Risk sharing for capital requirements with multidimensional security markets (Q2274226)

In this paper, the authors study an optimal risk sharing problem for capital requirements by considering heterogeneous agents who may adopt different capital adequacy tests and may concern different security markets. The conditions under which a representation agent exists are presented. Two classes of constraints, namely the polyhedral constraints and the distribution-based constraints, for capital adequacy are considered. Section 2 of the paper discusses risk measurement in the light of capital requirements, agent systems, optimal allocations and equilibria. Firstly, a risk measurement regime and its respective risk measure describing the attitude of individual agents towards risk are defined in Definition 2.1. Their definitions are based on an acceptance set, a linear security market with a finite-dimensional linear security space and a positive linear pricing functional, as well as the no-arbitrage condition in the security market. An example for a risk measurement regime is presented. Then the agent systems which describe the interaction of the individual agents and their respective capital requirements are described in an abstract single-period model, where aggregate net losses are modelled by a Riesz space. The heterogeneity in the risk assessments by the individual agents is described by the notion of order ideals in a Riesz space. Each of the individual agents is characterised by a risk measurement regime which consists of an acceptance set and a security market. Then an agent system consisting of n agents is defined in Definition 2.3 as an n-tuple of risk measurement regimes. Lastly, the optimal risk sharing problem and its solution for an agent system are discussed through introducing the concepts of attainable allocations and security allocations in Definition 2.5. The optimal risk allocation problem and its solutions for an agent system are then defined in Definition 2.6, where the concepts of Pareto optimality, an equilibrium allocation and an equilibrium price in the light of the optimal risk sharing problem are described. Section 3 discusses the mathematics for solving the optimal risk sharing problem. Specifically, the optimal risk sharing problem is related to an infimal convolution of the individual risk measures. The main results are presented in Section 3.1. Theorem 3.1 gives the representation of the infimal convolution as a market capital requirement for a representative agent. Theorem 3.3 and Proposition 3.4 give the conditions for the existence of a Pareto-optimal allocation and an optimal payoff. Theorem 3.5 gives the conditions for the existence of an equilibrium allocation and an equilibrium price. Section 4 discusses the determination of the optimal payoffs for the market, Pareto-optimal allocations and equilibria in polyhedral agent systems. The main results for optimal risk allocations are presented in Theorem 4.3 and Corollary 4.5. Section 5 considers the optimal risk sharing problem for agent systems with law-invariant acceptance sets. Theorem 5.3 and Corollary 5.5 give the conditions for the existence of an optimal payoff, a Pareto-optimal allocation and an equilibrium. The main result is then presented in Theorem 5.8. The existence of optimal portfolio splits is studied in Section 6. An attention is given to searching for cost-optimal portfolio splits in the presence of transaction costs. Theorem 6.1 gives the conditions for the existence of optimal portfolio splits. Corollary 6.2 provides the conditions under which the key assumption of Theorem 6.1 holds.