Mixed-Cause Delay and Labour-Market Access for International Protection Applicants: Causal Attribution, Proportionate Extension, and No Francovich Liability

1. Introduction

In LK v The International Protection Appeals Tribunal & Ors (No.2) [2026] IESC 26, the Supreme Court (Hogan J.) determined whether an international protection applicant was entitled to access the Irish labour market under the EU Reception Conditions regime where the first-instance protection decision had not issued within nine months. The central dispute was whether the relevant delay was “attributable to the applicant”, and whether Ireland’s implementing measure—Article 11(4)(b) of the European Communities (Reception Conditions) Regulations 2018 (SI No. 230/2018)—mis-transposed Article 15(1) of Directive 2013/33/EU.

The applicant (a Georgian national) challenged an IPAT decision (3 March 2021) upholding refusal of labour-market access. In the High Court, Heslin J. quashed the decision and held that Ireland had failed properly to transpose the Directive, awarding damages on a Francovich basis. On appeal, the Supreme Court—following a preliminary reference—allowed the State appeal, upheld the legality of the refusal, and rejected damages.

Key actors (by role)

  • The applicant: an international protection applicant (identified as “LK”/“Mr. K.” in the judgment).
  • Decision-makers: the Labour Market Access Unit (LMAU), a Ministerial Review Officer, and IPAT (Designated Member).
  • Respondents/Appellants: IPAT, the Minister for Justice, Ireland, and the Attorney General.

Core legal issues

  1. Attribution of delay: how to assess “delay” under Article 15(1) where causes are mixed (applicant conduct, State action/omission, and external events such as a pandemic).
  2. Transposition: whether the addition in the 2018 Regulations (“attributed, or attributed in part”) unlawfully diluted Article 15(1).
  3. Damages: whether any EU-law breach capable of founding Francovich damages was established.

2. Summary of the Judgment

The Supreme Court allowed the State appeal and held:

  • No defective transposition: after the CJEU ruling in Case C-742/24 Havvitt EU:C:2026:15, Article 11(4)(b) of the 2018 Regulations was consistent with Article 15(1) as properly interpreted.
  • Not a binary test: “attributable to the applicant” requires a causal-link inquiry and permits apportionment where delay has mixed causes; only delay caused by the applicant can count against the nine-month access rule.
  • Outcome correct despite flawed reasoning: although elements of IPAT’s reasoning (which treated large periods as attributable to the applicant) could not stand after Havvitt, the applicant still failed to show a nine-month period of non-attributable delay. The refusal of labour-market access was therefore legally correct.
  • No Francovich damages: since there was no breach of EU law/entitlement, damages could not arise.

3. Factual and Procedural Timeline (condensed)

Date/Period Event Relevance to “delay”
2 Sep 2019 Protection application lodged Start point for Article 15(1) nine-month clock
16 Sep 2019 IPO interview scheduled but applicant not notified Non-attributable delay (State/administrative mischance)
12 Dec 2019 IPO interview; questionnaire issued (in Georgian) Questionnaire return becomes key cooperation step
28 Jan 2020 Solicitor assigned Applicant cooperation assessed with legal support context
Mar–Apr 2020 COVID-19 impacts; translation difficulties Potentially non-attributable/excusable delay
End Apr 2020 onward No adequate explanation for continued delay Attributable delay (on Supreme Court’s assessment)
Jun 2020 LMAU labour-market access application Triggers domestic decision-making
28 Aug 2020 LMAU refusal decision Reference date used in the Supreme Court’s analysis window
25 Aug 2020 Questionnaire ultimately submitted Shows extended period without completed cooperation step
3 Mar 2021 IPAT appeal refused Decision challenged by judicial review

4. Analysis

4.1 Precedents Cited and Their Role

(a) Case C-742/24 Havvitt EU:C:2026:15 (CJEU)

Havvitt was decisive. The Supreme Court’s earlier judgment had referred questions on (i) what counts as “delay attributable to the applicant” and (ii) whether the Irish implementing text (“attributed, or attributed in part”) was permissible. The CJEU clarified:

  • Non-cooperation can justify exclusion: a complete lack of cooperation that prevents examination of the application (e.g., no information for more than nine months) can justify refusal at the nine-month point.
  • Partial non-cooperation cannot trigger total forfeiture: where the applicant’s failure to cooperate covers only part of the period, the consequence is extension of the nine-month period by the duration causally linked to that failure, not categorical denial.
  • Mixed causes require apportionment: where time lapse is explained by mixed causes (including externalities such as a pandemic), the decision-maker must identify the fraction attributable to the applicant.
  • Transposition: Article 15(1) does not preclude national legislation denying access where delay may be attributable “in part” to the applicant—because the correct EU-law approach already accommodates partial attribution via proportionate extension.

The Supreme Court treated Havvitt as eliminating the High Court’s transposition critique and reframing the attribution inquiry as one of causation and proportion, not an “all-or-nothing” allocation.

(b) LK v. International Protection Appeals Tribunal (No.1) [2022] IEHC 441 (Heslin J.)

The High Court had quashed IPAT’s decision on irrationality and fair procedures, and treated the “in part” wording as creating a materially different (and harsher) test than Article 15(1). On appeal, Hogan J. disagreed with the High Court’s reading of IPAT’s method (considering the Designated Member sought to apply the Directive), but ultimately held that—post-Havvitt—the legal foundation for the transposition critique fell away.

(c) LK v. International Protection Appeals Tribunal (No.2) [2023] IEHC 210 (Heslin J.)

This decision awarded Francovich damages for alleged failure to transpose and resulting loss of earnings. The Supreme Court’s reversal followed from its conclusions that (i) transposition was not defective and (ii) no EU entitlement to labour-market access had been established on the facts. Without an EU-law breach, Francovich liability could not be made out.

(d) LK v. International Protection Appeals Tribunal [2024] IESC 42 (Dunne J.)

This was the Supreme Court’s earlier judgment deciding to refer questions under Article 267(3) TFEU. It framed the uncertainty: some delay was not fairly attributable to the applicant, yet some appeared inexplicable; the legal system needed an EU-law rule for attributing mixed-cause delay. Havvitt answered that uncertainty; [2026] IESC 26 applies it.

(e) The State (Keegan) v. Stardust Victims' Compensation Tribunal [1986] IR 642

The High Court relied on Keegan to characterise IPAT’s attribution findings as irrational (“fundamental reason and common sense”). The Supreme Court’s approach is notable: rather than resolving the case primarily through domestic irrationality review, it resolved the dispute through the corrected EU-law attribution framework and a re-assessment of what delay could fairly be excused.

(f) Minister for Justice & Equality & Others v. Workplace Relations Commission & Others (Case C-783/17, EU:C:2018:979)

IPAT had cited this authority when noting a difference between the Directive and Irish regulations and stating it would apply the Directive wording. In the event, the Supreme Court held that the “difference” was not legally problematic once Article 15(1) was interpreted as permitting partial attribution/apportionment.

4.2 Legal Reasoning

(a) The governing EU-law rule: causation, not labels

The Supreme Court adopted Havvitt’s core move: “attributable to the applicant” requires a causal link between the applicant’s conduct and the delay in the first-instance decision. The legal consequence of applicant-caused delay is not necessarily refusal; it is to extend the nine-month period by the time attributable to the applicant. This converts the inquiry from a blunt eligibility bar into a structured calculation problem.

(b) Rejection of a binary approach and the need to segment time

Hogan J. held that some of IPAT’s reasoning “cannot now be stood over” because it treated periods as applicant-attributable that were not (e.g., where the applicant was not notified of an IPO interview). The Supreme Court segmented the relevant timeline and identified periods that could not fairly be blamed on the applicant, including:

  • 2 Sep 2019 to 12 Dec 2019: non-notification of a scheduled interview.
  • 12 Dec 2019 to 28 Jan 2020: until legal representation was assigned.
  • March/April 2020: disruption caused by COVID-19 and interpreter availability.

(c) The decisive factual conclusion: even generous excusals left applicant-attributable time

Even after excusing those periods “if anything” generously, the Court held there was no excuse whatsoever for delay after the end of April 2020. On that assessment, the applicant failed to establish the necessary nine months of delay not attributable (in whole or in part) to him. Accordingly, he was not entitled to labour-market access at the relevant time, and the refusal outcome was legally correct.

(d) “Right result, wrong reasons” in an EU-law context

A significant feature is the Court’s acceptance that IPAT’s reasoning was partly unsustainable after Havvitt, yet the decision was upheld because, applying the corrected legal test, the applicant still failed on the entitlement threshold. The judgment thus illustrates a pragmatic appellate approach: a decision may survive if the lawful application of the correct test leads to the same conclusion on the established or properly assessed facts.

(e) Transposition and the “attributed in part” wording

The High Court treated the phrase “attributed, or attributed in part” (Regulations) as diluting “cannot be attributed to the applicant” (Directive). Following Havvitt, the Supreme Court held the transposition claim “disappeared”: EU law itself requires partial attribution in mixed-cause cases, and so the Irish text did not introduce an impermissible additional ground of refusal. It aligned with the Directive’s effect as interpreted by the CJEU.

(f) Francovich damages: no breach, no liability

Heslin J. had applied the standard three-limb Francovich framework (rights-conferring rule; sufficiently serious breach; causal link). The Supreme Court’s reversal turned on the first step: no breach of EU law was established because (i) transposition was not defective and (ii) the applicant was not substantively entitled to labour-market access on the facts. Without breach, the damages claim necessarily failed.

4.3 Impact

(a) Decision-making standard for labour-market access under Reception Conditions

Irish decision-makers (LMAU/Review/IPAT) must treat “attributable delay” as:

  • Segmented (time periods may have different causes);
  • Causation-based (a demonstrated causal link is required);
  • Apportionable in mixed-cause scenarios (including systemic and external factors such as pandemics).

Practically, this encourages reasoned findings tying specific applicant acts/omissions to specific time loss, rather than global conclusions of “non-cooperation”.

(b) Litigation effects: narrowing of transposition challenges

The Supreme Court’s endorsement of the 2018 Regulations (in light of Havvitt) reduces the scope for future challenges premised solely on the “attributed in part” wording. Future disputes are more likely to turn on evidence and causation (who caused which slice of delay), rather than the abstract validity of the implementing text.

(c) Damages exposure for alleged delay misattribution

The judgment signals that Francovich damages in this field will generally require a clear demonstration that (i) the applicant actually satisfied Article 15(1) as interpreted (after apportionment), and (ii) a refusal unlawfully denied labour-market access. Where entitlement fails on a properly apportioned timeline, damages will not lie.

5. Complex Concepts Simplified

  • “Attributable delay”: time lost because of a party’s conduct. Under Havvitt, it is not enough to say an applicant “caused delay” in general—there must be a causal link between what the applicant did (or did not do) and a specific period of delay.
  • Mixed causes: delay can be caused partly by the applicant (e.g., late forms), partly by the State (e.g., administrative error), and partly by external events (e.g., COVID-19). The law requires an apportionment of responsibility.
  • Proportionate extension of the nine-month period: if an applicant causes, say, two months of delay, the nine-month threshold for labour-market access may effectively become eleven months (nine + two), rather than an outright loss of access.
  • Article 267 TFEU preliminary reference: when an EU-law question is unclear, a national apex court must refer it to the CJEU. Here, the reference produced the rule that resolved both attribution and transposition concerns.
  • Francovich damages: compensation from the State for breach of EU law. It presupposes an actual breach of an EU rule intended to confer rights; if there is no breach (because entitlement is not established), damages cannot be awarded.

6. Conclusion

[2026] IESC 26 establishes—through the application of Case C-742/24 Havvitt—that labour-market access under Article 15(1) of the Reception Conditions Directive requires a causal, proportionate assessment of delay, including apportionment in mixed-cause scenarios. It also confirms that Ireland’s 2018 Regulations do not mis-transpose EU law by referring to delay “attributed in part” to the applicant.

Although IPAT’s reasoning could not fully survive the clarified EU-law framework, the Supreme Court held the refusal outcome remained legally correct on a segmented timeline. Consequently, the High Court’s findings of defective transposition and the award of Francovich damages were overturned. The decision shifts future disputes away from the text of the implementing measure and toward disciplined, evidence-based attribution of discrete periods of delay.