[niekmaas]

For most journals I have experience with as a (co-)author a signed form is required for every (co-)author on the paper. This way it's pretty hard to deny you have knowledge of the paper.

I agree with you that people that have not contributed should not be on a paper. For most high impact journals thesedays explanation is required of who did what. Of course this is far from 100% secure but it's an interesting step. Lying about the contribution of a certain co-author is something else than just adding a name to the list. Lying is fraud whereas adding a name might not actually be fraud I think.