Vector Autoregressive Models With Measurement Errors for Testing Ganger Causality

TL;DR

This paper introduces a new method for estimating VAR model parameters with measurement errors, enabling reliable Granger causality testing even with noisy data, demonstrated through simulations and fMRI data analysis.

Contribution

It provides a non-iterative, consistent estimation approach for VAR models with measurement errors, addressing unidentifiability issues in Granger causality testing.

Findings

Empirical false positive rates are close to nominal levels.

Method performs well even with small samples.

Effective in real fMRI data analysis.

Abstract

This paper develops a method for estimating parameters of a vector autoregression (VAR) observed in white noise. The estimation method assumes the noise variance matrix is known and does not require any iterative process. This study provides consistent estimators and shows the asymptotic distribution of the parameters required for conducting tests of Granger causality. Methods in the existing statistical literature cannot be used for testing Granger causality, since under the null hypothesis the model becomes unidentifiable. Measurement error effects on parameter estimates were evaluated by using computational simulations. The results show that the proposed approach produces empirical false positive rates close to the adopted nominal level (even for small samples) and has a good performance around the null hypothesis. The applicability and usefulness of the proposed approach are illustrated using a functional magnetic resonance imaging dataset.