Road safety, speed cameras and speed humps

September 2008

In 2000 the Government set a target of a 40% reduction in the number of Killed and Seriously Injured (KSI) casualties by 2010.  The reduction was reference to average for the 5 years 1994 to 1998.  Hence it effectively applied to a 13.5 year period, namely, mid 1996 to the end of 2010.

How have they fared and how does that compare with the corresponding period ending 1996, prior to any significant effect from the camera campaign, which started in 1993?. 

Here are the numbers.  Detail is in Appendix 1.

Table 1: The record.

1. Annual reductions are on a compound interest basis.

Firstly, the rate of decline of deaths has fallen off from 3.4% per year to 1.8%, or by a factor of close to two.  However, there is an inconsistency; the downward trend in KSI has accelerated slightly rather than falling off.  That is highlighted by the change in the ratio of the KSI to Killed (down by 25%, from 14.2 to 10.4). There does not seem to be any reasonable explanation for the latter, beyond under-reporting of the serious casualties compared with earlier years, leading us to ask – is the reporting target driven?

In any event, if the trends continue then the reported reduction in KSI will be 43% by 2010, so achieving the Government’s target. However, (a) that target is not materially different from the 41% achieved for the previous period (when there were no special measures) and (b) the likely reduction in deaths, 22%, will be little more than half the 40% hoped for.

The collapse in the trend for deaths has happened despite the installation of tens, if not hundreds of thousands, of speed humps, endless traffic management measures and the speed cameras. The implication is that those policies are damaging, rather than helping, road safety.  In any event, had none of the present policies been in place, and had past trends continue unabated, then (a) both deaths and KSI casualties would have met the Government’s target and (b) there would have been 500 fewer deaths in 2007 than actually occurred.  That appears to make the DfT’s claim that the cameras save 100 lives per year difficult to sustain.

Not only do we have an apparently dreadful outcome in terms of lives saved but this punitive approach has undermined respect for the police and for the other authorities. Millions have been prosecuted, most of whom were, by any reasonable measure, driving sensibly.  Worse still, tens of thousands have lost their licenses and many of them will have lost their jobs or continued to drive - illegally and uninsured.

Support for that approach stems partly from the oft quoted statistic that a high (30%) number of road traffic accidents are due to speed (no doubt “excessive”).  We comment, it would perhaps be more honest to say that speed is a contributory factor in 100% of accidents, since if we were all stationary there would be none.  However, neither that nor the official line can be used in support of the speed cameras.  After all, the cameras only catch people who are breaking the speed limit and that is recorded as a cause in only a trivial proportion of road traffic accidents. Here is the background.

Traffic and Road Research paper LR 323 published in 1998 found that in only 7.3% of accidents was “excessive” speed recorded as a contributory factor.  Despite that Marie Taylor of the Traffic and Road Research Lab used the data to claim that 30% of accidents were speed related.  To do that she combined the following:

The combination was referred to as “inappropriate speed”.  It has morphed, in the minds of the policy makers and the public, to “speeding”(meaning breaking the speed limit) and to 30% when in reality breaking the limit was, on the face or it, present in only a trivial proportion of accidents.

Further, Table 6 from the paper with the title “Contributory Factors to Road Accidents” by the Road Safety Department of the DfT provides as follows (where, and  for technical reasons, the combination “Both above” does not match the total for the two classes).

Table 2. Speed as a contributory factor 2005

We comment, (a) those “Going too Fast” were within the speed limit and would not be influenced by the cameras (b) other contributory factors were also present, such as drunk driving, 20%, unknown percentages joy-riding, driving without due care and attention, etc.

Hence, the data cited above, particularly the combination of exceeding the limit and going too fast, creates a false impression as to the extent that driving within speed limits would have avoided any of those accidents.  Instead it may be reasonable to halve the numbers associated with Exceeding the Limit when considering the potential of the cameras.

Table 2 of the same report provides data on the causes of accidents.  That enables speeding (defined as breaking the speed limit) as a proportion of all causes to be calculated.  The percentages are: for fatalities, 5%: for the seriously injured, 3%: for the slightly injured, 1.8% and for all casualties, 2.2%.

Against that background it is difficult to see why the authorities continue to claim that (a) “speeding” is a major contributor to road traffic accidents (b) speed cameras have increased road safety. 

Meanwhile LR 421 and LR 511 have led to the often quoted statistic that a one mile per hour reduction in speed saves 5% of accidents, although there are different values for different classes of road.

However, the research first found that on fast roads the accident rate was lower than on slow roads.  The reason was that on the slow roads there were more hazards.  To overcome that the research divided the roads into 4 groups, each containing roads with similar characteristics.  The analysis then found a relationship between speed and accidents although that varied by class of road.  That result was almost inevitable since (a) it was the objective of the studies and (b) a relationship must exist since if we were stationary there could be no accidents. 

However it is reasonable to speculate that had the subdivision been (a) more extensive and (b) by the level of traffic flow then differences of speed and accident rate within each set of roads  would have vanished.  Putting it more bluntly, we have a method which, if taken to its limits, may prove “zero = zero”, which is no use at all.

Instead of that approach it would have been better to find the differing accident rates of motorists driving at different speeds on the same road. No doubt that would show those driving unduly fast or unduly slowly were causing the greater number of accidents.

Whatever the case, our view is that although a statement such as “a one mph reduction in speed saves 5% of accidents” may be good propaganda it does nothing to inform policy in the real world beyond providing a stick with which to beat the motorist.

Against that background we advocate that the present coercive and restrictive approach should be phased out in favour of one based on education designed to develop mature and deferential driver behaviour. Such a change might very well be met with a deafening cheer from motorists, balanced by a charge of cynicism and a furious outcry from the anti-car lobby.  However, that is a penalty that has to be paid if policy is to be based on the numbers especially when those numbers contradict previously held beliefs.

Note: The data in the Appendix shows that between 1980 and 2007 deaths per year declined by a factor of two and deaths per veh-km by a factor of 3.8 implying that driving today carries one quarter the risk per mile that arose previously.  The decline in deaths per passenger-km since 1965, is even more astonishing; down by a factor of 8.5.

Click here for the appendix...

Click here to download printable pdf...