can't see the ratings though
(not that i don't rationalize looking)
it definitely has an effect or it wouldn't be there... how much? no idea. i would assume it's a small-ish effect. development is quite volatile to know by the eye. i've had stretches where nearly everone develops well with max budget, then a few lemons.
if you turned off some of the competing and similar effects it would reduce sample size needed to know. e.g. turn off coaching and personality.
easy to assume this: ~baseline is a neutral effect... makes sense, may or may not be correct, of course. as far as going over i'd assume a slow reduction of ROI to the point of being negligible (limit of max % bonus)
i'd assume and equal negative effect possible, but the limit occurs above $0, i hope. just to be rational relative to context. not rational that a team spens "0" on dev. could be a reason for a much larger negative effect, too, if not equal... spend zero - so no pitching machines, no balls etc, lol