Efficient solution techniques for linear and nonlinear rational expectations models (Q1118314)

Large-scale econometric models with rational expectations incorporated have to deal with the problem of obtaining simple and reliable numerical solutions. Things are even worse in the nonlinear case. Up to now only one solution technique - the iterative technique proposed by \textit{P. A. Anderson} [J. Monetary Economics 5, 67-80 (1979)] and \textit{R. C. Fair} and \textit{J. B. Taylor} [Econometrica 51, 1169-1185 (1983; Zbl 0516.62097)] - is capable to solve nonlinear large-scale rational expectations models. The authors present a general class of iterative algorithms of which the above mentioned one is a special case. Sections 2 and 3 present the iterative solution technique to linear rational expectations models and discuss convergence conditions as well as efficient computational forms. Section 4 extends the analysis to the nonlinear model where instead of stationary now non-stationary iterations are applied. How to set terminal conditions to obtain a unique solution is briefly discussed in the following section. An application of the technique to two large-scale nonlinear models of the U.K. economy is found in the final part of the paper.