本文以美國 S&P500 指數的月報酬率實證分析適應性市場假說 (the Adaptive Market Hypothesis, AMH) 之下，效率與非效率兩種市場氛圍的交替轉換。我們以馬可夫轉換模型的 filtering probability 評估兩種市場的動態關係，並由避險與投資兩個面向切入，實證適應性市場假說。本研究針對馬卡夫轉換模型所辨認的兩個市場，採用不同的避險比率，其在兩個市場的避險效率表現都較傳統 (naïve) 和最小平方法 (OLS) 避險為佳。由適應性市場擇時之實證，我們亦發現若針對轉換模型指出之非效率期間採行不同的策略，所得之累積報酬率都較買入持有策略 (buy and hold) 為佳，且在進一步以馬可夫轉化模型評估實際波動度 (realized volatility) 來囊括波動度擇時後，其累積報酬率更隨 之增加，擇時策略的報酬率在考量 Fama-French-Carhart 的四因子與交易成本後，仍然有顯著的超額報酬，再次呼應適應性市場假說之存在。;In this studies, we have delved into the monthly logarithmic returns of S&P500 index to characterize the transition between efficient and inefficient market under the Adaptive Market Hypothesis (AMH). By utilizing the Markov switching regression, we identify the efficient and inefficient market regimes. Through implementing different strategy in each regime, we find evidence for the AMH in the aspect of hedging and investment. For hedging, we apply distinct hedge ratio in two market regimes, and the AMH hedge outperforms the naïve and the OLS hedge. And for investment, besides estimating switching model of conditional returns, we also construct the switching model of conditional Realized Volatility (RV), which is inspired by the volatility timing. By applying different investment strategy during inefficient periods detected by the Markov switching model, the AMH timing portfolio outperforms the B&H portfolio with the regime regression of conditional returns, and the cumulative returns even increase after including the volatility timing by implementing different strategy during the sentimental periods identified with the switching model of conditional RV. Besides appealing cumulative returns, the AMH timing portfolio also has significant cross-sectional returns after explained by a four-factor model and considering transaction costs.
