【作者】 王博；

【导师】 郭万山；

【作者基本信息】 辽宁大学 ， 金融工程， 2014， 博士

【摘要】 资本资产定价模型（Capital Asset Pricing Model简称CAPM）是由美国学者Sharpe（1964）、Linter(1965)和Mossin(1965)等在马科维茨投资组合理论的基础上提出的投资组合定价方法，该模型主要研究投资组合预期收益率与投资组合系统性风险的关系。CAPM模型的核心是只对投资组合的系统性风险进行定价，非系统性风险可以通过扩大投资组合规模予以消除，因而市场不予风险补偿。CAPM模型对投资组合定价是在无风险收益基础上，对投资组合系统性风险予以风险补偿。CAPM模型存在的问题主要在实证方面，经验研究表明CAPM实证效果较差，定价效率存在问题，出现了“股权溢价之谜”等金融市场异象。所谓“股权溢价之谜”是指由Mehra和Prescott(1985)在1979年运用1889-1978年的历史数据，发现美国存在着较高的股市收益率与较低的国库券利率长期并存的现象，股权溢价等于股权收益减去债券收益，股权溢价始终维持在6%左右，溢价幅度过高。在此之后许多学者从各种不同的角度对“股权溢价之谜”进行解释，取得显著性成果的有Epstein和Zin(1989、1991)、Campbell和Cochrane(1999)。指出CAPM模型无法解释“股权溢价之谜”的原因源于CAPM的核心假设：市场有效性和投资者理性与现实不符。CAPM模型实证效果较差的一个重要原因在于马科维茨最优投资组合理论的假设条件在现实中几乎无法得到有效满足。马科维茨投资组合要求投资组合只包含系统性风险，非系统性风险可以通过扩大投资组合规模予以消除，但没有考虑投资分散化成本问题。而实践中投资者必须要考虑投资分散化成本。从理论上来说，投资组合只包含系统性风险，不包含非系统性风险的条件是投资组合中所包含的股票数量n无限大。对于一个确定的股市来说，该条件是无法得到满足的。既使从现实投资角度来说，投资者所持有的股票数量往往是几只股票，持有太多的股票首先是增加了股票的交易成本。同时，也无法对投资组合进行有效管理。对机构投资者来说，所持有的股票组合所包含股票数量会更多一些，但也无法保证投资组合只包含系统性风险，不包含非系统性风险的要求。为了改进CAPM定价效率，Fama和French(1993)将系统性风险由整个市场层面分解至行业层面，提出了三因素模型。三因素模型就考虑到了行业风险问题，将CAPM模型的系统性风险表述拓展到行业层面，部分地克服CAPM以市场组合代替系统性风险的局限性，但投资组合中非系统性风险残值仍然偏大，三因素模型并未完全解决包含非系统性风险的投资组合定价问题。行为金融学将金融心理学引入到股市价格收益及股价波动领域，有效地解释了股市的“羊群效应”和价格波动的集聚性现象。本文在Fama和French三因素定价模型的基础上，从行为金融学角度，将投资者情绪引入资本资产定价研究，对资本资产定价模型进行了改进，构建了基于投资者情绪的资产定价理论模型。并检验了投资者情绪对投资组合价值的影响。采用理论分析和实证分析相结合、定性和定量相结合的研究方法。主要以定量分析为主，利用统计学和计量模型，并采用主成分分析方法，构建情绪综合指数。从理论和实证两个方面探讨投资者情绪对资本资产定价模型的改进问题，分析和检验投资者情绪对资本资产定价效率影响的显著性。本文选题在该研究领域属于前沿热点问题。理论意义在于使用中国股票市场的数据，检验了传统金融学关于市场有效性的假设是否成立。应用价值则主要体现在投资者情绪对沪深300指数收益率具有显著的影响，揭示了投资者行为对中国股市具有不可忽视的影响。本文具有重要理论价值和现实意义。本文研究的主要内容包括：本文首先系统的对国内外基于投资者情绪的资产定价理论及实证研究文献进行综述，指出投资者情绪与资产定价关系的理论争论。其次，文章论述了资本资产定价的基础理论。第三，本文对投资者情绪给出了清晰的界定，说明投资者情绪的影响因素和度量方式。重点说明了综合投资者情绪指数的构造方法。第四，在Fama和French三因子模型基础上，构建基于投资者情绪的资产定价理论模型，重点论述了理论模型的假设条件和构造方法。第五，对所构建的情绪资产定价理论模型进行实证估计和结果分析，并与耶鲁CCER中国股市投资者信心指数的定价效果进行对比分析，证明本文所构建情绪综合指数的合理性。最后对全文进行总结和归纳，阐述全文结论和下一步的研究方向，即构建基于信息、动态的情绪资产定价模型、开展情绪综合指数研究和加强投资者情绪理论与非线性资产定价和市场微观结构理论相结合。论文的主要结论是:1.沪深300指数受投资者情绪因子的影响最大，上证A股指数受投资者情绪的影响显著。2.小盘指数受投资者情绪因子的影响更大。3.从投资者情绪对行业影响的角度来看，投资者情绪对房地产行业、有色金属行业、医药生物行业、钢铁行业的影响存在较大的差异性。4.投资者情绪对上证50ETF和上证180ETF收益率影响显著。论文的主要创新之处:1.将投资者情绪引入资本资产定价模型，在Fama和French三因子模型基础上，对资本资产定价模型进行了改进,构建了基于投资者情绪的资产定价理论模型。2.采用主成分分析方法构造了投资者情绪综合指数。3.通过沪深300指数等投资组合，对所构建的基于投资者情绪的资产定价理论模型进行实证检验，检验了投资者情绪对投资组合收益率的影响是否显著。4.对本文构建的情绪综合指数与耶鲁CCER中国股市投资者情绪信心指数对所构建的基于投资者情绪的资产定价理论模型的实证定价效果，进行比较分析。更多还原

【Abstract】 The Capital Asset Pricing Model (CAPM) is by the American scholar Sharpe (1964),Linter (1965) and Mossin (1965), On the basis of Markowitz portfolio theory proposedportfolio pricing method, The model has researched portfolio expected return and the riskof portfolio system in the relationship. At the heart of the CAPM model is only to pricesystem of portfolio risk. Unsystematic risk can be eliminated by enlarging the scale ofinvestment portfolio.Stock market will not be risk compensation.CAPM model forportfolio pricing is based on risk-free return, risk compensation system for portfolio risk.CAPM model mainly in empirical aspect, The problems has existed in the empiricalresearch for showing that CAPM empirical effect is poorer for pricing efficiency problem,the "equity premium puzzle" vision financial markets. So called "equity premium puzzle"has been pointed by Mehra and Prescott (1985) in1979.Using the historical data from1889to1978, the U.S. stock market yields and the risk-free rate has conductedcomprehensive comparison, It has found that the United States there is a Treasury ofhigher stock market returns and lower interest rates the phenomenon of long-termcoexistence, the equity premium is equal to the return on equity minus bond yields, theequity premium is always maintained at about6%, the premium is too high. After thatmany scholars from different Angle explained the “equity premium puzzle”, significantachievements have been found by Epstein and Zin (1989,1991), Campbell and theCochrane (1999). Pointing out that CAPM model cannot explain why the "equitypremium puzzle" is the result of CAPM core assumption: market efficiency and investorrationality is not reality.CAPM model demonstration effect is an important reason for the difference is thatthe optimal portfolio theory of markowitz assumptions in reality almost impossible toeffectively meet. Markowitz portfolio contains only systematic risk portfoliorequirements, unsystematic risk can be eliminated by enlarging the scale of investmentportfolio to; But not the diversified cost into consideration. Investors must consider costof investment diversification and reality.In theory, portfolio contains only systematic risk,excluding unsystematic risk is the condition of stock portfolio contains the number n is infinite. For a certain stock, the condition cannot be satisfied. Is made from realinvestment perspective, the number of shares held by the investor, is often a few stockshold too many stock first is to increase the stock of transaction costs. At the same time,also can’t effective portfolio management. For institutional investors, stock portfoliocontains stock quantity will be some more, but there is no guarantee that the portfoliocontains only systematic risk, excluding unsystematic risk.In order to improve the CAPM pricing efficiency, Fama and French (1993) by thewhole market system risk level decomposition to the industry level, three factors modelis put forward. Three factors model took into account the industry risk, system riskrepresentation of the CAPM model to the industry level, partly overcome the limitationsof CAPM replace the system with the market portfolio risk, system risk portfoliochina-africa salvage value still slants big, three factors model has not fully includeunsystematic risk portfolio pricing problem.Behavioral finance the financial psychology into the stock market price earnings andshare price volatility, effectively explaining the "flock effect" of the stock market pricefluctuation and the agglomeration phenomenon. In this paper，Based on Fama and Frenchthree factors pricing model. From the perspective of behavioral finance.Investorsentiment has been taken into the capital asset pricing research,This capital asset pricingmodel has been improved.Constructing Asset Pricing Theory model based on investorsentiment. And examing the influence of investor sentiment to portfolio value. Bycombining the theoretical analysis and empirical analysis, the combination of qualitativeand quantitative research methods. Mainly quantitative analysis, the use of statistical andeconometric model, and use the principal component analysis method, composite indexbuild emotions. Investor sentiment were discussed from two aspects of theory andpractice to the improvement of the capital asset pricing model, analysis and test ofinvestor sentiment effect on capital asset pricing efficiency significantly. Subject belongsto the frontier hot issues in the research field. Theoretical significance lies in the use ofthe Chinese stock market data, test the hypothesis that the traditional finance aboutmarket efficiency is established. Application value is mainly reflected in investorsentiment has significant effect on the CSI300index yield, reveals the investor behaviorhas influence to the Chinese stock market. This paper has important theoretical value and practical significance.This paper studies the main contents include: The first system at home and abroadbased on investor sentiment for asset pricing theory and empirical research, pointing outthe relationship between investor sentiment and asset pricing theory. Secondly, the articleexpounds the basis of the capital asset pricing theory. Third, this paper provides a clearlydefined as investor sentiment, to explain the influence factors of investor sentiment andmeasure. Expounding the method of comprehensive index of investor sentiment. Fourth,on the basis of both Fama and French three factor model, build a model of asset pricingtheory based on investor sentiment, expounding the theory model of hypothesisconditions and construction method. Fifth, To build the sentiment of the empirical assetpricing theory model and the result analysis, and Yale CCER pricing effect of Chinesestock market investor confidence index were analyzed,It has been proved in this paper,The rationality of the sentiment composite index has been constructed. Finally, Tosummarize the full text and the conclusion.In this paper, the full text conclusion andfurther research direction, which are built based on the information, the sentiment ofdynamic asset pricing model, the research of sentiment composite index and strengthenthe investor sentiment theory and nonlinear combination of asset pricing and marketmicrostructure theory.The main conclusions of this paper is1. The sentiment factor has been the biggestinfluence on the CSI300index, The Shanghai A-share index is also significantly affectedby investor sentiment.2. Small-cap index more influenced by sentiment factors.3. Fromthe point of view of investor sentiment influence on industry, investor sentiment on thereal estate industry, non-ferrous metal industry, biological medicine industry, iron andsteel industry has significant differences.4. In this paper, the research conclusion hassignificance difference with foreign research,This article has found that investorsentiment significantly affecting the Shanghai50ETF and the Shanghai180ETF.This paper innovation points are:1. Using principal component analysis get the finalstandardization composite sentiment index.2. On the basis of both Fama and Frenchthree factor model, build a model of asset pricing theory based on investor sentiment3.Innovative discussions, this paper will be based on investor sentiment for Fama andFrench three factor asset pricing model empirical research to market value index, index of different industries, different style and different ETF investment funds, Research oninvestor sentiment for the difference of these combinations.4. For this paper acomparative analysis has been builded on the sentiment composite index of Chinesestock market and Yale CCER confidence on sentiment pricing effect for asset pricingtheory model.更多还原