Estimating Markov-Switching ARMA models with an Extended Algorithm of Hamilton
2007/09/11
研討會日期 : 2007-09-11
時間 : 15:00
主講人 : Prof. C.C Chen
地點 : B棟110室
演講者簡介 : Prof. C.C Chen 為國立政治大學金融學博士（2005）。
現為私立東海大學財金系助理教授。
其主要研究領域為財務工程、計量經濟學。
演講摘要 : This paper proposes two innovative algorithms to estimate a general class of N-state Markov-switching autoregressive moving-average (MS-ARMA) models with a sample of size T. To resolve the problem of possible routes induced by the presence of MA parameters, the first algorithm is built on the original algorithm of Hamilton (1989) and the idea of Gray (1996) by replacing the lagged error terms with their conditional expectations. We thus name it as the Hamilton-Gray (HG) algorithm. The second algorithm refines the HG algorithm by continuously updating the conditional expectations of MA terms and is named as extended Hamilton-Gray (EHG) algorithm. The computational cost of both algorithms are very mild because the implementation of these algorithms are very much similar to that of Hamilton (1989). The simulations show that the finite sample performance of the EHG algorithm is satisfactory and is much better than that of the HG counterpart. We thus apply the EHG algorithm to the issues of dating U.S. business cycles with the same real GNP data employed in Hamilton (1989). The turning points identified with the EHG algorithm resemble closely to those of the NBER's Business Cycle Dating Committee, and thus confirms the robustness of the finings reported in Hamilton (1989) about the effectiveness of Markov-switching models in dating U.S. business cycles.
