Restrictions and identification in a multidimensional risk-sharing problem (Q2249579)

The multivariate case of the risk-sharing problem is considered where a resource (i.e. a risky aggregate multivariate endowment, e.g. wheat and meat production) is to be shared by several expected utility maximizers. It is shown that there are some sharp restrictions on efficient risk-sharing among the expected utility maximizers that take the form of systems of nonlinear PDEs. Moreover, an identification theorem is proved, namely under a rank condition, the knowledge of an efficient risk-sharing enables to reconstruct some sharp information on individual preferences and Pareto weights.