2020/24 | LEM Working Paper Series | ||||||||||||||||
Identification of Structural VAR Models via Independent Component Analysis: A Performance Evaluation Study |
|||||||||||||||||
Alessio Moneta and Gianluca Pallante |
|||||||||||||||||
Keywords | |||||||||||||||||
Independent Component Analysis; Identification; Structural VAR; Impulse response functions; Non-Gaussianity; Generalized normal distribution.
|
|||||||||||||||||
JEL Classifications | |||||||||||||||||
C14, C32, E62
|
|||||||||||||||||
Abstract | |||||||||||||||||
Independent Component Analysis (ICA) is a statistical method that transforms a set
of random variables in least dependent linear combinations. Under the assumption
that the observed data are mixtures of non-Gaussian and independent processes, ICA
is able to recover the underlying components, but a scale and order indeterminacy. Its
application to structural vector autoregressive (SVAR) models allows the researcher
to recover the impact of independent structural shocks on the observed series from
estimated residuals. We analyze different ICA estimators, recently proposed within the
field of SVAR identification, and compare their performance in recovering structural
coefficients. Moreover, after suggesting an algorithm that solve the ICA indeterminacy
problem, we assess the size distortions of the estimators in hypothesis testing.
We conduct our analysis by focusing on distributional scenarios that get gradually
close the Gaussian case, which is the case where ICA methods fail to recover the
independent components. In terms of statistical properties of the ICA estimators, we find
no evidence that a method outperforms all others. We finally present an empirical
illustration using US data to identify the effects of government spending and tax cuts
on economic activity, thus providing an example where ICA techniques can be used
for hypothesis testing.
|
Downloads
|
|
| |
|
Back
|
|