The structure of ARMA solutions to a general linear model with rational expectations (Q1098538)

Forward looking linear rational expectations models are usually played with the problem of multiple solutions. Various authors (e.g. J. F. Muth, R. E. Lucas, O. J. Blanchard etc.) suggested techniques to obtain a unique solution. This paper contributes to this line of research by describing the set of all stationary and non-stationary ARMA solutions to a general univariate linear model, in which the endogenous variable depends on expectations of its values at arbitrary dates. The advantage of focussing on the ARMA solutions is that it admits an explicit parametrization in terms of a finite number of parameters. Part 2 of the paper discusses properties of ARMA processes, part 3 relates these results to linear rational expectations models and derives the class of ARMA solutions to these models. Examples are given in the final part of the paper.