Speed limits

The competent and careful actions of a majority of responsible people should obviously be considered legal

How should they be set?


Contrary to some ill informed opinions we welcome properly set speed limits, and welcome speed limit enforcement when speed limits are exceeded in a way that causes danger.

But how should speed limits be set?

Canadian common sense

This recent Canadian report (Review and Analysis of Posted Speed Limits and Speed Limit Setting Practices in British Columbia) has a clear view of the role of speed limits and speed enforcement. (home page) (actual report)

We'd rate this as a "must read", and it includes such straight forward common sense items as:

  • The majority of motorists drive at a speed they consider reasonable, and safe for road, traffic, and environmental conditions. Posted limits which are set higher or lower than dictated by roadway and traffic conditions are ignored by the majority of motorists.
  • The normally careful and competent actions of a reasonable person should be considered legal.
  • A speed limit should be set so that the majority of motorists observe it voluntarily and enforcement can be directed to the minority of offenders.
The 85th percentile

Much road engineering and speed limit design refers to the "85th percentile of vehicle speed".  What does it mean and why is it important? Refer to figure 1 below.

For any road situation there will be a spread of vehicle speeds. This spread of speeds is indicated by the green line. There's an average speed of vehicles and some travel faster and some slower. The 50th percentile speed is the speed which 50% of vehicles are not exceeding. (equally it's the speed which 50% of vehicles ARE exceeding).

The 85th percentile speed is the speed which 85% of the vehicles are not exceeding. The blue "accumulated frequency" curve shows how at each higher speed point the number of vehicles not exceeding that speed point increases.

Plenty of research has shown that the safest group of vehicles are travelling at or below the 85th to 90th percentiles. Research shows that crash risk alters with speed and this is shown by the red crash risk curve. At the 85th to 90th percentiles we tend to find drivers with above average skill and competence, and this is why their crash risk is the lowest. Above the 90th percentile we tend to find drivers exceeding safe limits and their accident risk increases as a consequence. Note that the "average" driver at the 50% percentile has a greater crash risk than the 85th percentile driver. Below the 30th percentile crash risk is significantly increased and these speeds tend to be used by less skilled and competent drivers.

This leads to engineering recommendations that speed limits should normally be set at around the 85th or 90th percentile speed of traffic under good conditions. There may well be some of the very safest drivers somewhat above the 90th percentile, but historically in the UK there has been Police discretion and an "enforcement margin" to avoid penalising these safe drivers. But how things have changed.

Does crash risk really increase for driver at and below average speed?

You should not need research to know that this view is true. Consider, for example, the crash risk of a car in lane 3 of a busy motorway. At extremely high speeds or low speeds it should be quite obvious that crash risk is elevated. There is ample scientific research to support the view too.

Safe speed has looked into the relationship between crash risk and travelling speed in some detail. We've realised that the curve shape of crash risk applies to an averaged population of vehicles and an averaged population of road conditions. The instantaneous crash risk for an individual vehicle is somewhat different. The speed contributes nothing to crash risk within a range of safe speeds. At the top end the crash risk rises sharply as the safe speed for the situation is exceeded. At the bottom end the curve splits. A position in the traffic flow will give an elevated crash risk at low speeds, but a position out of the main flow of traffic gives a reduced crash risk at low speeds. This gives us a "bath tub" curve.

When you add a lot of different bath tub curves (as you would when defining the risk attached to a population of drivers) you tend to create the smooth curves seen in the research into speed and crash risk.

Some research claims that accident risk simply increases with speed. The faster you go the greater the risk of crashing. Simply considering two situations will clearly demonstrate that this view is false.

  • Doddery old fool at 30 mph on a UK "A" road suitable for 60 mph. Of course he has an elevated crash risk. He does not know what he is doing. That's why he's doing 30 mph with a queue of increasingly frustrated traffic behind.
  • A single vehicle in lane 3 of a busy motorway. It's obvious that as the speed is reduced below 55 mph the crash risk will increase.
Motorway speed limits

Our 70 mph motorway speed limit is frequently exceeded and the 85th percentile speed in clear conditions is in excess of 85 mph. Since the theoretical prosecution threshold is 79 mph (=70mph + 10% + 2 mph) it is quite possible that the safest group of drivers will be prosecuted. This alone is sound justification to increase the motorway speed limit.

It is not as if the 70 mph motorway speed limit was set by any sort of science. As we understand it it was an arbitrary decision taken at a meeting at the then Ministry of Transport in 1965. In 1965 a typical new car was a Ford Anglia with an absolute top speed of about 85 mph and much inferior brakes.

Since 1965 cars have improved out of all recognition, yet the Motorway speed limit remains unchanged. It is certainly time for a review, and probably time for a change.

Rural speed limits

This is where we see the most problems with modern speed enforcement. The safe speed on British rural roads varies from 0 mph to over 150 mph, yet we have a "one size fits all" 60 mph national speed limit (70 mph on dual carriageways).

It's on rural roads where our system of speed limits misses the chance to warn drivers of hazards, and it's on rural roads that most of the speed camera injustices are perpetrated.

For example, our speed limit designers have missed out on the chance to warn drivers that they are leaving a section which is safe for 90 mph and entering a section which is only safe for 60 mph.

And in rural areas we've seen widespread and unjustified speed limit reductions usually policed by camera.

Raising and lowering speed limits

Referring to tables 10 and 11 of the Canadian report mentioned above we tend to see a pattern of raised speed limits leading to more accidents and vice versa. We don't generally believe that this should happen, and wonder at the possible error sources.

1) Regression to the mean effect. (speed limits were lowered when accidents were abnormally high and raised when accidents were abnormally low) (more on regression to the mean)

2) Deliberately biased reporting. We're seeing a lot of it in the UK. The claims from Iowa smell very much of this to us.

3) Speed limit changes tend to alter traffic volumes. After a rise in speed limit, more traffic uses the road and accidents rise in proportion to the traffic.

4) Urban speed limits revised downwards were genuinely set too high in the first place. This applied in Australia where an urban speed limit of 40 mph was probably too high, and was properly revised to 30 mph.

Other reports don't find this apparently consistent pattern. (see further reading below)
Urban speed limits

Our 30 mph urban speed limit is probably generally about right. It's existed since the 1930s, and during that time the roads in the UK have become roughly 20 times safer in terms of deaths per billion vehicle kilometres.

Sometimes 30 mph is far too fast, but because of the constantly changing road conditions we cannot expect to be able to generally impose a lower limit. We need instead a population of drivers who have the skill and inclination to slow down when necessary.  That's exactly what we had when we achieved (more or less) the safest roads in the World.

DoT Circular Roads 1/93

This important "circular" is available from (this page). The actual PDF is available (here)

It contains such important and sensible information as: 

5. The main purpose of specific speed limits is to provide for situations where it is appropriate for drivers to adopt a speed which is lower than the national speed limit. That limit does not imply that it is a safe speed under all conditions and drivers should adopt still lower speeds if conditions warrant. The establishment of speed limits is also a method through which legal sanctions can be brought to bear on those who drive markedly faster than is reasonable on that road. Specific speed limits cannot, on their own, be expected to reduce vehicle speed if they are set at a level substantially below that at which drivers would choose to drive in the absence of a limit.
It's well worth a read.
The speed limit trap

There's a tendency these days to think that vehicle speed problems in general will be solved by speed limits or by speed limit enforcement. They will not. Too much speed limit enforcement and emphasis is already leading to a reduced tendency for drivers to slow down when necessary.

Ideas like variable speed limits and differential wet/dry speed limits are talked about. These are mostly an attempt to paper over the cracks in the speed limit system. These ideas recognise the fact that different conditions demand different vehicles speeds, but fail to take account of the wide range of changes in instantaneous changes in conditions which drivers face daily. If drivers were not fully able to adjust their speed to the conditions would would have literally thousands of times the number of road accidents that we experience. 

People sometimes think that vehicle speed control is too important to trust to the drivers of vehicles, but trust them we must because we need them to slow (perhaps to a stop) whenever it isn't clear ahead. On every single road journey speed must be moderated far below the speed limit for serious and immediate safety considerations. This is an essential and unavoidable skill that all drivers must use if they are not going to crash almost immediately. It is fundamental to safe driving and cannot be replaced with speed limits and speed limit compliance. We call it safe speed behaviour. It is usually about being able to stop safely within the distance that you know to be clear. In real driving this clear distance ahead changes frequently (especially in town where most accidents happen).

It's wholly unrealistic to expect speed limits to replace proper speed setting decisions from drivers. That has never been the purpose of speed limits.

But modern oversimplified thinking keeps returning to the idea of a speed limit as a means to ensure that drivers use speed safely. This is the speed limit trap. It's the unfulfilled and unfulfillable promise that speed limits will make drivers choose safe speeds.

How should speed limits be set, and by who?

It should be obvious from all the above that speed limits should be set by skilled traffic engineers giving due consideration to 85th percentile speeds and local conditions.

Speed limits do little to modify the speed of traffic, and should never be used to attempt to modify the speed of traffic. The idea of lowering a speed limit and enforcing it by camera which is so common these days is a case of falling into the "speed limit trap". It's a dangerous mistake because it shifts vital responsibility away from drivers. (click here)

Speed limit setting should never be carried out by local councillors who lack skills in traffic engineering. It is quite absurd to think that safety can be improved with arbitrary (as opposed to properly engineered) speed limit reductions.

Further Reading

Speed limits - how they are set and your right to object (ABD)
Speed limit signs (ABD)
Bad speed limit signs (SafeSpeed)
Speeding (SafeSpeed)
Study Shows that  motorists drive at reasonable speeds (US Roads)
Effects of raising or lowering speed limits (US)
Speed Doesn't Kill (pdf of US research)
Synthesis of speed and safety research (US)


Speed limits based around 85th percentile of traffic speeds with discretionary enforcement served us very well in the UK up until about 1993 when the silly season started.

Now we have many speed limits set wrongly by the wrong people for the wrong reasons and enforced by cameras without discretion. This damages the credibility and usefulness of speed limits and leads to millions of pointless prosecutions, which in turn lead to hardship.

The bleatings about "we're only trying to save lives" do not stand scrutiny, because in camera infested Britain we're seeing the poorest overall road safety improvements in recent decades. (click here)

  • We must not allow our legal system to penalise the competent and careful actions of a majority of responsible people.
  • We must not allow speed limits to be improperly set. It is clearly dangerous to do so.
  • We must avoid the speed limit trap. (believing that speed limits can force drivers to set safe speeds)
And above all we must urgently return to the sensible speed limit and road safety policies which gave us (more or less) the safest roads in the World in the first place.
new Comments

Safe Speed encourages comments, further information and participation from our visitors. See our (forums).

Read about our comments policy (here).

Many pages (including this one) have a specific associated forum topic. You can (view) or (add comment) to the forum topic for this page. Posting in the forum requires simple registration.

You can't measure safe driving in miles per hour

We have a strict editorial policy regarding factual content. If any fact anywhere on this web site can be shown to be incorrect we promise to remove it or correct it as soon as possible.
Copyright © SafeSpeed 2003, 2004
Created 27/10/2003. Last update 7/03/2004