摘要：股票市场作为直接融资的场所，提高了社会金融资源配置的效率，对各国经济的发展起到重要的作用。但股票市场经常发生脱离经济基本面的过度波动甚至是异常波动，影响投资者参与股票市场的积极性，不利于股票市场支持实体经济功能的有效发挥。中国股票市场建立仅28年时间，己发生多次脱离经济基本面的过度波动。对股票市场过度波动现象的解释和波动成因的研究成为监管层、学术界和投资者关注的焦点和热点。 基于理性预期的传统资本市场理论认为理性套利者会消除股价的过度波动，使得资产价格处于基本面价值决定的均衡价格。其假设条件过于抽象和理想化，无法对股票价格长期脱离基本面的过度波动做出合理解释。学者们放松传统资本市场理论的假设，发展出新兴的资本市场理论，认为交易者不是完全理性和同质的，而是有限理性和异质的，不可能及时、准确的获得资产基本面的所有公开信息。即使面对同样的信息，交易者也会受到行为偏差的影响，对信息进行不一致的处理和解读。资产价格的动态波动是有限理性异质性交易者行为相互作用的直接结果。新兴的资本市场理论能够对资本市场的异常现象进行更加合理的解释，其前提假设也更加符合资本市场的客观实际，异质性交易者视角的资本市场价格理论受到广泛关注。 本文从中国股票市场经常发生的周期性过度波动特征事实出发，基于对资本市场价格研究国际前沿文献的追踪和把握，从异质性交易者行为分析视角出发，认为中国股票市场价格走势是市场上异质性交易者买卖交易行为的直接结果，制度等外部因素是通过影响异质性交易者的价格预期和交易行为，间接作用于股票价格的。因而本文首先分析中国股票市场投资者的结构演变、当前构成及行为特征，总结投资者行为特征背后的价格预期方法，按预期方法将投资者区分为基本面分析交易者、图表(技术)分析交易者、噪音交易者和内幕信息交易者。在当前文献区分异质性交易者类型及模型构建思路的基础上，结合中国股票市场的实际，详细剖析异质性交易者买卖交易行为的形成过程，研究其交易行为对中国股票市场价格走势的影响。 本文认为，不同类型的交易者有不同的价格预期与交易行为，影响股票价格的波动。中国股票市场基本面分析交易者是有限理性的，根据基本面预期规则计算股票的基本面价值，预期未来价格会朝向基本面价值决定的价格均值回归。图表(技术)分析交易者使用不同的图表(技术)规则，会产生不一致的价格预期，导致在每一个价格点位上存在向上和向下两种价格预期，同时存在两类的图表(技术)分析交易者：趋势跟随交易者和反向趋势交易者。趋势跟随交易者认为当前的价格趋势会继续，反向趋势交易者认为当前的价格趋势会反转。噪音交易者不具有专业的预期技能，根据证券经纪人、新闻媒体和身边投资者等消息面对股价进行幼稚预期，交易具有随意性和盲目性，在市场上追涨杀跌制造噪音，资金量小，对市场价格的走势不会产生决定性的影响。内幕信息交易者是市场上的信息优势方，拥有尚未公开的私人信息，提前建仓股票或人为制造虚假信息获利。随着股票市场的发展和监管的完善，内幕信息交易者逐渐减少，仅分布于极少数股本小的个股，对股票市场价格走势的影响微乎其微。 从整个股票市场来看，价格走势主要由基本面分析交易者和图表(技术)分析交易者决定，噪音交易者仅会放大由基本面分析交易者和图表(技术)分析交易者决定的价格走势的幅度。本文根据上述分析，构建中国股票市场异质性交易者价格模型，解释中国股票价格过度波动及其他异常现象，提出保障中国股票市场稳健运行的政策建议。 基于上述主要研究内容，本文按以下结构逐次展开： (1)根据国内外对金融市场研究的前沿理论和最新成果，结合中国股票市场价格运行波动剧烈、暴涨暴跌频发，信息不对称程度高等特点，选择从有限理性异质性交易者视角来研究中国股票市场价格的波动。 (2)梳理中国股票市场波动的特征事实，对中国股票市场进行非线性分形市场检验，为从异质性交易者视角研究中国股票价格波动的合理性提供实证基础。在此基础上梳理中国股票市场投资者的历史演变，分析机构投资者和个人投资者的交易行为，总结投资者主要使用的价格预期方法，根据预期方法将中国股票市场的投资者分为基本面分析交易者、图表(技术)分析交易者、噪音交易者和内幕信息交易者。 (3)对异质性交易者与股价波动的关系进行理论分析，提出本文模型构建的前提假设。分别建立基本面分析交易者、图表(技术)分析交易者、噪音交易者和内幕信息交易者基于价格预期的交易模型，根据对四类交易者预期策略和模型的分析，建立中国股票市场异质性交易者价格模型。 (4)对中国股票市场价格模型进行稳定性分析和数值模拟，研究模型是否能够产生和解释中国股票市场常见的过度波动和收益率序列相关现象，确定模型中交易者对解释股票价格走势的作用，并使用异质性交易者价格模型定性解释中国股票市场的其他异常现象。 (5)在股票的成交数据中，难以区分四类交易者的数据，对本文理论模型的直接实证检验存在困难。为增加本文理论模型结论的可靠性，本文经过分析，采用机构投资者作为基本面分析交易者的代理变量，实证检验基本面分析交易者、图表(技术)分析交易者对于股价波动的作用。使用中国股票市场的换手率数据以及股权分置改革之后不同市场的动态特征数据进行对比，统计分析中国股票市场异质性交易者预期规则转换和行为的变化，支撑本文关于异质性交易者结构转变的分析。 (6)统计分析中国股票市场主板、中小板、创业板的波动差异，在本文理论模型的前提假设下，应用模型结论对三个板块波动的差异性进行解释，预测三个板块异质性交易者的结构。由于缺乏相应异质性交易者数据，采用调查问卷这一方式调研中国股票市场主板、中小板和创业板上的异质性交易者结构，检验其是否与本文模型预测的结果一致。 (7)总结理论模型的政策含义及相应的配套建议，并对未来进一步的研究进行展望。通过对本文理论模型的模拟，从异质性交易者视角定量解释了股票价格的波动和收益的序列相关性。得出以下结论： 一是当市场上只有图表(技术)分析交易者时，不管图表(技术)分析交易者中趋势跟随交易者和反向趋势交易者的比例如何，股票市场价格系统整体运行都是不稳定的和无规则的，当基本面分析交易者出现以后，价格系统整体运行是相对稳定的，表明基本面分析交易者具有纠正由图表(技术)分析交易者导致的价格系统无规则运行的作用。 二是在基本面分析交易者可以使价格系统相对稳定运行的基础上，进一步分析表明，若基本面分析交易者具有短期的记忆，则市场价格仍会表现出相对过度波动的走势；若基本面分析交易者具有长期记忆，市场价格过度波动的现象消失，表明同短期的基本面分析交易者相比，长期记忆的基本面分析交易者具有更好的稳定市场价格运行的作用。 三是在基本面分析交易者、图表(技术)分析交易者共存的市场中，若更多的交易者转变预期规则，成为趋势跟随交易者，使趋势跟随交易者在市场中占比较大时，市场倾向于过度波动；若更多的交易者转变为反向趋势交易者，使反向趋势交易者在市场中占比较大时，市场运行较为稳健，股价波动较小，说明反向趋势交易者具有稳定市场运行的作用。 四是基本面分析交易者的学习机制(即记忆)对解释收益率的连续和反转具有重要的作用，当基本面分析交易者和图表分析交易者在市场共存时，基本面分析交易者的长期记忆会在短期会导致价格趋势持续(收益正的序列相关性)，在长期会增加价格运行的振幅(收益率存在负的序列相关性)。由此说明，具有长期记忆的基本面交易者在短期放大了由趋势跟随交易者引致的价格趋势，而在长期又放大了由反向趋势交易者所触发的价格反转行为，这在一定程度上验证出具有长期记忆的基本面分析交易者具有稳定市场价格运行的作用。 在使用模型模拟结论对中国股票市场价格波动进行解释的基础上，分析中国股票市场存在的卖空限制、涨跌停板限制和T+1交易制度对异质性交易者行为的影响，发现这些制度会促使更多的交易者转化为趋势跟随交易者，促使中国股票市场上图表(技术)分析交易者中趋势跟随交易者的力量进一步上升，再加上噪音交易者的作用，这会进一步放大中国股票市场价格的波动。同时，使用本文的理论模型，从异质性交易者相互作用的视角可以解释，中国股票市场存在的一般基本面信息的变化引起股票价格的剧烈波动，重大的经济基本面信息发生变化而股票市场反应平淡，无基本面变化但价格大幅波动，股价有时产生对基本面价值的长期、大幅度偏离，股票价格对中美经贸摩擦等某一重大基本面事件有着超乎寻常的剧烈反应等异常现象。 与现有的构建异质性交易者模型研究股票市场价格波动的文献相比，本文可能的创新点如下： (1)本文构建的中国股票市场异质性交易者价格模型不仅能够解释中国股票市场价格的波动，还能定性解释中国股票市场存在的一般基本面信息的变化引起股票价格的剧烈波动，重大的经济基本面信息发生变化而股票市场反应平淡，无基本面变化但价格大幅波动，股价有时对基本面价值的长期、大幅度偏离，股票价格对中美经贸摩擦等某一重大基本面事件超乎寻常的剧烈反应等异常现象。 (2)本文从投资者行为出发，较为全面地总结了中国股票市场投资者主要的预期方法，根据预期方法将中国股票市场的投资者划分为基本面分析交易者、图表(技术)分析交易者、噪音交易者和内幕信息交易者四种类型。基于这种划分，可以更为全面和系统地考察交易者行为，更贴近中国实际。 (3)在确定基本面分析交易者的基本面价值时，引入一个基本面分析交易者的学习机制(记忆)来刻画中国股票市场上基本面分析交易者的有限理性，更好地刻画了基本面分析交易者的有限理性行为。对图表(技术)分析交易规则的详细分析表明，使用的图表(技术)分析规则不同，交易期限长短不同，对同一点位价格走势预期不同。因而同一价格点位，同时存在两种不同的图表(技术)分析价格预期，将中国股票市场上的图表(技术)分析交易者分为趋势跟随交易者和反向趋势交易者两类，二者同时并存。与现有文献大多仅考虑趋势跟随交易者相比，这种划分更好的刻画了中国股票交易者多元存在及其交易行为的事实。 关键词：股票价格 异质性交易者 预期方法

 Abstract: As a place of direct financing, stock market has improved the efficiency of the allocation of social financial resources and has a significant role in promoting economic development of various countries. However, stock market often has excessive volatility from the economic fundamentals, which affects the enthusiasm of investors to participate in the stock market, which is not conducive to effective functioning of the stock market to support the real economy. The Chinese stock market has been established for only 28 years, and there have been many excessive fluctuations from the economic fundamentals. The researches on interpretations and causes of excessive volatility have become the focus and hotspot of the regulatory, academic and investor. The traditional capital market theory based on rational expectations holds that rational arbitrageurs will eliminate excessive volatility of stock prices, making asset prices long-term equilibrium prices determined by fundamental values. The assumptions are too abstract and idealized to explain the excessive volatility of stock prices over the long-term fundamentals. Scholars relax the hypothesis of traditional capital market theory and develop the theory of emerging capital markets. They believe that traders are not rational expectations and homogeneous, are bounded rational and heterogeneous. It is impossible to obtain all the disclosures of asset fundamentals information in a timely and accurate manner. Even if faced with the same information, traders will have inconsistent processing and interpretation because of behavioral bias. The dynamics of asset prices are a direct result of the interaction of bounded rational heterogeneous traders. Emerging capital market theory can explain the abnormal phenomenon of capital market more reasonably. The premise hypothesis is more in line with the objective reality of capital market. The theory of capital market price from the perspective of heterogeneous traders has been widely concerned and flourished. Based on the characteristics of the cyclical excessive volatility that often occurs in China's stock market, this paper studies the current theoretical literature on the frontier capital market price at home and abroad, choosing to follow the research framework of heterogeneous traders. Stock price is the direct result of trading behavior of traders, external factors such as institutions are indirectly affecting the stock price by affecting the price expectations and trading behavior of traders. Therefore, this paper analyzes structural evolution, current composition and behavior characteristics of Chinese stock market investors, summarizes price expectation methods behind investor behavior characteristics, and classifies investors into fundamentalists, chartists, noise traders and insiders. Based on the current literature to distinguish the type of heterogeneous traders and the construction of the model, combing with the actual situation of the Chinese stock market, the paper analyzes formation process of the heterogeneous trader's trading behavior, and its influence on Chinese stock market. It is considered that fundamentalists in the Chinese stock market is bounded rational. The fundamental values are calculated according to fundamental expectation rules. Fundamentalists expect that future price will return to the price determined by fundamental value, using different chart (technical) rules, chartists can produce inconsistent price expectations, resulting in up and down price expectations at each price point, while there are two types of chartists: trend followers and reversers. Trend followers expect that current price trend will continue, and reversers believes that current price trend will reverse. Noise traders do not have professional expected skills. They face naive expectations based on stock brokers, news media and investors around them. Their transactions are random and blind, and they are chasing up and down in the market to create noise. Their positions are so small that their transaction behaviors have little influences on the trend of market prices. Insiders are the information superiors in the market, have private information, build positions in advance or make profits by making false information. With the development of the stock market and the improvement of supervision, insiders are gradually declined and only distributed in a very small number of stocks. Their impacts on the stock market price trend is minimal. Therefore, from the perspective of entire stock market, price trend is mainly determined by the fundamentalists and chartists. The noise traders only amplify price determined by fundamentalists and chartists. Based on the analysis above, we construct a heterogeneous agent model in China's stock market to explain excessive fluctuations of China's stock prices and other anomalies, proposing policy recommendations to ensure the stable operation of China's stock market. this paper is developed according to the following structure: (1)According to the latest literatures in financial market research, combined with the characteristics of China's stock market price volatility, frequent surges and high information asymmetry, this paper chooses to study Chinese stock market from the perspective of bounded rational heterogeneous traders. (2)Based on characteristics of China's stock market volatility, non-linear fractal market test of China's stock market provides an empirical basis for studying the rationality of China's stock price volatility from the perspective of heterogeneous traders. On this basis, it sorts out historical evolution of investors in China's stock market, analyzes trading behavior of institutional investors and individual investors, summarizes price expectation methods mainly used by investors, and divides investors into fundamentalists, chartistsnoise, noise traders and insiders. (3)Based ont theoretical analysis of the relationship between heterogeneous traders and stock price fluctuations, putting forward premise hypothesis of the model construction. According to premise hypothesis, this paper establishes trading model of fundamentalists, chartists, noise traders and insiders to form a heterogeneous agent model in China's stock market. (4)Carrying out stability analysis and numerical simulation on the Chinese stock market price model, whether the model can generate and explain the phenomenon of excessive volatility and yield series in China's stock market, determining the role of traders in explaining the trend of stock prices. Using the heterogeneous agent price model explain other anomalies in Chinese stock market. (5)In the transaction data of stocks, it is difficult to distinguish data of the four types of traders. Thus, it's difficult to test the theoretical model directly. In order to increase reliability of conclusions of the theoretical model, this paper analyzes and uses institutional investors as the proxy variables of the fundamentalists. Using turnover data of China's stock market before and after the split share reform supports this model on heterogeneous agent analysis of structural transformation. (6)Applying model conclusions explain different fluctuations among Main Enterprises Board, Small and Median Board and Gowth Enterprises Board, and predict heterogeneous traders of the three sectors. Due to lack of corresponding heterogeneous trader data, the questionnaires are used to investigate the heterogeneous trader structure on Main board, SME board and GEM of China's stock market in order to test whether the traders' structure is consistent with the structure predicted by the model. (7)Summarize policy implications of the theoretical model and look forward to further research in the future. Main conclusions of the heterogeneous agent model are as follows: First, when there are only chartists in the market, regardless of the ratio of trend followers and reversers, the overall operation of stock market price system is unstable, when fundamentalists appear, the overall operation of the price system is relatively stable, indicating that fundamentalists have the effect of correcting the irregular operation of the price system caused by chartist. Second, based on the fundamental analysis, the trader can make the price system relatively stable. If the fundamentalists have short-term memory, the market price will still show a relatively excessive fluctuation trend; if fundamentalists have long-term memory, the phenomenon of excessive volatility disappears, indicating that compared with short-term fundamental analysis traders, the fundamentalists with long-term memory traders have a better function of stabilizing market prices. Third, in the market where the fundamentalists and the chartists coexist, if more traders change expected rule to become trend followers, the market is prone to excessive volatility; if more traders turn into reversers, stock price volatility is smaller, indicating that trend followers can lead to stock price volatility, reversers can stabilize market operations. Fourth, fundamentalists' learning mechanism plays an important role in explaining continuation and reversal of the rate of return. When fundamentalists and chartists coexist in the market, fundamentalists' long-term memory will lead to a sustained price trend in the short term and will increase the amplitude of the price run in the long run. This means that fundamentalists with long-term memory weill amplify price trend caused by trend followers in the short run and amplify price reversal behavior triggered by reversers in the long run. Considering the short selling restrictions, the price limit and the T+1 trading system in China's stock market, the analysis found that these systems will encourage more traders to convert into trend followers, which will further amplify the fluctuations in China's stock market prices. Compared with the existing literatures on heterogeneous agent models studying stock market price fluctuations, possible innovations of this paper are as follows: (1)The heterogeneous trader price model of China's stock market constructed in this paper can not only explain the fluctuation of China's stock market price, but also qualitatively explain the changes in the general fundamental information of China's stock market, causing the volatility of stock prices. The information on economic fundamentals has changed and the stock market has reacted dullly. The price has not fluctuated greatly when there is no fundamental change. The stock price has long-term and large deviation from the fundamental value. The stock price has exceeded a certain fundamental event such as Sino-US trade friction. An unusual phenomenon such as violent reaction. (2)Based on the behavior of investors, this paper summarizes the expected methods mainly used by investors in China's stock market, and divides investors in China's stock market into fundamental analysis traders and charts (technical) analysis according to the expected method. There are four types of traders, noise traders and inside information traders. This division is more comprehensive and systematic, and closer to the reality of China. (3)Introduce a fundamental analysis of the fundamentals of the trader's learning mechanism (memory) to characterize the fundamental rationality of the trader's bounded rationality in the Chinese stock market, and better characterize the basics. Analyze the limited rational behavior of traders. Through detailed analysis of the technical analysis of trading rules, it is considered that the chart (technical) analysis rules used are different, the trading period is different, and the price trend of the same point is expected to be different. Therefore, there are two different technical analysis price expectations at the same price point. The technical analysis traders in the Chinese stock market are divided into two categories: trend followers and reverse trend traders. They think that they coexist at the same time. Compared with the existing literature, most of them only consider the trend to follow the traders. Good portrays the reality. Compared with the existing research, the possible innovations of this paper are as follows: (1)The heterogeneous agent model of China's stock market can not only explain fluctuation of China's stock market price, but also explain a string of abnormal phenomenon qualitatively. For example, changes in general fundamental information cause excess volatility of stock prices, economic fundamental information has changed, but stock market has reacted dullly, stock price changes rapidly when there is no fundamental change, stock price has long-term and large deviation from fundamental value, stock price changes rapidly as a reaction on certain fundamental event such as China-US trade friction. (2)According to structural evolution and behavior characteristics of heterogeneous traders in China's stock market, this paper summarizes current methods of price expectations of investors. There are four categories: the first is fundamental analysis; the second is chart analysis; the third is based on news; the fourth is based on inside information. According to four expected methods, this paper divides investors in Chinese stock market into four types: fundamentalists, chartists, noise traders and insiders. (3)Introducing a learning mechanism to characterize fundamentalists' bounded rationality in Chinese stock market; A detailed analysis of the rules, considering the same price point, there are two different technical analysis expectations. Based on this, chartis are divided into two categories: trend followers and reversers , which portrays reality more better. Keywords: Stock price Heterogeneity traders Expectation methods