The Association of British Drivers'
Response to Public Consultation on
Updating Circular Roads 1/93

1. Introduction
1.1 This document sets out the response of the Association of British Drivers (ABD) to public consultation on proposed changes to Circular Roads 1/93 (Setting Local Speed Limits). It is written as a stand-alone document but has also been submitted to the Department for Transport (DfT) in conjunction with the consultation questionnaire, from which sections in this document have been cross-referenced.
1.2 Within each of the following sections, therefore, a discussion is followed by a list of summary points, for easy reference from the numbered questions in the consultation.
1.3 The ABD's response to the consultation has been formulated on the basis of the following principles concerning the use and setting of speed limits: 2. Policies and Main Issues
2.1 The ABD shares the DfT's concerns that traffic authorities in recent years have been applying different approaches to the setting of local speed limits, especially on rural roads, with the result that there are wide variations in speed limits between roads of similar character in different parts of the country. This is confusing to drivers and has increasingly brought the system of speed limits into disrepute. The ABD would support action to ensure that local speed limits are applied and set on a consistent basis, and also at levels that drivers can see are reasonable. The latter requirement is essential if a high level of compliance is to be achieved.
2.2 The ABD is not convinced that the proposed changes to the speed limit circular are likely to achieve these aims. A system of consistent and reasonable speed limits could be achieved if the guidance in Circular Roads 1/93 were applied equally by all traffic authorities. The fact that the circular is not legally binding has enabled authorities to adopt policies on speed limit setting that do not comply with the guidance. It appears to the ABD that the proposed changes have been designed to give retrospective legitimacy to the actions of these 'rogue' authorities and to encourage others to follow suit. Also, since the new circular will have no more legal force than the old, there will be nothing to stop authorities that are so minded from ignoring its provisions where they conflict with those authorities' predetermined policies.
2.3 The draft circular does not sufficiently emphasise that there should be a presumption against the setting of local speed limits, especially on rural roads, where they should be viewed as a last rather than a first resort. Speed limits should be used to alert drivers to marked changes in the road environment, such as when entering a built-up area or a section of road where its geometric characteristics demand a lower speed. In the latter case, a local speed limit should only be applied when the view of the road to the driver does not make it sufficiently clear that the characteristics have changed - in which case, signing or engineering measures should be considered first.
2.4 The need for a driver to exercise judgement in adjusting speed to the prevailing conditions is one of the fundamental requirements of the driving task. Speed limits should be used to assist drivers in that task, not to take the responsibility away from them. A proliferation of local speed limits carries the risk of drivers becoming reliant on external information for adjusting their speed, when the factors governing the safe speed under a particular set of conditions are too complex for any system of speed limits to replace a driver's judgement. This tendency towards 'driving by numbers' is eroding drivers' abilities.
2.5 The advice in paragraph 33 of the draft circular, that speed limits should not be used to solve accident problems at isolated hazards, needs to be given much more prominence. There are many examples where traffic authorities have introduced such speed limits and they should be required to review those limits, with the intention of replacing them with specific measures to improve safety at the hazards themselves.
2.6 Mandatory speed limits that apply around the clock should not be used to address circumstances that exist infrequently or for limited periods. Examples include roads that may be used by significant numbers of walkers, cyclists or equestrians, but only at limited times, such as weekends; or around schools at arrival and dispersal times during school terms. Drivers should be advised by signs that they may encounter vulnerable road users in these locations and should generally be relied upon to adjust their speed when they do so. Where a speed limit is considered necessary, it should apply only at the times when it is needed. The procedures for introducing such variable speed limits should be simplified to enable their use in these locations.
2.7 Paragraph 88 of the draft circular states that 45 per cent of serious and fatal casualties occur on rural roads, where accident reductions have been at a notably slower rate than in urban areas. It is claimed that speed is a major factor in this difference, implying that reductions in rural speed limits are justified. These statements fail to acknowledge that traffic levels are growing at a higher rate in rural areas, as traffic growth in urban areas is more constrained by lack of capacity. They also fail to recognise that lack of investment in upgrading major rural roads is a contributory factor to the relatively poor improvement in safety on those roads.
2.8 For the reasons given in paragraphs 2.3 to 2.5 above, the ABD opposes the use of lower speed limits in an attempt to reduce accident frequency on whole routes or over large areas. The use of speed limits in these circumstances is based on the flawed premise that lower average speeds lead to reductions in accident rates. Paragraph 15 of the draft circular repeats the claim that there is a 5 per cent change in accident rate for every 1mph change in average speed. This claim is based on the findings of a Transport Research Laboratory (TRL) report published in 1994. The report reviewed a large number of studies worldwide on the relationship between speeds and accidents and combined the results using various statistical techniques.
2.9 One of the most dubious methods used in that research was the way in which mean speeds and accidents were compared on roads claimed to be of similar type. The report used accident rates rather than accident numbers, since the latter are influenced very strongly by traffic flow. By using accident rates the report's authors claimed to have compensated for differences in flow. However, although accident numbers increase with traffic flow, they do not do so in direct proportion. Thus the accident rate decreases with increasing flow. But speeds also decrease as traffic levels rise. So it is wrong to claim that decreasing accident rates are the result of lower speeds. Rather, both accident rates and speeds decrease as a result of higher traffic flows.
2.10 Evidence to support the ABD's assertion that the claimed relationship is flawed comes from the TRL itself. TRL's published project report PPR026 is a technical guide for road safety engineers, to help them decide which rural roads should be given priority for accident analysis. Paragraph 4.15 in that report states: "The relationship between accidents and vehicle flows is not a linear one (e.g. see Walmsley and Summersgill, 1998). For this reason, it is recommended that roads with very different flow levels are not studied together." Yet this is exactly what was done in producing the average speed/accident rate relationship used to justify many speed limit reductions.
2.11 The assessment framework proposed within the draft circular gives further weight to the argument that, other things being equal, it is traffic flow rather than average speeds that determines accident rates (see paragraph 4.10 below).
2.12 Where local speed limits are applied to significant lengths of road, especially in rural areas, there will be sections where the limit can be exceeded safely and others where the maximum safe speed is below the speed limit. In the absence of a local speed limit, the national speed limit is not indicated by signs, except where drivers leave a built-up area. Drivers are not constantly reminded, therefore, that they are legally allowed to travel at up to 60mph, so they are more likely to vary their speed according to the conditions rather than automatically drive up to the speed limit. A lower speed limit requires repeater signs at regular intervals, so drivers are subject to constant reminders that it is permissible to travel at that speed. Some will take this as an indication that it is safe to travel at the indicated speed. This could lead to higher speeds in the hazardous locations, exactly the opposite of what is intended - an example of the erosion of driver skills mentioned in paragraph 2.4 above.
2.13 Although a majority of drivers will ignore speed limits that are seen to be set at unreasonably low levels, a few will endeavour to obey the law however absurd it may appear. These drivers will soon end up with a queue of vehicles behind them, causing frustration in the drivers of those vehicles. Inevitably, some will lose patience and attempt risky overtakes that may not succeed. The ensuing accidents may well involve the slow moving driver at the head of the queue. Thus those who obey the law can end up putting themselves and others at risk. Badly set speed limits may therefore create danger rather than reduce it.
2.14 Summary of main points: 3. The Use of Mean Speeds for Speed Limit Setting
3.1 One of the most important changes proposed in the draft circular is the replacement of the 85th percentile speed by the mean speed for determining an appropriate speed limit. The ABD opposes this change and questions the stated reasons for it. In particular, paragraph 37 of the draft circular states:
"Practitioners' thinking has evolved since then [the publication of the 1993 circular] and many have expressed concern that 85th percentile speed can be heavily influenced by excessive speeds travelled by a minority of drivers Ö The Department shares this view and therefore recommends that mean speeds be used in future assessments of appropriate speed limits."
3.2 On the face of it, this quote appears to show a lack of understanding of basic statistical definitions and methods. To obtain the 85th percentile speed of a sample of vehicles, the individual measured speeds must be ranked in descending order and the top 15 per cent discarded. The highest speed then remaining is the 85th percentile. The mean, on the other hand, is obtained by adding together all the individually measured speeds and dividing by the total number of observations. Consequently, a small minority of high speeds will affect the mean but not the 85th percentile.
3.3 The ABD does not believe that the DfT is unaware of the difference between a percentile and a mean. Rather, it considers that the quote from paragraph 37 of the draft circular is intended to conceal the real reason for moving to the use of mean speeds, which is to legitimise the speed limit policies already adopted by some traffic authorities. Where those policies demand imposing speed limits well below the 85th percentile level, then anyone driving above those predetermined limits is deemed guilty of 'excessive speed'. The 'minority' of such drivers will then be far greater than 15 per cent and will thus affect the 85th percentile speed.
3.4 The 85th percentile speed has formed the basis of guidance on speed limit setting for over 30 years, and with good reason. International research into the relationship between accident risk and distribution of vehicle speeds has shown that the lowest risk is associated with drivers travelling at around 90 per cent of the maximum recorded speed - not at the mean speed. Accident risk increases sharply above the 90th percentile speed, but it is also high at speeds at the lower end of the speed distribution. The 85th percentile has been used as the guideline level for speed limit setting as it provides a margin of safety below the point at which accident risk starts to rise.
3.5 The other major advantage of setting speed limits at the 85th percentile level is that it allows the majority of drivers, including the safest, to travel at the speed they would choose anyway, within the law. Speed limits set this way achieve the greatest level of compliance and they have a moderating effect on the speeds of the fastest drivers, who see that the majority respect the limits. Speed limits set significantly below the 85th percentile level do not achieve good compliance: they criminalize the safest drivers and encourage the fastest to ignore them altogether. Paradoxically, therefore, reducing speed limits can lead to an increase in the highest speeds.
3.6 The mean speed alone gives no indication of the range of speeds within a speed distribution: a measure of the variance of the distribution is needed as well. In Appendix C of the draft circular it is stated that for a 'typical' distribution of vehicle speeds on single carriageway rural roads, the 85th percentile speed is about 6mph above the mean where there is a 50mph speed limit, and 8mph above the mean speed where there is a 60mph limit. These may well be valid average figures, but there will be many cases where the relationship between the mean and 85th percentile speeds differs significantly from these averages. The 85th percentile speed automatically provides a measure of the variance of the speed distribution.
3.7 Appendix C also states that, for mean speeds to be an acceptable basis on which to set speed limits, the measured mean speed should be no higher than the proposed limit. Under the current guidance in Circular Roads 1/93, a speed limit may be set if the 85th percentile speed is up to 20 per cent, or 7mph (whichever is greater), above the proposed limit. The ABD accepts that, where the average relationships between the mean and 85th percentile speeds apply, there could be circumstances where the application of either the existing or the proposed guidance would result in the same speed limit being set. For example, on a road with mean and 85th percentile speeds of 48 and 54mph, a 50mph speed limit would be justifiable under either system.
3.8 Despite this, there are several issues of concern to the ABD. First, as stated in paragraph 2.2 above, some traffic authorities are already ignoring the current guidance and setting speed limits well below the 85th percentile level. Since the new guidance will not be legally binding either, there will be nothing to stop authorities setting speed limits below the mean speed if it suits them.
3.9 Second, as already pointed out in paragraph 3.6, there will be occasions where the average relationships between the mean and 85th percentile speeds will not apply. This is likely to be the case on those single carriageway roads where it is perfectly safe to travel at the national speed limit of 60mph when traffic flows are light, but where increasing congestion and/or a high proportion of heavy goods vehicles keeps speeds artificially low for part of the day. The use of mean speeds for speed limit assessment would almost certainly lead to many more 50mph limits being imposed on this sort of road, as is already happening in certain counties. Far from improving safety, these speed limits increase driver frustration and danger, as well as criminalizing those travelling at perfectly safe speeds when traffic flows are light.
3.10 Third, the ABD considers that the tolerance in the existing guidance, that allows speed limits to be set below the 85th percentile speed, is too great. The current 20 per cent tolerance allows, for example, a 50mph speed limit to be set where the 85th percentile speed is as high as 60mph. This is clearly not within the spirit of the 85th percentile principle, which the ABD would like to see retained as the basis of speed limit setting, but with the tolerance reduced so that a speed limit cannot be set more than 4mph below the 85th percentile speed, across the whole range of speed limits. This would mean that traffic authorities should not be permitted to set a local speed limit on any single carriageway road with an 85th percentile speed of 55mph or more.
3.11 Summary of main points: 4. Speed Assessment Framework
4.1 The ABD agrees with the statements made in paragraph 31 of the draft circular about the importance of how a road looks to a driver, including the presence of potential hazards, if drivers are to respect speed limits. As stated in paragraph 2.3 above, the ABD agrees that a principal aim of speed limits should be to alert drivers to changes in road geometry or environment.
4.2 The ABD is unhappy, therefore, that a speed assessment framework is being proposed that categorises rural roads according to their function, leading to the possibility that otherwise similar roads could have different speed limits, simply because of a bureaucratic decision about whether they are 'upper tier' or 'lower tier'. Drivers do not take such considerations into account in choosing the speed at which to travel. While opposing the two-tier framework in principle, the ABD has detailed comments to make on the proposed criteria within the framework.
4.3 For upper tier roads, the criteria would allow a 50mph speed limit to be introduced if the accident rate exceeds 35 injury accidents per 100 million vehicle kilometres and the mean speed is below 50mph; or if the mean speed is below 50mph but the accident rate threshold is not exceeded. With the criteria as written, therefore, the accident rate threshold is redundant, since a 50mph speed limit could be imposed on any road where the mean speed is already below 50mph.
4.4 If the difference between mean and 85th percentile speeds quoted in Appendix C of the draft circular is correct, the imposition of a 50mph speed limit on a road where the mean speed is already below 50mph would have little or no effect on higher speeds. It could not be expected, therefore, to have a downward effect on accident rates, even if the claimed relationship between average speeds and accident rates were valid (which the ABD disputes - see paragraphs 2.8 to 2.10 above). In practice, however, a 50mph speed limit could lead to an increase in speeds at the more hazardous locations, as drivers feel less responsible for adjusting their speed according to the conditions (see paragraphs 2.4 and 2.12 above). The application of a local speed limit in these circumstances also offends against the principle that speed limits should only be used where they alert drivers to changes in road geometry or environment.
4.5 The unwritten implication in the criteria is that where an upper tier road has an accident rate that exceeds the threshold of 35 injury accidents per 100 million vehicle kilometres, but the mean speed is above 50mph, a 50mph speed limit could be introduced if engineering measures were taken to bring the mean speed down to below 50mph. This again assumes a relationship between mean speeds and accident rates that is fundamentally flawed. Reducing speeds is no substitute for investigating and addressing the true causes of accidents.
4.6 For upper tier roads, a 40mph speed limit is recommended where there is 'substantial' development or there are 'considerable' numbers of vulnerable road users. These definitions are open to wide interpretation. In particular, speed limits should not be reduced to allow for circumstances that apply only during limited periods. Variable speed limits and/or signs warning drivers to take particular care where vulnerable road users may be encountered should be considered instead (see paragraph 2.6 above).
4.7 For lower tier roads, 60mph is recommended as the default speed limit for roads with a 'mixed' function, i.e. those partially used by through traffic. Appendix D states that: "In the longer term these roads should be assessed against upper tier criteria." This begs the question of why they are initially being categorised as lower tier if they should really be considered as upper tier. It suggests another concession to those traffic authorities that have imposed low speed limits on roads of this type and are to be allowed to retain them where they are not justified.
4.8 For roads that fully meet the definition of lower tier, i.e. they perform a local or access function, the maximum speed limit would be 50mph regardless of a road's geometric characteristics, the degree of frontage development, or its use by vulnerable road users. While paragraph 101 of the draft circular states that application of the criteria does not mean that speed limits should automatically be reduced, the reality is that many traffic authorities will use the criteria to justify the introduction of more unnecessary speed limits.
4.9 Furthermore, 40mph speed limits will be allowed on lower tier roads if the accident rate exceeds 60 injury accidents per 100 million vehicle kilometres. Once again this assumes, falsely, that accident rates can be reduced by lowering speed limits.
4.10 The proposed accident rate threshold for deciding between speed limits within the lower tier is 70 per cent higher than that proposed for the upper tier. This reflects the fact that roads in the lower tier will, in general, have lower traffic flows, and that accident rates are inversely proportional to traffic flow. It shows that, for the same mean speed, a higher accident rate would be expected on a lower tier road than on an upper tier one. This is further evidence that there is no direct relationship between mean speeds and accident rates, but that both are related to traffic flow.
4.11 For both tiers, the proposed speed limit in villages is 30mph. The ABD agrees that this is the appropriate speed limit for the centres of villages, where there is near-continuous frontage development and the greatest concentration of activity by pedestrians and other non-motorised road users. The definition of a village in Traffic Advisory Leaflet 1/04, however, is where there are 20 or more houses in a minimum length of 600 metres. The strict application of this definition could lead to 30mph speed limits being imposed on sections of road that are suburban in character, where a 40mph speed limit would be more appropriate and likely to achieve greater compliance. Speed limits in villages should not be set on the basis of the definition alone, therefore, but account should be also taken of the measured speed distribution.
4.12 Summary of main points:
5. Street Lighting and Restricted Roads
5.1 The ABD is pleased that the DfT is now advising traffic authorities that they should not in future use section 82 of the Road Traffic Regulation Act 1984 to impose restricted road status on roads that do not have street lighting. But this welcome guidance is negated by the extraordinary advice contained in paragraph 51 of the draft circular:
"Any fitting that throws light on a carriageway constitutes street lighting and traffic authorities should bear this in mind as in principle this could include any lights on adjoining buildings or other lights not specifically put in place for road illumination. Ultimately it would be for a court to decide whether 'street lighting' was in operation in a particular place."
5.2 This is tantamount to inciting traffic authorities to avoid their responsibilities under section 85 of the 1984 Act, which requires them to provide adequate guidance to drivers about speed limits they are expected to observe. On a restricted road, drivers are supposed to realise that they are in a 30mph speed limit when there is a system of street lighting - even during daylight. The street lamps act, therefore, in lieu of road signs so they need to be clearly visible at all times. It is quite absurd that drivers should be expected to guess whether there are lamps close to a road that could be construed as street lights for the purposes of determining the speed limit.
5.3 The advice also ignores the definitions of 'footway lighting' and 'road lighting' in section 270 of the Highways Act 1980. Lamps that conform to the definition of footway lighting cannot be deemed to be carriageway lighting.
5.4 The draft circular highlights the archaic concept of speed limits being linked to the presence of street lighting. There are many instances, especially in the less dense rural settlements, where street lamps are placed at irregular intervals, hidden by vegetation, or fixed to buildings, telegraph poles or electricity transmission poles. Their presence is often far from obvious, especially during daylight and, coupled with non-continuous frontage development, drivers can easily be led to believe that the speed limit is higher than 30mph.
5.5 Even where street lighting is obviously present, drivers are sometimes unsure of the speed limit. Many urban dual carriageways and suburban main roads, where drivers reasonably expect that a 40mph speed limit should apply, have been downgraded to restricted road status in recent years. The absence of repeater signs on these roads has led to many drivers being prosecuted for exceeding the speed limit, when they genuinely believed it was higher.
5.6 While drivers are supposed to assume that the absence of repeater signs on a street-lit road means that there is a 30mph speed limit, the failure of many traffic authorities to erect and maintain repeater signs at regular intervals where a higher speed limit applies adds to their uncertainty. For instance, there are many examples of street-lit sections of rural roads where the national speed limit applies but traffic authorities have failed to erect signs to diagram 671 (the 'end of maximum speed limit' sign), as they are required to do under direction 11(5) of the Traffic Signs Regulations and General Directions 2002.
5.7 The present system, coupled with the failure of many traffic authorities to meet their obligations under section 85 of the 1984 Act, is confusing to drivers and requires overhaul. While the ABD accepts that it is not necessary to install 30mph repeater signs in every residential road in the country, they should be required on all roads that carry a 30mph speed limit and perform the function of a local distributor road (or higher), regardless of whether they have street lighting. Only by doing so will drivers, especially those unfamiliar with an area, be able to know with certainty what the speed limit is. An amendment should be issued accordingly to the Traffic Signs Regulations and General Directions 2002.
5.8 Summary of main points:

August 2006
As was not unexpected by anyone who understands how Blair's government worked, the DfT announced they were going to use mean speeds as the basis for speed limit setting.
The ABD issued a press release on 11 Aug 2006 condemning the DfT for ignoring the rejection of this idea by the majority of the police forces who responded to the consultation.
SafeSpeed went one further, describing the DfT as "not fit for purpose".
ABD PR 636 — County Council Ignores Police Opposition to Speed Limit Cuts
Covers plans by Warwickshire County Council, the first authority to use means speeds as an excuse to reduce speed limits. Includes photos of roads with inappropriate reductions planned.