科学研究

This paper studies estimation and inferences in multiple regime panel threshold regression with unobserved individual-specific threshold effects. These effects are important from the practical perspective and define a distinguishing feature from traditional linear panel data models. It is shown that the withinregime demeaning in the static model or the within-regime first-differencing in the dynamic model cannot generate consistent estimators of the threshold, so the correlated random effects models are suggested to handle the endogeneity in such general panel threshold models. We provide a unified estimation and inference framework that is valid for both the static and dynamic models and regardless of whether the unobserved individual-specific threshold effects exist or not. Specifically, we suggest model selection based on sequential testing which allows for underestimation of the number of regimes, develop the asymptotic theory for the least squares estimator that is robust to such model misspecification, and propose new inference methods for the model parameters which have better theoretical properties than the existing methods. Simulation studies and an empirical application illustrate the usefulness of our new model selection, estimation and inference methodology in practice.