Consumer Protection Act 1987: No inferred “defect” from seatbelt unfastening where regulatory scheme contemplates inadvertent release and claimant cannot prove the safety shortfall

1. Introduction

This Outer House decision (Samantha Maguire against Black Isle Electrical Ltd and others [2026] CSOH 51, Lord Lake, 3 June 2026) arose out of a serious road traffic accident in which the pursuer (a front passenger) was ejected from a Nissan van during a rollover after her seatbelt became unfastened. The driver (an employee of the first defenders) lost control; liability to the pursuer was admitted and the principal claim settled for over £10.7m plus benefits, NHS charges and expenses.

The remaining dispute was between the defenders and the van producer (brought in as third party): whether the third party was jointly liable under section 2(1) of the Consumer Protection Act 1987 (“CPA 1987”) because the seatbelt/buckle was “defective” under section 3(1), such that the settlement sums should be apportioned.

The core factual issue at proof was narrow but technically complex: how the buckle released (and whether any such release pointed to a legally relevant “defect”).

2. Summary of the Judgment

  • No CPA 1987 liability established against the third party. The defenders failed to prove a “defect” in the seatbelt/buckle that caused the pursuer’s injuries.
  • The court rejected the defenders’ principal engineering theory (a 3-stage mechanism involving an “unbalanced zone” allowing withdrawal of the tongue without true actuation of the release mechanism).
  • The court accepted that the only evidence-based mechanism consistent with the buckle design was that the release mechanism was operated, likely by an inadvertent strike; but (i) the defenders had not pled a case that inadvertent strike-release was itself a CPA defect, and (ii) in any event, the court indicated it would not have found defect on that basis given the regulatory context.
  • Inference of defect from the mere fact of unfastening was refused: seatbelts are designed to open, and the applicable regulatory scheme implicitly tolerates some strike-release risk.
  • Obiter: if apportionment had arisen, Lord Lake would have apportioned 40% to the defenders / 60% to the third party, and would have awarded interest at the judicial rate.

3. Analysis

3.1 Precedents Cited

(a) Expert/factual evidence boundaries: Kennedy v Cordia Services LLP and Gubinas v HMA

A key preliminary point was whether evidence from an engineer employed by the buckle manufacturer should be treated as expert opinion (and potentially inadmissible for lack of independence) or as factual evidence of design/operation.

  • Kennedy v Cordia Services LLP [2016] UKSC 6, 2016 SC (UKSC) 59 was used to frame the distinction between (i) evidence derived from personal experience/observation (fact) and (ii) evidence sourced from research/other materials (expert opinion requiring independence and compliance with expert duties).
  • Gubinas v HMA [2017] HCJAC 59, 2018 JC 45 supported the permissibility of witnesses identifying what is visible in video recordings: such testimony can be treated as a form of factual assistance drawing attention to features, with the recording remaining the primary evidence.

Lord Lake held the manufacturer engineer’s description of buckle components and operation to be evidence of fact within personal knowledge, not expert opinion, and relied on it (not contradicted by the parties’ skilled witnesses).

(b) Regulatory compliance as evidence in CPA defect analysis: Wilkes v Dupuy International Ltd and Tesco Stores

Lord Lake adopted the approach in Wilkes v Dupuy International Ltd [2018] QB 627 that:

  • Non-compliance with mandatory standards is evidence of defect.
  • Compliance is not a complete defence but can be powerful evidence that the product met the safety level persons are entitled to expect in respect of the risks the regulation addresses.

The judgment also noted Wilkes’ discussion of Tesco Stores [2006] EWCA Civ 393 (via the quotation in Wilkes) to the effect that compliance can be cogent evidence against defect.

(c) Pleaded case limits and “new” theories: Burns v Dixon's Iron Works

The defenders attempted, in submissions, to pivot to a new basis: even if the buckle released by inadvertent button strike (the third party’s theory), that very possibility meant the buckle was defective.

They relied on Burns v Dixon's Iron Works 1961 SC 102 for the proposition that courts may allow variations/developments of a pled case. Lord Lake distinguished this: the defenders’ inadvertent-release-as-defect argument was not a permissible development but an entirely new case inconsistent with their pleaded denial and absent from their evidence, reports and notices.

(d) Competing causation hypotheses: Rhesa Shipping Co SA v Edmunds (The Popi M)

The third party invoked Rhesa Shipping Co SA v Edmunds (The Popi M) [1985] 1 WLR 948 to argue that if the defenders’ primary mechanism failed, the case necessarily failed. Lord Lake rejected that submission because the pleadings also advanced other alleged defects (mounting position and casing strength), requiring separate consideration.

(e) Inference of defect without proving the precise mechanism: Gee v Dupuy and product-fire cases

Lord Lake relied heavily on Gee v Dupuy [2018] EWHC 1208 (QB), [2018] Med LR 347 for two propositions:

  • “Defect” under the CPA is not the dictionary concept; it is a statutory safety yardstick.
  • A claimant may not need to prove the precise mechanism, but must establish what about the product’s state/behaviour/risk fell below entitled expectations; in some cases defect can be inferred from failure in normal use where a standard product would not fail.

Lord Lake then compared and distinguished situations where courts inferred defect because failure itself is strongly indicative of unsafe abnormality:

  • Ide v ATB Sales Ltd and Another [2008] EWCA Civ 424 (as cited in Gee)
  • Parks v LG Electronics UK Limited 2025 SLT (Sh Ct) 171 (phone charging fire; defect inferred)
  • Baker v KTM Sportmotorcycle UK Limited [2017] EWCA (Civ) 378, [2018] ECC 35 (galvanic corrosion permitting inference of a braking defect)
  • Al Iqra v DSG Retail Limited [2019] EWHC 429 (QB), [2019] ECC 23 (heater fire; defect inferred absent misuse)

Those cases involved products whose failure mode (fire/explosion/brake corrosion) was itself powerful evidence of a safety shortfall. Lord Lake held that seatbelt buckles are different: they are intended to release under actuation, and the regulatory framework implicitly accepts that certain strike scenarios may release a buckle. Therefore no “necessary inference” of CPA defect arose from the mere fact of unfastening.

(f) Scottish Inner House context: McGlinchey v General Motors Limited

Lord Lake addressed McGlinchey v General Motors Limited, [2012] CSIH 91, noting that although it referenced a dictionary meaning, it ultimately applied the statutory safety-performance test and was not inconsistent with Gee.

3.2 Legal Reasoning

(a) What “caused” the buckle to open: rejecting the “unbalanced zone” mechanism

The defenders’ principal case depended on a theory that a partial depression (2–3mm) of the release button created an “unbalanced zone” allowing the tongue to be withdrawn without full operation of the release mechanism. Lord Lake rejected this on a straightforward mechanical logic derived from accepted buckle operation:

  • The tongue remains retained until the locking bar travels around the “corner” of the L-shaped slot, retracting the latch tip from the tongue aperture.
  • A 2–3mm depression did not move the locking bar around the corner; the latch tip remained through the tongue; therefore the tongue could not be pulled free.
  • Vibration/damage speculation could not supply the missing step (movement around the corner) on the evidence.

This reasoning mattered legally because the pleaded “defects” (high rigid mounting, casing weakness) were advanced largely as enablers of the unbalanced-zone hypothesis. Once the hypothesis failed, so did the causal link between those features and the unfastening.

(b) The police officer’s post-accident “pull test”: not decisive evidence of defect

A police officer reported that 11 days post-crash the tongue “clicked” in but then pulled out with a gentle sharp pull, surprising two officers. Later joint inspections (2018 and 2021) found normal operation.

Lord Lake treated the officer’s observation as unexplained but not supportive of the defenders’ theory: it suggested the buckle may not have latched properly (contrary to the hypothesis that it had latched pre-crash), and its later normal functioning pointed to a transient obstruction (e.g., vegetation/debris visible in photographs) rather than inherent design defect.

(c) Interpreting UN/ECE Regulation 16: applicability to rollover and meaning of “inadvertent” release

A central aspect of the reasoning was how UN/ECE Regulation 16 (“UN/ECE16”) informs the CPA’s “entitled expectation of safety” test:

  • Rollover applicability: Lord Lake rejected the third party’s submission that UN/ECE16 effectively addressed only frontal impacts because the dynamic test is a sled test. The absence of a rollover test did not imply the substantive requirements were irrelevant to rollover accidents, especially given the regulations’ broad protective purpose and the vehicle handbook’s statement that belts reduce rollover injury risk.
  • But the defenders’ construction of 6.2.2.2 failed: the defenders argued any accidental/inadvertent opening breached the regulation. Lord Lake read UN/ECE16 as a whole, including its definitional treatment of enclosed release buttons (the “sphere of 40 mm” concept). That implied the regulatory scheme recognises some strike-release risk outside the enclosed-button parameter; construing 6.2.2.2 as banning all inadvertent strike-release would create an internal contradiction.
  • “Inadvertent/accidental” in 6.2.2.2: Lord Lake construed this as targeting unintended release arising from the wearer’s actions (e.g., other deliberate in-cabin movements) rather than forbidding all chance strike-release by objects in the compartment.

This interpretation drove the outcome on two fronts: it undermined the defenders’ attempt to equate chance strike-release with regulatory breach, and it reinforced the conclusion that the public is not entitled (under the CPA yardstick) to expect elimination of a risk the regulatory framework implicitly tolerates.

(d) Pleadings and fair notice: the “inadvertent strike-release is itself a defect” argument was not open

Even though the third party had advanced inadvertent strike-release as an explanation, Lord Lake held the defenders could not adopt, at closing, a new legal theory that inadvertent strike-release itself constituted a CPA defect. The case was commercial procedure, but the court emphasised that fair notice remains essential and may be provided by pleadings and/or procedural documents (issues notes, witness statements, expert reports). Here, none gave notice.

Importantly, Lord Lake nevertheless provided an alternative view: even if competently pled, he would not have found defect merely because an inadvertent strike could operate the release. Buckles must release readily, and the regulatory scheme’s implicit risk acceptance means the public is not “entitled to expect” a design that eliminates that risk absent a properly articulated case and supporting evidence as to feasible safer design.

(e) No res ipsa inference from unfastening alone

The defenders’ fallback—“it came undone, therefore defective”—failed because the court refused to treat unfastening as inherently abnormal in the CPA sense. Unlike unexplained fires or explosions, buckle opening is an intended function; and there was an evidence-based, non-defect explanation consistent with ordinary operation (activation of the release mechanism by a strike).

3.3 Impact

(a) CPA 1987 defect claims involving safety-critical components

The decision reinforces that in CPA 1987 claims the court will:

  • focus on the specific safety shortfall said to fall below entitled expectations, not merely on the occurrence of a bad outcome;
  • be reluctant to infer defect where the event can be explained by ordinary functioning in an unusual environment (here, a rollover with loose objects) and where the regulatory scheme tolerates residual risk.

(b) Regulatory frameworks as a “yardstick” for public expectation

The judgment is a practical Scottish illustration of Wilkes: detailed regulations do not create an automatic defence, but can strongly shape what “persons generally are entitled to expect.” Here, the court used the structure and internal logic of UN/ECE16 to reject an absolutist reading (“no accidental release in any circumstances”) and to deny an inference of defect from risk realisation.

(c) Litigation discipline: commercial procedure does not dilute fair notice

The ruling is also a warning against late-stage forensic pivots. Even where one party’s explanation emerges as the most plausible, an opponent cannot convert that into a new liability theory without having pled and signposted it.

(d) Evidence handling: manufacturer engineers and “fact” evidence

By treating a manufacturer engineer’s description of components/operation as factual evidence, the court confirms a route by which product-knowledge witnesses can be used without necessarily triggering the full Kennedy expert admissibility constraints—provided the evidence is confined to personal knowledge and demonstrable operation rather than opinion drawn from external research.

4. Complex Concepts Simplified

  • CPA 1987 “defect” (s.3): not “something went wrong,” but whether the product’s safety fell below what people generally were entitled to expect, considering circumstances including standards, warnings and likely use.
  • “Producer” liability (s.2): if a defective product causes damage, the producer is strictly liable (no need to prove negligence), but defect and causation must still be proved.
  • “Inference of defect”: courts sometimes infer defect when a product fails in a way that ordinarily would not occur (e.g., a charger bursting into flames). This case limits that move where the “failure” is the product doing what it is designed to do (opening), albeit at an unintended moment.
  • UN/ECE16: an international vehicle safety regulation adopted in Europe. Compliance is powerful evidence about acceptable safety levels, but it is not a complete defence.
  • “Unbalanced zone” theory (as alleged): a hypothesised intermediate buckle state where partial button depression would let the tongue pull free. The court rejected it because the latch tip still physically blocks withdrawal until the mechanism completes its designed travel.

5. Conclusion

Lord Lake’s decision is significant less for any novelty in CPA 1987 doctrine than for its disciplined application of the statutory “entitled expectation of safety” test to a real-world, high-stakes product incident. The court:

  • required a coherent, evidence-based account of the alleged safety shortfall;
  • used the internal logic of UN/ECE16 to resist absolute safety expectations inconsistent with the regulatory scheme;
  • refused to infer defect from unfastening alone where ordinary operation (button actuation) plausibly explained the event; and
  • reaffirmed that fair notice in pleadings/procedure constrains late-stage reformulation of liability theories.