ABSTRACT
Among the solution methods described in Sections 2 and 3, some may be extended to treat the case of (univariate or multivariate) models involving various expectations. The martingale differences approach allows for the substitution of any expectation in terms of a realization and a prediction error. Starting from the initial structural model, this method leads to a dynamic equation including observable variables and prediction errors. This equation is then used for completely solving the model. This section will begin with some examples. It will be emphasized that, when some dynamic complications arise, it is necessary to impose some constraints on the successive prediction errors. An illustration will give an intuition idea of the general result concerning the solution of set of any linear univariate rational expectations model. After describing this general result, we will apply it to the search for stationary solutions.
