size with terrain

Simulated state-size distribution log s (state size) log Pr (S > s) (cumulative frequency) log S ~ N(1.47, 0.53) MAE = 0.050

Applying Geosim to world politics Configurations Processes Qualitative properties Example 3. Democratic peace Example 4. Emergence of the territorial state Distributional Example 2. State-size distributions Example 1. War-size distributions

Simulating global democratization Source: Cederman & Gleditsch 2004

A simulated democratic outcome

Applying Geosim to world politics Configurations Processes Qualitative properties Example 3. Democratic peace Example 4. Emergence of the territorial state Distributional Example 2. State-size distributions Example 1. War-size distributions

The initial state of OrgForms

Modeling technological change

OrgForms: A dynamic network model Conquest Technological Progress Systems Change Organizational Bypass

Indirect rule in the “Middle Ages”

Replications with moving threshold and slope

Exploring geopolitics using agent-based modeling OrgForms GeoSim 0 GeoSim 5 GeoSim 4 GeoContest

Toward more realistic models of civil wars Our strategy: Step I: extending Geosim framework Step II: conducting empirical research Step III: back to computational modeling

Step I: Modeling nationalist insurgencies Target Fearon & Laitin. 2003. Ethnicity, Insurgency, and Civil War. American Political Science Review 97: 75-90 Weak states that cannot control their territory are more prone to insurgency Use agent-based modeling to articulate identity-based mechanisms of insurgency Will appear in Cederman (forthcoming). Articulating the Geo-Cultural Logic of Nationalist Insurgency. In Order, Conflict, and Violence, eds. Kalyvas & Shapiro. Cambridge University Press. Political economists, working mostly in comparative politics, have energized the literature on civil wars. This has been a very positive trend: it has raised the bar considerably: * More coherent theories are being proposed, thanks to modeling * More systematic evidence is presented Two main problems have appeared, however: 1. There is a bias in favor of logistical and power-related mechanisms at the expense of identity-driven ones; 2. The cross-national research designs obscure contextual effects, such as diffusion I will address the first of these problems with the current paper. Quote Sambanis: “The gap between micro-level behavior and macro-level explanation is large. It is magnified when micro-macro relationships are studied solely through cross-national statistical analyses. What is often lost in such studies is information about causal pathways that link outcomes with causes.” Follow political economists’ lead in proposing formal models. Use agent-based modeling to articulate an alternative ID-based mechanisms of insurgency that produce similar macro-level patterns as those observed by political economists. Formal modeling of a slightly different type: opportunity-driven, where the actual configuration is more important Structural but not reified: structuration?

Step I: Main building blocks National identities Cultural map State system Territorial obstacles 3##44#2# 32144421 All this requires a complex model with at least four basic building blocks. Can only give you an overview -- we’ll have to use the Q & A period to discuss details I propose a layered ontology: First, at the bottom are the artificial mountains, that introduce obstacles to power extraction and projection in rugged terrain Second, there is the political heart of the model: a state system with states as hierarchical organizations that interact and can fight wars. In addition, capitals dominate provinces, but the latter can rebel and secede from the states. Third, there is a cultural map based on a vector representation. Each site in the grid, provinces or capitals, have their own culture represented as a string of 8 digits with four values. This can be seen as an abstract formalization of culture, be it language, religion, dress code, or any other ethnic trait. Fourth, in order to get nationalist politics, we need to move one more step up to national identities, that are coalitions based on combinations of cultural traits, where only a subset of the possible cultural raw material needs to be relevant. Wild cards are introduced for traits that do not matter. This means that we can formally model the thickness of identity. Each capital or province can belong to one nation only, with capitals being more likely to create their own national identities.

Step I: An artificial system

Step I: Conclusions Important hunches: Going beyond macro correlations Developing mechanisms based on explicit actor constellations Focus on center-periphery power balance Location of ethnic groups crucial But the model is too complex and artificial

Step II: Empirical research Beyond fractionalization (Cederman & Girardin, forthcoming in the APSR) Expert Survey of Ethnic Groups (Cederman, Girardin & Wimmer, in progress) Geo-Referencing of Ethnic Groups (Cederman, Rød & Weidmann, just completed) Modeling Ethnic Conflict in Center-Periphery Dyads (Buhaug, Cederman & Rød)

Step II: Constructing the N* index State-centric ethnic configuration E*: Micro-level mechanism M*: p(1) s1 EGIP p(2) s0 p(i) s2 … p(n-1) sn-1 Here it is: let us first assume that we have a start-like ethnic configuration with an ethnic partisan state at the center. Here we are not assuming that everybody interacts with everybody, but that there are as many dyads as there are peripheral actors. See ABM Based on opportunity structures, we can now model the probability that violence will break out as one minus the product that each dyad will remain peaceful. We now need to specify the dyadic probability of conflict, p(i). Let us assume that violence is more likely when the power balance is favorable for the challengers. This can be done by letting a logistical function that approaches 1 as the relative power balance tips in favor of the peripheral group i. Two parameters r and k determine the shape of the curve: r tells us where the threshold is located, and k fixes the curve’s steepness r(i)=

Step II: N* values for Eurasia & N. Africa Here is the world according to the N* index. Irak and United Arab Emirates are highlighted, but Afghanistan and large portions of the near abroad are also selected.

Step II: Expert Survey of Ethnic Groups Project together with Luc Girardin (ETH) Andreas Wimmer (UCLA) Web-based interface in order to expand coding of ethnic groups and their power access to the rest of the world with the help of area experts

Step II: Geo-Referencing of Ethnic Groups Scanning and geo-coding ethnic groups Polygon representation Based on Atlas Narodov Mira (1964)

Step II: Ethnic Dyads Calculating distances from capital

Step II: Ethnic Dyads Calculating mountainous terrain

Step II: Results from dyadic model

Step III: GROWLab Technical approach Main features Discrete spaces Follow same tradition as other toolkits, but higher level of abstraction Tailored to geopolitical modeling, but might be useful to others Java based; targeted at programming literates Main features Support for agent hierarchies Support for complex spatial relationships (e.g. borders) Support for GIS data (raster with geodetic distance computation) Discrete spaces Integrated GUI Comes with 13 example models Batch runs (cluster support in development) Available at: http://www.icr.ethz.ch/research/growlab/ Domain specific

Step III: GROWLab