Here is a perfect example of what I tried to explain to you above. First off, I'm not aware of a source to get Manu's stats when paired with Hill as opposed to when he's paired with Parker, but that's really irrelevant.
Let's assume that Manu's individual stats are better with Hill than with Parker. You have concluded that Hill makes Manu better. An equally plausible conclusion, and one well supported by watching the games, is that Manu dominates the offense more with Hill and thus increases his personal stats.
Once again, correlation is not causation. Do you even understand that basic concept?