The System of Bus Safety Governance in Britain

A detailed evidence-led examination of fragmentation, accountability, transparency and the missing national framework

1. Introduction: the safety system the public assumes exists

Every day, across Britain, millions of people step onto buses with a quiet but reasonable assumption: that behind that service there is a coherent safety system.

They assume that incidents are recorded. They assume trends are monitored. They assume serious events are escalated. They assume lessons are learned. They assume drivers, passengers and the public are protected not only by individual operators, but by a wider system of oversight, governance and accountability.

They assume that if bus services are publicly funded, publicly specified, franchised, contracted, regulated, licensed or delivered under public transport policy, then there must be a clear framework behind them.

But when you start asking direct questions about that system, the picture becomes deeply uncomfortable.

Over recent months, I have used Freedom of Information requests, internal reviews, authority correspondence, published reports, regulatory responses and wider transport policy evidence to ask one central question:

How is bus safety actually governed in Britain?

Not simply how many collisions are recorded.
Not simply whether operators have internal procedures.
Not simply whether vehicles are inspected.
Not simply whether the police record road traffic collisions.

The question is bigger than that.

Who holds bus safety data?
Who standardises it?
Who publishes it?
Who reviews it?
Who identifies patterns?
Who learns lessons from serious incidents?
Who monitors passenger falls, assaults, near misses, fatigue and operational risk?
Who makes sure safety is not hidden behind commercial sensitivity?
Who ensures workers have a safe route to raise concerns?
Who owns the national picture?

The answer emerging from the evidence is stark.

There is no single coherent system of bus safety governance in Britain.

Instead, the evidence points to a fragmented and uneven landscape. In some areas, safety systems exist but are immature. In others, they are still being built. In some places, the authority does not hold the data. In others, information exists but is withheld. Some safety material is embedded in commercial franchising structures. Some national material is spread across multiple teams. Some policy development is happening behind closed doors. In several cases, responsibility is split between operators, transport authorities, regulators, government departments and police systems.

This is not a minor administrative problem.

It is a structural problem.

And it matters because bus reform is accelerating. Franchising, Enhanced Partnerships, rural pilots, devolution, combined authorities, integrated brands and public control models are all advancing. But the evidence shows safety transparency is not advancing at the same pace.

That is the central political issue.

Public control is increasing. Public accountability is not keeping up.

2. The national layer: DfT, STATS19 and the absence of a single national picture

The Department for Transport should be the first place to look for a national view of bus safety governance.

If there is a coherent national system, DfT should be able to explain it. It should be able to describe who collects what, how data is standardised, how risks are escalated, how authorities are expected to report, how lessons are learned and how bus safety is monitored across the country.

But the FOI evidence shows something very different.

DfT responses to multiple bus safety requests have aggregated numerous requests and refused them under section 12 of the Freedom of Information Act because compliance would exceed the cost limit. The important point is not just the refusal. It is the reason given. DfT explained that the requests covered closely related aspects of bus safety and would require searches across a wide range of policy teams, records and systems. That means the information is not sitting in one central system as a coherent national safety picture. It is distributed across the Department.

A further DfT response on rural bus franchising and reform pilots showed the same pattern in a wider bus policy context. The request concerned rural franchising, reform pilots, legal advice, policy analysis and barriers to franchising for non-mayoral authorities. Again, DfT said the material would require extensive searches across a wide range of teams.

That widens the concern. The fragmentation is not limited to safety data. It runs through bus policy and reform more broadly.

Then there is STATS19.

STATS19 is often treated as the national road safety dataset. It is useful, but it is not a bus safety governance system. It is a police-recorded road collision dataset. It tells us about reported injury collisions after the event. It does not provide an operational safety-management framework for buses.

STATS19 does not properly capture driver fatigue, duty design, excessive running pressure, toilet access, inadequate facilities, passenger falls not linked to police-recorded collisions, internal operator incident reports, non-collision injuries, near misses, verbal abuse, assaults not recorded as road collisions, or safety governance within franchising and Enhanced Partnership arrangements.

That distinction is crucial.

STATS19 counts some outcomes after the event. It does not govern bus safety.

This matters because the bus sector appears to sit in a gap between road safety statistics and public transport safety governance. Rail has a much more developed safety culture, including independent investigation and structured safety reporting. Aviation has a mature accident investigation and reporting culture. Buses, by contrast, are governed through a patchwork of road traffic law, operator licensing, vehicle compliance, local transport contracts, operator-held data, police collision data and authority-specific arrangements.

That is not a joined-up national safety system.

It is a set of disconnected parts.

3. The regulatory layer: many bodies, no whole-system owner

One of the most important findings from this work is that bus safety responsibility is split across multiple institutions.

The Traffic Commissioners regulate PSV operator licensing and local bus service registration. They can call operators to public inquiry, consider good repute, take action on operator conduct and impose regulatory consequences. That is important. But the Traffic Commissioners do not operate a public national bus safety data system. They are not the body publishing a national dashboard of bus passenger falls, assaults, near misses, safety-critical events, fatigue risk or authority-level safety performance.

DVSA enforces vehicle and operator compliance. It undertakes inspections, investigations, roadside checks and operator compliance activity. Again, that is vital. But DVSA compliance is not the same as a transparent public bus safety governance framework covering the whole passenger and workforce safety picture.

DVLA licenses drivers and vehicles. It is part of the wider road transport system, but it does not govern bus operational safety performance.

Police forces record collision data and enforce road traffic law, with national coordination through policing structures including the National Police Chiefs’ Council. But police-recorded collision data is not the same as continuous bus safety governance. It is incident-based, retrospective and limited by the scope of what is reported and recorded.

The Office of Rail and Road adds another layer. ORR has deep safety expertise from rail and has been involved in policy discussions relevant to cross-modal safety learning. Its internal review response confirmed that information within scope had been identified relating to engagement with DfT concerning possible future policy options on road safety, but the remaining material was withheld under section 35 because it related to government policy formulation. ORR recognised the public interest in transport safety and transparency, but concluded that policy development space should be protected.

That response is significant because it shows that cross-modal safety policy thinking exists, but the public cannot yet see it.

Taken together, the regulatory picture is not one of total absence. There are bodies with important roles. But there is no visible whole-system owner.

A collision may sit in STATS19.
A vehicle defect may sit with the operator or DVSA.
Operator conduct may reach the Traffic Commissioner.
Driver licensing sits with DVLA.
A passenger fall may sit in an operator incident system.
An assault may sit with police, operators or local ASB systems.
A franchising safety requirement may sit inside a commercial contract.
A safety governance framework may sit within a combined authority programme board.
A policy discussion may sit inside DfT or ORR and be withheld while policy develops.

Each piece has a home.
The whole picture does not.

That is the structural weakness.

4. Greater Manchester and the Bee Network: the flagship model under pressure

Greater Manchester matters because the Bee Network is the most advanced franchised bus system outside London. It is the model often pointed to as the future of bus reform in England.

That is why its safety transparency matters so much.

If the flagship cannot deliver visible, timely and robust safety transparency, then the rest of the country needs to pay attention.

The evidence shows that safety data publication has been delayed. There was a public expectation that Bee Network accident and incident safety data would be published within around 12 months of launch. That original expectation has not been met, and publication is now expected later, after the fully franchised network has been operating for a longer period.

That is not a minor detail. It goes to the credibility of public control.

Greater Manchester also exposes a split between GMCA and TfGM. GMCA has the political and governance role. TfGM holds transport operational data and manages the Bee Network. That distinction may make sense institutionally, but it creates a problem for public accountability. If governance sits in one place and operational data in another, who owns safety transparency? Who owns publication? Who is responsible for the promise made to the public?

The TfGM FOI response is important because it shows that safety systems do exist. TfGM released or referred to safety KPI summaries, operator submissions, the Bee Network Safety Plan, the Bus Operational Risk Register, accident and incident forms, serious health, safety and environmental reporting templates, HSE Bus Operator Forum minutes, hotspot meeting minutes and Executive Board material.

That is useful evidence of a developing safety framework.

But the same response also reveals serious limitations. TfGM confirmed that prior to January 2025, safety incident reporting by individual operators was undertaken on an ad hoc and goodwill basis before being formalised with Tranche 3. It also confirmed incomplete data for some operators, and that serious incident reporting using the SHSE form only began in November 2025. At the time of the request, official plans to publish KPI safety data supplied by operators were not in place. Operators were also not obliged to provide compliance audits under the current franchising agreement.

This matters because Greater Manchester is not an undeveloped authority. It is the flagship.

And yet the evidence shows:

Safety reporting has moved from ad hoc/goodwill arrangements to more formal reporting only during the development of the franchised system. Historic data is incomplete. Operator coverage has gaps. Serious incident reporting was formalised late. KPI publication planning was not in place at the time of the request. Compliance audits were not required in the way one might expect from a mature public-control safety regime.

That does not mean Greater Manchester is doing nothing. Clearly, it is doing a lot. But it does mean the system is still maturing.

Then there is GMCA’s section 14 vexatious refusal. GMCA refused a further request, citing the volume and repetition of requests and explaining that Bee Network bus safety data and operational safety information were most likely held by TfGM.

That response is important because it exposes the accountability split.

If GMCA says TfGM holds the data, and TfGM holds the operational systems, then what exactly does GMCA hold in relation to safety governance, public commitments, assurance and political oversight? If the public asks who is responsible for safety transparency, where should they go? GMCA? TfGM? The Mayor? The Bee Network Committee? Operators?

The danger is that accountability gets passed around the system.

Greater Manchester therefore gives us one of the most important lessons of the entire investigation:

Public control must mean public accountability.

A publicly controlled bus network without public safety data is not genuine reform.

5. West Yorkshire and the Weaver Network: when safety governance becomes commercially sensitive

West Yorkshire is one of the strongest and most concerning examples because it shows a different problem: not absence, but restricted visibility.

The WYCA internal review confirmed that material within the scope of the request existed, but upheld withholding under section 43(2), commercial interests. The authority said that the majority of material was not standalone safety documentation, but was embedded within the commercial, operational and governance architecture of the live Bus Franchising Programme. It said disclosure could reveal commercially sensitive programme design, internal risk frameworks and elements of procurement, mobilisation and operator oversight.

That is a major admission.

It means safety governance exists, but is embedded in franchising and procurement architecture. It is not being treated as a standalone public accountability framework.

That raises a fundamental public-interest question.

If bus services are publicly funded, publicly specified and moving towards public control, why is the framework for governing safety not publicly visible?

This is not about asking for commercially sensitive bid prices, personal data, live security vulnerabilities or confidential operator submissions. The request was about governance: roles, responsibilities, escalation, assurance, reporting and transparency.

That should be visible.

Safety governance should not be hidden behind commercial sensitivity.

West Yorkshire is important because it shows how safety can become wrapped inside commercial programme structures. As more authorities move into franchising, procurement and mobilisation, there is a real risk that safety governance becomes harder to access precisely at the moment public accountability should be getting stronger.

That is the wrong direction.

If reform creates more public control but less visible safety governance, then reform is not delivering the accountability the public was promised.

The core line is simple:

Safety data is public-interest information, not a commercial secret.

6. Liverpool City Region: building the system during mobilisation

Liverpool City Region adds another category: safety governance not yet finalised.

The response confirmed that no finalised recorded information was currently held within scope relating to the development of bus safety governance arrangements, safety performance monitoring for franchised services, incident reporting or escalation arrangements, or safety benchmarking and assurance frameworks. It also confirmed that work was ongoing as part of Tranche 1 mobilisation, with relevant information expected later.

That is important because mobilisation is not a distant concept. Mobilisation is when the operating system is being built.

If safety governance, incident escalation, performance monitoring and benchmarking frameworks are still being developed during mobilisation, the immediate question is: what is the baseline and what governs safety in the transition?

This is where pre-franchise safety baselines become essential.

Authorities moving into franchising must publish a baseline before the system changes. Without a baseline, there is no credible way to prove whether reform improves safety. If passenger falls, assaults, near misses, collisions, serious injuries and workforce safety indicators are not measured before reform, any future claim of improvement will be weaker.

Liverpool therefore proves a wider national point.

Every authority moving into franchising should publish a pre-franchise safety baseline and a safety governance framework before services transfer.

Otherwise, safety transparency is being built after the political decision has already been taken.

7. East Midlands Combined County Authority: responsibility before readiness

The East Midlands Combined County Authority is politically important because it is newly established and has taken on transport responsibilities across Derby, Derbyshire, Nottingham and Nottinghamshire.

The FOI response is one of the clearest pieces of evidence in the whole investigation. EMCCA confirmed that it held no information on draft KPIs for monitoring bus service contracts, no documents considering safety-related performance indicators, no documents describing how driver fatigue, operational safety performance or incident reporting may be reflected in contract monitoring, and no examples or case studies reviewed from other franchised bus systems. It explained that EMCCA is still in transition after taking over transport authority responsibilities on 1 February 2026 and is identifying and progressing work that needs to be done.

This is not a small gap. It is a readiness issue.

An authority has taken on transport responsibilities, but does not yet hold information on how safety will be monitored through KPIs, how fatigue and incident reporting may be reflected in contracts, or how other franchised systems have been benchmarked.

That does not mean staff are not working hard. It does not mean no one cares. But it does show that responsibilities can transfer before safety governance is visibly ready.

This is exactly why national standards are needed.

A new authority should not have to design bus safety governance from scratch. DfT should provide a minimum safety transparency framework. Every new authority should know from day one what safety data should be collected, what should be published, how operators should report, how risks should be escalated, how workers should raise concerns and how the public can scrutinise performance.

EMCCA has a chance to build this properly. But the FOI evidence shows why leaving each authority to invent its own model is dangerous.

Safety transparency must be built in from day one, not bolted on later.

8. South Yorkshire: public ownership must include safety accountability

South Yorkshire is another major piece of the developing national picture. The region is moving towards franchising and a “People’s Network” model, with major structural changes including public ownership of depots.

That matters because public ownership of assets is not only an economic or operational issue. It has safety implications.

Depots affect maintenance, charging infrastructure, vehicle reliability, staff facilities, operational resilience and fleet transition. Franchising affects scheduling, contract monitoring, operator requirements, incident reporting and public accountability. A public network model should strengthen safety governance, not leave it unclear.

South Yorkshire therefore raises key questions that should be answered before the system goes live.

Will South Yorkshire publish a pre-franchise safety baseline?
Will it publish safety KPIs once franchising begins?
Will passenger falls, assaults, near misses and serious incidents be reported?
Will driver fatigue and welfare be monitored as safety issues?
Will unions have a formal role in safety trend review?
Will independent confidential reporting be available for bus workers?
Will safety governance be public-facing, or embedded inside contract documents?

South Yorkshire could become a positive model. It has the opportunity to learn from Greater Manchester, West Yorkshire and Liverpool before its own system fully matures.

But the lesson must be acted on early.

Public ownership must not stop at depots and contracts. It must include visible public safety accountability.

9. West Midlands: contracts are not enough

The West Midlands evidence shows progress, but also an incomplete system.

WMCA/TfWM has indicated that minimum contractual safety reporting obligations are expected in future franchising contracts. That is positive. It means safety reporting is being considered as part of the contractual framework.

But there are still major gaps.

Independent confidential reporting, comparable to CIRAS in rail, has not yet been developed in detail. Public reporting arrangements are not yet finalised. The role of trade unions in reviewing trends and shaping safety governance is not yet secured as a permanent feature.

This distinction matters.

A contract can require an operator to send data. But a safety governance system must do much more than receive data.

It must analyse the data.
It must identify trends.
It must publish what matters.
It must involve workers.
It must protect those who raise concerns.
It must learn from incidents.
It must act before harm occurs.

West Midlands therefore demonstrates an emerging contractual safety approach, but not yet a full public safety governance model.

The warning is clear: do not mistake contractual reporting for safety governance.

Reporting is the start. Governance is what happens after the report arrives.

10. Cambridgeshire and Peterborough: safety recognised, but paused

Cambridgeshire and Peterborough provides another important case because it shows an authority that has identified safety as important, but where wider reform uncertainty has paused the work.

CPCA confirmed that it does not currently operate an Enhanced Partnership and is progressing towards franchising. Safety had been identified as a key part of workstreams as the authority moved towards franchising, but those workstreams and budgets were paused due to the Independent Review into Bus Implementation Pathways.

Board-level material is particularly important. It recognised safety, referenced Manchester as a lesson where safety was initially less of a focus, and acknowledged that responsibility currently lies with operators while the authority must embed safety into future workstreams such as procurement.

This is significant because it shows the lesson from Greater Manchester is already known.

If safety was initially less of a focus in Manchester, then every other authority should be required to correct that mistake before launch.

But CPCA also shows the fragility of safety workstreams. When the reform pathway is paused, safety governance can be paused with it. That should not happen.

Safety governance should not depend on which reform model is chosen. Whether an authority pursues franchising, an Enhanced Partnership, a rural pilot or another model, safety transparency should be mandatory.

Passengers and workers should not have to wait for institutional certainty before safety governance exists.

11. North East Combined Authority: reform without authority-held safety data

The North East Combined Authority has an active Enhanced Partnership and is undertaking a Franchising Scheme Assessment. However, NECA also stated that it does not hold bus safety data.

That contrast matters.

An authority can be actively considering or progressing reform while not holding a central bus safety dataset.

NECA also raised commercial NDAs and legal privilege in relation to franchising assessment material. That again shows the risk of safety-relevant governance material becoming entangled with commercial or legal protections.

The key distinction is between safety data and safety governance.

An authority may not hold raw incident data if operators hold it. But it may still hold governance documents, assurance arrangements, escalation protocols, risk registers, board papers, contract expectations or reform workstreams dealing with safety.

The North East example therefore demonstrates why FOI questions must be framed carefully. Asking for “safety data” may lead to an answer that the authority does not hold it. Asking for “safety governance” may reveal whether the authority is overseeing, requiring or planning anything in relation to that data.

The wider issue remains the same:

If operators hold the data, but authorities hold the reform powers, who brings the two together?

12. Hull and East Yorkshire: operational management is not safety governance

Hull provides one of the clearest local examples of the difference between operational coordination and safety governance.

The Hull FOI response confirmed that passenger injury documents were not held and that bus safety data was held by individual bus companies. The disclosed Bus Operators and Interchange minutes show regular coordination on operational matters: roadworks, infrastructure, bus stops, shelters, lighting, potholes, signage, CCTV, interchange issues, real-time information, events and delays.

There are occasional safety-related references, including risk assessments, evacuation arrangements, reversing signage, ASB spreadsheets and CCTV. But those references do not amount to a structured safety governance framework.

The Bus Alliance minutes show efforts to create governance structures, including subgroups and operational arrangements. But they also reveal data gaps, requests for outstanding performance data and references to needing clearer structures.

That is the point.

Hull shows a system that is being managed day to day. Meetings happen. Operators talk to the council. Infrastructure issues are discussed. Problems are logged. But that is not the same as a public-facing safety governance system.

A true safety governance system would show centralised authority-level data, consistent incident definitions, safety KPIs, trend analysis, escalation arrangements, published performance, lessons learned and public accountability.

Hull therefore asks a question that applies across Britain:

Is the bus system being managed, or is it being governed for safety?

Those are not the same thing.

13. West of England: the authority’s own audit identifies the fragmentation

The West of England evidence is especially powerful because it includes a commissioned Bus Passenger Safety Audit. This is not just an FOI response. It is a professional report produced for a public authority.

The WSP audit examined passenger safety across the whole bus journey: walking routes to stops, waiting environments, bus stations, staffing and procedures, customer relations and information provision. It recognised that safety is shaped by lighting, infrastructure, anti-social behaviour, harassment, gender, disability, ethnicity, age, socioeconomic status, personal experience and accessibility.

The data in the report is striking. WSP research cited in the audit found that 82% of women felt unsafe due to poor lighting, 79% felt unsafe or very unsafe using public transport at night, 62% felt unsafe waiting at a bus stop, and 52% felt unsafe walking to a bus stop or station. Local survey findings showed that 35% felt unsafe or very unsafe using buses at night, 15% felt unsafe waiting at a bus stop and 18% felt unsafe walking to a stop or station.

But the most important part for this investigation is the governance diagnosis.

The report identified lack of clarity over roles and responsibilities, silos between stakeholders, inconsistent training, disjointed approaches between operators and police, inconsistent incident classification, lack of centralised data, and absence of systematic monitoring and evaluation.

That is the national problem in one document.

This is not speculation from outside the system. This is a commissioned audit identifying fragmented roles, weak data structures and inconsistent reporting.

West of England proves that safety governance must be broader than collisions. Passenger safety includes the first and last mile, waiting environments, lighting, harassment, reporting confidence, staff training and accessibility. If bus safety governance ignores those areas, it is incomplete.

The West of England audit should be read nationally.

It shows what a real whole-journey safety framework must include.

14. Transport for Wales: data exists, but the system is partial and unpublished

Transport for Wales shows another form of fragmentation: partial capability without full public transparency.

TfW confirmed that it receives safety data from contracted Traws Cymru operators only, including collision data, passenger falls or near misses and assaults on staff. It also uses telematics equipment to monitor driver behaviour.

The KPIs include collisions per million miles, passenger incidents per million miles and driver telematics scores. The telematics system can monitor acceleration, braking, cornering, lane change handling, speeding, fatigue, smoking, mobile phone use and distracted driving.

That is potentially powerful information. It goes far beyond what STATS19 can show.

But TfW also confirmed that bus safety data is not currently published, summarised or reported publicly. It confirmed that monthly safety reporting began in January 2025, with no historic data before that other than telematics.

So the Welsh evidence shows a mixed picture.

Data exists. KPIs exist. Telematics exists. But the system applies to contracted Traws Cymru operators, is not a full all-Wales public bus safety governance system, and is not publicly reported.

That proves another point.

Technology alone does not equal transparency. Data collection alone does not equal governance. A dashboard that the public cannot see is not public accountability.

15. Northern Ireland and Translink: centralisation does not guarantee transparency

Northern Ireland is structurally different from the deregulated and franchising landscape in much of Britain. Translink operates as a central public transport provider, which might suggest a more coherent safety system.

But the evidence shows that centralisation alone does not guarantee transparency.

Translink’s response indicated that the information requested was not stored in a centralised, electronically searchable form and that retrieving it would require a manual exercise exceeding the cost limit.

That is a major governance point.

If information exists but is not searchable, then the organisation may not be able to provide a clear public safety picture without significant manual effort. That affects public accountability, trend analysis and transparency.

A safety system must be able to see itself.

Central public operation may reduce some forms of fragmentation, but it does not automatically create a public-facing safety governance framework.

Northern Ireland therefore provides an important warning:

Centralisation is not enough.
Transparency must be designed into the system.

16. Scotland: policy ambition without a clearly visible unified safety system

Scotland has a distinct transport policy environment. Transport Scotland, local authorities and regional transport partnerships operate in a different legislative and policy context. Scottish transport policy has often been more progressive on public transport, climate, accessibility and modal shift.

But the same safety governance question still applies.

Where is the single visible system for bus safety governance? Where are standardised national bus safety indicators? Where is public reporting of passenger falls, assaults, near misses and safety-critical incidents? Where is the formal worker reporting route? Where is the bus equivalent of rail’s more structured safety culture?

The evidence considered so far does not show a clear, unified, publicly accessible Scottish bus safety governance system comparable to what would be expected in rail or aviation.

That does not mean nothing exists. It means the system is not publicly visible in a clear and standardised way.

Policy ambition is not the same as safety transparency.

Scotland should not be excluded from this debate. A national conversation about bus safety governance must include Scotland, because passengers and workers need clarity wherever they are in Britain.

17. Rural bus reform and pilots: the next frontier of fragmentation

Rural bus reform is critical because the future of bus governance is not only metropolitan.

Rural and county areas are increasingly part of the reform landscape through franchising discussions, rural pilots, Enhanced Partnership development, devolution deals and DfT reform work.

Relevant areas discussed in this work include Hertfordshire, Cornwall, Cumbria through Cumberland and Westmorland & Furness, Cheshire West and Chester, York and North Yorkshire Combined Authority, and other county or rural areas as DfT programmes evolve.

Rural safety governance is not simply a smaller version of urban safety governance. Rural bus services present distinct risks.

Routes can be longer.
Stops can be isolated.
Lighting can be poor.
Services can be infrequent.
Waiting times can be longer.
There may be fewer staffed locations.
There may be fewer safe places to report or seek help.
Passengers may be older, disabled, isolated or highly dependent on the bus.
Drivers may face longer stretches without facilities or support.
Toilet access, layover arrangements, fatigue and operational pressure may become even more significant.

If rural reform is rolled out without a safety governance framework, the national patchwork will simply spread into areas where visibility may be even weaker.

Every rural pilot should include:

a pre-pilot safety baseline,
standard incident definitions,
passenger fall reporting,
staff and passenger assault reporting,
near miss reporting,
driver welfare and fatigue indicators,
toilet and facility consideration,
infrastructure safety audits,
public reporting,
operator obligations,
and clear authority accountability.

Rural reform is a chance to get this right early.

But if safety governance is not built into the pilots from the beginning, rural reform risks becoming another experiment without a safety framework.

18. The national comparison: the pattern is undeniable

When the evidence is compared across authorities and systems, the national pattern becomes clear.

Greater Manchester has the most advanced franchised model, but safety transparency has been delayed and the system is still maturing.

TfGM has safety data, forums, KPIs and risk registers, but reporting was ad hoc before formalisation, historic data is incomplete and publication planning was not in place at the time of the request.

West Yorkshire holds safety governance material, but it is embedded in commercial franchising structures and withheld under commercial interests.

Liverpool City Region is developing governance during mobilisation, but no finalised framework was held at the point of response.

EMCCA has taken on transport responsibilities but does not yet hold information on safety KPIs, fatigue monitoring, incident reporting or benchmarking.

South Yorkshire is moving towards public control and asset ownership, but must still demonstrate visible safety transparency from the outset.

West Midlands expects contractual safety reporting, but independent reporting and public transparency are not yet fully developed.

Cambridgeshire and Peterborough has identified safety as a workstream, but work has been paused and current responsibility remains operator-led.

North East Combined Authority is progressing reform but does not hold bus safety data centrally.

Hull shows operational coordination but not a visible central safety governance system.

West of England’s own audit identifies unclear responsibilities, inconsistent reporting and lack of centralised data.

Transport for Wales has partial data and telematics, but no public reporting.

Translink shows centralisation without accessible searchable transparency.

Scotland shows policy ambition but no clearly visible unified public safety governance system.

Rural pilots risk expanding reform before safety transparency is defined.

DfT does not hold a single integrated national picture.

ORR/DfT policy development exists but is withheld during formulation.

Traffic Commissioners, DVSA, DVLA and police each hold part of the picture, but no single body holds the whole.

That is not a coherent system.

It is a national patchwork.

19. Why this has happened

This fragmentation has not happened by accident.

It is the result of how bus safety has been allowed to sit between different systems.

First, bus safety sits awkwardly between road safety and public transport governance. It has not been given the same clearly structured safety architecture as rail or aviation.

Second, data ownership and governance ownership are separated. Operators often hold incident data. Authorities hold contracts or policy responsibilities. Regulators hold licensing and enforcement powers. Police hold collision records. DfT holds policy. No single layer brings it together.

Third, deregulation left a legacy in which operator responsibility dominated. Franchising changes who controls the network, but unless safety governance is redesigned, old gaps are simply transferred into new structures.

Fourth, commercialisation creates opacity. As safety requirements become embedded in contracts, procurement and mobilisation, authorities may treat safety governance as commercially sensitive.

Fifth, national guidance is weak. There is no mandatory national framework requiring all franchising and Enhanced Partnership authorities to publish comparable safety data.

Sixth, worker voice is not structurally embedded. Drivers and other bus workers see fatigue, pressure, facilities, assaults, unrealistic schedules, near misses and safety problems first-hand. Yet formal trade union involvement in safety trend review is not universally required.

Seventh, public reporting is not standardised. Some authorities may publish more than others. Some may hold data but not publish it. Some may not hold it centrally at all.

That is why the system looks fragmented.

It was never designed as one system.

20. Why this matters politically

Bus reform is being sold as a new era.

Public control.
Integrated networks.
Better services.
Stronger accountability.
Improved planning.
Simpler ticketing.
Better branding.

But safety transparency must be part of that new era.

A franchised network without published safety data is not enough.
A new combined authority without safety KPIs is not enough.
A rural pilot without a safety baseline is not enough.
A contract without independent reporting is not enough.
A national dataset that only records police-reported collisions is not enough.
A governance framework hidden behind commercial sensitivity is not enough.
A meeting with officials is not a substitute for Freedom of Information.
A public brand is not the same as public accountability.

The public are being asked to support bus reform. Drivers are being asked to operate within restructured systems. Passengers are being asked to trust new networks.

Trust requires transparency.

If authorities take control of networks, contracts, fares, branding, routes and performance, they must also take visible responsibility for safety governance.

That is the political test of the new bus era.

Public control must mean public accountability.

21. What a real bus safety governance system should look like

A real system would start with national definitions.

Every franchising and Enhanced Partnership authority should report the same core safety categories:

collisions,
fatalities,
serious injuries,
minor injuries,
passenger falls,
staff assaults,
passenger assaults,
verbal abuse,
physical abuse,
near misses,
bridge strikes,
RIDDOR reports,
vehicle defects,
dangerous occurrences,
safety-critical incidents,
and serious operational safety events.

The data should be published routinely, at least annually and preferably quarterly for major transport authorities.

But numbers alone are not enough.

Authorities should publish narrative and action taken:

what trends are emerging,
what risks have been identified,
what action has been taken,
what lessons have been learned,
what changes have been made to contracts, infrastructure, training, policing, staffing, scheduling or facilities.

Every authority moving into franchising should publish a pre-franchise safety baseline. Without that, reform cannot be measured.

Every franchising contract should include mandatory safety reporting obligations, including definitions, timescales, escalation thresholds, audit rights and publication expectations.

Every bus worker should have access to independent confidential reporting, equivalent to CIRAS in rail, so that fatigue, pressure, unsafe practices, facilities issues, near misses and safety concerns can be reported without fear.

Trade unions should have a formal role in safety governance. Not token consultation. A real role in reviewing trends, identifying risks and shaping solutions.

DfT should issue national guidance on bus safety transparency and minimum reporting standards.

Safety governance should not be hidden behind commercial sensitivity. Where genuine security or personal-data concerns exist, redact them. But do not withhold the governance framework itself.

There should also be serious consideration of an independent Bus Accident Investigation Branch or equivalent function. Serious bus incidents should be investigated for systemic lessons, not just individual blame.

The bus sector carries millions of people. Bus workers perform safety-critical work. The sector deserves a safety governance model that reflects that reality.

22. What should happen next

The next stage should be a national reset.

DfT should convene a bus safety governance review involving combined authorities, local transport authorities, operators, trade unions, DVSA, Traffic Commissioners, police, passenger groups, accessibility groups and independent safety experts.

That review should map:

who holds what data,
who has legal responsibility,
what is currently published,
what is not published,
where data definitions differ,
how serious incidents are escalated,
how lessons are learned,
how worker concerns are reported,
how public accountability is maintained,
and where the system currently fails.

From that review, DfT should create a national bus safety transparency framework.

Authorities already operating franchising should publish safety data and governance arrangements without further delay.

Authorities developing franchising should publish safety governance plans before contracts go live.

Authorities running Enhanced Partnerships should publish how safety is monitored under those arrangements.

Rural pilots should include safety transparency as a core evaluation measure.

Traffic Commissioners and DVSA should publish clearer aggregate data on PSV safety enforcement trends.

Police collision data should be linked more intelligently with operator and authority safety data, without pretending STATS19 alone is enough.

The new bus era must not be built on old opacity.

23. Final conclusion: say it plainly

The evidence gathered through FOI, internal reviews, authority correspondence and published reports points in one direction.

Britain does not currently have a single coherent system of bus safety governance.

It has fragments.

Some are national.
Some are regional.
Some are local.
Some are operator-held.
Some are hidden in contracts.
Some sit in collision datasets.
Some are still being built.
Some are not held.
Some are not searchable.
Some are withheld.

That is not good enough.

Bus safety cannot be left to a patchwork of partial systems.

If reform is accelerating, safety transparency must accelerate with it.

If authorities want public control, they must accept public accountability.

If safety data exists, it should be published.

If safety governance exists, it should be visible.

If systems are not yet built, they should be built before reform goes live.

If workers see risks, they must have a protected route to report them.

If passengers are harmed, lessons must be learned openly.

And if the state is serious about bus reform, then it must be serious about bus safety governance.

The test of the new bus era is not just whether buses are branded differently, contracted differently or controlled differently.

The test is whether safety is governed differently.

Because a safety system that cannot be seen cannot be fully trusted.