Evidence of Corruption from Parallel Criminal Proceedings: Ignore Rather than Follow? Lessons From Vitrus v. Thales and Veolia

Evidence from parallel criminal proceedings can be both difficult to reconcile with arbitral procedure and highly persuasive when relied upon in arbitration proceedings where corruption allegations have been raised. The difficulty stems from the differences between the procedural frameworks governing criminal and arbitral proceedings, and arbitral tribunals have been increasingly confronted with the question of how to treat evidence that originates from, or is used in, parallel criminal proceedings. This question is made more pressing by the absence of a unified evidentiary standard: applicable procedural rules instead grant arbitrators wide discretion to determine the admissibility, relevance, materiality, and weight of any evidence in a case.

As previously discussed on this Blog, investment arbitration tribunals have recognised that tribunals can reach conclusions on corruption that are contrary to national investigations or criminal proceedings, for example because domestic findings function in “different legal spheres”, are subject to “diverging standards of proof”, do not meet the basic requisites of res judicata, are tainted with bias, or are arbitrary. There is therefore an established consensus that domestic findings are not binding on investment tribunals, although they may be treated as persuasive evidence and one of several factors in the tribunals’ evaluation of corruption claims.

While the discussion above concerns investment arbitration, recent case law shows a similar tension in international commercial arbitration, with arbitrators increasingly deferring to, or aligning with, national authorities on corruption allegations. This post therefore examines that trend through two recent commercial arbitration awards, before returning to the question: how far is too far?

Recent Illustrative Case Law

Vitrus v. Thales

The judgment of the Amsterdam Court of Appeal of 24 June 2025 (“Judgment”) has brought up this issue into the spotlight again, as it found that an ICC arbitral tribunal relied on evidence that had been previously declared inadmissible by national proceedings, in violation of Dutch public policy.

The Court of Appeal in that case suspended the annulment proceedings brought by a Brazilian consultancy company, Vitrus Consultoria de Mercados S.A. (“Vitrus”), in relation to an ICC arbitration award of 7 June 2023 in its dispute against a Dutch holding company within the global group operating in the fields of defence, aerospace and information technology, Thales International Latin America Holding B.V. (“Thales”), for the construction of a new satellite system in Brazil.

Thales suspended its payment obligations under the consultancy agreements after allegations of corruption emerged in the press against an officer and shareholder of Virtus, who allegedly made illicit payments of approximately EUR 1.5 million to Brazilian authorities as part of the Odebrecht corruption scandal (named after Odebrecht S.A., a major Brazilian construction conglomerate involved in a multi-national bribery and corruption scheme).

The arbitral tribunal found that Vitrus had breached its contractual guarantees as to ethical conduct and dismissed Vitrus’s claims (Award, para 246, as referred to in para 4.20 of the Judgment). In reaching that conclusion, the tribunal relied on the statements made by Odebrecht employees during plea bargaining with the police and the printout data from Odebrecht’s internal systems, even though this evidence had twice been declared inadmissible by the Brazilian Federal Supreme Court because the chain of custody and technical soundness of that evidence were compromised. The tribunal, however, did not clarify in its award how it considered these circumstances or why it regarded the grounds of inadmissibility as merely “technical” (Judgment, paras 4.11-4.12).

The Court of Appeal held that an arbitral award based on such evidence, without sufficient justification and adequate reasoning, is contrary to Dutch public policy (Judgment, para 4.15). Notably, it did not annul the award but rather suspended the annulment proceedings for six months and remitted the award to the tribunal, providing an opportunity to address the public policy violation given that the tribunal’s reasoning referred to other admissible means of evidence (Judgment, para 4.16). On 16 December 2025, a revised arbitral award was issued “in compliance with the decision of the Amsterdam Court of Appeal”, according to the Order of the Superior Court of Justice of Brazil No. 10929 from 10 February 2026, although the content of the revised award is not publicly available.

Veolia

A similar issue of overlapping evidence arose in relation to the award of 30 November 2023 in Veolia & Vilniaus Energija v. Vilniaus šilumos tinklai (Lithuania) & Vilnius City Municipality, SCC Case No. 2016/183 (“Veolia”). The respondents argued that Veolia’s representatives “engaged in bribery systematically and reflexively” with numerous Lithuanian officials, and that Veolia had failed to implement basic compliance controls to conduct meaningful investigations (Award, para 314). The evidence of the alleged corruption scheme included both direct evidence (black accounting documents and wiretaps transcripts) and circumstantial evidence, the former of which had been collected during a six-year investigation by national prosecution authorities in Lithuania. That investigation was closed without any finding of corruption, and the direct evidence was not available during the subsequent arbitration proceedings.

The arbitral tribunal in Veolia chose not to decide whether the parallel national investigation precluded it from re-examining the same facts and evidence, or bore a res judicata effect. Instead, the tribunal opted for a vaguer formula, finding itself bound to consider the general impact of the national investigation (Award, para 334). It relied on the fact that the direct evidence under its consideration had already been scrutinised by national authorities and found insufficient to establish corruption (Award, para 333). While acknowledging that circumstantial evidence can support a finding of corruption where direct evidence is unavailable, the tribunal decided that without direct evidence of bribery payments the respondents’ case was “bound to fail as insufficiently proven” (Award, para 332).

Takeaways

Both Veolia and Vitrus v. Thales are examples of the arbitrators’ inclination to follow the national authorities’ conclusions, whether as to the admissibility of evidence or its sufficiency to justify a finding of corruption.

These cases, however, raise at least three questions which remain unanswered:

1. Which law, and which evidentiary rules and principles embodied therein (e.g. standard of proof) apply to the assessment of corruption evidence in the arbitration proceedings? In Veolia, the tribunal did not engage with the parties’ disagreement on this issue, even though both agreed it was bound by the fundamental principles of evidence in international arbitration enshrined in instruments such as the UNCITRAL Model Law on International Commercial Arbitration and the IBA Rules on the Taking of Evidence in International Arbitration (including due process, equality of arms, and the adversarial principle). As a result, the tribunal did not address the respondents’ argument that the very high standard of proof applied by Lithuanian authorities in bribery cases should not apply in arbitration. The claimants, for their part, argued that the respondents’ position was not grounded in any applicable standard of proof, but instead rested on the premise that arbitral tribunals bear an inherent duty to root out corruption so as to uphold the values of the international community. Arbitral tribunals should therefore take greater care in explaining which evidentiary principles they apply, and which legal system governs their content, since that content can differ significantly between legal systems.

2. The principle of equality of arms means that evidence can only be relied upon once its authenticity is sufficiently established. However, if that evidence is not available during the arbitration, as in Veolia, could a tribunal still rely, without further analysis or justification, on the findings of a criminal investigation without taking steps to gain access to such evidence, especially where circumstantial evidence indicating corruption is available? Similarly, if the evidence used in a criminal investigation was considered tampered with, as in Vitrus v. Thales, should a tribunal question whether such inaccuracies distort the essence of that evidence, or are mere irregularities in the chain of custody that suffice to render it inadmissible? Arbitral tribunals should ensure that they independently assess the evidence in question rather than blindly follow or discard the findings of national authorities without further analysis.

3. In Vitrus v. Thales, the findings of the Brazilian Supreme Court were made in unrelated criminal proceedings, though based on the same underlying documents as the indictment of Vitrus’s officer. In Veolia, it was unclear whether the allegations assessed by Lithuanian authorities were the same as those raised in the arbitration, whether the underlying facts were the same, whether the standard of proof applied by national authorities in finding the evidence “insufficient” matched that applied by the tribunal, and whether the evidence in question was exactly (rather than “largely”) the same as that used in the national investigation. Given this uncertainty, could the tribunal’s conclusion truly be as unequivocal as it purported to be?

Conclusion

Ultimately, both Vitrus v. Thales and Veolia show that arbitral tribunals cannot treat findings from parallel criminal proceedings as a substitute for their own independent assessment of the evidence. Where a tribunal aligns with, or defers to, the national authorities’ conclusions, it should clearly articulate the legal framework applied, satisfy itself that the underlying facts, allegations, and evidentiary standards are genuinely comparable, and explain why procedural irregularities in the parallel proceedings do not undermine the reliability of the evidence relied upon. Absent such reasoning, awards leaning on national findings – whether to establish or dismiss corruption allegations – remain vulnerable to challenge on public policy grounds, as the Amsterdam Court of Appeal’s decision demonstrates.