This paper has two goals. First, I argue that contrary to what has been assumed by many previous works, the two characteristics of long-distance reflexivization, long-distance binding and subject-binding, should not be treated as two inseparable consequences of a single LF operation. While long-distance binding indeed results from LF movement, subject-binding must be analyzed separately. Secondly, I show that it is not necessary to make the unjustified assumption that bare reflexives always involve X0-movement while compound reflexives move only as phrases. Given Huang's observation that there is no Subjacency effect at LF, the contrast between long-distance binding and local binding can be accounted for by uniformly moving N0 (=head of a reflexive) at LF regardless of the actual form of the reflexive. The only constraints are the independently motivated ones such as Lasnick and Saito's ECP and Li's theory of X0-binding.