We concluded PvE Holy paladins were ending fights with too much mana relative to other healers, so we raised the mana cost of their single-target heals.
This is a silly metric, even if it's true (which again, I very much doubt). Tank healing requires less healing than raid healing, and the burden of preventing damage is less distributed. You should balance around what classes can do and are doing with their mana, not how much they have left over. I suspect whatever paladins you were looking it used LoH late in the fight, and the 20k extra mana provided by a late LoH (which has been nerfed itself, anyway) shows up at the end of a fight more than the 20k extra mana druids and Disc priests start with. This is just an absurd way to balance.
In any case, if your previously cited "40% mana" is actually what you were seeing and meant what you thought it meant, then healers being within 40k mana expenditure of each other would be the most balanced spell costs and regen have ever been. I feel like you guys don't understand the full range and interactions of everyone's differing regen, cost reduction and efficiency mechanics, just as you didn't during Ulduar.
You've already raised and then lowered paladin spell costs in turn this tier. What is even going on? Are you going to do this again before the next tier, as paladins are casting more Divine Light and everybody is is still spamming CH/WG/Rejuv/PoH/CoH?
We tweaked Holy Light because we wanted it to compete a little more with Divine Light, especially when used with Beacon of Light.
Is the intention for Holy Light to be some sort of ghetto raid heal with this change? This doesn't change spell selection, anyway: if there is enough healing to warrant a DL, paladins will use DL. All this does is set up another arcane spell/target selection rule for paladins (and paladins already have plenty of these): never cast HL on your Beacon, ever.
We buffed Word of Glory for three reasons: We felt Holy Power was mattering less to Holy paladins than it did at Cataclysm launch. We wanted to provide more uninterruptible healing in PvP. We knew Light of Dawn was trumping Word of Glory in almost all cases in 25-player raids.
HoPo matters less to Holy paladins because Holy Shock heals for so little, not because WoG healed for so little. Light of Dawn continues to trump WoG in almost all cases in 25-player raids. For a paladin with my gear, in 4.2, glyphed WoG generates 45,007 healing per cast if used off-Beacon. Glyphed LoD generates 83,608 healing per cast if used off-Beacon. 10-man paladins don't glyph LoD and 25-man paladins don't glyph WoG, so the difference is actually even greater. Light of Dawn trumps WoG in 25-mans because you can actually use LoD in 25-mans. Holy Power is essentially twice as worthwhile to 25-man paladins as it is to 10-man paladins.
We buffed Holy Radiance both to help PvE paladins feel like they could make larger contributions to raid healing (especially in light of the single-target nerfs) but also as part of a significant buff to Speed of Light to let Holy paladins have more mobility in PvP.
PVE paladins don't want to feel like they can make larger contributions to raid healing, they want to actually make larger contributions to raid healing. This means improved targeted raid healing through an improved LoD. Furthermore, it's very cute that you call the HL/DL/FoL cost increases "single-target nerfs." What do you think paladins raid heal with when they've already pressed Holy Radiance?
We made a tweak to Holy’s mastery to allow its bubble to stack, so it would be wasted less often when healing a single target.
Seven months after we told you it was necessary. Progress, sure, but mastery is still worth much less to paladins than it is to all other healing specs.
We changed all heals to have 200% crits because we wanted to make crit a more attractive stat to healers. Anything random is already at a disadvantage when viewed by a healer, and one point of haste just flat out increased throughput more than a point of crit.
It still does. Haste continues to be significantly more throughput per point than both crit and mastery, which is almost uniquely a paladin problem.