Accumulated Impulse response
Posted: Tue Jul 15, 2014 3:44 am
Good Morning
I would appreciate to have some help on the following issues:
1) I'm esitmating a VAR with 3 varialbes, one in growth rates (Dlogx) and 2 in log levels (log y and log z). I want the IRF of the 3 variables in Levels to a one sd shock to logx (ie Dlogx=epsilon) . I'm computing the accumulated response of the first variable to get the response of log x to a sd shock to logx. IS this correct. Is there a way to compute the IRF for all variables in levels without calculating the IRF twice (accumulated to get the response of log x and not accumulated to get the response of logy and log z) .
I'm doing this in a loop estimating the same IRFs for 30 countries.
2) when I save the IRF of the three variables to a one sd shock to logx with rmat, I get hte IRF of log x , a columns of 0's and the IRF of log z . Why I don't get hte response of logy? I'm using the command
var.impulse(40,m, rmat=xshock_{%y}) Dlogx_{%y} logy_{%y} logz_{%y} @ Dlogx_{%y} @ Dlogx_{%y} logy_{%y} logz_{%y}
3) can I combine the IRFs for the 30 countries in one signle matrix
Thank you in advance for your help
Alfonso
I would appreciate to have some help on the following issues:
1) I'm esitmating a VAR with 3 varialbes, one in growth rates (Dlogx) and 2 in log levels (log y and log z). I want the IRF of the 3 variables in Levels to a one sd shock to logx (ie Dlogx=epsilon) . I'm computing the accumulated response of the first variable to get the response of log x to a sd shock to logx. IS this correct. Is there a way to compute the IRF for all variables in levels without calculating the IRF twice (accumulated to get the response of log x and not accumulated to get the response of logy and log z) .
I'm doing this in a loop estimating the same IRFs for 30 countries.
2) when I save the IRF of the three variables to a one sd shock to logx with rmat, I get hte IRF of log x , a columns of 0's and the IRF of log z . Why I don't get hte response of logy? I'm using the command
var.impulse(40,m, rmat=xshock_{%y}) Dlogx_{%y} logy_{%y} logz_{%y} @ Dlogx_{%y} @ Dlogx_{%y} logy_{%y} logz_{%y}
3) can I combine the IRFs for the 30 countries in one signle matrix
Thank you in advance for your help
Alfonso