I made a mistake today.
I answered a question incorrectly. (A C# question, no less. My head is truly hanging in shame.) Unfortunately, by the time this was pointed out, the answer had already been accepted.
Another answer has been posted which is correct. It has more votes than mine, and mine has some downvotes (still a balance of upvotes, but...) Unfortunately because the answer has been accepted, I can't delete my answer. I've edited it so that it's correct - it won't mislead anyone any more - but I'd far rather delete it to let the original correct one float to the top.
In this case I suspect it won't be a real problem - the question author is reasonably regular, so hopefully he'll transfer acceptance to another answer reasonably soon. However, I can easily foresee situations where this wouldn't happen.
Could I suggest that the author of the accepted answer should be able to delete it, if the answer has received some threshold of downvotes (e.g. 3)? Deleting the answer straight away would have been preferable to editing it to be correct, effectively duplicating the effort already put in by Luis Filipe in this case.
I understand the "we don't want to lose value" argument for not letting a question author delete their own question after it's had good answers, but if an answer has received downvotes (so at least some of the community think it's harmful) and the answer's author thinks it should be removed, I think that should override the questioner's choice.