Regime Switching Models and Multiple Thresholds Cointegrations.

摘要

Threshold cointegration has been a vibrant research topic in finance and statistics. Estimation procedures of threshold cointegrated models are usually based on the so-called threshold vector error correction forms (TVECMs) for one threshold case. In this thesis, we investigate two estimators for multiple thresholds cointegrations via TVECMs, namely the least squares estimator and the smoothed least squares estimator. The convergence rate of the least squares estimator is obtained and limiting distribution of the smoothed least squares estimator is developed. To assess the performance of these two estimators, we conduct a simulation study, the result of which supports the asymptotic theories developed. We study the term structure of interest rates by a two thresholds cointegration as an example.;Finally we also investigate the least squares estimator of smooth transition cointegration and establish the limiting distribution.