You can't just ignore the consequences of vehicles hitting bollards, you have to weigh the likelihood of cars hitting them and the severity of those incidents against the likelihood of cars going past where the bollards would be and the severity of that scenario both when there are or aren't pedestrians that could be struck.
I'm not saying the status quo is correct, but I am saying that the author's tone does not strike confidence that they are approaching this from an objective and rational viewpoint that accounts for all the factors, at least in the case of bollards in locations where there's a good chance of high speed collisions with them.