Font Size: a A A

Study On Fractal Based Trend Analysis And Asset Pricing Model In The Stock Market

Posted on:2013-05-26Degree:DoctorType:Dissertation
Country:ChinaCandidate:J L SunFull Text:PDF
GTID:1109330362967304Subject:Computer applications
Abstract/Summary:PDF Full Text Request
The dissertation starts with three questions:1) Does the Chinese stock marketdeserve to investment, can we share the benefits of economic growth in China throughthe investment of Chinese stock market?2) Is there any difference between Chinesestock market and developed country’s mature market?3) Can we explore a new in-vestment theory or model which is suitable for Chinese stock market? On the otherhand, the Efficient Market Hypothesis (EMH) has been discussed for several decades.According to EMH, it is impossible for any investor to get risk-free excess return fora certain long period. However, many asset pricing anomalies were discovered con-tinuously. A lot of researchers try to develop new pricing models to explain theseanomalies. In such a case, many interdisciplines began to emerge such as behavioralfinance, econophysics and so on. The purpose of this dissertation is to identify thefractal features of the Chinese stock market and to test the trend following models inthe Chinese stock market. At last, we proposed a new pricing model that was suitablefor the Chinese stock market. The primary task and contribution of this dissertationare as follows:First of all, the dissertation starts with a review of development from traditionalfinance theory to econophysics. Successively the review covers efficient market hy-pothesis, capital asset pricing model, anomalies, fractal and multifractal theory andfinally the trend filtering techniques.Second, we introduced the Fractal Market Hypothesis (FMH) and the fractal dis-tribution. The result of DFA analysis shows that the movement of market prices, trad-ing volume and volatility of the Chinese stock market behaves long-term correlations,which indicates that the linear efficient market hypothesis can not be adapted to deal-ing with the nonlinear phenomenon of the Chinese stock market. Then, we review the development of the technical analysis methods and trendfollowing techniques. We proposed a new trend filtering method which is based on thetime series’ fractal behaviors. The new method is applied to the Chinese stock marketpricing series, ranging from daily index data to high frequency stock future data andeven high frequency option data. The result shows that during the past several years,our new trend filtering method could earn stable excess return from Chinese stockmarket. With this evidence, we believe that there exists speculation premium in theChinese stock market which can be measured by the Hurst index.Finally, we proposed a Hurst index based fractal capital asset pricing model (F-CAPM). The traditional Markowitz’s portfolio selection theory studies how can therational investors choose portfolio on the σ r plain. Analogously, we study how canthe rational invertors choose portfolio on the h r plain and in the σ h r space. Wediscuss the characteristics of the feasible region and efficient surface in the σ h rspace. At last, we proposed the Hurst index based fractal capital asset pricing modeland established a new investment model for Chinese stock market. Empirical resultshows that our new model can get stable excess return from the Chinese stock market.In conclusion, financial study is about the human nature. It’s quite difficult toestablish a model that will work well under all the circumstances. This dissertationfocus on the Hurst index as a measurement of the speculation premium and try toestablish a pricing model based on the Hurst index. We believe that our study willuseful to some other researchers.
Keywords/Search Tags:Econophysics, Hurst index, Trend following, Fractalmarket hypothesis, Fractal capital asset pricing model, Chinese stockmarket
PDF Full Text Request
Related items