Much of the discussion of Blizzard's controversial decision to allow player-to-player item sales for real money in Diablo III has focused on the tired "pay to win" argument. Like Ferrel, Keen, Rohan, and others, I'm not so interested in this side of the equation because how other players obtain their gear does not affect me directly. I'm more concerned about the indirect effects of putting the developer in the business of making money when players replace their gear.
Runes of Magic Gear Upgrades
The best example from my current experience is Runes of Magic and its cash store gear upgrades. While both models do allow players to get what they want in-game with some (possibly prohibitive) amount of time investment, I would argue that the presence of gear boosts in the cash shop means that Runewaker/Frogster see more revenue specifically when new gear is added to the game, much as DIII's auction fees will almost certainly provide more revenue for Blizzard when new gear is added to the game. Any MMO expects to make money off of new content somehow (box sales, renewed subscriptions, item shops), but this approach biases the payment model towards additions that replace players' gear.
If you look at ROM's content history, you'd see what my model predicts - compared to traditional subscription games, the ROM level cap has increased far more frequently, but with fewer additional levels at each increase, allowing players to park and farm up new gear at each new level range before the cap rises again. There are benefits to this model - players get content more frequently, and all of the new content gets used while it is the current level cap, where games that raise the cap by 10 levels at a time often end up with underutilized content in the middle of that level range. However, you also see insane vertical progression, with all the problems that this entails - my level 53 Druid has fewer than 3,000 HP in solo gear, while the Druid Encyclopedia recommends 5K-15K for dungeons in my level range.
Moreover, you will also see dramatic swings in relative class balance. The recently concluded Chapter 3 era saw Scouts topping the DPS charts by as much as 3-fold, while the talk in Chapter 4 is that Rogues are the new scouts, just as the game added a third class slot to let those disgruntled Scouts take up Rogue-ing. Every game has class balance issues and flavor of the month builds. Not every game makes more money when players replace their gear, and nothing makes ROM players replace their gear faster than changing archetypes outright.
The Post DIII Era
Bobby Kotick's comments about exploiting his franchises aside, I don't expect that Blizzard would take this in a direction that immediately ruins a flagship product. You can't sell virtual goods to people who aren't playing your game, and a game that's balanced assuming that players are fully decked out in the rarest of the low drop rate gear from the cash auction house will drive players away. We can probably expect extremely rare items powerful enough to affect game balance, especially on higher difficulty levels, but that was a feature of the original game anyway.
If anything, I think a fair number of would-be-Gevlons will waste a lot of time and real world money on posting fees - and, gods help them, attempts to buy low and sell high - as demand to get real cash for DIII loot outstrips willingness to pay real cash for all but the rarest of DIII loot. That said, I think the longterm effects of having publishers demand the revenue from this feature will only worsen the vertical progression issues that are already causing huge problems in MMO's.
Also, as Stabs points out, the legal implications are potentially staggering. Can you sue a ninja looter? How long before someone decides to test the EULA in court by suing Blizzard for a nerf or server crash that cost them real money? Can Blizzard flag accounts to ensure that the rare drops go to people with a proven history of selling them, rather than keeping the loot and depriving Blizzard of their auction fees on an item that might not drop again for weeks? Would they have to disclose it if they do so? Is this system effectively online poker with the killing of monsters substituted for the random drawing of cards?
Though I don't think that pay to win is the direct problem, I think Scott Jennings is right when he says that this will take games in a direction that is bad for players, but that we've already lost the battle against it because the market will tolerate the new model. It's not the apocalypse, but it's also not a good day for online gaming.