Linear regression is one of the primary tools for econometric and statistical analysis. There is, however, considerable evidence that nonlinear modeling is sometimes appropriate, especially in the analysis of macroeconomic relationships that are subject to regime change.

We describe here EViews tools for switching regression models—linear regression models with nonlinearities arising from discrete changes in regime. We consider settings with both independent and Markov switching where the sample separation into regimes is not observed. Dynamics specifications are permitted through the use of lagged dependent variables as explanatory variables and through the presence of auto-correlated errors.

While this chapter focuses only on models where the regimes are unobserved, related models with observed regimes are discussed in the Heckman chapter.