This paper provides a method to formally evaluate models with calibrated parameters. We examine whether the restricted VAR representation implied by the model is consistent with the data. Two types of restrictions are analyzed: one due to the fact that there are less forcing processes than endogenous variables and one obtained factoring out the first type of restrictions. We also propose a way to evaluate the performance of a model relative to a competitor. An application to the model analyzed by Burnside, Eichenbaum and Rebelo
(1992) is considered.