Identity and Trust in Covert Networks

This post is the first in a series I will be posting on models of identity, persuasion and leadership in conjunction with a seminar on the subject I am currently taking. I welcome any and all criticism on these initial research ideas.

Global criminal and terrorist groups do not organize in traditional top-down hierarchical structures of command and control. Rather, these covert organizations operate through sparse network structures, which have evolved naturally to support their security and operational needs. At the same time, terrorist and criminal groups face a significant asymmetric threat from military, law enforcement, etc., and thus are in constant need of new membership to replenish lost personnel and increase their level of available resources (e.g., financial, intellectual, physical, etc.). Given this asymmetry, and the intense scrutiny that terrorist networks receive, trust is the primary currency used to adjudicate both the formation of new of ties and the expulsion of “bad” ties in these covert networks. As a matter of practice, therefore, covert networks face a serious dilemma: under constant siege from outside infiltrators, network members must attempt to grow the network in order to survive while at the same time minimize their potential exposure to infiltrators.

In this brief research proposal I present a simple model of interactions among a layered set of actors, first belonging to one of two macro-types; A or B, as well as one of three micro-groups; coverts, enforcers or citizens. The model attempts to examine how members of covert networks cope with this dilemma by using both the perceived identity of actors and their current social network to explore how groups bound by the requirements described above emerge successfully or fail. This research is inspired by the agent-based framework proposed in Hammond and Axelrod 2006 and the network-based laboratory experiments presented by Kearns, et al 2009, and draws on substantive and methodological themes from both. The paper proceeds as follows, first the model of interaction among these groups is presented. Next, two potential implementations are proposed, a pure agent-based simulation and a laboratory experiment with human subjects. Potential advantages and disadvantages to these techniques are then discussed in the conclusion.

A Model of Covert Network Emergence

As stated, there are two types of players; either A or B, which can be members of one of three different groups; coverts, enforcers or citizens. As a practical matter, I restrict coverts to be of only one type[1], thus there are total of five type-group possibilities for players. The distribution of players in the game will model the asymmetry described above, such that the number of players from each group will follow coverts<enforcers<citizens and type will be distributed B<A. Before play begins, nature endows citizens players with a level of personal resource , coverts with , and enforcers ; therefore, coverts will have the lowest endowments on average, enforcers the highest, and citizens in between both. Player interactions occur on an N*N torus\footnote{This geometry is used so that all players have an identical number of immediate neighbors}, which is divided into four quadrants starting with 1 in the upper-left corner, and following 2 through 4 clockwise. Citizen players of both type are dispersed randomly throughout all quadrants. Coverts are dispersed randomly in quadrant 1, type A enforcers are dispersed randomly in 2, and type B enforcers likewise in 4, thus leaving quadrant 3 populated strictly by citizens at the outset. The use of space in this way is meant to model the proximal asymmetry for coverts versus enforcers; that is, ceteris paribus coverts should have a higher probability of forming ties with each other than any other group.

When play begins, each player must decide whether to form a tie with its immediate neighbors. Since type is common knowledge but group membership is not, players can observe its neighbors’ types, their endowment , and their ego-networks; meaning they can observe the type and endowments of those their neighbors are connected to and the density of ties among those second order connections[2]. Once the initial ties have been formed, coverts can move in any of the six directions available to them in two-dimensions in order to find new neighbors or form ties with their neighbors-neighbors. The play of enforcers and citizens resembles that of the queen and pawn players in chess respectively; such that enforcers can move to any position on the play grid at any round, and connect with one of their neighbors, while citizens move in one direction at a time and form ties with neighbors randomly. The primary conflict in the game is that coverts seek to form ties with each other coverts, then co-opt citizens to increase their resources; while enforcers seek to form ties with coverts in order to break up and destroy their networks. In each round of the game, coverts lose resources, which can only be replenished by creating ties with other coverts, or converting citizens players. If a covert enters a round with zero resources they “die” and are removed from the game and all network connections are severed. In order for coverts to “grow” their networks, they must first find another covert and form a dyad, then begin to close triangles with citizens. When a citizen is connected with two or more coverts it becomes a covert and the resources are shared. This is a key element of the game, as coverts must work to grow their networks in order to survive, but at the same time remain cognizant of invasion from a enforcers. If a covert forms a tie with an enforcer it is removed from the game and all of its network ties are severed.

To explore how identity and trust effect the success or failure of covert groups in this simple game I propose two different strategies for covert players. First, there is a resource maximizing covert that attempts to form ties with only those nearest neighbors that present the highest expected benefit based on their known level of , and their neighbors’. With such a strategy a covert would calculate the expected payoff from forming a tie with one of its neighbors based on the potential wealth the neighbor and its network neighbors could provide. The second strategy is a trust maximizing player, wherein a player decides to base a tie on the perceived ability to trust a neighbor, rather than the resources. There are many possible ways to calculate trust, but one simple way would be to only form ties with neighbors whose ego-networks meet some threshold of type membership. In this game, therefore, trust maximizing coverts would would only form ties with neighbors of type A that also some threshold meeting number of second order ties to type A’s.[3] The assumption driving this strategy is that identity plays a major role in trust, which has been observed across disciplines in the identity literature.

From this basic framework it may be possible to explore how identity and trust act as support mechanism for covert networks and allow them to emerge successfully, even in the face of constant threat from infiltrators. My hypothesis is that trust maximizing agents will emerge as the more successful agents, despite their disregard for the other player’s resources. In the next section I discuss two possible implementations of this model; first as a pure agent-based simulation, and then as a laboratory game played with human subjects.

Model Implementation

Given that this model seeks to explore the emergence of covert networks given two different strategies an agent-based implementation is a natural extension of the model. With the various types, environmental rules and player strategies defined, a simple computer simulation could be developed test the behavior of agents within the system and the emergence of covert networks. This would follow much of the work that has been done in agent-based modeling, particularly that of Axelrod, Hammond, Axtell, Laver and Fowler. There are several off-the-shelf tools that could be used to develop this ABM. Also, given the simplicity of this model, an independent software solution could be developed using any of several modern programming languages.

An alternative implementation would be to assign human subject to the type-group assignments above and play the game in a laboratory experiment. In this case, strategies would not be assigned to players, but rather game play would be observed to see what human strategies produce the best outcomes for covert types. It would also be interesting to have human enforcer players, as it is a well established in artificial intelligence research that human players outperform computer based strategies for even the simplest tasks. The implementation of human enforcer players also adds another dimensions of complexity, as human may develop survival strategies not predicted by the model as enforcers can please as they wish.

Each of these implementations presents several advantages and disadvantages, and I conclude by discussing a few of these issues. First, the ABM design is highly desirable because the model easily and completely fits into the framework. As with many evolutionary game theory models, where different play strategies are pitted against each other, the ABM design allows for quick prototyping and high scalability. Also, given that play is conducted on a static two dimensional surface, various ABM technologies could be used to visualize the emergence or decline of various player types in real-time. One significant disadvantage, however, is the difficulty in gathering and testing true comparative statics from the model. That is, one major criticism of the ABM methodology is that results are often “programmed in” because researchers report results with no counterfactuals. This would be an area of particular focus in the model described here, though new methods for fully exercising ABM models, as proposed by Laver, may be brought to this issue.

The laboratory experiment overcomes some of the limitations of the ABM design, but at a high cost. While ABM models are cheap and easy to implement, a laboratory experiment is difficult, time consuming, and expensive. The major advantage is the possibility that non-predicted strategies emerge as the best performers, which would allow for additional hypotheses and future tests. When comparing the two methods, however, the ABM designs seems to be the better starting point given its relative ease to build and relatively high inferential value.

[1]This assumption models the frequently observed identity divergence between members of terrorist or criminal organizations and those of the military or law enforcement communities. It is also a useful simplification, however, a more complex model could add a second covert type with competing interests.

[2]For a detailed definition of the ego-network see Wasserman and Faust 1994, section 2.3.3

[3]The actual operationalization of resource maximization and trust maximization are left open, as there are several different possible implementations, and the intent of this piece is to motivate the idea, rather than define specific strategy mechanisms.

Automatically Generated Related posts:

1. Identity and Trust in Covert Networks: A Laboratory Experiment
2. Networks, Collective Action, and Building State Capacity
3. The Influence of Secrecy on Covert Network Dynamics
4. Terrible Covert Network Structure?
5. Cultural Cleavages and Federalism in Iraq