Non-stationary iterative methods for solving macroeconomic numeric models

Macroeconometric modeling was influenced by the development of new and efficient computational techniques. Rational Expectations models, a particular class of macroeconometric models, give raise to very large systems of equations, the solution of which requires heavy computations. Therefore, such models are an interesting testing ground for the numerical methods addressed in this research. The most difficult problem is to obtain the solution of the linear system that arises during the Newton step. As an alternative to the direct methods, we propose non-stationary iterative methods, also called Krylov methods, to solve these models. Numerical experiments conducted by authors confirm the interesting features of these methods: low computational complexity and storage requirements.