基于Nelson-Siegel族模型对国债利率期限结构的动态拟合与预测研究

发布时间：2018-03-28 13:10

本文选题：国债利率期限结构　切入点：Nelson-Siegel族模型　出处：《吉林大学》2014年硕士论文

【摘要】：近年来，中国国债发行量逐年增大，国债在金融市场上的地位逐渐凸现出来。随着经济的发展，利率市场化的推进，国债利率期限结构研究的重要性也日益显现。众所周知，国债利率期限结构更是金融产品设计、研究货币政策效用及其传导机制、金融资产的定价、金融市场利率风险管理的关键。很多学者及相关政策制定者都期待着能找到合适的方法来拟合及预测中国的国债利率期限结构，目前大部分学者认为Nelson-Siegel模型及其扩展模型（统称为NS族模型，，包括NS、DNS、SV、ASV模型）较为适合中国国债利率期限结构，但是由于实证检验中，方法选取不适、数据样本量过小亦或没有模型的对比研究，对于哪种模型更适合中国国债利率期限结构，一直未能得到统一的结论。鉴于此，本文将基于Nelson-Siegel族模型对中国国债利率期限结构的拟合及预测进行对比研究，以期待能找到最适合拟合及预测中国国债利率期限结构的最优模型。本文在国债利率期限结构重要性不断加大这一背景下，首先对利率期限结构的传统理论及现代理论进行研究，得出较为适合中国国债利率期限结构研究的模型——NS族模型。然后，基于NS族模型，先分别对上海交易所市场和银行间市场上2005年—2010年国债利率期限结构的月度数据进行拟合，采用均方误差及绝对误差指标比较四种模型的动态拟合效果；然后，利用拟合过程中所形成各模型的动态参数，采用被学者们认为预测效果较好的一阶自回归（AR（1））方法得到2011年—2013年的参数预测值，将真实参数值与预测值进行比较，通过参数偏差率指标比较各模型的预测效果。实证结果表明，在四种模型中，Nelson-Siegel模型对我国国债利率期限结构的动态拟合及预测效果均为最优，这在某种程度上解释了很多国家都运用Nelson-Siegel模型进行本国货币政策制定的原因。最后，基于实证检验结果，本文从三个方面对中国国债市场的体制建设提出了相关的政策建议，希望可以进一步完善国债市场，并对日后国债利率期限结构的研究奠定基础。
[Abstract]:In recent years, the issuance of national debt in China has been increasing year by year, and the position of national debt in the financial market has gradually emerged. With the development of economy and the promotion of marketization of interest rate, the importance of studying the term structure of national debt interest rate is becoming increasingly apparent. The term structure of national debt interest rate is the design of financial products. It studies the utility of monetary policy and its transmission mechanism, the pricing of financial assets, The key to interest rate risk management in financial markets. Many scholars and relevant policy makers are looking forward to finding a suitable way to fit and predict the interest rate maturity structure of China's government bonds. At present, most scholars think that the Nelson-Siegel model and its extended model (collectively called NS family model, including NSD NSN / Nelson-Siegel model) are more suitable for the term structure of interest rate in China's treasury bonds. However, the method is not suitable for the empirical test. There is no uniform conclusion as to which model is more suitable for the term structure of the interest rate of Chinese government bonds, because the data sample size is too small or there is no model. Based on the Nelson-Siegel family model, this paper makes a comparative study on the fitting and forecasting of the term structure of the interest rate of China's treasury bonds, in order to find the best model for fitting and forecasting the term structure of the interest rate of China's national debt. Under the background of the increasing importance of the term structure of national debt interest rate, this paper firstly studies the traditional theory and modern theory of the term structure of interest rate. Then, based on the NS family model, the monthly data of the term structure of the interest rate in the Shanghai Stock Exchange and the interbank market from 2005 to 2010 are respectively fitted. The mean square error and absolute error index are used to compare the dynamic fitting effect of the four models, and then, the dynamic parameters of each model formed during the fitting process are used. The first order autoregressive method, which is considered by scholars to be better, is used to get the predicted values of the parameters from 2011 to 2013, and the real values are compared with the predicted values. The empirical results show that the Nelson-Siegel model is the best in the dynamic fitting and forecasting of the term structure of national debt interest rate in China. To some extent, this explains the reasons why many countries use Nelson-Siegel model to formulate their monetary policy. Finally, based on the empirical results, this paper puts forward some relevant policy recommendations on the institutional construction of China's national debt market from three aspects. It is hoped that the government bond market can be further improved and the research on the term structure of interest rate of government bonds in the future will be laid a foundation.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F832.51;F224

【参考文献】