A0843
Title: Threshold regression for large datasets with common stochastic trends
Authors: Daniele Massacci - Kings College London (United Kingdom) [presenting]
Lorenzo Trapani - Cass Business School (United Kingdom)
Abstract: Inference is studied for threshold regression in the context of a large panel factor model with common stochastic trends. We develop a Least Squares estimator for the threshold level, deriving almost sure rates of convergence and proposing a novel, testing-based, way of constructing confidence intervals. We also investigate the properties of the PC estimator for the loadings and common factors in both regimes, and develop a procedure to estimate the number of common trends in each regime. Although the main focus is on common stochastic trends with mean zero, we show that our technique can be applied even in the presence of common factors with drifts and/or trend stationary common factors. The theoretical findings are corroborated through a comprehensive set of Monte Carlo experiments. Finally, the analysis of long-run returns from a set of buy and hold strategies shows the usefulness of our model for empirical work.
