Figure 1: The first stage in an accident
Here is a thought experiment. A red car is travelling along a road, when another white car pulls out of a side road. Figures 1, 2 and 3 show the progression of the accident.
The accident starts with cause of the accident - this is rarely only due to speed. In this example, the cause is 'Looked but did not see', or 'Failed to judge speed correctly'. This is shown in Figure 1.
Figure 1: The first stage in an accident
The red car driver emergency brakes, and tries to get out of the way, as shown in Figure 2.
Figure 2: The second stage in an accident
However, despite this, the two cars collide, as shown in Figure 3. The red car is travelling too fast to stop, because the speed of the car is excessive. This part of the accident is very speed dependent, since if the red car was travelling slower, then the accident wouldn't have happened.
Figure 3: The third stage in an accident
When a collision like this occurs, kinetic energy is released. Kinetic energy is the energy that an object has because it is moving. The kinetic energy of an object goes up with the mass of the object, and exponentially with the speed of the object.
Ek=1/2mv2
The bodywork of the cars will absorb some of the kinetic energy. The remainder is delivered to the occupants of the cars. It is the kinetic energy which causes injuries and fatalities. This is shown in Figure 4.
Figure 4: The production and dissapation of kinetic energy
The situation is much worse if one of the parties is a motorbike or a cyclist. The more vulnerable party will receive most of the kinetic energy, and this kinetic energy will only be slightly reduced by any protective equipment.
This analysis shows why speed, and the kinetic energy produced by it, causes injuries and fatalities on the roads. Speed does indeed kill.
The notion that only inappropriate speed kills is fatuous. The killing agent, kinetic energy, has no emotions, and cannot be bargained or reasoned with. The appropriateness of the speed they were doing, or lack thereof, in the mind of the driver of the red car, is therefore entirely irrelevant.
All car drivers already slow down when they see a hazard, and speed up afterwards, yet most pay attention to the speed limit. Not paying attention to the speed limit has three major flaws in it.
A lot of roads have hidden hazards. For example, an entirely straight road may be marked with a 40mph speed limit. The car driver thinks - "I can do this road at 60mph!". However, the road has houses along its length, set back from the road. Cars and pedestrians, including children, can enter the road at any point, making a 40mph speed limit entirely appropriate.
Secondly, roads are designed. In the fantasy world that some petrolheads inhabit, the highway engineers spend all of their time with their feet up, drinking cappucinos. The reality is that the roads are designed for a certain speed limit, with braking distances, vertical and horizontal alignments, and visibility splays designed for the speed limits. Driving at a faster speed places the vehicle outside the safety envelope provided by the road.
Thirdly, the perception of speed in a car is determined by the surroundings. If the surrounding features are tall, and the road is narrow, the car driver feels boxed in, and the speed that they are actually doing feels faster, so they tend to slow down. If the surroundings are low, and the road is wide, the car driver feels relaxed, and the speed that they are doing feels slower, so they tend to speed up. This effect is used in traffic calming, e.g. the village gateway and dragons's teeth (Figures 5 and 6 - source DfT). Any car driver that uses their perception of speed to determine how safe they are is using an optical illusion as their guide - a crazy approach. This is why cars has speedometers fitted, and why the driver has to check their speed from time to time. The slogan 'Safety is more than a number' is just that - a vacuous slogan.
 |
 |
| Figure 5: Dragon's teeth | Figure 6: Gateway |
Figure 7 shows the number of fatalities in the UK, and more importantly - since more cars means more accidents - the number of fatalities per billion kilometres. There is a reduction in both measures in 1967, when seat belts and the breathalyser were introduced, although the numbers rebound in subsequent years - did the drivers get used to these measures, and reacted by driving more carelessly? Reductions in fatalities between 1983 and 1994 were probably due to making the use of front seatbelts mandatory, better car design, and major reductions in drink-driving fatalities.
Figure 7: Fatalities and fatalities per billion km on the roads of the UK
Speed cameras were introduced into the UK from the year 2000. They have increased rapidly in number, as is shown in Figure 8 (Source: Hansard).
Figure 8: The number of speed cameras in the UK by year
At the same time, the number of traffic officers has decline markedly, as shown in Figure 9.
Figure 9: The number of road traffic officers in the UK by year
There are two possibilites. The first is that speed cameras have radically increased the safety of the roads in the UK. The second is that speed cameras have reduced safety, and that because the fatality rate hasn't dropped recently, there is a 'fatality gap', where people are dying unnecessarily.
There is no evidence for either idea.
The fatality rate per billion km is still falling, and the rate of which it has been falling has changed slowly over time, leading to a smooth curve. Equally, there is no major drop in the fatality rate per billion km, which is what we would expect if speed cameras were really having a major impact.
In fact, it appears that we are approaching an irreducible minimum in the fatality rate per billion km. This is in fact what we would expect, since as the risk of something falls, fewer and fewer people will put themselves out to lower it further.
The only artifact is an apparent reduction in the slope of the fatalities per billion km around 1994. However, this also follows a major drop in the number of fatalities, for reasons suggested above. This cannot be then used to argue that something happened then which has made the roads more dangerous that they need be. It is also some six years before the first speed camera partnership was set up.
The decline appears to be roughly following an exponential curve, with a residual of 4 fatalities per billion km. A graph of ln(fatalities per billion km-4) against ln(year-1950) is shown in Figure 10.
Figure 10: Linear regression curve for fatalities per billion km
The line through the graph has a correlation (r2) value of 96%. The formula appears to be similar to this:
F=4+94.5e-0.064*(year-1950)
It would appear that despite the major benefits which accrue in a particular point from speed cameras (which may be of great benefit to local residents), on the road network as a whole the decline in road deaths is continuing in its own way, regardless. This is called 'Risk Compensation', as discussed by John Adams in his book 'Risk' (ISBN 1-85728-068-7). As one aspect is made safer (airbags, brakes, etc.) the public adjusts its behaviour to get the overall risk undertaken back to the level that each citizen feels is acceptable. Speed cameras increase the legal risk over only a small distance, leading to lower speeds, but has no effect outside this extent of road. Car drivers would then make up the speed outside of the area. The only way to get a substantial reduction in fatalities in a short time period is to increase the level of legal risk over the entire road network, thereby substantially reducing the fatality risk. This, speed cameras cannot do.
Firstly, speed cameras can only be deployed where there there is already an identified accident black spot. However, the speed camera then it left to manage speed at that point. This is like someone going to their doctor with a broken leg, and the doctor advises them not to stand on it. Just as we would expect the doctor to set the broken leg, so we should expect the road to be improved at that point. Each fatality costs £2m to the country, and so preventing a fatality by improving the road has a very high rate of return - more than that of most other economic activity. Then, the speed camera can be removed. Only where it is impossible to make an improvement, for example, a village with a narrow main road, and no pavements for pedestrians, should the speed cameras be retained.
Another example of an appropriate use is to cover roadworks, for the safety of the engineers and road workers at that location. The speed cameras can then be removed at the completion of the work.
Secondly, speed cameras are used as part of an automated system for punishing car drivers. Judging from some of the self-justifying remarks made in letters and on web sites, not all drivers caught by the speed cameras see the punishment as fair. If people see a fine as a 'gotcha', then the speed cameras will lose their ability to restrain drivers from speeding - since if they only restrain car drivers at certain discrete points on the road network, and the car drivers are not challenged as to whether or not the speeds at which they drive are reasonable, then they will be in general ineffective - as can be seen in Figure 7 above. The car drivers will likely slow down for the cameras, and speed up again afterwards. One of the more effective ideas is to allow the car driver to swap points on their license for a training session - more interraction between the authorities and car drivers caught by the cameras is required.
There is an alternative technology, which aims to inform rather than to punish. This is called 'Vehicle Activated Signs' . The technology is supported by the government, and is effective in most situtation, causing significant and sustained reductions in speed. The author's experience of this technology is that it works as described, with most drivers checking their speed. An example is given in Figure 11. It would be better if the number of violations was also recorded, so that a highway patrol presence could be provided if necessary. This would involve the addition of a rear-facing speed detector and a data logger to the device.
Figure 11: Vehicle Actuated Signs (source: DfT)
Real safety cameras would look very different.
We need to estimate the chance that a fatality will occur in any one journey. If we assume that the average journey cover five miles, then the chance of a fatality will be (roughly) 50 parts in a billion. The car driver would have to drive many lifetimes to stand a good change of being killed. This is why so many car drivers get away with driving in a careless or reckless manner (and a good thing, too).
Eventually, someone's luck will fail, and then someone will be hurt or killed. At this point, we are supposed to chorus,'It was an accident, it couldn't have been prevented'.
The truth is mixed. The accident could very likely have been prevented if the car driver responsible had been advised or banned over previous recklessness. Yet, even when such behaviour is observed by other car drivers, it is not reported to the police - either through a misconceived loyalty to the other car driver, through concerns for how busy the police are, or because it is hard to see and remember a number plate when seen only for a few moments.
A safety camera would consist of a camera mounted to the front and rear of every car on the road. The rear camera is there to record tail-gating, a very dangerous behaviour. The front camera is there to capture what the car driver can see through their windscreen. A police database would record the complaints by the public about other driver's behaviour, and those drivers complained about the most can be followed in an unmarked police car to see how exactly they drive. Almost all drivers will have complaints made against them, but some car drivers will be reported far more than others. The very presence of these cameras has already been shown to have a major deterrent effect
This technology would increase the legal risk to those who drive dangerously, reducing the fatality rates across the entire road network.
