That sounds exactly like an urban myth along the same lines as 'seatbelts kill more people than they save'. Has it ever been proved?

Worth also noting that rear-end crashes are generally less dangerous than t-bones. For similar reasons, roundabouts can be preferable to standard crossroads even if they result in a higher number of collisions.

"That sounds exactly like an urban myth along the same lines as 'seatbelts kill more people than they save'. Has it ever been proved?"

I hope you will look further into this - you may be interested to learn (I was) that a nearly perfect intervention for red lights being "run" is to lengthen the yellow signal.

It costs nothing and works as well, or better, than red light cameras - and with none of the artificial, unexpected driving behaviors that can cause rear-end accidents.

http://saferstreetsla.org/679/case-studies-longer-yellow-lig...

https://www.motorists.org/issues/red-light-cameras/yellow-li...

Yellows have been shortened on purpose to raise revenues from tickets. If cities cared about accidents and congestion, we'd have less lights and more traffic circles, but that means less tickets and more thought and effort than lowering the yellow light time.

See https://www.motorists.org/issues/red-light-cameras/increase-... for various sources showing that red light cameras increase the number of accidents. Within The Netherlands there is a restriction on where such cameras can be placed due to this effect; basically the number of current accidents has to be pretty high.

Fun fact. Anticipated behavior is less likely to cause harm, whatever it is. If someone anticipates that you'll fire a gun at them, they'll have a much better reaction than someone who has a gun fired at them without that anticipation.

In California it is legal to be in an intersection on red as long as you entered it on green or yellow. The slam-on-the-brakes effect should at least be smaller than in states where being in the intersection on red for any reason is an offense.