This section describes methods of estimating the parameters of systems of equations. We describe least squares, weighted least squares, seemingly unrelated regression (SUR), weighted two-stage least squares, three-stage least squares, full-information maximum likelihood (FIML), generalized method of moments (GMM), and autoregressive conditional heteroskedasticity (ARCH) estimation techniques.

Once you have estimated the parameters of your system of equations, you may wish to forecast future values or perform simulations for different values of the explanatory variables. “Models” describes the use of models to forecast from an estimated system of equations or to perform single and multi­variate simulation.