Skip to main content

Litigation privilege in the context of Deferred Prosecution Agreements (DPAs) – an analysis of the High Court's treatment of litigation privilege issues in ENRC and the consequent Court of Appeal judgment



United Kingdom

There was relief at the rejection of the High Court's judgment especially in relation to the ruling that litigation privilege can attach to internal investigation reports and their underlying...

This article was first published in Fraud Intelligence.


Last month saw the long awaited release of the Court of Appeal's judgment in the case of Director of the Serious Fraud Office v Eurasian Natural Resources Corp Ltd [2018] EWCA Civ 2006 ("ENRC"). There was obvious relief at its wholesale rejection of the High Court's judgment especially in relation to the ruling that litigation privilege can attach to internal investigation reports and their underlying documents. However, the judgment of Andrews J, in the High Court, had done more than call just this one aspect into question. She had attempted to reduce the scope of documents and communications to which litigation privilege can apply in a number of ways which was likely to have serious repercussions for the incentives offered by the DPA regime to companies had the Court of Appeal agreed with her ruling.

The concept of cooperation with a government authority when things start to smell a little fishy has always involved a bit of a leap of faith. If that fishy smell turns out to be indeed a great pile of rotting sharks within your organisation then that leap might pay off with that authority treating you that bit more leniently than if they had had to do all the investigative legwork. The key word here is, of course, "might" – there is no guarantee of leniency post the required soul baring. The balance of the DPA regime is, as such, a delicate one. The incentives to self-report and cooperate must be sufficiently high to overcome the potential penalty.

The incentive to self-report was severely eroded in early 2017 when Rolls Royce, despite not self-reporting, was given a DPA by the SFO. Rolls Royce had cooperated with the SFO from the time it learned of the SFO's knowledge of suspected wrongdoing, but this alone should not have been enough. Companies which become aware of possible wrongdoing within their organisation must now ask themselves why should they self-report when instead they could simply cooperate upon discovery by an authority.

The judgment of Andrews J in ENRC had also significantly reduced the size of the carrot of lenient treatment by casting doubt over one of the most basic legal protections, that of litigation privilege. Litigation privilege is a key protection in the context of internal investigations where establishing the facts might include the creation of documents and records by third party forensic accountants or interviews with employees of the organisation.

Andrews J took quite a number of aims at limiting the circumstances in which litigation privilege can apply which have since been put right by the Court of Appeal, but let's take a look at what happened while breathing a huge sigh of relief.

Litigation must be in progress or in contemplation – the evidence threshold and who must know about it

In dealing with the first limb of the test for litigation privilege, as set out in Three Rivers DC v Bank of England (Disclosure) (No.4) [2004] UKHL 48, [2005] 1 A.C. 610 ("Three Rivers"), Andrews J drew a distinction as to whether the potential proceedings are civil or criminal as the latter has an evidence threshold which must be overcome before an action can be commenced. She found that in order for criminal proceedings to be in contemplation that the evidence test must be met and, critically, she found that the converse is also true – until sufficient evidence is unearthed in an investigation to meet the evidence threshold, litigation cannot be considered to be in contemplation.

Pausing to let that sink in - this would mean that the investigation stage would have been more or less completed and proceedings all but imminent before the party under investigation could be said to be in contemplation of litigation. This flies in the face of common sense. And, thankfully, this is what the Court of Appeal found, stating that

"whilst a party anticipating possible prosecution will often need to make further investigations before it can say with certainty that proceedings are likely, that uncertainty, in our judgment, does not in itself prevent proceedings being in reasonable contemplation".

Flaux J in R v Paul Jukes [2018] EWCA Crim 176 ("Jukes"), a case heard between the ruling of the High Court and the Court of Appeal in ENRC, concurred with Andrews J that the evidential test must be met before litigation can be contemplated finding that

"there is no evidence … that at the time that these investigations…, any of them had enough knowledge … that they appreciated that it was realistic to expect the Health and Safety Executive to be satisfied that it had enough material to stand a good chance of securing convictions."

The Court of Appeal in ENRC commented on the judgment in Jukes finding the reference to Andrews J's judgment on this point to be obiter.

Andrews J, and Flaux J to an extent, also struggled with who must hold the knowledge that the internal investigation has unearthed sufficient evidence. Logically the requisite knowledge must be held in the mind of the prospective defendant. However, Andrews J appeared to also require that, in order for the prospect of litigation to be more than a general apprehension, the prosecutor must also be aware to some degree (she did not comment on the extent of that degree) that something has been unearthed and that the prospective defendant is aware of the prosecutor's knowledge (to quote a classic Friends episode that this brings to mind – "they don't know that we know that they know that we know…"). Andrews J commented on information that became available to ENRC during its internal investigation causing that investigation to be widened in its scope stating that

"there is no evidence that this further allegation had become known to anyone outside ENRC or that ENRC thought that it had. There could be no prosecution for a criminal offence, unless and until the prosecutor became aware of the matters which could give rise to such a prosecution."

Of course there can be no prosecution until the prosecutor knows the facts, but that is not the question. The question was can the prospective defendant contemplate litigation in respect of facts that he knows, but are not yet known to the prosecutor? Extrapolating from Andrews J's tautology she appears to be saying that until the prosecutor knows something, a prosecution cannot be commenced and therefore cannot be contemplated by a defendant even though the defendant might know, in relation to the newly uncovered material that it would surpass the evidential threshold required.

Where that which is uncovered in the course of the investigation is entirely separate from the remit of the existing investigation, was unexpected and the prospective defendant knows that the prosecutor has no way of knowing or suspecting of its existence and the prospective defendant knows that he is not going to inform the prosecutor then perhaps this is correct, but this is not a realistic context in which internal investigations are conducted.

When conducting internal investigations it is commonplace for the scope to be widened to encompass more avenues of exploration. It is rare that these avenues are so separate and cut off from the the original remit of the investigation that it can be said that that which is on one side of the line is covered by litigation privilege as the prosecuting authority is aware of it and that on the other is not as the prosecutor has not (yet) been informed of it and that it is known or assumed that the prosecuting authority has no knowledge as to its existence. It is often not possible for the defendant to say what the prosecutor knows and what it does not know. It cannot be that until the prosecutor has been informed of the widening in scope and the preliminary results yielded from it that litigation privilege will not apply.

The sole or dominant purpose of conducting litigation – documents created to be shown to the other side and documents created to settle prior to litigation or avoid litigation

Perhaps with one eye on the third limb of the Three Rivers test for litigation privilege Andrews J again got caught up wrestling with the question of where an investigation ends and litigation begins in relation to the second limb of the test for litigation privilege.

When addressing the requirement that the communications be for the sole or dominant purpose of litigation Andrews J relied upon the judgment of Goldberg J in the Australian case of Bailey v Beagle Management Pty [2001] FCA 185. Part of the quotation from this case that Andrews J relied on in her judgment is as follows:

"Properly characterised, it is not correct to say that a document is brought into existence for the purpose of conduct of litigation, and so is privileged from production, if it is brought into existence, albeit to try and settle the litigation but for the purpose of being shown to the other side."

The principle Andrews J drew from this and followed was that where a document is created for the purpose of being shown to the other side it does not attract litigation privilege. (That is indeed the conclusion of the Australian case.) The document in question in that case was an "independent chartered accountant's report" commissioned for the purpose of exploring settlement. It was based on a number of attachments including

"assessments of, and advice by, solicitors for [the organisation] and [the organisation's] paralegal staff and matters relating to the prospects of success of proceedings to which [the organisation] was and is a party. It was said that these matters are the foundation of the document and the conclusion expressed in it."

Notable is that the party seeking disclosure was not seeking disclosure of the attachments – only the document itself.

In ENRC the documents sought included the underlying documents upon which the report produced for the SFO by ENRC is based, which includes, for example, notes of the evidence of employees, ex-employees and employees of subsidiaries and materials generated by external forensic accountants. Andrews J ruled that these documents were created for reports to be based on and those reports were made with the knowledge and expectation that the SFO might wish to verify the reports by way of an audit of the underlying documents and therefore

"no legitimate distinction can be drawn between the reports and the underlying materials in terms of the purpose for which they were created."

This was of serious concern as the vast majority of internal investigative reports are concerned with such material.

Goldberg J's talk of the documents in the Australian case being related to settlement also, somewhat understandably, caused Andrews J some difficulty. In that case it was stated that

"[a]lthough the document came into existence for the purpose of exploring the possibility of settlement … I am satisfied that the purpose for which the document came into existence was to be passed on to the trustee for the purpose of … settlement ..."

On plain reading of Goldberg J's words the issue of whether settlement pre or post commencement of proceedings is or is not part of litigation was not of concern to him. The question Goldberg J was wrestling with was whether litigation privilege attaches to documents created with the purpose of being shown to the other side.

Andrews J, however, interpreted Goldberg J's as saying that, firstly, litigation privilege does not attach to documents created for the purpose of being shown to the other side, but, secondly, that it can attach to documents created for settlement purposes provided that proceedings are in train, when, in fact, he did not come to this conclusion much less ask himself the question. Andreww J concluded that

"[t]he conduct of ongoing proceedings embraces litigation tactics, and must include bringing them to an end by agreement short of trial [EMPHASIS ADDED]."

It was a strange conclusion to draw that use of the full gamut of litigation tactics during litigation proceedings attracts litigation privilege, but in the pre-litigation stage when litigation is contemplated the use of such tactics does not.

This point was addressed by Sir Geoffrey Vos in Bilta (UK) Ltd (in Liquidation) & Ors. v (1) Royal Bank of Scotland Plc, (2) Mercuria Energy Europe Trading Limited [2017] EWHC 3535 (Ch) ("Bilta"). The judgment of Bilta was handed down between the ENRC judgment of the High Court and that of the Court of Appeal. Bilta turned on slightly different facts as once HMRC considered that it had sufficient evidence to overcome the evidentiary threshold it wrote a letter found by the Court to be akin to a letter of claim to the company setting out the law and its application to the facts. Both parties from that moment were aware of the facts considered by HMRC at least to be sufficient to pass the evidential test for prosecution therefore allowing litigation to be contemplated as per the first limb of the test for litigation privilege according to Andrews J's interpretation.

The documents over which litigation privilege was claimed in Bilta were created after the date of that letter, but still before the commencement of proceedings. The company took various steps to indicate it was gearing up for litigation, such as instructing external solicitors specialising in tax litigation. However, the relationship remained cooperative. The company provided HMRC with updates as to its internal investigation. Its external solicitors also created a report, which was the subject of the judgment, which was intended to be provided to HMRC in response to the letter akin to a letter before action. The company argued that such cooperation was for the purpose of complying with its duties as a taxpayer and in order to seek to persuade HMRC to change its mind that is to avoid litigation.

It is to these actions to which Sir Geoffrey Vos sought to reconcile Andrews J's judgment in ENRC. Sir Geoffrey Vos neatly sidestepped taking Andrews J's judgment head on instead taking the approach adopted in Highgrade Traders Ltd, Re [1984] BCLC 151 stating that a communication for a subsidiary purpose, such as avoiding litigation, can be subsumed into the dominant litigation purpose. Sir Geoffrey Vos pointed to the unusual facts of ENRC as the basis for Andrews J's bizarre conclusion that attempts to avoid litigation that has not yet commenced cannot be considered to be for the sole or dominant purpose of litigation saying that he would not "accept… that one can properly draw a general legal principle from her approach to those facts".

The Court of Appeal rejected the approach of Andrews J unequivocally stating that

"[i]n both the civil and criminal context, legal advice given so as to head off, avoid or even settle reasonably contemplated proceedings is as much protected by litigation privilege as advice given for the purpose of resisting or defending such contemplated proceedings."

The litigation must be adversarial, not investigative or inquisitorial.

On the third limb of the test for litigation privilege Andrews J drew a distinction between the SFO's role as investigator and its role as prosecutor. She held that an investigation by the SFO is not adversarial litigation stating that

"documents generated when there is no more than a general apprehension of future litigation cannot be protected by litigation privilege regardless of the inevitability of an investigation."

Flaux LJ in Jukes agreed with Andrews J's reasoning also stating that at the time of the relevant statement in that case

"no decision to prosecute had been taken by the Health and Safety Executive and matters were still at the investigatory stage".

This potentially attempts to constrict to adversarial litigation to communications created after the decision by the prosecuting authority to start proceedings.

The Court of Appeal rejected this separation of the SFO's roles concurring with Etherton LJ in Westminster International BV v Dornoch [2009] EWCA Civ 1323 ("Dornoch") that while not "every SFO manifestation of concern would properly be regarded as adversarial litigation" nor that "once an SFO criminal investigation is reasonably in contemplation, so too is a criminal prosecution", but that each case must be assessed on its own facts.


The SFO has indicated that it does not intend to take the case to the Supreme Court. While this is a pragmatic decision this is an area that could benefit from judicial opinion.

The three cases of ENRC, Bilta and Jukes, referred to above, have all been heard within the last year and all concerned litigation privilege and its application to third party documents created in the context of investigations arising from the interface between government authorities and private entities.

All three of the cases sought to apply the test for litigation privilege as set out in Three Rivers. However, the three cases have attributes not present in Three Rivers which have raised difficulties in its application. All three involved scrutiny by government authorities of corporates; all three involved internal investigations either with external investigations being conducted in parallel or threatened by the government body concerned; and two involved cooperation and coordination with the government authority with a view to being treated with leniency.

The Three Rivers test for litigation privilege fails to cater for the circumstances surrounding internal investigations occurring at the interface between government authorities and private entities against the backdrop of cooperation with the constant threat that the authority acting as external investigator may at some point rip off its hat of investigator to don, with fire and fury, that of prosecutor.

For the DPA regime to be successful and for its incentives to self-report and cooperate this is an area where the legal protection of privilege, as well as when and how it applies, needs to be clearly understood by all parties to any DPA. It is, therefore, somewhat disappointing that this case will not be taken to the Supreme Court as we could do with some definitive guidance as to how privilege operates in this area, but it is understandable that Lisa Osofsky, the new Director of the SFO (and from a jurisdiction with great regard for privilege) is choosing not to start her tenure at the SFO challenging this popular Court of Appeal judgment.


Sign up to our email digest

Click to subscribe or manage your email preferences.