Despite their simple formulation, VARs are
very successful in capturing such stylised facts about economic
time series as the decaying pattern in autocorrelations’
values with increasing lag order, strong autocorrelations
at an annual frequency, as well as dynamic linear
interdependencies between time series. Therefore, VARs
are well suited to model data that collects vectors of observations
of N variables, denoted by yt, for the time subscript
t going from 1 to T.
The basic VAR equation is given by