First, Provoke itself only generates 1 enmity if you use it while you're target's target, and only generates one enmity when using it for pulling.
However, if you are not the target's target, and someone else is, provoke matches you to the enmity level of the highest person in the party. If I'm stunbotting Ifrit (expression) and I provoke him, it matches me to the MT's enmity. If the MT just died, it matches me to whoever is next.
In many situations, following this with a Shield Lob or Tomahawk (ranged attacks for tanks) is enough to create enough enmity for the target to get to you and for you to establish.
However, high healing situations, such as picking up an add on Garuda HM, where medica and medica 2's and regens galore are producing a lot of enmity can be rough to pick up a mob from.
Is it necessary?
No, it's not necessary, few things actually are, but they would make important tasks a bit easier, and easier on one person = easier on the group.
Could it be "abused"? It could be manipulated for advantage early in the fight, OT provokes, MT provokes for a small threat lead, but really, it wouldn't be a great gain. Things could be made provoke-immune for even 5 seconds after being provoked and the ability to abuse this would fly out the window.
My suggestion? Make it equal to our defense. My defense is 732 without protect. Make Provoke generate 732 * (enmity modifier of Shield Oath/Defiance (if they're up))
So, if a tank used Provoke to pull, he'd get 732 * eMod on the target he provoked.
If a tank used Provoke to pull off another player, he'd get that persons hate + (732 * emod)
If a tank used Provoke on a target he was already tanking, he'd get 0 (or 1, whatever) hate.
A better geared tank would get more out of provoke than I would. A worse geared tank would get less, and that gives us an ability with that scales with our gear.