multiple cross section
Posted: Tue Jul 02, 2013 2:49 pm
Hi,
my analysis revolves around the estimation of the influences of both bond variables and cds variables on the mutual relationship, the cds bond basis.
The problem I am having is that I have a stacked cross section of the CDS relating to about 120 companies and timeseries of about 300 unique bonds (another stack). So 1 company can have multiple bond issues. Eviews gives me the opportunity to either identify a cross section for bonds or for CDS and this creates a mismatching problem. My teacher gave me as much as a hint: use fixed effects. I'm not sure how to do this. Sprecify 120 variables in the common fixed effects section?
As far as I can see these are my options:
- match each CDS variable to each bond, which would create duplicate measures for the CDS on occasion but not sure if eviews would then calculate them as a unique match within the cross section or generate a bias
- somehow merge the bonds. So I have a bond identifier being ISINID (example: AET16, AET17) and a cds identifier being ticker (AET). Both are included as fields in the file. I don't know how to do this however and a few copy paste special attempts have failed.
I've also tried to define groups within the separate stacked series but not successful.
Would greatly appreciate advice!
my analysis revolves around the estimation of the influences of both bond variables and cds variables on the mutual relationship, the cds bond basis.
The problem I am having is that I have a stacked cross section of the CDS relating to about 120 companies and timeseries of about 300 unique bonds (another stack). So 1 company can have multiple bond issues. Eviews gives me the opportunity to either identify a cross section for bonds or for CDS and this creates a mismatching problem. My teacher gave me as much as a hint: use fixed effects. I'm not sure how to do this. Sprecify 120 variables in the common fixed effects section?
As far as I can see these are my options:
- match each CDS variable to each bond, which would create duplicate measures for the CDS on occasion but not sure if eviews would then calculate them as a unique match within the cross section or generate a bias
- somehow merge the bonds. So I have a bond identifier being ISINID (example: AET16, AET17) and a cds identifier being ticker (AET). Both are included as fields in the file. I don't know how to do this however and a few copy paste special attempts have failed.
I've also tried to define groups within the separate stacked series but not successful.
Would greatly appreciate advice!