- 學位論文
以非對稱動態相關模型探討我國股價及匯率報酬之相關性
A Study of the Dynamic Correlation between the Returns of Currency and Equity Markets in Taiwan by ADCC-MVGARCH Model
摘要
本研究主要利用Engle（2002）提出之DCC-MVGARCH模型與Cappiello（2004）發展出的ADCC模型探討近10年來台灣股市與匯率間報酬波動度的動態相關性。 過去大部分討論股價與匯率間關係的研究，均著重在報酬率（一階動差）關係上，較少探討波動性（二階動差）的關係。然而，在財務理論上，不論在資產評價或動態避險等，波動度（Volitility）均扮演很重要的角色。以往採用固定相關係數來衡量不同金融商品彼此間的影響，不僅不符合經濟直觀，且不被實際資料所支持，無法及時有效地掌握其變化。Engle在2002年發展出動態條件相關模型（Dynamic Conditional Correlation Model），簡稱DCC模型，允許相關係係數可隨時間而變動（time-varying）。而後至2004年Cappiello將對稱的DCC模型加以擴充，發展出非對稱的動態條件相關模型（Asymmetric Dynamic Conditional Correlation Model），簡稱ADCC模型，此模型將訊息衝擊效果的不對稱性納入模型中考慮。 本論文之研究結果發現在建構股價與匯率報酬率波動度的相關係數模型上，非對稱的ADCC模型表現較對稱的DCC模型為佳，前者較能捕捉市場上對於負面消息衝擊的影響，顯示兩者間有非對稱性的效果存在，即負面（壞消息）衝擊造成市場上變異的增加比正面（好消息）衝擊要來的大。
關鍵字
非對稱動態相關並列摘要
In this paper, we apply the Dynamic Conditional Correlation Multivariate GARCH (DCC MV-GARCH) model, proposed by Engle (2002), and Asymmetric Dynamic Conditional Correlation（ADCC）Model , proposed by Cappiello(2004), to investigate the effects of return volatility between currency and equity markets in Taiwan. Past studies mostly focus on the first moment (return), however, volatility(the second moment) plays a key role in many areas of finance, especially in asset pricing and dynamic hedging strategies. The hypothesis of a constant correlation of volatility among markets is likely to be incorrect and unfit for real data. It’s unable to grasp the changes effectively promptly. Engle developed the Dynamic Conditional Correlation（DCC）Model in 2002. Then Cappiello expanded the symmetrical DCC model to the Asymmetric Dynamic Conditional Correlation（ADCC）Model in 2004. This model integrates the impact effect's asymmetry in the model to consider. Our results show that the ADCC model is superior to the DCC model on construction return volatility between stock price and exchange rate in Taiwan. The ADCC model can catch the negative news impact influence in the market and shows that the asymmetric effect does exist in both markets.