Network Based Corruption Modeling
MetadataShow full item record
Development of relationships enabling corrupt behavior can be a major hindrance to the productivity of an organization with a limited capacity for agent activity monitoring. This paper presents the Network Based Corruption (NBC) model, offering a perspective on preventing malignant activity of connected agents. The agents are assumed to be strategically embedded into a directed peer-to-peer reporting network. Corruption is taken to be a threat whenever an agent and their supervisor(s) are all corruption-prone (ready to go corrupt); once they identify each other as such, which takes time, they can hurt the organization financially, going unnoticed. Under the NBC model, this paper addresses the policy-maker's problem of fixing the agent reporting structure and rotation timing so as to maximize the expected long-time profit. The impact of the NBC model parameters – proportion of corruption-prone agents in the network, productivity rate, costs, the rate of corrupt tie formation, etc. – on the choice of an optimal policy is analyzed. To enable real-world NBC analyses, it is shown that the problem of learning the hidden model parameters can be addressed by “feeling the system”, i.e., observing what financial outputs an organization generates when operating under different policy settings. Thus, an organization can experiment, and then, identify an optimal policy to limit NBC.