A0680
Title: Application of non-Gaussian SVAR model to the analysis of Japans quantitative easing monetary policy
Authors: Tadashi Nakanishi - Hokkaido University (Japan) [presenting]
Koichi Maekawa - Hiroshima University of Economics (Japan)
Takashi Senda - Hiroshima University (Japan)
Abstract: In recent years the independent component analysis (ICA), originally developed in machine learning, has been introduced to time series econometrics. The number of papers using ICA is increasing. The advantage of the ICA approach is that it can give an alternative way to avert identification problem under non-Gaussian disturbances. Another advantage is that if the contemporaneous coefficient matrix in Structural VAR is lower triangular, then the causal order of economic variables can be easily detected. This paper attempts to check the effectiveness ICA-based SVAR model by Monte Carlo experiment and to compare the performance of several existing macro models. Furthermore, we apply this model to detect the effect of quantitative easing monetary policy by the Bank of Japan.
