Harry Miller (R on the application of) v The College of Policing

The central issue raised in the appeal is the lawfulness of certain parts of a document entitled the Hate Crime Operational Guidance (the Guidance). The Guidance, issued in 2014 by the College of Policing (the College), the respondent to this appeal, sets out the national policy in relation to the monitoring and recording of what are described in the Guidance as non-crime hate incidents. At the root of the challenge is what is called perception-based recording. Specifically, the policy that non-crime hate incidents must be recorded by the police as such (against the named person allegedly responsible) if the incident is subjectively perceived by the “victim or any other person to be motivated by a hostility or prejudice against a person who is transgender or perceived to be transgender” and irrespective of any evidence of the “hate” element. The lawfulness of the relevant parts of the Guidance is challenged as contrary to the appellant’s right to freedom of expression, both at common law and as protected by Article 10 of the European Convention on Human Rights (the Convention).

The appeal is from the Order of Julian Knowles J dated 14 February 2020. By that Order, Julian Knowles J:

Granted the application of the appellant, Mr Harry Miller, for judicial review of the Chief Constable of Humberside’s recording of a non-crime hate incident (in respect of Mr Miller) under the Guidance, and the subsequent actions taken in relation to him by officers under the Chief Constable’s command (which included seeking to prevail on Mr Miller not to continue tweeting about proposed reforms to the Gender Recognition Act 2004); but

Dismissed Mr Miller’s application for judicial review against the College, in respect of the lawfulness of the Guidance itself.

By the same Order the judge granted a ‘leapfrog’ certificate under section 12(3A) of the Administration of Justice Act 1969, certifying a point of law of public importance in relation to the lawfulness of the Guidance. The judge also gave Mr Miller permission to appeal to the Court of Appeal in the event that the Supreme Court refused permission to appeal. By an Order dated 30 July 2020, the Supreme Court refused permission to appeal on the ground that it considered that the appeal should be first heard by the Court of Appeal. The judge also refused the second defendant permission to appeal, and that application was not renewed.

The judgment under appeal provides a detailed account of the facts and the legal framework to which reference can be made where necessary. References to specific paragraphs from it are contained in square brackets. The judge held in summary (i) that Mr Miller’s broad-based challenge to the legality of the Guidance under the common law and Article 10 of the Convention, failed; (ii) that the mere recording of a non-crime hate incident based on an individual’s speech is not an interference with his or her rights under Article 10(1) of the Convention, but that if it is such an interference, it is prescribed by law and done for two of the legitimate aims in Article 10(2), namely the prevention of crime and the protection of the rights of others; (iii) that the Guidance does not give rise to an unacceptable risk of a violation of Article 10(1) on the grounds of disproportionality; but (iv) that the police’s treatment of Mr Miller after a complaint was made against him that was recorded as a non-crime hate incident, disproportionately interfered with his right of freedom of expression.

Mr Miller is a shareholder in a plant and machinery company in Lincolnshire. He is a former police officer who holds a number of degrees and formerly taught in higher education.

The College is a company limited by guarantee, established in 2012. It is owned by the Secretary of State for the Home Department but operates at arms-length from the Home Office. The College’s work is limited to policy. It has no operational role. According to the evidence of David Tucker, its Faculty Lead for Crime and Criminal Justice, the College is the professional body whose purpose is to provide those working in policing with the skills and knowledge necessary for effective policing; and its purpose is to support the fight against crime and to protect the public by ensuring professionalism in policing. Amongst its principal responsibilities are setting standards and developing guidance and policy for policing; building and developing the research evidence base for policing and supporting the police, other law enforcement agencies and those involved in crime reduction. The National Police Chief’s Council (NPCC), which replaced the Association of Chief Police Officers (ACPO) has responsibilities regarding the operational side of policing.

Sections 123 to 124 of the Anti-Social Behaviour, Crime and Policing Act 2014 give power to the College to issue regulations and codes of practice in support of its functions. In addition, the College issues manuals of guidance and advice called Authorised Professional Practice (APP) which set out standards that police forces and individual officers should apply when discharging their duties.

Article 10 of the Convention provides as follows:

The (non-statutory) Guidance was developed by ACPO and adopted by the College. It was published in 2014.

The genesis of the Guidance can be traced back to the recommendations made by the report produced by the late Sir William Macpherson in February 1999 (the Macpherson Report) following a public inquiry that was established after the murder in April 1993, of Stephen Lawrence. Amongst the recommendations made by the Macpherson Report were that:

The police adopt the definition of a racist incident as “any incident which is perceived to be racist by the victim or any other person” (recommendation 12);

“The term “racist incident” must be understood to include crimes and non-crimes in policing terms. Both must be reported, recorded and investigated with equal commitment” (recommendation 13);

Codes of practice be established by the Home Office, in consultation with police services, local government and relevant agencies, to create a comprehensive system of reporting and recording of all racist incidents and crimes (recommendation 15);

The Metropolitan Police Service review their procedures for the recording and retention of information in relation to incidents and crimes to ensure that adequate records are made by individual officers and specialist units in relation to their functions, and that strict rules require the retention of all such records as long as an investigation remains open (recommendation 21).

ACPO produced its first ACPO Hate Crime Manual in 2000, which was republished in 2002 and 2005. This set out the national policy response in relation to hate crimes and hate incidents and incorporated the recommendations of the Macpherson Report. Following a review by Sir Adrian Fulford in 2006, a shared definition across the criminal justice sector of hate crimes and non-crime hate incidents was developed; and in 2007, this definition was adopted by the Crown Prosecution Service (the CPS) and other criminal justice agencies. That definition is incorporated in the Guidance which superseded the 2005 Hate Crime Manual.

In October 2020, after therefore the handing down of the judgment under appeal, the Guidance was fully replaced by a new Hate Crimes Authorised Professional Practice (the Revised Guidance). The Revised Guidance retains the same essential features of the Guidance, including recording non-crime hate incidents, based on complainant perception, but its contents contain some amendments to the Guidance. These include a (stronger) warning against police taking a disproportionate response to reports of a non-hate crime incident, the inclusion of a link to the judgment under appeal and detailed further guidance on how police should contact people “accused” of non-crime hate incidents. We received very brief submissions from the parties on the implications of those amendments. We were also provided after the hearing with a table comparing the Guidance and the Revised Guidance. Unless stated otherwise, this judgment concerns the Guidance as promulgated in 2014.

The Guidance is principally concerned with hate crime rather than non-crime hate incidents. However Section 6 of the Guidance (headed “Responses to hate incidents) is exclusively concerned with “Non crime hate incidents”. Nonetheless, matters relevant to non-crime hate incidents are also addressed in Section 1 (headed “Defining hate crime”).

Mr Miller’s challenge is centred as I have said on the requirement of perception- based recording, and in particular, the requirement in Section 6.3 that a hate incident (defined in Section 6.1) must be recorded as such “irrespective of whether there is any evidence to identify the hate element”. Mr Ian Wise QC who appears for Mr Miller, was at pains to point out however that Mr Miller has never objected in principle to the police recording non-crime incidents, so long as they are relevant to police functions, nor is any objection made to a record being made of the fact that a complaint of this nature has been made. His complaint is centred on the uncritical recording (and therefore acceptance) as it is submitted to be, of such as non-hate crime incidents without any investigation or inquiry as to their veracity and regardless of whether there is any objective evidence to support the complaint that is made.

The Guidance is described as Operational Guidance. The introductory words to Section 1 state that “The police service must provide an appropriate level of service to victims of hate crime. All officers and staff must have a clear understanding of what constitutes a hate incident and a hate crime.”

The five monitored strands of hate crime are then identified. These are disability, race, religion, sexual orientation and transgender.

Section 1.1 identifies the rationale for collecting data on what is described as monitored hate crime, namely that it helps to provide an accurate picture of the extent of hate crime and to deliver an intelligence-led response.

Section 1.2 is headed Agreed Definitions. Section 1.2.1 is headed ‘Monitored hate crime’ and contains a box with three headings: Title, Definition and Included Subjects. “Hate motivation” is defined in this way:

As for “Hate incidents”, a hate incident in relation to people who are transsexual, transgender, transvestite and those who hold a gender recognition certificate under the Gender Recognition Act 2004 is defined as:

Section 1.2.2 headed Hostility, says:

Section 1.2.3 which is headed “Perception-based recording of hate crime” provides that:

Para 1.2.4 which is headed “Other person” provides that:

police officers or staff

witnesses

family members

civil society organisations who know details of the victim, the crime or hate crimes in the locality, such as a third-party reporting charity

a carer or professional who supports the victim

someone who has knowledge of hate crime in the area – this could include many professionals and experts such as the manager of an education centre used by people with learning disabilities who regularly receives reports of abuse from students

a person from within the group targeted with the hostility eg a Traveller who witnessed racist damage in a local part

Paragraph 1.2.5 which is headed “Malicious Complaints” provides that:

Section 1.5. deals with Secondary victimisation. This is:

Further, consistent with the approach to perception-based recording Section 1.5 of the Guidance goes on to say:

Section 6 of the Guidance provides guidance on responses to hate incidents. It states as follows:

prescribed by or in accordance with the law

necessary in a democratic society

in pursuit of one or more legitimate aims specified in the relevant Article

proportionate.”

It is necessary to see how the recording by the police which is under scrutiny in this case, fits in to the legal and regulatory framework more generally.

It is not controversial in this appeal, and the judge correctly stated at [156], that the police have common law powers to obtain and store information for policing purposes, including for the maintenance of public order and the prevention and detection of crime. There is, moreover, a comprehensive legal framework that regulates the retention of that data: see R(Catt) v Commissioner of Police of the Metropolis [2015] UKSC 9 at para. 7; Catt v UK app. No. 43514/15 at para. 34. The police are subject to the Data Protection Act 2018 (the DPA 2018) with regard to the processing of personal data. Part 3 of the DPA 2018 contains numerous safeguards over the processing (including the retention and disclosure) of data, including specific obligations under Part 3 of the DPA 2018 with regard to the processing of personal data by chief constables of police forces for law enforcement purposes (“Law enforcement processing”): see para. 5 of schedule 7 and section 30(1)(a). Law enforcement purposes “are the purposes of the prevention, investigation, detection or prosecution of criminal offences or the execution of criminal penalties, including the safeguarding against and the prevention of threats to public security”: see section 30(1) DPA 2018. The collection of non-crime hate incident data falls within this provision.

The recording of incidents by the police which may become crimes or may remain non-criminal is regulated by the National Standard for Incident Recording (the NSIR), for which the NPCC now has responsibility. The NSIR (issued in 2011) states in its introduction that it “was introduced to replace the wide variety of incident recording (and non-recording) that differed from force to force so that common understanding and recording practices would result in effective data provision and use.”

An incident is defined under the NSIR as: “a single distinct event or occurrence which disturbs an individual’s, group’s or community’s quality of life or causes them concern.” (para 1.4). Chapter 2.5 of the NSIR sets out a list of qualifiers, which are data descriptions used to capture key aspects and characteristics of an incident. One such qualifier is “hate and prejudice”.

The NSIR is to be distinguished from the National Crime Recording Standard (NCRS) which is incorporated into the Home Office Counting Rules for Recorded Crime (the Home Office Counting Rules) and provides for the national standard in relation to the recording of crimes that are notifiable to the Home Office for monitoring purposes. Para 2.1 of the Home Office Counting Rules provides that:

Relevant guidance as to what is recorded in an incident report is provided in a (publicly available) APP, “Information Management Collection and Recording” produced by the College. This provides that the minimum data standards to be complied with when recording information on an incident record are time and date of the report received, method of reporting, time and date the report was recorded, an incident unique reference number, details of the person making the report (name, address and telephone number), sufficient information to describe the location and nature of the report, opening and closing category and time and date of initial and closing classification.

A statutory Code of Practice on the Management of Police Information (the MoPI Guidance) was published by the Secretary of State for the Home Department in July 2005, under sections 39 and 39A of the Police Act 1996 and sections 28, 28A, 73 and 73A of the Police Act 1997, and provides at para. 1.1.7 that it will be supported by more detailed guidance defining information standards required within forces, which must be framed in compliance with the principles established by the MoPI Guidance. The more detailed guidance to which this refers is to be found in the APP, “Information Management – Retention, review and disposal” published by the College in 2014, which provides further information on the collection, recording, evaluation and processing of police information, including non-crime hate incidents. Records of non-crime hate incidents data are not monitored nationally or held on the Police National Computer but as the Guidance itself makes clear, are held and monitored by police forces locally. Such incidents either fall into Group 3 under the MoPI APP, information that must be retained for an initial period of six years, following which it may be subject either to review at five-yearly intervals or automatic time-based disposal, or Group 4, in which case it must be retained for a minimum period of six years.

One feature of the argument in this case has concerned the fact that non-crime hate incidents may be disclosed in certain circumstances, notwithstanding the absence of any underlying objective assessment of the accuracy of such a report; and the chilling effect this may have on freedom of expression.

Non-crime hate incidents cannot be disclosed on ordinary criminal record certificates. They may however be disclosed on an “enhanced criminal record certificate” (ECRC) which may be requested from the Disclosure and Barring Service (DBS) by employers in relation to a list of certain positions, such as teachers, social workers and carers. The decision whether to include information relating to a non-crime hate incident will be taken by a chief officer in each local police force. In this connection, section 113B of the Police Act 1997 provides that:

The Secretary of State for the Home Department published the Statutory Disclosure Guidance under section 113B(4A) of the Police Act 1997 in August 2015. Principles for the chief officer to follow include: not disclosing information if it is trivial or simply demonstrates poor behaviour or poor lifestyle (para 15); ensuring information is sufficiently credible before disclosure (para 18); and considering whether the applicant should be offered the opportunity to make representations before the information is submitted (para 26).

There is also a right to raise a dispute over any proposed disclosure; and a right to appeal to the Independent Monitor, who can require the DBS to issue a new certificate omitting information considered to be not relevant for the purpose sought: see section 117A and 117A(5) of the Police Act 1997.

It is to be noted that there is no requirement to notify a person that a record of a complaint has been made about them. The decision on notification is a matter for the individual police agencies.

Mr Miller holds what are sometimes described as gender critical beliefs which are encapsulated in his belief that trans women are men who have chosen to identify as women. Mr Miller considers that conflating sex (which he considers to be a purely biological classification) with gender, poses a risk to women’s sex-based rights.

Mr Miller is particularly concerned about potential reforms to the Gender Recognition Act 2004 which would enable men to obtain legal recognition as women based on self-identification (as opposed to undergoing the current exacting process which that Act requires). Specifically, the Government’s 2018 consultation on reforms to the Gender Recognition Act (Reform of the Gender Recognition Act – Government Consultation July 2018) proposed replacing the current requirements for obtaining a Gender Recognition Certificate with an approach that places a greater emphasis on the self-identification by a person of their gender. For the past few years, Mr Miller has tweeted extensively about these matters.

His views on gender re-assignment and the proposed reforms are set out in his witness statement made in support of these proceedings. There he says:

The topic on which Mr Miller was tweeting and its broader implications are plainly important matters of public interest on which strong views are held and publicly expressed. Some are concerned as Mr Miller is, that enabling men to obtain legal recognition as women based on self-identification would carry risks for women because, for instance, it may make it easier for trans women to use single-sex spaces such as women’s prisons. Others, however, consider it is of paramount importance for trans individuals to be able to obtain formal legal recognition of the gender with which they identify more easily.

In the context of this case, the judge said that Mr Miller’s views are supported by many women’s rights campaigners and academics, including Professor Kathleen Stock, Professor of Philosophy at Sussex University at the material time, and Jodie Ginsberg the CEO of Index on Censorship, both of whom have provided evidence in support of Mr Miller’s claim. This includes evidence of the pressure by various means put on Professor Stock to prevent her from expressing her views through the operation of what has come to be known as the “cancel culture” and the climate of fear that has arisen around the public expression of views about these issues. Thus, Professor Stock describes the hostile climate facing gender-critical academics working in United Kingdom universities; Ms Ginsberg gives evidence of the concern that Twitter is stifling legitimate debate and provided examples of the police taking action because of posts about transgender issues on Twitter [248/9]. Some of those involved in the debate label those with different views as transphobic or displaying hatred when they are not; and are intolerant of different views, even when expressed by legitimate scholars whose views were not grounded in hatred etc but are based on legitimately held different value judgments and form part of mainstream academic research [250].

In her evidence on this topic, Professor Stock gave an example of three utterances: “Trans women are men”; “Trans women aren’t women” and the use of the pronouns “he/him” rather than “she/her” in referring to a trans woman in the third person. Professor Stock describes these as utterances which are intended in the mouths of many people as simple observable facts, and non-evaluative utterances, along the lines of “water boils at 100 degrees” or pillar boxes in the UK are red.” So here, she says, the failure or refusal to use of a preferred pronoun of a trans woman is not an expression of hostility but an indication of a descriptive, non-evaluative belief, that the trans woman is biologically male; and the fact that such readings tend to be heard as transphobic is not therefore a reliable guide to the true nature of the utterances. On the other hand, trans advocacy groups, such as Stonewall explicitly define transphobia as the fear or dislike of someone based on the fact that they are trans, including the denial or refusal to accept their chosen gender identity.

The specific facts relating to these proceedings were dealt with by the judge in detail at [18] to [100]. Most of those facts were not contentious, and the judge’s limited conclusions on the evidence are not the subject of challenge in this appeal. The position in summary is as follows.

The 31 tweets which gave rise to these proceedings were posted by Mr Miller between November 2018 and January 2019 using the Twitter handle @HarrytheOwl. A selection of the tweets, together with Mr Miller’s explanation of them, were set out by the judge at [23] to [57]. The judge described Mr Miller’s tweets as, “for the most part, either opaque, profane or unsophisticated” but emphasised that he was “quite clear that they were expressions of opinion on a topic of current controversy, namely gender recognition.” Unsubtle though they were, the judge said, Mr Miller “expressed views which are congruent with the views of a number of respected academics who hold gender-critical views and do so for profound socio-philosophical reasons” (see [251]). The judge relied on the evidence of Ms Ginsberg as illustrating that many other people hold concerns similar to those held by Mr Miller.

Mrs B (Footnote: ¹) who describes herself in her witness statement in these proceedings as a post-operative transgender woman, had Mr Miller’s tweets drawn to her attention by a friend. The judge found that Mrs B made a voluntary choice to read the tweets; they were not directed at anyone in particular or the transgender community, but were simply posted on Twitter to be read by Mr Miller’s Twitter followers or anyone else who might come across them. Further, there was no evidence of what Mrs B’s friend thought of them, or that anyone other than the friend, or Mrs B read them or found them to be offensive or indecent or complained about them [271].

Mrs B was offended by them, however, and made a complaint to Humberside Police via an online system called True Vision, choosing that force because Mr Miller was chairman of a company based in its area. The complaint asserted that Mr Miller had been making transphobic comments on his Twitter account, “designed to cause deep offence and show his hatred for the transgender community.”

The tweet that appears to have most concerned Humberside Police was a “re-tweet” by Mr Miller of a verse written by a feminist song writer which Mr Miller considered “reveals the sentiment that many feminists feel – that male privilege is now encroaching on womanhood” and which reads as follows:

Following the complaint made by Mrs B, Humberside Police decided to record Mr Miller’s Twitter activity as a “hate incident” pursuant to the impugned provisions of the Guidance: see [69]. The judge found that the police did so “simply on the say so of Mrs B and without any critical scrutiny of the tweets or any assessment of whether what she was saying was accurate”: see [70]. Humberside Police then created various crime reports in which Mr Miller was described as a “suspect” and Mrs B a “victim”: see [71] to [72]. The Crime Report created by Humberside Police for the incident, under the heading “Modus Operandi Summary”, recorded that the “suspect”, i.e., Mr Miller, was “posting transphobic comments on Twitter causing offence and showing hatred for the transgender community." It referred to the “Primary Offence” as “Hate incident - transgender”. It recorded, without qualification, Mrs B’s view that “These comments [i.e. Mr Miller’s Twitter comments] are designed to cause deep offence and show his hatred for the transgender community”. Language and labels matter: anyone reading the Crime Report might understandably assume that Mr Miller was suspected of committing a hate crime against the transgender community.

In addition to recording the report by Mrs B in this way, the decision was made to allocate the case to an investigating officer. This was P.C. Gul, an officer from Humberside Police. P.C. Gul went to Mr Miller’s Head Office on 23 January 2019 to speak to him. Mr Miller was not there but was told about the visit by his Managing Director and subsequently spoke to PC Gul on the telephone. During that call PC Gul told Mr Miller that the police had been contacted about 30 tweets Mr Miller had posted and which were said to be transphobic and which Humberside Police had recorded as a non-crime hate incident under the Guidance. PC Gul specifically referred to the re-tweet. He said Mr Miller had not committed a crime but told him if his behaviour escalated then it may become criminal, and the police would need to deal with that appropriately. He strongly advised Mr Miller to cease tweeting gender critical messages. When Mr Miller asked why PC Gul was wasting his time, the officer said: “I need to check your thinking.” Mr Miller was not told who had made the complaint, or, apart from the re-tweet, which of his tweets were complained about (He found this out later, after making a subject access request on 9 May 2019).

Mr Miller lodged a formal complaint with the Humberside Police Professional Standards Department about their conduct three days after the visit by PC Gul, and also tweeted about what had happened. A publicly posted statement from Assistant Chief Constable Young of Humberside Police on 28 January 2019 defended the actions of PC Gul. It described Mr Miller’s tweets as “transphobic”, said that the force aimed to ensure that “hate related incidents” do not “escalate” and that “the correct decision was made to record the report as a hate incident.”

There was subsequent contact between Mr Miller and Acting Inspector Wilson of Humberside Police, by telephone and by letter of 28 March 2019, in which the police’s action was defended as proportionate. The letter said that Mr Miller had been spoken to, to help him “understand the impact [his] comments could have on others and to prevent any possible escalation in the future”. In addition, Humberside Police issued a statement to the Hull Daily Mail, which said in part: “There was never any suggestion he shouldn’t engage in politics or debate around the subject, he was just asked why he would want to, knowing it would cause distress and upset to others in society”. Mr Miller was not satisfied with the police response to his complaints. He referred the matter for resolution to a local resolution procedure and requested that the recording of a hate incident be removed from police records. Mr Miller’s complaint was finally dismissed by the Humberside Police Appeals Body by a letter dated 18 June 2019. That letter said amongst other things that in recording the tweets as a hate incident, it was clear that the officer was following the Guidance.

The judge found that the Police’s actions “led [Mr Miller], reasonably, to believe that he was being warned not to exercise his right to freedom of expression about transgender issues on pain of potential criminal prosecution”, without the police ever explaining “on what basis they thought that the [Mr Miller’s] tweets could ‘escalate’ to a criminal offence” ([100]). The judge also found that there was “not a shred of evidence that [Mr Miller] was at risk of committing a criminal offence” ([259]). The judge accepted, that as Mr Miller said in his evidence, in consequence of the actions of Humberside Police, he felt compelled to withdraw from involvement with his company. In addition, he and members of his family were subjected to threats and intimidation from a number of individuals on Twitter, causing them briefly to leave the family home. However, after much deliberation, Mr Miller decided to continue tweeting about transgender issues.

Mr Miller’s claim for judicial review was commenced on 4 June 2019 and permission to apply was granted by Order of Cockerill J dated 5 August 2019. Mr Miller sought (i) a declaration that Section 6.3 of the Guidance was unlawful in that it breached his Article 10 and common law right to freely express his views on transgender issues and the proposed changes to the Gender Recognition Act; and (ii) a Mandatory Order requiring Humberside Police to expunge the non-crime hate incident from the record, or alternatively to recognise that the record made in January 2019 was wrongly made. The claim was made on the ground that the recording by the police of tweets sent by him as ‘non-crime hate incidents’ was a violation of his fundamental right to freedom of expression found in Article 10 of the Convention and the common law. This violation, it was said, was all the more egregious and unjustifiable given that the tweets concerned controversial proposed changes to legislation which had been the subject of widespread public consultation.

Mr Miller did not claim that the events which gave rise to these proceedings or any part of them had violated Mr Miller’s rights to respect for his private and family life under Article 8 of the Convention.

The judge heard the claim over 3 days in October 2019. Fundamental to the complaint against the College, as I have already said, was that the Guidance requires police forces to record non-crime hate incidents (which are defined as “any non-crime incident which is perceived by the victim, or any person to be motivated (wholly or partially) by hostility or prejudice” directed at one of five “monitored strands”, namely “race, religion, sexual orientation, disability and transgender”) regardless of whether or not the complainant is the victim, and irrespective of whether there is any evidence to justify the hate incident. The judge handed down a judgment of 289 closely reasoned paragraphs on 4 February 2020, dismissing the claim against the College, but granting relief against the Chief Constable of Humberside (the second defendant).

In terms of remedy, the judge granted Mr Miller a declaration that the following actions by officers under the Chief Constable of Humberside’s command in combination, unlawfully interfered with his right to freedom of expression under Article 10(1) of the Convention: recording pursuant to the Guidance, 31 tweets posted by him on Twitter between November 2019 and January 2019 as a non-crime hate incident; pursuant to the said recording, the visit by PC Gul to Mr Miller’s place of work to speak to him about the tweets; PC Gul warning Mr Miller in the course of a subsequent telephone conversation that he might be prosecuted if he escalated matters; and subsequently, twice in written communications by Assistant Chief Constable Young and Acting Inspector Wilson, referring to risks of escalation and directly or indirectly, to the risk of prosecution. The judge also made a mandatory order that Humberside Police amend its record of a non-crime hate incident involving Mr Miller’s tweets to record that the Court had made a declaratory order in the above terms. For the avoidance of doubt however, the Order recorded that the mere recording by the Humberside Police of Mr Miller’s tweets, of itself, pursuant to the Guidance, did not violate Article 10(1) of the Convention.

The record made of the particular non-crime hate incident therefore remains in place, albeit accompanied by the Court’s declaratory Order.

There are five grounds of appeal. These are (1) the judge was wrong to hold that the principle of legality is merely a principle of statutory construction, and therefore wrong to decide that the impugned provisions in Chapter 6 of the Guidance are lawful in the absence of express statutory or established common law authorisation; (2) the judge was wrong to hold that the approach in the Guidance to the mandatory recording of ‘non-crime hate incidents’ in the absence of any evidence of hate is lawful as a matter of common law; (3) the judge was wrong to hold that the impugned provisions in Chapter 6 of the Guidance involve no interference with the right to freedom of expression under Article 10(1) of the Convention; (4) the judge was wrong to hold that the impugned provisions in Chapter 6 of the Guidance satisfy the Convention requirement of “foreseeability”, with the consequence that he was incorrect to conclude that any interference with Article 10(1) of the Convention arising from those provisions is “prescribed by law” for the purposes of Article 10(2) of the Convention; and (5) the judge was wrong to hold that the impugned provisions in Chapter 6 of the Guidance are “necessary in a democratic society”.

The judge held (at [152]-[172]) that the Guidance was lawful under domestic law because the police have the power at common law to record and retain a wide variety of data and information; and the cases make clear that no statutory authorisation is necessary in relation to non-intrusive methods of data collection, even where the gathering and retention of that data interferes with Convention rights. Collecting details of hate crime and non-crime hate incidents formed one aspect of the police’s common law duty to keep the peace and prevent crime. This did not contravene the principle of legality formulated by Lord Hoffmann in R v Secretary of State for the Home Department, ex parte Simms [2000] 2 AC 115 as that principle was limited to one of statutory construction and did not operate as a free-standing control over the exercise of non-statutory powers such as the College was exercising when it issued its Guidance in 2014.

The argument for Mr Miller on Ground 1 proceeds as follows. Any action which may interfere with the fundamental right to freedom of expression, unless clearly authorised by primary legislation, is unlawful. This is an aspect of the principle of legality, and insofar as the judge dismissed this challenge on the grounds that that principle was inapplicable because no issue of statutory construction arose, he was in error. The cases he cited in support of the point are either wrongly decided or cannot stand in the light of the decision of the Supreme Court in R (Miller / Cherry) v Prime Minister [2019] UKSC 41, [2019] 3 WLR 589. In the absence of any statutory power permitting such an interference, the judge should therefore have found that the Guidance was unlawful at common law to the extent that it interfered with freedom of expression. In the alternative, he should have found that the Guidance was unlawful in the absence of any “clearly defined” common law power permitting such interference.

I would reject this aspect of Mr Miller’s challenge. In my view, the judge was right to reject this ground of challenge for the reasons he gave. As the judge found, and is accepted by Mr Miller, the police have the power at common law to record and retain and use a wide variety of data and information, including images of information and personal information. Putting aside the language used to describe the incidents themselves, I can see no basis for the argument that the common law powers do not extend to the collection and retention of details of non-crime hate incidents. Indeed, I regard the various arguments which are advanced to the contrary to be misconceived.

First, it is clear that no statutory authorisation is necessary in relation to non-intrusive methods of data collection, even where the gathering and retention of that data interferes with Convention rights. See for example, R (Wood) v Commissioner of Police of the Metropolis [2010] 1 WLR 123, R (Catt) v Association of Chief Police Officers [2015] AC 1065 and R (Bridges) v Chief Constable of South Wales Police [2019] EWHC. Secondly, as Mr Jason Coppel QC for the College points out, there are a great many areas of common law, which lawfully interfere with free speech rights, and are unsupported by primary legislation - common law contempt of court and common law breach of confidence being two such examples. It is not the case, therefore that as a matter of principle, any interference with free speech can only be lawful if there is a statutory basis for it. Indeed if there were such a principle, this would be a much more stringent restriction upon executive action than Articles 8 and 10 of the Convention, and large swathes of the common law would become inoperable.

Thirdly, the boundaries of the principle of legality may be fertile ground for academic debate, but for the present, there is weighty authority, including authority binding on this court, that the principle is one of statutory construction. As it was described by Lord Hoffman in ex parte Simms at p.131, the principle is a constraint on the legislative actions of Parliament, preventing general words in legislation being construed incompatibly with fundamental human rights unless the express language or the necessary implication shows that Parliament intended those words to override such rights. It is not a free-standing ground for control of all types of action by public bodies, particularly the exercise of non-statutory power. See further AJA v Commissioner of Police of the Metropolis [2013] EWCA Civ 1342, [2014] 1 WLR 285 at para 28; R (Al-Saadoon) v Secretary of State for Defence [2017] QB 1015 at para 198 and R (Youseff) v SSFCO [2012] EWHC 2091 (Admin), [2013] QB 906 where Toulson LJ deprecated the “tendency on the part of lawyers…to seek to use the ‘principle of legality’ as a developmental tool providing an additional ground of challenge in a case purely involving questions of common law, i.e. not a case where the defendant is seeking to justify his action by reference to a statutory power.”

Returning to this case, the exercise by the police of their common law powers to record information is extensively regulated by statute including by data protection legislation and the Human Rights Act 1998 (the HRA). Pursuant to the HRA, the actions of the police, and guidance issued by the College in respect of those actions, must be sufficiently foreseeable as to be prescribed by law (addressed under Ground 4 below). Any further issue as to the degree of intrusion into freedom of expression, if there is one, falls to be analysed under the Article 10(2) issue of proportionality (addressed under Ground 5 below). There is no common law principle which overrides that analysis and requires a finding in favour of Mr Miller without him establishing his case on these important issues.

The decisions in R (Miller / Cherry) v Prime Minister and R (UNISON) v Lord Chancellor [2017] UKSC 51, [2017] 3 WLR 409 on which reliance is placed, take Mr Miller’s argument no further in this connection. Miller/Cherry was an exceptional case on its facts, concerning Parliamentary sovereignty as a limit on the prerogative power to prorogue Parliament. It neither addressed in terms nor sought to distinguish the principle of legality as identified in ex parte Simms (a principle which is consistent with Parliamentary sovereignty). As for the decision in R (UNISON) the central issue in that case was whether the Employment Tribunals and the Employment Appeal Tribunal Fees Order 2013 was within the powers conferred by Section 42(1) of the Tribunals, Courts and Enforcement Act 2007, see para 65.

An alternative submission, advanced in writing, by Mr Wise QC but barely touched on in argument, is that any common law power (to interfere with freedom of expression) must be an “exception” which is “clearly defined”: see Re Central Television Plc per Hoffman LJ. However, in that case, the court was concerned to explain why the protection of the privacy of a child did not justify adding new exceptions to existing statutory exceptions to the freedom of expression of the media. That case provides no support, even tangentially, for the introduction of wide-ranging restrictions upon well-established common law powers, including those of police to record information.

The judge said at [171] that Mr Miller’s common law challenge that it was irrational to record non-crime hate incidents based on the perception of the complainant, “in reality is an argument about proportionality which is to be analysed as part of the [Mr Miller’s] challenge to [the Guidance] under Article 10”. Albeit Mr Wise QC described this Ground as a free-standing rationality challenge his very brief submissions did not address or explain why the judge was wrong to take the course he did. In the absence of any good reason to do otherwise, I propose to deal with it in the same way as the judge.

There is no challenge to the judge’s conclusion that the Guidance had a basis in domestic law, because it fell within the police’s general common law power to collect, use, retain and disclose information for the purposes of preventing and detecting crime [191]; or that the Guidance satisfied the accessibility test, as it is available to all with access to the internet on the College’s website [192]. It is not in issue either, that as the judge found, the impugned parts of the Guidance pursue the legitimate aim of preventing disorder and crime and protecting the rights and freedoms of others [211].

What is in issue in this appeal however, is the judge’s conclusion that the mere recording of non-hate crime speech did not interfere with Mr Miller’s right to freedom of expression within the meaning of Article 10(1); and his secondary findings (which he went on to make, in case he was wrong about the issue of interference) that the interference, such as it was, was justified within the scope of Article 10(2) because it was sufficiently foreseeable, and therefore prescribed by law and because it was necessary in a democratic society.

The judge dealt with this issue at [175] to [185]. The judge acknowledged the argument that the mere act of recording speech may have a chilling effect on the speaker’s right to freedom of expression, but he said the mere recording without more was too remote from any consequences to amount to a Handyside (Footnote: ²) restriction and had no real consequence for the individual such as Mr Miller [176-7]. In coming to that conclusion, the judge rejected Mr Miller’s submission that the recording of an incident pursuant to the Guidance was or was analogous to the administrative warning which was found to constitute an interference with a publisher’s rights in Balsyte-Lideikiene v Lithuania Application, App no 72596 (ECHR, 4 November 2008). This was because the penalty in that case (punishment accompanied by the confiscation of the publication in question) was not “relatable” to the kind of record-keeping prescribed by the Guidance [178-180]. Instead, the judge accepted the evidence of the College that recording is primarily an administrative process to build an intelligence picture based on statistics; and that the Guidance does not mandate the police to take any form of action in response to a report of a non-criminal hate incident. The judge said, at [177]:

As to the risk of disclosure on an ECRC, the judge accepted [181-182] that non-crime hate incidents, are in principle, disclosable information, and there may be circumstances in which this might occur, for example, if Mr Miller applied for a job which would bring him into contact with vulnerable transgender individuals.

Furthermore, the legal framework relating to the disclosure of non-conviction data on an individual’s ECRC is tightly drawn and subject to safeguards to prevent against arbitrary unfairness including the statutory framework under ss 112-127 of the Police Act 1997 and the Statutory Disclosure Guidance issued by the Secretary of State for the Home Department under section 113B(4) of the Police Act 1997 (see [184] - [185]).

With respect to the judge, I am unable to agree with his conclusions that the impugned provisions of the Guidance do not constitute an interference with Mr Miller’s right to freedom of expression.

The concept of a chilling effect in the context of freedom of expression is an extremely important one. It often arises in discussions about what if any restrictions on journalistic activity are lawful; but in my judgment it is equally important when considering the rights of private citizens to express their views within the limits of the law, including and one might say in particular, on controversial matters of public interest. The familiar words of the European Court of Human Rights (ECtHR) in Handyside v The United Kingdom merit repetition in the present context.

When dealing with the case against the police, but not the College, the judge said the following amongst other things. The tweets did not even arguably amount to hate speech of the sort that transgresses the criminal law, and which might fall outside the protection of Article 10, however offensive Mrs B had found them to be. Further, the context of the debate to which Mr Miller was contributing was complex multi-faceted and important, as was the evidence of Professor Stock and Jodie Ginsberg, which assisted in understanding its contours ([240]-[246]). Contextual evidence was relevant because in the Article 10 context, special protection is afforded to political speech and debate on questions of public interest [252]. The fact that Mr Miller’s tweets were “for the most part, either opaque, profane or unsophisticated” did not rob them of the protection of Article 10(1) [251]. Mr Miller expressed the sort of views that are also held by many academics as part of a multi-faceted debate [265]. When considering the proportionality of the police’s response to Mrs B’s complaint the judge said: “Mrs B profoundly disagrees with [Mr Miller’s] views but such is the nature of free speech in a democracy. Professor Stock’s evidence demonstrates how quickly some involved in the transgender debate are prepared to accuse others with whom they disagree of showing hatred, or as being transphobic when they are not but simply hold a different view [280]. The judge also said that Mrs B’s reaction to the tweets was, at times, at the outer limit of rationality [280].

The judge was right to say as he did, that comparatively little official action is needed to constitute an interference for the purposes of Article 10(1) [255]; and was thus able to find that what the police had done (going to his place of work, warning Mr Miller that he would be at risk of criminal prosecution if he continued to tweet etc.) constituted an interference with Mr Miller’s Article 10(1) rights, even though he was not made subject to any formal sanction. The judge went on to say that the police’s submissions impermissibly minimised what occurred and did not properly reflect the value of free speech in a democracy; there was not a shred of evidence that Mr Miller was at risk of committing a criminal offence [259]. It was nothing to the point that Mr Miller continued to tweet afterwards; warning him that in unspecified circumstances he might find himself being prosecuted for exercising his right to freedom of expression on Twitter had the capacity to impede and deter him from expressing himself on transgender issues. It had a chilling effect on his right to freedom of expression [260-1].

In my view, the principles of law that protect freedom of expression and which underpinned much of what the judge said in support of his conclusion that the police had infringed Mr Miller’s rights were matters that should have led to the same conclusion against the College.

The respondent argues that different outcomes on the issue of interference as between the College and the Chief Constable are justified because of a distinction between the requirement to categorise on the one hand and the requirement to record (emanating from the NSIR) on the other. However, this approach is an artificial one and I am unable to accept it. Paragraph 6.3 of the Guidance says, “where any person… reports a hate incident… it must be recorded regardless of whether or not they are the victim and irrespective of whether there is any evidence to identify the hate element”. The document goes on to make clear that the mechanism for local recording varies with some forces using their crime recording system to do this but also makes clear that hate incidents must be recorded. In the absence of the Guidance, it may be that an incident reported by a member of the public would be recorded as a complaint or incident, but the nature of the record is very different: the record must include time and date of report, method of reporting, details of the person making the report and sufficient information to describe the location and nature of the report (see paras 22 to 25 above). There is no suggestion that the person reporting is referred to as a victim, nor that the person against whom a complaint is made is referred to as a suspect. Moreover ‘incident’ suggests some objective fact being recorded and no necessary stigma from the word hate being attached to it. The Guidance by contrast, requires things like speech to be categorised and recorded as a ‘hate incident’ when no objective incident has necessarily occurred, apart from the speech itself. The contention made on behalf of the College therefore, that the Guidance does not require recording, and that recording happens separately is unsustainable in the circumstances.

If that is right, the question is whether the mere act of categorising and recording speech as hate speech, interferes with Article 10 rights. As the judge said when discussing the actions of Humberside Police, the Strasbourg court has made clear that there is wide protection for all expressive activities by virtue of a very broad understanding of what constitutes an interference with freedom of expression. That is particularly so in the context of political speech and debate on questions of public interest and the Strasbourg court has emphasised that there is “little scope under article 10(2) of the Convention for restrictions on political speech or on the debate of questions of public interest”: see Vajnai v Hungary, App no 33629/06 at [47], quoted by the judge at [252] but not in the context of the Guidance interfering with these rights. The absence of any analogous Strasbourg authority is neither here nor there. In any event, the judge identified a case at [178] where an administrative warning intended to serve as a preventative measure was regarded by the Strasbourg court as an interference with Article 10 rights. The distinction drawn by the judge between the administrative act of categorising which he attributed to the College and the requirement to record which he did not, is what enabled him to distinguish that case from this one. I do not accept that distinction. In short, the recording of non-crime hate incidents is plainly an interference with freedom of expression and knowledge that such matters are being recorded and stored in a police database is likely to have a serious ‘chilling effect’ on public debate.

It is instructive to consider the approach of the Grand Chamber in Altug Tane Akcam v Turkey, (2016) 62 E.H.R.R. 12. In that case, a professor of history had published numerous books and articles on the historical events of 1915 concerning the Armenian population, a subject of great sensitivity in Turkey. He published an editorial opinion in a Turkish-Armenian newspaper criticising the prosecution of the late editor of that newspaper for the crime of “denigrating Turkishness” under article 301 of the Turkish Criminal Code. A complaint was then made against him which led to a criminal investigation, that in publishing that opinion, the applicant himself had breached various provisions of the criminal code, including article 301. The applicant claimed that he had directly been affected by the criminal investigation and there was an ongoing risk that he would be subject to further investigation or prosecution for his opinions on the Armenian issue. The Turkish Government asserted that the investigation had been terminated by a non-prosecution decision by the local public prosecutor and the applicant therefore lacked victim status under Article 34 of the Convention. In addition, a legislative amendment to the text of article 301 in 2008, meant there was no longer any risk of prosecution that the applicant would be prosecuted for expressing such opinions.

The Grand Chamber held the objection about the applicant’s status was inextricably linked to whether there had been an interference with his freedom of expression: see paras 49-51 and went on to hold that the ongoing threat of prosecution interfered with the applicant’s Article 10 rights: see paras 68-83. In reaching those conclusions, the Grand Chamber said amongst other things that the applicant was directly affected because he had shown he was actually concerned with a public issue and was involved in the generation of specific content targeted by article 301. In any event, the applicant could contend that the law had violated his rights if he was required to modify his conduct because of it or risk being prosecuted, or if he was a member of a class of people who risk being directly affected by the measure. The fear of sanction had a chilling effect on the exercise of freedom of expression, given the likelihood of the fear discouraging future similar statements (see paras 65-68). The applicant’s research interests, concerning as they did the sensitive historical events of 1915 concerning the Armenian population, belonged to a group of people who could be stigmatised for their opinions on this topic and subjected to investigations or prosecutions under article 301. The public prosecutor’s decisions of non-prosecution did not necessarily mean that the applicant would be safe from investigations of that kind in the future. Even though, therefore, the impugned provision has not yet been applied to the applicant’s detriment, the mere fact that in the future an investigation could potentially be brought against him had caused him stress, apprehension and fear of prosecution. This situation has also forced the applicant to modify his conduct by displaying self-restraint in his academic work in order not to risk prosecution under article 301: see paras 71-72.

The risk Mr Miller runs is not one of prosecution. Nonetheless in my view, the parallels between this case and that of Altug Tane Akcam are significant. Mr Miller belongs to a group of people who could easily be stigmatised for their opinions and be subject to complaints by those offended by his views. He is able to contend that the impugned provisions have violated his rights as he was required to modify his conduct because of them or risk having a “non-crime hate incident” being recorded against him; and he is member of a class of people who risks being directly affected by the measure. In this case, not only is there a chilling effect on future similar statements because of the fear of a record being made, a matter evident from the evidence the Professor Stock as well as Mr Miller’s wife, to which Mr Wise QC drew our attention, but there is also, at the very least, a non-trivial risk that in future such a record might be disclosed on an enhanced ECRC. The judge’s suggestion that if this happened it would not be a result of the Guidance is, with respect to him, artificial. But for the hate incident record, there would be no such risk. The record means that this risk is there and that inevitably has a chilling effect. The safeguards in relation to disclosure may mean that the interference is proportionate but do not avoid the conclusion that there is an interference.

In considering whether the interference was prescribed by law, Mr Miller’s challenge focusses on the requirement of foreseeability.

Before the judge, Mr Miller argued that the perception-based definition of non-crime hate incidents is such that people cannot foresee the consequences of making a given statement; it is uncertain whether there is a discretion not to record non-crime hate incidents, and, if there is a discretion, its scope is unclear.

The judge accepted what the College submitted in response (see [196] to [210]). He said:

The judge went on to say that the fact that the Guidance means that a complainant should not be called upon to provide evidence that his attacker was in fact hostile to him because of a protected characteristic:

The judge said this conclusion was consistent with Humberside Police’s evidence ([204]-[205]). And there was no inconsistency between the positive requirement to record and proceeding on the basis that the police have a discretion whether to record such incidents, to be exercised by reference to whether doing so would be an “overreaction” and/or the considerations in Section 1.2.4. This was because the mandatory duty to record had to be read as subject to the overarching duty which all public authorities have to abide by the Convention (see Section 6.4), which the Guidance sufficiently delineated ([208]-[209]).

Against that background, Mr Wise QC submits the judge’s analysis was inconsistent and wrong. His importation of the requirement of a rational basis for concluding there is a hate element, and therefore a discretion not to record, was inconsistent with the mandatory recording required by the Guidance, as he had earlier described it. Further, taking the facts of this case as an example, he submits that it is makes no sense to suggest that someone can reasonably predict the reactions of a person acting at the outer margins of rationality (which is how the judge characterised the complaint by Mrs B) to something that was said, so as to regulate their conduct or foresee the consequences of their acts. He also points to the Altug Tane Akcam case where the Grand Chamber concluded, at para 95, that article 301 of the Turkish criminal code, which criminalised “denigrating Turkishness” did not meet the “quality of law” since it’s unacceptably broad terms (too wide and too vague) resulted in a lack of foreseeability as to its effects; and thus constituted a continuing threat to the exercise of freedom of expression. Amongst other things, the Grand Chamber said it was clear from the number of investigations and prosecutions brought under the provision that “any opinion or idea that is regarded as offensive, shocking or disturbing can easily be the subject of a criminal investigation by public prosecutors” (see para 93).

Mr Coppel QC supports the reasoning of the judge. He submits the core answer to this ground of appeal lies in Catt at paras 7 to 17. The Guidance must be read in conjunction with the wider system of legislation, statutory and non-statutory guidance of which it is a part. Retention of data by the police force is in accordance with the law. Afortiori, the recording is in accordance with the law, as is the categorisation of information which the police have recorded. It is plainly foreseeable in the light of the framework that does exist, that the police will retain and categorise information, including information categorised as non-crime hate incidents. Mr Miller’s fire is therefore directed at the wrong question. Whether Section 6.3 of the Guidance is too broad, is a complaint about proportionality. But even if the breadth of the Guidance is relevant to foreseeability, it is material that the interference in this case is at a relatively low level. This was not the case in Altug Tane Akcam where the applicant faced a serious risk of prosecution. Moreover, the purpose of the Guidance has to be borne in mind: it is not lengthy, it does not provide for every eventuality, but it does not have to. This is an area where precision cannot be achieved and flexibility is required, and guidance relating to offensive and discriminatory views, will necessarily be widely drawn. He says further describes as “surprising” the submission that safeguards on potential interference (the discretion not to record) are unlawful.

The general principles applicable to the “in accordance with the law” standard, were helpfully summarised by the Divisional Court in R (Bridges) v Chief Constable of South Wales [2019] EWHC 2341 (Admin) at [80]: and approved on appeal in the same case: see [2020] EWCA Civ 1058 at [55] – [56]:

Foreseeability therefore means, using the Strasbourg formulation, that the relevant provision must be formulated with sufficient precision to enable the citizen to regulate his conduct and foresee to a degree that is reasonable in the circumstances, if need be with appropriate advice, the consequences which a given action may entail: see Perinçek v Switzerland, App no 27510/08, ECHR 2015 at para. 131 amongst many other cases. Built into this formulation is a realistic appreciation of how particular provisions work in the real world. Those consequences need not be foreseeable with absolute certainty: experience shows this to be unattainable. Again, whilst certainty is highly desirable, it may bring in its train excessive rigidity and the law must be able to keep pace with changing circumstances. Accordingly, many laws are inevitably couched in terms which, to a greater or lesser extent, are vague and whose interpretation and application are questions of practice: see Sunday Times v United Kingdom para 49. Similarly, the scope of the concept of foreseeability depends to a considerable degree on the content of the instrument in question, the field it is designed to cover and the number and status of those to whom it is addressed.

Though I do not agree with two aspects of the judge’s reasoning, in my view he was right to conclude, broadly for the reasons he gave at [196] to [198] that perception-based recording does not mean the Guidance fails to meet the criteria of foreseeability. The starting point is that the Guidance clearly identifies what must be recorded. “For recording purposes, the perception of the victim, or any other person (see 1.2.4 Other person) is the defining factor in determining whether an incident is a hate incident or recognising the hostility element of a hate crime. The victim does not have to justify or provide evidence of their belief, and police officers and staff should not directly challenge this perception. Evidence of hostility is not required for an incident …to be recorded as a hate crime or hate incident.” Further, as Lord Sumption said in Catt what is required is law which is "reasonably predictable, if necessary with the assistance of expert advice" but, except perhaps in the simplest cases, this does not mean that the law has to codify the answers to every possible issue which may arise. "It is enough that it lays down principles which are capable of being predictably applied to any situation." It is to be borne in mind that the Guidance is not a legislative instrument, but is non-statutory operational policy guidance, designed to assist and guide the police in carrying out their functions in an exceptionally difficult and sensitive area. As Mr Coppel QC submits, rules relating to offensive and discriminatory views, will, of necessity, be widely drawn. This is not an area where precision could be achieved, and flexibility is particularly important where the provisions fall to be applied to a large number of cases: see Silver v UK (1983) 5 EHRR 347, para 88.

True it is, that there may be cases recorded, and the judge obviously considered Mr Miller’s case to be on the cusp in this respect, where there is no rational basis for concluding there is a hate element. But it seems to me, this is an inevitable consequence of the straightforward point that there may be a difference on occasion between perception (which is what is considered to matter and be of central importance here, and which the recording is aiming to capture) and objective reality. True it also is, that it follows from this, and other associated wording in the Guidance that the scope of what must be recorded is very broad, but these issues are it seems to me ones which as Mr Coppel QC submits can properly be addressed when considering proportionality.

As to the aspects of the judge’s reasoning with which I would disagree, the first is whether a “common-sense discretion” on the part of the police not to record is to be read into the Guidance.

Bearing in mind of course that the Guidance should not be construed like a judgment or a statute, and that regard should be had to how it would be interpreted by those who have to apply it, the starting point is that there is nothing in the Guidance itself (or indeed the updated version which we were shown) which overtly suggests let alone implies that police have any such discretion. On the contrary, the requirement to record is expressed in unambiguous and trenchant terms. This is not coincidental. If such a discretion were to be read into the Guidance, this would undermine the principal ‘Macpherson’ rationale of perception-based recording, namely, to remove the individual decision-making of the police from the process, and to place the perception of the complainant at the centre of the reporting/recording system. Complaint in other words is followed by recording. Looking at the matter from the other end of the telescope, the mandatory requirement to record (based on the perception of the complainant, with few specified exceptions, and the associated need to avoid secondary victimisation) is consistent with the underlying principles which gave rise to the concept of perception-based recording, and avoids placing the police in the cross-hairs of the difficult and sensitive issues they would otherwise have to confront when deciding whether and what to record, presumably in many, if not all cases.

A discretion not to record is specified in the Guidance in Sections 1.2.4 and 1.2.5, but the scope of that specified discretion is clear. The co-existence however of a mandatory requirement to record with a nebulous and apparently unfettered common-sense discretion not to record would, in my opinion, provide a recipe for uncertainty and confusion about what could or should or must be recorded and would diminish to an unacceptable degree the possibility of a person foreseeing the consequences of the Guidance for themselves.

The facts of this case provide something of an illustration of how difficult such decisions can be. Prior to the case coming to court, as has been noted above, the case had passed through the hands of a number of officers, including at a senior level, who had concluded that the complaints by Ms B were rightly recorded as a non-crime hate incident. And this was a stance Humberside Police maintained at the trial (by the College too, who made common cause with Humberside Police on this issue). The judge however took a different view.

At [281] the judge said:

I would not agree with the judge that for recording purposes, “hate implies a high degree of animosity”. When Section 1.2.2 is read as whole, the statement is caveated, including by reference to the much broader CPS definition of hostility (see para 16 iv) above). The result is that the threshold of “hostility” for these purposes is at a much lower level. That aside, despite his obvious and strong reservations, the judge did not express a definitive view as to whether Mrs B’s complaint about the tweets should have been recorded as non-crime hate incident or not. Albeit the judge said it was not necessary for him to decide the point, had the matter had been a straightforward one to resolve I would have expected him to do so, because this case is about recording, and part of the relief Mr Miller had asked for was a declaration that the record made in January 2019 was wrongly made. In the event, this important issue from Mr Miller’s perspective, was never resolved and the record made of Mrs B’s complaint remains in place.

The judge considered that the apparent contradiction between the requirement to record on the one hand and the discretionary recourse to common-sense not to do so on the other, could be resolved by reference to what the Guidance says in Section 6.4, viz, where reference is made to an “overreaction”. On my reading of the Guidance, the relevant part of Section 6.4 is concerned not with recording, which must be done, but counsels the police to be proportionate in their response to an incident that has been recorded (“it is important that officers do not overreact to non-crime incidents…The circumstances of any incident dictate the correct response, but it must be compatible with section 6(1) of the Human Rights Act 1998.”) Though we are concerned with what the Guidance says rather than the intention of those who drafted it, I note that Mr Giannasi (who was the Hate Crime Advisor to the NPCC and heavily involved in the drafting of the Guidance) said in his evidence that the Guidance is not intended to guide police on their response to non-crime incidents (for example, as regards how they should deal with the “alleged perpetrator”); but that (by reference to Section 6.4) it is clear on the need for measured and proportionate responses.

The judge concluded that the Guidance was a proportionate interference with the right to freedom of expression having regard to the aims pursued (see [212] to [236]). The test adopted by the judge for the systemic challenge to the Guidance was whether it gave rise to an unacceptable risk of unlawfulness – which he said was not, without more, established by individual instances of an unlawful result. Thus, Mr Miller had to show that the Guidance creates a real risk of more than a minimal number of cases where Article 10(1) will be unlawfully infringed (see [216] to [219]).

The judge accepted as important and weighty the aims in Article 10(2) justifying any potential interference, namely the prevention of disorder or crime and the protection of the rights of others. On the evidence, the specific aims of the Guidance (which needed to be viewed in the context of the Macpherson Report) were preventing or taking steps to counter hate crime and hate incidents and building confidence in policing minority and marginalised communities [221]. The Guidance helped to achieve those aims because monitoring hate incidents helps inform police action to protect minorities and marginalised groups which in turn assists in building confidence in some communities (see [222] to [223]). So that, for example the evidence of Mr Giannasi is that low-level incidents are often pieces in a local jigsaw of information and intelligence that enables police to be aware of community tensions and take action to prevent minor issues or a series of minor issues escalating into something more serious. The Guidance also helps the police to fulfil their public sector equality duty under section 149 of the Equality Act in relation to protected groups that are particularly vulnerable and in need of protection [225].

The judge was satisfied that the aims and objectives of the Guidance justified the limitation it imposes on freedom of speech because its aims are extremely important; and as against that, the level of interference with freedom of expression is low [226]. Moreover, he said the Convention itself gives only limited protection to hate speech (properly so called), so even if Article 10 is not excluded by Article 17, any restriction upon genuinely hateful speech has been generally easier to justify, as necessary in a democratic society than other forms of speech [226]. He decided the recording of non-crime incidents was rationally connected to the objectives it serves [227]. It was important for the reasons identified by the defendants in their evidence, that the police have adequate records of potential hate incidents to inform their work; and the evidence is that recording of non-crime hate incidents is provided for by the Home Office’s National Standard for Incident Recording (NSIR). Further, no less intrusive measure could have been used without unacceptably compromising the achievement of the objective [228]. In this context the judge referred to two factors. First, “[t]he recording of non-hate crime incidents barely encroaches on freedom of expression, if it does so at all.” And second, “that key elements of the Guidance have been derived from sources which should command great respect and weight, and it can be concluded that they are what is thought necessary to achieve the Guidance’s aims: these included the Macpherson Report, ACPO Hate Crime Manuals” and Sir Adrian Fulford’s Race for Justice Taskforce Report of 2006.

Finally, the judge said the impact of the rights infringement was not disproportionate to the likely benefits brought by recording non-crime hate incidents. First, the mere recording of non-crime hate incidents barely impacts on the right to freedom of expression [230]. Set against that was the considerable evidence about both the necessity of the Guidance’s measures in relation to the benefits which they bring [231], and the safeguards that are in place as to how information is recorded and retained [233]. The safeguards he identified in the Guidance were first, an element of discretion as to whether to record - it has to be applied in a common-sense manner by police forces; secondly, the Guidance expressly provides that it must be applied in a proportionate and Convention-compliant manner [234]; thirdly, retention is subject to safeguards, e.g. the NSIR, the DPA 1998 [235] and fourthly, on disclosure, there is a framework of laws and policies in place, the legality of which has been upheld [236]. As to that, he said that disclosure is only permissible in principle where an individual may be in contact with vulnerable individuals, and because of the test of relevance, where those individuals may belong to the group against whom it is complained the applicant was hostile. “It is right that employers, who themselves must uphold their own equality duties in relation to their staff and service-users, may be informed about the potential prejudicial and discriminatory views of prospective employees.” The judge considered that there were important safeguards in place to protect job applicants, who had the right to request that information held about them be removed from the police’s record.

Mr Wise QC submits that the judge’s conclusions on proportionality were heavily based on his earlier flawed conclusion as to interference. Having discounted the extent of any intrusion on the right to freedom of expression, the judge did not then subject the Guidance to the requisite stringent analysis. In particular, he failed to grapple with the fact that the “hate incident” provisions of the Guidance give rise to an interference with freedom of expression on subjects of political controversy which requires strong justification. Further, the limited protection given by the Convention to hate speech was irrelevant; and the judge failed to distinguish between instances of hate speech, which it is accepted, would warrant recording, and legitimate contributions to matters of public debate. The question of whether less intrusive measures could have been taken, was, moreover, never properly addressed, as the judge manifestly failed to carry out this analysis for himself. As for the judge’s reliance on safeguards that are in place in relation to how information is recorded and retained under the Guidance, it is Mr Wise QC submits, impossible to see how the existence of such safeguards can provide a justification, let alone a convincing one for interference with speech that contributes to a public debate and cannot (on any view) be characterised as hate speech.

Mr Coppel QC submits the judge made no significant error of principle that would justify this court making a fresh determination of proportionality. Even if the threshold for a finding of interference is surmounted, the low level of that interference should play an important role in the proportionality exercise. The judge was right to find the interference is proportionate to the legitimate aims pursued by the Guidance. The aims are important and weighty ones, and the interference (if any) is at a low level and is outweighed by the importance of the aims. The judge’s reference to the limited protection afforded to hate speech by the Convention is entirely relevant, given that the Guidance concerns the recording of “hate incidents” and the incident in question in the present case is a form of speech. The judge was also justified in placing reliance on the views of the authors of the Macpherson Report etc., when considering whether less intrusive measures could have been taken; but did not defer to their views on the proportionality issue. Additionally, the judge’s reference to safeguards was plainly relevant to the type of information being stored under the Guidance. He was justified in relying on the authorities cited at [184] of the judgment, which concerned the safeguards relating to disclosure. (Footnote: ³) Crucially, the disclosure regime is subject to safeguards against arbitrariness in the form of the framework in the Police Act 1997 at ss 112-127, the Statutory Disclosure Guidance and the Quality Assurance Framework. In each case, the chief officer carries out an assessment of the facts and considers relevance to the purpose of the DBS check, and to an offence or risk of harm to the vulnerable. Individuals can also make representations before disclosure as explained in the Statutory Disclosure Guidance.

The question we have to address in the first instance is whether the assessment made by the judge was wrong, bearing in mind that this appeal consists of a review rather than a re-hearing.

I have already said that knowledge that the police are categorising and recording as non-crime hate speech, speech of the kind with which we are concerned, has the potential to create a chilling effect on public debate on issues of controversy and public importance in an area of expression where the scope for lawfully restricting such debate is very limited. As I have also said, in my view, the interference, while not at the highest end of the spectrum, is nonetheless real and significant. I note that when considering the level of interference, the judge had some regard to the low level of protection given by the Convention to hate speech. It is true that, broadly speaking, by virtue of Article 17 of the Convention (Footnote: ⁴) and its prohibition on the protection of acts aimed at the destruction of Convention rights, freedom of expression cannot be invoked in respect of statements directed against the Convention’s values. This case is not concerned, however, with the sort of extreme speech that might fall into that category. It is concerned with non-crime hate speech, the recording of which does not depend on any objective evidence of “hate” or meet the criteria for criminal investigation or prosecution, and which engages the protection of Article 10.

The judge’s view that the impugned parts of the Guidance did not encroach on freedom of expression, or if he was wrong about that, they did so barely at all, was an error of law that in my opinion skewed his approach at each stage of the analysis: see [226], [228] and [230]. This error is sufficient on its own to require us to re-do the proportionality assessment. With respect to the judge’s careful approach however, in my view there were two other errors in his assessment.

First, the court is obliged to consider for itself whether any less intrusive measure could have been used without unacceptably compromising the achievement of the legitimate aim. I accept of course that key parts of the Guidance derived from the views of “sources which should command great respect.” The factual case advanced to justify perception-based recording included statements for example from various members of the Government’s Independent Advisory Group on Hate Crime, a body of voluntary sector organisations with the key function of ensuring that the voices of victims, advocates and experts are heard, and which played an important role in the drafting of the Guidance. But this does not obviate the need for the judge to consider the issue for himself. It seems to me however that he failed to do this, probably because of his view on interference (there would be little scope for less intrusive means if there was no interference, or if there was, it was only at a low level). Secondly, whilst the judge was right to have regard to the framework of safeguards in place for information recorded and retained under the Guidance, he included two factors which in my view, for the reasons already explained, were not part of that framework, namely the broad discretion not to record and the application of proportionality to recording.

In considering afresh the balance to be struck between the Article 10 rights that are engaged and the interference with those rights, we have to keep in mind that this is a systemic challenge, made on the basis that the policy gives rise to an unacceptable or inherent risk of unlawfulness and where the focus of the inquiry must be on the Guidance itself. The broad test for systemic illegality was recently considered by the Supreme Court in R (on the application of A) v Secretary of State for the Home Department [2021] UKSC 37; [2021] 1 WLR 3931 at [38]-[48] (Footnote: ⁵) and in R (on the application of BF (Eritrea)) v Secretary of State for the Home Department [2021] UKSC 38. (Footnote: ⁶) In A v SSHD the Supreme Court set out the principles governing the test that should be applied when considering the lawfulness of policies. The standard of judicial review of a policy issued by a public authority derives from Gillick v West Norfolk and Wisbech Area Health Authority [1986] AC 112. Such a policy must not direct officials to act in a way which is contrary to their legal obligations. Guidance in a policy should not sanction or positively approve or encourage unlawful conduct: see A v SSHD at [38]. (Footnote: ⁷)

The question in this case can therefore be framed as follows. Does the Guidance sanction or positively approve or encourage unlawful conduct viz, conduct which violates Article 10? In my judgment it does.

If an interference with Article 10(1) rights is to be justified, it must meet the well-known four-part test identified in Bank Mellat v Her Majesty’s Treasury (No 2) [2014] AC 700. As I have said, there is no issue between the parties in respect of the first two parts of that test.

Perception-based recording has a legitimate aim (or series of aims) linked to the prevention of disorder or crime and the protection of the rights of others and this aim is sufficiently important to justify interfering with the fundamental right to freedom of expression. The evidence is that recording of non-crime incidents has been a feature of policing for many years and fulfils part of the core purpose of policing which is to prevent crime and protect citizens. Non-crime hate incidents may be an indicator of community tensions and of hostility which can escalate to more serious and criminal behaviour if unchallenged. Such incidents can damage the victims of them even if the behaviour is not criminal; and a failure to take them seriously can foster distrust in the police and a disinclination to report. Recording non-crime hate incidents allows the police to respond appropriately. This can be through monitoring, collecting intelligence, assessing the risk of escalation, mitigating risk or making informed decisions about the use of resources. Recording can also assist in relation to potential “course of conduct” offences, under the Protection of Harassment Act 1997 for example and provide evidence relevant to the motivation of those who go on to commit criminal offences. It is not difficult to see that there is a rational connection between the measure in question and the aims it seeks to achieve.

The difficulty arises so it seems to me when considering whether less intrusive means could have been used to achieve those legitimate aims without unacceptably compromising the achievement of them. In this case, whether the impact of the rights infringement is disproportionate to the likely benefit of the impugned measure seems to me overlap with the question whether less intrusive means could have been used, so the measure will be disproportionate if any more limited measure was capable of achieving the objective.

It is convenient to start with the breadth of the recording requirement.

Taking the Guidance at face value, the scope of the conduct that must be recorded by the police as a non-crime hate incident is extraordinarily broad, and deliberately so. This is not simply because a non-crime hate incident is one which is perceived by the victim or any other person to be motivated by hostility or prejudice. The threshold for hostility is low (it can include ill-will, ill-feeling, spite, contempt, prejudice, unfriendliness, antagonism, resentment and dislike). Further, the police are responsible for managing an interaction to ensure that the victim has no residual feelings of secondary victimisation, which can include feeling they have experienced indifference or rejection from the police when reporting a crime or incident. Once again, the issue is not to be grounded in an objective assessment of the evidence. As the Guidance also says in this context, it doesn’t matter whether there is victimisation or not, or even whether it is reasonable for the victim to feel they have been victimised.

The exceptions stipulated in the Guidance, relate to cases where the report is from someone who has no knowledge of the victim, crime or the area and who may be responding to media or internet stories or who are reporting for a political or similar motive; and where the complaint may be a malicious one against celebrities and political figures. Subject to these exceptions, the Guidance appears not merely to contemplate but to require that an incident must be recorded as a non-crime hate incident if the perception of the victim or any other person is that it is motivated for example by, spite or ill will against a protected strand, irrespective of whether there is evidence to support that perception or not. Thus, the Guidance contemplates on its face, the recording by the police of incidents as non-crime hate incidents, which are, to put it shortly, non-crime non-hate incidents.

Such issues do not arise in relation to those parts of the Guidance dealing with hate crime. This is for two reasons. First, in respect of the crime element, an incident can only be recorded as a hate crime incident, where the alleged actions of the perpetrator amount to a crime under normal crime recording rules. There is, in other words, an objective assessment of the evidence to determine whether a crime has been committed or not. Secondly, though the perception of the victim or any other person is still the defining factor in recognising the hostility element of a hate crime for recording purposes, there is a further evidential step contemplated in that connection. Before there can be a conviction for a hate crime, and an associated uplift in sentencing, it is necessary for the prosecution to prove to the criminal standard that the offender has either demonstrated hostility based on race, religion, disability, sexual orientation or transgender identity or that the crime has been motivated by such hostility: see the Crime and Disorder Act 1998 and section 66 of the Sentencing Act 2020.

In defending perception-based recording, the respondent naturally enough has placed enormous weight on the recommendations for perception-based recording made in the Macpherson Report and their continuing critical importance for the role of the police in the protection of vulnerable members of the community as outlined above. All that is understood. However, the recommendations for the recording of non-crime hate incidents at the time that the Macpherson Report was produced in 1999, confined as they were to racial incidents, were part of a suite of recommendations, one of which was an equally strong recommendation that criminal and non-criminal hate incidents should be investigated with equal commitment: see recommendation 13, cited at para 10 above. Since those recommendations were made however, the internet and social media has profoundly changed how people communicate and infinitely increased their ability to use and abuse those means of communication. As might be expected therefore, and as Mr Giannasi has pointed out in his evidence, the sheer volume of abusive messages on social media means it is impossible for the police to investigate and make decisions on the motives of everyone before deciding whether to record such incidents, unless policing and its resources changed fundamentally. Indeed Mr Giannasi gives this as one of the reasons why perception-based recording is so important. (Footnote: ⁸)

The Guidance applies without distinction to the protected strands identified within it (disability, race, religion, sexual orientation and transgender). But the prism provided by this case has been the public debate concerning transgender issues and the adverse impact that the Guidance has on freedom of expression in that context, specifically, the chilling effect that it has on an important issue of legitimate public interest. As to that, the judge found that a distinction could be drawn between gender critical views held for proper socio-philosophical reasons on the one hand and views that are grounded in prejudice and hostility on the other. See further, the judge’s discussion of the evidence of Professor Stock and Ms Ginsberg at [240] – [249]. At [249] for example, the judge cited the following passages from Ms Ginsberg’s evidence:

The judge’s conclusion at [250] was this:

The position in the round can therefore can be encapsulated as follows. The net for non-crime hate speech is an exceptionally wide one which is designed to capture speech which is perceived to be motivated by hostility against one of the protected strands, regardless of whether there is evidence that the speech is motivated by such hostility. The volume of non-crime hate incidents is enormous, and the police do not have the resources or the capacity to investigate all the complaints that are made. It is reasonable to suppose therefore that most will remain uninvestigated. There is in any event very little that can be filtered out of that net, subject to specified limited exceptions. In particular, there is nothing in the Guidance about excluding irrational complaints, including those where there is no evidence of hostility, and little, if anything to address the chilling effect which this may have on the legitimate exercise of freedom to expression. Even so, where the perception of the complainant is that speech is motivated by hostility towards one of the protected strands, the Guidance says it must be categorised as a non-crime hate incident; and the language used (of a non-crime hate incident and a victim) is capable of unfairly stigmatising those against whom such a complaint is made. There is no provision for proportionality to be applied to recording. And the Guidance says nothing about the language to be used in any such record, or whether someone should be notified that a record, flagged as a hate incident has been made of a complaint against them, leaving such issues to individual forces to decide.

There are, as the judge said, various safeguards in the legal framework relating to the retention, recording and disclosure of information by the police as described at paras 19 to 25 above, which apply to the records of non-crime hate incidents that are made. Their principal objective, however, seems to be balancing the Article 8 (Footnote: ⁹) rights that are engaged when the police collect and use personal data rather than the implications this may have on the right to freedom of expression protected by Article 10.

The respondent has argued that there is nothing wrong in legal terms with the Guidance, but that in any event, the Revised Guidance described at para 12 above) is “Even clearer” than the Guidance: it warns against disproportionate action by the police, it provides a link to the judgment of Julian Knowles J, and it contains detailed guidance as to how the police should contact people accused of non-crime hate incidents.

The Revised Guidance says:

Under the heading “Responding to hate”, the Revised Guidance says: “Note: The terms ‘victim’ and ‘suspect’ are used throughout this Authorised Professional Practice (APP) to refer to the person reporting an allegation and to the alleged perpetrator. These terms do not mean that a crime has been reported or that an investigation into a crime is taking place.” Under the heading “Responding to non-crime hate incidents”, it says: “There may be an overlap between a perceived non-crime hate incident and the legitimate exercise of rights and freedoms conferred by the Human Rights Act 1998.” Under the heading “Recording non-crime hate incidents”, the Revised Guidance says “The recording system for local recording of non-crime hate incidents varies according to local force policy. Managers should have confidence that all incidents are being recorded correctly. See data recording for further information on how information should be managed.”

These revisions (with their greater emphasis on proportionality) appear to be designed to meet the criticisms made of the police conduct in this case, and it could be said that at the very least they demonstrate there was scope for less intrusive measures. But in my opinion they do not go very far, or not nearly far enough to address the chilling effect of perception-based recording more generally. The position is therefore that less intrusive measures could be used to achieve the legitimate aims of such recording, without unacceptably compromising the achievement of those aims. That is not to say that perception-based recording of non-crime incidents is per se unlawful, but that some additional safeguards should be put in place so that the incursion into freedom of expression is no more than is strictly necessary.

It is not for this Court to attempt any redraft since, subject to the views of my Lord and my Lady as to the outcome of this appeal, that is a matter for the College. It is to be noted however that though the judge said the police have a common-sense discretion not to record irrational complaints and the police say they exercise such a discretion, nothing is said about this in the Revised Guidance. This may be because of the tensions between perception-based recording on the one hand and a discretion not to record on the other which I have discussed at paras 89 and 90 above. Providing a link to the lengthy judgment below merely ducks this issue it seems to me. The Guidance should truly reflect what the police are expected to do and should not mislead by omission either the police who have to use it or the public. I do not think the tension is an impossible one to resolve. The Guidance and the revised version already provide for limited exceptions which demonstrate that a derogation from perception-based recording of non-crime hate incidents can operate where the principle of perception-based recording is abused. And as I say, as the position of the College is that such a common-sense discretion exists and is exercised, no doubt a Guidance which better reflects in suitable terms what the College says is in fact the position, is achievable.

For the reasons given, I would allow this appeal under Ground 3 and 5.

I agree.

I also agree.