Foreseeability Is Not Breach in Highway-Design Negligence: A Structured Evaluative Assessment Is Required

1. Introduction

Transport for London v Young & Anor [2026] EWCA Civ 1095 concerned two personal injury claims brought by experienced cyclists after night-time accidents on a stepped cycle track (“SCT”) on Farringdon Road forming part of Cycle Superhighway 6 (“CS6”), constructed by Transport for London (“TfL”).

Each cyclist collided with a raised kerb (a vertical upstand of about 5–6 cm) separating the carriageway from the SCT, having mistaken the kerb for a white line and not appreciating the SCT was raised above the carriageway. The claims were tried together in the County Court. The trial judge found TfL primarily liable, with substantial contributory negligence (65% and 70%).

TfL appealed on liability. The Court of Appeal treated the appeal as turning not on novel duty principles, but on whether the trial judge performed the legally required evaluative breach assessment, rather than moving too directly from foreseeability to negligence.

Key issues

  • The nature and scope of a highway authority’s common law duty where it has created a highway feature.
  • The role of “obviousness” and the fact that risk may arise mainly for inattentive or mistaken road users.
  • Whether the judge could properly base breach on a failure to install “wands” (flexible bollards/segregators) along the kerb.
  • Whether the judge wrongly elided duty, foreseeability, and breach.

2. Summary of the Judgment

The Court of Appeal (Lady Justice Yip, with Lord Justice Stuart-Smith and Lord Justice Bean agreeing) held:

  • Duty: TfL did owe a common law duty of reasonable care because the SCT was a positive act affecting road safety, and the duty was owed to all road users, including careless ones.
  • Breach: The trial judge’s breach reasoning was legally inadequate. He moved too quickly from (i) foreseeable risk and (ii) availability of a preventative measure (wands) to a finding of negligence, without conducting the full evaluative assessment required.
  • Procedure: It was not procedurally unfair to consider “wands”; the pleaded case and trial evidence were sufficient.
  • Disposal: The appeals were allowed; liability findings were set aside; both claims were remitted for a fresh hearing before a different Circuit Judge, with no findings preserved (including contributory negligence).

3. Analysis

3.1 Precedents Cited and Their Influence

(a) The “no general duty to improve the highway” line: Stovin v Wise and Gorringe v Calderdale Metropolitan Borough Council

The Court reaffirmed the established distinction between:

  • Omissions / failure to improve: where the complaint is essentially that the authority failed to make the highway safer, failed to warn of ordinary features, or failed to exercise a statutory power. In such cases, road users “take the highway network as they find it”. This is the core teaching of Stovin v Wise [1996] AC 923 and Gorringe v Calderdale Metropolitan Borough Council [2004] 1 WLR 1057.
  • Positive acts creating danger: where the authority’s conduct has created or materially contributed to a hazard; ordinary negligence principles can apply. Gorringe itself recognises this (Lord Hoffmann at [13], as referenced in the judgment).

This categorisation was central to the Court’s rejection of any supposed “tension” between authorities: different cases apply to different fact-patterns.

(b) Duty where a hazard is created: Yetkin v Mahmood & London Borough of Newham

Yetkin v Mahmood & London Borough of Newham [2010] EWCA Civ 776; [2011] QB 827 was treated as the paradigm example of the “positive act” category. The Court emphasised two points drawn from Yetkin:

  • A highway authority that creates a hazard owes a duty “not to create a hazard on the highway which would affect the safety of road users”.
  • That duty is not limited to careful road users. As stated in Yetkin (Smith LJ at [17], quoted in this judgment), the extent of the duty is fact-sensitive and bounded by reasonableness.

In the present case, this meant TfL could not avoid duty merely because a careful cyclist would likely have understood the markings and layout.

(c) Foreseeability does not equal breach: Tomlinson v Congleton Borough Council

The Court relied on Tomlinson v Congleton Borough Council [2004] 1 AC 46 to underline a crucial control mechanism in negligence: even where harm is foreseeable, reasonable care does not require every possible precaution; in some circumstances it is reasonable to do nothing further. This was the doctrinal anchor for the Court’s criticism of the trial judge’s approach to breach.

(d) Recent illustration, not a new rule: Braithwaite v Lewisham LBC

Braithwaite v Lewisham LBC [2025] EWHC 782 (KB) was treated as an illustration of established principles rather than a source of new doctrine.

3.2 Legal Reasoning

(a) The Court’s core clarification: keep duty, foreseeability, and breach distinct

The Court accepted the trial judge’s formulation of duty: TfL’s construction of the SCT was a positive act, so TfL owed a common law duty not to create a reasonably foreseeable risk of injury, subject to what was reasonable. However, the Court held that parts of the judgment below blurred:

  • Foreseeability of harm (a threshold concept),
  • Existence of duty (whether negligence principles apply at all), and
  • Breach (whether TfL fell below the standard of reasonable care).

The legal error was not in recognising that confusion was foreseeable (it was, and was flagged in Road Safety Audit material), but in treating the availability of an additional precaution as effectively determinative of breach.

(b) Why foreseeability mattered—but did not decide the case

The Court accepted that:

  • It was foreseeable that some cyclists might mistake the kerb upstand for a white line, especially at night.
  • The risk was in fact foreseen during design/construction (Road Safety Audits).

But the Court stressed that a foreseeable risk—especially one arising mainly from inattention or mistake—does not automatically mean the authority acted unreasonably in not adding further measures such as continuous physical barriers.

(c) The required “evaluative assessment” of breach

The Court set out the factors that must be weighed to decide whether reasonable care required further precautions. The list (expressly identified as relevant circumstances) included:

  • Magnitude/likelihood of the risk (including risk measured against volume of use).
  • Gravity of potential injury.
  • Social utility of the facility (here, segregated cycling infrastructure and efficient road-space use).
  • Standards/guidance and the role of professional design judgment.
  • Accident/complaints history (and what it truly indicates).
  • Cost, practicality, and disadvantages of further precautions (including safety consequences of barrier solutions).

The Court found the trial judge did not sufficiently engage with these considerations—particularly the scale of risk relative to the very high usage of the route, and the broader design purpose and consequences of installing a line of wands.

(d) Procedural point: “wands” were not an unfair ambush

TfL argued that breach was found on an unpleaded basis. The Court rejected this: the pleaded allegations were broad enough to encompass the contention that TfL should have taken further steps to make the kerb more distinguishable and/or prevent entry at that point; TfL’s witness was cross-examined about wands; no objection was taken; and the judge himself recognised when a point would be unfair (late-raised painted kerb markings) and excluded it. The wands point, by contrast, was properly in play.

(e) Remedy: why the Court remitted rather than re-deciding breach

Although the Court identified the proper approach, it declined to substitute its own breach assessment. The breach decision was “fact-sensitive and evaluative” and interlinked with causation and contributory negligence. Given how the case was tried and reasoned, the Court considered it unfair to reconstruct the evaluation at appellate level. Accordingly, it ordered a full retrial with no preserved findings.

3.3 Impact

(a) No new duty principle—but a firm appellate warning on methodology

The Court explicitly framed the law as “settled and clear”. The practical significance lies in method:

  • Claimants can rely on Yetkin to establish that a duty may exist where the authority created the hazard, even if the claimant was careless.
  • Defendants can rely on this decision to insist that breach must be justified by a structured proportionality-style evaluation, not by “foreseeability + an available fix”.

(b) Design and infrastructure litigation: audit concerns are important but not conclusive

Road Safety Audit materials identifying potential confusion are relevant to foreseeability and to scrutiny of design choices. But the Court cautioned that such material does not, by itself, resolve whether it was unreasonable not to adopt additional measures—particularly where those measures may undermine the design purpose or create other risks/costs.

(c) Consequences for future SCT/segregation claims

The judgment pushes future litigants and trial courts toward:

  • Quantifying/characterising risk in context (including usage rates and accident frequency).
  • Engaging with trade-offs inherent in road design (including whether barriers create new hazards or reduce utility).
  • Separating “could have been done” from “should reasonably have been done”.

The Court also downplayed any suggestion that the decision set a broad precedent threatening “the general viability of stepped cycle tracks”, stressing the location-specific nature of the dispute.

4. Complex Concepts Simplified

  • Duty vs breach: “Duty” asks whether negligence law applies at all. “Breach” asks whether the defendant fell below reasonable care. A defendant can owe a duty and yet not be in breach.
  • Positive act vs omission: If an authority merely fails to improve an existing road, liability is hard to establish (see Stovin, Gorringe). If the authority builds or alters something that creates danger, ordinary negligence rules can apply (see Yetkin).
  • Foreseeability: That harm might happen is necessary but not sufficient for negligence. The law still asks whether additional precautions were reasonably required.
  • Contributory negligence: Even if a defendant is negligent, damages can be reduced if the claimant’s lack of care contributed to the injury (as the trial judge held in large percentages here—though those findings were set aside on remission).
  • Evaluative assessment: A balancing exercise weighing risk, seriousness, utility, standards, history, and practicality/cost of precautions. It is not a box-ticking exercise and must be explained.

5. Conclusion

[2026] EWCA Civ 1095 confirms orthodox principles but delivers a pointed appellate correction: where a highway authority has created a feature that foreseeably risks confusing some road users, a finding of negligence cannot rest on foreseeability and a conceivable remedial measure alone. The trial court must conduct—and explain—a full evaluative assessment of whether reasonable care required further precautions, considering risk magnitude, design purpose and trade-offs, standards and professional judgment, accident history, and the practicality and consequences of proposed measures.

The immediate outcome was procedural and practical rather than substantive: TfL’s appeals were allowed and the cases remitted for a complete retrial with all issues reopened.