Study On The Dynamic Correlation Between The Term Structure Of Interest Rate And Macroeconomic Factors

With the rapid development of China’s economy in recent years,China’s economy is in urgent need of a sound and stable financial market as its support.Interest rate liberalization is an important part of China’s financial reform.With the beginning of interest rate liberalization reform in 1996,China’s interest rate liberalization has made remarkable achievements,and is now in an overall sprint stage.In an economic environment where interest rates are determined independently by the market,central Banks often use the Taylor rule to maintain the stability of short-term interest rates,thereby affecting long-term interest rates and finally achieving the purpose of regulating social output and inflation.After the outbreak of the American financial crisis,the effect of interest rate term structure on macro economy has been further strengthened,such as after the outbreak of the financial crisis in 2008,in order to get rid of the economic crisis,the US federal reserve modulated the federal benchmark interest rate to the lowest level in history,and by means of"quantitative easing" changed the term structure of interest rate to achieve its economic goals.Therefore,studying the term structure of interest rate and its related issues has important academic value and practical significance for analyzing and judging the macroeconomic situation and expectation.The construction of term structure model of interest rate can be divided into static model and dynamic model,however,due to the financial market become more and more complex,neither the estimation method of static model nor dynamic model can meet the needs of modern market in terms of accuracy and stability.The deviation of the fitted yield will make us unable to accurately grasp the true change characteristics of the term structure at the micro level,and it is difficult for the government to provide the corresponding empirical support at the macro level.So this paper from the estimation method of term structure of interest rates,improve the estimate of the static model and dynamic model method,improve the accuracy of the term structure of interest rates describes and stability of algorithm,at the same time,based on the optimization of the accuracy of the term structure of interest rates,closely combine macroeconomic information,analyze dynamic characteristics of the term structure of interest rates and macroeconomic factors,which can help us to analyze and judge the macroeconomic situation and expectation.This paper is divided into seven chapters,mainly doing the following five aspects of work:First,we elaborate the relevant theoretical basis of the term structure of interest rates,and summarize the relevant literature at home and abroad.First,a systematic introduction of the basic theory of term structure of interest rates,straighten out the development,then elaborate about static term structure of interest rates,dynamic term structure of interest rates,and macro-financial model on the related literature,summarizing their advantages and disadvantages,and is pointed out that the innovation of this paper compared with the previous studies.Second,we study the static model estimation method of the term structure of China’s national debt interest rate.Firstly,introduce the static model of the term structure of China’s national debt interest rate.The cubic spline function,Nelson-Siegel and its extended form model,which are commonly used in static models,are introduced.It is pointed out that the estimation method of interest rate term structure model based on cubic spline function,NS model and Svensson model has low parameter estimation accuracy,and the overall fitting effect needs to be optimized.To solve the problem that the model fitting accuracy is not high,this paper proposes an improved genetic algorithm,which is applied to the dividing point of cubic spline function and parameter estimation of NS model and SV model.Through the improved genetic algorithm,the problem that the traditional genetic algorithm is easy to fall into the local optimal solution can be solved,and the accuracy of the overall algorithm is improved.And the data of Chinese national debt is used for empirical analysis and research.The research results show that the static model obtained by the improved genetic algorithm has greatly improved the fitting of in sample data and out of sample data.And the related parameters in the algorithm are discussed in the polynomial spline function model,and the optimal range of relevant parameters is obtained.Third,we study the estimation method of the no-arbitrage Nelson-Siegel model.At present,use Kalman filter to estimate parameters in no-arbitrage NS model,However,in the Kalman filtering algorithm,it is more dependent on the setting of the initial value of parameters.At the same time,as the number of estimation gradually increases,the mean value of error and the covariance of estimation error will gradually enlarge in the iteration process,which will reduce the accuracy of filtering estimation.Therefore,this paper proposes an improved adaptive Kalman filtering method.Through empirical comparison analysis,it is found that the spot interest rate based on the improved adaptive Kalman filter estimation is closer to the true value than the spot rate based on the traditional Kalman filter estimation,and it can better describe the variation characteristics of the term structure of China’s national debt.Finally,by using the improved adaptive Kalman filter to extract the three state factors in the model,we find that the three factors of level,slope and curvature are highly coincident with the empirical proxy variables.Through the correlation analysis between the latent factor and its own lag term,it is concluded that the horizontal factor represents the long-term interest rate,and the affected duration is longer when subjected to external shocks;the persistence of the curvature factor is higher than the slope factor,but lower than the horizontal factor.It represents the medium-term interest rate;the slope factor has the shortest durability and the highest volatility.Fourth,we study the correlation between term structure of national debt interest rate and macroeconomic variables.In this paper,the macroeconomic variables are added in Gaussian affine model to construct the macro financial affine model of the term structure of the interest rate of China’s government bonds,and the parameters in the macro financial affine model are estimated by the improved adaptive Kalman filtering method proposed in this paper.The research focuses on the correlation between macroeconomic factors such as economic condition,fiscal policy and monetary policy and the term structure of interest rate.The research is based on the no-arbitrage affine model based on potential factors and macroeconomic factors,and use China’s national debt interest rate data and macroeconomic data to empirically the study.Through empirical research,it is found that fiscal factors and monetary factors have a relatively positive impact on the term structure of China’s national debt interest rate;economic state factors have a relatively positive effect on China’s short-term national debt,but have less effect on long-term national debt;The inflation factor has a positive effect on the short-term maturity spot rate,but as the term increases,the effect gradually weakens and even a weak negative effect.This conclusion is consistent with the actual economic situation.This is because from 2014,China’s economic growth began to slow down,and the slowdown in economic growth offset the positive impact of inflation growth on interest rates;At the same time,the potential factor coefficient is larger than the macroeconomic factor coefficient,indicating that the lag term of interest rate has a greater impact on the term structure of interest rate than the impact of China’s macroeconomic factors on the term structure of interest rate,indicating that the transmission mechanism between bond yield rate and macroeconomic variables is still not perfect.Fifth,we further study the relationship between term structure of interest rate and macro-economy.Since the macro financial affine model can only contain a few macroeconomic variables,the term structure of interest rate cannot correctly reflect its close correlation with the macro economy,and it is also impossible to know the influence and transmission effect of monetary policy expectations on the term structure of national debt interest rate.Therefore,based on the FAVAR model,this paper incorporates more suitable macroeconomic variable information,including expected monetary policy and unanticipated monetary policy indicators into the model.In the model,the improved adaptive Kalman filter proposed in this paper is firstly used to estimate the term structure of spot rate,and then the dynamic correlation between the term structure of interest rate and macroeconomic variables and the impact of monetary policy expectations on the term structure of interest rate are investigated.The empirical results show that the improved adaptive Kalman filter is more effective in the macro-financial model based on FAVAR than the traditional Kalman filter;The macro variables have a relatively stable impact on the term structure of interest rates,and there is no large fluctuation.The three factors of interest rate term structure have great influence on macro variables,and the dynamic correlation between curvature factor and macroeconomic variables is similar to the horizontal factor;during the economic slowdown,the correlation between macro variables and the term structure of interest rates is unchanged,but the effect of the impact is greater,and the duration of impact is shortened;the impact of unanticipated monetary policy on the term structure of interest rates is much greater than that of expected monetary policy.The innovation of this paper lies in:Firstly,an improved genetic algorithm is proposed to examine the effect of the improved genetic algorithm on the static model of the term structure of interest rate of China’s government bonds.The algorithm not only adds a hierarchical mode to the selection operation,but also improves the dynamic crossover probability proposed by Srinivas in the crossover process,and proposes the concept of "mean neighborhood"According to the "mean neighborhood",the crossover probability can be dynamically adjusted more effectively.Through two improvements,the excellent genes of individuals in the population can be further preserved,and the optimization effect of the overall algorithm is improved.In this paper,the improved genetic algorithm is applied to the parameter estimation of polynomial spline function model,Nelson-Siegel model and Svensson model in the static model of interest rate term structure,which greatly improves the fitting effect of the above static model and gets a more accurate estimate of interest rate term structure.Secondly,the estimation methods of the no-arbitrage dynamic Nelson-Siegel model and the macro financial affine model are improved,and an improved adaptive Kalman filtering method is proposed.The algorithm adds an exponential attenuation factor when calculating the mean square error of the prediction,which avoids the phenomenon of filter divergence in the filtering process,and then obtains more optimized parameter values.The fitted spot interest rate value is closer to the actual value,and Based on this,the correlation between the term structure of interest rates and macroeconomic variables is studied.Thirdly,the improved adaptive Kalman filter is used to estimate the macro financial model based on FAVAR form.The general macro-financial model can only contain a few macroeconomic variables,and it is impossible to conduct a deep analysis of the relationship between the term structure of interest rates and the macroeconomic factors.This paper uses FAVAR model to take into consideration a variety of macro variables including expected monetary policy and unexpected monetary policy,and not only further investigates the dynamic correlation between the term structure of interest rate and macroeconomic factors,but also investigates the impact of monetary policy expectations on the term structure of interest rate.At the same time,the improved adaptive Kalman filter is used to accurately estimate the parameters in the macro financial model based on FAVAR form.