Hamed Amini
University of Florida
Room: M3 3127
Ruin Probabilities for Risk Processes in Stochastic Networks
We study networked multidimensional risk processes in which agents, located on a large sparse random graph, incur losses from their neighbors. To reduce the dimensionality of the system, agents are classified according to an arbitrary countable set of types. The ruin of an agent generates losses for its counterparties, and the resulting loss-revelation mechanism follows a Poisson process with a general intensity function that scales with the network size. As the network size tends to infinity, we derive explicit limiting fractions of defaulted agents and characterize the corresponding ruin probabilities for a representative agent of each type at the end of the loss propagation process. These limits depend on the agents’ types, the network structure, the loss distribution, and the specification of the reveal intensity. The talk is based on joint work with Zhongyuan Cao, Andreea Minca, and Agnès Sulem.