Research on impulse response and variance decomposition analysis of co-integrated systems

Abstract

The cointegration test explores whether a relationship of long-term equilibrium exists between the multi-dimensional series. It helps to exclude the possibility of spurious regression. As for the vector auto-regressive (VAR) model, it describes the interrelationship between endogenous variables. With impulse response and variance decomposition, the VAR model can further capture the dynamic response path of the dependent variable after being impacted by other variables, and the contribution rates of different variables to the impulse response of the dependent variable. In this paper, impulse response and variance decomposition analysis of cointegrated systems are adopted based on the VAR model to evaluate measurable data of RMB internationalization. The framework consists of four steps. First, the long-term equilibrium between variables after confirming the existence of cointegration relationship is explored. Second, the VAR model is adopted to consider the endogeneity between variables. Then, through the impulse response function, the dynamic response path of the system impacted by unit impulse is observed. Finally, combined with variance decomposition, further analysis is taken to realize the contribution rates of variables to the impulse. The results show that foreign influencement has the greatest impact on RMB internationalization, followed by foreign exchange reserves. However, the corresponding paths of RMB internationalization still requires to be adjusted.