Recent LIBOR manipulation scandal inspired discussions of the proper design of survey-based benchmarks. I solve a version of the survey as a game of incomplete information in which the benchmark administrator is unaware of the distribution of private signals. I characterize a standard survey in...