Why (I believe anyway) god is not necessary for an ethical system.
GOD IS GOOD:
Let's begin with the idea of goodness. When the theist claims that his god is necessary for the idea of goodness, just what does he mean? To say that god is good, is just uttering a truism. It's a substitution on par with the statement, "good is good." There are some problems with this definition of good.
1. It is circular. There is no information provided as to what "good" is.
2. There are unacceptable results which stem from this definition. Usually, since we do not have complete information about a given "Act of god", there is room the theist to claim that some sort of unknown higher purpose is served. Suppose however, this same deity made it known that it performed an act of cruelty, for *no* reason whatsoever. Given the current definition of "god is good", even behavior that is admittedly wanton, must be considered good. I think that it is clear that this is not what is meant when people use the term good.
GOD'S NATURE IS GOOD.
Another theistic response to the problem of defining good/evil is to claim that god's nature is good. This does not help at all, because it admits that god is *not* necessary to the notion of good; in fact, the god concept becomes completely unrelated to the definition of good. To say that god's nature is good is essentially a comparison between some pre-existing notion of good (independent of a god), and the nature of the proposed deity. This does not further the ethical absolutist's cause.
THE COMMANDS OF GOD ARE GOOD.
There are two forms of this argument: form A is where the theist asserts simply that god's commands are good. This has the exact same problems mentioned above (god is no longer necessary to the determination of good and evil, even if it's commands form a paradigm of good). Form B, is where the theist asserts that something is good because god commands it. There are some problems with this notion.
1. A deity that commanded something *without reason* (let's just say that it said, "I do this without reason") would still be issuing a command that is "good", even if it results in a command that is usually considered evil. This could be accomplished simply by god forbidding murder without cause, then later commanding a murder without cause (to simplify matters, we can assume that the deity let's us know that it has no cause for this murder). We then have 2 mutually exclusive commands which both must be considered good. A contradictory state of affairs that renders useless the term good.
2. If we grant that it is true that something is good, because god commands it, there are some peculiar ramifications. It is quite possible that what a deity commands ends up being the exact same thing as say, what a professional, atheistic, philosopher comes up with. In this case, we have 2 commands, one issued from god, one issued from man, both of which are good because they are the exact same thing. It should be obvious that the philosopher's notion of goodness is *not* dependent on a god. This is a case where the proposed definition does not uniquely exclude members of the set of notions we understand as good.
3. If we grant that it is true that something is good, because god commands it, then it must necessarily be true that something is good because a god would command it. In a world without a god, it would be possible to formulate ideas about goodness based on the notion of what a god would command, if it did exist. As in 2 above, we can arrive at a definition of good, wholly independent of god; it isn't even necessary that a god exists.
Other concerns.
Given the inherent ambiguities in communication, how does a god communicate the idea of good to mankind? If at any point, the commands issued are unclear, then the problem is not settled at all; the notion of good once again becomes dependent on man's understanding, not on the commands of god. In fact, I think this is the underlying problem to the whole idea that a god is necessary for the definition of good. "Good" and "Evil" are concepts that change, among humans, over time. Without people, it is pointless to argue that there is some sort of "eternal" or "Absolute" ethical standard; there is no way to test such a claim. Since people are ultimately, the final referrent in any system of morality, the inherent ambiguities in human communication result in a necessary subjectivity in all moral structures. This is evidenced in that even among inerrantists, there are disagreements about the notion of sin, precepts of duty, etc.