For 31 years, I've been editing the Journal of Economic Perspectives. At the most basic level, editing is about pushing the author to have a point in the first place, and to make it clearly. Sounds simple, perhaps? On complex subjects, meeting those criteria can be a high hurdle to cross. 

Albert Jay Nock, in his 1943 Memoirs of a Superfluous Man, tells a story along these lines in his description of the editorial policy at a magazine he had edited called The Freeman (p. 172):
"In one way, our editorial policy was extremely easy-going, and in another way it was unbending as a ramrod. I can explain this best by an anecdote. One day Miss X steered in a charming young man who wanted to write for us. I took a liking to him at once, and kept him chatting for quite a while. When we came down to business, he diffidently asked what our policy was, and did we have any untouchable sacred cows. I said we certainly had, we had three of them, as untouchable and sacred as the Ark of the Covenant. He looked a bit flustered and asked what they were. 
"The first one," I said, "is that you must have a point. Second, you must make it out. The third one is that you must make it out in eighteen-carat, impeccable, idiomatic English."
"But is that all?" 
"Isn't it enough for you?" 
'Why, yes, I suppose so, but I mean, is that all the editorial policy you have?" 
"As far as I know, it is," I said, rising. "Now you run along home and write us a nice piece on the irremissibility of postbaptismal sin, and if you can put it over those three jumps, you will see it in print. Or if you would rather do something on a national policy of strangling all the girl-babies at birth, you might do that—glad to have it." 
The young man grinned and shook hands warmly. We got splendid work out of him. As a matter of fact, at one time or another we printed quite a bit of stuff that none of us believed in, but it all conformed to our three conditions, it was respectable and worth consideration. Ours was old-school editing, no doubt, but in my poor judgement it made a far better paper than more stringent methods have produced in my time.
I especially like the comment in the closing paragraph about how they "printed quite a bit of stuff that none of us believed in."  For editors, agreeing with authors is overrated and in fact unnecessary.