Statistical Research On Interaction Between Term Structure Of Credit Spreads And Macroeconomic Factors

In this dissertation, an estimation approach based on the genetic algorithm (hereafter GA) is introduced for the extended Nelson-Siegel model to improve its fitness, which is proved by an empirical comparison between the method established above, the bootstrap method based on cubic spline interpolation and the extended Nelson-Siegel model based on nonlinear regression. Then the model introduced above is used to construct the yield curves for both defaultable and Treasure bonds, based on which the term structure of credit spreads and then a composite index and level, slope, curvature factors of credit spreads are constructed. Then, using the indexes constructed above, on one hand, the influence of different macroeconomic factors on the term structure of credit spreads is analyzed by stepwise regression to find the macroeconomic factors with significant influence, and the dynamic lagged effects of the macroeconomic factors above on the term structure of credit spreads are studied through impulse response and variance decomposition. On the other hand, the predictive power of the term structure of credit spreads for the future economic growth is analyzed by the GMM method under a regression framework.Through statistical analysis of the influence of macroeconomic factors on the composite index of credit spreads, it is found that the slope of the yield curve is the most influential, and the default-free rate comes next, while M2 and the stock market value are the least. The slope of the yield curve, M2 and the stock market value are all negatively related to credit spread, while the default-free rate is positively related to it, which is opposite to the theory. Furthermore, the statistical analysis of the influence of macroeconomic factors on the shape factors of the term structure of credit spreads reveals that the slope of the yield curve and the default-free rate are main influential factors on the level factor, while the volatility of stock market value and M2 on the slope factors, and the default-free rate on the curvature factor.The statistical research on the predictive power of the composite index of credit spreads for the future economic growth shows that the composite index of credit spreads has significant but varying predictive power for the future economic growth at different forecast horizon, which is more powerful for the medium-term horizon while less for the short and long horizons. Furthermore, the statistical study of the predictive power of the shape factors of the term structure of credit spreads for the future economic growth reveals that the level factor has modest predictive power for the future economic growth at medium and short term horizon, while little for the future growth at long term horizon. On the contrary, the slope factor has significant predictive power for the future economic growth at long term horizon, while little for the future growth at medium and short term horizon. The empirical result also shows that the curvature factor has significant predictive power only for the future economic growth at five-month horizon.