Block exogeneity in Panel BVAR

[oscarfergue98]

I am trying to estimate a panel BVAR model in which I have 5 unit-specific variables, and an oil price shock (which is the same for all units). I would like to estimate the effect of this oil price shock on all units by imposing block exogeneity, given that the variables of all units should not affect the oil price. So far, I am using the hierarchical prior scheme as in Jarocinski (2010). Unfortunately, the oil price shock seems to be affected by the variables of the units once I plot the IRFs. I would very much appreciate it if someone could help me with this.