【作者】 解洪涛；

【导师】 林少宫； 周少甫；

【作者基本信息】 华中科技大学 ， 数量经济学， 2009， 博士

【摘要】 本文在全面梳理不完全契约理论、行为金融理论和市场微观结构理论对基金投资行为解释基础上，对我国证券投资基金的资产集中策略、选股策略(投资风格)、活跃交易投资策略轮动进行了实证分析，着重分析了2005年10月至今的一轮市场周期中我国开放式股票基金的证券投资基金是否存在反馈交易行为，采用门限思想建立了非线性的三因子模型，测定在不同市场周期中风险偏好改变，验证了投资基金对市场稳定性的影响；结合我国投资基金最新的产品创新方向，提出了对基金流动性风险监管的对策体系。主要研究内容和结论如下：(1)通过对证券投资基金(机构投资者)资产配置集中策略的研究发现，当前开放式基金资产配置中较高的个股集中度对收益贡献为正，而产业上的集中对收益贡献为负。基金持股的集中度也具有周期性，在牛市中持股较为集中。基金资产集中投资与持股趋同符合基金经理的短期利益目标，而一旦市场转向则会造成羊群效应下的市场巨幅波动。(2)采用基金资产持有收益与公告收益，构造了收益差值指标RG，测度了我国开放式证券投资基金在公告期间是否存在频繁的交易行为，及这些交易特征在不同市场周期中的差异。结果发现我国基金存在普遍的季度内频繁交易行为，即大多数基金都具有积极的交易策略，这种交易活跃程度在市场转折周期最为显著，对于基金个体而言灵活的交易策略增加了其收益，对于市场而言，这也会增加市场周期波动。(3)采用threshold面板数据模型，验证了中国证券基金对小盘股和成长股偏好对市场收益的依赖。结果发现Fama—French模型一般具有非线性形式，对应着不同基金的风险偏好差异，及其策略取向。这符合前景理论假说，说明基金等机构的反馈交易具有内在的心理特征基础。(4)进一步，采用状态空间模型，得到了不同投资风格基金的时变β，通过构造β的变异系数指标(即季度内β标准差与均值之比)，进一步分析了基金的风格漂移程度，提出了对风格漂移监管的策略。(5)针对基金对股市稳定作用的争论，采用多元GARCH模型研究了基金指数与股指的波动相关性，及其溢出效应的非对称性。指出基金资产配置与大盘成分股高度雷同，基金对股市的稳定作用有限。采用前面得到的基金投资策略指标，发现基金短期交易活跃度、仓位变动以及投资风格漂移都增加了市场波动。更多还原

【Abstract】 This dissertation consists of 4 comparatively self-contained papers devoted to econometric analysis of mutual fund’s investment Tactics Rotation in China’s stock market. The research presented has been focused on studying threshold panel data model and state space model, and their empirical application for bahavior finance theory to explain investment behavior of China’s mutual funds.Firstly , This Chapter three examines the relationship between investment performance and concentration in Chinese stock equity mutual funds portfolios, using the dynamic unbalanced panel data model developed by Bruno. (2005) .Through a panel data model of 95 stock equity mutual funds, the study shows a positive relationship between fund performance and portfolio concentration of stock, but a negative relationship between fund performance and portfolio concentration of industry. I also find funds with better performance tend to be those having smaller aggregate assets and lower turnovers.Secondly, because of the present disclosure requirements, the trade actions of Chinese stock equity mutual funds are unobserved. It strengthens managers’ principal-agent behavior of mutual funds and induces more manipulation. This forth chapter estimate the impact of unobserved actions on mutual funds’ returns using the return gap, which was defended as the difference between the reported fund return and the return of a portfolio that invests in previous disclosed holding by Kacperczyk, Marcin and Sialm (2006) .Through a panel data model of 30 stock equity mutual funds in China, the study shows a positive relationship between fund returns and return gap of stock during bull market, but a negative relationship during bear market. I also find funds with better return gap tend to be those having smaller aggregate assets and more active investment styles.Thirdly, based on prospect theory provided by Kahneman and Tversky (1979, 1983) , the fifth chapter gives a new explanation on investment style drifting of China’s mutual funds. Furthermore, adopting Hansen (1999) threshold panel data econometrical model, we estimate the modified Fama-French (1993, 1997) three factors model. The result shows the mental accounting of mutual funds’ manager is not related with their present return, but related with market return adjusted according with their stock equity asset’s proportion, which is the reason of investment style cyclical drifting and convergence. The estimation also shows HML index and SML index have different inference on excess return different under different threshold values. It describes the changing cycle difference of value-growth investment style and small-big cap investment style. At last, we brought forward a series of policy suggestion.Morever, behavior finance theory shows investors’ risk preference shift under the different market environment. Mutual funds investment style drifting damages the interest of mutual funds’ share holders. It also intensifies the volatility of stock market .Because the mutual funds’ short-term investment tactics is unobserved. A statistical state space approach is used to estimate the model. Recent algorithms are adopted for the computation of observation weights for forecasting based on state space models. The methodology is illustrated the investment style shifting of China’ stock equity mutual funds among different investment styles.Finanly, based on BEKK-GARCH model and structure vector auto regression model, this seventh chapter intends to discover the aggregate relationship between China’s stock equity mutual fund and the sock market. The bivariable GARCH model results suggest the two index volitility are highly correlative .The Granger causality test shows the feedback relation of CSI stock fund index and CSI 300 index. The structured VAR empirical results give the restricted estimation pattern. Further, the impulse responses display the stock equity mutual funds enlarge the short-term volatility of the stock market, but in longer term the mutual funds are beneficial to keep the stability of the stock market.更多还原