The Security of Mediation and “The Interests of Justice”

by Tony Allen

1. Mediation – A Common Law Non-Statutory Concept

The abiding value of the common law has always been in its functioning as a flexible tool for legal development, with principles being applied to new sets of facts, sometimes leading to new law without the need for the precision of statutory language or the necessary investment of Parliamentary time.  The law of England & Wales has never embraced systematic codification typical of civil law jurisdictions, and judicial precedent is particularly important for mediation, which is not in any way a creature of statute or rule made under the force of statute.  We have no Uniform Mediation Act or similar legislation here as a framework to regulate mediation such as exists in a number of US States.  Mediation here sits firmly within both the common law and civil procedure, based on principles found in contract, evidence, and breach of confidence.  This in turn necessitates a close watch on court judgments whenever they impinge upon the status of mediation, either by developing its preferred status or possibly undermining it.

No topic is more important for the health of mediation than its status as a broadly confidential forum.  It offers to participants the opportunity for secure discussion of the possibility of settlement in a way that protects the parties from any adverse consequences.  This protects parties whatever they may do or say or indeed refuse to do or say within the confines of the process, and whether or not they settle.  Even if on objective analysis it was unreasonable for a party not to settle on offered terms within a mediation, that decision is kept from the judge at a later trial. Second only to the significance of confidentiality is the question of the extent that mediators are protected from being called as potential witnesses as to what happened during a mediation, which itself might lead to disclosure of confidential material.  Questions over whether courts can order mediation or might impose costs sanctions for unreasonable failure to engage are of subsidiary, albeit not trivial, significance.   If a party could suddenly find what they said and did during a mediation, or their refusal to settle on offered terms, laid before a court as evidence against them at a litigated trial at the unilateral behest of their opponent, it would seriously undermine the secure fabric of the mediation process.  Equally, for a party to find the mediator compelled by another party to give evidence in open court, however unwillingly, about what was said and done would also undermine trust in the mediator and the mediation process for the future.

It cannot be said that the law as it has developed in relation to mediation in England & Wales is free from doubt, as will emerge from this article.  Recent decisions amount to a patchwork of cases at first instance which are not always easy to reconcile, though the attempt needs to be made.  This article has been prompted by three recent first instance decisions which touch upon mediation confidentiality – EMW Law v Halborg[1], Briggs v Clay[2] and Berkeley Square Management v Lancer Properties[3], the last of these to be re-evaluated in the light of the appeal hearing which followed[4].  However, an understanding of these decisions requires a careful look at the foundations on which mediation confidentiality stand.   Each of these recent decisions included a similar review of past cases, and the debate generated by such reviews (including this one) is not free from complexity, inconsistency and perhaps even error.


2. Without Prejudice Privilege

There is no doubt that, as a process intended and designed as a means for exploring settlement of disputes and claims, every mediation is at common law automatically protected by the evidential privilege usually called “without prejudice” (for convenience abbreviated in this paper to WP, including in quotations from judgments).  What this privilege does as a minimum is to prevent evidence of any admissions as to any justiciable issue in litigation made during a mediation (or indeed any WP negotiations to settle a dispute) being admitted as proof of the truth embodied in that admission.  These are called “admissions against interest”:  if admissible, they would “prejudice” the case of one (or possibly both) parties, because the opposing party could adduce the admission to prove the fact admitted.  When someone admits a material fact openly and on the record, it is admissible as evidence to prove that fact, even when it is technically hearsay.  But if that admission is made for the purposes of exploring the settlement of a dispute, whether within or outside litigation, it is automatically privileged from being given in evidence, as long as there is a real dispute which could be litigated if not settled.  This will cover the vast majority of important inter-party exchanges leading towards and made within, and even following, a mediation. So offers by a claimant to settle for less than what is openly claimed, or to pay more than what a defendant openly accepts (or indeed to pay anything at all when liability is openly denied) will all be protected from being given in evidence in subsequent litigation between the same parties.  Preparatory expressions of willingness to consider settlement discussions or a proposal to mediate might also be regarded as admissions at least of the possibility of compromise and therefore attract WP privilege.  But the essence of mediation confidentiality was tellingly expressed by William Wood QC in his lively essay on the boundaries of mediation confidentiality “When Girls Go Wild”[5]:

If the mediation does not succeed then nobody will later be able to ask the question: “Mr. Jackson, if you are so sure that the widgets you supplied were compliant with the contract why did you offer to pay £60,000 to settle the case in a mediation last month?”.

There is no need to preface or mark such communications as “WP” to create the privilege, as it arises by operation of law when negotiations to settle a dispute take place[6].  For the same reason it is not strictly necessary to spell it out as applicable in a formal mediation agreement, though this is normally done.  To do so establishes WP privilege by express agreement between the parties on a contractual basis, one of the two recognised rationales for WP privilege.  The other identified rationale for WP privilege does not depend upon express or implied agreement between the parties, but instead on a principle of public policy – the need to encourage and allow parties to explore settlement without adverse consequence.  This basis is consequently of wider scope and effect, as it may affect potential access to evidential material by third parties, despite their being external to WP privilege created by contract between other disputing parties.

For WP privilege to be waived, both or all parties to the privilege, whichever basis is applicable, must agree to such waiver. However, in relation to a mediation, neither the mediator or any mediation provider organisation has any say over waiver, even though both of the latter normally sign the mediation agreement which sets up the process.  Parties have on occasions mutually chosen to waive privilege protecting a mediation, allowing the judge to be told of offers made within the mediation process, each trying to convince the judge that they were the reasonable party in the mediation.  One of them has usually been proved to be misguided in doing so and faced a costs sanction because the judge thought that party was unreasonable[7]!

Several problems remain.  Firstly, is the scope of protection narrowly limited only to admissions against interest or is it wider?  Discussions during settlement meetings do not simply involve admissions against interest, even if these are what might lead to settlement, or be what an opposing party would want to utilise as evidence at trial because negotiations have failed.  They may include representations as to fact, views as to the applicable law, assertions of opinion on facts or rights intended to influence or alter another party’s approach, speculation as to possible consequences of different courses of action, suggestions about concessions for the purposes of negotiation, and many other types of exchange.   Are all these ancillary exchanges capable of being admitted as evidence rather than protected by WP privilege?  And who does WP privilege protect and against whom?  What exceptions to WP privilege are recognised and how are they relevant to mediation?  Is there a difference in the way courts treat WP privilege as between two parties to the same litigation (“two party situations”) as opposed to where a third party outside the scope of WP privilege (“three party situations”) seeks access to WP material?  Does it make any difference in a “three party situation” if the challenge to privilege is raised by a party in the same piece of litigation or where it is raised in a different separately generated claim?

Additionally, there is the question as to the scope and effect of contractual confidentiality as created by the mediation agreement before each mediation, and what if any is the difference in effect between that and WP privilege.  And finally, what practical difference does the current state of law (whatever it turns out to be) make to the confidentiality and security of the mediation process and its conduct? Each of these will be considered in turn, looking first, and at length, at WP privilege.


3. The Scope of WP Protection

The modern law as to WP privilege originates in Cutts v Head[8] and Rush & Tomkins v GLC[9].  In the former case, the Court of Appeal established that the rule is based partly on the public policy which encourages parties to try to settle without adversely affecting their asserted case, and partly on express or implied agreement between the parties that what they exchange will be protected from the trial judge.  Either or both might apply on given facts.  In two party disputes it is likely to be by agreement, and the usual provision in a mediation agreement makes it explicitly so.   Whereas if a party to WP discussions is seeking to resist access by a third party to those normally privileged discussions (a three-party situation), the third party was not privy to that agreement as to WP privilege, so that privilege can in those circumstances depend only on the extent of public policy protection.

In Cutts v Head, a two-party case, one party sought to adduce in evidence an offer they had made in a WP letter which had specifically reserved the right to rely on it in relation to the costs award after trial, at which the offeree did worse than the offer made earlier.  For the first time, the Court of Appeal in Cutts recognised what are now known as “Calderbank offers” as having force outside the Family Division, admitting the offer in evidence in dealing with costs issues because the right to refer to it had been specifically reserved.  But if the offer had been in a letter simply marked “WP” (with no reservation as to costs) they found that it would not have been admissible to determine costs issues.  This established that parties can agree to define the limits of WP protection by express or implied agreement.

Rush & Tomkins v GLC and P.J. Carey, was a three-party case, although all were parties to the same piece of litigation.  R&T compromised a dispute with GLC over a building contract in a mediation, but PJC, who were the second defendant in the same construction proceedings but were not at the mediation, wanted access to documents produced within the mediation to see whether their outstanding claims against R&T had been fairly treated in the deal between R&T and GLC.  R&T disclosed the settlement agreement but declined to disclose any intra-mediation documents.  Lord Griffiths delivered the lead judgment in the House of Lords, and said:

The rule applies to exclude all negotiations genuinely aimed at settlement whether oral or in writing from being given in evidence. A competent solicitor will always head any negotiating correspondence WP to make clear beyond doubt that in the event of the negotiations being unsuccessful they are not to be referred to at the subsequent trial. However, the application of the rule is not dependent upon the use of the phrase WP and if it is clear from the surrounding circumstances that the parties were seeking to compromise the action, evidence of the content of those negotiations will, as a general rule, not be admissible at the trial and cannot be used to establish an admission or partial admission. I cannot therefore agree with the Court of Appeal that the problem in the present case should be resolved by a linguistic approach to the meaning of the phrase WP.  I believe that the question has to be looked at more broadly and resolved by balancing two different public interests namely the public interest in promoting settlements and the public interest in full discovery between parties to litigation.

Nearly all the cases in which the scope of the WP rule has been considered concern the admissibility of evidence at trial after negotiations have failed. In such circumstances no question of discovery arises because the parties are well aware of what passed between them in the negotiations. These cases show that the rule is not absolute and resort may be had to the WP material for a variety of reasons when the justice of the case requires it. It is unnecessary to make any deep examination of these authorities to resolve the present appeal but they all illustrate the underlying purpose of the rule which is to protect a litigant from being embarrassed by any admission made purely in an attempt to achieve a settlement.

He goes on to consider a list of exceptions to the WP rule, which are later repeated more extensively by Robert Walker LJ in Unilever v Proctor & Gamble, discussed below.

Then turning to the case in hand he says:

I cannot accept the view of the Court of Appeal that Walker v Wilsher[10] is authority for the proposition that if the negotiations succeed and a settlement is concluded the privilege goes, having served its purpose. In Walker v Wilsher. the Court of Appeal held that it was not permissible to receive the contents of a without prejudice offer on the question of costs and no question arose as to the admissibility of admissions made in the negotiations in any possible subsequent proceedings. There are many situations when parties engaged upon some great enterprise such as a large building construction project must anticipate the risk of being involved in disputes with others engaged on the same project. Suppose the main contractor in an attempt to settle a dispute with one subcontractor made certain admissions, it is clear law that those admissions cannot be used against him if there is no settlement. The reason they are not to be used is because it would discourage settlement if he believed that the admissions might be held against him. But it would surely be equally discouraging if the main contractor knew that if he achieved a settlement, those admissions could then be used against him by any other subcontractor with whom he might also be in dispute. The main contractor might well be prepared to make certain concessions to settle some modest claim which he would never make in the face of another far larger claim. It seems to me that if those admissions made to achieve settlement of a piece of minor litigation could be held against him in a subsequent major litigation it would actively discourage settlement of the minor litigation and run counter to the whole underlying purpose of the WP rule. I would therefore hold that as a general rule the WP rule renders inadmissible in any subsequent litigation connected with the same subject matter proof of any admissions made in a genuine attempt to reach a settlement. It of course goes without saying that admissions made to reach settlement with a different party within the same litigation are also inadmissible, whether or not settlement was reached with that party.

So WP privilege was found to exclude a third party from access to the content of settlement discussions in a mediation even though they had an identifiable interest in their outcome.  Furthermore, the fact that agreement had been reached on the issues debated in the mediation did not subvert or terminate WP protection – the WP discussions remained inaccessible to that third party.

One example of reliance on this authoritative statement is Instance v Denny Bros[11], when it was cited with approval by Lloyd J in granting injunctions to restrain the use of WP correspondence, communications during settlement meetings and also during a mediation on CEDR standard terms from use in later proceedings in the USA relating to wider issues between the same or related parties.


4. A Narrow Approach to Privilege? – Hoffman LJ in Muller v Linsley & Mortimer

The breadth of the Rush & Tomkins view, encompassing apparently both two-party and three-party (and indeed multi-party) situations, was encouraging to those who sought to be able to attempt settlement within WP privilege without adverse consequences, and with only limited exceptions arising.  Its breadth of application is perhaps only qualified by Lord Griffiths’ remark that:

as a general rule the WP rule renders inadmissible in any subsequent litigation connected with the same subject matter proof of any admissions made in a genuine attempt to reach a settlement. It of course goes without saying that admissions made to reach settlement with a different party within the same litigation are also inadmissible, whether or not settlement was reached with that party.

Rush & Tomkins involved three parties in considering whether WP privilege applied, but all were parties to the same litigation.  Would the breadth of this approach apply where A and B settle, and B then starts a fresh action against C to recover a shortfall in the settlement between B and A, even if arising out of the same dispute?  This question provoked a major challenge to the broad approach to WP privilege in a 1995 Court of Appeal decision, for a while actually unreported, which suggested that the scope of this privilege was indeed tightly limited to prevent disclosure only of what might be strictly viewed as “admissions against interest”, interpreted narrowly. This principle was asserted by Hoffman LJ in his judgment in Muller v Linsley & Mortimer[12],  a three-party case.

Muller was a director and shareholder in a computer software company S Ltd.  He consulted the defendant solicitors L&M about a dispute with the other shareholders and directors, worried that he might be dismissed and then have to sell back his shares at significantly reduced value. L&M advised that while still in employment he should transfer his shares to his wife. S Ltd’s Articles required the Board to register such transfer but only if duly stamped. Because time was short, L&M presented the transfer unstamped, with an undertaking to have it stamped.  The Board rejected it and purported to dismiss Mr Muller and activate the compulsory sale. Mr Muller and his wife (MM) commenced proceedings against the other shareholders and the company. Negotiations followed leading to settlement comprising compensation for loss of office and loss of shares and an issue of new shares in the company which became worth over £2 million when the company was floated later, plus a contribution to legal costs.

MM then chose to sue their former solicitors L&M in fresh proceedings for negligent advice over the ineffective share transfer, alleging unrecovered losses. Their new lawyers pleaded that the settlement amounted to proper mitigation of loss, and in their list of documents asserted privilege in respect of “letters, telexes, faxes, and other communications and the notes of communications passing between [MM], third parties or [MM]’s solicitors and others marked WP or otherwise sent or made in an endeavour to settle the proceedings”.  As between MM and S Ltd, WP privilege was contractual, but as between MM and L&M the basis for the existence of WP privilege could only be founded on the “public policy” basis. MM did however disclose the settlement terms agreed between themselves and S Ltd, as had happened in Rush & Tomkins.  L&M challenged the claim for privilege for documents related to the mediated settlement discussions in a summons for disclosure which was dismissed by both the Master and Judge below. However, the Court of Appeal declined MM’s assertion of privilege and allowed L&M’s interlocutory appeal, with Hoffman LJ giving the following reason:

The public policy aspect of the rule is not in my judgment concerned with the admissibility of statements which are relevant otherwise than as admissions, i.e. independently of the truth of the facts alleged to have been admitted.

And later:

“If this is a correct analysis of the rule, then it seems to me that the WP correspondence in this case falls outside its scope. The issue raised by paragraph 17 of the statement of claim is whether the conduct of [MM] in settling the claim was reasonable mitigation of damage. That conduct consisted in the prosecution and settlement of the earlier action.

The WP correspondence forms part of that conduct and its relevance lies in the light it may throw on whether [MM] acted reasonably in concluding the ultimate settlement and not in its admissibility to establish the truth of any express or implied admissions it may contain. On the contrary, any use which the defendants may wish to make of such admissions is likely to take the form of asserting that they were not true and that it was therefore unreasonable to make them. 

I do not think that interpreting the rule in this way infringes the policy of discouraging settlements…”

His suggested principle that WP privilege does not protect anything more than pure admissions against interest has, as we shall see later, since been discredited and replaced by a wider scope to WP privilege, as set out in Unilever v Proctor & Gamble[13]But even so, this passage from Hoffman LJ raises immediate questions as it stands, even before those subsequent judicial doubts emerge.  His thesis is that relevant statements made within normally privileged exchanges are unprotected by WP privilege when based simply on public policy as between those seeking and those denying access to them, except that a statement which is purely an admission or concession made to further or explore settlement is privileged from being adduced in evidence in order to prove the fact embodied in that statement.  So on the facts of Muller, Hoffman LJ would presumably exclude comments in documents made by MM insofar as they amounted to specific concessions either as to willingness to compromise or as to heads of claim, but would admit comments for the purpose of showing that there was or was not a genuine attempt to mitigate loss and reduce the potential liability of L&M for any shortfall in restoring MM to where they should have been if L&M had not been negligent.

If that were to be right, it might be feared that little said and done within what was thought to be a privileged discussion would be free from examination.  It would also appear that the burden is on the person seeking exclusion of material.  But what does Hoffman LJ really mean?  Does he mean that the same statement made by MM in a document might be inadmissible as a way for L&M to prove that MM had admitted a weakness in their case, but admissible if it shed light on whether the settlement terms amounted to reasonable mitigation?  Unfortunately, Hoffman LJ makes no reference to actual instances of statements that might be taken one way or the other, or indeed both ways at once.  He does not refer back to MM’s list of documents for which privilege was claimed to discuss their actual nature or possible probative force.  What he does do is to comment that L&M might want to make use of admissions and assert they were not true.  But if they were admissions (whether true or false), they should not be admissible or considered as free from privilege in any event, according to his own test.  Surely he has missed his own point in that comment.   It is as if he has proposed a purposive test:  if the purpose for admitting the document is to rely on the admission, it is not admissible; but if the purpose is to show whether or not damage has been mitigated, it is admissible.

Nor does he deal with oral admissions (privileged) as opposed to oral non-admissions (which he would say were admissible).   As we shall see below, and as rightly observed by Robert Walker LJ in Unilever v Proctor & Gamble, oral exchanges which might be one or the other are much the likelier mode of communication within a mediation rather than documents.  Although Hoffman LJ hoped not to have discouraged secure settlement discussions by his judgment, he has certainly sown a degree of confusion.

The other judges purported to agree with Hoffman LJ but sought also (or instead) to explain that their agreement with Hoffman LJ’s judgment ordering disclosure was based on MM’s having waived privilege by seeking to sue L&M, in the course of which they put in issue the reasonableness of the settlement, trying perhaps to draw a parallel (though they did not say so) with waiver of legal professional privilege when a client sues his solicitor for negligence.  This is however an entirely different concept, and their approach in Muller to waiver has also been criticised by later judges, because effective waiver has to be made by both (or all) parties to the WP privilege[14].  Indeed in Muller, far from waiving WP privilege, MM positively asserted it, and S Ltd remained silent on the topic.

Muller and Rush & Tomkins are both three-party cases concerned with an interlocutory discovery application by a third party for access to a specific class of documents generated for a mediation which had settled part of what was in dispute.  Neither case was an appeal from a substantive trial.  Hence the full range of evidence, oral and documentary, had not been heard, when other admissible, relevant and non-privileged evidence to prove or disprove each parties’ assertions might have emerged.  In each case one litigant wanted a decision in advance of trial over access to evidence which was the subject of express or implied agreement as to privilege, to which access was being denied on the public policy justification.  The judges had to decide just on the narrow issue of discovery before them in each case.  The only discernible difference between the two is that all those involved in Rush & Tomkins were subsumed into the same piece of litigation as parties, whereas in Muller, solicitors who were being sued in negligence in separate proceedings wanted access to documents used in an earlier proceeding, albeit that they had themselves advised the Mullers at the pre-litigation stage of the same case.  Is there a material difference?  Arguably not.  In Rush, R&T had issued one action but against different parties under different contracts – the main contract with GLC and the sub-contract with PJC.  In Muller, MM might in theory at least have chosen to bring S Ltd and L&M into the same action[15].  Each case was really about different aspects of the same broad dispute.

A key question to pose (albeit unanswerable) is whether on his general Muller principle, Hoffman LJ would have allowed P.J Carey to have access to the R&T/GLC privileged documents. If he would have done, this must mean that he regarded Rush & Tomkins as wrongly decided, despite its being a House of Lords decision binding on the Muller court, unless distinguishable.  The Muller court was certainly not entitled to do that without making the basis for its doing so crystal clear.  While he cites Rush & Tomkins in his Muller judgment, Hoffman LJ does not spell out any disagreement with its outcome, despite the close similarity of facts, nor does he seek to distinguish them in any way.  Having quoted Lord Griffiths verbatim, exactly as set out above in this paper, he makes no comment as to whether he feels bound by it, or disagrees with it, except to suggest that Lord Griffiths was only dealing with admissions in his judgment.  But what difference was there between Carey’s wish to have R&T’s settlement with GLC explained, and L&M’s wish to have MM’s settlement with S Ltd explained? Hoffman LJ simply embarks upon his narrow principle, again in the precise words cited above.  Dare one suggest that Lord Hoffman chose to invent his own narrow rule and replace it for the broad view taken by the House of Lords in Rush, despite being technically obliged to follow or at least clearly to distinguish it in principle as well as on the facts and the application of principle to those facts.

Another puzzle to a mediator about what happened in Muller relates to what might have been in the privileged documentary evidence that would have helped L&M.  The issue to be litigated was whether MM had under-settled at the mediation sufficiently to justify recovering the shortfall from L&M.  MM asserted that they had mitigated their loss, and they produced the settlement terms to show what was agreed. Presumably they would also have produced the pleadings in the claim against S Ltd.  So both the basis of the claim against S Ltd and the basis of its settlement were available for consideration by the court in the proceedings against L&M, and, one might have thought, a proper assessment could then be made as to whether those settlement terms amounted to fair mitigation of damage.  If the full panoply of open evidence available at the trial had been heard, recourse to WP material might simply not have been necessary.  There would have been no need to manufacture a new (now discredited) principle or even indeed a possible new exception to WP privilege, and the regime established by the House of Lords in Rush & Tomkins would have been respected.

In any event, the claim in professional negligence against L&M would almost certainly be framed as being the loss of a chance of a better outcome.  This would have given the trial judge ample discretion when determining what damages to award. At trial, close scrutiny of the settlement outcome would have been undertaken, set against the nature and strength of the whole claim, and MM would have to shift the evidential burden of alleged failure to mitigate, with L&M seeking any sign of failure to do so.  The settlement terms are not res judicata, so would be open to easy challenge by L&M and judicial criticism if shown to be surprisingly generous to S Ltd.  Indeed in respect of the grant of new shares in the company which seem rapidly to have achieved greatly enhanced value, the settlement terms went way beyond what a court could have awarded in damages against S Ltd anyway.  MM would have had to bring those benefits into account when seeking damages against L&M.

The judgments in Muller simply list the documents sought to be protected, a list which gives no hint of the precise nature of the documents or their possible usefulness, or whether any or all of them might have constituted admissions anyway and thus (according to Hoffman LJ) unavailable to L&M for that very reason.

What is also striking to a mediator’s eyes, insofar as current mediation practice is concerned, is that few WP documents are actually produced for the purposes of the mediation only or during the mediation day.  Often the only ones are each party’s position statement and occasionally schedules and expert reports disclosed specifically WP, so that if settlement is not reached, these can still be updated and altered without criticism.  Maybe some notes and calculations will be roughed out or typed into a laptop.  Otherwise, parties produce mediation bundles of on-the-record documents to argue the strengths of their own case or the weaknesses of the opposing case on open material, all discoverable in any event.  Then settlement terms almost always have to be put in writing and signed by each party, as required by the agreement to mediate.  Tomlin v Standard Telephones[16] makes it quite clear that any such concluded settlement agreement is disclosable for all subsequent purposes and this was readily conceded by the parties in both Rush & Tomkins and Muller, who both disclosed their respective settlement terms.  There may be letters and e-mails exchanged later if a dispute is not settled during the mediation, but this did not arise in the cases now under consideration.

Additionally, would not the only relevant material capable of assisting L&M have amounted merely to concessions by MM against S Ltd, to the effect that in some respects their claim fell short and therefore merited a risk-discounted settlement?  If so, these would almost certainly have only been made orally within the mediation and not evidenced in writing.  And surely any such concessions would have been “admissions against interest” of the kind that Hoffman LJ accepted were protected anyway – in other words, an acceptance against S Ltd that in some respects they would be unable to bring their case home and were thus willing to accept a risk-discounted figure.  It would inevitably have to be at MM’s risk as to whether they could prove that the shortfall remained properly recoverable from L&M.  Mediators are entirely used to encountering parties who accept a risk-discounted figure to achieve a settlement, with such discounting going unexplained even to the mediator, let alone explained in detail to an opponent.  Nothing would prevent a party who then seeks to recover a shortfall from a lawyer who advised earlier (whether at the mediation, or in advising on the subsequently disputed contract) from arguing that it was reasonable to reduce a claim which might not succeed.  Why should the privileged content of the discussions be necessary to make or disprove that point, especially when (as Lord Griffiths said in Rush & Tomkins) privilege continues to apply to WP discussions even after settlement is reached?  For instance, does Hoffman LJ suggest that evidence might be adduced as to what S Ltd’s counsel said during the mediation to urge on MM’s counsel any weaknesses in MM’s case?  Or should MM’s counsel have been able to say that he was persuaded by such arguments that the claim against S Ltd was weak, and had decided instead to seek recovery from L&M? The settlement terms are a fixed point which surely can easily be measured against the claim and assessed by evidence other than the content of conversations during a mediation.

So in a number of respects, Muller v Linsley & Mortimer looks puzzlingly out of line, and questions arise, both on principle and in practical terms, as to whether it is a sound authority for what it seeks to decide.  This is debated more fully below in the context of the three recent cases which have revisited the decision.

The narrow Hoffman view has been specifically followed from time to time.  A case in point relating to a mediation was Cattley v Pollard[17], also a three-party case on interlocutory disclosure, a decision of Master Bragg.  Here Mrs Pollard sought disclosure of the document bundle relating to a mediation between executors and a law firm (effectively with that firm’s insurers, SIF) in which her husband had been a partner and had admittedly misappropriated funds.  The executors had settled claims against the law firm at a mediation at which Mrs Pollard was not present, and she sought access to the mediation bundle to check that the claimants against her had not already been compensated for what they were now claiming against her in the settlement with SIF. Again, the settlement agreement was disclosed by the executors to reveal the terms of settlement.  The mediator too attended to make representations to resist admission of this material based on the confidentiality of the process, though in effect the argument was based essentially on WP privilege and not contractual confidentiality.  SIF wrote to confirm that they supported the privilege asserted by the claimants and the mediator and did not waive it.  Master Bragg relied very specifically on Muller, without (arguably) clearly distinguishing the conflicting issues in Rush & Tomkins, Unilever v Proctor & Gamble[18] (decided six years before Cattley) and Muller.  He saw that the facts were similar to Muller and simply applied its authority to order disclosure.  This was again a disclosure summons and not a substantive trial hearing.  Again, it is not explained what any disclosed documents could do to shed light on the point at issue: that point is effectively assumed.

However, the Hoffman view formulated in Muller has also been subject to very considerable challenge since.  Doubts were expressed by other Law Lords when Lord Hoffman repeated his narrow principle in Bradford & Bingley BS v Rashid[19]He asserted that an acknowledgement of a debt was not an admission even if in WP correspondence, and thus admissible, but the other Law Lords based their opinions on different grounds in that case.

Positive dissent to his views were expressed by four members of the House of Lords in Ofulue v Bossert[20], where the broader approach explained by Robert Walker LJ (who was sitting as Lord Walker in Ofulue) in the earlier Court of Appeal decision in Unilever v Proctor & Gamble – discussed fully below – was approved and adopted.  In Ofulue, Bossert had been a tenant of Mr and Mrs Ofulue but was allowed to stay on unchallenged for well over twelve years after which Bossert claimed full title by adverse possession.  During the twelve year period his solicitors wrote a letter marked WP saying that the Olufues would “at the most . . . be entitled to six years arrears of rent”, and concluding it by saying “In these circumstances, our client would be willing to make an offer of £35,000 to your client for the purchase of the property”. This offer was promptly rejected by the Ofulues’ solicitors.

This letter clearly acknowledged the Ofulues’ title, but in doing so it was also suggested settlement terms which would normally attract privilege.

Four of the Law Lords in Ofulue decided that privilege did attach to the WP offer, which prevented the letter from being admissible as an acknowledgement (which otherwise it undoubtedly was) to stop limitation from running to prevent acquisition of title by adverse possession.   Lord Hoffman was not sitting in Ofulue, though his approach found the support of one Law Lord: in his dissenting judgment, Lord Scott said:

If Hoffmann LJ’s view as expressed in the Muller case is correct, this appeal must be allowed. 

Here is a prime example of the same sentences in a WP letter operating in a double capacity. The Ofulues wished to use the letter as an acknowledgement and Bossert argued that it was a WP offer.  In this direct clash of purposive interests, privilege won.   The majority decided (4-1) that the Muller view was indeed incorrect and rejected the Hoffman analysis, though with the usual courtesy that characterises the judiciary when disagreeing with a colleague.  Unfortunately, despite the House of Lords in Ofulue again being the superior court in precedental terms to the Court of Appeal in Muller, they declined to say that the Hoffman approach was simply wrong and should not be followed, or whether the decision was still right on its special facts, and this has led to complications in this field ever since. One might perhaps have expected Muller to have sunk as an authority long since in the light of Ofulue, but it has persistently resurfaced, as we shall see.


5. A Broad Approach to WP privilegeUnilever v Proctor & Gamble

Unilever v Proctor & Gamble[21] is now the leading authority on the extent of WP privilege.  This was a two-party patent dispute in which Unilever sought to challenge a patent asserted by Proctor & Gamble, who had warned Unilever that P&G would litigate to protect it, an oral remark made in a clearly WP meeting.  Unilever argued that the Muller principle allowed the warning to be given in evidence as it was not an admission against interest.  Certainly, under Hoffman LJ’s strict test, such a warning of an intention to enforce patent rights could not have been found to be protected as an admission.   It was an assertion of rights coupled with a declared intention to enforce them by action.  But at first instance, the late and greatly lamented Laddie J, who rejected Unilever’s claim to use this remark and found WP privilege to apply to it, somewhat scathingly said this:

There is a tendency to treat the judgments and speeches, particularly judges of great eminence, as if they were statutes, applying to them the same meticulous verbal analysis which is frequently deployed in construing statutes.  It appears to me that that is the approach which Mr Hobbs [for Unilever] has adopted in relation to the Muller case. I think that the decision should be looked at rather differently.  There is no record of the arguments advanced by counsel and the authorities cited in support.  The judgments themselves make reference to Rush & Tomkins and Cutts v Head and Hoffman LJ said that his analysis of the without prejudice doctrine was consistent with those two cases.  He did not suggest that he was seeking to distinguish or refine, let alone overturn, any previous decision of the Court of Appeal.  Yet if Mr Hobbs’ interpretation is correct, it would seem that a number of earlier decisions of the Court of Appeal will have to be reconsidered.

This was perhaps another way of saying that Muller is indeed inconsistent with Rush & Tompkins, despite being bound by it as a House of Lords decision.  Unilever appealed against the decision of Laddie J, but their appeal was dismissed.  Robert Walker LJ again rejected the narrow approach of Hoffman LJ in Muller, in favour of a wider interpretation of WP privilege.

Looking at the real world of settlement discussions, Robert Walker LJ said:

I have no doubt that busy practitioners are acting prudently in making the general working assumption that the [WP] rule, if not “sacred”…has a wide and compelling effect. This is particularly true where the WP communications in question consist not of letters or other written documents but of wide-ranging unscripted discussions during a meeting which may have lasted several hours. At a meeting of that sort, the discussions between the parties’ representatives may contain a mixture of admissions and half-admissions against a party’s interest, more or less confident assertions of a party’s case, offers, counter-offers and statements (which might be characterised as threats or as thinking aloud) about future plans and possibilities.


In his view, the modern cases show:

“that the protection of admissions against interest is the most important practical effect of the rule. But to dissect out identifiable admissions and withhold protection from the rest of WP communications (except for a special reason) would not only create huge practical difficulties but would be contrary to the underlying objective of giving protection to the parties, in the words of Lord Griffiths in the Rush & Tompkins case ‘to speak freely about all issues in the litigation both factual and legal when seeking compromise and,  for the purpose of establishing a basis of compromise, admitting certain facts.’  Parties cannot speak freely at a WP meeting if they must constantly monitor every sentence, with lawyers or patent agents sitting at their shoulders as minders.” 

This judgment firmly swings the balance back towards privileged protection of a broad class of communications as being the right general approach, though allowing for clearly identifiable exceptions which Robert Walker LJ’s judgement goes on to define.  Whether he would actually have regarded the decision to admit WP material in Muller as correct is not spelt out, although he allows for the possibility (perhaps reluctantly) that Muller constitutes an exception in its own right, as discussed below.  Robert Walker LJ’s remarks as quoted above could have ended the Hoffman doctrine at a stroke, particularly dealing as they do with the reality that most privileged exchanges are likely to be oral and not contained in documents.   But Muller has proved difficult to kill off, even in relation to Hoffman LJ’s wide principle.  And it is phrases like “(except for a special reason)” from Robert Walker LJ’s Unilever judgment, underlined above, or Lord Griffiths remark in Rush & Tomkins that “the rule is not absolute and resort may be had to the WP material for a variety of reasons when the justice of the case requires it” that still give pause.

We must now look in detail at the established exceptions to WP privilege, when privilege is lost and WP material is received in evidence at a trial or hearing.


6. The Established Exceptions to WP Privilege      

A number of exceptions to WP protection have been developed in caselaw over the years, and where one is found to apply, it may well lead judges to admit otherwise privileged material generated for or within a mediation.  Judges have usually treated mediation as merely subject to WP privilege[22].  There is an additional feature of mediation specific to its creation and nature, namely explicitly contracted confidentiality which is special to mediations, where settlement discussions are almost always founded on a formal contract to mediate.  The extent to which contracting for confidentiality might create wider protections for parties and mediator alike is discussed further in Section 10 below.

The possibility of an applicable exception to WP privilege can be raised by one party to the privilege, or even by someone outside the privilege, and if found to apply, may allow judicial access to privileged material even if the other party or parties to the privilege object.  Of course, as we have seen above, if all parties to the privilege waive it, then such normally privileged material is admissible, and the mediator has no standing to prevent any such mutually agreed waiver.

Thus oral misrepresentations, threats, or commitments upon which the other party relies may be conveyed during any settlement discussion and might lead to agreed terms.  These are capable of being set aside if untrue or unduly oppressive, or on the basis of an estoppel.  As we shall see, all of the above fall within established exceptions to WP privilege, and almost certainly will be taken by judges to apply to the admissibility of evidence as to what happened during mediations, as well as other types of settlement discussion.  This is subject only to whether the parties effectively exclude the operation of those exceptions, and also to whether mediation confidentiality as contracted in the mediation agreement might prevent its admission.[23]

The best source for these exceptions to WP privilege is Robert Walker LJ’s list in Unilever v Proctor & Gamble, together with two additional exceptions identified in the subsequent Supreme Court decision in Oceanbulk Shipping v TMT[24].  The Unilever exceptions are [with my comments put in square brackets]:

1. To ascertain whether without prejudice communications have resulted in a concluded compromise agreement, those communications are admissible: see e.g. Tomlin v. Standard Telephones and Cables[25] [This exception was utilised in Brown v Rice & Patel[26], which despite its title was a two-party case.  It should be noted that mediation agreements normally provide that no terms negotiated during a mediation will bind parties unless put in writing and signed.  It is arguable, in the light of the subsequent Supreme Court decision in MWB v Rock[27], that unless a party can produce written signed settlement terms, no settlement can  have been achieved, perhaps not even allowing a collateral contract to overrule such a term.  This could well have produced a different result in Brown v Rice if MWB had preceded it, but there is as yet no specific English authority on this point in relation to the effect of such a clause in relation to mediations[28]].

2. Evidence of the negotiations is also admissible to show that an agreement apparently concluded between the parties during the negotiations should be set aside on the ground of misrepresentation, fraud or undue influence [this might also be regarded as comparable to exception 4 below and would include duress as alleged in Farm Assist v DEFRA (No.2)[29] and Ferster v Ferster[30], the latter being found to be an instance of “unambiguous impropriety”.

3. Even if there is no concluded compromise, a clear statement which is made by one party to WP negotiations, and on which the other party is intended to act and does in fact act, may be admissible as giving rise to an estoppel: See Hodgkinson & Corby Ltd v. Wards e Mobility Services Ltd.[31]

4. Apart from any concluded contract or estoppel, one party may be allowed to give evidence of what the other said or wrote in without prejudice negotiations if the exclusion of the evidence would act as a cloak for perjury, blackmail or other “unambiguous impropriety”: see Foster v Friedland[32], and more recently Ferster v Ferster[33]This exception should be applied only in the clearest cases of abuse of a privileged occasion[34].

5. Evidence that WP negotiations took place may be admitted, for instance, on an application to strike out proceedings for want of prosecution in order to explain delay or apparent acquiescence: see Walker v Wilsher[35].

6. [This exception is set out as it appears in Robert Walker LJ’s judgment verbatim,  as it is a source of continuing controversy] :

In Muller[36] (which was a decision on discovery, not admissibility) one of the issues between the claimant and the defendants, his former solicitors, was whether the claimant had acted reasonably to mitigate his loss in his conduct and conclusion of negotiations for the compromise of proceedings brought by him against a software company and its other shareholders. Hoffmann L.J. treated that issue as one unconnected with the truth or falsity of anything stated in the negotiations, and as therefore falling outside the principle of public policy protecting without prejudice communications. The other members of the court agreed but would also have based their decision on waiver.”

The inclusion of this as an exception as worded certainly smacks of judicial courtesy over not wanting openly to disagree with another colleague, rather than a ringing endorsement of it, or confirmation of it as a clear exception. Indeed, Robert Walker LJ merely articulates it on the basis of Hoffman LJ’s narrow approach to WP protection which the effect of the Unilever decision is to challenge and undermine.  As noted above, the House of Lords in Ofulue specifically rejected Hoffman LJ’s formulation of general principle in similar terms and did nothing to endorse Muller in any modified form, even as a mere additional exception to the WP rule.  The alternative waiver basis used by the other judges was also shown to be doubtful, because even if MM had in their view waived privilege (in fact they strongly asserted it!), S Ltd had not.[37]  In the much later decision of King Security Systems v King[38], Miles J quotes the above paragraph verbatim and then characterises this exception (which he was perhaps generously prepared to treat as binding on him) as the “mitigation” exception.  But that paragraph does not in any way identify that if a mitigation issue is raised, this justifies the withdrawal of WP privilege.  On the contrary, the reasons given (i.e. the technical precedental ratio) for the decision are Hoffman’s discredited theory about admissions against interest, and the flawed suggestion by the other judges that waiver applied. Yet here and in the Court of Appeal judgment in Berkeley Square Holdings v Lancer, some credence is given to the idea of a “mitigation” exception based on Muller, as we shall see.

7. Where parties expressly vary the effect of WP protections, e.g. by making offers WP except as to costs.  See Cutts v Head[39], in which Fox LJ also said that “what meaning is given to the words ‘without prejudice’ is a matter of interpretation which is capable of variation according to use in the profession. It seems to me that, no issue of public policy being involved, it would be wrong to say that the words were given a meaning in 1889 which is immutable ever after, suggesting therefore that parties might be able to agree to vary the effect of WP in other ways”[40].

8. In matrimonial cases, a distinct privilege exists which extends to communications received in confidence with a view to matrimonial conciliation: see Re D[41].  The exception to what is effectively regarded as wider than WP privilege is that evidence would be received as an exception to that privilege if it related to past harm of threats of future harm to a child and it was in the public interest to receive it.

To these exceptions must be added the following (continuing the numbering for convenience, and adopting the additions recognised in Berkeley Square Holdings v Lancer[42], discussed further below):

9. Where otherwise privileged material is admissible as part of the factual matrix or the surrounding circumstances to aid construction of the agreement concluded by those negotiations – see Oceanbulk Shipping v TMT Asia[43].  This is an important recent slight extension of WP privilege, though it could be interpreted as a slightly wider interpretation of exception 1.  This was a two-party dispute in which OS and TMT discussed at a WP meeting (not a formal mediation) settlement of forward freight agreements over which they were in dispute.  A date for settling these liabilities was set, but the market rates moved adversely after that to the detriment of OS.  OS argued that TMT had failed to perform one of TMT’s settlement obligations in time, resulting in substantial losses to OS.  But TMT sought (again in a interlocutory disclosure application) to have evidence in a WP e-mail and exchanges at a WP meeting admitted to prove that the fact that OS had “sleeved” FFAs (i.e. hedged them against potential loss) was known to both of them and affected the proper interpretation of the agreement.  In the Supreme Court, Lord Clarke SCJ again reviewed the earlier cases, emphasising the decision in Ofulue and making no comment on Muller, other than listing it in Robert Walker LJ’s Unilever list.  He then summarised TMT’s case as follows:

…it was submitted on behalf of TMT that facts which (a) are communicated between the parties in the course of without prejudice negotiations, (b) form part of the factual matrix or surrounding circumstances and (c) would, but for the without prejudice rule, be admissible as an aid to construction of a settlement agreement which results from the negotiations should be admissible in evidence by way of exception to the rule because the agreement cannot otherwise be properly construed in accordance with the well-recognised principles of contractual interpretation and because there is no distinction in principle between this exception (“the interpretation exception”) and, for example, the rectification exception.

He accepted that submission and found for TMT, with the concurrence of five other SCJJ.  Lord Phillips SCJ summarised the outcome very succinctly as follows:

When construing a contract between two parties, evidence of facts within their common knowledge is admissible where those facts have a bearing on the meaning that should be given to the words of the contract. This is so even where the knowledge of those facts is conveyed by one party to the other in the course of negotiations that are conducted “without prejudice”. This principle applies both in the case of a contract that results from the without prejudice negotiations and in the case of any other subsequent contract concluded between the same parties.

Put like that, this exception sounds closely related to exception 1, in trying to find out whether there was an agreement at all and if so what the agreement means.

10. Where required to decide whether rectification of the settlement agreement is appropriate – also confirmed by the Supreme Court in Oceanbulk [this is arguably just an extension of exceptions 1, 2 and 9 above][44].

We now return (at last!) to the three recent decisions of EMW Law v Halborg[45], Briggs v Clay[46] and the first instance and appeal decisions in Berkeley Square Holdings v Lancer[47], in all of which the true status of Muller is discussed. Startlingly, despite the firm rejection [48] of Hoffman LJ’s general principle limiting WP privilege to admissions against interest, in each of these cases counsel sought to rely upon it explicitly as justifying the admission of WP material.


7.  Recent Developments (1) – EMW Law v Halborg

In EMW Law v Halborg[49], Halborg acted as solicitor for his parents (HH) against architects SH, and instructed EMW, another law firm, to act as his agent on a conditional fee basis, with EMW’s right to be paid base costs and a success fee dependant (surprisingly perhaps as a pre-condition) on final agreement and payment of HH’s costs first by SH.  HH settled their claim against SH by accepting a Part 36 offer of £350,000, but costs were disputed by SH’s solicitors as being excessive and a detailed assessment process started.  When EMW sued HH and Mr Halborg for failure to pay the costs due to EMW under the conditional fee agreement, they asserted their right to payment on the basis that the costs dispute between HH and SH had been settled. Halborg denied this and declined to disclose documents written WP in relation to this issue passing between himself (on behalf of HH) and SH’s lawyers.   EMW applied for specific disclosure.  The Master ordered disclosure, and Halborg appealed.

Reducing the decision to the relevant issues dealing with WP privilege[50], EMW asserted (and the Master agreed) that disclosure of relevant WP material, in this case, was justified because:

  • It fell within the WP exception as to whether an agreement had been reached between HH and SH over costs; and
  • the Muller exception applied.

In the first basis above for admitting the evidence, Newey J found that what he called the “concluded agreement exception” did apply on the facts in EMW to justify admission of the WP material, against HH’s wishes and without waiver or agreement from SH, even though agreement over costs (which HH denied) was a matter between HH and SH, with the privilege being theirs.  Newey J noted that the main authority for the concluded agreement exception was Tomlin v Standard Telephones and Cables[51], where it was the parties to the dispute themselves who were at odds over whether they had agreed a deal.  This was true also in Brown v Rice and Patel[52] Halborg’s counsel argued that to apply the concluded agreement exception when a third party (i.e. EMW Law) sought access to WP negotiations (i.e. between HH and SH) would lead to “remarkable consequences: it would mean that a stranger to without prejudice negotiations could, by asserting that there had been a settlement, obtain communications which the parties themselves were entitled to assume enjoyed the protection of privilege”.  Despite that, Newey J was prepared to find that it applied, noting that EMW had a “legitimate interest” in the privileged material in discovering whether agreement over costs had been reached and if so, what this meant for EWM.

This was in effect the ratio of the decision, but Newey J could not resist discussing the Muller exception.  Having reviewed many of the authorities discussed in this paper, including the criticism of Muller in Ofulue, he said:

On the other hand, I was not referred to any criticism of the actual decision in the Muller case. To the contrary, Lewison LJ observed in Avonwick Holdings Ltd v Webinvest Ltd[53] (at paragraph 22) that it was “hardly surprising that … the court ordered disclosure of the negotiations leading to the settlement”. Thanki, “The Law of Privilege”, 2nd ed., states (at paragraph 7.22):

“Whilst the result in the Muller case was undoubtedly correct, it is best analysed as another exception to the without prejudice rule.”

In the circumstances, I ought, as it seems to me, to proceed on the basis that Muller v Linsley & Mortimer was correctly decided on its facts.

Is he merely saying that Muller was right on its own facts, or identifying it as the basis for a new exception to WP prejudice?  In the end, Newey J says:

I have concluded that, to echo Lord Walker in Ofulue v Bossert and Lord Clarke in the Oceanbulk case, justice clearly demands that an exception to the without prejudice rule (whether that encompassing the facts of the Muller case or another, comparable, exception) should apply even aside from the question of whether an agreement has been reached with [SH].

He lists several reasons, among which are his finding that HH had in effect made assertions about the WP negotiations (i.e. that they were incomplete, and also that EMW’s costs had been challenged by HH’s opponents), albeit a negative assertion as opposed to the positive assertion of mitigation of damage made by MM in Muller. In suggesting that to find an exception here would not chill WP privilege as a principle, he said (by now accepting that Muller does indeed represent a further exception to WP privilege):

The existence of the Muller exception, moreover, means that communications otherwise protected by the without prejudice rule may become disclosable and admissible because the other party to negotiations unilaterally chooses, for reasons of his own, to put forward a case about the negotiations in litigation with a third party.

On the facts of EMW, Newey J presumably means that Halborg (presumably as “the other party”) chose to make the negotiations with SH an issue in the litigation brought by EMW.  This sounds rather like a reworking of his decision on the concluded agreement exception.   If his first decision was right, was there any need to consider the effect of Muller at all?  But once again, the use of the phrase “justice clearly demands” again raises some concerns. If this is to be used as a fallback if or whenever a judge is confronted with the need to do justice as the judge sees it, rather than being constrained by principle, this might well have wider ramifications than the case before the judge if breached or extended.

As a parting shot on EMW, it would be sad if this decision, with its particular facts, made an unwelcome impact on the status of WP prejudice at the behest of those engaged in this case.  Mr Halborg had hired EMW as agency solicitors because he felt insufficiently competent to conduct his parents’ litigation without their help and should have been prepared to pay their proper fees, whether or not they were fully recoverable from SH.   Extracting £350,000 from SH was on any view a “win” for the purposes of a CFA. But EMW Law seem to have entered into an extraordinarily badly draft conditional fee agreement which surrendered control over their remuneration and its timing to Mr Halborg.  Neither party covered themselves with glory – an instance perhaps of a bad case making hard law.


8.  Recent Developments (2) – Briggs v Clay

Briggs v Clay[54] (simplifying the facts considerably) concerned the failure of a pension scheme designed by Aon for a number of participating firms.  The scheme trustees had brought earlier proceedings against participating firms to ascertain the validity of the pension deeds.  It was expected that any shortfall would be visited later upon Aon if the deeds were invalidated as a result of their negligent advice.  The deeds were found to be invalid, so fresh proceedings were brought against Aon for consequential losses.  An appeal was later lodged against the first adverse decision on the status of the deeds, which was compromised in WP discussions between the trustee claimants and the participating firms, though the claim against Aon continued.  Up to this time, the claimant trustees had been advised and represented by solicitors (GG) and leading counsel (QC), in both the separate WP negotiations between the claimants and the participating firms, and also between the claimants and Aon.  Aon then amended their defence to allege that GG and QC themselves had negligently failed to advise the trustees of an argument which would have reduced the losses incurred and joined GG and QC as defendants in a counterclaim.  As a result the claimant trustees also added GG and QC as additional defendants by the claimants.  As in Muller, the parties were seeking to recover some of their perceived shortfall from the lawyers, though this time the lawyers had been brought into the core litigation.

GG and QC then sought to have the content of both the WP discussions which led to settlement of the appeal between the trustees and the participating companies and also the trustees and Aon admitted in evidence.  They wanted to show that Aon and Aon’s lawyers had been kept fully informed of developments in the course of those discussions and so had been in a good position to raise the point which now Aon and the trustees both alleged had been negligently overlooked by GG and QC.  They made it clear that they were not relying on the WP discussions as admissions or asserting their truth or falsity but simply as evidence of Aon’s knowledge of the position during the negotiations.   They were thus relying heavily on Muller in its unregenerated form, despite all the criticism of Hoffman’s narrow interpretation raised in Ofulue and other decisions.

Fancourt J reviewed all the authorities referred to in this article and several more, including EMW v Halborg[55].  The key paragraphs in Fancourt J’s judgment in Briggs read as follows:

49. There therefore appear to be two bases for Hoffmann LJ’s decision. The first is that the without prejudice rule only applies to protect admissions, not facts that are relevant independently of their truth or falsity, and the defendant was not seeking to rely on the content of the without prejudice negotiations to prove any admissions. The second basis is that the plaintiff himself had raised (or “put in issue”) the reasonableness of the negotiations; that issue could not be determined without disclosure of the negotiations, and the public policy underlying the rule was not infringed by ordering disclosure in favour of the defendant for the purpose of the second claim. The shareholders were not parties to the second claim or (apparently) affected by its outcome.

50. Swinton Thomas LJ, while agreeing with both Hoffmann LJ and Leggatt LJ, preferred to rest his decision on the basis that by putting his own conduct in issue in the second claim, the plaintiff had waived any privilege attaching to without prejudice negotiations. Leggatt LJ accepted Hoffmann LJ’s “thesis” that the without prejudice correspondence fell outside the scope of privilege, but also held that in any event the plaintiff could not both assert the reasonableness of the settlement and claim privilege for the negotiations; he further held that by disclosing the letter before action and the settlement agreement the plaintiff had waived any privilege in all the other documents relating to settlement.

51. The decision in the Muller case has given rise to considerable comment in later cases and still generates controversy today. In so far as the judgment of Hoffmann LJ rests on the first basis identified above, it has been substantially discredited in later cases, both as to the rule only applying to protect admissions and as to there being any general exception to the rule where negotiations are relied upon to establish a “collateral fact”. The true ratio of the second basis for the decision is controversial. A narrow interpretation would be to say that it is a decision based on its own facts, or limited to a case where the issue is the reasonableness of a negotiated settlement, but no judge subsequently considering the case has said so.

52. There has also been judicial disagreement with the basis for the decision preferred by Swinton Thomas and Leggatt LJJ. The privilege conferred by the without prejudice rule cannot be waived unilaterally by one party only to the negotiations, in the way that the sole owner of legal professional privilege can waive the privilege. There was no suggestion in Muller that the shareholders had expressly or impliedly agreed to give up their privilege. Accordingly, waiver, in its true sense – voluntarily giving up privilege that exists and is protected by the without prejudice rule – could not have arisen: see per Lewison LJ in Avonwick v Webinvest Ltd [2014] EWCA Civ 1436 at [21] and per Newey J in EMW Law LLP v Halborg [2017] EWHC 1014 (Ch); [2017] 3 Costs LO 281 at [62]. It is however clear that both Swinton Thomas and Leggatt LJJ considered it material to their decision that the plaintiff had raised an issue on which the court could not adjudicate unless the negotiations were disclosed.

53. Despite the criticism of the building blocks of the three judgments in Muller, it is generally accepted to have been rightly decided, as an exception to the without prejudice rule. The difficulty lies in deciding the true ratio of the decision and the extent of the exception thereby established. 

So this approach to Muller confirms its demotion (if that is the right word) from a general concept –   limiting WP privilege generally to protecting admissions against interest from disclosure, to (at its highest) a mere exception to the WP rule.  Fancourt J rejected the proposition that Aon had waived their right to rely on WP privilege and decided that the only basis on which WP material on which the lawyers sought to rely could be admitted was if they could bring themselves within an exception based on a proper interpretation of Muller.  Fancourt J summarised the essence of the Muller exception thus:

….the general principle that bringing a claim or making an allegation does not disentitle a party to rely on without prejudice privilege may well be qualified where an issue is raised that is only justiciable upon proof of without prejudice negotiations. Indeed, in cases where the Muller exception has been applied, the judges have emphasised that the claim would otherwise be non-justiciable. A claimant (or defendant) cannot at one and the same time raise an issue to be tried and rely on without prejudice privilege to prevent the court from seeing the evidence that is needed to decide it.

So, it is said, this exception arises when a party to WP communications who raises an issue (such as claiming to have properly mitigated their loss in settling, as in Muller, or in EMW v Halborg[56], where Halborg denied that settlement had been reached) yet seeks to shelter behind that privilege in relation to a matter which is objectively not justiciable unless the privilege is removed.  Oddly, though, the “only justiciable” point is not made by Hoffman LJ at all, but by Swinton Thomas and Leggatt LJJ, whose judgments are both somewhat tainted by their wrongly finding an effective waiver by the Mullers.  Nor did any of the Muller judges explain why it is only by disclosure of WP material that the reasonableness of the settlement between the Mullers and S Ltd could be assessed.  Swinton Thomas and Leggatt LJJ simply accepted the assertion of L&M’s counsel that the reasonableness of the settlement was not justiciable without the WP material, without questioning precisely what it was and how it might cast light on the issue of mitigation.

Hoffman LJ circumvented Rush & Tomkins by saying that Lord Griffiths’ views were limited to the exclusion of admissions.  But four years on from Muller, Robert Walker LJ in Unilever described the kind of exchanges that might be protected by WP privilege as consisting “not of letters or other written documents but of wide-ranging unscripted discussions during a meeting which may have lasted several hours – at a meeting of that sort the discussions between the parties’ representatives may contain a mixture of admissions and half-admissions against a party’s interest, more or less confident assertions of a party’s case, offers, counter-offers and statements (which might be characterised as threats or as thinking aloud) about future plans and possibilities”.   Might this description not have covered much of what was exchanged WP during the mediation between the Mullers and S Ltd?  After Unilever, such material would have to be excluded when considering the basis on which the Mullers compromised their claim against S Ltd.  So, as suggested earlier, was it not possible for the eventual trial judge in Muller to look at the terms of settlement as disclosed and to hear arguments as to whether those terms, objectively considered, amounted to fair mitigation of loss?  Or was that issue truly “only justiciable” if the judge heard details of the negotiations which led to the settlement terms?  That seems rather far-fetched, especially if the WP material to which Rush & Tomkins gives protection is interpreted as re-defined in Unilever.

At any rate, Fancourt J bravely sought to make it clear that the Muller exception as redefined is not merely a matter of judges being able to mobilise “the interests of justice” in order to excuse the trumping of WP privilege to enable them more easily to decide the case in front of them:

It is clear, on authority, that there is no exception to the without prejudice rule merely because justice can be argued to require one on the facts of a particular case. In EMW Law, Newey J did not conclude that disclosure should be given because justice required it: he concluded that it was just to regard an established exception to the without prejudice rule, whether the Muller exception or a comparable one, as applying on the facts of that case. 

This is perhaps a slightly charitable interpretation of Newey J’s explanation for his decision, as quoted above, and ignores the fact that the first basis for his decision was that the concluded agreement exception applied.  It also slightly smacks of special pleading to justify his own decision in Briggs.

Furthermore, and equally bravely, having reviewed the facts and likely future course of the complex Briggs litigation, Fancourt J decided (unlike the Court of Appeal in Muller or the judges in Cattley v Pollard and EMW Law v Halborg) that the issues were justiciable without access to the WP material and declined to admit it.  In doing so, he acknowledged that he was denying the trial judge access to material that might have been relevant or probative (as probably any WP material might be in any hearing), yet he did so because it had been created WP between people engaged in the same overall dispute.  They – especially the claimants’ former lawyers, GG and QC – must be taken to have known that WP privilege might prevent it from being admitted, and “the interests of justice” alone did not justify ignoring the privilege that belongs to both parties who contracted to it or who operated under the umbrella of public policy.  Of course the trustees will have waived legal professional privilege of material passing between themselves and GG and QC as the trustees’ lawyers, but they were not acting for Aon during the WP negotiations, so waiver of WP privilege will have required the consent of both GG and QC and Aon, and Aon firmly resisted disclosure.

In excluding access to the WP material, Fancourt J at least undertook the task of seeing whether the issues in the case could be determined without it, something which was not undertaken in Muller.   Judges are even able to perform the mental juggling act of looking at privileged material and then deciding to ignore it because it is privileged[57].   But an inevitable continuing worry, despite Fancourt J’s cautionary note, is whether it is still too easy for a judge to use his interpretation of Muller to decide that an issue is “only justiciable” if WP material is admitted, perhaps under the cloak of “the interests of justice”. We shall return to this.


9.  Recent Developments (3) – The Trial of Berkeley Square Holdings v Lancer

Recent review of the true significance of Muller has not ended there.  It returned to the fore in Berkeley Square Holdings v Lancer Property[58],  in which the court (uncomfortably for mediators at first sight) admitted the contents of undoubtedly privileged mediation position papers, finding that exceptions to WP privilege applied.  Put simply, the claimants BSH owned a large and valuable property portfolio in London and in 2005 their representative AA contracted by deed with the defendants LP to manage it.   A later side letter authorised further fees to LP and allowed payments to be made to BB, another company in fact controlled by AA.  A further deed of variation was entered into by BSH with LP in 2011, again through the agency of AA, confirming AA’s authority and ratifying payments by LP to BB.  In 2012, LP sued BSH for underpayment of fees under the original agreement, and this dispute went to a CEDR mediation.  In their two position papers for that mediation, LP stated as facts to BSH that BB was a company controlled by BSH’s manager, AA.  They also referred to the side letter and said that £27 million had been paid to BB by LP under the existing arrangements.  This dispute was settled at the mediation, leading to the signing of new deeds relating to the deal between BSH and LP. In 2015, AA was replaced as BSH’s representative, and LP’s appointment was terminated by BSH’s new manager in 2017.  In 2018 BSH sued LP for colluding in fraud allegedly perpetrated by AA on BSH, claiming that both the side letter and the 2012 deeds which were executed as a result of the mediated settlement, plus payments made thereunder, were all consequently void.  LP riposted that BSH and their solicitors had been appraised of the relevant facts as to AA’s control of BB, and as to payments made to BB by the facts asserted in LP’s (admittedly privileged) position papers for the 2012 mediation.  LP sought to have them admitted in evidence to demonstrate these facts.  BSH objected on the grounds that they were subject to WP privilege, as they were an “owner” of (as a party to) the expressly contracted privilege set up by the mediation agreement.

So in effect LP were seeking to uphold the validity of the 2012 deeds that emerged from the mediated settlement, while BSH were seeking to have them declared invalid as a result of alleged fraud, although the fraud was said to have preceded the mediation and not been directly committed during it. They thus sought to overturn the settlement documentation. LP argued that the misrepresentation/ fraud exception (Unilever exception 2 above) applied, though in an obverse way.  They argued that since a false representation by them within the WP mediation (presumably including in their position papers) could have led to the 2012 deeds being invalidated, it would be contrary to justice if they could not rely on true representations made within the WP privilege covering the mediation and which if accepted would render the deeds enforceable.  Roth J agreed this reversal of the misrepresentation exception:

In my judgment, the statements here are admissible either under this exception, properly interpreted, or by reason of a small and principled extension of it to serve the interests of justice. If Lancer had misled the Claimants by misrepresentation in the mediation, then the Claimants could rely on that in challenging the 2012 Deeds. It seems to me contrary to principle to hold that where Lancer was truthful in the mediation, their statement cannot be admitted to rebut a case that the Claimants were deceived by Lancer as to the true state of affairs. 

He drew support from the exceptions to WP privilege relating to rectification and Oceanbulk Shipping.  And on this point he concluded:

In the present case, the mediation papers are being looked at to determine what were the facts of which both parties were aware, on a dispute as to whether the contracts they concluded were made in ignorance by one party of certain key facts. Furthermore, there is no conflict here with the fundamental principle that parties should be encouraged to speak freely in negotiations, without concern that what they say may be used against them in litigation. The Defendants are seeking to adduce evidence of what was said by the First Defendant, not of anything said by the Claimants. 

The facts are really very similar to Oceanbulk, where again it was the admission of information to explain the understanding on which the deal was based that was sought.

However, having in effect disposed of the dispute in LP’s favour, Roth J still went on to consider the other exceptions on which LP had relied.  Strictly what he says on these further points is obiter, therefore.  But in that his further thinking seeks in any way to legitimise Muller as a proper exception to WP privilege, at the same time using it to justify admission of the content of privileged mediation position papers, it may be viewed as regrettable.

On LP’s allegation that BSH had set up an estoppel by silence in apparently ignoring the information given to them by LP in their mediation position papers, Roth J declined to accept this as a valid reason for admitting the contents of the position papers.  BSH had not made a representation during privileged discussions on which LP had relied – indeed most of the alleged silence post-dated the mediation, and he would not have permitted evidence of what was and was not said during the mediation to be admitted in evidence,

since parties seeking to compromise a dispute would then have to take care to controvert in the negotiations any statements made by the other side, which is not an approach conducive to open and constructive discussion. 

This view will comfort mediators.  However, before concluding, Roth J returned to the familiar territory of the Muller exception, as this had also been argued by LP.  Commenting that Muller is “problematic” and accepting that, although rejected as the source of a general principle that WP prejudice is strictly limited to “admissions against interest”, he noted that the decision nevertheless has been accepted as rightly decided, and one which can be treated as establishing a specific WP exception, if properly defined.  Roth J added to the efforts of Fancourt J in Briggs v Clay in trying to ascertain the true definition of the Muller exception.  He did so as follows [with my underlining]:

I respectfully agree with Fancourt J’s analysis of the Muller exception, which I gratefully adopt. The question then arises what is meant by “fairly justiciable[59].” This of course does not mean justiciable in the sense applied to an act of State or a claim to title over foreign land. In my judgment, it means that the evidence is so central to an issue which the party resisting disclosure has introduced that there is a serious risk that there will not be a fair trial if that evidence is excluded. Hence in Muller, the issue was whether the Mullers had acted in reasonable mitigation of loss by settling the proceedings in the amount that they did. Plainly, that issue could be determined without seeing the content of the WP negotiations, since the court would see the letter before action, the pleadings and the terms of the settlement. But to reach a fair decision, the court would need to see the WP negotiations which led to the settlement. This is the point made in the short judgment of Swinton Thomas LJ who, although justifying the outcome in terms of waiver, said:

“It is the plaintiffs who have brought the reasonableness of their conduct in issue…. [T]hat allegation made by the plaintiffs would in reality not be justiciable without the court having sight of the without prejudice negotiations and correspondence.” 

The same applies, it seems to me, to EMW’s allegation in EMW Law that Mr Halborg had failed to make reasonable efforts to secure agreement by Savage Hayward to cover its fees. 

The underlined passage goes unexplained.  Why did the court have to see “the WP negotiations” to reach a fair decision, if the issue could be determined without access to them?  And in any event, Muller was about documents, not the content of oral WP negotiations.

[Counsel for BSH] submitted that the Muller exception applies only in a three-party case, where the other party to the WP negotiations is not a party to the action and so cannot waive the privilege. I do not agree. Although that was the case on the facts in Muller and EMW Law, it was not the basis on which Fancourt J rejected application of the exception in Briggs v Clay and I see no logical basis for such a limitation on the exception if its justification is as set out above. The justification may apply as much in a two-party case, where the other party to the negotiations is a party to the action but, in its own litigation interest, refuses to agree to a waiver. 

Roth J went on to apply this principle to the facts in Berkeley Square Holdings.  He commented:

A fundamental issue in the trial of the claim will be whether the Defendants, as the Claimants assert, acted dishonestly; and therefore whether the Claimants were indeed unaware of these key facts before May 2017, and more particularly before entering into the 2012 Deed of Settlement and Deed of Variation. 

He continued:

Since the Claimants rely strongly on their lack of knowledge, I consider that this is an issue, and indeed a potentially critical issue, raised by the way the Claimants have advanced their case. In my judgment, this issue is not fairly justiciable if the Defendants cannot put in evidence of what the First Defendant (Lancer) told the Claimants in its mediation statements in September 2012. Put another way, I do not see that the Claimants can fairly advance a case based on their ignorance until May 2017 of certain key facts while excluding evidence that they were told those facts some five years earlier. 

He pointed out that no reliance was being placed upon an admission by BSH, merely on what LP themselves had tendered as background information as true, and were seeking to rely on what they themselves, not BSH, had said.  In that sense it was the reverse of an admission against interest; rather it was an assertion supporting their interests.  Surely in that sense it is very close in nature to the assertion made by Proctor & Gamble in Unilever.  Roth J does not seek to reconcile that decision with Muller. And again, could this have been proved by other means at a trial, by open documents or oral evidence?  Once more the judge was being asked to decide admissibility as a preliminary point.  To have tried to use Muller again obiter to explain what happened in Berkeley Square Holdings at an interlocutory stage is perhaps unfortunate, especially as it raises again the “interests of justice/only justiciable” approach to removing WP protection, this time to admit evidence from a mediation position paper.  The first ground on which Roth J based his decision sounded far more comfortable to the ears of mediators than his third.


10.  Recent Developments (4) – The Appeal in Berkeley Square Holdings

To some extent, the Court of Appeal in Berkeley Square has produced a degree of relief over such concerns.  BSH appealed the admission of the otherwise privileged contents of the mediation position papers on both Unilever ground 2 (the misrepresentation/ true representation exception) and ground 6 (the Muller exception); while LP cross-appealed Roth J’s refusal to allow admission by estoppel (Unilever ground 3) and also sought to canvass a previously unargued exception in the 1794 decision of Waldridge v Kennison, where a WP communication was admitted as an “independent fact” to prove the handwriting of the writer.  The cross-appeals were both dismissed, the more important estoppel argument being rejected for the same reasons as deployed by Roth J at first instance.  But full consideration was given in the judgment of David Richards LJ to the BSH appeals on grounds 2 and 6.  Put shortly, the Court admitted the content of the mediation position papers based on ground 2 but rejected BSH’s reliance on the Muller exception, in terms which criticised the analysis of Muller as undertaken by both Fancourt J in Briggs and Roth J in Berkeley Square, insofar as their decisions rested on being admitted in evidence if justiciability was otherwise impossible.  He appeared to recognise that Muller (in particular based on the judgments of Swinton Thomas and Leggatt LJJ, but not Hoffman LJ’s) did establish an exception to WP privilege where one party clearly puts in issue the content of WP discussions (or a mediation) as the basis for a claim or defence in separate proceedings (i.e. in a three party situation).  He doubted whether it could be used in two party cases, where privilege and waiver thereof is a matter directly between the parties to the litigation and the WP discussions which were associated with it.  He agreed with Miles J’s view, expressed in the intervening case of King Security Systems v King,[60] that Fancourt and Roth JJ erred to that extent in delineating a “justiciability” exception in Briggs and at first instance in Berkeley Square.  These decisions need fuller examination, especially as the net effect remains that a normally privileged mediation position paper was found to be admissible as evidence, on the face of it still a theoretical worry for the mediation world.


11.  Confidentiality as Contracted in Mediation Agreements: Does this Add to WP Privilege? 

The whole of the above debate has concentrated on WP privilege and its exceptions.  But mediations are also formally convened under the umbrella of contractual confidentiality, with the parties and the mediator (and any mediation provider) each signing a written agreement to mediate which provides inter alia that they will all treat the content of mediation discussions and documents produced for the mediation as confidential.  It also usually requires that for settlement terms to be binding, they must be in writing signed by the disputing parties. This formality distinguishes mediation from most WP settlement discussions or correspondence. To contract confidentiality is to create positively enforceable rights for signatories to it, whereas WP privilege is by contrast “merely” an exclusionary rule of evidence which regulates whether certain material can be laid before a judge. Breach of contracted confidentiality confers a cause of action on an affected party, entitling them to an injunction preventing disclosure or seek damages for breach.  Judges have however a discretion to require that they hear confidential material “only if disclosure is necessary for the attainment of justice in the particular case”[61] .

Two cases have considered what that contracted confidentiality might mean.  The first is Cumbria Waste Management (1) and Lakeland Waste Management (2) v Baines Wilson[62], a decision of HHJ Kirkham sitting in the High Court.  On the face of it the facts were very similar to Muller.  CWM and LWM were in dispute with DEFRA over payments for foot-and-mouth carcass disposal contracts but settled each claim at separate mediations.  They then sued their former solicitors BW hoping to recover the shortfall allegedly arising from BW’s poorly drafted contracts.  BW sought disclosure of documents relating to the two mediations over whether CWM and LWM had under-settled.  According to HHJ Kirkham, BW argued that:

If the claimants settled with DEFRA on the basis that there was a risk that the unmeritorious construction advanced by DEFRA would be upheld by the court, then that was an unreasonable basis for the claimants to settle.  Further, if the claimants settled with DEFRA on the basis of concerns (whether legal or commercial) other than the construction of the contract, then the defendant cannot be held responsible for any shortfall between the settlement monies and the amounts invoiced by the claimants.  

HHJ Kirkham declined to follow Muller, despite its close factual similarity, pointing out that WP privilege was being asserted there by DEFRA as an owner of the privilege and with their own legitimate interest to protect (namely fear of impact on other similar disputes they were defending), whereas in Muller it was the claimants who sought to enforce the privilege when seeking to recover their alleged shortfall[63].  So, as in Rush v Tomkins v GLC and Unilever v Proctor & Gamble, someone external to a mediated settlement was refused access to privileged material as a means of proving their case.  Each mediation agreement signed by CWM and LWM with DEFRA expressly contracted for WP privilege and also for confidentiality of what passed at the mediation.  By odd coincidence, the mediator in CWM had also been the mediator in Cattley v Pollard a year or two earlier: perhaps stung by the Cattley outcome, he took a neutral stance this time, leaving it to the parties to argue the issues.  The mediator in LWM was rather more robust in asserting the right to confidentiality and, as we shall see below, she took the same robust line in the next case to be considered, when as mediator she intervened in Farm Assist v DEFRA (No. 2).  Although HHJ Kirkham’s decision that WP privilege applied at DEFRA’s behest was enough to dispose of the application, she went on to consider (again arguably obiter) the effect of the contract for confidentiality set up by each mediation agreement, and said:

Had I not concluded that the defendant’s application failed for the reasons given above -that is, as not falling within one of the exceptions to the without prejudice rule – I should have concluded that DEFRA would be entitled to rely on an exception to the general rule that confidentiality is not a bar to disclosure.  DEFRA was a party to the confidentiality agreement and wishes its provisions to be honoured.  In any event, I am persuaded that, for the reasons identified in 17-016 above, documents within a mediation should be protected from disclosure. 

For the first time[64], a judge was clearly, if tersely, indicating the confidentiality contracted in a mediation agreement might render evidence of what happened at a mediation unavailable to a judge, as well as the rest of the world.  But before discussing this decision, Ramsey J’s decision in Farm Assist v DEFRA (No. 2)[65] needs to be considered.  Here DEFRA startlingly reversed tack and not only sought full disclosure of what happened at a mediation with Farm Assist (presumably thinking that it was in their interests to do so) but even wanted the mediator to be called to give evidence of all oral exchanges during the mediation, both jointly between the parties and in also private meetings with each of them.   Ramsey J found that the parties had waived WP privilege, which the mediator could not influence.  This meant that the only remaining basis for exclusion of evidence was the confidentiality agreement, over which the mediator intervened to assert that mediators have a say over its enforcement as well as the parties.   Ramsey J’s general view was as follows: 

I consider that, in the context of mediation and in the absence of an express provision, a similar implied confidentiality would arise, but that evidence may be given of those matters if the court considers that it is in the interests of justice to do so. In this case DEFRA and FAL have agreed with the Mediator to treat the mediation as confidential. That, in my judgment, is an obligation which is binding as between the parties and the Mediator but that the court can permit the use of or order disclosure of the otherwise confidential material if it is in the interests of justice to do so. Whilst it is possible for the confidentiality to be waived, that hasto be with the consent of all parties. This means that, in my judgment, FAL and DEFRA cannot waive confidentiality in the mediation so as to deprive the Mediator of her right to have the confidentiality of the mediation preserved. 

Ramsey J considered Cumbria Waste and the views expressed by HHJ Kirkham without disagreeing with them.  However, on the facts, he came to the unfortunate conclusion that:

whilst the mediator has a right to rely on the confidentiality provision in the Mediation Agreement, I consider that this is a case where, as an exception, the interests of justice lie strongly in favour of evidence being given of what was said and done. 

Farm Assist is primarily about whether a mediator should be called to give evidence of what passed at the mediation, orally and even in private meetings, a topic which the judge did not address specifically in balancing out “the interests of justice”.  After all, the parties had waived WP privilege, sensibly or not.  Luckily the proceedings were forthwith discontinued, and the mediator never had to give evidence.  The extent to which mediators are compellable as witnesses remains a challenging topic, but one to be debated elsewhere than this already lengthy paper[66].  However, Ramsey J’s decision apparently to admit widespread evidence of what was said throughout the mediation in the “interests of justice” despite opposition from the mediator based on her acknowledged right to enforce its confidentiality certainly smacks of a wish to admit anything that makes it easier for a judge to reach a conclusion, to the exclusion of any other kind of interest.  The security of mediation (or other settlement discussions) or the neutrality of the mediator, seem not to have come into consideration, and this creates cause for concern.  It might be characterised as an insistence on “best evidence”, yet surely there are other factors that ought to be weighed in such a balancing exercise, as will be discussed below[67].

For completeness, it is important to note that the mediator additionally contracts to treat what is said in private meetings with each party separately as confidential and will only convey such material from one party to the other if specifically authorised to do so. This entirely separate and deeper layer of contractual confidentiality is extremely important in generating party trust in the mediator. It is one of the features of the mediation process that enables it to work so well.   Any breach of this duty by the mediator would seriously damage that trust and indeed each party’s perception of the mediator’s neutrality.  It would probably be negligent conduct by the mediator to disclose unauthorised material, as well as in breach of that contracted obligation.  It is all the more distressing that Ramsey J appeared to contemplate requiring a mediator to break that obligation in Farm Assist (No 2), in effect ordering a mediator to break a contractual obligation.

The main significance of Cumbria Waste and Farm Assist in the context of this paper is that they represent two cases in which contractual confidentiality has been identified as apparently adding further protection to the privacy of disclosure of mediation material.  Additionally the mediator (and maybe also any mediation provider which also signs the mediation agreement) is identified as having locus standi to enforce it on the other signatories.  In theory at least, it would appear from these two cases (albeit at first instance) that contracted confidentiality might of itself keep confidential material from a trial judge.  The problem with the “interests of justice” proviso utilised by Ramsey J is that this is still subject to a degree of judicial discretion if the judge perceives a need to admit confidential material.


12. The Continuing Debate

Where does this lengthy debate leave us in relation to the security of mediation evidence against the modern law of WP privilege?  There remain tantalising questions.  Are Muller and Rush & Tomkins inconsistent with each other?  Would Cattley v Pollard have been decided the same way if it had followed Ofulue?  Would Muller have been decided the same way if Ofulue or Oceanbulk had preceded it?  Do EWM v Halborg, Briggs v Clay and Berkeley Square Holdings v Lancer make more sense of the debate?  Would MWB v Rock have altered the judge’s approach in Brown v Rice & Patel? What difference might contractual confidentiality make to the issue, ignored in most of the decisions discussing these issues.  Such are the hit-and-miss delights of the common law, in which the courts cannot choose the order in which they are asked to answer questions and define developments in thinking.

A few points seem clearly established now.  In relation to WP privilege generally, the wider Muller doctrine that would limit protection only to strict admissions against interest has been discredited and is simply not right.  This makes it particularly surprising is that counsel in EWM v Halborg, Briggs v Clay and Berkeley Square Holdings v Lancer all tried to resurrect the general Muller approach, despite its having been firmly discredited by the Supreme Court.  It must surely be right that the broad Unilever approach to what is privileged material is taken by English judges who should regard WP protection as paramount, and only derogate from that if one of a few defined exceptions is clearly made out. Furthermore, judges should be slow to identify any further exceptions.  Roth J’s judgment in Berkeley Square Holdings quotes Teare J in in Single Buoy Moorings Inc v Aspen Insurance Ltd:[68] 

In my judgment an exception can only be allowed where it is of the same character as one already established or where it is an incremental but principled extension of an existing exception, as was the exception in Oceanbulk v TMT.

The Oceanbulk exception, and its related minor extension to include rectification, are indeed the only new ones which have been accredited in the 21st century so far since Unilever was decided in 2000.  But even so it seems that there remains the need to look once again at the Muller exception in the light of the attempts to rework it in EMW Law v Halborg, Briggs v Clay and Berkeley Square Holdings v Lancer.   Does what happened in Muller require its continued existence as a separate exception to WP privilege?  Or on closer analysis does it properly belong within another class of exception, so as to make greater sense of WP privilege as the general rule, without needing to widen the scope of the exceptions to privilege?

It is obviously desirable to establish with a high degree of certainly (if possible) the parameters of the exceptions to the WP rule, so that parties engaging in WP settlement discussions of disputes, including mediations, can do so with as clear an understanding as possible of what, despite normal expectations as to the application of WP privilege, might yet be revealed to a judge because of an exception.  I would suggest that simply giving judges a discretion on such matters based on their view of “the interests of justice” or that “this case is only justiciable if I admit this WP material” would be dangerous and subversive to the value of WP privilege.   So far as possible, exceptions need to be clear and defined, though always subject to slight adjustment or explanation to suit new facts, a process very familiar to common law in England & Wales.


13.  Further Analysis of the so-called “Muller exception”

We start by listing the decisions which supported WP privilege when challenged by one party against the wishes of another, and those which did not[69]:

Group A: Cases where WP privilege prevented disclosure

Rush & Tomkins v GLC and Carey House of Lords 1988

Instance v Denny Bros Commercial Court 2000

Unilever v Proctor & Gamble Court of Appeal 2000

Cumbria Waste Management v DEFRA QBD 2008

Ofulue v Bossert   Supreme Court 2009

Briggs v Clay   Ch D 2019 

Group B: Cases where WP privilege did not prevent disclosure

Muller v Linsley & Mortimer   Court of Appeal 1995

Cattley v Pollard   Ch D Master 2006

Oceanbulk v TMT   Supreme Court 2010

EMW Law v Halborg   Ch D 2017

Berkeley Square Holdings v Lancer  Ch D 2020, as varied by the CA in 2021

One explanation which reconciles the last three cases in Group B under a common principle is to suggest that they are all part of, or minor extensions to, Unilever exception 1 – what Newey J called the “concluded agreement” exception in EMW Law v Halborg.   EMW and Berkeley Square Holdings are both really examples of this, as slightly broadened by Oceanbulk.  Indeed, the first ground given in each of them for admitting the WP material was specifically related to the “concluded agreement”/Oceanbulk exception.

As noted above, the House of Lords in Oceanbulk effectively ignored Muller in reaching its conclusion that evidence to explain the factual matrix at the time of negotiation is admissible to explain what it means, whether in open or WP communications. Ironically, in his lead judgment in Oceanbulk, Lord Clarke did adopt dicta from Lord Hoffman in support of his approach, but drawn from another Hoffman judgment – in Chartbrook v Persimmon Homes[70] and not from Muller.   Berkeley Square is best explained by the first of the three reasons given by Roth J for admitting the facts in the mediation position papers, associated closely with Oceanbulk.  The facts asserted were part of the factual nexus needed to explain why BSH agreed to the new deeds, which they were now trying to have set aside as void or voidable for fraud.  EMW is more closely related to Muller and Cattley, all being three party cases[71].  In EMW there was a need to adjudge whether – and if so, what – agreement had been reached within WP negotiations or at a mediation, at the understandable behest of an outsider possessing what Newey J called a “legitimate interest” in the true position.

Could the remaining Group B cases – Muller and Cattley – be explained similarly as being covered by the concluded agreement/Oceanbulk exception, rather than necessitating the existence of a rather dubiously defined Muller exception in its own right?  Linsley & Mortimer arguably had a “legitimate interest” in the terms of the mediated agreement because of the mitigation of damage issue, and Mrs Pollard similarly had a legitimate interest in whether there had been double counting by Cattley.  The same is true in EMW Law.  So perhaps all that is needed to understand these three cases is that access can be afforded under the “concluded agreement-rectification-factual nexus” exception which Oceanbulk effectively defined, but which is available not just to the parties to WP privilege, but also to outsiders with a proper interest in the issue.  It should not be difficult for judges to identify whether someone has a legitimate interest sufficient to claim access.

The only problem remains in reconciling these cases with Rush & Tomkins v GLC, a decision made by and with the authority of the House of Lords in 1988 on quite similar facts.  WP privilege was upheld by the House of Lords to prevent a party from looking behind the veil of privileged negotiations in which material relevant to that party’s case had been discussed.  As a reminder, the ratio in Lord Griffiths’ speech was this:

I would therefore hold that as a general rule the “Without Prejudice” rule renders inadmissible in any subsequent litigation connected with the same subject matter proof of any admissions made in a genuine attempt to reach a settlement. It of course goes without saying that admissions made to reach settlement with a different party within the same litigation are also inadmissible whether or not settlement was reached with that party.

It is certainly hard to see how the decision in Muller can stand against this proposition as a separate exception, or how the facts in Muller are sufficiently distinguishable, especially when we remember that the meaning of “admissions” as used in Rush & Tomkins has to be expanded to encompass Robert Walker LJ’s wider meaning which he gave in Unilever – to include letters or other written documents….. wide-ranging unscripted discussions during a meeting which may have lasted several hours…..the discussions between the parties’ representatives may contain a mixture of admissions and half-admissions against a party’s interest, more or less confident assertions of a party’s case, offers, counter-offers and statements (which might be characterised as threats or as thinking aloud) about future plans and possibilities.

 He gave WP protection to all of these.  So what of the new formulation above of the “concluded agreement” exception?  After all, P.J Carey had a perfectly “legitimate interest” in the terms of agreement as between Rush & Tomkins and GLC, really no less than Linsley & Mortimer, Mrs Pollard and EMW Law.  This raises the argument that, rather than trying to strongarm EMW, Muller and Cattley awkwardly into the concluded agreement /Oceanbulk exception, it would be more honest simply to declare that Muller was simply per incuriam and wrongly decided, rather as Laddie J was frank enough to suggest at first instance in Unilever, and that Cattley was also understandably wrong as being wholly reliant on Muller.

Perhaps the approach taken by Fancourt J in Briggs v Clay casts some light on this.  It will be remembered that he found that the WP material which Aon wanted to keep privileged was certainly relevant and might well have cast light on what the defendant lawyers wanted to prove.  He also found that it was caught by WP privilege unless an exception applied.  He might well have considered it to be caught by a broad interpretation of the “concluded agreement-rectification- factual nexus” exception (particularly the factual nexus aspect, in showing that Aon and the lawyers GG and QC must have known about what was going on in the negotiations).  But he ruled that it was not right or necessary to admit such WP material for the trial judge to be able to adjudge the issues fairly.  In searching for a coherent explanation of Muller, Fancourt J found that the issues were justiciable without the WP material.  In Berkeley Square Holdings, Roth J found that the issue before the trial judge in that case would not be justiciable unless the WP material was admitted.  But a similar justification for the decision in Rush v Tomkins might have been available too, though in truth it was never hinted at by Lord Griffiths, whose formulation of the principle reads as formidably comprehensive in effect, especially when modified authoritatively by Unilever. 

Does the fact either that the Mullers asserted that they had mitigated their loss in settling, or that the defendant solicitors (who were external to the WP privilege) alleged failure to mitigate against the Mullers’ claim to recover a shortfall, explain the decision adequately?  It presumably matters not whether the assertion was made in the claim, as in Muller, or was raised as a defence, as happened in EMW Law.  It was clear that the WP documentation might possibly have shed light on that issue (though this was never explained) and was thus evidentially relevant.  But as Lord Griffiths suggested in Rush & Tomkins, and as Fancourt J said in Briggs v Clay, WP material almost always would be relevant, but that does not of itself justify its being admitted or the cloak of privilege being pulled aside.  Whether or not the Mullers had mitigated their loss was undoubtedly a live issue.  But, as queried above, was admission of the mediation documents necessarily going to cast light on it, and were there not other ways to decide that issue?  And cannot the same points be made about Cattley v Pollard?

Maybe another question to ask is whether any of these decisions actually seem wrong on the facts from a common-sense view?  The answer is probably No, though a mediator might have felt that insufficient weight was placed on the interests of protecting mediation as a process.  In each of them it can be understood why the judge felt that the outcome was right and fair.  One might even put it that “the interests of justice” were served in each case.  Putting it another way, except for Briggs v Clay, each of those cases in which WP material was admitted was regarded as being “only fairly justiciable” if the WP material was admitted. It was only in Briggs that Fancourt J considered the issues to be determined at trial and decided that they could properly be decided without admitting the WP material.  He for one resisted the unprincipled lure of the judge’s easy way out.

All of these cases were interlocutory applications for disclosure, and not the substantive hearing.  One party wanted if possible to forestall the presentation of what it presumably feared was damaging evidence, or, in Berkeley Square Holdings, by getting favourable evidence admitted.  Therefore, in deciding the interlocutory point, the judge had not heard the full range of evidence, oral and documentary, which might impinge on the issues so as to decide whether what was privileged was necessary to make the case “justiciable” at trial.  In Briggs, Fancourt J bravely reviewed the issues in advance so as to make such a decision, perhaps weighing in the balance the need to protect WP privilege if at all possible and only to declare an exception if there was no other way to decide the case.  Judges are remarkably skilled at hearing allegedly privileged evidence and yet still ruling that it is inadmissible if so persuaded and going on to ignore it.  Maybe they should do so more often, rather than considering such questions in advance at a disclosure hearing.

A rather similar argument might have arisen over whether the mediator could or should have been called to give evidence, and about what, in Farm Assist v DEFRA (No. 2).  Was it the only way that duress could be made out and fairly adjudicated upon “in the interests of justice” for the judge to order that the mediator should give evidence of what passed between the parties conveyed by her and even what one party said to her in private conversations which she had undertaken to keep confidential? Ramsey J ordered the mediator to give evidence “in the interests of justice”, overriding her identified and justified right to confidentiality.  Was that “in the interests of justice”?  I would certainly argue No to that question.  In Farm Assist, the parties had waived WP privilege and were perfectly capable of telling a judge what each said to the other and what the mediator had conveyed, without the judge necessarily needing to hear from the mediator.  There were other “interests of justice” at play in that situation, especially the desirability of keeping mediation as a broadly safe environment in which parties need not fear that a mediator’s promise of confidentiality and neutrality will be suborned by order of the court.

The interests of justice must be founded on principle and precedent which can be carried from one decision to another.  So can a consistent but restrictively drawn exception to WP privilege be discerned from these decisions which can be safely labelled “the Muller exception”, and added in its amended form to the Unilever list as point 6? There is some parallel between these situations and the loss of legal professional privilege when a client sues a former lawyer., rendering what would otherwise be confidential material available to the judge to decide whether negligence is made out.  That quite neatly solves the Muller case itself, where the defendants were former solicitors.  But it did not apply to the other “former solicitors” cases – Cumbria and Briggs – where an earlier WP privilege was protected.  And it has no bearing on Rush & Tomkins or Cattley, neither of which involved professional defendants.

So are we really left with another exception to WP privilege as being that “WP material may be admitted, regardless of any objection or lack of support from all those with the benefit of that privilege, if it relates to a fundamental aspect of a case put at issue by either a party to a privilege or a third party with a legitimate interest in the WP material, which is incapable of fair decision (i.e. not justiciable) without it”?  This is a dangerously wide proposition.  The “interests of justice” are suddenly elevated to a degree of which Fancourt J might well disapprove and almost certainly Lord Griffiths too.  Such a formulation was not necessary to decide Berkeley Square Holdings, where the decision accords with common sense, and where what was really a version of the Oceanbulk exception was the real ratio, with the application of Muller actually obiter.  Similarly, in EMW Law, the first basis for admitting the WP material was that it was covered by the concluded agreement exception, albeit applied in a rather wider way than previously applied.  And in Briggs, the material was excluded, it not being necessary to look behind the WP privilege to decide the case. Once again, Briggs might be fairly interpreted as another instance of the Oceanbulk principle, though rightly subject to exclusion within the discretion of the judge.

I would argue that it is time that Muller and decisions based on it were consigned to the history of procedure and regarded as inconsistent with the Group A cases, particularly Rush & Tomkins, Unilever and Ofulue.  At most it can be regarded as just another example of the “concluded agreement” exception, broadly drawn.    Subsequent judges have acknowledged – or perhaps merely assumed for their own purposes – that Muller was rightly decided on its facts, though never giving it an enthusiastic endorsement.  There is certainly more than a hint that later judges have tacitly accommodated Muller without choosing to remark that it is simply in conflict with the strong decisions in Group A.  Any principle that subsequent judges have sought to wring out of the decision was not one of those on which Hoffman, Leggatt and Swinton Thomas LJJ actually decided the case.  Might it simply be wrong, as comparison with Rush & Tomkins suggests, with subsequent judges (apart from Laddie J at first instance in Unilever) not having suggested as such simply out of deference to a distinguished judge?  After all, the Court of Appeal certainly does get things occasionally wrong, as is witnessed in relation to mediation by Halsey v Milton Keynes NHS Trust in finding that ordering mediation offends Article 6 of the ECHR.

What Briggs also demonstrates is that it is perhaps only at a trial that a judge can tell whether there is enough material to decide the issues, balancing out the relative force of non-privileged oral and open documentary evidence.  Fancourt J had to undertake an imaginative exercise as to whether the WP material might be needed and ruled (on an advance application for disclosure) that it was not.  How much easier it would have been to decide that when as trial judge he know what other non-privileged evidence had been given.


14.  The Effect of this Debate on Mediations

Those concerned for the wellbeing of mediation and the general security of settlement discussions should not despair at this debate.  Thousands of mediations and tens of thousands of settlement discussions are conducted every year within the safety of WP privilege and from which no controversy arises.  The framework of privilege is generally sound, making it possible for a very high percentage of cases issued in the courts of England & Wales – perhaps as high as 95% – to be settled before trial.  This is just what Lord Woolf’s Access to Justice reforms and the Civil Procedure Rules 1998 aimed to achieve.  Litigated trial is indeed the last resort now.  Use of mediation as a dispute resolution process has grown exponentially in recent years, significantly due to judicial encouragement and direction.  There is at least anecdotal evidence of high satisfaction levels with mediation, and such formal surveys as have taken place have confirmed these.  So as usual judges are asked to deal with difficult and rare sets of facts over which the parties have chosen to, and could afford to, litigate.  Once again, the haphazard nature of common law development is striking.

But complacency is unwise, hence this close examination of the very few cases where that privilege has been challenged and on occasions set aside.

There are five provisions normally in formal mediation agreements which are caught by this debate:

  • the express agreement that WP privilege will cover mediation content, which includes documents prepared for the mediation (such as position papers), and specifically disclosed WP for the purposes of the mediation, but not documents which are in any event disclosable;
  • overall confidentiality contracted by the parties and the mediator;
  • each mediator’s contracting to observe confidentiality as to the content of private meetings with each party during the course of the mediation, coupled with the mediator’s undertaking to abide by a Code of Conduct which requires independence, neutrality, and absence of conflict of interest:[72];
  • The parties’ undertaking not to call the mediator to give evidence about the dispute or the mediation in any subsequent proceedings; and
  • The usually inserted contractual formality that for any settlement terms to be binding, they must be recorded in writing and signed by each party (though not the mediator as a matter of practice).

The overarching comment to make about the law as it has dealt with mediation is that it has hitherto only come before the courts extremely rarely in England & Wales.  Really very few controversies over the above terms have emerged, and not all the cases considered in this paper arose after a mediation.

Judges certainly regard WP privilege as attached to mediation material, but they are also likely to regard mediations no differently from any other WP process initiated to explore settlement of a dispute and, as noted earlier, have been ready to subject mediation to the same exceptions to WP privilege.  So mediators and parties cannot be surprised if, in a controversial case, a party or someone with a legitimate interest in the outcome of a mediation conducted WP seeks to argue that the court is entitled to receive in evidence otherwise WP material in order to decide whether agreement was reached, and if so what that agreement was, or what it means (the “concluded agreement exception”).

The main unresolved question about that is whether the requirement for any binding agreement to be in writing and signed (point 5 above) means that, in the absence of any such written signed agreement, there cannot have been a binding agreement[73].  Nor can the mediation community be surprised if courts are prepared to look behind the curtain of privilege to see if an agreement has been based on misrepresentation, fraud, duress or other “unambiguous impropriety”, so long as coherent grounds for doing so are raised by an owner of the privilege or someone with a legitimate interest.  Also, no senior court has yet decided what the true extent and effect of the confidentiality clauses typical of mediation agreements can mean – whether they add to WP prejudice, especially if they are enforceable at the suit of the mediator or mediation provider who signs the agreement. or whether they effectively run in parallel to WP privilege and are likely to be enforced or set aside in similar circumstances.  What is clear is that mediation in England & Wales cannot be regarded as a sealed-off area into which judges cannot gain access.  Even though such access has been sought on very few occasions, and sometimes on uncertain or unclear grounds, the rights of parties under the common law must be subject to external policing by the courts.

The cases above mostly deal with documentary evidence, which may be of surprise to mediators, who, as noted above, are mostly looking at documents which are openly disclosable in the litigation.  The main privileged documents to be encountered are so-called position statements drafted and usually exchanged prior to a mediation, though these are quite often not particularly conciliatory.  Further WP documents which might excite the interest of a court might comprise e-mail exchanges and draft documentation after the mediation day if the case is not settled, and in a few cases these have been seen by a judge, though never so far against the wishes of an owner of WP privilege[74].   There would be concern about wholesale admission of what was said orally during a mediation between the parties and especially in private to a mediator who has undertaken a deeper level of confidentiality about such private discussions.  The idea apparently accepted by Ramsey J in Farm Assist v DEFRA (No.2) that a mediator could be asked about private conversations with each party during a mediation is shocking, and it is to be hoped that one day a senior court will outlaw such an idea.  In terms of such matters as alleged misrepresentation, duress, or estoppel, it is surely only the material that is exchanged between parties direct or through the agency of the mediator that could possibly affect the basis for settlement[75].

If judges seek to hear evidence of what was said, there will continue to be a risk that they might want to hear evidence from mediators.  So far only two mediators have ever given evidence in an English court[76], and it is not clear whether they simply acceded to being asked or were ordered to do so.  It is to be hoped that mediators will be brave enough to resist such invitations in the future so as to clarify their legal position further, preferably with the support of mediation provider organisations in a position to intervene and argue the case for mediation as a proper concern for “the interests of justice”.

A slightly different lesson can be drawn from the situations that led to the difficulties discussed in this paper, which relates to mediation design.  Wherever possible, all those involved in the ramifications of a given dispute need to be encouraged to participate in a mediation, so that all possible rights and consequences can be worked out through one united effort.  If L&M had been a party in the mediation with the Mullers and S Ltd, they would have been within the WP loop, as would EMW Law in EMW v Halborg, or Mrs Pollard with Cattley and SIF.  If for any reason they refused to participate, they could hardly then be heard to complain about not knowing what transpired in WP negotiations.  In Oceanbulk and Berkeley Square Holdings, the relevant parties were involved in the WP discussions, as were the lawyers (in both sets of negotiations) in Briggs v Clay.


15.  A Need for Further Reform?

In relation to WP privilege, this paper does not suggest a need for broad reform.  It is more a plea for judicial awareness and care when making rulings which might impinge on mediation and its usefulness as an integral part of dispute resolution in England & Wales.  Perhaps the main purpose of this paper is to remind judges that the interests of justice may well properly be assessed in the light not just of easing the way towards their decision-making, but also encompasses the health and wellbeing of dispute resolution processes outside the courts, and in particular in relation to the two levels of confidentiality which apply to formal mediation.  They should continue the trend that many have articulated that exceptions to WP privilege should be rare and not extended further than the existing state of the law of evidence.  They should resist early decision-making as to whether WP material has to be admitted for a case to be justiciable, as often it will not be needed to determine whether one party properly mitigated their loss or double counting has occurred or their interests have been properly look =ed after in a costs assessment process – the circumstances which arose in Muller, Cattley and EMW Law.  Judges are well able to consider such matters in the light of open evidence at trial and will usually be able to preserve WP material from scrutiny in reaching those decisions.  More importantly, parties seeking early access to such WP material should still be able adequately to assess their risks of not succeeding on such arguments in the light of disclosed settlement terms to choose whether to settle their claims or fight them.  At that stage they may well be glad to be able to use the safe environment of a mediation to sort out their differences.

When judicial decisions on mediation such as in Brown v Rice & Patel and Farm Assist v DEFRA (No 2) emerged, there was considerable concern expressed by some mediators about the risks to the security of the mediation process posed by the kind of inquiry that such cases seemed to authorise.  But others, notably William Wood QC in his two articles[77] When Girls Go Wild and Mediation Privilege, pointed out that users of mediation would be uncomfortable about entering a wholly unreviewable process if it meant that rights would be lost.  He noted, for instance, that under the largely impenetrable confidentiality screen in California, courts could not even rectify a typographical error in a settlement agreement where a zero had been wrongly omitted from the settlement sum.  The present state of the law in England & Wales would permit rectification of such a mistake, and rightly so.  But even so, Wood pleads for proper protection for mediators when engaged in private exchanges with each party.

Ten years or so on from that stage of the debate about mediation confidentiality, we can at least feel relieved that the value of mediation has not been subverted by the various occasions and sets of facts that have drawn judges into that debate.  Indeed, it flourishes and grows.  This is in part due to the fact that many in the judiciary now will themselves have participated in mediation as advocates and even as mediators before appointment.   However, the powerful, perhaps unintended, effect that Halsey v Milton Keynes NHS Trust[78] had in chilling its growth is a reminder that it only takes one ill-considered decision (dare one say with hindsight, a view of Halsey now shared by many judges) to make a dramatic impact on mediation’s availability and usefulness.  This paper has sought to clarify the rules relating to judicial access to the private exchanges that characterise mediation without in any way suggesting that judges have no place in correcting misunderstandings and misfeasance which takes place within that environment.  So long as they are clear, mediation will not be surprised that they are occasionally mobilised “in the interests of justice”, so long as “the interests of justice” are interpreted broadly and not simply to make it easier for judges to decide cases.  Judges might always feel that seeing WP material will help them, and it might.  But if it was created in order to explore settlement, there really is a genuine public policy interest in protecting parties from disclosure of what might have been provisional, exploratory or even a try-on.  It should suffice in the vast majority of cases to know what the settlement terms were without needing to see how they were reached.  Similarly (and with great respect to Ramsey J in Farm Assist (No 2), in which he seemed willing to allow evidence from the mediator, even of private oral exchanges with each party) there is a really significant public interest in minimising the use of mediators as witnesses.  Once a mediator is seen to have taken sides (even if this is because a judge has effectively compelled them to choose which party’s case to support in the witness box), their neutrality and quite probably their reputation will be compromised.

As to any need for further reform or protection which does not at present clearly exist, to echo and quote William Wood in his Mediation Privilege article of 2009:

The reform which may be needed in this area is the recognition by the Courts of the need for special protection for the confessional exchanges between mediators and individual parties, the truly private core of mediation activity.  This is now the cutting edge of the mediation confidentiality issue.

He wrote this in 2009 and in support referred to the articles by Lord Briggs (then Sir Michael)[79], who himself suggested that a kind of mediator privilege might be judicially developed in relation to such private exchanges, as an extension of Re D.  Nothing has happened with this idea so far.  We may have to wait until another chance set of relevant facts bubbles up, in the way so familiar to the haphazard development of the common law, when a litigant is stubborn enough and funded enough to ask the Court of Appeal to consider this.

[1]   [2017] EWHC 1014 (Ch)

[2]   [2019] EWHC 102 (Ch)

[3]   [2020] EWHC 1015 (Ch)

[4]   [2021] EWCA Civ 551

[5]   Mediator Magazine December 2008, about which further comment follows towards the end of this paper

[6]   See Rush & Tomkins v GLC discussed fully below

[7]   See Chantry & Vellacott v The Convergence Group [2007] EWHC 1774 (Ch)

and Earl of Malmesbury v Strutt & Parker [2008] EWHC 24 (QB)

[8]   [1984] Ch 290

[9]   [1989] AC 1280

[10]    23 Q.B.D. 335

[11]  [2000] FSR 869

[12]   [1996] PNLR 74, so still not a in a major law reports series

[13]  [2000] 1 WLR 2436, and in Ofulue v Bossert [2009] HL 16

[14]  In Briggs v Clay [2019] EWHC 102 (Ch)and in other cases cited therein

[15]  Indeed as will be see below, this is what happened in Briggs v Clay

[16]   [1969] 1 WLR 1378

[17]   [2006] EHWC 3130 (Ch)

[18]   [2000] 1 WLR 2436, decided before Cattley and discussed fully below.

[19]   [2006] HL 37

[20]   [2009] HL 16

[21]  [2000] 1 WLR 2436

[22]   See for instance Stuart Isaacs QC sitting in the Chancery Division in Brown v Rice & Patel [2007] EWHC 625 (Ch) when he described mediation as “assisted without prejudice negotiation”, which perhaps somewhat oversimplifies its status, as it overlooks the implications of the formal agreement to mediate which rarely if ever is used to frame other WP negotiations: in particular the effect of explicitly contracted confidentiality.

[23]   For the suggestion that a very carefully worded agreement to mediate could prevent judges from applying exceptions to WP privilege, see Dow v Bombardier Supreme Court of Canada [2014] SCC 35.  For contractual confidentiality in mediations see section 10 below and the cases discussed there.

[24]   [2010] UKSC 44

[25]    [1969] 1 WLR 1378

[26]    [2007] EWHC 625 (Ch).

[27]   [2018] UKSC 24

[28]    See my article The requirement for written signed agreements for a binding mediated settlement available through CEDR or from the author

[29]    [2009] EWHC 1102 (TCC)

[30]    [2016] EWCA Civ 717

[31]  [1997] F.S.R. 178 at 191

[32]   Unrep 10 Nov 1992 CA

[33]   See above

[34]  See Savings and Investment Bank v Fincken [2003] EWCA Civ 1630

[35]  [1889] 23 QB 335

[36]  [1996] PNLR 74

[37]  See Fancourt J in Briggs v Clay [2019] EWHC 102 at [52] (Ch) and also Avonwick v Webinvest Ltd [2014] EWCA Civ 1436 at [21] and per Newey J in EMW Law LLP v Halborg [2017] EWHC 1014 (Ch);

[38]  [2020] EWHC 2996 (Ch)

[39]  [1984] Ch 290

[40]    Interestingly in the Supreme Court of Canada decision of Dow v Bombardier [2014] SCC 35, it was held that parties could actually agree to exclude the effect of all exceptions to the WP rule and make the privilege absolute if they so chose and used very clear language to do so.

[41]   [1993] 2 All E.R. 693, at 697.

[42]   [2020] EWCH 1015 (Ch)

[43]   [2010] UKHC 44

[44]   One further rather dubious exception was canvassed by Lord Griffiths in Rush & Tomkins v GLC based on the ancient decision in Waldridge v Kennison (1794) 1 Esp 142, where a without prejudice letter was admitted solely as evidence of the writer’s handwriting.  It was doubted as being sound both there and by Lord Neuberger in Ofulue v Bossert, and is only mentioned here for completeness because it was unsuccessfully raised in the cross-appeal of Lancer Properties in Berkeley Square Holdings v Lancer [2021] EWCA Civ 551 at para 95.

[45]   [2017] EWHC 1014 (Ch)

[46]   [2019] EWHC 102 (Ch)

[47]   [2020] EWHC 1015 (Ch)

[48]   by the Court of Appeal in Unilever and the House of Lords in Ofulue

[49]  [2017] EWHC

[50]  There were also issues relating to relevance and to legal professional privilege

[51]   [1969] 1 WLR 1378

[52]   [2007] EWHC 625 (Ch)

[53]  [2014] EWCA Civ 1436: an emergency appeal in which Lewison LJ said he would not embark on a full discussion of WP privilege.  The Court found that there was no dispute on which to found WP privilege, and although the outcome of Muller did not on its facts generate surprise, as cited above, Muller was disapproved of by Ofulue, and was wrong to distinguish between protected admissions against interest and other WP material.

[54]  [2019] EWHC 102 (Ch)

[55]   See above:  [2017] EWHC 1014 (Ch)

[56]   See discussion above of this case

[57]   As was done for instance in Brown v Rice and Patel

[58]   [2020]  EWHC 1015 (Ch)

[59]  By which he presumably means “only fairly justiciable”

[60]  [2021] EWHC 2996 (Ch)

[61]  Toulson & Phipps Confidentiality 2nd edition Ch. 3.

[62]   [2008] EWHC 786 (QB)

[63]  But not an argument that later stopped Newey J from admitting WP material in EWM v Halborg against the wishes of Halborg’s clients (admittedly they were his parents) who claimed not to have settled their costs dispute with Savage Hayward and might have been prejudiced by admission of WP material as a result. Cumbria Waste is not referred to by Newey J as having been cited to him.  This is yet another anomaly to note in this complex discussion.

[64]   Although the point did come up before Lloyd J in Instance v Denny Bros [2000]. It was also accorded some validity in Mason v Walton-on-Thames Charity [2010] EWHC 1688 (Ch)

FSR 869.

[65]   [2009] EWHC 1102 (TCC)

[66][66]   See this author’s paper Mediators as witnesses and evidence about mediators in Mediation Theory and Practice: 4.1 (2019)

[67]  Under CPR 32.1(1) the court has extensive powers over admission or exclusion of evidence, being able to exclude admissible evidence

[68]    [2018] EWHC 1763 (Comm) at [54]:

[69]   Note that in Farm Assist v DEFRA (No 2) privilege had been expressly (if unwisely) waived by both parties.  Ramsey J was therefore free to hear evidence of alleged duress in what happened at the mediation.  The issue there was whether the mediator could be required to give evidence about what was said and done at the mediation. Similarly, privilege was waived by both parties in Earl of Malmesbury v Strutt & Parker and Chantry Vellacott v the Convergence Group.  The question considered here is confined to unilateral attempts to open up WP material. And in Avonwick Holdings v Webinvest, the ratio was based on the absence of a dispute, with a rather dismissive mention of Muller, which was not the real basis for that decision.

[70]   [2011]  UKHL 38

[71]   although Roth J in Berkeley Square Holdings rejected any difference of approach being required in principle as between tw0- and three-party cases

[72]   both of which are theoretically enforceable by suit against the mediator, though this has so far not happened in this jurisdiction

[73]   See MWB v Rock Enterprises on the validity of no oral variation clauses in commercial contracts

[74]   See for instance AB v CD, Ferster v Ferster and Universal Satspace v Govt of Kenya, and this authors discussion of this topic in Mediation Theory and Practice (4.1 (2019)).

[75]  In two articles in the April 2009  New Law Journal entitled Mediation privilege, Lord Briggs (then Sir Michael) suggested that courts might well be prepared to discern a kind of privilege as to what mediators were told in private, maybe developing this from Re D

[76]  See AB v CD and Abberley v Abberley.  There was never a trial in Farm Assist v DEFRA, so the mediator never gave evidence despite being ordered to do so.  The mediator made submissions in Cattley v Pollard.

[77]  In The Mediator Magazine 2008 and 2009

[78]  [2004] EWCA Civ 576

[79]   In the New Law Journal April 2009

Tailored Insights for You

Sign-up to receive regular insights on topics ranging from effective Conflict Management and Negotiation to Commercial Mediation and ADR Thought Leadership.