ABSTRACTSeveral techniques for solving dynamic rational expectations models have been proposed. This paper puts forward an alternative method for discrete time models, which is significantly simpler to use in practice. That solution is used to derive and compute optimal policy selections (incorporating ‘noncausal’ effects) which, by exploiting the discrete time framework, are also time consistent when sequentially reoptimized. Those decisions are shown to contain an optimal open loop component plus an innovations dependent correction mechanism. A numerical example is used to verify these properties, and to demonstrate the superiority of this policy selection technique over recursive methods (e.g. dynamic programmi