Coherent and convex monetary risk measures for unbounded càdlàg processes. (Q2488485)

The authors prove a theorem which shows that the requirement that coherent convex risk measures be real valued is too restrictive and it is better to let them take values in \((-\infty,\infty]\). This leads to definition of the coherent and convex monetary risk measures on \(\mathcal{R}^0\). The main result gives a characterization of coherent and convex monetary risk measures on \(\mathcal R^\infty\) that can be extended to the coherent and convex risk measures on \(\mathcal R^0\). Two examples of coherent and convex risk measures on \(\mathcal R^0\) are given. The first one is related to the Cramér-Lundberg approach to measuring the risk of an insurance company. The second one is motivated by results of F. Delbaen, where the time consistency properties of the dynamic risk measures that depend on one-dimensional random variables are studied.