Chapter 3 Computational methods for macroeconometric models

This chapter discusses computational methods for the estimation and analysis of macroeconometric models. The chapter focuses on methods that, while possibly computationally routine, are not trivial. Most of the methods discussed are methods for complete models. The results reported in Fair and Taylor (1990) using the methods discussed in this chapter are encouraging regarding the use of models with rational expectations. FIML estimation seems computationally feasible using the procedure of computing derivatives, for the expectations and stochastic simulation , when done in the manner described in the chapter. FIML estimation is particularly important, because it takes into account all the nonlinear restrictions implied by the rational expectations hypothesis. It is hoped that the methods discussed in this chapter will open the way for many more tests of models with rational expectations.