Red light cameras in many US cities increased rear-ending collisions hugely and they've been taken out in some cities due to it. Points only would stop any claims of it being a revenue measure
There are several issues with the study, and they admit that there are issues, particularly in regards left on red, observation etc. They also talk about 8 months of data collection but in reality, they only capture one month of data after citations started, and that was within the first three months after they were installed. The one overwhelming thing that can be said is that placement of red light cameras reduces red light running. Accidents fluctuate in their type but seemingly not so much in their severity.