资产价格系统演变模型的理论与实证研究

【作者】 郑醒尘；

【导师】 蒋振声；

【作者基本信息】 浙江大学 ， 农业经济管理， 2006， 博士

【摘要】 资产价格分析一直是金融理论创新与实证研究的重要切入点。本文重点研究真实资本市场中资产价格的演变机制。本文认为真实资本市场中的交易者和资产价格相互作用,推动了资产价格的不断演变,并具有一定的趋势性。为此,本文建立相应的市场系统模型来分析交易者和资产价格之间的相互关系。 根据本文研究重点,本文以刻画、拟合真实市场中资产价格演变过程为切入点,提出了交易者决策模型——交易者盈利期望模型,并以此为组成单元,建立市场系统模型——资产价格系统演变模型。围绕这一模型,本文进行相关的理论分析和实证研究,并就模型在预测领域的应用进行探索。 根据研究角度的不同,目前国内外关于资产价格演变机制的分析可分为微观和宏观两大类。在微观层面,主要着眼于对交易者行为特征的刻画。传统的理性人假设已经被行为金融学家证明离现实有很大距离,为此,行为金融理论对传统的理论模型进行了发展,Kahneman和Tversky提出了展望理论(Prospect theory)等。这类理论依赖于对市场参与者心理的先验性假设,为人们认识理解资产市场提供了新的切入点,但对真实市场进行定量分析的难度较大。在宏观层面,人们就市场群体行为建立了相应的仿真模型进行模拟,如市场行为模型(BSV,DHS)、统一理论模型,以及人工多智能体模型,在国内人们也进行了大量的仿真研究。这些研究通过模拟产生类似真实市场中的自组织进化、羊群效应等现象,刻画出了真实市场中的价格演变机制,但这几类模型并不能直接用来预测真实市场中的价格演变趋势,对真实资产价格演变过程的拟合能力有限,这是其不足之处。 在价格演变规律分析的具体应用领域,价格预测一直是热点问题。除了传统的线性预测和非线性预测方法,人们已经广泛地把其它领域的数学工具,如遗传算法、人工神经网络、模糊理论等人工智能计算方法和混沌原理、小波分析理论等相关领域的新成果迅速应用到价格预测中。但是,人们往往从实用角度偏重于预测模型的精度,忽略了模型对微观经济运行机制的解释能力,预测模型的工具化倾向十分明显。如现有的人工神经网络模型具有较强的拟合能力,但模型本身像一个黑箱,人们对据此得到的结果不知其所以然,难以从理论上进一步加深对资产价格运动机制的理解。换而言之,预测技术的进步并没有为金融理论研究提供实证支持。 对于上述情况,本文在相关的文献回顾中进行了介绍评价。这些是本文第二章的主要内容。本章先就资产价格运动的理论假设进行分析,说明本文采用分形市场假设的缘由和依据;随后,分别对有关交易者投资决策行为和市场系统模型的理论文献进行回顾分析,并对现有价格预测方法进行总结分析,指出优点与不足;在本章的最后部分,以相关研究文献为依据,以混沌理论为基础,对资产价更多还原

【Abstract】 Asset price analysis is always the important issue of financial theory innovation and positive research. The study of this paper is mainly around the mechanism of the price evolvement in financial market. We consider that the reaction between traders and asset prices promoting the price evolvement in real financial market, and this evolvement have its tendency. So it is needed to build a market system model to describe the relationship between traders and price of assets.Due to the key issue of this paper, we develop Trader Profit Expect Model and Asset Price System Evolvement Model to describe and simulate the price evolvement process. Related theory analysis and positive study will be used to sustain the idea, and the model will be used to develop new method of price forecasting. There two ways to describe the process of price evolvement in model foundation. In microcosmic way, peoples usually deal with describing the behavior of trader. For example, it is proved that the rationalistic hypothesis is not so tally with the real world by the scholar of behavior finance. There for, classic models have been developed. Kahneman and Tversky developed Prospect Theory in 1979. By now, because those new theories are under the mentality analysis of traders, so it is very difficult to testify those theories by demonstration researches. In macroscopic way, peoples develop artificial simulation models to describe the colony behavior of traders in financial market. Such as BSV, DHS, Unified Theory Model, and Multi-agent Model. By those models, people can simulate the herd behavior of trader in financial market, and also can describe the mechanism of the price evolvement. But those models can’t be used in forecasting.In the field of price analysis, forecasting is always the important issue. Peoples have adopted varies kind of math tools to heighten the accuracy of price forecasting model, such as Genetic Algorithms and Artificial Neural Network model and Fuzzy Theory are used in prediction. The wavelet theory and chaos theory are also used in prediction. It should be emphasized that peoples always pay their attention to the accuracy of prediction and don’t care about the capability in explaining the micro economy process of forecasting model. For example, when the ANN is used in prediction, people can only know the result of forecasting, but can’t get information about the process of pricing. So such a forecasting model will not help people to develop financial theory. It means that financial theory can’t get help from mathematic tools in prediction.Those facts are introduced in the part of literature review. It will be the main content of chapter 2. By the way, the reason why the Fractal Market Hypothesis (FMH) is adopted but not Efficient Market Hypothesis (EMH) firstly. Then, the reports about the decision behavior of traders and the models of market system will be introduced and estimated. At the last of this chapter, it will be concluded that the asset price can be predicted in short period due to the chaostheory, and mathematic theory about chaos will be introduced.The main mission of chapter 3 is to study the mechanism of asset price evolvement, develop model to describe the process of asset price evolvement, and to provide new method of price forecast. The trader Profit Prospect Model will be developed to analysis the relationship between traders and price of assets, and Asset Price System Evolvement Model will be developed to describe the process of prices evolvement. Related theory deduction and demonstrations about those two models indicate that those models can tally with the real price evolvement process well, and can get information about the traders and the relationship between traders and price of assets, can also be used to predict the price.The main mission of chapter 4 is to sustain the viewpoint of this paper about price evolvement and its models developed in this paper by positive researching which is mainly about price prediction. By the way, a new method of price prediction is also developed. The key idea of the forecast method is that we should make the most use of information in financial market, and make the forecast model to learn the principle of price evolvement. Then, we can make the model to forecast the future price of asset. Due to the new forecast method, we can forecast a set of prices at same time, and we can also forecast future price of an asset by information of other assets. The can also be used in orbit forecast according with chaos theory. In the part of demonstration research, random sample of 9 stocks are all proved in chaos, and we also prove that except the manipulation stock in slumping, price of most stock in financial market can be forecasted efficiently.In chapter 5, we summarize the paper. Conclusion will talk about issues as follows. Due to the reaction between traders and asset price, this paper develops Trader Profit Prospect Model and Asset Price System Evolvement Model under theories of fractal market hypothesis, behavioral finance and chaos. Duo to those two models, new method of price prediction is developed which can predict several asset price in future. Relative positive examples sustain the issue of this paper about the relationship between traders and asset price. Innovations of this paper are as follows.Firstly, Trader Profit Prospect Model is developed. It is assumption that there is a special relationship about the change of stock price between real world and the prospect of trader, which can be expressed by the S function of logistic curve. The difference between this model and Prospect theory is that this model can be used to determine the characteristic of a kind of trader.Secondly" Asset Price System Evolvement Model is developed in this paper. This model can approach the process of price evolvement. The difference of this model from BSV and DHS is that this model contain subdivision unit of traders, and this model can be used in price prediction. The difference between this model and ANN isthat this model can approach the real process of price evolvement and determine the market construction and the mechanism of price evolvement.Thirdly, new method of price prediction is developed. The larruping characteristic of this model is that it can prediction several asset price in future. During the working of demonstration, it is found that most of price series of 9 stocks have different length of non-periodic cycle in different period of time in according with chaos theory. It means that the principle of price series is in changing along with the evolvement of market system. At the same time, we find that composing of traders with different profit prospect and sensitivity to price changing can reflect the tendency of price evolvement, and can also influence the future price evolvement. In general, on the basis of relative literature, we develop Profit Prospect Model which can be used to determine the composing of trader with different profit prospect, and develop Asset Price System Evolvement Model which can be used to describe the mechanism of the price evolvement in financial market, and develop new forecast method on the basis of those models in according with chaos theory. The relative theory analysis and demonstration research prove that those new models and method are feasible and valuable in real market.更多还原