Universal estimators of a vector parameter (Q791240)

\(X_ 1,...,X_ n\) are independent and identically distributed random variables, the common distribution depending on unknown parameters \(\theta_ 1,...,\theta_ m\). The problem is to find a characterization of the distributions of \(X_ 1\) and the generalized prior densities for \((\theta_ 1,...,\theta_ m)\) for which the generalized Bayes estimator of \((\theta_ 1,...,\theta_ m)\) is the same for all loss functions in a certain class of symmetric loss functions. For the case \(m=1\) the author solved this problem in his paper Ann. Stat. 6, 1345-1351 (1978; Zbl 0392.62023). The present paper gives the solution for general m.