本文主要在探討美國銀行產業在利率風險和匯率風險下的投資組合動態避險績效。除了投資組合效果之外，本文同時考慮在避險文獻上較少被考量的狀態轉換效果，並採用Billio 與 Caporin (2005)提出的多變量馬可夫狀態轉換動態條件相關GARCH (Multivariate Markov Switching Dynamic Conditional Correlation GARCH, MS-DCC GARCH)模型來比較幾個在文獻上常被使用的模型之避險績效。實證結果顯示：投組避險相對於個別避險，在樣本外有較佳的避險績效；而且，當我們考慮狀態轉換的效果後，此績效能夠更進一步的提升。因此，當銀行在建構利率和匯率的避險策略時，確有必要同時考量投組避險和狀態轉換這兩個效果。 This paper investigates the dynamic portfolio hedging effectiveness of US banks which are exposed to both interest rate and currency risks. To date, the investigation focusing on bank’s hedging effectiveness under portfolio effect and regime switching effect is limited. This paper tries to fill this literature gap by applying a Multivariate Markov Switching Dynamic Conditional Correlation GARCH (MS-DCC GARCH) model to examine whether taking into account of both effects improves bank’s hedging effectiveness. Empirical results show that portfolio hedging is superior to separate hedging out-of-sample in terms of both return maximization and risk minimization. The hedging performance is further improved when we incorporate the regime switching effect. This shows the importance of incorporating both effects of portfolio and regime switching for bank while constructing hedging strategies to hedge both interest rate and currency risks.