A Reliable and Computationally Efficient Algorithm for Imposing the Saddle Point Property in Dynamic Models

This paper describes a set of algorithms for quickly and reliably solving linear rational expectations models. The utility, reliability and speed of these algorithms are a consequence of 1) the algorithm for computing the minimal dimension state space transition matrix for models with arbitrary numbers of lags or leads, 2) the availability of a simple modeling language for characterizing a linear model and 3) the use of the QR Decomposition and Arnoldi type eigenspace calculations. The paper also presents new formulae for computing and manipulating solutions for arbitrary exogenous processes.