The Epidemics of Corruption

Abstract

We study corruption as a generalized epidemic process on the graph of social relationships. The main difference to classical epidemic processes is the strong nonlinear dependence of the transmission probability on the local density of corruption and the mean field influence of the overall corruption in the society. Network clustering and the degree-degree correlation play an essential role in corruption dynamics. We discuss phase transitions, the influence of the graph structure and the implications for epidemic control. Structural and dynamical arguments are given why strongly hierarchically organized societies like systems with dictatorial tendency are more vulnerable to corruption than democracies. A similar type of modelling can be applied to other social contagion spreading processes like opinion formation, doping usage, social disorders or innovation dynamics.