The Court of Appeal has confirmed that, when assessing whether or not a communication is privileged, a “dominant purpose” test must be applied not only to claims of litigation privilege, but also to claims of legal advice privilege.

The court also held that a communication which simultaneously seeks legal advice from a lawyer and commercial comment from a non-lawyer will not normally be privileged, unless the purpose of the commercial comment is specifically to settle the instructions to the lawyer. In reaching this conclusion, Hickinbottom LJ gave useful practical pointers as to how such communications should be analysed and on the approach to be taken to email attachments.

Finally, the court clarified that, when a party voluntarily waives privilege in a communication, the scope of any collateral waiver of privilege in related communications is determined by the purpose of the original voluntary waiver, the aim being to ensure that the court is not given a misleading or unfair impression of the issue in question.


R ( v Civil Aviation Authority (CAA) [2020] EWCA Civ 35 is an interim decision in judicial review proceedings which Jet2 has brought against the CAA over the latter’s decisions to publish (i) a press release criticising Jet2 and other airlines for opting out of an industry ADR scheme and (ii) an exchange of correspondence between the CAA and Jet2 after Jet2 objected to the press release.

One of the issues for disclosure in the judicial review claim was the CAA’s reasons for publishing the correspondence. An email which the CAA disclosed in relation to this issue showed that one executive had favoured an aggressive response to Jet2’s complaint. In order to put this in context, the CAA voluntarily disclosed a later email, referred to in the judgment as “the 24 February email”, which showed that this approach was not shared by others involved in the decision. This email attached a draft response to Jet2’s complaint and was sent to an in-house lawyer for legal comment as well as to a number of non-legal executives for commercial input.

At first instance, Morris J held that the 24 February email was not privileged and that, even if it had been, by voluntarily disclosing it, the CAA had waived privilege in all communications concerning its response to Jet2’s complaint.

Issues in the appeal

The CAA’s appeal focused on four issues:

  1. Should a dominant purpose test be applied when considering whether or not the 24 February email was protected by legal advice privilege?
  2. How should the court approach the privilege analysis where a communication was addressed to both lawyers and non-lawyers?
  3. Should emails with attachments be considered as a single communication, or should each component be analysed separately?
  4. What is the scope of the collateral waiver which may result from a voluntary disclosure of a privileged document?

Issue 1: Legal advice privilege and dominant purpose

Giving the leading judgment, Hickinbottom LJ conducted an extensive review of English and other common law (especially Australian) authorities on dominant purpose. On this basis, he concluded that there was no distinction between litigation privilege and legal advice privilege in this regard, and the dominant purpose test applied to both. Although a different Court of Appeal panel in SFO v ENRC [2018] EWCA Civ 2006 had suggested such a distinction, the comments were obiter and made without the benefit of a review of previous authorities on the issue.

Issue 2: Multi-addressee communications

In analysing the 24 February email, Hickinbottom LJ set out a suggested practical approach to multi-addressee communications that include both lawyers and non-lawyers, which can be summarised as follows:

  • The first step is to establish why the lawyers are involved in the chain of communication in the first place. If they have become involved in order to give their commercial views, that will generally be taken to be the purpose of their involvement throughout, unless there is a specific request for legal advice later in the chain. However, this is only a starting-point, and the dominant purpose test must be applied separately to each communication in the chain.
  • If the lawyers have been included in order to give legal advice, the next step is to consider whether this is the dominant purpose. It may be helpful to ask whether or not the communication would have been privileged if it had been sent to the lawyer alone. If not, the communication clearly will not be privileged and that will be the end of the analysis. If yes, the effect of adding the non-lawyers has to be considered. Usually, privilege will only apply if any non-legal input is being sought specifically in order to settle the instructions to the lawyers. If non-lawyers are being asked to comment on any other aspect of the matter, then the dominant purpose will generally be considered to be commercial.
  • However, if as expected the lawyers then respond with legal advice, that response will almost certainly be privileged, even if the original request was for mixed purposes and even if the non-lawyer recipients continue to be copied. This is because the Three Rivers line of authorities (which the court questioned, but was bound to follow) takes a more lenient approach to the dissemination of legal advice than it does to the collation of the instructions on which the advice is based.
  • Responses from the non-lawyers will not be privileged, unless they are likely to disclose the tenor of the lawyers’ legal advice. The test for “likelihood” in this context is one of “realistic possibility”, not the balance of probabilities.

Issue 3: Separate consideration of emails and attachments

The court reiterated the established principle that a document which is not privileged at the time of its creation does not become privileged merely because it is later sent to a lawyer. This means that some degree of separate consideration of emails and their attachments will inevitably be required.

Issue 4: Collateral waiver

Strictly speaking, having concluded that the 24 February email was not privileged, the court did not have to consider whether disclosing it gave rise to a collateral waiver.However, Hickinbottom LJ took the opportunity to clarify that voluntary disclosure of a privileged document waives privilege only in relation to the issue in relation to which the voluntary disclosure has been deployed. This is not the same as (a) the subject matter of the disclosed communication, (b) the category of documents to which the communication belongs, or (c) all documents that would be “relevant” to the issue using the standard test for relevance in the disclosure context. Rather, the relevant documents are those which are required to avoid giving a misleading or unfair impression to the court. In this instance, the 24 February email was deployed in order to show that there was no general intention on the part of the CAA to take an aggressive approach. On the facts, fairness did not require any additional disclosure, and therefore there was no collateral waiver.


This judgment provides welcome clarity on the practical application of legal advice privilege. It will be particularly helpful for in-house lawyers who may be asked to advise on both legal and commercial issues or to provide training and policies for the business on communications protocols.

In light of the comment that mixing legal and commercial communications will usually result in an assumption that the dominant purpose is commercial, a key principle in such policies should be that commercial and legal discussions should be kept separate, even if this means starting two email conversations or holding two separate meetings. Any attempt to intertwine the two is highly risky and is more likely to lead to losing privilege in the instructions to in-house lawyers than to shield the commercial discussions. However, mistakes inevitably happen, and provided lawyers who receive a mixed-purpose communication take care only to address the legal aspects of the request in their reply, there is a good chance that the legal advice itself will still be protected.

Businesses should also be aware of the danger of collateral waiver when sharing a document that may be privileged, and should always discuss the risk-benefit analysis with internal or external lawyers before making a voluntary disclosure.

Finally, the judgment spends considerable time (going beyond what was strictly required in order to resolve the issues in the case) highlighting the difficulties that have been caused for both courts and large businesses by the Three Rivers line of cases, which restricts privilege to a very narrowly defined “client”. In doing so, it lends further support to the calls already made in SFO v ENRC for the Supreme Court to reconsider this area at the earliest opportunity. It will be interesting to see whether this case is itself appealed to the Supreme Court – although, given that much of what is said about Three Rivers in the judgment is obiter, that will not necessarily guarantee a definitive ruling.