powered by
Convert VHAR process to infinite vector MA process
VHARtoVMA(object, lag_max)
VMA coefficient of k(lag-max + 1) x k dimension
A vharlse object
vharlse
Maximum lag for VMA
Let VAR(p) be stable and let VAR(p) be \(Y_0 = X_0 B + Z\)
VHAR is VAR(22) with $$Y_0 = X_1 B + Z = ((X_0 \tilde{T}^T)) \Phi + Z$$
Observe that $$B = \tilde{T}^T \Phi$$
Lütkepohl, H. (2007). New Introduction to Multiple Time Series Analysis. Springer Publishing.