【作者】 梁巧；

【导师】 傅强；

【作者基本信息】 重庆大学 ， 金融市场与证券投资， 2011， 硕士

【摘要】 2005年以来,我国在外汇市场和股票市场相继实施了汇率机制改革和股权分置改革。本文基于GARCH簇模型,使用了2005年7月至2010年12月的日交易数据,分别考察了我国汇率市场和股票市场在收益率和波动性方面所具有的特征,并进行了两个市场间的动态时变相关性检验。研究结果对于金融市场监管和投资者的组合构建具有一定的理论价值与现实参考意义。论文的主要特色和结论如下:(1)通过建立具有虚拟变量的修正GARCH模型,同时在收益率和波动性上检验了汇率和股票市场所存在的周内效应。结果表明:在收益率序列方面,人民币美元汇率存在着周二和周四的负效应(升值),人民币欧元汇率仅存在周四的正效应(贬值),上证指数则同时存在着周二和周四的负效应(下跌);在波动性方面,人民币欧元汇率和上证指数均存在着周二的负效应(波动减小),人民币美元汇率的周内效应并不显著。(2)通过采用GARCH-M模型,分别探讨了汇率和股票市场的收益率和波动性之间的相互作用关系。研究揭示出,汇率市场中仅在人民币美元汇率的收益率与波动(风险)之间呈现显著风险与收益的负向关系,反映出风险越高则人民币美元汇率越容易升值,这可能是汇率市场投资者的自适应预期所导致;上证指数的收益率与波动之间呈现显著的负向关系,意味着中国投资者并没有因为承担高风险而获得补偿性的高收益。(3)GARCH模型能够较好地扑捉股市和汇市所具有的波动持续性。如上证指数波动冲击对股价造成的影响在30个交易日后尚有约88.7%,这说明股市一旦受到异常冲击,波动在短期内难以得以消除。(4)通过利用多元动态条件相关GARCH模型(DCC-MVGARCH)和多元常相关GARCH模型(CCC-MVGARH),检验了2005至2010年期间我国股票与汇率间的动态相关和常相关性。研究发现,股票与汇率间相关性具有动态时变特征,但其相关性较弱,更接近于显著为负的常相关关系。研究揭示出,由于股票市场和汇率市场均具有对于国家经济繁荣和衰退的指示器作用,使得两者之间保持着密切的联系。理论上,周内效应与有效市场假说相背离而被称为“异象”,本文检验出我国人民币汇率和股市存在不同程度上的周内效应,揭示出我国这两个市场均具有非有效性。因此,健全外汇和股份管理体制、优化汇率和股票形成机制将是我国汇率和股票改革的重点任务和工作目标。未来的研究包括深入揭示汇率和股票周内效应的其他影响因素,比如国内外财政货币政策等。同时,开展对我国汇率和股票的非线性复杂运行特征的探讨,也将会是非常具有理论价值及现实意义的工作。更多还原

【Abstract】 Since 2005,exchange rates reform and split-share structure reform have been implemented in China.By using GARCH cluster models and adopting the data from July 2005 to December 2010,this article attempts to observe the characteristics of returns and volatility in Chinese foreign exchange market and stock market .The research has useful information for the portfolio management,market regulation and risk controls.The main distinctive research work and conclusions are as follows:Firstly,when it comes to the day-of-week effect of returns sequence,RMB-USD exchange rate exits Tuesday and Thursday negative effects,while RMB-EUR exchange rate only exists Thursday effect during2005～2010,its economical meaning is that there are significant revaluation features in RMB-USD ,which is on Tuesday and Thursday , and the RMB-EUR is easier to devalue on Thursday.Shanghai compose index also appears negative effects on Tuesday and Thursday.Meanwhile, RMB-EUR exchange rate and the Shanghai index all exist negative effects on Tuesday as mentioned on the volatility ,which means the decline of volatility.Secondly,In foreign exchange markets,there is significant negative relationship between the returns and fluctuation only in RMB-USD during the whole sample,stating that the more risk the more RMB-USD increase in value,which may be caused by the currency market investor’s adaptive expection .Simultaneously,the Shanghai stock indix exhibits significant negative relationship between the returns and fluctuation,which means that Chinese investors will not get more returns as bearing high risk.Thirdly,the stock market and exchange market all have strong continued volatility which can well be captured by GARCH models.For example,once the Shanghai index is sufferred from a exogenous shock,the impact will still remain 88.7% after 30 trading days. Thus,it is difficult for stock market to eliminate abnormal fluctuations in the short term.Finally,after utilizing DCC-MVGARCH model and CCC-MVGARCH model,the article studies the relationships between foreign exchange market and stock market. The empirical results reveal that the relationship exhibits time-varying feature but more close to constantly negative correlation due to the role of economic indicator which exchange rate and stock index all possess. Theoretically, day-of-week effect is so called " abnormal phenomena" because it departures from efficient market hypothesis. The results reveals two markets are all not of effectiveness.Therefore,the reform task and working targets of foreign exchange and stock marekts would be perfect management system and optimized formation mechanism.Of course,much more work is needed.One challenge for future research will be planned to investigate other influence factors of day-of-week effect in foreign exchange market and stock market,such as fiscal and monetary policy.Meanwhile,we are going to study the nonlinear complex dynamic of exchange rates and stock indices.更多还原