Download This PaperInference for Factor-Augmented Forecasting Regressions with Threshold effects
50 Pages Posted: 5 Jun 2019 Last revised: 27 Apr 2020
Date Written: May 17, 2019
Abstract
This paper introduces a factor-augmented forecasting regression model in the presence of threshold effects. We consider least squares estimation of the regression parameters, and establish asymptotic theories for estimators of both slope coefficients and the threshold parameter. Prediction intervals are also constructed for factor-augmented forecasts. Moreover, we develop a likelihood ratio statistic for tests on the threshold parameter and a sup-Wald test statistic for tests on the presence of threshold effects, respectively. Simulation results show that the proposed estimation method and testing procedures work very well in finite samples. Finally, we demonstrate the usefulness of the proposed model through applications to forecasting stock market returns and the annual growth rate of industrial production, respectively.
Keywords: Factor-augmented regression; Threshold parameter; Likelihood ratio statistic; Forecasting error
JEL Classification: C12, C13, C23
Suggested Citation: Suggested Citation