This paper studies optimal policy in models with multidimensional uncertainty and endogenous observables. We first consider a very general setup where the policy-maker does not observe the realisations of the shocks that hit the economy, but only some aggregate variables that are endogenous with...