Senedd Cymru | Welsh Parliament
Pwyllgor y Bil Atebolrwydd Aelodau | Member Accountability Bill Committee
Bil Senedd Cymru (Atebolrwydd Aelodau ac Etholiadau) | Senedd Cymru (Member Accountability and Elections) Bill
Ymateb gan Jeremy Horder, London School of Economics | Evidence from Jeremy Horder, London School of Economics
I have sought to answer the questions asked in the letter of invitation to give evidence. The answers follow the questions below.
The first point to make about the current draft of 2(A) is that it ties the hands of Welsh Ministers: requiring them not merely to address the issue of false or misleading electoral statements but to make provision for prohibiting such statements.
Clearly, legal steps must be taken in relation to false and misleading electoral statements. However, the proposed phrasing runs counter to the ‘permissive’ character of section 13 of the 2006 Act, which generally empowers Welsh Ministers in relation to the conduct of elections but does not require them to take any particular step, or to take a step in any particular direction.
My initial suggestion – which I hope is not based on any constitutional misunderstanding on my part - is some of this flexibility is preserved.
In an ideal world, 2A would simply empower Welsh Ministers to take such steps under section 1(a) as are reasonable and proportionate to discourage and prevent the making or publishing of false or misleading statements of fact before or during an election etc.
In that regard, it does not make sense to limit the duty of Welsh Ministers to enacting ‘prohibitions.’
For example, it might be as or more effective for Welsh Ministers to have the power to issue codes of conduct, to pass (non-prohibitory) regulations, to set up election disinformation units to counter disinformation, and so on.
A second point concerns the effect or impact of prohibitions of this kind. Almost all jurisdictions world-wide have some kind of prohibition on false election statements. However, even in authoritarian states, prosecutions tend to be few in number. There is also no evidence that prohibitions on the making or publishing of false statements are effective in preventing systematic online propagation of falsehoods through the use of bots, multiple (fake) social media accounts, and so on, especially when such malign campaigns are orchestrated from outside the jurisdiction.
Ironically, there is also some evidence that a prohibition on false electoral claims is itself capable of generating new species of false claims. This can happen when the false claim law is itself weaponised to smear an opponent. For example, politician X may be keen to discourage attention being drawn to some fact about him. So, he uses his army of supporters to create a wave of well-publicised complaints to the police/electoral authority that a ‘false’ claim about him is being circulated by his opponents. The matter is not resolved before the election, and the complaints are dropped thereafter.
Moreover, there is a significant risk that an over-broad false statement law will have an unduly chilling effect on the speech of conscientious people and media outlets fearful of transgressing the law, whilst failing to do much to deter those who see in the use of such laws against them an opportunity to gain public prominence and possibly political martyrdom.
Careful thought will also have to be given to the role of law enforcement in this process. There have been concerns since the 19th century about the politicisation of law enforcement bodies when such bodies are asked to police and adjudicate on political questions. One only has to consider the controversy that continues to be generated over the policing of ‘hate crime’ in the UK to appreciate that there are further risks of politicisation of law-enforcement bodies from an over-broad false statement crime.
A third point concerns the intended reach of the offence. In one way, 2A seems too narrow: the focus is seemingly solely on electoral impact on a candidate’s return; but it would surely make sense to extend the scope of any prohibition to the result of any official referendum (this is the position in England).
In another way, 2A risks being too broad, putting aside for now the question of the place of ‘misleading’ statements in any prohibition.
(a) So far as statements about individual candidates are concerned, English law, in common with some other jurisdictions, confines false electoral statement crimes to false or misleading statements that are about a candidate personally (personally defamatory statements), excluding political claims about the candidate (politically defamatory statements). Whilst the political-personal statement distinction is not always entirely clear – e.g. a politician can make their personal life central to their political campaign – the distinction has served the law well for 140 years.
In essence, whilst English law (Representation of the People Act 1983, section 106) has tried to protect election candidates from personal attack through the propagation of falsehood, it has taken the view that political falsehoods about candidates must be dealt with through a so-called ‘’more speech’ solution. In other words, the answer to political falsehoods is robust denial and reliance on counter-evidence. It must be kept in mind that it is typically much easier to deal swiftly and immediately with a false claim by countering it publicly online and in the media, than it is to take action though the police/courts. That matters, in what is commonly a fast-moving information environment at election (or referendum) time.
(b) English law (Representation of the People Act 1983, section 114A) also prohibits the use of fraud when someone does any act designed to deceive a person in relation to the administration of an election, ‘for the purpose of— (a)inducing or compelling a person to vote in a particular way or to refrain from voting, or (b) otherwise impeding or preventing the free exercise of the franchise of an elector or of a proxy for an elector.’
Section 114A is somewhat limited in that it focuses on the impact of a fraud on the right to vote, excluding frauds about election officials. However, the key important limitation here is that the fraud must relate to the ‘administration of an election’: what I call in my book ‘electoral process fraud’. Section 114 does not extend to frauds in relation to a candidate’s policies or personal conduct and views: what I call ‘political viewpoint fraud’.
The point is that it is right – indeed essential – to prevent people (through the use of the criminal law if need be) being deceived about their right to vote, about the feasibility or consequences of voting, or about the identity of candidates and which party (if any) they represent: electoral process frauds. By contrast, a false claim that (say) a party or politician has broken a manifesto commitment - a political viewpoint fraud - should be met with denial or counter-evidence only, not the use of criminal or regulatory prohibitions.
In short, on this third point, criminal prohibitions should be confined to false (or misleading) electoral claims about candidates that are personally defamatory in nature, and to false (or misleading) claims about elections or referendums that relate to the administration of the process and thereby imperil the right to vote and the free exercise of the franchise.
-------------------------------
In itself, ‘statement of fact’ is an ordinary English phrase and so should not be defined in law. However, it would be helpful, being mindful of obligations to respect free speech, to make it clear that a statement of opinion is not to be regarded as a statement of fact, even when – as is commonly the case – the opinion relies on underlying factual assumptions. A reliance, in expressing an opinion, on underlying factual assumptions is not as such the making of a ‘statement’ of fact.
--------------------
This is a very difficult question admitting of no easy answer and opinions will differ. In short, I would support the inclusion of misleading statements, as prohibited alongside false ones, when the statement concerns the administration of an election (the right to vote, or the integrity etc of election officials). However, it going too far to prohibit purely misleading statements when these are made about political or personal matters relating to candidates.
Currently the law is focused only on false statements (excluding misleading ones) in relation to statements about the personal character of candidates, but it focuses more broadly on fraud in relation to election administration. So far as the latter is concerned, ‘fraud’ encompasses (alongside false statements) statements that the maker knows are, or realises may be, misleading.
Any extension in the Bill that criminalises misleading – as well as false - statements about candidates themselves (or their politics) will enter into new and difficult territory, not least because whether a statement is adjudged to be ‘misleading’ necessarily involves a value judgement and not just empirical evidence, in relation to what may be highly contested issues.
The challenge posed by a prohibition on misleading personal or political statements at election time is that such a prohibition would be in tension with the desirability or necessity of making election statements to the voting public that are clear and succinct, attention-grabbing and (very commonly) vehement and politically rousing. Such statements are part of the warp and weft of politics even if, to the politics scholar or analyst (as well as to political opponents), they can be described as misleading. Sometimes, a complex political point must be presented in a perhaps somewhat misleading way if the gist of it is to be understood by the vast majority of voters.
Moreover, in some instances, as with the interpretation and presentation of statistics, what counts as a ‘misleading’ (political) statement may depend on the point of view from which the statement is made, and there may be differing legitimate points of view.
By contrast, a prohibition – as exists at present - on misleading statements solely about election administration that the maker knows to be misleading, or realises might be misleading, is defensible. This is in part because the scope for controversy and political contention over such statements is far more restricted. Questions about the election process itself - about when, how and where to vote, about who is standing where and for which party, and so on – are generally not the subject of political contention and intense disagreement. Further, a failure to prohibit misleading statements (when made intentionally or recklessly) about such matters threatens to undermine protection of the right to vote.
-----------------------
It might assist to qualify any prohibition on misleading statements by requiring them to have been (say) ‘gravely or seriously misleading’. Likewise, a prohibition on false statements could be confined to ‘serious and significant’ falsehoods.
Otherwise, reliance should be placed on the fact that both ‘false’ and ‘misleading’ are ordinary words that any lay person can understand perfectly adequately. Further legal definition risks undesirable ‘satellite’ litigation over the scope of legal prohibitions.
-----------------------
I have already indicated that the Bill should make it clear that an exemption should be made for the expression of political opinions, but this is obviously a matter falling under 2B(g): exemptions and exceptions.
It might be helpful to include in 2B an indication of time limit not just for the application of the prohibition (2B(c)), but also for commencing a prosecution. Whilst, in general, English law does not place time limits on prosecutions, there is an argument that the highly context-specific nature of a false election or referendum claim means that it should not be possible to commence proceedings more than, say, one year after the end of the election.
For example, it would seem quite wrong that, as ‘grist to the mill’ in their current election campaign, a candidate (X) could start false statement proceedings against a rival (Y) respecting statements made by Y in the election in which both candidates were involved five years earlier.
Further, a question arises over whether it should be possible to bring a private prosecution for violation of a false statement law. Common sense suggests that this should not be permitted. Candidates already have private law defamation rights. Further, there have been cases in the USA where wealthy candidates have used false statement prosecutions to intimidate and harass less well-off opponents.
Something that perhaps needs to be addressed in 2B is whether it will be possible for someone to obtain a restraining order in respect of false or misleading statements, to prevent them being made at all, or to prevent further dissemination. In practice, there will rarely be time for this: a well-advised propagator of false information will maximise its impact by making any false statement shortly before voting commences, to frustrate the impact of the law. Even so, the issue must be considered. It must be kept in mind that ‘prior restraint’ is the most likely kind of measure to fall foul of Article 10.
----------------
I have suggested that the Bill should broadly track the existing law, in terms of limitations on the kinds of false statements that are to be prohibited (personal attacks on candidates; fraud relating to election administration).
However, there is one further reform that it would be very important to consider.
Currently, criminal liability can be incurred for making a false personal statement about a candidate when this occurs merely through carelessness (RPA 1983, section 106). This is arguably a disproportionate response to the problem of false statements.
Whether a statement is false or misleading, mere carelessness in the making of the statement should probably not trigger potential liability. Election processes are fast-moving and there may often simply be insufficient time fully to check the accuracy of important claims before making or sharing them. Criminalisation should be focused on knowing (in the case of misleading statements) or knowing/reckless (in the case of false statements) statements intended to affect elections or referendum outcomes
------------
No comment to make in this.
---------------
A false statement law is likely to contravene human rights law in the following circumstances:
(a) If a prohibition on false or misleading statements does not require proof of knowledge or recklessness in the making of the statement.
(b) If a prohibition is accompanied by wide powers of prior restraint in the hands of an aggrieved party.
(c) If penalties are disproportionate (as when they include the possibility of imprisonment).
(d) If the period in which the prohibition applies is unclear.
(e) If the extension of the law to include false or misleading information about the substantive political views of candidates (as opposed to information of a personal nature or about the election process) is likely to have a chilling effect on free speech generally.
(f) If inadequate or unclear guidance is given on the circumstances in which a prosecution may be undertaken.
Additional points:
2A speaks of the focus of prohibitions as being the ‘making or publishing’ of false or misleading statements. Whilst the ‘making’ of a statement might seem clear enough, clarity in fact needs to be provided on whether this extends to forwarding a statement (possibly without reading it), or to copy-and-paste use of the statement in various contexts.
Secondly, in the modern world, ‘publish’ has the potential to be too narrow and uncertain in scope as a focus. It would be better to make clear that what is meant by this – if indeed this is what is meant – is that what is prohibited is making a statement available to the public or to a section of the public. Messaging within private online groups (of friends etc) should ordinarily be excluded from scope, as it is an intrusion on private conversation.
*Author of Jeremy Horder, Criminal Fraud and Election Disinformation: Law and Politics (Oxford 2022).