That's not really bad math, as I would define it. That has more to do with inconsistency, which is or isn't bad on a case-by-case basis. I think one of the really nice things 3e did was provide more consistency in stat bonuses, but taking inconsistent things and forcing them into a systematic structure doesn't always work that well. Clerical spheres and putting druid spells into that framework are a prime example of that. All of that incorporated an ambitious idea of providing an overall framework. But the first attempt at it needed a lot of revision to make the druid work as well as it did in 1e. 3e was, I believe, right to back off of the overall structure and give druids back their own, unique spell list.
I think bad math would better characterize structures where the game seems to intend one thing, but delivers another because the numbers don't match at a certain point. For example, the saving throws in 3e sometimes get accused of involving bad math because they really don't keep up with the challenges faced in the game, particularly when the save DCs from the monsters tends to be based on their hit dice, which grow faster than the CR (and thus faster than the PCs' saves), and because large monster stats and optimized caster stats rise faster than defensive stats. By comparison, 2e's saves just become stronger as the character becomes more experienced... a result more people seem to favor.