I can't help but wonder why nobody's looking at this from a logistics point of view.
1) When the cycle of creating a game starts, there are a lot of people with grand ideas for gameplay concepts and all the hows and whats and whatnots, who all pile their ideas into a great big pile.
2) Next the more level headed people takes that big pile of ideas and start the weeding process where the most outrageous or disjointed ideas are tossed aside and only the seemingly possible ideas and concepts are left. The flow of the game is being determined and at the end of the process, the game goes into Alpha state or "feature lock".
3) Next the level designers and engine coders work on implementing the game and along the way a bit of content might change (certain areas may be cut, expanded or otherwise changed for various reasons though most commonly to keep a certain pace to the flow of the game). When the game reaches Beta stage, the game is essentially finished and the entire focus is on QA and bug fixing.
4) And finally the game goes gold and the code is handed over to the printing press.
Now there are of course numerous sub levels of each of the 4 major points listed above but that is irrelevant. The point is that each time a milestone is reached and the development moves into a new phase a lot of people will be done and ready to move on the next project (the content creators are seldom the same people who does the engine coding). A high profile game like Dragon Age reaches the Alpha stage years before launch and all that time, if given the opportunity, the content creators can create tons of new content, especially now that the limits of what can and what can not be done in this particular game are a lot better defined than it was initially.
When the game reaches Beta and the level designers stop creating new levels/areas, if given the opportunity, can begin implementing the new content, especially now that tons of meshes, objects and models have already been created for the game.
And the list goes on. Instead of reassigning everyone to different projects the minute their particular task on a particular project is over, a portion of them could stay on the project and try to find a way to get that great idea that was cut at the last moment to everyone's dismay into the game after all. Or come up with new ways to use a particular gameplay element that turned out better than expected. Or try something completely new.
This whole notion that developers are trying to sell you an empty wardrobe only to have you pay for the clothes separately afterwards is a bit paranoid (at least at this point in time it is). Yes, the suspiciously locked door in the middle of the Mage Guild in Oblivion that turned out to be the entrance to the Orrery DLC did indeed reek of being left out of the main game deliberately but the addition of the Orrery DLC had no impact on the game whatsoever. In fact the one specially added feature that has annoyed me the most was the pre-order bonus merchant in Baldur's Gate 2 that had the best sling in the game.
Other than the bonus merchant (which was not a DLC), the only issue I have had with DLC's is that a few have felt a bit overpriced, but such is the nature of any consumer product that somehow didn't live up to your expectations. I could just have refrained from buying it.