【作者】 王郧；

【导师】 张宗成；

【作者基本信息】 华中科技大学 ， 数量经济学， 2010， 博士

【摘要】 市场波动性与投资者交易行为关系一直是学术界所关注的重要问题,然而当前对其的研究还较为缺乏,本文则是针对这一问题所做的前瞻性研究。本文研究的目的在于探讨投资者行为与期货市场市场波动的传导机制,学术价值在于丰富期货市场的研究内涵和进一步实证验证行为金融理论,而应用价值则在于促进期货市场投资者结构的改善和监管方政策的制订,为期货合约的定价和新品种上市提供建议。研究的特色在于将金融学与数学、软件编程技术三者融会贯通。针对中国期货市场黄金、黄豆、强麦、沪深300股指期货这四种代表性期货合约的波动性研究发现：四种期货合约的收益序列均平稳；部分合约存在自相关性,商品期货的残差具备条件异方差性,收益的波动存在非对称性,负收益减小波动而正收益增大波动,黄豆、黄金、沪深300的正负收益波动的结果是增大了波动率,而强麦的正负收益波动的结果是减小了波动率；过去的交易、未平仓合约、大单交易三个因素对合约的波动均有不同影响。在对波动特性进行分析之后,本文将两期世代交叠模型引入到期货市场,建立了基于期货市场的投资者交易行为与市场波动的数理模型,并将其扩展到投资者完全信息和缺乏信息两种情况,然后通过均衡分析对期货市场的波动进行了考察,求解出了一阶条件和二阶条件。数理模型证明发现：债券的无风险利率,投资者的风险厌恶系数,合约供给的状态以及风险溢价的波动都会对期货合约价格的波动产生影响,文中的5个命题总结了全部的情况。建立在基于期货市场的数理模型基础之上,本文进一步推导出了7个假说,实证研究了期货市场上的动量与反转效应和日历效应这两种市场异象。针对动量和反转效应的实证研究发现：中国期货市场上确实存在此类异常现象,富有信息的投资者(如机构投资者),通常采用反转交易模式；而缺乏信息的投资者(如散户),则通常采用动量交易模式。针对日历效应的实证研究发现：中国期货市场上同时存在日内、周内、月份这三种日历效应,大豆、沪铝、硬麦期货的日内趋势证实了期货市场存在日内效应；大豆、沪铝、天胶三种期货合约的周内趋势证实了期货市场上存在周内效应；沪铝、天胶、硬麦三种期货合约的月份趋势证实了期货市场上存在月份效应。而造成市场异常现象的原因是什么,本文尝试从投资者的认知和行为偏差方面给出解释。而导致和驱动此类偏差的主要因素是投资者的过度自信心理和羊群行为心理,其中前者属于认知偏差,而后者属于行为偏差。针对过度自信的实证研究表明：与股票市场的研究类似,中国期货市场确实存在过度自信现象,投资者对私人信息反应过度,而对公开信息反应不足,市场短期的波动主要不是来源于公开信息,而是来源于私人信息。针对羊群行为的实证研究表明：中国期货市场上存在一定程度的羊群行为,并且在期货价格下跌时表现尤为明显,但并没有证据显示有大规模羊群行为存在,因而中国期货市场运行较为有效。在上述市场波动-市场异象-投资者认知与行为偏差的传导机制的框架之下,本文得出了主要结论,指出稳定市场的根源在于纠正投资者认知和行为偏差,并从进一步完善和优化期货市场投资者结构,加强投资者教育力度,对上市期货新品种的规划与建议这三个方面给出了政策建议,最后对本文研究的不足,行为金融研究和期货市场研究的前沿问题,以后需要进一步研究的方向做了展望。更多还原

【Abstract】 The relation between investor trading behavior and market volatility is always the main issues in capital market research, but it is insufficiency in Chinese future market. So this dissertation deals with it. The purpose of the research is discussing the transmission mechanism between the investor trading behavior and market volatility, and the academic value is enrich the connotation of the future market research and test the behavioral finance theory, and the application value is improving the investor structure and preparing the supervisory policy, which give the advice for pricing and new products launches. The Characteristic of this dissertation is making the finance theory, mathematics, and software programming techniques to a whole research.We use moonbeams, wheat, aurum and hs300 index futures and find four contracts are stationarity, and partial contracts are auto-correlative. The commodities futures contracts are all heteroskedasticity; the volatility of revenue are asymmetric negative revenue amplify the volatility and positive revenue reduce the volatility. The lag volume, open interest, and large volume have different effect on volatility.After the study of volatility, we introduce the OLG model into future markets. We set the investor behavior model based on future contract price, which can also be extended to complete and incomplete information. We provide the equilibrium solution and give the first-order or second-order condition. The mathematical model present follow findings:the Bonds risk-free interest rate, the Investor’s risk aversion coefficient, the supply of contracts conditions and the risk premium volatility all give impact on the volatility of future contract prices. The 2-period OLG model based on future market is consistent with the practical situation; the five propositions in the article summarize the whole situation.Based on the mathematical model in the future markets, this dissertation further induce seven hypotheses and give an empirical research on the two market abnormality which included momentum or reversal effect and calendar effect. First we find that Chinese future market exists momentum and reversal effect, the sufficient information investors such as institution adopt reversal trading patterns generally and the insufficient information investors such as private adopt momentum trading patterns in general. Second we find that intraday-effect, weekly-effect and month-effect are all exists in Chinese future market. Soybean, aluminum and wheat futures ’intraday-trend confirms the intraday-effect; soybean, aluminum and nature rubber futures weekly-trend confirm the weekly-trend; aluminum, nature rubber and wheat future’s month-trend confirm the month-effect.This dissertation attempts to give reason from the deviation of investors’ cognition and behavior, and the main explanation of the deviation is investor’s overconfidence and herding behavior. The research on overconfidence finds that similar with the case in stock markets, investors underreact to public information and overreact to private information in Chinese futures market. Private information shocks may bring great volatility in the short-run, while private information shocks are weak and cannot persist, i.e., investors have overconfidence features indeed. The research on herding behavior shows certain extent herding behavior and it is obvious in decline future market. But there is little evidence of systemic herding and the market is relatively efficient.In framework under market volatility-market abnormality-investors’cognition and behavior, the dissertation gives the main conclusion. We point out that stable market come from the reduction of investors cognition and behaviordeviation.we give the Policy recommendations from the Consummate and optimization of investor structure, investor’s education efforts, the planning of new future contracts. At last, we look ahead the lack of research, the front of behavior finance and future markets, the direction of further study.更多还原