There has been previous comment here on the various stages of Banks v Cadwalladr, and the facts have been well-stated. However, there has so far been no comment on the substantive decision of the Court of Appeal in the case. This post aims to fill the gap. There is a lot going on in the Court of Appeal’s decision. It brings some clarity to novel issues like the judicial definition of SLAPPs, the meaning of ‘echo chamber’ for the purpose of s 1 of the 2013 Act, and the need to reassess serious harm once the public interest defence falls away.

It also contains some interesting ambiguities on the role of inference of harm in cases involving grave imputations with scope of publication, and the relevance to the serious harm test of factors which cause the public interest defence to fall away. It is upon those ambiguities which this comment will focus.

The role of inference in the serious harm test

The Court of Appeal essentially overruled Steyn J’s decision on three grounds. First, it was decided that the trial judge made an error of principle in inferring that harm was diminished because the statement was published to the defendant’s own ‘echo chamber’. The Court adopted a more literal definition of ‘echo chamber’ [57], and cited (at [59]) the rule in Associated Newspapers Ltd v Dingle [1964] AC 371 that evidence of harm to the claimant’s reputation depends on the testimony of witnesses who know the claimant and can testify to the relevant sector of their reputation, and its continued relevancy to the serious harm test according to the Supreme Court in Lachaux.

Second, the trial judge was also held to have made an error of principle in inferring that harm to the claimant’s reputation was diminished because the publication was made to people ‘of no consequence’ to him [64]. Any previous reference to publishees ‘of no consequence’ applied only to the specific and very different facts of Ames, with its limited publication in the jurisdiction in an application to strike out which was nonetheless denied.

The Court of Appeal’s ruling in that regard is relatively uncontroversial. However, they also overruled the trial judge on another basis, although the reasoning here is less clear. Along with the defendant’s own ‘echo chamber’ and publishees of ‘no consequence’, the trial judge (at [97]) also based her decision on an inference that the online publication would have ‘peaked’ by the time that the public interest defence fell away and the cause of action arose:

‘The normal inference, in the absence of reasons to think otherwise, would be that views of the TED Talk will have peaked at or shortly after the time of first publication, and certainly within the first year of publication. The numbers given in paragraph 66 above [summary of claimant’s evidence regarding the scale of publication] show there is no reason to think otherwise in this case. In addition, the fact that the TED Talk is on the subject of the EU Referendum, which took place nearly six years ago, increases the likelihood that the extent of publication will continue to diminish.’

The Court of Appeal does not explicitly denote this an error of principle, but nonetheless overrules it on the basis of an ‘inevitable inference’ that the ‘gravity of the allegation’ and the ‘scale of publication’ in the relevant period caused serious harm to the claimant’s reputation. While the allegation in Banks may have been ‘not at the most serious end of the scale’ (according to Steyn J in the High Court [61]), for the Court of Appeal it was a serious enough one of ‘repeated dishonesty about a secret relationship with a potentially hostile foreign power.’ In terms of the scale of publication, the Court accepted the summary of the claimant’s evidence of 100,000 viewings in the relevant period, and considered this ‘broadly equivalent to circulation of a broadsheet national newspaper.’[68]

It is this broad equation of the online publication in Banks with the ‘circulation of a broadsheet national newspaper’ which invites a problematic uncertainty into the judgment. The courts have previously recognised the specific context of online communication as a ‘fast moving medium’, from which publications ‘disappear from view quite swiftly’, and come to be ‘buried’ in a greater complexity, tempo and scale of content. Much of that relates to social media, of course. Banks involved publication on the TEDTalk website and on YouTube. Perhaps they have a bit more staying power and gravitas in the minds of publishees, but they must still be examined in the specific context of online communication marked by greater ‘content pollution’, and a more fluid ‘attention economy’. There is no doubt in this regard that a year-old publication of a video online is contextually distinct from the daily issue of a broadsheet national newspaper, freshly committed to print and on the newsstands.

This is not to favour Steyn J’s inference of non-harm over the Court of Appeal’s inference of serious harm to the claimant’s reputation. Even taking on board repeated viewings and the likely ‘peaking’ of the online publication by the time the cause of action arose, the statement complained of may still have caused a significant number of people to think less of the claimant. However, what should be favoured is the more contextual approach to the serious harm test. In Lachaux, Lord Sumption held that ‘[w]hile a court may be willing, where the imputation is a particularly grave one, to draw an inference of serious harm without the need for any additional evidence, it remains the case that it is for the claimant to prove on the balance of probabilities that serious harm to their reputation was caused or was likely.’ [14] In Economou, the Court of Appeal cautioned that, even if an inference of serious harm can be drawn in an appropriate case, there must still be a ‘causation analysis’, and that ultimately the judge has to be ‘satisfied that particular publication concerned had caused the claimant serious harm’ [41].

There was always a danger that, with its neglect of the special context of the online publication and somewhat understated causation analysis on serious harm, the Court of Appeal’s decision would be interpreted by some as permitting a regression to a common law standard of ‘inherent tendency of harm’. In a recent case in the High Court, the claimant sought to rely on the Court of Appeal’s decision in Banks to assert such a common law standard in satisfaction of s 1 of the 2013 Act. Saini J had to correct that interpretation (at [26]), underlining that it is ‘important not to dilute the “serious harm” element of the cause of action using the inference of harm argument to such a degree that it becomes so easy to satisfy the requirement that it becomes meaningless.’ In another recent case, Nicklin J (at [23]) interpreted the key point of the Court of Appeal’s decision in Banks, so far as it relates to the serious harm test, to be that the statement is ‘not defamatory unless’ the claimant has proved as a matter of fact that the publication has caused, or is likely to cause, serious harm to reputation, and that ‘inherent qualities’ alone do not suffice for this (even if the Court only emphasised this in relation to the issue of the assessment of serious harm once the public interest defence fell away (at [42])).

This should be the enduring interpretation of the Court of Appeal’s decision in Banks. The decision does not overrule the established requirement that an inference of serious harm to reputation for the purpose of s 1 must be supported by a contextual analysis of the specific facts of the case. 

The serious harm test and the anomaly of the joint statement

The late cause of action in Banks also raised another more perplexing anomaly in relation to the relevancy to the serious harm test of the statement of the Electoral Commission of 29 April 2020 (the ‘joint statement’), which publicly stated that there was no evidence to suggest the claimant had breached electoral funding law. Neither the Hight Court or Court of Appeal considered this relevant to the assessment of serious harm once the public interest defence fell away, but on reflection it is hard to see why even this should be completely excluded from the contextual analysis.

The joint statement also enjoyed some scope of publication and carried some gravity of imputation. It represented the conclusion of government agencies tasked with investigating the very substance of the defamatory allegation, and a categorical statement that they had found no evidence that the claimant had committed any of the offences alleged by the defendant. The natural inference there is that it would have undermined the veracity of the statement complained of in the minds of the relevant publishees, and therefore mitigated, at least to some degree, any harm to the claimant’s reputation.

Thus, we have the curious situation where the factor which caused the public interest defence to fall away and the cause of action to arise also made serious harm to the claimant’s reputation less likely. In some ways the case reflects the problem posed by Parkes QC in Hodges v Naish. There (at [165]), the judge imagined a scenario where ‘X tells an assembly of people at a meeting that he has evidence that Y has murdered Z, his wife, who has been missing. As they leave the meeting, shocked and appalled, they see Z walking down the street.’ Is the point to assess serious harm the point of publication? What, he asks, if Z were to ‘reappear a day later, or a month, or a year. Would that make a difference, and if so, why?’.

In the Court of Appeal judgment in Lachaux, Davis LJ rejected the idea of an ‘ambient cause of action, drifting in and out of actionability’ [60]. The unique facts of Banks nonetheless made that a necessity. Once the public interest defence fell away, it was necessary to reassess serious harm. However, the fact which triggered the failure of the public interest defence under s 4(2)(b) also made it less likely, to some degree at least, that the defendant’s statement caused serious harm to the claimant’s reputation.

Why then did the joint statement not figure in the assessment of serious harm to the claimant’s reputation in the relevant period? Why did it not offer any ‘counterweight’ to the ‘natural inference’ that there was serious harm in phase two? The admission of such evidence does not appear to violate the letter of the rule in Dingle. That excludes reliance on other publications which may have already tarnished the claimant’s reputation. However, perhaps it offends the spirit of the rule? That is, if the defendant is able to rely on other publications to negate an inference of serious harm to the claimant’s reputation, she should be able to point to evidence that such publications have in fact mitigated harm to the claimant in the relevant sector of their reputation. The only problem with this approach is that, in practice, it would be even harder for a defendant to prove that as it would be for a claimant to prove serious harm to reputation—and as interpreted by the Supreme Court, the burden of proof should fall on the claimant in relation to the serious harm threshold under s 1.

All things considered, there is no easy answer here. Although this specific issue may only present in relatively few cases, it is clear from Banks and other cases now that the general issue of burden of proof in relation to serious harm under s 1 is more complex than the drafters envisaged and that it still requires some fine tuning by the courts. To some extent it is understandable that the Court of Appeal did not engage with this issue in Banks. It did not come up in the trial judgment, and it was not argued by counsel. And even if it did make harm to the claimant’s reputation in the relevant period less likely, on evidence such harm may still have been likely enough. But again, it was another missed opportunity to address through principled reasoning the persisting complexity about the balance of rights between the parties in defamation claims.

Dr Mark Hanna is Assistant Professor in Media Law at Durham University. This comment is based on a forthcoming article in the Journal of Media Law.