As m.k. says in the comments, in order for the axioms to even make sense (as they are usually presented) it is necessary to state identity before inverses. Other than that the axioms may be stated in any order one wishes. (As William says, it is possible to write down the identity $e$ as a distinguished symbol instead of to merely posit the existence of the identity as an axiom, and in that case it is not even necessary to do this.)
However, I think the order you've stated them in is a good one because, as Hurkyl says, taking an initial segment of the axioms leads to other theories of independent interest:
- Taking only associativity gives you a semigroup.
- Taking associativity and identity gives you a monoid.
Just as groups abstract collections of symmetries of a set, semigroups and monoids abstract collections of functions from a set to itself (not necessarily invertible), and for that reason they are also natural and important to study. However, it is often harder to say anything useful about them. For example, finite groups have some nice number-theoretic structure to them because of results like Lagrange's theorem, and Lagrange's theorem fails very badly for finite monoids.