【作者】 郭尧琦；

【导师】 黄健柏；

【作者基本信息】 中南大学 ， 管理科学与工程， 2012， 博士

【摘要】 从2002年开始,我国已超过美国,成为了世界上最大的有色金属生产国和消费国。随着城镇化、工业化快速发展和内需拉动,我国有色金属资源需求量将继续增长,其基础战略资源地位日益重要。有色金属产业能否持续健康发展己成为影响我国经济增长和国防安全的重要因素。然而,近年来以铜为首的有色金属价格波动剧烈和频繁,形成了巨大的市场风险,也严重影响了我国资源安全保障。因此,在这一背景下,分析有色金属价格的波动行为,测度市场蕴含的波动风险具有重要的现实意义。然而,现有研究基本是以传统的有效市场假说为基础,其较为严格的假设条件使得该理论不能够全面有效地解释诸多金融现象。而分形理论作为市场复杂性研究的前言理论,则为研究市场价格行为提供了一个新的理论框架。在这一背景下,本文将分形市场理论引入到对有色金属价格的波动行为的分析中,从一个全新的视角对有色金属价格波动行为进行识别和刻画,并将分形特征参数应用于对市场蕴含的波动风险的测度中,具有重要的理论和现实意义。通过上述研究,本文得出了以下结论：(1)金属市场是一个具有明显分形特征的复杂系统。沪铜和沪铝收益序列均呈现出明显的“尖峰、厚尾、有偏”的特征,不服从正态分布,这说明传统的有效市场理论将不能更好的对金属市场进行更深层次的刻画及理解。而不同时期的金属收益序列的分形维均处于1-2之间,则说明金属市场不遵循随机游走过程,是一个具有分形结构的市场,其价格行为是一个趋势加噪声的有偏随机游走过程。而市场存在分形特征的原因正在于市场的长期记忆特征和“尖峰厚尾”分布。因此,本文将分形市场理论引入到对有色金属价格波动行为的分析中,具有重要的理论和现实意义。(2)金属期货市场仍需进一步发展完善。通过对2004年1月至2007年12月、2008年1月至2008年12月和2009年1月至2011年7月等三个阶段内我国金属市场的量价相关性所展现的多重分形程度进行比较发现,2008年量价相关性的多重分形特征最为明显,这表明该阶段内金属期货市场量价相关关系波动更为剧烈。而这可能是由于受金融危机及全球商品市场影响,我国金属市场同样出现恐慌,市场噪声增加,量价相关性由于市场剧烈的波动行为变得更为复杂。相比较而言,2009年至今,我国金属期货市场量价相关性的多重分形程度逐渐减弱,这说明我国期货市场逐渐趋于相对有效的市场状态。但与LME金属期货市场相比,我国金属期货市场量价相关性的分形特征更为明显,这说明尽管我国金属期货市场逐渐趋于有效,但有效性依然要低于成熟的LME金属期货市场的有效性,期货市场依然需要进一步发展完善。(3)金属市场量价相关性同样存在着长期记忆和多重分形特征。量价关系的多重分形特征的存在,意味着价格与交易量之间存在着非线性依赖关系,以有效市场假说为前提对量价关系进行研究同样可能是不正确的。此外,在理解和分析市场行为时,应该同时考虑二者的影响,而不是单独量化。仅仅讨论这些变量中的一个,而忽略另一个变量,对市场的理解可能是存在偏差的。(4)将分形特征参数引入到对金属市场的风险测度是可行且有效的。本文将基于多重分形特征参数的波动率测度指标应用于VaR模型,并对不同分位数下的有色金属市场风险进行测度,同时与基于已实现波动率的VaR模型的风险测度结果进行比较。结果表明,多重分形的特征参数能够较好的识别金属收益序列中的行为信息和规律,在某些分位数下,基于多重分形特征参数的VaR模型具有优异的风险识别和测度能力,将多重分形特征参数引入到对金属市场风险管理中具有一定的可行性和有效性。更多还原

【Abstract】 Since2002, China has gone beyond the United States as the world’s largest producer and consumer of non-ferrous metal. With the rapid development of urbanization and industrialization, the pulling of domestic demand, China’s demand for nonferrous metal resources will continue to rise and its basic strategic resources will be in an increasingly important position. Whether non-ferrous metal industry can keep healthy development has become the important effect on our country’s economic growth and National Defense Security. However, in recent years, the price fluctuation of nonferrous metal, especially the bronze’s price, has been high and frequent, which formed a huge market risks, and also serious impacted on the China resources security. Therefore, in this context, analyzing of the fluctuating behavior of non-ferrous metal prices, measuring the fluctuations of the market risk has the important practical significance. However, the current research is basically the traditional effective market hypothesis as the foundation, and its strict assumptions cannot fully effectively explain many financial phenomenon. At the meantime, the fractal theory as frontier theory of a market complexity research provides a new theoretical framework for the study of the market price act.In this context, this paper introduces the theory of fractal market to non-ferrous metal prices fluctuating behavior analysis, identifies and characterizes the non-ferrous metal prices fluctuating behavior from a new angle of view, and then applies the fractal characteristics parameters in the measurement of the fluctuations risk contained in the market. It of course has great theoretical and practical significance. Through the research, this paper concluded that the following conclusions:(1) The metal market is a complex system containing obvious fractal characteristics. The return series of copper and aluminum in Shanghai were shown significant "peak, fat tail, partial" characteristics, which are not normal distribution, indicating that the traditional efficient market theory will not be able to better understand the deeper characterization of the metal market. The return series of metal fractal dimension is between1-2in different periods, which indicates the metal markets do not follow a random walk process but is a fractal structure of the market. It’s price behavior is a random walk process containing trend and noise. The reason of market fractal characteristics is the long-term memory characteristics and the "peak fat tail" distribution. Therefore, the introduction of fractal market theory to the analysis of the behavior of non-ferrous metal price volatility has an important theoretical and practical significance.(2) Metal futures market still needs further development and improvement. The multifractal extent of the volume and price of the metal market in China from January in2004to December in2007, from January to December in2008and from January in2009to July in2011and so on three stages showed that in2008, the volume and price of the multifractal characteristics is the most obvious, which indicates that the phase correlation between the metal futures market volume and price fluctuations is more intense. This may be due to the impact of financial crisis and the global commodity markets, china’s metal market is also panic in the market, an increase in noise, volume and price become more complex due to the sharp market fluctuations in behavior. In comparison, since2009, multifractal degree of market volume and price correlation gradually weakened in China’s metal futures, indicating that China’s futures market is becoming more and more relatively effective state.Comparing LME metal futures market, the fractal characteristics of volume and price correlation are more obvious, indicating that China’s metal futures market is becoming effective gradually, but the effectiveness is still lower than the mature LME metals futures market. The effectiveness of the futures market still need to be further developed.(3) the correlation of quantity and price in metal market also has the long-term memory and multifractal characteristics. The existence of multifractal characteristics of price-quantity relationship means that the price and trading quantity exist non-linear dependencies, the premise of the efficient market hypothesis to study the relationship between quantity and price may be incorrect. In addition, with the understanding and analysis of market behavior, we should also consider the impact of price and quantity, rather than separate quantization. Only discussing one of these variables, ignoring the other variables, the understanding of the market may be biased.(4) The parameters of the fractal characteristic introduced into risk measure of the metal market is feasible and effective. Volatility Measure indicators based on the multifractal characteristics parameters are used in the VaR model, and to measure risk under different quantile of non-ferrous metals market, compare the result of risk measure in VaR model based on realized volatility.The results show that the multifractal characteristics parameters are able to better identify the metal in the return series of behavioral information and laws.Under some quantile, since VaR model based on the multifractal characteristics of parameters has excellent risk identification and measurement, multiple fractal Parameter introduced into the metal market risk management has the feasibility and effectiveness..更多还原