市場效率性長久以來一直是全球財務學者所感興趣的議題，也是各國政府極力想達成的目標。本研究藉由美國國家不動產投資信託協會所編製之美國不動投資信託（Real Estate Investment Trusts）指數，探討2001年01月03日至2008年10月31日間，其四種REITs指數在傳統OLS模型、GARCH模型與GARCH-Level模型下，市場報酬與波動性之星期效應（Day-of-the-week Effect）變化為何？是否存在所謂的週一效應（Monday Effect）或稱週末效應（Weekend Effect）。此外本篇研究加入次級房貸風暴之虛擬變數以探討次級房貸風暴對於REITs的報酬與波動性是否存在影響性。本篇研究也特別考慮利用VaR模型進行風險值的預測，以比較最適模型在應用上是否有其一致性。 實證結果顯示，使用GARCH-Level模型對於市場報酬與市場波動度的解釋能力優於傳統OLS模型與GARCH模型，解釋能力相較於其他模型有明顯的提升。在星期效應的實證上發現與故去文獻研究有一致性。在波動度星期效應的研究上發現，All REITs指數與Equity REITs指數在週二、週三與週四的波動度有顯著的異常現象。Hybrid REITs指數在週一與週五的波動度有異常性。Mortgage REITs指數則是五天波動度皆有異常性。形成此一現象的因素可能來自受到次級房貸風暴的影響而有此一結果。此外次級房貸風暴對於REITs的報酬並無顯著的影響，但對於波動度卻有明顯的影響性。另外在VaR的檢驗上，GARCH-Level模型在考慮波動度後，其風險預測能力相較於傳統GARCH模型來的準確，顯示GARCH-Level模型在實務的應用上依然為最適模型。 Market efficiency has been an interesting financial issue. We examine the day-of-the-week effect of return and volatility for American REITs (Real Estate Investment Trusts) index data from January 3, 2001 to October 31, 2008 by traditional OLS model, GARCH model, GARCH-Level model and VaR model. And then we also examine that the Subprime Mortgage event affects the return and volatility of REITs. Our results show that using GARCH-Level model has better explanation of market return and market volatility than OLS model and GARCH model, that is, all REITs index, equity REITs index and hybrid REITs index have abnormal return on Monday, Wednesday and Friday, but mortgage REITs index have on Monday and Tuesday, and this finding is not only consistent with other studies but also the explanation is improved than other models. The results also indicate that Subprime Mortgage Crisis has significant effect to volatility of REITs. Besides, the examination of VaR model shows that the prediction ability of risk of GARCH-Level model is more accurate than other models, which means GARCH-Level model still is the optimal model for practical application.
