RATIONAL EXPECTATIONS SOLVER

In previous versions of EViews, models assumed that the structure of the model allows us to solve each period of the model in sequence. This will not be true in the case where the equations of the model contain future (as well as past) values of the endogenous variables.

Since solving the model for any particular period requires both past and future values of the endogenous variables, it is not possible to solve the model recursively in one pass. Instead, the equations from all the periods across which the model is to be solved must be treated as a simultaneous system, and solving the model will require terminal as well as initial conditions.

EViews 14 introduces three methods for solving this class of models: Gauss-Seidel, E-Newton, and E-QNewton. All three are iterative procedures that attempt to reduce the change in the endogenous variables to zero as the model’s equations are solved repeatedly.

© 2024 S&P Global