Richard Bartle: "[without permanent death] Newbies (and not-so-newbies) feel they can never catch up. The people in front will always be in front, and there's no way to overtake them. The horizon advances at the speed you approach it." [2]
[1] http://en.wikipedia.org/wiki/Permanent_death [2] Designing Virtual Worlds, 2003
There are other ways to allow newbies to catch up and to incentivize or disincentivize players that don't rely on 'ruining your month of gameplay'.
Games like Realm of the Mad God, DayZ, and Minecraft (its survival based multiplayer) make this perma-death a part of the game (though Minecraft, you keep the character). You feel defeated, but not dissuaded towards playing.
The way to keep newbies is to make the gear they can earn secondary to the gameplay.
In Monopoly, I can go bankrupt. In Nethack, I can (and often do) die. In Super Mario Bros 1, I can lose all of my lives. In football, I can get knocked out of the Champions League (by a red card).
In all of these cases, the fact that I can "lose" creates meaning - I care more about the outcome and I appreciate my achievements more. And one never loses everything - the skills stay with you. "Next time I can do better."
Counterbalancing, for a different mood: works like Ocarina of Time are still great, but they sit separately - they're about the journey, not about the achievements.
The stakes are a little different in a long-term, persistent world MMO. The Mario you died with is, for the most part, the same Mario you started with. This is not the case in an MMO. Punishing you, irrevocably, for dying in that situation is a dreadful, fun-destroying design choice.
Even in platformers, it's probably telling that one of the most successful recent ones is Jon Blow's Braid where you not only don't die but you get to reverse time.
PvP corpse looting works exceptionally well in Eve Online. You don't really have permanent death, there, though.
And if a month of your gameplay gets ruined by getting instacane blapped in Eve, you're doing it wrong. Never undock with what you can't afford to replace.
If the entry point is too high (say several weeks of grinding to get even a mid-level player) people will be less likely to venture out and play the game. The reward is very low for the risk. If the entry point is too low then it'll just be a FPS. Striking that balance is really tough.
In my opinion, a permanent death MMO it should take about 20 hours of grinding to reach the top 2%, 60 hours of grind to reach the top 1%, and 400 hours to reach the top %0.01 (and so on).
You can self replicate perma death for yourself in any game you play but I know most people would not, they would rather be water cooler hard cores.
Having recently played XCOM which has a perma death for your troopers I lost my taste real quick because of bugs. Players certainly won't pay for a perma death MMO where they died because of bugs or cheating; there will be cheating. A free to play perma death MMO might survive but even then its a fine line before the bugs push them over the edge.
I really don't want to be in another MMO like Eve, where areas of the game are off limits or worse, certain groups go on nerd rages and just grief people who are no threat/not even interesting but convenient.
Dude, if no one likes you, that isn't their fault.
There's nothing like invading a big alliance space, get chased by multiple pilots, and by some skilled (and lucky!) piloting managing to escape alive. All the while knowing that one screwup would cost the whole ship.
UO had a skill cap, which you could reach rather quickly (less than a month if you had no life, a few months playing casually). After that, there was no difference.
And you could even close the gap faster, by focusing on skills which made the most difference. You alchemy could wait.
Also, there was a stat cap. There wasn't that much difference between a new player and a veteran. The veteran would have like 2x the amount of health compared to the newbie. This is in constrast to WOW, where it is not uncommon to have a 140x difference, or more. You actually had a decent shot against veterans, if you knew what you were doing.
That's all my way of saying Permanent death may sound great in theory, but it takes a hell of a lot of balancing against to make it work well in the long term. Maybe making it easier to get up to speed with any single character? Or maybe making it worth your while to be altruistic to other players (I'm not sure what that mechanic would be, but it would be interesting to see something that make people more social as opposed to the antisocial element that PvP tends to foster).
Never add in features willy nilly because it sounds good. Figure out why they work well with your design.
Playing through Diablo III on all but the hardest difficulties, I must have died about 100 times due to extended lag spikes and only a few times due to my own mistakes.