Recursive solution methods for dynamic linear rational expectations models

This paper develops recursive solution methods for linear rational expectations models. The underlying structural model is transformed into a state-space representation, which can then be used to solve the model and to form the Gaussian likelihood function. The recursive solution method has several advantages over other approaches. First, the set of solutions to the model are summarized by a set of parameters that appear in the state-space representation but are unspecified by the structural model. Next, the likelihood function is formed as a byproduct of the solution to the model. Finally, modifications in the likelihood function necessary to incorporate complications arising from temporal aggregation, dynamic errors-in-variables, etc. are straightforward.