非线性范式下汇率行为与汇率管理机制研究
发布时间:2018-04-30 01:33
本文选题:汇率行为 + 非线性 ; 参考:《湖南大学》2012年博士论文
【摘要】:汇率行为是汇率系统复杂规律的外在表现,是在特定的管理机制下对外部影响因素变化进行感知和响应的客观结果,其特征能够反映出经济体的发展水平和金融市场的成熟程度。当前,受到全球性金融危机的影响,各国汇率行为的异动增多,呈现出更加显著的动态复杂性,由此也引发了学术界和实务界对汇率管理机制的高度关注。因此,全面总结和提练汇率行为的非线性特征,探索非线性范式下的汇率行为组合预测模型,并定量研究汇率管理基准和管理行为有效性,均具有重要的理论价值和现实意义。 本文依照汇率行为特征与汇率管理机制相一致的研究框架,遵循由理论探究到实证分析、再到政策建议的研究路线,针对各国汇率行为特征及其汇率安排的特点,进行了系统的规范化研究。通过构建非线性模型并融合数字信号处理技术、机器学习算法以及数理统计与计量经济方法,对非线性汇率行为与汇率管理机制进行了若干具有一定创新性的研究工作。主要的研究内容如下: 首先,本文界定了汇率行为与汇率管理机制的基本内涵,强调汇率管理机制是中央银行管理调控本国汇率的根本机理和操作准则,既包括汇率管理基准的确定,也包括以此基准为参考对本国货币汇率行为的管理调控机理。同时,系统地梳理了汇率行为与汇率管理机制的基本理论,并对相关的参数与非参数研究方法进行了比较分析,借此阐明了非线性研究方法在汇率行为与汇率管理机制研究中的优势。其次,通过综述汇率行为非线性特征的研究进展,归纳总结出汇率行为普遍具有非正态性、波动聚集性、非对称性以及长记忆性等典型的非线性特征。针对上述特征,分别应用Jarque-Bera统计检验方法、ARCH统计检验方法、ARMA-GJR-GARCH模型以及MR/S模型等对4个发达国家和5个新兴市场国家汇率的非线性行为特征进行了实证检验与度量。比较分析发现,新兴市场国家的汇率行为具有更加显著的非线性特征和杠杆效应。此外,通过对上述汇率行为非线性特征的检验与度量发现,在外汇市场中有效市场假说并不成立。再次,本文依据汇率行为的非线性基本特征,提出了基于独立分量分析方法与支持向量机技术的汇率行为组合预测模型。利用人民币兑美元的汇率对该模型的预测效果进行了实证检验,并与其它典型的预测模型进行了比较。实证结果表明,本文所提出的IC-SVM模型具有较高的预测精度和稳健性。接着,首次提出人民币名义货币篮子和实际货币篮子的概念,较好地解释了官方宣称的货币篮子构成与学术界研究结论之间的差异。在此基础上,迭代应用上确界F检验,较准确地侦测到汇率政策调整对汇率管理基准的影响,从而获得了更加有效的汇率管理参考基准。在此基础上,提出了基于工具变量法和广义自回归条件异方差模型的组合范式,修正了由汇率收益与管理调控行为之间的共生性所导致的估计偏误,并对汇率管理调控手段的有效性及其影响因素进行了实证研究。实证结果表明,央行外汇市场操作是最为直接有效的汇率管理手段,但同时也会加剧汇率波动,可能带来额外的货币风险。最后,在上述实证研究的基础上,本文系统地阐述了人民币汇率形成机制的历史发展进程和阶段性内涵,讨论了当前人民币汇改进程中面临的核心问题与挑战,同时提出推进汇改的政策建议,并展望了人民币汇率的发展前景。
[Abstract]:The exchange rate behavior is the external manifestation of the complex law of the exchange rate system, the objective result of the perception and response to the changes of external factors under the specific management mechanism. Its characteristics can reflect the development level of the economy and the maturity of the financial market. The increase, showing a more significant dynamic complexity, has also aroused the high concern of the exchange rate management mechanism in the academic and practical circles. Therefore, the nonlinear characteristics of the exchange rate behavior are summarized and practiced in a comprehensive way, and the combination forecast model of the exchange rate behavior under the nonlinear paradigm is explored, and the exchange rate management benchmarks and the effectiveness of the management behavior are studied quantitatively. All of them have important theoretical and practical significance.
In accordance with the research framework consistent with the exchange rate behavior characteristics and exchange rate management mechanism, this paper follows from theoretical to empirical analysis, and then to the research route of policy recommendations. The systematic research on the characteristics of exchange rate behavior and exchange rate arrangement is carried out systematically. The nonlinear model is constructed and the digital signal processing technology is fused. The machine learning algorithm and mathematical statistics and econometric methods have carried out a number of innovative research work on the nonlinear exchange rate behavior and exchange rate management mechanism. The main contents are as follows:
First, this paper defines the basic connotation of exchange rate behavior and exchange rate management mechanism, and emphasizes that the exchange rate management mechanism is the fundamental mechanism and operating criteria of the central bank management and regulation of the national exchange rate, including the determination of the exchange rate management benchmarks and the management and regulation mechanism of the exchange rate behavior of the domestic currency as a reference. The basic theory of exchange rate behavior and exchange rate management mechanism is combed, and the relative parameters and non parameter research methods are compared and analyzed. In this way, the advantages of the nonlinear research method in the study of exchange rate behavior and exchange rate management mechanism are clarified. Secondly, the exchange rate is summarized and summarized through the review of the research progress of the nonlinear characteristics of exchange rate behavior. Behavior generally has the typical nonlinear characteristics such as non normality, volatility aggregation, asymmetry and long memory. According to the above characteristics, the Jarque-Bera statistical test method, ARCH statistical test method, ARMA-GJR-GARCH model and MR/S model are applied to the nonlinear behavior of the exchange rates of 4 developed countries and 5 emerging market countries. The empirical test and measurement are carried out. The comparative analysis shows that the exchange rate behavior of emerging market countries has more significant nonlinear characteristics and leverage effects. In addition, the effective market hypothesis in the foreign exchange market is not established by testing and measuring the nonlinear characteristics of the exchange rate behavior. Again, this paper is based on the exchange rate behavior. A combination prediction model of exchange rate behavior based on independent component analysis and support vector machine is proposed. The prediction effect of the model is tested with the exchange rate of RMB against the dollar, and the comparison with other typical forecasting models is made. The results show that the IC-SVM model proposed in this paper has a good result. Then, the concept of the nominal currency basket and the actual currency basket is first proposed, which explains the difference between the official declared currency basket and the academic research conclusion. On this basis, the iterative application is confirmed by the F test, and the exchange rate policy adjustment is more accurate. On the basis of this, a combination paradigm based on the tool variable method and the generalized autoregressive conditional heteroscedasticity model is proposed on the basis of the influence of the management benchmarks, and the estimation error caused by the symbiosis between exchange rate income and management regulation behavior is amended, and the control means of exchange rate management are also adjusted. The empirical results show that the operation of the central bank's foreign exchange market is the most direct and effective means of exchange rate management, but it also aggravates the fluctuation of exchange rate and may bring additional currency risk. Finally, on the basis of the above empirical study, this paper systematically expounds the formation mechanism of the RMB exchange rate. The historical development process and the stage connotation, discussed the core problems and challenges facing the current RMB exchange reform process, and put forward the policy proposals to promote the exchange reform, and looked forward to the prospects for the development of the RMB exchange rate.
【学位授予单位】:湖南大学
【学位级别】:博士
【学位授予年份】:2012
【分类号】:F831.6;F224
【参考文献】
相关期刊论文 前6条
1 干杏娣;杨金梅;张军;;我国央行外汇干预有效性的事件分析研究[J];金融研究;2007年09期
2 李稻葵;尹兴中;;国际货币体系新架构:后金融危机时代的研究[J];金融研究;2010年02期
3 韩峰;谢赤;孙柏;;基于IV-GARCH模型的外汇干预有效性实证研究[J];金融研究;2011年06期
4 谢赤;杨妮;孙柏;;汇率时间序列混沌动力学特征及实证[J];系统工程理论与实践;2008年08期
5 孙柏;谢赤;;金融危机背景下的人民币汇率预测[J];系统工程理论与实践;2009年12期
6 张在美;谢赤;孙柏;;外汇干预反应函数非线性FTR模型的提出及其实证[J];中国管理科学;2010年04期
相关会议论文 前1条
1 许友传;;金融时间序列R/S长记忆性检验的有效性[A];第十二届中国管理科学学术年会论文集[C];2010年
,本文编号:1822503
本文链接:https://www.wllwen.com/guanlilunwen/huobilw/1822503.html