Convert VHAR process to infinite vector MA process
Usage
VHARtoVMA(object, lag_max)
Arguments
- object
A vharlse object
- lag_max
Maximum lag for VMA
Value
VMA coefficient of k(lag-max + 1) x k dimension
Details
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
$$B = \tilde{T}^T \Phi$$
References
Lütkepohl, H. (2007). New Introduction to Multiple Time Series Analysis. Springer Publishing.
