George Monbiot has written a column this morning about the increase in road casualties in Oxfordshire following the deactivation of speed cameras. The critical point, as he notes is that you can’t learn much from such a small sample:
“So far, the sample size is too small and the period too short to be sure that the deaths and injuries around the county are linked to the switch-off. The experiment would have to run for longer and be conducted over a wider area. Any volunteers?”
After all, Swindon switched off their cameras and saw no increase in accident rates. The problem is that he casts speed cameras the safe option, and switching them off as the experiment. But in many ways speed cameras are themselves an experiment as road safety policy was focused overwhelmingly on enforcing speed limits, with less attention paid to other problems. We wrote a report last year aiming to test how that experiment had worked out.
If you look at the pattern of casualties before and after cameras were introduced, it looks very much like it didn’t go well:
To see if there was a significant change in the trend, we used a statistical test called a Chow Test. The results suggested there was.
Some people got very overexcited when we released our report. They looked at the graph and imagined extrapolating forward the blue pre-1990s trend line to a time when there would be negative deaths. Of course, at some point improvement was going to slow and then stop. But there is no sign that is what had happened in the early nineties. If it was, you would expect improvement to be slowing before speed cameras were introduced – it wasn’t – and to continue to slow afterwards – it didn’t. There is no particular reason to think that, in the early nineties, diminishing marginal returns drove the slowdown in road safety improvement. The line is on that graph to show the extent to which the two trends differ, not to make any grand prediction about an inevitable and eternal trend of road safety improvement.
Just switching speed cameras off doesn’t constitute a road safety policy. But there is good reason to think that the cameras, and the associated monomaniacal focus on speed limits, have been a mistake. We got better results before with a more balanced approach which we should try to reconstruct over time.