Dear all,
I am dealing with a SVAR model and I have two models: baseline and extended model. When I run the baseline model, I receive relatively satisfying results, both theoretically and intuitively. However, when I run my extended model (just including more variables), I receive pretty weird results. First, the estimated elements of the lower diagonal in the A matrix are nearly 1. Depending on this (I guess), when I drew the IRF of variables with their confidence intervals, I realized that the values of IRF, lower and upper are the same! Is this a little strange? Let me show you my estimation process.
Matrix A:
When I create the IRF graphs, the confidence intervals do not appear. At first, I thought it might be related to the charting code. However, when I looked at the results, I noticed that the confidence intervals are not visible since the lower and upper values are the same as the IRF values.
Estimation results:
Response of Inflation to PIMP shock:
What could be the reason for this? Thank you.
I am dealing with a SVAR model and I have two models: baseline and extended model. When I run the baseline model, I receive relatively satisfying results, both theoretically and intuitively. However, when I run my extended model (just including more variables), I receive pretty weird results. First, the estimated elements of the lower diagonal in the A matrix are nearly 1. Depending on this (I guess), when I drew the IRF of variables with their confidence intervals, I realized that the values of IRF, lower and upper are the same! Is this a little strange? Let me show you my estimation process.
Code:
matrix A2 = (1,0,0,0,0,0 \ .,1,0,0,0,0 \ .,.,1,0,0,0 \ .,.,.,1,0,0 \ .,.,.,.,1,0 \ .,.,.,.,.,1 ) // Causal Ordering: PIMP, HICP, r, wage, markup, PS svar d_ln_pimp ln_inf ln_r d_ln_wage ln_markup ln_PS, lags(1/8) aeq(A2) //where ln=log, d: first difference matlist e(A) irf create model2, set(var1.irf) replace step(16)
When I create the IRF graphs, the confidence intervals do not appear. At first, I thought it might be related to the charting code. However, when I looked at the results, I noticed that the confidence intervals are not visible since the lower and upper values are the same as the IRF values.
Estimation results:
Response of Inflation to PIMP shock:
What could be the reason for this? Thank you.
