How to compute confidence intervals for scaled impulses
Posted: Sat Dec 23, 2017 3:43 pm
Hi Tom,
I am working on a FAVAR model. I am studying the effect of exchange rate shocks on US economy using 2 step PCA, and identification based on recursive ordering of variables.
If you see my code attached, I have first grouped variables into similar variables. There are 5 groups –economic activity, prices, interest rate, money supply, and exchange rate. Economic activity has 95 variables; for prices there 23 variables, finally for money supply there are 7 variables. For each of these groups I extract the first principal component.
In step 2, I estimate the VAR using recursive ordering – economic activity (factor used), prices (factor used), federal funds rate (observable), money supply (factor used), and exchange rate (observable). After estimation, I accumulate the impulses and then I use the factor loadings to compute the impulse responses for all the variables at the disaggregate level. (I am basically scaling them by their factor loadings)
I could compute the confidence intervals for the main 5 variables that enter my VAR using @MCVARDODRAWS and @MCGraphIRF. My question is that how can I compute the confidence intervals for all the variable at the disaggregate level. I have done a scaling there so I am not able to figure out how to have confidence intervals for the scaled responses too.
Could you please help me out. I am not sure if my question is clear. Basically how do I calculate confidence intervals for the responses in a(), c(), and d()
I would be extremely grateful for your help.
Many Thanks
I am working on a FAVAR model. I am studying the effect of exchange rate shocks on US economy using 2 step PCA, and identification based on recursive ordering of variables.
If you see my code attached, I have first grouped variables into similar variables. There are 5 groups –economic activity, prices, interest rate, money supply, and exchange rate. Economic activity has 95 variables; for prices there 23 variables, finally for money supply there are 7 variables. For each of these groups I extract the first principal component.
In step 2, I estimate the VAR using recursive ordering – economic activity (factor used), prices (factor used), federal funds rate (observable), money supply (factor used), and exchange rate (observable). After estimation, I accumulate the impulses and then I use the factor loadings to compute the impulse responses for all the variables at the disaggregate level. (I am basically scaling them by their factor loadings)
I could compute the confidence intervals for the main 5 variables that enter my VAR using @MCVARDODRAWS and @MCGraphIRF. My question is that how can I compute the confidence intervals for all the variable at the disaggregate level. I have done a scaling there so I am not able to figure out how to have confidence intervals for the scaled responses too.
Could you please help me out. I am not sure if my question is clear. Basically how do I calculate confidence intervals for the responses in a(), c(), and d()
I would be extremely grateful for your help.
Many Thanks