風險與報酬具不對稱性關係－以市場輪廓理論分析八年期貨資料之實證

Abstract

自法瑪（Fama）定義效率市場假說以來，研究學者們紛紛發表了各種實證市場效率性的研究，而由於研究方法、資料期間，和研究對象及標的不同，是否具有市場效率性的結論也不盡相同，因此對於是否具有效率性市場的結論仍然莫衷一是。 本研究以臺灣期貨市場，八年期間長期的期貨資料為研究對象，根據市場輪廓理論，採用人工智慧方法中的倒傳遞類神經網路作為評估未來市場方向性的工具。期能在期貨市場中，發現低風險高勝率的交易位置，並探討臺灣期貨市場中風險與報酬間的關係，進而檢視臺灣期貨市場是否符合弱式效率市場假說。 研究結果顯示，以市場輪廓指標變數，輸入倒傳遞類神經網路進行學習及評估期貨市場方向的結果，的確在臺灣期貨市場上，發現了相對安全的低風險勝率高的進場價位。此實驗結果，不僅驗證了市場輪廓指標的有效性，也證實於臺灣期貨市場中存在風險與報酬不對稱的現象，因此推論臺灣期貨市場不符合弱式效率市場假說。

Since Fama defines the efficient market hypothesis, the researchers have published a study of empirical analysis of efficient markets, but because the different kinds of research methods, research data, and the research object, whether the market is efficient or not may also have different conclusions. Based on the theory of market profile, the study intends to analyze the eight-year Taiwan Stock Index Futures data using back-propagation neural network as an assessment tool for the future market direction. The study tries to find free exposure in the futures market, so as to achieve better profit point but low-risk trading position and discuss the relationship between risk and return in Taiwan futures market, then examine the Taiwan futures market meet the weak-form efficient market hypothesis. The research results show that indeed it finds out the relative low-risk trading price in the futures market. The experimental results not only verify the effectiveness of the market profile indicator, but also support the asymmetric relationships between risk and return exists in the Taiwan futures market, so that the Taiwan futures market does not meet the weak-form efficient market hypothesis.