Hello dears,
this is my first post and I hope to fulfill all requirements, such that you can understand my question.
I am currently using a structural VAR (= restrictions in a VAR equation). (Similar to "The Effect of Monetary Policy on Commodity Prices: Disentangling the Evidence for Individual Prices - from Carbales, Castro and Joya)
This restrictions for cointegration are in a 5x5 matrix.
Usually the following is what I did:
Model:
Ayt=C1yt−1+⋯+C3yt-3+et
C=AAj
et =Aut
matrix A = (1,.,0,0,.\.,1,.,.,0\0,0,1,.,0\0,0,0,1,0\.,.,.,.,1 )
matrix B =(.,0,0,0,0\0,.,0,0,0\0,0,.,0,0\0,0,0,.,0\0,0,0,0, .)
matrix D = (1,0,0,0,0\.,1,0,0,0\.,.,1,0,0\.,.,.,1,0\.,.,.,.,1 )
Matrix D is a simple lower triangular matrix
As the order of variables are important for cholesky decomposition there is according to SIMS 1986 also a general version, where I can ignore the order, this is important as I want to have contemporanous effects according to A:
The order is:
Federal Fund Rate
M2(money supply)
Consumer price index
GDP
Commodityprice
. svar d1usgdpad d1cpiusad d1fftrna d1m2usd d1aluminium, aeq(A or D) beq(B) lags(1/3)
2 very crucial questions:
1. I can run a SVAR model but only with the given order like matrix D. How can I run a model with the given order stated in the table? (When I use it in Stata it gives me that the rank of matrix A is 49 but 50 are needed, so i receive an error.
2. How can I decide what structure matrix C has? I mean A-1Cj = to what? to the beq(B)? How do I have to define beq? Can somebody give me an explanation for that? or should beq= I?
Thank you so much for your help in advance!!
Best regards,
Christian
this is my first post and I hope to fulfill all requirements, such that you can understand my question.
I am currently using a structural VAR (= restrictions in a VAR equation). (Similar to "The Effect of Monetary Policy on Commodity Prices: Disentangling the Evidence for Individual Prices - from Carbales, Castro and Joya)
This restrictions for cointegration are in a 5x5 matrix.
Usually the following is what I did:
Model:
Ayt=C1yt−1+⋯+C3yt-3+et
C=AAj
et =Aut
matrix A = (1,.,0,0,.\.,1,.,.,0\0,0,1,.,0\0,0,0,1,0\.,.,.,.,1 )
matrix B =(.,0,0,0,0\0,.,0,0,0\0,0,.,0,0\0,0,0,.,0\0,0,0,0, .)
matrix D = (1,0,0,0,0\.,1,0,0,0\.,.,1,0,0\.,.,.,1,0\.,.,.,.,1 )
Matrix D is a simple lower triangular matrix
As the order of variables are important for cholesky decomposition there is according to SIMS 1986 also a general version, where I can ignore the order, this is important as I want to have contemporanous effects according to A:
| 1 | . | 0 | 0 | . |
| . | 1 | . | . | 0 |
| 0 | 0 | 1 | . | 0 |
| 0 | 0 | 0 | 1 | 0 |
| . | . | . | . | 1 |
Federal Fund Rate
M2(money supply)
Consumer price index
GDP
Commodityprice
. svar d1usgdpad d1cpiusad d1fftrna d1m2usd d1aluminium, aeq(A or D) beq(B) lags(1/3)
2 very crucial questions:
1. I can run a SVAR model but only with the given order like matrix D. How can I run a model with the given order stated in the table? (When I use it in Stata it gives me that the rank of matrix A is 49 but 50 are needed, so i receive an error.
2. How can I decide what structure matrix C has? I mean A-1Cj = to what? to the beq(B)? How do I have to define beq? Can somebody give me an explanation for that? or should beq= I?
Thank you so much for your help in advance!!
Best regards,
Christian
