Title: Semi-parametric single-index predictive regression models with cointegrated regressors
Authors: Weilun Zhou - University of Cambridge (United Kingdom) [presenting]
Jiti Gao - Monash University (Australia)
David Harris - University of Melbourne (Australia)
Hsein Kew - Monash University (Australia)
Abstract: The estimation of a semi-parametric single-index predictive regression model in the presence of cointegration among the multivariate predictors is considered. This model is useful for predicting financial asset returns, whose behaviour is compatible with a stationary series, when the multiple predictors are nonstationary, and also allows for nonlinear predictive relationships. The single-index specification, which contains the cointegrated predictors, not only solves the problem of unbalance in the predictive regression, but also avoids the problem of the curse of dimensionality associated with fully nonparametric multivariate models. An orthogonal series expansion is used to approximate the unknown link function for the single-index component. We consider the constrained nonlinear least squares estimator of the single-index (or the cointegrating) parameters and the plug-in estimator of the link function, and derive their asymptotic properties. In an empirical application, we find some evidence of in-sample nonlinear predictability of U.S. stock returns using cointegrated predictors. We also find that the single-index model in general produces better out-of-sample forecasts than both the prevailing mean model and the linear predictive regression model.