The Application of Regime Switching Copula Models In Asset Allocation

[[abstract]]本文研究MSCI 香港指數對MSCI 臺灣指數與對MSCI 太平洋指數(日本除外)的相 關性影響，並藉由此相關性進行投資組合。其樣本期間為2002 年1 月4 日至2012年12 月28 的週資料。為了捕捉事件發生後資產間相關性的變動，本研究模型以Regime Gaussian Copula GARCH 模型並比較Gaussian Copula GARCH 模型。本研究發現在1/n、MV、CVaR、MSR 投資組合的策略下，Regime Switching GaussianCopula GARCH 模型的的平均年報酬率明顯優於Gaussian Copula GARCH 模型，這表示若考慮到資產的狀態變化，可使投資績效變得更好，且無論是全樣本時期或金融風暴時期，CVaR 策略皆表現良好。[[abstract]]In this paper we study the dependence structure between the MSCI Hong Kong index and MSCI Taiwan indexor MSCI Pacific (ex Japan) index during 2002 to 2012 and use weekly data To capture the change of dependence structure after event we use Regime switching Gaussian Copula GARCH model and compare Gaussian Copula GARCH model We then adoptthe equal-weighted portfolio(1/n) the minimum-variance portfolio(MV) the maximum Sharpe ratio portfolio(MSR) and the conditional Value at Risk (CVaR) portfolio to find optimal weights Final according to four portfolio strategies the annual average return ofRegime switching Copula GARCH modelare much better than Copula GARCH model meaning that the portfolio performance can be better if we consider about two state of financial asset and the CVaR portfolio has good performance during entire period or financial crisi