Gubarev & Anor v Orbis Business Intelligence Ltd & Anor

This is the second case to come before this Court as a result of an article published by

The first claimant (“Mr Gubarev”) is a businessman and entrepreneur, born and brought up in Russia, but currently resident in Cyprus. Mr Gubarev is the ultimate beneficial owner of approximately two thirds of the XBT Group of companies. The main business of the XBT Group is the provision of server capacity to individual and corporate customers. This is a worldwide business, carried on by a network of companies in Europe, America, and Asia. The corporate claimant (“Webzilla Ltd”) is a member of the XBT Group. Webzilla Ltd was incorporated in Cyprus in 2005, as an enterprise hosting company.

Mr Gubarev and Webzilla Ltd seek damages and other remedies in respect of words contained in a document published alongside the BuzzFeed Article. The article referred to “A dossier, compiled by a person who has claimed to be a former British intelligence official, [which] alleges Russia has compromising information on Trump.” The “dossier” of “reports” referred to was embedded in the article by means of a document viewer. The reports took the form of numbered Intelligence

The full text of the BuzzFeed Article is set out in Appendix A to this judgment. The full text of the December Memorandum is set out in Appendix B. The words complained of are as follows:

“[redacted] reported that over the period March-September 2016 a company called XBT/Webzilla and its affiliates had been using botnets and porn traffic to transmit viruses, plant bugs, steal data and conduct “altering operations” against the Democratic Party leadership. Entities linked to one Alexei GUBAROV were involved and he and another hacking expert, both recruited under duress by the FSB, Seva

KAPSUGOVICH, were significant players in this operation. In Prague, COHEN agreed contingency plans for various scenarios to protect the operations, but in particular what was to be done in the event that Hillary CLINTON won the presidency. It was important in this event that all cash payments owed were made quickly and discreetly and that cyber and other operators were stood down / able to go effectively to ground to cover their traces.”

Lawyers acting for the claimants promptly sent letters of claim to BuzzFeed, and on 3 February 2017, Mr Gubarev and two XBT Group companies brought defamation proceedings in the United States (Florida) against the operator of the BuzzFeed website, BuzzFeed Inc, and its editor, Benjamin Smith (“the Florida Proceedings”). Later the same day, BuzzFeed redacted references to Mr Gubarev and the XBT Group companies from the article, and made a public apology for including them. On 19

Letters of claim were also sent to the defendants in the present case, Orbis Business Intelligence Ltd (“Orbis”), and Christopher Steele (“Mr Steele”). The claim was issued on the same day as the Florida Proceedings, 3 February 2017. Originally, there were four claimants: Mr Gubarev, Webzilla Ltd, and two other XBT Group companies, XBT Holding SA and Webzilla BV. The other companies have since discontinued their claims in this Court. The claim is brought in respect of the publication of the BuzzFeed Article within the European Union. It is not against BuzzFeed or Mr Smith, but against Orbis and Mr Steele only.

Mr Steele is the person who created the memoranda in the “Dossier”, which is why it has become known as “the Steele Dossier”. Mr Steele is a former Crown servant, a British citizen, resident in England. In compiling the Dossier, he was acting in his capacity as a director and/or employee of Orbis, a company incorporated in England and Wales which provides corporate intelligence services. The PEM were created pursuant to a commission from a Washington DC consultancy called Fusion GPS (“Fusion”), which was acting on the instructions of a law firm, Perkins Coie LLP. The ultimate client was a person or body in the upper echelons of the Democratic Party. The December Memorandum was produced by Orbis outside the scope of that commission.

The case pleaded in paragraph 6 of the Particulars of Claim is that the defendants

“published and/or caused to be published to the world at large, and thereby to vast (and unquantifiable) numbers of people across the European Union”, the words set out above. The claimants’ case is that those words bore the following natural and ordinary meaning:

“that the Claimants had deliberately and without consent hacked into the IT systems of the leadership of the United States Democratic Party and had used such unlawful access to transmit viruses, plant bugs, steal data and alter files and programs.”

The claimants allege that there was widespread publication in the EU of the words complained of, conveying this imputation, and causing each of them serious reputational harm. Mr Gubarev complains of hurt and distress. At one stage, Webzilla Ltd advanced a substantial claim for special damages. After four amendments of the Particulars of Claim, that is no longer pursued. Webzilla Ltd does however maintain a claim for general damages, reflecting what it contends is serious financial loss.

It is not in dispute that the words complained of referred to both claimants, bore meanings defamatory of both of them, and caused Mr Gubarev serious reputational harm. No issues of foreign law have arisen. (In principle, a claimant suing for libel in respect of words published abroad must establish that the offending words are civilly actionable in each foreign jurisdiction relied on; but the trial was sensibly conducted, and I proceed, on the common assumption that foreign law is no different from English law (reflecting the so-called “presumption”)). The defendants do not advance any substantive defences to the claim: they do not maintain that any defamatory meaning which the offending words might convey about the claimants was true, or an expression of honest opinion, or a reasonable publication on a matter of public interest. At one stage a defence of qualified privilege was relied upon, but a change in the way the claimants put their case has led to that issue falling away.

By the end of the trial, there were only four main issues:

I heard oral evidence from Mr Gubarev, and from two further witnesses for the claimants: Nikolay Dvas (CEO of XBT Holding SA, the parent company of Webzilla Ltd) and Roman Grinin (of Hilltop Ads Ltd, a customer of Webzilla Ltd (“Hilltop”)).

Much of the factual background is common ground. The overall factual framework is reflected in an agreed joint chronology, a copy of which is at Appendix C to this judgment.

The claimants’ case on responsibility for publication relies on some of these agreed facts, and on documents. It is the documents, relied on by the claimants for this purpose, that make up the majority of the voluminous trial paperwork. Most of the remaining documentation relates to the allegations of serious financial loss.

As required by CPR 33.2, the claimants served hearsay notices identifying statements in the documents that would be relied on as evidence of the truth of their contents, and explaining why the maker of the statement was not being called to give evidence (see CPR 33.2). There are four such notices. Some of the documents relied on are memoranda, notes, emails or other electronic messages. Some are media reports, or passages in books (for instance, “A Higher Loyalty – Truth, Lies and Leadership” by James Comey, former Director of the FBI). But prominent among the documents relied on are written transcripts of testimony, depositions and declarations made by witnesses who gave evidence in the US, to Congressional hearings or in the Florida Proceedings, in 2017 and 2018, on matters relevant to the issue of responsibility for publication.

There are thirteen such witnesses. Taking their testimony in date order, the witnesses, and the testimony they gave to other Courts or bodies that was adduced in this way, are:

The reasons given by the claimants for not calling these witnesses at this trial are that their “memories are likely to have been more reliable” at the times when they gave their testimony in the contexts I have identified, and that it would be “disproportionate, in all probability unproductive and impracticable” to call them.

The claimants also rely on aspects of two documents in respect of which they have served no hearsay notice, the position being that this is not legally required, as these are formal published reports and/or judgments dealing with matters of a public nature and open to public inspection:

For the defendants, Mr Steele gave evidence in court and was cross-examined at length, and in detail, by Mr Caldecott QC. The defendants also called evidence from

This is a short and relatively straightforward issue, which it is convenient to address first.

The defendants dispute the meaning advanced by the claimants. They submit that, read in context, the natural and ordinary meaning of the words complained of was that:

“there were grounds to investigate whether the Claimants had been coerced by Russia into hacking the computers used by the Democratic Party leadership, transmitting viruses, planting bugs, stealing data and conducting altering operations.”

There is therefore a common core to the rival meanings: both reflect the fact that the words complained of link both claimants to the activities of hacking, transmitting viruses, planting bugs, stealing data and conducting “altering operations”. There are two main differences between the parties’ meanings. The first is that the claimants’ meaning is a “Chase Level One” meaning, whereas the defendants’ meaning is at “Chase Level Three”. The second main difference is that the defendants suggest the impression conveyed to the ordinary reasonable reader was that the claimants may have been “coerced by Russia” into the hacking and other misdeeds described in the offending words.

The nature of a “natural and ordinary meaning”, and the well-established principles for identifying that meaning in a defamation claim, are explained in Stocker v Stocker [2019] UKSC 17 [2020] AC 593 [34-38] and the judgment of Nicklin J in Koutsogiannis v The Random House Group Ltd [2019] EWHC 48 (QB) [2020] 4

Mr Millar QC has emphasised the importance of context in this case, reminding me of authority in the form of Charleston v News Group Newspapers Ltd [1995] 2 AC 65, and Dee v Telegraph Media Group [2010] EWHC 924 (QB) [2010] EMLR 20 [2932] (Sharp J, DBE), as well as the discussion in Gatley on Libel and Slander 12^th ed (2013), ¶3.33 – 34. He submits that the BuzzFeed Article and the December Memorandum are closely connected material published at the same time on two different web pages of a news website, and the ordinary reader must be assumed to have read both in their entirety. I agree.

Mr Millar’s purpose in making this submission is to support his clients’ case that the publication complained of bears a lower meaning than the one advanced by the claimants. The principle he relies on is clear. Where a document written by one

For the claimants, Mr Caldecott QC has made seven main points:

For the defendants, Mr Millar makes the following main submissions:

In my judgment, the natural and ordinary meaning of the words complained of, in their overall context, is that:

Viewed in isolation, the December Memorandum would convey a Chase Level One meaning: that Mr Gubarev and Webzilla Ltd were – along with others - guilty of engaging in the conduct reportedly attributed to them by the unnamed source. This conclusion must follow from the unequivocal nature of the source’s allegations as reported in the Memorandum, and the application of the repetition rule (the rule that reports of allegations made by a third party will generally bear the same meaning as the underlying allegations themselves: see Brown v Bower [19-32]). Put shortly, the Memorandum presents the source’s allegations as fact, and that is how the ordinary reader would understand the position. The mere fact that the Memorandum is styled as an “Intelligence Report” does not affect this conclusion. The contents of the Memorandum itself do not suggest that there are any grounds for doubt about the veracity or reliability of the “intelligence” that is being reported. Rather the contrary. The December Memorandum, on its face, follows an earlier report on “secret meeting/s in Prague”. It contains “Further details” of the “secret dialogue” including contacts with “associated hackers”. The details are extensive, plausible on their face, and names are provided. Nothing is said to suggest that the source, or the information provided by the source, is or may be unreliable.

I do not believe that what I have said so far is controversial. But the reader through whose eyes I am looking at this material would come to the December Memorandum with a mind conditioned by the BuzzFeed Article. The article contains a mixture of information about the Dossier.

Some of that information would tend to portray the Dossier as a reliable source of information. The reader is told that the Dossier contains specific allegations made by someone who not only “has claimed to be” a former British intelligence official (subheadline), but is also “understood to be” a former British intelligence agent ([4]) – in other words, a professional intelligence expert, who has served a major Western power. That status tends to confer an aura of reliability on the source information. The reader is also told that the allegations have been circulating among journalists, lawmakers and intelligence officials for months. On the face of it, the allegations are being taken seriously in those circles (for instance, by David Corn, Senator Harry Reid and Senator John McCain). There is nothing in the BuzzFeed Article to suggest that anyone in these categories had dismissed the Dossier as unreliable. Equally, there is nothing in the article to undermine or cast doubt on any of the specific factual assertions contained in the December Memorandum. The “errors” identified in the article are not mistakes in that Memorandum, but in other memoranda. The reader would see that. Moreover, as Mr Caldecott submits, the identified errors are not central but minor.

On the other hand, there are the quoted denials from the Trump transition team. These are a factor in any reasonable assessment. I doubt that the ordinary reader would place any great weight on these. They are sweeping and general denials that deploy cliché (“fake news”), and attacks on the motives of those who have commissioned and reported them, rather than addressing the allegations by way of detailed factual rebuttal. Even so, the denials would have some impact.

More significant is the article’s odd and unusual mixture of sensationalism and caution, which would strike the ordinary reader. BuzzFeed was adopting a studiedly neutral stance in relation to the “explosive” allegations it was placing before its readers. It was plainly not setting out explicitly to endorse the veracity of any of those allegations. It was manifestly setting out to highlight factors that might cast some doubt on the allegations generally. BuzzFeed pointed out that the Dossier was commissioned by the President-elect’s political opponents. Repeatedly, the article stresses the unverified nature of the information and the fact that the “report” (meaning the Dossier generally) contained errors. Readers are told that investigations into some allegations by BuzzFeed reporters on two continents had failed to establish whether they were true or false. It said some were “potentially unverifiable”. It was expressly presenting the full document “so that Americans can make up their own minds about allegations … that have circulated at the highest levels.”

This approach does seem inherently contradictory: if the allegations were unverifiable, or in any event unverified, and BuzzFeed’s own investigations had failed to establish whether or not they were reliable, how were readers supposed to form a view? But that is another matter. My point here is that the article invited its readers to approach the Dossier with a degree of scepticism or at least caution, and gave them specific reasons for doing so; and that will have affected, by mitigating it, the defamatory meaning taken away from the article and the December Memorandum, read together.

Nobody could suggest, and Mr Millar has not argued, that this is a “bane and antidote” case – one where the poison of the defamatory allegations is wholly neutralised by other material in the same publication. But there are articles, and this is one of them, which report unequivocal allegations of wrongdoing in a context that

I should mention two further points, by way of explanation of my conclusion:

The law, since 1 January 2014, is that a libel claim cannot succeed unless the claimant establishes that the statement complained of satisfies each of three requirements: (a) the common law requirement, that the statement should have a defamatory tendency; (b) a requirement, emanating from statute, that the publication of the statement must have caused actual damage that is more than minimal; and (c) a further, and more demanding, statutory threshold for actual defamatory impact.

The common law requires that the offending statement should have a tendency to cause a substantial adverse effect on the attitude of other (right-thinking) people towards the claimant: Thornton v Telegraph Media Group [2010] EWHC 1414 (QB) [2011] 1 WLR 1985 [94] (Tugendhat J). This is an objective test, depending on the extent to which the meaning of the words has an inherently harmful character. The requirement of more than minimal actual damage was recognised by the Court of Appeal in Jameel (Yousef) v Dow Jones & Co Inc [2005] EWCA Civ 75 [2005] QB 946, where the Court held that the Human Rights Act 1998 imposed on it a duty to dismiss a libel claim which was so trivial that its continuation would involve a disproportionate interference with freedom of expression.

The higher statutory threshold was laid down by s 1 of the Defamation Act 2013, which contains what I have called the serious harm requirement:

“1 Serious harm

(1)

A statement is not defamatory unless its publication has caused or is likely to cause serious harm to the reputation of the claimant.

(2)

For the purposes of this section, harm to the reputation of a body that trades for profit is not ‘serious harm’ unless it has caused or is likely to cause the body serious financial loss.”

I am concerned only with this requirement, and only in relation to Webzilla Ltd. Otherwise, it is either conceded or I find that the requirements I have mentioned are satisfied.

The correct interpretation of s 1 has been litigated as far as the Supreme Court, which has now confirmed that section 1:

“not only raises the threshold of seriousness above that envisaged in Jameel (Yousef) and Thornton, but requires its application to be determined by reference to the actual facts about its impact and not just to the meaning of the words.”

Lachaux v Independent Print Ltd [2019] UKSC 27 [2020] AC 612 [12] (Lord Sumption, with whom the other Justices agreed). The burden of proof lies, of course, on the claimant. The issue for my decision is whether Webzilla Ltd has discharged that burden, to the civil standard.

There has been relatively little case-law on the meaning of s 1(2), but the following points made by Lord Sumption in Lachaux at [15] were an integral part of the chain of reasoning that led to the above conclusion:

“The financial loss envisaged here is not the same as special damage, in the sense in which that term is used in the law of defamation. Section 1 is concerned with harm to reputation, whereas (as I have pointed out) special damage represents pecuniary loss to interests other than reputation. What is clear, however, is that section 1(2) must refer not to the harm done to the claimant’s reputation, but to the loss which that harm has caused or is likely to cause. The financial loss is the measure of the harm and must exceed the threshold of seriousness. As applied to harm which the defamatory statement “has caused”, this necessarily calls for an investigation of the actual impact of the statement. A given statement said to be defamatory may cause greater or lesser financial loss to the claimant, depending on his or her particular circumstances and the reaction of those to whom it is published. Whether that financial loss has occurred and whether it is “serious” are questions which cannot be answered by reference only to the inherent tendency of the words.”

So, reputational harm that would otherwise be regarded as serious cannot be so regarded unless the claimant establishes, as a matter of fact, that the harm caused, or is likely to cause, serious financial loss. I do not think this means that proof of financial loss is the sole requirement in cases that fall within s 1(2). As Lord Sumption points out, there is a distinction between reputational harm, and financial loss that results from that harm. It is not hard to envisage cases in which a publication demonstrably causes a business serious financial loss, but the defamatory (or actionable) component of the publication is not serious, or cannot be shown to be causative of that loss. In principle, all cases that fall within s 1(2) require proof of serious reputational harm that results from the statement complained of and financial loss that is (a) serious, and (b) consequent on the reputational harm: see my observations in Undre v London Borough of Harrow [2016] EWHC 931 (QB) [2017] EMLR 3 [40].

In this case, what is complained of is publication within the EU. So, it is loss caused by that publication with which I am concerned. The company cannot rely on, or recover damages for, any loss it may have suffered as a result of publication in any other territory. Financial loss is not necessarily the same thing as loss of revenue; the Court must be concerned with the company’s overall position, asking itself whether a loss of profit has been established: see Undre [49]. The loss must be “serious”. This is an ordinary English word which requires no elaboration here; whether loss is serious depends on context: Brett Wilson LLP v Person(s) Unknown [2015] EWHC 2628 (QB) [2016] 4 WLR 69 [30].

There is room for inference rather than strict proof. As ever, the Court is entitled - if not bound - to draw sensible conclusions from evidence which it accepts. If that conventional proposition needs support in authority, it can be found in Lachaux, and in Brett Wilson. In Lachaux, Lord Sumption said this at [21]:

“The judge’s finding was based on a combination of the meaning of the words, the situation of Mr Lachaux, the circumstances of publication and the inherent probabilities. There is no reason why inferences of fact as to the seriousness of the harm done to Mr Lachaux’s reputation should not be drawn from considerations of this kind.”

A similar process of inferential reasoning led to my conclusion in Brett Wilson, on a default judgment application, that the claimant had established serious financial loss as a result of allegations published on a website called solicitorsfromhell. The allegations were, by the common law standard, highly defamatory. The uncontradicted allegation was that they had been widely published. They had (on the claimant’s uncontradicted case) demonstrably led to the loss of one prospective client. I accepted that on the face of the statement of case, the claimant had made out the inference that other clients had been lost to the firm.

But inference is not the same thing as speculation; there must be a sound evidential basis on which to infer that the publication is more likely than not to have caused serious financial loss. Proof that a statement with a seriously defamatory tendency

Webzilla Ltd’s pleaded claim, as it now stands, is that the publication complained of “will inevitably have caused [it] serious financial loss and is likely to do so in the future”. In support of that contention, two specific matters are relied on: (a) a decision of January 2017, by a “substantial UK-based customer”, not to proceed with a planned agreement significant to expand its dealings” with Webzilla Ltd; (b) messages received from “third parties in the marketplace” in December 2017 that are said to indicate “that the reputational damage caused by the publication … is continuing and has deprived it of profitable contracts”, or at least the opportunity to secure them. The Court is invited to infer, from all of the circumstances, “that these two instances … are examples of a wider pattern in which the Third Claimant has failed to obtain business that it would otherwise have done, and thereby suffered serious financial loss.”

This is not a claim for the recovery of damages for specific financial losses. As I have said, there was at one stage a special damage claim. Claims for special damages in defamation have an inglorious history. Many are brought, but few are established. Some fail at trial. Others find themselves abandoned along the way. The reasons for this are many and various, but there can be difficulty in identifying with precision the changes in a company’s financial position that followed a publication, and there is often insuperable difficulty in establishing causation. Even if a loss of revenue is established, there can be problems in assessing whether it led to a loss of profit, which was not made good by other means.

The difficulties can be compounded in cases where there is global business activity, using complex corporate structures. This is such a case, as will be evident from the corporate organogram produced by Mr Dvas, which is attached as Appendix D to this judgment. This (and Mr Dvas’s explanatory evidence, which I accept) reveal that XBT Holding SA is not a trading company, but the ultimate holding company of the group. It is hard to see how it could ever have advanced a tenable claim for financial loss. Webzilla BV is not a trading company in the ordinary sense of the term, either. Its primary functions are to hold the tangible assets of the Webzilla companies, and to sub-lease servers to other group companies, including Webzilla Ltd. The evidence is that, of the original corporate claimants, it is Webzilla Ltd that is the customer-facing company in Europe, offering customers dedicated servers as well as datacentre space. The other two corporate claimants have, understandably in the circumstances, abandoned their claims altogether.

Webzilla Ltd’s special damage claim was changed and reduced and then dropped. That happened in stages, culminating with the statement in the Re-re-re-Amended Particulars of Claim dated 19 June 2020 that “The Claimants no longer rely on the voluntary particulars of special damage served on 19^th June 2018.” The defendants point to some of the detail of this process, in support of their overall contention that

It is not necessary to set out exhaustively the claims and allegations that have been made and then withdrawn, amended, or dropped. It is sufficient to list the following as contentions about the financial consequences of publication that were previously advanced on behalf of Webzilla Ltd but have now been abandoned.

The abandonment of the special damages claims has been explained in the claimants’ evidence. The claim by XBT Holdings SA was discontinued in December 2017. The evidence is that this resulted from advice from expert forensic accountants. It is trite law that a mere holding company cannot pursue a claim for loss of trading profits, or additional expenses of trading. But Mr Gubarev says, “We had not fully appreciated the position until the accountant began analysing the claim.” At the same time, additional particulars were filed on behalf of the other corporate defendants, and further particulars in June 2018. Both Mr Gubarev and Mr Dvas acknowledge that the particulars of June 2018 contained a number of errors, discrepancies and assertions that could not be supported by the company’s own documentation. Shortly before the deadline for disclosure, Webzilla BV’s claim was discontinued and that of Webzilla Ltd, “refined”. Later, it became apparent that even the refined version of the case for Webzilla Ltd could not be sustained, and the case mutated to take its present form. I shall come to the explanations provided for these changes of position. For present purposes they are of no great consequence.

What is material is that, looking at the residual claim in this historical context, it is clear that the pleaded claim does not now involve any assertion that the offending publication led to the loss of any existing customer, or caused any such customer to cancel any contract, or harmed the company’s access to credit, or led to a downturn in revenue, or an increase in costs. The claimants’ pleaded case is not that the publication complained of caused the company’s financial position to be worse than it was before the publication complained of in these, or any other, respects. Rather, it contends that but for the publication it would have been better off than before, by gaining extra business.

It follows from this analysis that some of the evidence of the claimants’ witnesses on the issue of loss and damage has no relevance to the pleaded claims of Webzilla Ltd, as they now stand. For instance, although the claimants have abandoned any claim based on a downturn in revenue, the witness statement of Mr Dvas identifies that

Other aspects of the witness evidence have only indirect relevance. Mr Dvas’s statement contains evidence that explicitly relates to group companies other than Webzilla Ltd. He describes dealing with all incoming messages from customers of servers.com, denying the allegations and reassuring the customers. Messrs Gubarev and Dvas both speak, more generally, of the firefighting they had to do in the immediate wake of the offending publication, dealing with nervous customers of XBT group companies who needed reassurance. But neither confines himself to points that bear directly on the two specific allegations in support of the Webzilla Ltd claim.

I accord due weight to the evidence of Messrs Gubarev and Dvas about events they had to deal with in the immediate aftermath of publication including – in particular – the way that XBT group customers responded. This evidence does provide some context, when attempting to assess the likely position in relation to the relevant

“constituency”, namely people who read the defamatory message about Webzilla Ltd in the EU at a time when they were or would otherwise have been minded to place additional business with the company. Servers.com, according to the evidence of Mr Dvas, is a Cyprus-registered company that “offers similar products to Webzilla Ltd … but more standardized services.”

The value of this class of evidence for this purpose is, however, qualified, for two main reasons. First, there is the fact that the evidence does not expressly or impliedly tie any of this reaction to any customer of Webzilla Ltd, or even to publication in the EU. Secondly, the claimants have adduced no evidence that those who did express concern, and were reassured, went on to withhold business from any XBT Group company.

The law does recognise that those in whose eyes a claimant has been damaged are not likely to come forward and say so. Mr Dvas fairly makes the point that people who decide not to deal with a company do not generally explain their reasons. This is an

There are some general points to make about these categories of evidence.

This has important limitations.

Mr Mishra, or Raj, might have been an important witness, and was at one stage expected to give evidence at the trial. He did not do so. The evidence of Messrs Gubarev and Dvas makes it easy to understand why. Their witness statements give some detail - though not a great deal – about the problems with the Webzilla companies’ claims, how they emerged and why they led to the claims being dropped or amended. Broadly, Mr Gubarev and Mr Dvas both disclaim any responsibility. The explanation offered is that the task of working up these claims was given to Raj. Mr Gubarev had a light supervisory role only. He had no in-depth knowledge of the financial statements, and no reason to doubt the figures provided to him. Mr Dvas was not involved. Raj left the business in September 2018, before the process of collating all the relevant documents was complete. It was that process, carried out under the supervision of Mr Dvas and with the assistance of forensic accountants, that led to the emergence of the problems. Once the disclosure exercise was complete, it became apparent that Raj had (in Mr Gubarev’s words) “lost his ability to focus on fundamental parts of his day to day job” including his analysis in support of the claims. The claimants’ team concluded that even the refined special damages claim had to go.

In his witness statement, Mr Gubarev attributes Raj’s “loss of focus” to unspecified family medical problems. But he also explains that Raj had “overstated his qualifications”. That is a euphemistic way of putting it. In cross-examination, he confirmed that Raj had claimed to be an accountant, but agreed that Raj’s CV contained a false claim to have acquired a CPA (a US accounting qualification), and a false or misleading claim to have gained an MBA (Raj himself deposed that this was “not a real MBA”). On the evidence before me, the conclusion that the CV was a dishonest piece of misrepresentation is all but inescapable.

It is a notable feature of the history that the claimants only discovered any of this through the process of depositions in the Florida proceedings in 2018. In his witness statement, Mr Gubarev said that “we did not hire him based on his qualifications alone”, but on the strength of proposals Raj had made to improve the profitability of the XBT Group. It would have been more accurate for Mr Gubarev to say that he did not rely on Raj’s alleged qualifications at all. In cross-examination, he confirmed that he had not seen or asked for Raj’s CV, or taken up any references, explaining:

“We did not see his CV and I don’t know what was written on his profile on LinkedIn. Maybe I just did not go there. … He came from an Indian billionaire who wanted to buy our company. We trusted him. We took him on for a trial period ... He has proven himself … and after that we offered him a job. … The only one person I spoke to was an Indian billionaire who wanted to buy us.”

Mr Gubarev did not express or exhibit any shock at the news that his trusted CFO was an unqualified and dishonest individual. He suggested that the company’s financial statements were reliable, on the basis that other personnel were involved in their preparation, and they were audited by KPMG. Others no doubt were involved, but audit is not the same thing as the preparation of financial statements, for which Raj was the man with principal responsibility. In my judgment, this evidence not only casts a shadow of doubt over the reliability of Webzilla Ltd’s financial statements, it also indicates a lax attitude by Mr Gubarev towards the importance of accurate accounting by, or under the supervision of, professionals with established relevant expertise.

No evidence has been adduced from the expert forensic accountants who were evidently involved at an earlier stage. In the absence of any such evidence, and in the absence of the officer responsible for the implementation of proper accounting processes at the material time, Webzilla Ltd relies on evidence from Mr Gubarev and Mr Dvas.

As to the former, his own evidence is that he is not an accountant; that from 2016 onwards he had very little involvement with the financial aspects of any XBT Group company; and that he saw only general reports for the group as a whole, and would look at the overall revenue for the holding company, not the details of each company.

“Even now I don’t know what the revenue for … EU customers was”, he told me. This would help explain his inability to identify the false and unsustainable assertions contained in the claimants’ earlier statements of case, on the issue of damages. It necessarily means his evidence on this issue is of little value. Beyond this, there is the laxity I have identified, and his unabashed evidence that he deliberately deceived one customer. In cross-examination, he was shown an email he sent on 29 January 2017 to a Dutch-based representative of leasing companies, asserting that “We haven’t lost any customers”, and claiming that revenue had increased since the BuzzFeed Article. His evidence was that this was not true. He said it because this was an important customer, and “we wanted to look better than was actually the case.” This was one of a number of emails sent at the time, making the same point.

Mr Dvas appears to have been brought into the witness team to replace Raj. He describes himself as a technology executive. He holds no finance or accounting qualifications. Although I assessed Mr Dvas as a conscientious and honest witness, who was doing his best to assist the Court, the value of his evidence is restricted. He is in a position to give reliable evidence of fact about the aftermath of the publication complained of, and to produce the financial statements and other accounting documents of Webzilla Ltd. He presented as a thoughtful individual, with a good understanding of corporate financial statements. But although he worked for Webzilla Ltd from April 2013 to late 2015, his role then was that of a project manager, and he was not at the company at the relevant times when – on his account - he had only

“limited insight” into the company. He was not an expert witness. In particular, he is in no position to give any expert analytical evidence in support of the financial loss claim. As he concedes, “I am not an expert in quantification of damages”.

Mr Grinin cannot give evidence as to the financial impact on the company of the offending publication, and he is not relied on for that purpose. His evidence is concerned with how he responded, on behalf of Hilltop.

The pleaded case has not been established. The evidence makes clear that, although Hilltop is incorporated in England and Wales, with a registered office in London, it has scant connection with this jurisdiction. It was controlled and managed at the material times by Mr Grinin, who was its 100% beneficial owner. Mr Grinin is Russian. He lives in St Petersburg, and has done for 30 years. The London address was for correspondence only, which was forwarded by London lawyers to Mr Grinin. Asked what physical presence the company had in the UK in January 2017, he answered “none”.

The Particulars of Claim said that the claimants “cannot be certain” that the loss alleged under this head resulted from publication within the EU. In opening, Mr Caldecott rowed back from this, conceding that this could not be established to the civil standard. Mr Grinin’s witness statement did not support any such allegation. It referred to the publication of “the dossier, alleging Webzilla’s Aleksej personal involvement in criminal activity” (sic), but gave no information as to how Mr Grinin came to know this. His oral evidence undermined the pleaded claim. It was that “some of [his] acquaintances” had sent him a link to the BuzzFeed Article; that when that happened “I remember for sure I was in Russia”; and that he could not say where the acquaintances were.

There are difficulties, as well, with the claimants’ pleaded case that publication led

In cross-examination, he made clear there had been no contract for expansion. He said, “we had a preliminary agreement about … expansion” but there was nothing in writing. He was clear that he had not called a halt to the scheme after learning of the “scandalous story”. He said, “I suggested we pause with this expansion for a while”. He entered into a contract with Hetzner which cost only $233 a month at first, growing to US$3,600 per month in 2019. This was a cheap and basic service for “low capacity servers”, not a substitute for the much more expensive backup scheme he had earlier discussed with Mr Gubarev.

No evidence has been adduced about the likely profit on this or any new business. The financial statements of Webzilla Ltd show total revenue for 2016-2018 of between US$14m and US$18.5m. After deducting the cost of sales, the gross profit margin was between about 35-50%. But administrative, selling, and other operating expenses reduced that to an operating profit of some 3-9%, before tax. The accounts barely provide a sufficient basis to infer that additional income of $30,000 a month would have translated into a profit, the loss of which would count as “serious”, for

I approach Mr Gubarev’s evidence on this issue with some caution, for the reasons already given, and because of his obvious lack of independence. But even Mr Gubarev does not suggest that Hilltop had made a firm commitment to doubling its spend with Webzilla Ltd. The discussions that took place are not documented in any way. Three months after they began, it seems that nothing had been put in writing. Mr Grinin was in the best position to give reliable evidence of what would have happened. He could have said, if that was the case, that he would otherwise have gone ahead with the backup plan which the parties had discussed. He did not say that. Hilltop was not contractually committed. It is far from clear that it was committed as a matter of business policy.

In the event, Hilltop did not sever its connections with Webzilla Ltd. It made a strategic decision, “weighing all factors”, to pause and not to have a full backup, at least for the time being. It retained Webzilla Ltd’s existing services, adding a modest supplementary arrangement with another enterprise, at a cost that was 10% or less of the price of the full backup service that had been discussed. Hilltop has evidently carried on business in that way to its own satisfaction for more than 3 ½ years, thereby saving itself some US$25,000 per month or $300,000 a year. On the balance of probabilities, my finding is that this strategic decision would have been taken anyway. If the sums spent with Hetzner, or the profit on such revenues, were the measure of what Webzilla Ltd lost that could not, in the context of the company’s overall financial position, count as “serious” financial loss. The December 2017 Skype messages

Webzilla Ltd relies on messages sent to Mr Dvas on 6 and 7 December 2017 by a business colleague, Isaac Douglas. He was a member of the sales team at one of servers.com’s rivals. The case pleaded in the Re-re-re-Amended Particulars of Claim is that “the Third Defendant was sent messages … indicating that the reputational damage caused by the publication complained of is continuing …” It is no longer said that the messages went to Webzilla Ltd. The messages are now said to indicate that an enterprise called PrivateInternetAccess was “not willing to do business with the XBT group” due to the publication. The reason for the change of position is that these messages were addressed to Mr Dvas in his capacity as incoming CEO at servers.com and XBT Group. It is accepted they “cannot be specifically linked” to Webzilla Ltd.

The claimants’ case at trial was that the messages show “the likely long-term effects of publication in the industry generally”. Even that broad proposition seems rather ambitious. The messages encouraged Mr Dvas to “really put some effort into repairing the damage done from the court cases and stuff that have been going on”, and stated that Mr Douglas “would have had a 300 servers deal lined up for you … but the guy couldn’t see passed that issue” (sic). Mr Douglas did not give evidence. The messages are not crystal clear about the “stuff” or the nature of the “issue” which is said to have blocked the 300-server deal. They certainly do not specify the publication of the December Memorandum. It is very hard to gauge how likely it is that any deal could have been done, but for the “issue”. In cross-examination, Mr Dvas disclosed that he had earlier spoken to Mr Douglas about the court cases, and that he probably replied to the messages (the content of one of the messages supports

The matter goes beyond that, as there is no evidence from Mr Douglas (or anyone else) that the prospective customer was based in the EU, or responding to anything published there. Mr Dvas conceded that it might have been the US or Canada or Europe.

There is a feature of Mr Douglas’s messages that serves to underscore this point. Mr Douglas was urging Mr Dvas to “really put some effort” into damage repair, “if I was you and wanting to move into Western markets….” (my emphasis). This is puzzling, and unhelpful to the claimants’ case. It may indicate that PrivateInternetAccess was a “Western” company – at least, in Mr Douglas’s estimation. But it clearly suggests that – in the estimation of a trusted business contact of Mr Dvas – the XBT Group was not yet “in” Western markets, which is an essential ingredient of its claim to have sustained serious financial loss. This would be consistent with my analysis of the position in respect of Hilltop. I note, also, that I have not been provided with any evidence about the identity of any of the other 31 EU-based customers to which Mr Dvas referred in his evidence, or the nature of their businesses. I am not clear that he possessed that information.

In all the circumstances, I could not find that Mr Douglas’s messages are evidence that the publication in the EU that is complained of caused Webzilla Ltd to lose an opportunity to sell servers to PrivateInternetAccess. And I am not persuaded that the messages are evidence to which I should attach any significant weight when assessing whether publication in the EU caused the company serious financial loss.

A third, unpleaded, matter was advanced as an “illustration” of the impact of publication. Messages on LinkedIn, from as recently as May 2020, appear to show a prospective employee making, and then pulling out of, a job application, citing “the article of BuzzFeed about the 2016 cyber-attack on the US Democratic party …”. Again, this documentation does not relate to Webzilla Ltd. The job was with

“Webzilla.com”, seemingly a trading name of servers.com. It is conceded that this is not evidence of any actual financial loss. I have only the LinkedIn exchanges (in Russian, with a translation). There is no evidence from the job applicant, and no explanation of why there is no such evidence. His name and contact details are of course available to the claimants (and in the papers). I have failed to detect any evidence that he was responding to publication in the EU. The job applicant was Russian by name, wrote in Russian, and was applying for a job as an engineer in Dnipro.

Webzilla Ltd’s inferential case

The claimants accept that proof of the damaging nature of the defamatory imputation, coupled with the extent of publication, is not of itself enough to satisfy the statutory threshold. But Mr Caldecott submits that, where those two factors are very potent (as

I can readily accept that the defamatory imputation conveyed by the words complained of is a grave one, with a seriously harmful tendency. Webzilla Ltd might have had little difficulty in establishing a claim under the common law rules, which presumed the existence of some damage. But the law is now considerably more exacting. Scrutiny of the evidence about this company’s finances, and what actually happened, leaves me unpersuaded that I should infer that EU publication caused Webzilla Ltd serious financial loss. The company’s residual claim has to be viewed in the light of its failed attempts to prove special damage, which the claimants concede are open to serious criticism. The evidence I now have about the company, its business, its customers and its finances, is limited and unsatisfactory. For the reasons I have given, the company has failed to prove its case as to each of the three specific factors relied on to support the inference of serious loss. Against this background, and in any event, I do not find the company’s broader circumstantial case persuasive.

The pleaded case is that the words complained of were published to “vast” or “very substantial” numbers of people “across the European Union” (paragraphs 6 and 8.3 of the Particulars of Claim) and that thereafter those words or “the allegation they conveyed” were further published on the internet “countless times”, and read by “further very substantial numbers of people” across the EU (paragraph 8.4).

There is no presumption of law that material which has been published online has been read in this jurisdiction, or at all. Nor is there a presumption that every reader of a newspaper or news website reads everything it contains. Whether and by how many people words complained of as libellous have been read is a matter for evidence and, where appropriate, inference: Al Amoudi v Brisard [2006] EWHC 1062 (QB) [2007] 1 WLR 113 [32-36] (Gray J). Further, although the presumption that the ordinary reader of a newspaper article reads the whole of that article is a long-established and key feature of the law of meaning, it has never been held to be an ingredient of the law of damages for defamation. As explained in Charleston v News Group, the presumption is a necessary consequence of the single meaning rule, but the reality is that

“Everybody reads selectively, scanning the headlines and turning the pages. One reader whose interest has been quickened by an eye-catching headline or picture, will pause and read an article. Another, with different interests or less time, will read the headline and pass on, leaving the article unread.”

(Lord Nicholls, 73E-F).

I have no direct evidence as to the scale of BuzzFeed’s readership in the EU (or indeed generally). I can take judicial notice of the fact that it is a reasonably wellknown news website, which publishes online and therefore internationally. I can infer from those facts, and the evidence adduced by the claimants, that the website’s readership within the EU was substantial, not trivial or insignificant. But the words complained of are not prominent in the article. On the contrary, they are (or were, prior to the redaction undertaken by BuzzFeed) to be found in a document embedded in it. I can and do infer from the evidence as a whole that a number of readers will have read the words complained of, and that the reputations of Mr Gubarev and Webzilla Ltd, will have suffered serious injury as a result of that publication, and of the “percolation” effect which is recognised in the authorities, in the EU. But the pleaded contentions are not supported by the evidence.

As to republication, my conclusion is that there was some, from which reputational harm is to be inferred, but that the pleaded case is overstated.

The most important aspect of the case for present purposes is not whether serious reputational harm was sustained, but whether it is more likely than not that the publication and/or republication within the EU of the words complained of or their gist led to Webzilla Ltd suffering financial loss of a serious nature, by causing one or more of its customers to hold back from placing additional business with the company, that would have been profitable. I do not consider the evidence justifies that inference.

The big picture does not help. The financial statements for the calendar year 2017 show a year-on-year reduction of some $2m in gross turnover, but in all the circumstances I do not consider I can place confidence in those figures as an accurate representation of the true position. The circumstances include Mr Gubarev’s own evidence that Raj was not fit to perform his role as CFO at the material time; the cloud that hangs over Raj’s integrity; the fact that the company concedes, despite this feature of the accounts, that it cannot establish a general downturn in business resulting from the publication complained of; and the striking rebound in the revenues for the year 2018, when the year-on-year revenue figures increased by $4m, and profits soared. No explanation for this has been offered. These features of the accounts make it as likely as not that the 2017 figures represent an accounting “blip”, showing a falsely negative picture.

A narrower look does not make the claimants’ task easier. The evidence is that only 10% of Webzilla Ltd’s revenue is “European”, derived from 32 customers. The company’s own evidence is that its accounting methods do not enable it to present any analysis of what happened to the revenue or profitability of its “European” business. The only “European” customer, about which I have any substantial evidence, is

Standing back from the detail, it would be naïve and unreal to suppose that these serious allegations had no impact on Webzilla Ltd, and its business. But it is for the company to establish the nature and scale of that impact, and it has failed to do so. If the company did lose out on business it would otherwise have gained, the likelihood is that such loss - or the bulk of it - was felt in respect of its business outside the EU. But the financial statements leave it quite unclear whether or not there was any impact on the bottom line. The claimants’ own case is that group companies were “quite successful” in their efforts to reassure customers who expressed concern. It has not been suggested that the allegations went unnoticed in the EU. I am satisfied that there was substantial publication and republication within the EU of the words complained of or their gist, on a scale which I can infer will have caused serious injury to the reputation of Webzilla Ltd (and, as is conceded) Mr Gubarev. But it has not been shown that this publication brought the allegations to the attention of anyone at any of the 32 customers of Webzilla Ltd who are said to have been “European”. More importantly, it has not been shown that, if that did occur, it led or is likely to lead to substantial financial loss for Webzilla Ltd. The modern law does not permit a body that trades for profit to complain of reputational injury unless it is shown that this has caused serious financial loss, or that it is likely to do so in future. The evidence here falls short.

I have outlined the background to the creation of the Steele Dossier at [7] above. Much of the more detailed history of events has become common ground, as reflected in the agreed chronology at Appendix C. The focus of attention is the period of about 8 months beginning in late May 2016, when Orbis was commissioned by Fusion to investigate the alleged links between Russia and the presumptive Presidential candidate Donald Trump, and ending with the BuzzFeed Article of 10 January 2017. To summarise some key features of the story, as they are now known:

This is another aspect of the claim that has evolved over time, in the light of disclosure and the numerous third-party depositions, none of which were available to the claimants when this action was brought.

The original Particulars of Claim, apparently settled by the claimants’ solicitors, were filed and served with the Claim Form on 3 February 2017, less than three weeks after the publication of the BuzzFeed Article. It is plain that the claimants had a very slender evidence-base at the time. Paragraph 6 contained the primary allegations, quoted at [8] above. Paragraph 8.2 put the claimants’ overall case in support of those contentions as follows:

“Pending disclosure, the Claimants freely admit that they do not know the precise identities of those to whom the Defendants originally provided the December Memorandum. However, the Defendants prepared and initially published the December Memorandum intending that its contents should be republished to the world at large; further or alternatively in circumstances such as it was reasonably foreseeable that its contents would be republished to the world at large.”

The reference to “contents” was elaborated later in the Particulars of Claim, which complained of republication “of the words complained of (further or alternatively the allegation they conveyed).”

Paragraphs 8.2.1 to 8.2.5 contained particulars in support of those broad propositions. These alleged that the general subject of the Steele Dossier was one of enormous topicality; that the memoranda were prepared by the defendants to be provided to third parties, knowing they would be used for strategic purposes, and knowing that such use would be highly likely to include making public the information contained within them; and that by 13 December 2016, (a) the defendants were aware that some or all of the Memoranda had been published within media organisations and were being discussed and (b) Mr Steele had himself given informal interviews to journalists, and discussed his desire that the contents of the Dossier should be published more widely. Most if not all of this was information or allegations that were in the public domain at the time the Particulars were served, as a result of media coverage. The claimants conceded that

“Pending disclosure it is not known whether the Defendants themselves directly provided any of the memorandums to media organisations or journalists.”

The claimants’ pleaded case on this issue has undergone two rounds of formal amendment. When the pleading was amended for the third time, on 19 February 2020, two new strands were added to the general case on responsibility for publication. In the further alternative to the case that republication was (1) intended by the defendants, or (2) a reasonably foreseeable consequence of their actions, it was said that republication was

“[3] impliedly authorised by them and/or [4] it is just and equitable that the Defendants should be held liable for the republication.”

I have added the numbering here.

Additional factual matters were pleaded in support of the overall case, as thus amended. For the most part, these stemmed from the testimony given by Mr Kramer, in December 2017 (see [16(5)] above), and Mr Bensinger, in February 2018 ([16(8)]), the records of which had by now been “unsealed”. The amendments included new allegations:

When the pleading was amended, for the fourth time, on 19 June 2020, the following was added:

“8.6

Further by widely circulating the PEM, or their gist, beyond the FBI to journalists, state department officials, retired persons of influence (such as Strobe Talbott), politicians (including the President elect’s political opponents), all of whom were likely to share the information with others, it was intended and/or foreseeable in the circumstances, as happened, that:

8.6.1

the President and President elect would have to be briefed on the dossier’s headline allegations;

8.6.2

that fact would be reported by the media; and

8.6.3

sooner or later this would lead to publication by the media of the PEM and/or their detailed allegations and with them the December memorandum and/or its detailed allegations.

8.7

In this context the Claimants also rely on the fact that the Defendants (via at least Mr Kramer) continued to communicate with the media on issues relating to the memoranda (including the December memorandum) up and until publication by BuzzFeed.”

The claimants’ case also drew in facts and matters alleged in the Reply. The claimants’ Skeleton Argument for trial contained a helpful 10-point summary of the overall case, as it then stood. It was this (internal citations omitted):

A much more detailed case has been advanced in the course of the argument and cross-examination. It is the claimants’ case that, as Mr Caldecott put it: “the devil is in the detail”. I shall come to some of that. The claimants also invite me to draw adverse inferences from the defendants’ conduct of their case, including their disclosure. Their case on this point was elaborated by Mr Caldecott in opening, in cross-examination of Mr Steele, and in closing. Its main thrust was to suggest that the defendants had deliberately suppressed a number of facts or documents unhelpful to their case, and given false accounts to assist in the cover-up. In support of this aspect of their case, the claimants rely on fine-grained analysis of a variety of written communications that have been disclosed in the course of the action, identifying what they say are significant gaps. Accompanying their Skeleton Argument, for instance, was a “message log” showing communications between Mr Steele and Bruce Ohr between July 2016 and November 2017.

The defendants submit that this is a simple case, which has been hugely overcomplicated by the claimants’ construction of what Mr Millar has called “a labyrinthine inferential case” the strands of which “are difficult to discern and follow with clarity”.

The defendants deny that they intended, foresaw, or authorised the publication complained of, or that it would be just to hold them liable for that publication. They maintain that they only disseminated copies of the December Memorandum in strict confidence, so that the information in it was known to the US and UK governments at a high level by persons with responsibility for national security; the defendants did not provide or authorise anyone to provide the December Memorandum (or any of the PEM) to the media, or intend them to be circulated to the media; they were not aware that the PEM had been published to or within media organisations; Mr Steele gave some off-the-record briefings to a small number of journalists, and spoke to representatives of the Democratic Party, but it was not suggested that the PEM or the December Memorandum should be made public.

The claimants’ allegations about the role of Mr Kramer are denied in paragraph 32 of the Re-re-re-Amended Defence, which gives this account:

The defendants describe the claimants’ case, that they are liable for what BuzzFeed chose to do, as “novel and ambitious”, making the following points, among others:

The claimants in the Aven case attempted to persuade me that Orbis was responsible for the publication by BuzzFeed of the memorandum of which those claimants complained - “Memorandum 112” - and/or for the damage caused by that publication. Reliance was placed on Mr Kramer’s deposition in the Florida Proceedings. The suggestion was that Mr Steele had put Mr Bensinger in touch with Mr Kramer when he knew or should have foreseen that this would lead to the provision of the Dossier to, and its ultimate publication by, BuzzFeed. I dismissed this contention on the twin bases that it was procedurally unsound (the case had not been pleaded, and no hearsay notice had been served in respect of the Kramer transcript), and in any event the claimants had not persuaded me that any disclosures of Memorandum 112 by Mr Kramer represented the processing of data by or on behalf of Orbis: see [55-61], [190], [198].

At [60-61] I said this:

“60.

The deposition of Mr Kramer is not a satisfactory basis for an invitation to reject Mr Steele’s evidence and find Orbis liable for disclosure and publication of Memorandum 112 made by others. Besides the procedural shortcomings I have identified, the deposition is provided to me shorn of its context. I am told nothing else about the Gubarev v Buzzfeed litigation, and very little about Mr Kramer except that (as is obvious) he had a clear motive for tailoring his evidence. In any event, knowledge that a person intends to make a disclosure is not enough to bring home liability. And the substance of Mr Kramer’s evidence, so far as Buzzfeed is concerned, is this. Mr Steele asked him to meet Mr Bensinger, but without asking him to provide a copy of the Dossier; Mr Kramer did not provide Mr Bensinger with a copy, but left him in a room with the memos for 20-30 minutes, on the agreed basis that Mr Bensinger would use the time to read them; in that period, Mr Bensinger took photos of the documents, without Mr Kramer’s knowledge or consent; and Mr Kramer only found out about this when he saw the Buzzfeed Article, and did not intend the Dossier to be published. [Counsel for the claimants], having effectively called Mr Kramer as his witness, could not and did not question this account. It undermines the case he sought to advance.

61.

On the basis of this evidence, I see no room for concluding that Mr Kramer made a disclosure to the Washington Post or Buzzfeed of the personal data contained in Memorandum 112 which amounted to processing of those data by or on behalf of Orbis, still less that the publication of those data by the Washington Post and Buzzfeed represented, or even resulted from, processing by or on behalf of Orbis….”

My findings in the Aven case do not bind the claimants in this action. Indeed, they have no legal relevance. As Mr Caldecott points out, these were conclusions drawn between different parties, in a different case, on evidence that was different and more limited (he calls it “exiguous”), and where the legal issue arose under the data protection legislation. Although there are evidential overlaps – for instance, Mr Kramer’s deposition is relied on by the claimants in this case - my conclusions in the present case must be based on the totality of the fuller evidence and argument presented to me at this trial. The defendants have rightly approached the matter on that footing.

The claimants’ case is that the author(s) of a document (the December Memorandum) are liable for the defamation consequent on the republication of that document by others (BuzzFeed Inc). It is clear law that where a defamatory document written by the defendant is voluntarily republished by someone else the defendant is liable as a joint wrongdoer if he (1) intended the republication or (2) authorised it: see the decision of the Court of Appeal in Speight v Gosnay (1890) 60 LJQB 231, Gatley¶6.52, and the more recent decision of Nicklin J in Turley v UNITE the Union [2019] EWHC 3547 (QB) [84-89]. See also Watts v Times Newspapers Ltd [1997] QB

There is no need to elaborate on what is meant by intention. Authority to republish may be given expressly or by implication (see the examples at Gatley ¶6.53). The case which the claimants added by amendment in February 2020 is one of implied authority: see [94] above. The matter of republication is approached as one of substance, not form. Mr Caldecott points out that a defendant who intends or authorises it may be liable for the republication of the gist or sting of what he said or wrote; liability is not limited to cases where the defendant’s words are reproduced verbatim: see Gatley ¶6.54 and Al Refai v Dar Al Arkan Real Estate Development Co [2013] EWHC 1630 (Comm) [34] (Andrew Smith J). But there is a distinction between, on the one hand, authorising a third party to perform an act and, on the other, merely facilitating that act, or creating conditions in which that third-party act becomes more likely, or reasonably foreseeable.

A third basis for holding a defendant liable for another’s republication of his statement has been pleaded: that the defendant is liable if the republication was a reasonably foreseeable consequence of the original publication. In opening, Mr Caldecott advanced a different contention: that a defendant is liable for republication which is a “natural consequence” of publication by the defendant. He cited the judgment of Lopes LJ in Speight v Gosnay, where those words are used, and told me that the claimants would invite me to apply this principle, subject to a “more refined analysis”. In his closing submissions, that invitation was withdrawn. Rightly so, in my view.

Mr Millar cited a number of additional authorities, and identified the potential chilling effect on free speech which might result from a laxer regime as another reason to hold fast to these principles in the media law context; but the stance ultimately adopted by the claimants means it is unnecessary for me to address these further arguments.

I do, however, make clear that I accept the defendants’ submission, that aspects of the way the claimants’ case on this issue was opened are, in the final analysis, unsustainable. It was argued (in paragraph 33 of the Skeleton Argument) that “on the third basis” the Court is “required to consider” the nexus between the original publication and the republication, and such matters as (a) whether the republication would not have occurred but for the original publication (the ‘but for’ test); (b) whether the intervening actor was a stranger to the defendant and, if not, the nature of their relationship where relevant; (c) whether the republication (of the relevant defamatory allegations) was reasonably foreseeable by the defendant; and (d) the objective likelihood of such a republication occurring. These are matters of causation and remoteness of damage, not touchstones of liability. At best, some of them may

The fourth pleaded basis for holding the defendants liable for the BuzzFeed republication was not pursued. No argument has been presented that liability should be imposed on the basis that it would be “just and equitable” to do so.

The claimants’ case inescapably involves a frontal assault on the honesty and integrity of the defendants, and the veracity of Mr Steele. These are serious allegations, so it is necessary to recall that although the standard of proof remains the ordinary civil standard – the balance of probability – the claimants must present evidence commensurate with the gravity of the misconduct that is alleged against the defendants. More precisely:

“When assessing the probabilities the court will have in mind as a factor, to whatever extent is appropriate in the particular case, that the more serious the allegation the less likely it is that the event occurred and, hence, the stronger should be the evidence before the court concludes that the allegation is established on the balance of probability. Fraud is usually less likely than negligence. …

Although the result is much the same, this does not mean that where a serious allegation is in issue the standard of proof required is higher.”

Re H (Minors) (Sexual Abuse: standard of proof) [1996] AC 563, 586D-F (Lord

The claimants’ case of wilful suppression of documents must also be looked at in the light of Re H. As Mr Caldecott put it in closing, there is a “relatively high standard” to be satisfied in that respect. In relation to this part of the claimants’ case, I also remind myself of the Lucas direction given in criminal cases. I must ask myself if the claimants have established to my satisfaction (to the civil standard) that the defendants lied in one or more respects. If I find that the defendants have lied, I must beware of leaping to the conclusion that this is because they are guilty of what is alleged against them. There can be “innocent” reasons for lying: reasons that do not stem from a wish to conceal guilt.

The claimants have called no witness who can speak from personal knowledge of what took place in 2016-2017. They rely instead on some, relatively limited, contemporaneous documentation, and a substantial number of documentary records of what others have said, later on, about what happened, in books, articles, and testimony for proceedings in Court and to various political enquiries (see [15-18] above). Mr Caldecott fairly stresses the importance of contemporaneous records. In general, a Court will place particular weight on such documents, which tend to be a more reliable guide to the truth than oral evidence which conflicts or is hard to reconcile with the records: see the discussion in R (Dutta) v General Medical Council [2020] EWHC 1974 (Admin) [40].

When it comes to hearsay, the position is different. I need to bear in mind a number of general principles:

Section 4 of the 1995 Act contains a rule of law, but its application will turn on the facts of the individual case. So will that of the other two principles I have mentioned, which are rules of thumb, not principles of law. As Dyson LJ made clear in Welsh v Stokes [2007] EWCA Civ 796 [2008] 1 WLR 1224 [23]: “Where a case depends entirely on hearsay evidence, the court will be particularly careful before concluding that it can be given any weight”; but it is permissible to advance a case based entirely on uncorroborated hearsay evidence, and there is no rule against a Court giving weight to such material. As Baroness Hale observed in Polanski v Conde Nast Publications Ltd [2005] UKHL 10 [2005] 1 WLR 637 [74] “The court is to be trusted to give the statement such weight as it is worth in all the circumstances of the case.”

Among the factors that bear on the reliability of the hearsay evidence in this case are the following:

There are some hearsay statements which find support in, or are at least consistent with, the objective facts established by other, reliable evidence. One illustration, offered by Mr Caldecott, concerns the book “Russian Roulette” (2018), by Michael Isikoff (of Yahoo News) and David Corn (of Mother Jones). The passage relied on asserts that in September 2016 Mr Steele orally briefed Mr Isikoff, in detail, about the alleged secret meetings between Mr Carter Page and high-ranking Russian officials to discuss a deal for the lifting of sanctions. This hearsay is supported, says Mr

As a rule, a party who adduces evidence from a witness is bound by what they say.

In my judgment, having assessed the relevant evidence as a whole, it has not been shown that that Mr Steele authorised the publication of the words complained of, or words to that effect, by BuzzFeed, or that he intended that those words, or the gist of them, should be published by BuzzFeed or to the public at large.

He is the only person who was “in the room where it happened” - or did not happen - and who appeared before me as a witness. He has been consistent in his position, that he did not authorise the publication of the December Memorandum to the general public, or intend that such publication should take place. The account given in his witness statement is reasonably detailed, and inherently worthy of belief. That is to say, it does not lack credibility as a version of events. No allegations of general bad character have been made. Sir Andrew Wood has given evidence of Mr Steele’s positive good character as a “serious and responsible intelligence professional and an individual of the utmost honesty and integrity”, who had a “positive professional reputation”. There is also evidence of Mr Steele’s good character in Mr Kramer’s Florida deposition (at p88 ll.18- 25, p115 l.22 – 116 l.19). He is entitled, as a starting point, to be treated as someone of good character. He gave his evidence calmly and with a degree of confidence, withstanding sustained challenge in cross-examination without becoming flustered or confused.

That said, he is an intelligence professional whom one would expect to be capable of dealing with challenge. I must, in any event, guard against basing my decision on an assessment of a witness’s demeanour, or presentation. The evidence must be assessed as a whole, comparing the witness’s account with other versions of events, contemporary documents, and what the evidence as a whole suggests about the inherent probabilities. I need to consider the detailed criticisms of the way the defence case has been conducted, and what inferences I should draw.

The main features of the claimants’ factual case are summarised in the ten propositions I have set out at [97] above. But there is more to it, and some general observations can be made.

In the light of what I have just said, it is as well to start with this. As Mr Caldecott acknowledged in opening, “Ultimately the case turns on the December Memorandum.”

Mr Steele. The account given in his witness statement is, in summary, as follows:-

Cross-examined, Mr Steele:

Sir Andrew Wood His witness statement confirmed the account of his role provided by Mr Steele. He had believed it was clear, in his discussions with Mr Kramer and Senator McCain in Halifax, and in his later communications with Mr Kramer, that the PEM were not for publication. He had considered Mr Kramer to be honourable. He also thought very highly of Senator McCain, whom he described in cross-examination as “the very soul of truthfulness, honour and courage”. So far as the December Memorandum is concerned, Sir Andrew said that he and Mr Steele did not discuss it, and he was not aware of its existence until after these proceedings were commenced. He had not expected any of the PEM to be published, and understood this to be contrary to the wishes and interests of Mr Steele and Orbis. This evidence was not challenged.

Mr Kramer The account he gave in the Florida Proceedings, so far as relevant, includes the following:

The account given in Mr Kramer’s interview to the Russia Committee hearing, 6 days after this deposition, was consistent with the above. It covered other ground:

Mr Bensinger In his deposition and declaration in the Florida Proceedings, he gave accounts of his receipt of the Dossier from Mr Kramer. He made the following relevant assertions:

Mr Smith His declaration in the Florida Proceedings corroborates Mr Bensinger’s account of how he came to contact Mr Kramer. Mr Smith confirms that it was on 23

Observations This evidence makes clear that the source from whom BuzzFeed obtained the Dossier was Mr Kramer. But the evidence could not sustain a finding that Mr Kramer intended or authorised the publication by BuzzFeed of the Dossier, or any words quoted from it. The claimants’ own evidence is to the contrary. Mr Kramer did not intend or authorise such a thing, and was shocked when it happened. On Mr Kramer’s evidence, any notion Mr Bensinger may have entertained, that he had been authorised to take photographs of the Dossier, or to quote from it, was mistaken. Photography was expressly prohibited. Mr Steele’s evidence of his own shock at the BuzzFeed publication is not only supported by Mr Kramer’s account it is also (as I shall explain) corroborated by contemporaneous records, and is not challenged. It cannot be, and is not, suggested that he intended or authorised the publication that in fact took place. The primary version of the original pleaded case fails.

That much has been accepted by the claimants. Their case now is limited to the proposition that Mr Steele and Orbis are, nonetheless, liable for BuzzFeed’s publication of the defamatory imputations about them, because they intended or authorised the publication of something to that effect – an argument which might be labelled “reverse Al-Refai”. This case was pleaded as an alternative from the outset, but it has evolved in a number of ways. The evidence I have summarised poses a number of difficulties for this line of argument. The following points stand out.

Several documents are relied on by one side or the other in respect of the December Memorandum. I shall address the more important ones, in chronological order.

Pre-publication documents.

Post-publication documents.

In addition to what the documents say, I attribute importance to what they do not say. The Kramer/Steele messages do not, for instance, contain anything from Mr Kramer after his meeting with Mr Bensinger saying, “I have done as you asked”. In January 2017, Mr Kramer says a good deal about other journalists and their enquiries, but neither he nor Mr Steele say anything at all about that meeting. Nor is there anything in the Bensinger/Steele messages of early January to indicate that information as to the contents of the Dossier was discussed when they met. The Kramer/Steele messages exchanged privately after the BuzzFeed publication contain no hint that they were satisfied with the publication of the allegations. There is nothing on the lines, for instance, that “Bensinger went too far”.

These topics consumed a sizeable portion of Mr Caldecott’s opening speech, a good deal of his cross-examination of Mr Steele, and also his closing submissions. I have considered these with care. I am left unpersuaded that I should draw any adverse inferences against Mr Steele, for any of the reasons advanced. Particular attention was focussed on allegations that the Defence and/or the defendants’ disclosure involved misrepresentation or suppression in respect of (a) the Bensinger/Kramer dealings; (b) the fact of the Bensinger/Steele meeting; and (c) the existence and content of the Bensinger/Steele messages.

As for the Defence, it is necessary to beware overly sophisticated textual analysis. Such documents are the product of many hands, and when complex matters are at issue apparent flaws can creep in. I do not accept the claimants’ submission that paragraph 32(d) of the Re-re-re-Amended Defence has been shown to be deliberately false or misleading. I am not persuaded that statements in the Defence that Mr Steele was unaware that Mr Kramer had been contacted by journalists “about the ‘dossier’” or that he “did not instruct, encourage or permit Mr Kramer to discuss the [PEM] or the December Memorandum with Mr Bensinger” are incriminating lies. Nor do I accept that there was an onus on the defendants to disclose the Steele/Bensinger meeting in this paragraph of the Defence. The allegation the defendants were addressing was about contact between Mr Bensinger and Mr Kramer, not Mr Steele. It would be unfair to draw an adverse inference from a failure to address an unpleaded allegation. As Mr Caldecott submitted in opening “It is central to our system that a party’s case on the primary facts is set out in the statements of case… [which] are the starting point.”

Turning to disclosure, a similar caution needs to be applied. Flaws in disclosure are common, but it is relatively rare for Court to be able to conclude that they are deliberate, and incriminating. The claimants’ criticisms in this case fall to be considered against a truly exceptional background. That is apparent from the statement of Mr Mathieson, the partner in charge of the defendants’ case since midFebruary 2020. He says this: “the partner at RPC with responsibility for this case until mid-February 2020 actively, seriously and repeatedly misled the Defendants about these proceedings”. That summary is amply borne out by the more detailed evidence of the partner’s misconduct. It is unnecessary to set it out in full here, but important to identify that it included, though it was not limited to: (a) withholding the defendants’ disclosure from the claimants for many months, in breach of a court order, and without telling the defendants; (b) failing to show the defendants their own disclosure statement and list dated 29 July 2019; (c) signing that document herself and serving it on the claimants on 12 December 2019, without showing it to the defendants. The evidence of Mr Mathieson also discloses that, when Mr Steele first learned of the partner’s misconduct, he was shocked and dismayed. The specific criticisms advanced relate to events after the replacement of the partner concerned, but the context cannot sensibly be ignored, as will become clear.

Some of the claimants’ criticisms are, in my judgment, lightweight. It was suggested, for instance, that there was something suspicious about the fact that the defendants disclosed some messages by way of screenshots taken on Mr Steele’s phone, and the screenshots omitted parts of some of the messages. But the fact that they were incomplete was plain and obvious on the face of the disclosure, and the omissions tended to favour rather than harm the defence case. This was manifestly sloppiness,

The main criticism is that there was wilful suppression of the fact of the Steele/Bensinger meeting. It arises from a single sentence in a long letter from RPC, sent about 3 weeks after Mr Steele learned of the partner’s wrongdoing. This stated that apart from some specified face-to-face meetings Mr Steele did not conduct communications with “Schedule A/B individuals”, of whom Mr Bensinger was one. This was of course incorrect, as Mr Steele readily accepted when cross-examined. I can easily see why the claimants advanced the argument that this was a deliberate falsehood. But in the end that contention rests largely on Mr Bensinger’s message of 14 January 2017. Its existence was later disclosed by the defendants. More pertinently, I am told by Mr Millar on instructions (and Mr Caldecott made clear he would accept such an account) that Mr Steele had disclosed the existence of the Bensinger messages to the RPC partner by 19 September 2017 at the latest, and that he had told her of the London meeting by 7 February 2018. All of this was a long time before the defendants’ disclosure statement and list. That, as I have mentioned, was created in July 2019 and served in December 2019. The conduct of disclosure in this respect was clearly deficient, but I am not persuaded that the statement in the letter or any other statement in this respect was, or reflected, wilful misrepresentation by Mr Steele.

The defendants invite me to conclude that the claimants’ case is inherently improbable. They had nothing to gain, and a great deal to lose, by causing or authorising media publication of detailed allegations from the Dossier. I see some force in this. The publication of the Dossier, or detailed allegations from it, was liable to undermine trust in Orbis’ business, which depended to a significant extent on its ability to maintain confidentiality. There was also a risk of litigation, as has resulted. The suggestion that Mr Steele was so keen to undermine trust in Mr Trump that he was willing to take such risks is unconvincing. It is hard to see what else the defendants could have had to gain.

There is certainly a wealth of evidence that Mr Steele had, before the election, been keen to engage media interest in the fact of the allegations in the Dossier by conducting briefings. That is not seriously in dispute. But despite the detailed care and attention devoted to this aspect of the story on behalf of the claimants, I do not consider that it lends material assistance to their inferential case. It certainly does not outweigh the factors I have mentioned already, that count against that case.

Three general points can be made:

Turning to the ten-point summary, in so far as I have not already dealt with it: (1) The allegation that Mr Steele briefed the media on the detailed allegations in the PEM is overstated at best. I do not accept that he briefed the detailed allegations in the December Memorandum “via Mr Kramer”. Mr Kramer’s own evidence, in his deposition, is that he was not asked to do this. (2) Nor do I accept the claimants’ case that Mr Steele knew from September 2016 that the media had published detailed allegations from his briefings without investigation or (put another way) that his briefings “were liable to be reported without further investigation”. This aspect of the claimants’ case focused in the end on three publications in the Autumn of 2016. I take them in date order, and will deal at the same time with allegation no (3).

Returning to the ten-point list, allegation no (4) concerns the Kramer/Bensinger meeting, which I have disposed of above. (5) I do not accept that Mr Steele’s conduct in sharing the PEM with Strobe Talbott, and other senior officials before the election, affords a reasonable basis for concluding that, in December or in January 2017, he intended or authorised media publication of the allegations of which these claimants complain. (6) Briefing the media, in 2016, that the FBI was investigating the Dossier would be at best a weak basis for the inference the claimants invite. But I accept Mr Steele’s evidence that he did not brief the media to that effect. The hearsay evidence relied on is not enough to justify rejection of his evidence. (7) The fact that some of the allegations were not capable of verification by the media is not disputed, but does not lend any material support to the claimants’ case. (8) Nor does any knowledge that allegations “were circulating”, or that the Democrats were calling for publication. This is a long way from demonstrating that he foresaw, still less authorised or intended, media publication of the detailed contents of any of the memoranda. (9) Is a point about foreseeability which cannot bear any real weight. (10) The lack of a confidentiality warning is a point of no consequence. It was explained by Mr Steele as an omission, but these were not documents designed for dissemination to ingenues, and the confidential nature of their content was self-evident.

For all these reasons, I reject the claimants’ case that the publication by BuzzFeed of the defamatory allegations conveyed by the Memorandum was something intended or impliedly authorised by the defendants. With the modest qualifications I have indicated, I accept Mr Steele’s evidence. The essential facts as I find them are as follows.

As for the additional matters relied on to support the inferential case, these do not assist.

In the light of the conclusions just stated, this issue does not arise. It would not have arisen anyway, so far as Webzilla Ltd is concerned. If the claimants had established the defendants’ responsibility for the BuzzFeed publication, Mr Gubarev would have recovered a substantial award, but I need not examine what it would have been, or whether any and if so what injunctions would have been appropriate.

The words complained of, in their context, meant that there were good reasons to suspect the claimants of having, under duress from the Russian Secret Service, taken part in hacking the computers used by the Democratic Party leadership, and using the access they unlawfully gained in that way to transmit virus, plant bugs, steal data and alter files and software. That imputation is defamatory of Mr Gubarev at common law, and its publication in this jurisdiction and the EU caused serious harm to his reputation. No substantive defence has been advanced. He would have been entitled to substantial damages, if he had proved that the defendants are responsible in law for the publication complained of. But he has failed to prove that. So, Mr Gubarev’s claim must be dismissed. Webzilla Ltd’s claim must also be dismissed, for the same reason. That claim would have failed in any event because, although the words complained of are clearly defamatory of the company at common law, it has failed to establish that their publication caused it serious financial loss. Its claim therefore falls short of the statutory threshold set by s 1 of the Defamation Act 2013.

Agreed Chronology