A bit of a correction there. VE accumulates at three times the rate of CE, with respect to damage done (as well as damage taken and damage cured). On level 120 mobs (T6 VWNMs) every point of damage done is 1 CE and 3 VE. So VE will cap long before CE will, despite decay.
Reaching VE cap is remarkably easy. A pld can cap it in 4 actions after getting on hate list: Sentinel, Provoke, Divine Emblem+Flash.
CE, on the other hand, is far more difficult. It's essentially impossible to build CE without dealing with damage in some way (damage done or damage cured), and is also affected by damage taken. As noted above, you get 1 CE per damage done. In additional, assuming ~1800 HP, you lose 1 CE for every point of damage taken. Therefore for a pld to cap CE, they have to do 10k more damage to the mob than the mob does to them to maintain capped CE.
Now, the damage done to the pld will be far less than the damage done to any other player (eg: when stacking defense the way SE apparently suggests), and that in itself reduces CE decay. The problem is that in putting up sufficient defense that way, you've removed pretty much everything offensive a pld can do. A pld all Defendered up, with Gallant's Roll and Minnes and Tacos or whatever else, is sacrificing opportunities to improve their damage output with things like Chaos Roll and Marches and various attack/accuracy foods. Therefore their damage output is miserably pathetic and they'll sit at pretty much 0 CE indefinitely.
As an alternative, they can cure themselves, but that's a fairly weak option. Put in terms of DPS (ie: damage done per second, which directly correlates to CE generated per second), it tops out around 100 DPS using the current cure formulas. However, even with generous MP recovery conditions (including using 300 TP Chivalry), that can't be maintained for more than about one minute, which is still only halfway to the CE cap.
A DD, on the other hand, is going offensive. Assuming the pld is sitting at 10k VE all the time, and given the 1:3 ratio of CE:VE, a DD matches the pld's total enmity when the DD hits 2500 CE/7500 VE. That amount of CE/VE is done with a mere 2500 damage. After that it's just a question of who's doing more damage: the DD or the mob. If the DD is doing more, their CE increases; if the mob is doing more, the player's CE decreases.
In general, any decent DD will reach 2500 damage within ~30 seconds. In cases where they have TP built up before the fight (eg: Voidwatch), they can hit 2500 damage in 10 seconds or less. So that pretty much defines the window in which the pld can claim to be the tank.
The second threshhold is when CE and VE are both capped. That essentially takes 10k damage minus whatever AOE damage the DD takes (aside from whoever is at the top of the hate list, who might take more). This is actually a slightly more difficult threshhold to hit. A VWNM might have ~190k HP. 10k HP is 5.3% of the NM's entire HP pool. Decent DDs will likely do at least ~10% of the mob's HP over the course of the fight, while extremely strong DDs might be upwards of 25%-30%. Typically, though, you could expect a DD to reach cap within a couple minutes.
Essentially, what's needed is some way for a heavily defensive pld to be at some sort of advantage in terms of generating CE. At present you have an either/or situation: either you're good offensively and can gain CE, or you're good defensively and can't generate CE, which completely undermines the entire point of having a tank.
People have clamored for native Provoke on pld for years, but that would never solve the problem. Pld doesn't have issues generating VE, which is all that Provoke gives. What it needs is sort of the complete opposite: a tool that can generate a significant amount of *CE* that can be used on a relatively frequent basis (30-60 seconds). I'd say that if it were constructed like Cannonball (which does damage based on your current defense), and generated CE based on your defense, things would be a lot more feasible.
Consider how much defense is feasible to get: 565 base+gear; Protect V w/ring +65; Defender +157; Gallant's +157; double Minne +67+61; Tacos +150 = 1223 defense. Even removing Gallant's + Minnes, that's 937.
I think an ability that could generate CE equal to twice the pld's current defense, or maybe equal to the pld's defense since it would also be affected by +enmity (and reaching double under the effect of Sentinel) would work. If it's equal to the pld's defense+enmity, probably have it on a 30 second timer. It would still take a few minutes to cap CE even under optimal conditions, but that's supplemented by any damage done plus cures.
Maybe have the shield being used be an additional scaling factor: an extra 10% CE per size of the shield (so +10% for bucklers, +30% for kite shields, +50% for Aegis/Ochain).
So some sort of "Gallant Charge" or whatever, with emphasis on the shield. Doesn't do damage or stun or anything; just generates CE. The stronger your defense, the more effective it is. I think it might actually be a workable solution, at least on the pld side...
Then if they just added some way for thf to dump some of the enmity they gain from Accomplice/Collaborator, the tools for a more controlled hate structure would mostly be in place.