透過您的圖書館登入
IP:3.19.30.232
  • 學位論文

碎形市場假說下分析工具之研究與探討

The Study of Analytical Technique under Fractal Market Hypothesis

指導教授 : 邱顯比

摘要


本研究選取1999到2008十年間之台灣加權股價指數為研究對象,利用碎形市場假說下的兩大分析工具 ─冪係數與赫斯特指數─ 進行研究與探討。碎形市場假說的主要目的為修正效率市場假說的許多不合理假設,並發展出穩健的分析工具以捕捉金融市場的兩大特性:巨幅波動、正持續性;其中巨幅波動程度以冪係數衡量之,序列持續性則以赫斯特指數衡量之。 除計算兩數值並解釋其意涵外,本研究以碎形布朗運動建立一模擬模型模擬未來指數的走勢,並從兩方面去檢視模擬結果是否良好: 1) 針對特定模擬長度比較碎形布朗運動與幾何布朗運動所模擬1000次的結果,並求算兩模擬模型涵蓋真實指數下方風險的機率; 2) 分別計算兩模型模擬指數之冪係數,並檢定是否與真實冪係數存在顯著差異。若無顯著差異,則隱含模擬指數結果貼近真實指數。 實證結果顯示,採用過去四年的歷史資料為樣本,建立模擬模型模擬未來指數的走勢,比較碎形布朗運動與幾何布朗運動所模擬1000次的結果,有考慮時間序列長記憶性的碎形布朗運動模擬結果較無考慮時間序列長記憶性的幾何布朗運動佳。此外,本研究也針對模擬指數進行α值的計算,並檢定模擬指數之冪係數是否與真實冪係數存在顯著差異,證明碎形布朗運動模擬指數之波動程度較幾何布朗運動更貼近實際指數波動情形,能更真實且相對準確地模擬指數。職是,α值與H值為相輔相成之分析工具,碎形布朗運動模擬指數的結果不僅考慮了長期記憶性,亦考慮了價格波動性,均較以往的分析工具更符合真實世界的應用。

並列摘要


This paper analyzes the daily returns of Taiwan Weight Stock Index over the 10-yr period 1999-2008, by applying the analytical technique under Fractal Market Hypothesis (FMH) -- power coefficient (α) and Hurst exponent (H). The purpose of FHM is to remove the unrealistic assumptions under Efficient Market Hypothesis (EMH). FMH tries to develop some robust analytical tools to catch the two characteristics in financial markets – large price fluctuation and the long-memory effect. Besides calculating the two statistics, I establish a simulation model with fractional Brownian Motion (FBM), and examine it from two aspects: 1) compare the 1000 routes simulated by FBM and GBM (Geometric Brownian Motion) respectively to the real index at the specific dates, and calculate the probabilities if the simulated outputs contain the down-side risk of real index; 2) calculate the α coefficients of the two simulation models, and test whether they are different from the real world α significantly or not. The empirical results show that the simulations with FBM dominate the ones with GBM using the four-year historical data, because the FBM simulation models take the long-memory effect into account. Furthermore, the fluctuation of the simulations with FBM is closer to the real index. That is to say, the FBM models could simulate the index efficiently and realistically better than GBM models in both price fluctuation and long-memory effect aspects.

參考文獻


6. 張正利 (2007),「利用赫斯特指數及有效報酬衡量國內股票型共同基金績效表現」,國立臺灣大學財務金融研究所碩士論文。
5. 貝諾.曼德伯 (2007),《股價、棉花與尼羅河密碼》,台北:早安財經文化。何信慧 譯自Mandelbrot, B. B. (2004), The Misbehavior of Markets: A Fractal View of Risks, Ruin, and Reward.
7. 尼可拉斯.塔雷伯 (2008),《黑天鵝效應》,台北:大塊文化。林茂昌 譯自Taleb, N. N. (2007), The Black Swan: The Impact of the Highly Improbable.
1. Bachelier, L. (1900), “Theorie De La Speculation,” Ann Sci. Ecole Norm. Sup., Vol. 17, pp. 20-86.
2. Backus, D. K. and S. E. Zin (1993), “Long-memory Inflation Uncertainty: Evidence from the Term Structure of Interest Rates,” Journal of Money, Credit and Banking, Vol. 25, pp. 681-700.

被引用紀錄


胡勝綸(2013)。利用局部赫斯特指數預測國內股票市場〔碩士論文,國立臺灣大學〕。華藝線上圖書館。https://doi.org/10.6342/NTU.2013.01014

延伸閱讀