“Expose to sanctions” in EPS voyage charter clauses means a real risk, not a balance-of-probabilities breach

Case: Tonzip Maritime (Singapore) PTE Ltd v 2 Rivers PTE Ltd
Citation: [2026] EWCA Civ 641
Court: England and Wales Court of Appeal (Civil Division)
Date: 22 May 2026
Judges: Coulson LJ, Zacaroli LJ, Foxton LJ (lead)
Appeal from: [2025] EWHC 2036 (Comm)

1. Introduction

The appeal concerned a voyage charter on an amended ExxonMobil VOY2005 form. The owners of the MV CATALAN SEA refused the charterers’ order to load crude oil where the shipper was a Russian oil company said to be associated with a designated individual (referred to here as Mr G) under EU/UK Belarus-related sanctions. The owners relied on an “EPS Sanctions clause”, particularly sub-clause (C), permitting refusal where, in the owners’ reasonable judgment, compliance is prohibited by sanctions or “will expose” relevant persons to sanctions.

The High Court accepted the owners’ construction that “expose … to sanctions” could be triggered by a reasonable apprehension of sanctions risk, but held that—on the facts—the owners had not formed a reasonable judgment. The Court of Appeal (i) upheld the risk-based construction and (ii) allowed the owners’ appeal on reasonableness, holding the first-instance judge misdirected himself and that the owners’ refusal was objectively reasonable.

Parties and roles

  • Appellant: the shipowners (“the Owners”).
  • Respondent: the voyage charterers (“the Charterers”).

Key legal issues

  • Construction: whether “expose … to sanctions” requires a balance-of-probabilities conclusion that sanctions will be breached, or a reasonable judgment of a real risk.
  • Application: whether, on the material available in November 2021, the owners’ refusal to load was a reasonable judgment under sub-clause (C).
  • Ancillary: the relevance/admissibility of an EU General Court decision (ultimately not determinative on appeal).

2. Summary of the Judgment

  • The Court of Appeal dismissed the Charterers’ cross-appeal on construction and upheld that “expose … to sanctions” in clause (C) is satisfied by a reasonable judgment that compliance creates a real risk of sanctions liability, not that breach is more likely than not (paras 33–41, 58–60).
  • The Court of Appeal allowed the Owners’ appeal on reasonableness, holding that the first-instance judge:
    • misapplied Litasco SA v Der Mond Oil and Gas Africa [2023] EWHC 2866 (Comm) (paras 65–67); and
    • wrongly evaluated the case as though the owners had to decide that Mr G in fact controlled the shipper, rather than that there was a real risk of that control continuing (paras 72–74).
  • On the evidence (screening reports; the timing/nature of the purported transfer of interests to a close family member; and the limitations/assumptions in the legal materials supplied), a reasonable owner could conclude there was a real sanctions risk (paras 83–89).
  • The Court did not need to decide the disputed questions about the EU decision’s admissibility or effect (para 82).

3. Analysis

3.1 The EPS Sanctions clause and what changed in the law

The central doctrinal contribution of [2026] EWCA Civ 641 is its clear endorsement—on orthodox contractual interpretation grounds—of a risk-trigger for refusal clauses drafted around “expose … to sanctions” coupled with an express “reasonable judgment” standard, at least where the clause structure itself equates “exposure” with “risk”.

Clause (C) (core wording)

“THE OWNERS SHALL NOT BE OBLIGED TO COMPLY … WHICH IN THE REASONABLE JUDGEMENT OF THE OWNERS, IS PROHIBITED BY SANCTIONS OR WILL EXPOSE … TO SANCTIONS. IN THE EVENT THAT SUCH RISK ARISES …”

The Court treated the internal cross-reference—“such risk”—as powerful confirmation that “expose … to sanctions” is being used in a risk sense (paras 38–40). This anchors construction in the text rather than in generalised debates about sanctions clauses.

3.2 Precedents cited and how they influenced the decision

(a) Presumptions against derogation from “normal” obligations

  • Modern Engineering (Bristol) Ltd v Gilbert-Ash (Northern) Ltd [1974] AC 689
    The appeal was argued on the basis that sub-clause (C) potentially engages the principle that, absent clear words, parties are presumed not to derogate from normal contractual rights/obligations (para 32). The Court did not decide whether the presumption applied, because it found the language sufficiently clear in any event (para 40).

(b) “Pick and mix” drafting and limited cross-clause inference

  • MS Amlin Marine NV v King Trader [2025] EWCA 1387
    Cited for the general approach where different standard clauses overlap: they must be read together to resolve inconsistency (para 47). Here, the Court rejected using unrelated standard clauses (war/piracy) to reshape the meaning of the EPS sanctions clause (paras 46–49).
  • Prestcold (Central) Ltd v Minister of Labour [1969] 1 WLR 89 and Lindsay (WN) & Co Ltd v European Grain & Shipping Agency Ltd [1963] 1 Lloyd's Rep 437
    These authorities were used to caution against assuming uniform drafting intent in standard forms that evolve piecemeal (para 47).
  • International Entertainment Holdings Ltd v Allianz Insurance Plc [2024] EWCA Civ 1281
    Quoted for the “pick and mix” point: stitched clauses do not justify strong inferences of consistent terminology (para 48). This supported the Court’s refusal to infer that because other clauses used “may expose” etc, “expose” alone must mean a balance-of-probabilities outcome.

(c) Authorities on “expose” in other contractual contexts

  • Mamancochet Mining Ltd v Aegis Managing Agency Ltd [2018] EWHC 2643 (Comm)
    In that insurance case, “would expose that (re)insurer to any sanction…” was construed as applying only where payment was prohibited (paras 51–52). The Court distinguished it: there, “expose” was not paired with “such risk”, and the clause affected accrued payment obligations; here, clause (C) operates prospectively and preserves the charterers’ ability to give alternative voyage orders (paras 52–53).
  • Pacific Basin IHX Ltd v Bulkhandling Handymax AS (The Triton Lark) [2011] EWHC 2862 (Comm) and [2012] EWHC 70 (Comm)
    The piracy clause required assessment of exposure to piracy risk; Teare J treated “exposure” in risk/danger terms. The Court of Appeal held the wording and context differed materially from the EPS sanctions clause (paras 54–57), limiting its interpretive value.
  • Ceto Shipping v Savory Shipping [2025] EWHC 2033 (Comm)
    Factually closer. There was no dispute that “expose … to any sanction” required risk assessment; the debate was the level of risk. Cockerill J rejected “virtual certainty” and accepted “serious possibility” as sufficient. The Court treated this as consistent with its own construction (para 58).

(d) Appellate restraint and interference with evaluative conclusions

  • Subesh v Secretary of State for the Home Department [2004] EWCA Civ 56
    Cited for the general approach to appeals: the appellate court must find objective grounds requiring a different view, not merely preferring one (para 61).
  • Henderson v. Foxworth Investments Ltd [2014] UKSC 41
    The Court applied Lord Reed’s formulation limiting interference with primary findings absent identifiable error (para 62). Here, identifiable errors were found, warranting reconsideration (paras 63–74).

(e) Sanctions cases relied on below: correcting misapplication

  • Litasco SA v Der Mond Oil and Gas Africa [2023] EWHC 2866 (Comm)
    The first-instance judge treated this as establishing that “speculation” cannot support a reasonable determination. The Court of Appeal held this was a misdirection: Litasco concerned whether there was a triable defence of breach of sanctions, not a contractual “reasonable judgment” about risk; risk assessment necessarily involves imponderables (paras 66–67). This was a key reason to revisit the trial judge’s conclusion.
  • Vneshprombank LLC v Bedzhamov [2024] EWHC 1048 (Ch)
    That case concerned the meaning of “reasonable cause to suspect” in penal legislation. The Court of Appeal regarded it as of limited assistance to a contractual risk-judgment clause, but accepted one specific misstep: using it to suggest a lack of positive belief was fatal, when the clause required only a reasonable judgment of risk (paras 68–71).

(f) Contractual decision-making standards (“process” vs “outcome”)

  • Braganza v BP Shipping Ltd [2015] UKSC 17
    Introduced the modern framework for implied constraints on contractual discretions, with public-law rationality concepts sometimes used to police the decision-making process (para 76).
  • Lehman Brothers International (Europe) v. Exxonmobil Financial Services BV [2016] EWHC 2699 (Comm)
    Cited as an example of resistance to importing procedural review into fast-moving commercial determinations (para 77).
  • Pacific Basin IHX Ltd v Bulkhandling Handymax AS (The Triton Lark) and Ceto Shipping v Savory Shipping
    Both were used to highlight that where a clause requires an “objectively reasonable judgment”, the core question is usually the reasonableness of the outcome; flaws in enquiries may go to good faith or evidential weight but do not necessarily invalidate an otherwise objectively reasonable conclusion (paras 78–80).

(g) Prior judgments and admissibility (not decided but framed)

  • Hollington v Hewthorn & Co [1943] KB 27
    Raised as part of arguments about whether the EU General Court decision could be relied on as evidence of facts underpinning reasonableness. The Court assumed (in the Charterers’ favour) that it was inadmissible and decided the appeal without it (para 82).

3.3 Legal reasoning

(A) Construction: “expose … to sanctions” = real risk

The Court’s reasoning combined textual analysis with commercial common sense:

  • Textual structure and internal reference: the second sentence (“IN THE EVENT THAT SUCH RISK ARISES…”) refers back to both “prohibited” and “expose”, but must at least capture exposure’s risk dimension (paras 38–39). This supported the owners’ reading that “expose” is risk-based.
  • Commercial context: owners must make fast, prospective decisions with limited information; sanctions regimes are complex and multi-jurisdictional; regulators decide later with more time and information. A “more likely than not” threshold would be commercially and operationally inapt (paras 41(i)–(v)).
  • Limited utility of other “expose” cases: decisions like Mamancochet turned on different drafting and different commercial consequences (accrued payment obligations), and so did not compel a balance-of-probabilities standard here (paras 51–53).

(B) Application: why the owners’ judgment was objectively reasonable

Having corrected the legal lens (risk, not proof), the Court assessed reasonableness by reference to the material known at the time and the nature of what was being asked of the owners.

Key features supporting “real risk”
  • Timing and relationship red flags: Mr G had previously held a majority stake; after designation, that interest was said to have been transferred to a close family member and long-time business partner (paras 83(i)–(iii), 85(v)). The Court treated the inference of possible “cosmetic” restructuring as common sense, even without later guidance documents (para 83(iii)).
  • Screening outputs: the Refinitiv/World-Check report described the company as “associated to sanctioned individual” (para 83(v), and see factual summary at paras 13–14). The Court considered this plainly capable of supporting a risk assessment.
  • Weakness of “comfort” materials: the shipper’s letter was brief, Russian-law framed, and opaque as to source and detail (para 85(i)). The legal opinions were addressed to the shipper, heavily assumption-based, and reliant on unverified self-provided information—precisely what would be expected even if control had not genuinely changed (paras 85(ii)–(iii)).
  • Baker McKenzie memo flagged the risk: critically, the memo itself acknowledged that authorities could conclude de facto control persisted, and that different factual interpretations could not be excluded (para 85(iv)). The Court viewed this as tending to reinforce, not remove, the owners’ sanctions risk concern (para 86).
Why the first-instance reasoning was untenable on the correct test
  • The judge’s focus on absence of evidence “which evidences … control” and the owners’ inability to “confirm” control treated the clause as requiring something closer to proof of control rather than a reasonable assessment of risk (paras 72–74).
  • Reliance on Litasco as a bar on “speculation” was a category error (paras 66–67), because contractual risk judgments in sanctions contexts inevitably involve contestable and partially hidden facts.
Additional sources (Kommersant; Infospectrum) did not negate risk

Even considering the Kommersant article and the Infospectrum report (not fully engaged with by all decision-makers), the Court held they did not eliminate the real risk; indeed Kommersant expressly flagged uncertainty about whether the EU would treat the restructuring as circumvention (paras 87–88).

3.4 Impact

Practical consequences for charterparty performance

  • Lower evidential burden, but bounded by reasonableness: owners relying on clauses like EPS (C) do not need to establish, on balance of probabilities, that sanctions would in fact be breached. They must show a reasonable commercial judgment of a real risk of sanctions exposure.
  • Enhanced relevance of compliance tools: credible sanctions screening outputs (e.g., Refinitiv/World-Check) can rationally ground a “real risk” conclusion, especially where counterparty-provided assurances are assumption-heavy and self-referential.
  • Drafting lesson: “such risk” language materially strengthens a risk-based construction. Parties who intend a stricter threshold (e.g. “more likely than not prohibited”) should say so.

Litigation consequences

  • Appellate review: while reasonableness is evaluative, errors in legal framing (risk vs proof; misreading authorities) justify appellate intervention (paras 63–74).
  • “Process” challenges remain open: the Court surveyed authorities on process-vs-outcome review (paras 75–80) but did not resolve the outer boundaries for sanctions decisions. Future disputes may test whether inadequate enquiries can, without more, make an otherwise plausible outcome “unreasonable” under similar clauses.
  • Use of foreign/regional judgments: the Court avoided deciding the admissibility/effect of the EU General Court decision, leaving room for future clarification on when such decisions can evidence “reasonableness” at an earlier time (para 82).

4. Complex concepts simplified

  • “Expose … to sanctions”: in this clause, it means “put at real risk of sanctions consequences” (e.g., being treated as dealing with a sanctioned person/entity), not “will definitely (or probably) breach sanctions”.
  • “Reasonable judgment”: an objective standard. The question is whether a reasonable shipowner could reach that judgment on the available material and in the commercial context—not whether the judge agrees with it.
  • Gilbert-Ash principle (from Modern Engineering (Bristol) Ltd v Gilbert-Ash (Northern) Ltd): a presumption that contracts do not lightly cut down usual rights/obligations unless clear words do so. Here, the Court found the words clear enough.
  • Braganza / Wednesbury rationality: sometimes courts police discretionary decisions for irrationality or improper considerations. In fast-moving commercial contexts, courts often focus on the objective reasonableness of the outcome rather than auditing the decision-making process in public-law style (paras 76–80).
  • Hollington v Hewthorn: a general evidential rule (with exceptions) limiting the use of prior judgments as proof of the facts decided. Its application to the EU decision was argued but not decided (para 82).

5. Conclusion

Tonzip Maritime (Singapore) PTE Ltd v 2 Rivers PTE Ltd [2026] EWCA Civ 641 establishes a clear, commercially grounded approach to EPS-style sanctions refusal clauses in voyage charters: where the clause speaks of “expos[ing]” parties to sanctions and then refers to “such risk”, the contractual trigger is a reasonable judgment that compliance creates a real risk of sanctions liability, not proof on the balance of probabilities that sanctions will be breached. Applying that standard, the Court of Appeal held the owners’ refusal to load was objectively reasonable on the material available at the time, and corrected first-instance errors that effectively demanded proof of continuing control rather than a reasonable assessment of sanctions risk.