An Algorithm of Propagation in Weighted Directed Graphs with Applications to Economics and Finance

This paper puts forward an algorithm that computes the diffusion of events and actions across networks of economic agents, an algorithm that is applicable when such networks can be represented as weighted directed graphs. The functioning of the algorithm is shown in three applications. First, the algorithm is applied to a model of diffusion of innovation driven by the agents' imitation of their neighbors' behavior. Second, a graph-theoretic model of financial networks is introduced, and the corresponding algorithm is used to compute the so called domino effect, i.e., the diffusion of losses and insolvencies caused by the initial default of one or more agents. Finally, the algorithm is applied to the transfer of deposits operated by interbank liquidity networks. (C) 2010 Wiley Periodicals, Inc.