Because the number of confounding factors if you look at whole countries is so huge that you couldn't attribute it to alcohol even if you found an association.
Well, that's begging the question; how do we know that the other differences don't relate to physical health if we must do comparative studies to know if one difference (alcohol consumption) does?