Financial risk measures for a network of individual agents holding portfolios of light-tailed objects (Q2274222)

Studying a network of agents which share financial risks by holding portfolios with different objects that are subject to random losses, the authors focus on a system where individual agents, e.g. insurance companies or investment funds, form a financial network by holding portfolios of risky objects. As holding risky portfolios may lead to extreme losses, it is of importance to quantify the tail behaviour of the portfolio losses in this network context. The paper explores risk aggregation and risk sharing issues for portfolios of light-tailed losses in financial networks. The authors consider independent light-tailed object losses and only assume that they are asymptotically exponentially distributed. They develop a novel concept of functional exponential mixtures. Then the theoretical findings are applied to analyse extreme loss situations and to present expressions on marginal quantile-based risk measures. The interdependence of individual and system risks within the network is quantified by deducing results on conditional risk measures. Finally, the findings are summarised for portfolios of light-tailed losses and compared with those established for heavy-tailed ones.