中美国债市场溢出效应研究——基于集合经验模态分解

Research on Spillover Effect of Government Bonds between China and USA:Based on Ensemble Empirical Mode Decomposition

集合经验模态分解(EEMD)是目前国际公认的可以有效处理非平稳非线性时间序列的方法。在中美融合不断加深的背景下,首次将该方法用于中美国债指数的分解,筛选出的低频分量及趋势项能够有效表征原始序列。基于此建立了二元VAR-GARCH-BEKK模型,从均值与波动两个维度考证了中美国债市场的溢出效应,结果发现:面对重大事件及经济政策的冲击,两国国债之间表现出双向的均值溢出,但影响程度不对等,美国对中国的溢出相对较强,而波动溢出效应则是对称的;国债市场长期运行过程中,两国具有双向的波动溢出,但仅存在美国对中国的单向均值溢出。中国应进一步完善国债市场交易机制,丰富国债交易品种,加强金融市场监管等,来有效对冲美国市场对中国的溢出效应,防范与化解系统性风险。

The ensemble empirical mode decomposition is an internationally recognized method which can process non--stationary and nonlinear time series effectively. In the context of Sino--US convergence, this paper firstly decomposes the Sino--US Treasury bond index with EEMD, and filter out low frequency components and trends respectively which can represent the original time series exactly. Then it builds the VAR--GRACH--BEKK model to verify spillover effect of government bonds exist or not between China and USA from the mean and volatility. The results show as follows： when unexpected events occur and economic policies change, bilateral mean spillover effect exists, but the extent is not symmetrical, US＇ bond spills over into China＇s strongly, while volatility spillover effect is symmetric. In the long process of government bond＇s intrinsic development, volatility spillover effect is symmetric, but mean spillover effect exists from US to China unidirectionally. China should further improve the bond market trading mechanism, rich the government bond＇s varieties, and strengthen financial market regulation, so as to hedge the spillover effect from US to China effectively, prevent and mitigate the systemic risk.