Speed Limits: their correct use, setting, and enforcement
Apart from a brief period in the early 1930s, speed limits have been imposed on public highways in Britain for as long as there have been mechanically propelled vehicles. Their use is so well established that few people ever stop to consider their value as an aid to road safety. The introduction of camera enforcement technology in the 1990s has led to a massive increase in speed limit enforcement, and this has occurred at the same time as many highway authorities have been introducing new or reduced local speed limits, often against the guidance issued by Government.
The controversy surrounding these developments has led to calls for reviews of both enforcement policy and speed limits themselves. This briefing paper is intended to inform decision makers on the way in which speed limits affect driver behaviour, and on how it is essential that speed limits are used and enforced correctly to ensure that road safety is enhanced, not reduced.
The ABD believes that excessive emphasis on the importance of speed limits is actually damaging to the safety culture that gave us the safest roads in the world in the first place.
The Purpose and Limitations of Speed Limits
Speed limits are so familiar that it is easy to forget that their proper purpose - and the only purpose of enforcing them - should be to improve road safety. Yet for some people the enforcement of speed limits has become an end in itself: since it is against the law to exceed speed limits, they insist that the law must be enforced to the letter, regardless of whether to do so is actually beneficial to safety.
It should be easy to see how a system of fixed speed limits has its limitations. Drivers are faced with constantly changing conditions as they travel along a road: changes in carriageway width, and in horizontal and vertical alignment; the presence of junctions, accesses and on-street parking; pedestrian and cyclist activity; all of these and more must be taken into account in selecting a safe speed at which to travel. On the same section of road, different speeds may be appropriate at different times of the day or night, on different days of the week and at different times of the year, and as weather conditions change.
Even where speed limits are set reasonably, therefore, they cannot provide more than an approximate guide to what might be a safe speed under average conditions: at some times or locations the speed limit might be too high; at others it might be too low. Furthermore, speed limits in Britain are set at round-number intervals of 10mph and, to avoid driver confusion, frequent changes in speed limit are discouraged. Given all these factors, it is clear that speed limits are very blunt tools and drivers must not assume that it is always safe to drive up to them - at the same time it is fatuous to assert that it is always dangerous to exceed them. Enforcement of speed limits should be flexible enough to allow for these limitations.
Despite the limitations of speed limits, when set correctly they provide a positive contribution to road safety in the following ways: Modern road safety thinking tends to over-value speed limits, and assumes that they can provide benefits far in excess of their capabilities. As a result, better approaches to road safety are ignored or undervalued and, far too frequently, speed limits are seen as a panacea for a wide range of road safety issues.
The Effects of Speed Limits on Driver Behaviour
On any given stretch of road, drivers will travel at different speeds, and the spread of these speeds will take the form of a 'normal' or 'bell' distribution, as shown by the green curve in Figure 1. The vertical line through the crest of the curve represents the average or 50th percentile speed (the speed below which 50 per cent of drivers travel). Another vertical line, to the right of the graph, shows the 85th percentile speed - the speed not exceeded by 85 per cent of drivers.

The red curve shows the accident risk of drivers in relation to their chosen speed within the speed distribution. This curve is based on the results of studies in various countries of speed and accident involvement1. It can be seen that, contrary to what might be intuitively expected, accident risk does not rise in a simple relationship with speed but is actually lowest for drivers travelling in the 80th to 90th percentile speed range. At speeds above the 90th percentile, accident risk rises sharply, but it also rises at lower speeds and, indeed, the slowest drivers are at similar risk to the fastest.
This is a very important phenomenon to understand: drivers who travel at around the 85th percentile speed are the safest and most competent. Among the very fastest drivers are those who tend to be the most reckless, lacking the experience or skills required to recognise the presence of hazards dictating the need for a lower speed. At the other end of the scale, the slowest drivers are the least confident and often have poor vehicle handling skills.
It is because of these findings that the 85th percentile speed is recognised by traffic engineers as the optimum level at which to set speed limits. When speed limits are set in accordance with the 85th percentile, it means that the majority of drivers, including the safest, are travelling within the law. This is an important principle, which is summed up well by the following quote:
The normally careful and competent actions of a reasonable individual should be considered legal. 2
Speed limits based on the 85th percentile reduce the spread of the speed distribution, especially by reducing the number of drivers travelling at the highest speeds: since the speed limit has been set in accordance with the actions of the responsible majority, the remaining drivers are more likely to accept that it is reasonable. The converse is also true: if speed limits are set below the 85th percentile speed, they are largely ignored, even by the safest drivers; the reckless minority then see that the limit has no respect amongst the majority of drivers, so they ignore it altogether. Thus speeds can actually increase when unreasonably low speed limits are introduced, and can decrease when speed limits are raised. Evidence to this effect was presented in an annex to the Department of Transport circular to local authorities on the setting of speed limits in 19803. The annex is reproduced in Appendix A to this paper.
There are other adverse effects of unrealistically low speed limits. Those drivers who are fastidious about always obeying the law, no matter what, will create a queue of impatient drivers behind them. Eventually, frustration may lead to a driver attempting an unsafe overtaking manoeuvre, putting all road users at risk, including the law abiding driver at the head of the queue.
Correct and Incorrect Use of Speed Limits
It has already been explained how changing conditions dictate the need for drivers to vary their speed continually along a road. At the same time, it would clearly be confusing if speed limits changed too frequently. Where speed limits are set, therefore, they will always represent a compromise, so it is important that they are applied to lengths of road that are reasonably homogenous in character. These conditions are most often met in urban areas, where speeds themselves and the spread of the speed distribution are generally low.
In rural areas it is rare for roads to have consistent characteristics, except in the case of purpose-built, modern motorways and dual carriageways. Most single-carriageway rural roads have frequently changing alignments, often with individual hazards (difficult junctions, sharp bends, crests, etc) joined by safer sections. In these circumstances, it is wrong to apply a blanket speed limit along the whole length of the road, on the assumption that it will reduce accidents at the hazardous locations: if the speed limit were set at a level dictated by the individual hazards, it would be too low on the safe sections, so it would be ignored by most drivers. The road would thus become more dangerous, not less. Similarly, if a limit were set that was reasonable on the safe sections, it could lead to the less experienced drivers increasing their speed through the hazardous sections, as the spread of the speed distribution reduced; again, this would create more danger.
Accident locations on rural roads need to be treated with engineering measures tailor-made for those sites. These could include junction redesign, carriageway realignment, or additional signs and markings to warn drivers of the hazard; inter-active signs that light up when a driver approaches a hazard above a pre-set speed have been found by the Transport Research Laboratory to be very effective4.
Another incorrect use of speed limits that is becoming increasingly common is to extend town or village speed limits into open country on the edges of those settlements, sometimes for considerable distances. The thinking behind this is presumably to encourage drivers to slow down before they reach the built-up area itself, but they often have the opposite effect: the speed limit is too low when drivers enter it so they ignore it; when they reach the point where the limit is actually required, the less experienced drivers fail to adjust their speed. This is an example of the need for speed limits to be applied to homogenous lengths of road; speed limits should only change at points where drivers can relate the change of speed limit to a change in the road environment. In situations such as these, a more effective approach would be to erect count-down boards at hundred-yard intervals, for the 300 yards approaching the start of an urban speed limit, as a warning to drivers of the speed limit ahead.
Another practice in some areas is to apply speed limits on the basis of a road's status in a 'hierarchy', depending on the perceived function of the road. This can lead to speed limits being set that are well below the 85th percentile, so they do not command respect: drivers adjust their speed according to the physical environment of the road they are travelling on, not the status or classification assigned to it by bureaucrats.
Similarly, speed limits are sometimes reduced in the name of protecting pedestrians, cyclists or horse riders, even though few such road users may actually use the road: it is sometimes argued that more people would walk, cycle or ride if speeds were lower. Drivers do reduce their speed when vulnerable road users are around - hence the low speeds found in busy town centres or outside schools when children are arriving or leaving. It is not reasonable, however, to expect drivers to travel slower than the actual conditions dictate at the time, just because someone might want to walk, cycle or ride along the road at some point in the future. Drivers should always be prepared to encounter vulnerable road users and respond appropriately when they do so, but blanket speed limits will, as previously explained, increase rather than reduce danger.
Where there is a genuine reason to reduce traffic speeds below the natural level indicated by the 85th percentile, then the visual environment as seen by drivers must be altered through engineering measures, in order to reduce the 85th percentile speed to the level required. The only sustainable and self-enforcing speed limits are those set in accordance with the 85th percentile principle.
These principles for the correct use of speed limits are set out in the current Department for Transport circular on the setting of local speed limits5. Regrettably, many local highway authorities are now ignoring this guidance and setting unsound speed limits that are having a detrimental effect on road safety.
Enforcement of Speed Limits
It has been shown above that speed limits, even when set in accordance with sound scientific principles, are at best no more than a guide to a safe speed under average conditions. At the same time, speed limits are mandatory - it is an absolute offence in law to exceed them. These two facts do not sit well together. In the past, the police have traditionally used a degree of discretion in enforcing speed limits, but this has been reduced with the introduction of camera technology, at the same time as the scientific basis for speed limit setting is being abandoned.
These two trends are combining to create a crisis of confidence in the speed limit system and in the way enforcement is carried out. Speed cameras operate around the clock and make no allowance for changing road and traffic conditions. Furthermore, the rules of the speed camera hypothecation scheme mean that cameras can only be installed at places where the measured 85th percentile speed is well above the speed limit6. Since the 85th percentile is the speed chosen by the safest drivers, this means that cameras are installed at locations where it is safe to exceed the speed limit, not at locations where it is dangerous. This is clearly absurd and unjust.
When speed limits were enforced by trained traffic officers, they would take into account the circumstances at the time in deciding whether prosecution was warranted. With speed camera enforcement, no such discretion exists and most forces apply the guidelines issued by the Association of Chief Police Officers (ACPO) in setting the level above the speed limit at which enforcement begins. When the existence of these guidelines was publicly acknowledged, however, Transport 2000 threatened a legal challenge on the basis that ACPO was endorsing law breaking - it wanted no tolerance to be allowed whatsoever.
Although ACPO officially withdrew the guidelines, they are still in use, but it has been suggested that the threshold might be reduced, leading to further criminalizing of safe drivers. Tighter, less flexible enforcement has served to emphasise the flaws in a system of rigid, mandatory speed limits.
Current enforcement policy is giving out the false message that speed within the limit is safe while speed above the limit is dangerous. This leads to some drivers making false assumptions of safety simply because their speed is within the speed limit. A radical change of direction is needed.
  1. A return to the scientific system of speed limit setting, based on the 85th percentile principle, is essential throughout the country. Highway authorities should not be permitted to adopt local policies that contravene this principle. All new speed limits introduced since 1993, when the current guidance on speed limit setting was issued, should be subject to an independent audit and limits removed or adjusted if they do not comply with that guidance.
  2. The hypothecation scheme for the self-financing of speed cameras should be scrapped and the rules for the siting of cameras changed, so that they cannot be located where it is safe to exceed speed limits. Cameras should only be permitted at accident black spots where at least 85 per cent of drivers observe the speed limit, indicating that it is potentially unsafe to exceed it. Speed limit enforcement by means other than cameras should be targeted in the same way. Consideration should be given to the increased use of interactive signs warning drivers of the speed limit, as an alternative to police enforcement. Police traffic patrols should be restored to former levels and more emphasis given to careless driving within speed limits, which speed limit enforcement does not address.
  3. The use of variable speed limits should be made easier, so that limits could be reduced during periods when slower speeds are appropriate, e.g. outside schools at arrival and dispersal times, but raised at other times.
  4. Police traffic patrols should be given clear guidelines about enforcing speed limits in circumstances where danger is actually being caused, taking into account the prevailing conditions and the manner in which a vehicle is being driven. In the longer term, consideration should be given to the possibility of making some speed limits advisory rather than mandatory. The national speed limits might be most appropriate for this treatment in the first instance. (The ABD has already produced a submission for an increase in the motorway speed limit7.) The onus would then be on a driver involved in an accident while exceeding the advisory limit to show that his speed was not excessive for the conditions; conversely, if he was within the limit, it would be up to the police to show that he was travelling too fast. This would be a radical change in thinking, but it would bring an end to enforcement for its own sake.


1. An example can be found at www.tfhrc.gov/safety/speed/speed.htm.
2. Establishing Speed Limits - A Case of Majority Rule. Arizona Department of Transportation. 1999.
3. Circular Roads 1/80 - Local Speed Limits. Department of Transport. 1980.
4. TRL Report TRL548. Vehicle-activated signs - a large scale evaluation. Transport Research Laboratory. 2002.
5. Circular Roads 1/93 - Local Speed Limits. Department of Transport. 1993.
6. As revealed by Paul Smith of SafeSpeed, Motor Cycle News, 28 January 2004. See www.safespeed.org.uk/rules.html.
7. The submission can be viewed at www.abd.org.uk/motorwayspeedlimit.htm.
  1. It is a common but mistaken belief that drivers allow themselves a set margin over the prevailing speed limit, and that if a limit is raised by 10 mph, they will travel 10 mph faster. In fact, an increase in an unrealistic speed limit rarely brings an increase in traffic speeds. ("Unrealistic" is here used to mean "substantially below the 85 percentile speed"). It is much more likely that there will be no change, or even a fall. It seems that drivers relieved of the frustrations of too low a limit rarely abuse the higher one. Indeed it is not unusual for the accident rate to fall when a poorly-observed limit is raised. This may mean that reduced frustration leads to changes in driving behaviour conducive to accident reduction.
  2. The evidence for asserting that speeds and accidents do not increase in proportion to an increase in speed limit comes from studies made before and after unrealistic local limits have been raised. Some of the main evidence is summarised in paragraphs 3-9 below.
  3. In 1960, a Departmental Road Safety Committee reporting on the results of the experimental introduction of 40 mph speed limits in the London area concluded that the raising of the limit had resulted in no appreciable change in speeds, while the accident rate remained substantially the same. The committee considered that the higher limit had achieved its purpose of removing unjustifiably low speed limits, and encouraging a proper standard of enforcement.
  4. A before and after study carried out at 20 locations through Kent, where the limit had been raised from 30 mph to 40 mph, showed a fall in speed, or no change, in 80% of the measurements taken, and a small increase in the others. The total number of accidents fell by almost 20%.
  5. In 1973 the Metropolitan Police produced the results of a study on six sections of trunk road where - in accordance with the Department's criteria - speed limits had been raised from 40 to 50 mph, or from 30 mph to 40 mph. At four locations the 85 percentile went up by less than 2 mph and at two locations it went down. Allowing for a general decrease in accidents, the reduction in the number of accidents at these places was 15%.
  6. When the speed limit in Park Lane, London W1, was increased from 30 mph to 40 mph in 1970, the 85 percentile speed fell from 43.6 mph (measured in 1970) to 39.2 mph (measured in 1974).
  7. In 1974, the Midland Road Safety Unit reported the results of a study of a large number of speed limit changes from 30 mph to 40 mph. Their conclusion that there had been no significant increase in either speeds or accidents was in line with the conclusions from a similar exercise for cases in other parts of the country carried out within the Department.
  8. The Department has recently conducted a survey of the effects of changing the levels of speed limits in various parts of the country. The results indicate that raising speed limits has little effect either on the speeds of vehicles or the rates of accidents.
  9. The following examples of local speed limit changes from 30 mph to 40 mph illustrate this point.
    CountyRoad85%ile SpeedAccident Rate
    West Yorkshire A5840/4347/521.450.65
    WarwickshireB4453 42/4443/433.21
    The table above shows that, in the one instance in which speeds rose, the accident rate went down.
  10. With the removal of the energy conservation speed limits in June 1977, the national speed limit went up from 50 to 60 mph on single carriageway roads and from 60 to 70 mph on dual carriageway roads. This afforded an ideal opportunity to judge if traffic responds to national speed limit changes in the same way as it does to local changes. A survey of speeds at 49 points throughout the country made in July 1977, compared with a similar survey in July 1976, showed that for cars and motorcycles with a headway of at least 5 seconds there was no change in the mean speed on single carriageway roads and a 1 mph on single carriageway roads and 2 mph on dual carriageway roads (sic). An analysis of national accident rates in the months following the changes shows no evidence that raising the limits caused any increase in the number of accidents.