In this paper, we examine the cost of insurance against model uncertainty for the Euro area considering four alternative reference models, all of which are used for policy-analysis at the ECB. We find that maximal insurance across this model range in terms of a Minimax policy comes at moderate costs in terms of lower expected performance. We extract priors that would rationalize the Minimax policy from a Bayesian perspective. These priors indicate that full insurance is strongly oriented towards the model with highest baseline losses. Furthermore, this policy is not as tolerant towards small perturbations of policy parameters as the Bayesian policy rule. We propose to strike a compromise and use preferences for policy design that allow for intermediate degrees of ambiguity-aversion. These preferences allow the specification of priors but also give extra weight to the worst uncertain outcomes in a given context.