Asymptotic Distributions of Impulse Response Functions and Forecast Error Variance Decompositions of Vector Autoregressive Models

TLDR

Vector autoregressive models are widely used in economics to analyze interrelationships, yet impulse response functions and forecast error variance decompositions often lack reported estimation uncertainty. This article seeks to summarize existing asymptotic distribution results and supply missing links to facilitate computation of standard errors and test statistics. The authors compile and synthesize asymptotic distribution theory from disparate publications into a coherent framework for impulse response functions and forecast error variance decompositions. © 1990 MIT Press.

Abstract

In recent years, vector autoregressive models have become standard tools for economic analyses. Impulse response functions and forecast error variance decompositions are usually computed from these models in order to investigate the interrelationships within the system. However, sometimes no measures of estimation uncertainty are provided by authors. One reason may be that the relevant asymptotic distribution theory is distributed over various publications. In this article, the available results are summarized and the missing links are provided in order to facilitate the computation of standard errors and test statistics. Copyright 1990 by MIT Press.