07-10-2012 02:52 PM
I'm having trouble creating a conditional correlation matrix. I think I need to use a GARCH model in order to create a conditional variance matrix, but I'm not sure.
I'm trying to check the correlations between riskfactors and check the correlation between returns in two different time periods, so in total I need 4 matrices. I can't run the GARCH model in SAS becasue there are too many variables and an error message pops in the log window stating the "model is not full rank". Also I'm not trying estimate a dependent variable using the GARCH model; I'm strictly trying to figure out the correlations.
Is the GARCH model the process I should take if I have no dependnet variable and want to know the conditional correlation or is there a different procedure I should use.
Note: I wanted to use the GARCH model in the first place because there was autocorrelation in the variables.
Thank you for your help,