Hi Statalists,
I'm estimating panel VAR with -xtvar- (a user-contributed code by Cagala(2014)?). I'm trying to generate the cumulative IRF chart but the default chart is IRF only. Just wondering how can I generate the cumulated IRF chart with 95%CI? I can see the IR figures at each period can be saved, but a bit unsure how should I work out the 95%CI for the cumulative ones.
Alternatively I had tried to replicate the process with -pvar-, but I understand the two codes are based on slightly different methods. And as my dataset is fairly small (N15xT10), functions of -pvar- does not work properly so I didn't manage to replicate. Further, the help note said the option -pvarirf, cum- may be combined with option -oirf-
My code with -xtvar- is
The embedded functions will generate the IRF and FEDV with default 8 steps
My code with -pvar- is
Any help and suggestion are appreciated!
I'm estimating panel VAR with -xtvar- (a user-contributed code by Cagala(2014)?). I'm trying to generate the cumulative IRF chart but the default chart is IRF only. Just wondering how can I generate the cumulated IRF chart with 95%CI? I can see the IR figures at each period can be saved, but a bit unsure how should I work out the 95%CI for the cumulative ones.
Alternatively I had tried to replicate the process with -pvar-, but I understand the two codes are based on slightly different methods. And as my dataset is fairly small (N15xT10), functions of -pvar- does not work properly so I didn't manage to replicate. Further, the help note said the option -pvarirf, cum- may be combined with option -oirf-
My code with -xtvar- is
Code:
xtvar y1 y2 x1-x4, mc lag(2)
My code with -pvar- is
Code:
pvar y1 y2 x1-x4, lags(2) pvarirf, oirf mc(200) imp(x4) res(y1 y2) cum byoption(yrescale)
