Understanding Conflict of Interest Networks
Social Network Analysis can be used to understand a wide variety of systems such as research, biological or technological networks. In particular, it is a great tool to observe and analyze conflicts of interest and assess the risks that arise in the evolving relationships between individuals or institutions. By using these tools, one could not only analyze patterns but also understand observed behaviors in networks of individuals.
In this blog post we will show how Social Network Analysis can be used to understand conflict of interests. For this purpose, we will use a real example of the network formed by the Management Board of the 50 largest non-financial companies in Colombia. We will describe several properties related to the topology of this network, as well as the possible implications of these metrics.
The original idea and diagnosis of the companies network was published in a joint effort of aentrópico and La Silla Vacía. A brief description can be found here and the original piece (in spanish) can be obtained here.
The main idea is the following: The most important connections inside the 50 largest companies in Colombia are revealed and explained. Starting with a demographical stratification of the different companies, we explored some important aspects of their DNA, including education levels of their board members as well as their female representation levels. A network view of the participation of highly connected individual across several boards is presented on the article.
A subset of the complete network is shown below and will serve for demonstration purposes of how Social Network Analysis can be used to understand conflict of interests. In this network companies are depicted as blue nodes while board members are orange. Centrality of a node is illustrated by increasingly darker tones of orange.
Management board networks
Interpretation of network properties
The network topology can give important insight on different aspects, including, but not limited to the structure of communities or information flows.
Clusters and bridges
A closer look to the top of the network reveals that Grupo Mundial and Grupo Argos have many members that belong to both boards, which reveals high connectivity between these two companies and thus they form a highly connected cluster. In this type of cluster, individuals tend to adopt the behavior of other individuals close to them and are resistant to outside influences.
There are other types of links offer different interpretations. If we look at Isaac Yanovich (darkest node) we see that he behaves as a bridge between two parts of the network. A bridge is a type of social tie that connects two different groups.
These nodes are of particular interest because they are central in the network in the sense that any information flow is expected to pass through them. In other words, any new information received by some node is very likely to come from a friend connected through a local bridge. Local bridges are important because they compose the shortest path between pairs of nodes in different parts of the network.
Nodes in a local bridge have riskier interactions in the network due to potentially contradictory norms and expectations from the different adjacent nodes associates.
Empirical studies of managers in large corporations have shown correlations of individual success within a company to their access to local bridges. Standing at one edge of a local bridge can also empower creativity and promote combination of multiple ideas.
Given their privileged access to a wide array of information sources, bridge nodes act as social gatekeepers and even prevent formation of triangles.
Triangles or triadic closures are important in networks as they are the simplest structure of a community. If two people in a social network have a friend in common it is very likely that they will become friends and start behaving similarly.
Note that in the network of Colombian companies we can notice two types of clusters:
The first type, as shown in the previous example, is formed because of the similarity between companies (e.g. Grupo Mundial and Grupo Argos).
However, other clusters can be formed due to the presence of highly connected individuals. That is the case of Mónica de Greiff, Henry Navarro, Fernando Gómez and Ricardo Bonilla who are all linked to each other by simultaneously belonging to the management board of four different companies (Promigás, EEB, Emgesa, Codensa).
Clusters or closed communities are resistant to outside influences. Hence behavioral changes like the adoption of a new technology or the modification of an existing social norm can be slowed down and even blocked by the boundaries of a densely-connected community.
However, if there are incentives to adopt behaviors from neighbors, things can change dramatically and cascading effects can emerge. Given the appropriate conditions, certain behaviors can easily propagate through the network. This is the case of corruption incidents, where generally incentives to change the prevailing social norm tend to be much higher when these changes can benefit all the individuals in the network.
By using peer pressure, one could promote adoption of behaviors inside a community, enhancing the diffusion of a certain social norm for all individuals of the network.
Conflict of interest emerge mainly in bridges, where the structural balance of a group or community can be compromised because of emerging behaviors pushed by adjacent communities. Sources of stress are often related to unbalanced triangles, where among 3 individuals, 2 adopt a behavior but the third one is still reluctant. However unbalanced triangles are not the norm since people try no minimize them by either changing their behaviors or breaking up links.
This post aims to be the starting point for a discussion on future research of conflict of interest based on Social Network Analysis.
Network Science Glossary
Affiliation network Two mode networks that allow one to study the dual perspectives of the actors and the events (unlike one mode networks which focus on only one of them at a time).
Bipartite graph: A graph that does not contain any odd-length cycles.
Bridge: An edge whose removal would lead to two distinct components. An edge is a local bridge whenever it is not in a triadic closure.
Cascading effect: An unforeseen chain of events due to an act affecting a system.
Centrality The various types of measures of the centrality of a vertex within a graph determine the relative importance of a vertex within the graph.
Cluster: Individuals that have a lots of connections with each other forming a closed community. This behavior has been observed in several networks: diseases, gossip, technology, etc.
Triadic Closure: The property among three nodes A, B, and C, such that if a strong tie exists between A-B and A-C, there is a weak or strong tie between B-C.
Topology: The arrangement of the various elements (links, nodes, etc.) of a network.