How to regress VAR model with few parameters to estimate?


In the VAR process

y_t = c + A y_{t-1} + epsilon

A is made up of actually almost all 0s.
You see, models like unemployment is only related to the CPI and not others; while CPI is only related to CASH M2 and not others.

Is there a way that I can specify some of the elements in A to be always zero?