|Abstract: ||In this paper we introduce a new notion, that of "conditional risk measure", which has properties reminiscent of the conditional expectation, in order to deal with situations of partial information, or asymmetric information. We do so in a general context of uncertainty, where there is not necessarily an a priori given probability.
Important results of this paper are theorems of representation.
We study in detail a particular example important in finance, that
of a conditional risk measure associated to a loss function.