Since collinearity is a problem in the predictor variables (you don't need to know what the dependent variable is, mostly), a work around would be to use your second MODEL statement in PROC REG, with the VIF and COLLIN options in the MODEL statement.
SteveDenham
... View more
