Written evidence submitted by Sarah Simpson
BA(Hons) MSc(Eng) CEng FCIHT (RSM0074)

Executive Summary

1                        Introduction

1.1.1          In June 2019, a road collision occurred on the M1 All Lane Running (ALR) smart motorway in South Yorkshire, in which two people died; Mr Mercer and Mr Murgeanu. The solicitors for Mr Mercer’s family since instructed me, as a transport planning expert at Royal HaskoningDHV with some 19 years’ experience, to prepare a report to review the design development and implementation of smart motorways. As a result, an extensive review has been undertaken of the historical design development of smart motorways, hazard risk management processes, the evidence and data underpinning the 2020 Smart Motorways Evidence Stocktake and Action Plan, and the means by which smart motorway roads have been implemented in countries elsewhere in the world.

1.1.2          The subsequent report, entitled Independent Review of All Lane Running Motorways in England (“the Review”), was submitted to Highways England and the Secretary of State for Transport in March 2021. A copy was provided to the Transport Select Committee, and to the Office for Rail and Road (ORR) at the latter’s request.

1.1.3          This evidence now submitted to the Transport Committee under its call for evidence provides a summary of the key findings from that research with regards to four of the six questions asked by the Committee. The Review provides transparent and comprehensive details on the findings summarised here.

Q1. The benefits of smart motorways, for instance to reduce congestion on busy sections of motorway, and how necessary they are.

1.1.4          The development of smart motorways was in response to a perceived need to increase road capacity on the SRN using a ‘predict and provide’ model. By contrast, there has been a long-established understanding that building new roads induces traffic and that, at some point, a shift to a ‘decide and provide’ model will be more appropriate. For this reason, and given the material change in travel patterns which have resulted from Covid-19 and which are forecast to persist for some time[1], the basis on which the current road strategy is predicated is undermined and the extent to which smart motorways are necessary is weakened.

1.1.5          The evidence presented in the 2020 Evidence Stocktake and Action Plan[2], shows that Controlled Motorways are the safest form of smart motorway. By virtue of controlling and regulating traffic speeds and lane usage, Controlled Motorways also increase capacity by reducing headway between vehicles on the running lanes unless or until traffic levels increase to such a degree that all available road space is in use.

1.1.6          I note that ALR provides additional physical road space in addition to the provision of the active traffic management. However, this clearly does not reduce congestion in perpetuity in the absence of a mode shift strategy, rather it merely delays the point at which all road space is in use and congestion is again a common user experience. For this reason, smart motorways in Germany and the Netherlands have so far been largely used as temporary measures prior to more traditional road building schemes. In terms of the English roads strategy, the natural consideration relates to what comes after smart motorways.

1.1.7          In the absence of a detailed multi-modal strategy for strategic inter-city movements, I am led to assume that more traditional road building is anticipated despite legislative obligations to reduce carbon emissions and to meet climate change targets. However, I would submit that a ‘decide and provide’ approach is warranted now and that the changes in commuting travel patterns that are forecast on motorways over the coming years as a result of the Covid-19 pandemic, will provide the rare opportunity to do so.

1.1.8          In conclusion: there are some benefits of smart motorways in relation to easing congestion and providing enhanced capacity. However, these benefits come at the cost of reduced comparative safety in dynamic hard shoulder (DHS) schemes, and the lowest level of safety in ALR. By increasing road space they also induce more traffic. The recent changes in travel patterns weaken the need for increased road space, arguably reducing the necessity for DHS and ALR schemes with their additional physical road space.

Q2. The safety of smart motorways, the adequacy of safety measures in place and how safety could be improved.

1.1.9          As is established in the Review, I am in no doubt that the ALR has the lowest level of intrinsic safety than any form of motorway and certainly the lowest level of any type of smart motorway. Comparing the data for fatalities or serious injuries (KSI) on roads before and after the implementation of ALR, the Evidence Stocktake shows that the KSI rate increases not only compared with the predecessor roads, but also with the counterfactual. While there are some subtleties to this data (including the change in collision reporting from STATS19 and the fact that there is less than three years’ worth of post-opening data available for these ALR schemes) these nuances only reinforce the need for a precautionary approach to decision making in relation to ALR schemes.

1.1.10     This is supported by the adoption of Safe Systems by Highways England, and the overarching safety objective to reduce the numbers of people who are killed or seriously injured on the SRN. In summer 2020 a report for the ORR found significant shortfalls in the way that Highways England has implemented Safe Systems. This latter report also found that improvement schemes are consistently identified on the basis of capacity and not on the basis of improving safety.

1.1.11     The lack of consideration of the safety of the ALR environment is explained by the 2015 comparison of ALR with all purpose trunk roads (APTR). While I agree in part with the comparison between ALR and APTR with regards the removal of the hard shoulder, it is my opinion that these roads are of two distinct and different orders reflecting the weight of traffic and strategic hierarchy of the routes concerned. This is borne out by motorways’ designation as Special roads, with limits on the types of vehicle that can use them, the absence of at-grade pedestrian crossings and no Public Rights of Way traversing them. Thus these road types are not directly comparable. I would assert that outcomes on one are not fully generalisable onto the other.

1.1.12     A more fundamental consideration, however, is whether our road infrastructure should be required to be intrinsically safe as per the Safe Systems approach. One of the reasons smart motorways elsewhere in the globe use dynamic hard shoulders rather than ALR, is that the DHS system is more readily fail-safe. In the event of a system failure, the hard shoulder is physically still there, and the default is that it cannot be used as a running lane. By contrast the UK’s ALR has a distinctly reduced level of intrinsic safety compared with conventional motorways, Controlled Motorways, and DHS due to the absence of a permanently present, continuous place of relative safety along the carriageway’s length. There is therefore a heavy reliance on imperfect technology to mitigate these risks, but even once these are implemented, there will continue to be less mitigation than is used for example on Australian ALR schemes.

1.1.13     I have found that a traditional approach to road safety has been taken throughout smart motorways’ design development, with a mathematical balance being sought across all risks. In practical terms, the risk of being in a live lane breakdown increases by 216 percent in an ALR scheme over and above the level of risk for the same hazard on a standard motorway. However, instead of addressing this risk in physical design terms, the risk assessment approach used the reduced risks associated with other elements of ALR to mathematically offset this risk. It is unfortunate that this approach has persisted, despite the findings of multiple Transport Select Committee and All-Party Parliamentary Group inquiries which have repeatedly raised substantive concerns relating to road safety in smart motorways.

1.1.14     Thus the dialogue and debate regarding smart motorways is ill-informed and overly reliant on reduced numbers of slight collisions resulting from using variable mandatory speed limits (VMSL) being used to offset, what is in multiple cases, an increase in serious or fatal collisions. This offset is embedded in the safety objective for ALR which, unlike that for Controlled Motorways, has until very recently required only that the ALR road is no worse in safety terms than its predecessor road. This mathematical risk management approach accords with the traditional approach and contrasts with the Safe Systems’ focus on eliminating KSIs. If the SRN is truly subject to a Safe Systems approach, then it could be reasonably expected that any decision to change that network should be based on reducing KSIs. In undertaking the Review, I have found no evidence that this is the case for the decision to increase the miles of ALR motorways on the SRN. Indeed, the type of smart motorway which is associated with the highest rate of KSI casualties has been identified as the preferred and only option for future implementation. 

1.1.15     It is clear from the data that people are more likely to be involved in the most dangerous of hazards on ALR schemes: a live lane incident. When this happens, people are more likely to die or be seriously injured as a result. There is abundant evidence to show that the real-world implications of this approach to risk assessment have been understood by Highways England for some time. However, the Review found that decisions made in developing design criteria for smart motorways have consistently weakened and compromised their intrinsic level of safety and removed prompts for designers to consider foreseeable behaviour on these roads.

1.1.16     There are also concerns identified in 2016 research that MIDAS does not work in the way that has been previously anticipated, and Stationary Vehicle Detection technology identifies only 65 percent of stopped vehicles in a live lane[3]. It is concerning that technology that is being relied upon to make the ALR safer is not as reliable as it could be.

1.1.17     Throughout the design development and risk management processes for ALR, I have consistently found that peak period live lane breakdowns have not received attention due to the assumption that MIDAS will provide queue protection. MIDAS technology does not reliably provide the required mitigation and even when it does, this mitigation is negated by human behaviour for which enforcement has only recently been legislated. As a result, ALR has a higher rate of KSI collisions than predecessor roads. It is telling that more people are killed or seriously injured on ALR in peak period, congested conditions when most controls are in place, than in off-peak conditions which are typically associated with higher speeds. In the absence of evidence of the expected risk management outcomes of implementing or improving these technologies, it is unclear how ALR can be designed to ensure fewer people die or sustain life-changing injuries. 

1.1.18     It is my opinion that it is beyond doubt that the removal of the hard shoulder in a motorway environment results in an intrinsically less safe road than any other form of smart motorway. There is evidence to show that this has been understood since at least 2015 and this has been demonstrated most recently in the evidence associated with the 2020 Evidence Stocktake. ALR is not a maximally safe road system, and risk management and assessment methodologies have not resulted in effective mitigation of the hazards resulting from the removal of the hard shoulder.

1.1.19     To improve safety of our smart motorways, ALR should be reviewed in detail across all five Pillars of the Safe Systems approach. A decision could made quickly to use the Controlled Motorway system until the ALR is reviewed, with the variable mandatory speed limits and variable message signs on ALR potentially useful to expedite any change.

1.1.20     In conclusion: ALR has the lowest level of intrinsic safety of any type of motorway and certainly of any form of smart motorway. The technology which is being implemented to make these roads safer is unreliable and the forecast reduction in KSIs is not evidenced. The only means by which these roads can be made safer is for the Safe Systems approach to road safety to be fully brought to bear in reviewing English smart motorways.

Q3. Whether All Lane Running is the most suitable type of smart motorway to roll out or if there are better alternatives.

1.1.21     It is clear from the evidence in the Review that ALR is the least safe form of smart motorway. The safest form, Controlled Motorways, should be meaningfully revisited as part of the Roads Strategy going forward. This is due to the well-evidenced and substantial safety benefits associated with Controlled Motorways which also, because of the active traffic management technology in place, provide journey time and congestion management improvements. 

1.1.22     In the event that any form of smart motorway other than the Controlled Motorway, is identified as the preferred option, it is imperative that the design criteria are reviewed on the basis of the Safe Systems approach. This review should be transparent to increase public confidence and the outcome must be that a substantial and demonstrable safety benefit can be immediately identified as a result of its implementation, specifically in relation to KSI outcomes.

Q4. Public confidence in using smart motorways and how this could be improved.

1.1.23     There is a long-established, codified practice of public consultation to inform decision making large and small. Given that local highway authorities frequently consult on new bus stops, and Highways England consults on major changes, such as the A303 at Stonehenge, it is my opinion that the public reasonably expects meaningful consultation for proposals which affect material change to the transport infrastructure they use.

1.1.24     Based on my reading, aside from the M42 pilot and despite the need for strong public engagement being identified as long ago as 2009, this has not been provided in relation to smart motorways. Indeed, based on my review of the available literature, it is my opinion that a meaningful consultation has never taken place with the public or stakeholders in relation to smart motorways. As a result, it was foreseeable (and foreseen in documents associated with early smart motorway design) that the public would have a poor understanding of smart motorways and their associated hazards. That this is now the case is well evidenced by work undertaken by Highways England[4], Transport Focus[5] and others[6] 

1.1.25     As the Highway Code has not fully incorporated smart motorways, drivers have never been educated in their use as part of the licencing system[7]. Added to this the fact that there have been seven iterations of smart motorway design guidance in the past 12 years, it would be reasonable to believe that drivers may not be entirely confident in using smart motorways in general. 

1.1.26     It is clear that there is little understanding of the risks inherent to ALR smart motorways, and specifically those relating to the removal of the hard shoulder and live lane breakdowns. Safe Use of roads infrastructure is a core part of the Safe Systems approach and therefore public campaigns and driver education has a clear role in enabling road safety. However, as the ORR review identified in 2020, there is an over-reliance on public campaigns to increase road safety. As a result, driver information cannot be to the exclusion of specific consideration of other measures and, in particular, the roads’ design.  

1.1.27     The development of smart motorways was in response to a perceived need to increase road capacity on the SRN using a ‘predict and provide’ model which remains evident as recently as RIS2. By contrast, there has been a long-established understanding in the transport planning profession that building new roads induces traffic and that, at some point, a shift to a ‘decide and provide’ model will be more appropriate.

1.1.28     Given the current safety levels of the existing smart motorways, as well as the limited efficacy of the technology they rely on, I would submit that a ‘decide and provide’ approach is warranted now and that the changes in commuting travel patterns on the SRN that are forecast over the coming years, will provide the capacity and opportunity to do so.

1.1.29     Overall, with respect to the existing context, I conclude that the development of smart motorway design standards and guidance was carried out over a period in which there were fewer obligations in relation to Climate Change, the Safe Systems approach to road safety was not embedded in Highways England, there were no coroner verdicts relating to smart motorways, and the world was untouched by Covid-19. Given the wider context in which the SRN now operates it is my view that a fresh approach on the continued implementation of smart motorways would be timely and, given the evidence, necessary to reduce avoidable deaths.

1.1.30     In conclusion: the public rightly trusts designers and highway authorities to design with safety in mind as evidenced by Transport Focus’s qualitative research[8]. It is my view that the public’s confidence will be more likely gained with greater transparency in decision making, continued and greater use of driver information campaigns, and a more clear focus on reducing the numbers of people who are killed or seriously injured throughout the design and implementation of smart motorways

 

April 2021

 

Endnotes


[1] https://www.creds.ac.uk/publications/report-at-a-crossroads-travel-adaptations-during-covid-19-restrictions-and-where-next/

[2] https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/936811/smart-motorway-safety-evidence-stocktake-and-action-plan.pdf

[3] https://s3.eu-west-2.amazonaws.com/assets.highwaysengland.co.uk/Knowledge+Compendium/2016-17/Stationary+Vehicle+Detection+System+(SVD)+Monitoring.pdf

[4] https://s3.eu-west-2.amazonaws.com/assets.highwaysengland.co.uk/specialist-information/knowledge-compendium/2014-2015/M25+J23-27+SM+ALR+Monitoring+12+Month+Evaluation+Report_v2.0_Final.pdf

[5] https://d3cez36w5wymxj.cloudfront.net/wp-content/uploads/2020/12/04163914/All-lane-running-smart-motorways.pdf

[6] https://www.brake.org.uk/files/downloads/Reports/Direct-Line-Safe-Driving/Motorways-Green-Flag-Safe-Driving-Report-2020.pdf

[7] I have made a submission to The Highway Code consultation, in advance of its closing date, as instructed by the Mercer family’s solicitors.

[8] https://d3cez36w5wymxj.cloudfront.net/wp-content/uploads/2020/12/16171410/All-lane-running-agency-report.pdf