Interpreting the result of VAR model using Matlab

1 view (last 30 days)
The result I got from runnning a VAR model looks like this:
Value StandardError TStatistic PValue
Constant(1) 0.0019 0.0015 1.2429 0.2139
Constant(2) 0.2972 0.0926 3.2091 0.0013
Constant(3) 1.3198 2.8587 0.4616 0.6443
AR{1}(1,1) 0.3564 0.0671 5.3065 1.1177e-07
AR{1}(2,1) -1.6369 3.8915 -0.4206 0.6740
AR{1}(3,1) 276.18 120.1 2.2996 0.0214
AR{1}(1,2) -0.003 0.0011 -2.7079 0.0067
AR{1}(2,2) 1.3372 0.0644 20.753 1.1475e-9
I know that the 1 in AR{1} means lag period. Could anyone please explain how to interpret the remaining elements of the result, like the term (1,1) in AR{1}(1,1)?

Answers (1)

Pratyush Roy
Pratyush Roy on 2 Dec 2021
Hi Eric,
A VAR equation with multiple response series can be expressed in the following manner:
So the term AR{1}(1,1) stands for the coefficient a11 which we want to estimate with the VAR model. If we had considered the lag factor to be higher than 1, then more such matrices would have come up and we could estimate them as well.
Hope this helps!


Find more on Conditional Mean Models in Help Center and File Exchange




Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!