A Correlation Study Between The Term Structure Of Interest Rates And Macroeconomic Variables

Interest rate term structure is composed of spot rates of different maturities and has different forms at different time points, it contains a wealth of economic information, so it is an important reference for the central bank’s macro-control. The government explicitly proposed "Accelerate the marketization of interest rate and perfect the treasury yield curve" which was included in the core national development strategy. Therefore, it has been a long time searching for better fitting and forecasting evolution model of term structure of interest rates in academic and practical circles. Further, we study the term structure of interest rates from a combination of micro and macro points, which is important for understanding the trend of interest rate term structure and the relationship between the interest rate term structure and macroeconomic variables.Firstly, in this paper, we consider the traditional dynamic Nelson-Siegel model and possible conditional heteroscedasticity in different time and different term to get the dynamic Nelson-Siegel model with GARCH effect. The GDNS model describe the dynamics of the term structure through the affine model of "horizontal", "slope" and "curvature" potential factor, the model is not only simple structure but also has clear economic implications. After fitting the treasury spot rate monthly data between the Bank of China form March 2006 to March 2015, we discover that the GDNS model is better than the traditional dynamic Nelson-Siegel model in fitting and forecasting.Then, on the basis of GDNS model, we further use the kalman filter to obtain three main factors which can explain the changes in term structure.Those three factors are called "level factor" "slope factor" and "curvature factor", we not only do some relevant empirical analysis for these three latent factors but also verify the economic implications of the three latent factors, namely:the level factor represents the long-term interest rates, the slop factor represents the Long-short spread, the curvature factor represents the medium-term interest rates.Finally, in order to study the relationship between macroeconomic variables and the term structure, this paper do some relevant empirical analysis for these three latent factors and three macro variables, correlation analysis shows that the correlation between slope factor and three macro factors are more obvious than the level factor and the curvature factor. The slope factor is highly negatively correlated with the broad money supply growth rate and also highly positively correlated with the CPI. This result suggests that the slope factor can be used as an indicator of the state’s macroeconomic performance metrics, reflecting changes in monetary policy and inflation. When the short-long spreads become big, it means that the market participants predict a big press on CPI, and the country will take loosen monetary policy, the same to the opposite condition. Considering the curve contains the information about the trend of the CPI, It is useful for the country to make policy. Next, the paper established a VAR model that contains the potential factors and macroeconomic variables and do the granger causality test, impulse response analysis and variance analysis. The VAR model shows that the correlation between inflation>. the growth rate of the broad money supply and three latent factors are more obvious than industrial production growth. Granger causality test shows that the correlation between the interest rate term structure and the macroeconomic variables but the broad money supply growth rates are not bidirectional, Term structure is not only affected by the broad money supply, but also reflects the changes of the broad money supply in the future. In addition to the broad money supply, the term structure also has a significant impact on inflation and industrial production, which reflects the effect of macroeconomic indicators of the term structure. Impulse response analysis and correlation analysis conclusions are basically the same. Variance decomposition results show that variance contribution rate of macroeconomic factor for predicting short-term yield curve is not strong, but the contribution rate gradually increased with the forecast period increasing.By considering the conditional heteroscedasticity in different time and different term, this paper fit the interest rate term structure by the GDNS model, which improves the fitting effect. But this paper did not compare different methods of fitting, so it did not draw the most suitable fitting method for China’s yield curve, which is the shortcomings of this paper.