I would assume it would have some benefit in reducing banding and artifacts in any area with large areas of flat color. A good test would be simply to shoot 10 seconds of a blue sky with a wide enough FOV so you go from light to dark.
Also guessing it might be beneficial when shooting green screen. has anyone done tests to confirm this?
If you're talking about CBR, the encoder will pick higher qualities if set to larger sizes. It can't really control the data rate per se.
The difference between Q16, 0, -16 is big but -10 vs -16 people disagree on. I think it makes a difference. 120mbps video vs 40mbps video.