金磚五國之期貨避險績效－－動態Copula-GJR-GARCH模型應用

The Hedging Performance for BRICS Futures-Applying the Dynamic Copula-GJR-GARCH Model

本文主要採用動態的Copula GARCH模型估計現貨與期貨報酬的最小變異避險比率，並和傳統避險模型、固定條件相關係數之CCC-GJR-GARCH模型以及動態條件相關條數之DCC-GJR-GARCH模型進行各個模型的避險績效之比較，找出最適的避險比率和最佳的模型。實證應用以金磚五國的現貨與期貨為研究資料，包括巴西IBOVESPA指數、俄羅斯RTS指數、印度S&P CNX NIFTY指數、中國CSI3 00指數以及南非FTSE／JSE Sharehold Weighted Top40指數。實證結果發現，除了中國CSI300指數外，其它四國在樣本內及樣本外的績效評估檢定下，均以動態Copula GJR-GARCH模型的避險績效較其其他模型佳。

This paper applies dynamic copula GJR-GARCH models for the estimation of the minimum-variance hedge ratio and compares their effectiveness with the conventional static, the constant conditional correlation (CCC) GJR-GARCH, and the dynamic conditional correlation (DCC) GJR-GARCH models. In empirical study, we uses the BRICS futures and spot data, including the IBOVESPA index, the RTS index, the S&P CNX NIFTY index, China Securities Index 300 (CSI 300) index and FTSE/JSE Shareholder Weighted Top40 index spots and futures. Empirical results show that no matter in the training sample or test sample, the dynamic copula GJR-GARCH models can outperform other models except for CSI 300 index.