If Accidents Are Up, the Cameras Must Be Working
According to the Houston Chronicle, a city-commissioned study found that "the number of crashes at Houston intersections with red-light cameras doubled in the first year after their installation." That looks like evidence that ticket-wary drivers are causing crashes by stopping short at intersections or speeding up at yellow lights, as predicted by camera critics, who argue that such surveillance systems compromise public safety for the sake of revenue. Yet camera advocates say the study supports their argument that monitoring intersections reduces accidents by deterring drivers from running red lights.
How so? The cameras, which so far have been placed at 50 intersections in Houston, do not cover every direction at each intersection, and the increases in accidents were sharpest in the lanes on which the cameras were not trained. Accidents in the camera-covered directions, by contrast, "remained relatively flat or showed only a slight increase." Camera supporters argue that the total number of traffic accidents in Houston rose last year and that the monitored lanes would have seen even bigger increases in crashes without the cameras. "Collisions are going up all over the city," a co-author of the study says, "but red-light cameras have held back that increase at approaches where they have been installed." (D.C.'s police chief deployed a similar argument a few years ago.)Camera critics say there is no evidence that "collisions are going up all over the city." According to Houston Police Department data, accidents have been falling since 2004.
Another problem with the argument that the cameras have improved traffic safety: Can it really be that they are working only in the directions on which they are trained? That assumes drivers know exactly which lanes are monitored by the cameras at a given intersection, as opposed to being generally aware that the intersection has cameras, which seems more likely to be the case.
Previous reason coverage of the red-light camera controversy here, here, here, here, and here.
[Thanks to Derek Ashworth for the tip.]