Predicting Board Interlock Network Ties Using Proximity Measures

Authors: Panagiotis Giannakis*, University of Arkansas, Jackson Cothren, Advisor
Topics: Business Geography
Keywords: Proximity, networks
Session Type: Interactive Short Paper
Day: 4/11/2018
Start / End Time: 1:20 PM / 3:00 PM
Room: Edgewood AB, Sheraton, 4th Floor
Presentation File: No File Uploaded


Board interlocks are an important mechanism that enables the flow of information and resources among organizations, having an impact on organizational structure, strategy, and performance. Despite the rich literature regarding the benefits of board interlocks, limited is the research regarding the factors that trigger the creation and/or dissolution of these ties. The purpose of this proposed study is to shed light on this phenomenon by drawing on the concept of proximity and by employing advanced contemporary analytical tools that require massive computational power. The five dimensions of proximity, namely geographic, organizational, institutional, social, and cognitive, can possibly provide answers, as they incorporate indicators of the individuals’ (e.g., social) and organizational (e.g., organizational and institutional) proximity, thus being capable of explaining the evolutionary behavior of the interlocks network. All five dimensions, are expected to decrease the monitoring and agency costs between two connected organizations, facilitating a better cooperative relationship with less frictions on the flow of information and resources, in an overall more trustful environment. More than that, it is theorized that the geographic proximity will positively interact with every other aspect of proximity, enhancing their effects on the network. Taking into consideration that networks are not static over time, it is argued that the future structure of the network may depend on its current state and the existence or absence of ties between the actors. Due to this auto-dependence, exponential random graph models, a specific class of social network analysis methods, will be employed, since they incorporate such inter-dependencies.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login