The Electronic Collar Manufacturers Association & Anor, R (On the Application Of) v The Secretary of State for the Environment, Food And Rural Affairs

This is an appeal from a decision of Morris J (‘the Judge’). The Judge dismissed an application for judicial review by the Appellants (‘the As’) of a decision of the Respondent (‘the Secretary of State’) made on 27 August 2018 ‘to implement a ban on the use of remote-controlled hand-held electronic collar devices for cats and dogs in England’ (‘the Decision’). I will refer to those devices as ‘e-collars’, although they are called ‘shock collars’ in some of the Respondent’s documents. The mechanism for such a ban would be a statutory instrument made under section 12 of the Animal Welfare Act 2006 (‘the Act’). No such instrument has yet been made.

On this appeal, the As were represented by Mr Wise QC and Ms Proud. Mr Turney represented the Secretary of State. I thank counsel for their written and oral arguments.

There were five grounds in the claim form.

After a hearing lasting five days, the Judge dismissed all the As’ grounds.

There are, in effect, two grounds of appeal, each with two aspects.

The abandonment on this appeal by the As of their first three original grounds of challenge means that there is no dispute on this appeal about three points.

Davis LJ gave permission to appeal ‘on all grounds’. His first observation was ‘It is not arguable, or argued, that a decision to ban e-collars is itself irrational’. The question is whether the Secretary of State’s reasoning in arriving at such a conclusion to ban e-collars is (arguably) demonstrably so flawed as to vitiate that conclusion. He said that he was persuaded that ground 1 was ‘realistically arguable particularly in the circumstances of (1) the volte face between January 2018 and February 2018 (the rationale for which seems to be obscurely evidenced or explained) and (2) containment systems not being prohibited’. He noted that it is not irrational to change one’s mind, and that it might be that e-collars and containment systems are not identical. The combination of those two factors and the Judge’s view that the issue was not straightforward meant that permission should be given. He considered that it was also appropriate, in the circumstances, to give permission on grounds 2 and 3.

The Secretary of State seeks to uphold the Judge’s decision on A1P1 on additional grounds. These, in short, are that the Decision did not, in accordance with the approach of the European Court of Human Rights (‘the ECtHR’), amount to an interference with the As’ possessions. The relevant legislation

The short title of the Act describes it as an Act ‘to make provision about animal welfare; and for connected purposes’. Section 4 is headed ‘Unnecessary suffering’. It makes it an offence to cause a protected animal, or to allow a protected animal to be caused, ‘unnecessary suffering’. Section 4(3) lists the considerations which are relevant to a decision whether suffering is unnecessary. They include whether ‘the conduct which caused the suffering was in compliance with any relevant enactment or any relevant provisions of a licence or code of practice issued under an enactment’ (section 4(3)(b)). Section 9(1) imposes a duty on a person who is responsible for an animal to ensure its welfare, by making it an offence not to take such steps as are reasonable in the circumstances to ensure that its needs are met to the extent required by good practice. The relevant needs are listed in section 9(2). They include needs for a ‘suitable environment’, ‘to be able to exhibit normal behaviour patterns’ and ‘to be protected from pain, suffering, injury and disease’.

Section 12 is entitled ‘Regulations to promote welfare’. Section 12(1) gives ‘the appropriate national authority’ (in England, the Secretary of State) a wide power to ‘make such provision as the authority thinks fit for the purpose of promoting the welfare of animals…’ Regulations under section 12(1) can make provision ‘imposing specific requirements for the purpose of securing that the needs of animals are met’. That power to make regulations includes power to ‘provide that breach of a provision of the regulations is an offence’ (section 12(3)(a)). Before making regulations under section 12(1), the authority must consult ‘such persons appearing to the authority to represent any interests concerned as the authority considers appropriate’ (section 12(6)).

This summary of the facts is based on the Judge’s judgment (‘the judgment’). The first Appellant (‘the first A’) is an unincorporated trade association. It represents those who make and sell e-collars and other electronic training aids for dogs in the United Kingdom. The second Appellant (‘the second A’) makes and sells e-collars. Its parent company is based in the United States of America. The different types of device

There are three relevant types of device, each of which uses an e-collar, which is put round the animal’s neck. They are

In the first case, the owner, who can be at some distance from the animal, uses a handheld device which makes the e-collar give an electric shock to the animal or spray an unpleasant substance at it. An owner can use a hand-held device to deter unwanted behaviour, such as chasing cows or sheep, or fighting with other animals. The owner can choose when, and how often, to give the animal an electric shock. The owner can also choose to vary the strength of the shock. Some brands of e-collar make a warning sound before the shock (or spray) which gives the animal a chance to stop doing what it is doing and thus to avoid the deterrent.

The second type of arrangement, a containment system, is used to keep an animal in a defined space, such as a garden, so as to stop it straying onto a busy road, or defecating in a neighbour’s garden. In this second case, if the animal approaches a cable which is buried in the ground on the boundary of the defined space, the e-collar automatically gives the animal an electric shock, or sprays an unpleasant substance at it.

The third device, a bark collar, detects vibration from a dog’s voice box when the dog barks. It then automatically gives the dog an electric shock or applies some other deterrent to the dog, such as an unpleasant spray.

The principal relevant difference between an e-collar remotely controlled by a user with a hand-held device, on the one hand, and a containment system, and a bark-collar on the other, is that the first is triggered by, and at the option of, the owner, whereas the second and third are triggered automatically by an animal’s behaviour. The only human intervention in those two cases is the setting up, and calibration, of the system.

The first A’s evidence was given by Ms Critchley and Mr Penrith in their witness statements, and summarised by the Judge in paragraphs 31 and 32 of the judgment. The first A’s evidence was that over 500,000 people in the United Kingdom own and use, or have used, either an e-collar or an electronic containment system to train dogs or to stop them barking. They have been banned in Germany, Wales and Denmark. There is no consensus among ‘Western developed nations’ about banning them, or about whether a ban is an effective or sensible way of regulating their use. The first A strongly supports the regulation of e-collars which are used for training dogs. It has worked closely with the authorities in the Netherlands and the State of Victoria to devise systems for regulating their use, and with the Chief Veterinary Officer for Scotland (‘the CVO’) to devise options for regulation. I note from the Kennel Club’s response to the consultation that e-collars are also banned in Denmark, Norway, Sweden, Austria, Switzerland, Slovenia and in some territories of Australia.

Ms Critchley’s evidence was that most dogs can be trained by positive reinforcement, so that e-collars are unnecessary. A significant minority of dogs have intractable behavioural problems which cannot be changed by a system of rewards. If e-collars are banned, such dogs will either have to be put down, or kept closely confined, with adverse effects on their welfare. The Secretary of State had not taken into account that if e-collars are banned, owners may use crueller ways of controlling their dogs’ behaviour, such as choke collars or prong collars. Those could be more harmful to dogs than e-collars.

Mr Penrith trains dogs, particularly badly behaved dogs. He is very experienced. He had worked with police dogs for ten years. The police use choke collars. His view was that choke collars are worse for dogs than e-collars. He had found an e-collar very useful in 2011 when training his own dog. He used it to suppress the dog’s innate ‘predatory’ behaviour. E-collars are the only thing which works with some dogs. He trains between 100-150 dogs a year with them. He supports regulation of their use.

As the Judge recorded, these devices have been contentious for some years. Their use was not controlled until the Code of Practice for the Welfare of Dogs (‘the Code’) was promulgated in 2010. Animal welfare groups have campaigned for e-collars to be banned, on the grounds that they are cruel and unnecessary. There has been research about their effects. The conflict of views between the first A and the welfare organisations led the Secretary of State to decide, in 2007, to commission the University of Lincoln to research the effects of e-collars. The Lincoln research

That research was described by the Judge in paragraphs 30 and 88-90 of the judgment. There were two projects which ran from September 2007 to November 2010, and from October 2010 to June 2011. Reports about both were published on 10 June 2013. The Judge referred to the two reports as ‘Lincoln 1’ and ‘Lincoln 2’, respectively. The Judge also described, at paragraph 91, a summary by the Lincoln researchers of their work, and, in paragraph 92, a report by the Companion Animal Welfare Council (‘the CAWC’), which was published in 2012.

Lincoln 1 described the variables which are in play when an e-collar is used (the individual characteristics of different dogs, shock strength and the differences between different types of e-collar). These meant that the effects of e-collar on different dogs varied a good deal. There were significant differences, however, between dogs on which e-collars were used, and those which were trained using positive methods. The former group had ‘higher levels of physiological and behavioural arousal’. At least a proportion of those dogs experienced ‘negative emotions’. This led the researchers to suggest that ‘the use of e-collars in training pet dogs leads to a negative impact on welfare, at least in a proportion of dogs…’

Lincoln 2 described Lincoln 1 as uncovering evidence ‘that experience of e-collars had long-term negative welfare consequence in some dogs…’, but as not showing that there was a causal relationship. Lincoln 2 sought to fill that gap, and to examine recommended practice for using e-collars, and to contrast such practice with positive training methods. Lincoln 2 noted that the owners recruited for Lincoln 1 reported that their use of e-collars varied a good deal and that they did not always follow best practice as recommended by the first A. Most had not had any formal training, and instruction manuals varied considerably (judgment, paragraph 89).

Lincoln 2 said that further research about the welfare consequences of the different levels of skill of those using e-collars would be useful. Lincoln 2 nevertheless found that there was behavioural evidence that e-collars had a negative impact on the welfare of some dogs, even when the people using them were professional trainers ‘using relatively benign training programmes advised by e-collar advocates’. Lincoln 1 concluded that even if owners followed recommended best practice, some dogs would have ‘more negative emotional states (including anxiety and aversion)…’. Those effects lasted while the dogs were being trained. There was some evidence of elevated arousal if the dogs were, later, subjected to more training. One of the most common justifications for using e-collars was to deter dogs from chasing livestock. There were no statistically significant or clinically relevant differences between the effectiveness of positive and negative training to correct that behaviour.

The Judge also referred, in paragraphs 69 and 197 of the judgment, to research done by Lincoln University in 2016, which showed that there was no evidence of long-term ill effects on cats which had been exposed to containment systems. That research was mentioned in paragraphs 43, 53 and 56 of, and footnotes 4 and 6 to, Mr Casale’s first witness statement. Mr Casale, who is a Senior Civil Servant who made three witness statements for the Secretary of State. The 2016 research is also referred to in the ministerial submission (which I describe more fully in paragraphs 48-59, below), in the discussion of the pros and cons of option 2, the option which the Secretary of State adopted.

On 3 September 2014, the authors of Lincoln 1 and 2 summarised their research in a journal called PLOS One. The research was said to show that training with an e-collar caused behavioural signs of distress in dogs, particularly when they were used at high settings. Following the manufacturers’ best practice mitigated signs of poor welfare detected in Lincoln 1, but behavioural differences were consistent with the view that animals trained with e-collars had a more negative experience than those trained by other methods. Training with an e-collar did not result in ‘a substantially superior response’ to the response seen from training using positive methods. It seemed that ‘routine use of e-collars even in accordance with best practice (as suggested by collar manufacturers) presents a risk to the well-being of pet dogs. The scale of this risk would be expected to be increased when practice falls outside of this ideal’.

The CAWC report was published in June 2012. It was a review of the results of a number of other published studies. More research was necessary. The research so far did not support broad conclusions about the impact of e-collars on the long-term welfare of dogs, if they were used properly. A broader, ethical view should be taken, since it was not possible to advance evidence-based arguments against or in favour of the use of e-collars. There was, however, an unnecessary risk to animal welfare if the use of e-collars was not regulated as e-collars could cause harm if the devices did not have safety features, or were used by less competent trainers. The report made suggestions for the regulation of the manufacture and use of e-collars, if the Government supported their continued use. The effect of a ban on the As

In paragraphs 93-95 of the judgment, the Judge considered the evidence about the effect which a ban would have on manufacturers and suppliers. The Department’s Regulatory Triage Assessment (‘RTA’) estimated that the business impact of a ban would be a loss of profits of about £800,000 a year. That was considered to be an over-estimate, as businesses could focus their efforts, instead, on other markets, and any loss of business might be offset by increased demand for less punitive devices.

Ms Critchley, for the first A, estimated that the loss of the market in England would have an impact on sales of £1-2m a year, although that figure included containment systems. The long-term adverse economic effects of a ban on the reputation of the products in the market and on sales in the United Kingdom and elsewhere could last for many years. The evidence of Mr Stone (who is a solicitor acting for the As), was that in the relevant financial year after the announcement of the proposed ban, the second A’s sales had gone down by £51,000 compared with the previous year. A French company which belonged to the first A had lost about £10,000 in sales since the publication of the consultation document. A third member reported a dip in sales of £500,000 since that date. The ban in Wales

The Judge considered the position in Wales in paragraphs 96-102 of the judgment. In 2010, the Welsh Ministers made regulations under section 12 of the Act banning any e-collar which gave an electric shock to an animal. It was known that the Lincoln research was being done, but at that stage neither Lincoln 1 nor Lincoln 2 had been published. The As applied for judicial review of those regulations in 2010. Beatson J (as he then was) upheld the ban (R (Petsafe Limited and ECMA) v Welsh Ministers [2010] EWHC 2908 (Admin) (‘Petsafe’). The Judge noted, in particular, Beatson J’s conclusions that the option of training and licensing did not address concerns that the shock caused pain and that e-collars could cause pain and distress if used by people who were not trained and that it was open to legislators to form their own judgment about whether e-collars caused enough suffering to justify a ban, because it was for the Ministers to set the level at which animals should be protected. The ‘soft’ ban in Scotland

In paragraphs 103-106, the Judge considered the position in Scotland. The Scottish Government consulted on four options, including a ban, in 2015-2016. It considered that the Lincoln research did not show clearly that e-collars were inherently harmful

In 2013, the Secretary of State’s position, described in an email dated 10 June 2013, was that while the Lincoln research showed no evidence that e-collars caused longterm harm to dogs, the Secretary of State wanted to ensure that they were used appropriately and made to a high standard. The Secretary of State wanted to work with the first A to produce guidance for dog owners and trainers about how to use ecollars properly and a manufacturers’ charter to ensure that e-collars were made to high standards. The Secretary of State considered that a ban could not be justified because ‘the research provided no evidence that e-collars pose a significant risk to dog welfare. For a ban to be introduced, there would have to be evidence showing that they were harmful to the long-term welfare for dogs’.

In paragraph 34 of the judgment, the Judge summarised this position as ‘evidence of harm to animal welfare is required to justify a ban on e-collars, but that Lincoln 2 did not provide such evidence’.

In August 2017, the Secretary of State decided that the conclusions of the Lincoln research did not warrant a ban on e-collars. The Code was amended, instead. The Judge quoted the relevant passages of the amended Code in paragraph 35 of the judgment. The Code said that positive reinforcement was the ‘preferred’ method of training dogs and that training which involved punishment ‘may cause pain, suffering and distress. These techniques can compromise dog welfare, lead to aggressive responses and worsen the problems they aim to address’. The Judge recorded that this form of words had been agreed with the first A and animal welfare organisations. I infer from paragraph 35 of the judgment that the language was toned down at the request of the first A.

On 5 February 2018, the Secretary of State wrote to various interested parties describing the Government’s position on e-collars. The Government knew that there were strong feelings on the subject, but ‘the Government would need to be satisfied that such a ban was in the public interest and could be supported from an animal welfare point of view’. The Lincoln research was evidence that ‘electronic aids can have a negative impact on the welfare of some dogs, but not all’. That evidence was said to be ‘not strong enough to support a ban’. The claim that they were no more effective than other training methods was not in itself a reason to ban them or to restrict their use. The Government’s advice was that electronic training aids ‘should only be used as a last resort and on the recommendation of a professional…and should only be used by competent operators’.

The Judge identified three features of the Secretary of State’s position in early February 2018.

The evidence before the Judge showed that shortly after that letter, the Secretary of State’s position changed. ‘In February 2018 Ministers indicated that they wished to consider a ban on e-collars and sought the advice of officials’. The evidence also showed that officials very quickly responded to that request with four options, which were described by officials on 21 February 2018. Ministers chose the second option (the promulgation of a statutory instrument banning the use of e-collars). Officials then asked for, and got, the approval of the relevant Cabinet Sub-committee to consult on a proposed ban. On instructions, counsel for the Secretary of State told the Judge that the email exchanges by which Ministers had communicated their wish to consider a ban on e-collars contained no reasons for that apparent change of heart. On the basis of that assurance, the Judge refused an application by the As for the disclosure of those emails. Mr Casale made a second witness statement confirming that.

When the hearing resumed after an adjournment (between 24 May and 10 June 2019), Mr Casale had made a third witness statement, in which he tried to explain the apparent change of heart. Mr Turney told us in his oral submissions that the Judge asked for this witness statement. Mr Turney’s position was that Mr Casale’s witness statement was made in order to comply with the Judge’s request, but that it was, nevertheless, unnecessary, and that he had made that submission to the Judge.

Mr Casale said, in short, that the research evidence had not changed, but the attitudes of society on the subject might change, so that society was no longer willing to tolerate a risk of harm which it had once tolerated. ‘It is legitimate for views on the desirable policy which relates to that research to evolve over time with the changing views and attitudes of society, and for consultation to inform what, if any, changes to the policy should be’. Mr Casale explained that the Lincoln research had identified ‘some adverse effects on the welfare of dogs from the use of electronic collars in some circumstances’. The initial position of Ministers was that this evidence did not support a ban. Having taken into account the current views of society about animal welfare, they considered that there was a case for a change in policy, subject to the outcome of the consultation. The letter of 5 February 2018 reflected the Department’s policy on e-collars when it was written; the decision to consider a change in approach ‘followed short afterwards’. The Judge summarised the effect of that evidence in paragraph 41 of the judgment.

The Decision was preceded by a consultation between 12 March 2018 and 27 April 2018, announced, and described, in a five-page consultation document entitled ‘A ban on electronic training collars for cats and dogs in England: March 2018’ (‘the consultation document’). The consultation document was published on 11 March 2018. The Judge summarised the consultation document in paragraphs 42-48 of the judgment. It did not refer to bark collars. It described the position in Scotland (it was then proposed that the guidance there be amended to highlight the welfare effects of e-collars, and that it clarified what offences might be committed by using them). There was a ban in Wales. It said that ‘In the light of growing concerns regarding use and potential misuse of e-collars in England, and in order to protect the welfare of cats and dogs, we wish to ban their use here…’ It was proposed that it should be an offence to use an e-collar, to put one on an animal, or to be in charge of an animal which was wearing one. ‘This will dry up sales of these punitive devices’.

The consultation document referred to the Lincoln research which was said to show that e-collars could have ‘detrimental welfare effects on dogs and can cause harm and suffering’. It also showed that many owners did not read the manufacturer’s instructions before using them. Many animal welfare organisations and trainers opposed them not just because of the harm they caused, but because they are ‘a negative form of training’. Positive methods should be used, rather than punitive methods which did not necessarily work and which could be counter-productive, by making a dog show other problematic or dangerous behaviour. The consultation document recorded the view that e-collars could be a last resort for badly-behaved dogs which would otherwise have to be put down. It said that ‘[r]elatively little evidence’ had been provided in support of that view, ‘although the evidence about the harm which e-collars inflict on pets has been growing’.

Having weighed up the evidence of their impact and taking into account public concerns that ‘we should treat all our pet animals with appropriate reward and respect’, the Department had decided that e-collars should now be banned, consistently with the position in Wales. The consultation document recorded an argument by manufacturers that when used correctly, e-collars made it possible to train difficult dogs which do not respond to positive training. They also argued that containment systems could keep animals in a defined area, where other methods were not possible, which would keep them safe, and prevent ‘unwanted behaviours…in other places’. The consultation document considered the effect on manufacturers. It was anticipated that owners would use other methods of managing the behaviour of their pets. Businesses which made and sold e-collars ‘may experience a reduction in profits’. But a rise in demand for ‘other pet training aids and implements’ was also anticipated. Manufacturers and suppliers might decide to focus on other markets where e-collars could still be used.

The consultation document invited responses on the proposed ban. Consultees were invited to comment on the ‘expected impacts’ of the proposed ban. The on-line response facility allowed responses to nine questions, by tick box and with the opportunity to use free text boxes. The responses to the consultation

There were 7334 responses, including detailed representations from the first A. Ms Critchley explained that the As had not been sounded out on the proposed ban before the publication of the consultation document. In the past, by contrast, the Department

In paragraphs 52-61 of the judgment, the Judge summarised three responses opposing the proposed ban (from the Countryside Alliance, Ms Caroline Furze, and the first A:

The first A recommended an expert analysis of the comparative impacts of the various proposals for reform and regulation. The first A explained which issues should be considered in more depth (see paragraph 58 of the judgment). The first A supported ‘sensible regulation’ but was concerned about the way in which a ban was being proposed. It was not supported by any proper analysis. The first A referred to successful regulation in other jurisdictions, for example the Netherlands. The first A also objected to the short period for responses and to inaccuracies in the consultation paper. The first A’s representations about regulation occupied eight detailed pages of that document. The first A said that there was an ‘integrated model of regulation’ in the Netherlands. The Government ‘turned their minds’ to this topic in 2010. A ministerial decree was promulgated in 2016 but was not due for ratification until some time in 2018. It provided for a ban on e-collars, with a wide range of exemptions. That approach was ‘broadly consistent’ with what were then understood to be the principles of the proposed Scottish model. I observe that it is self-evident that the first A could not rely on any evidence of the effectiveness of either model, since neither had been put into practice. The only regulatory system to which the first A referred which had been in operation was the system in the State of Victoria. The first A said that that had been operating ‘successfully since 2008’.

The Department had meetings with consultees after the end of the consultation period but before the consultation responses had been fully analysed. Some consultees made further written representations during that time. The Judge described those in paragraphs 65-68 of the judgment, and the analysis of the consultation responses, in paragraph 69. That analysis was evidently elaborate. It took four months. The ministerial submission

The ministerial submission (‘the submission’) is dated 3 August 2018. It is addressed to Lord Gardiner and to the Secretary of State. There were three annexes: annex A, which described the pros and cons of four options, annex B, which was a draft letter to the Social Reform (Home Affairs) sub-Committee, and annex C, which I refer to in more detail, below, at paragraphs 62-64, described as a ‘draft consultation response document’ (‘the response document’).

The summary referred briefly to the positions in Wales and in Scotland. The consultation showed that animal welfare groups were strongly in favour of a ban, whereas the majority of (individual) respondents was not. This reflected, in part, the difference in views about banning hand-held training systems and banning containment systems. There was a strong lobby in favour of allowing containment systems, especially for cats. Many felt that these did not give rise to the same risks to animal welfare as hand-held devices, and, at the same time, they avoided risks of harm by preventing straying. The providers of containment systems had separately shown their systems to officials, and stressed how keen they were to provide ‘a full service which involves training and ensures systems are set up properly for the pet and premises. Such providers whose business model involves providing a full service account for almost all market share’.

The submission recorded that officials considered, ‘on balance…that a more proportionate approach would be to ban hand-held devices’ and to update the Code to say that if they are needed to prevent risks of harm, containment systems should be installed by professionals, with suitable training to minimise risks to animal welfare. A ban on hand-held devices would be easier to take through Parliament.

Officials recommended that the Secretary of State sign the attached letter (I infer that this means the letter in Annex B) and approve the response document. The RTA was said to confirm that ‘business costs are low and that a full impact assessment was not required’. It is not clear to me that the RTA was attached to the submission, although the Judge (see paragraph 80 of the judgment) thought that it was. I say a little more about the RTA in paragraph 28, above.

Under the heading ‘Discussion’, the submission accurately summarised the consultation, in a passage quoted by the Judge in paragraph 70 of the judgment. Officials noted that most of those who opposed the ban asked why the Government had changed its stance, ‘given that no new major evidence to support a ban has emerged recently’.

Paragraph 3 of the submission referred to the summary of the options in Annex A, and explained officials’ recommendation. The four options are listed in paragraph 71 of the judgment. The Judge summarised officials’ assessment of the four options in paragraphs 72-75 of the judgment.

The Judge quoted officials’ conclusions in paragraph 76. In short, officials considered that there was an animal welfare justification for banning hand-held devices. There was enough evidence that they could cause ‘avoidable pain and suffering especially when not used properly’. There were positive training methods which could produce the same, or better, results. There were ethical reasons for banning hand-held devices, but any legislative ban under the Act ‘would need to be based solely on animal welfare grounds’. Officials considered that the animal welfare justification for banning containment systems was ‘less clear’. They were usually installed by professionals for the site in question. The service from installers often included training, minimising the shocks during training, so that, if the system worked as intended, the animal would avoid shocks in future. Containment systems could avoid

The views of stakeholders were summarised in the next section of the submission. Those were the RSPCA, the Kennel Club, the first A, Feline Friends, an individual whose name is redacted and who is described as a dog trainer (I infer it is Mr Penrith), and another entity (the identity of which is also redacted). The manufacturers are said, ‘unsurprisingly’, to oppose a ban. The first A indicated that a ban would be likely to result in a legal challenge. A lobby group which opposed a ban on containment systems had emerged. Officials repeated that the providers whose business model included providing a full service accounted for almost the whole market.

The submission then described the position in Wales and in Scotland.

Almost a page of the submission concerns communications. Officials suggested that if Ministers chose option 2, that decision be framed as it was in an answer the Secretary of State had given to a question in Parliament. ‘This would demonstrate that the government has listened carefully to stakeholders [sic] concerns on invisible fences, but remains committed to banning the remote control devices which can be abused’.

The Secretary of State adopted option 2. As described in the submission, the pros of this option included that it was supported by Lincoln 1 and 2, which showed that ecollars compromised the welfare of some dogs. Option 2 nevertheless allowed dogs and cats to be kept in a specific space, without physical fencing, which might be inappropriate. Many attacks on livestock were caused by the escape of dogs from poorly maintained fences. Invisible fencing can be more effective. It was also supported by recent research by Lincoln University which showed that containment systems did not have a negative impact on cats. The cons of this option were that it would not have unqualified support from ‘the animal welfare and veterinary group lobby’. It removed a way of controlling dogs and potentially of stopping ‘livestock attacks in motion’. There was a risk of legal challenge from the first A. It needed to be handled carefully, ‘to ensure that the arguments we deploy to distinguish containment systems do not undermine the welfare arguments needed to justify the ban’ on ecollars.

Officials also summarised the pros and cons of regulation instead of a ban (option 4). The pros were that it aimed to tackle the identified problem of improper use of ecollars, enabled the continued use of e-collars to control wayward dogs (eg around livestock), enabled containment systems to be retained, with their advantages, and was supported by the first A and by the Countryside Alliance. The cons were that it might not have much support from animal welfare or veterinary stakeholders, it might be difficult to get regulations through Parliament, a new regulatory system might be burdensome, and might be hard to apply if hand-held devices were got from abroad, new regulation for containment systems might add little to current robust selfregulation, a new regulatory system might be unpopular if owners had to apply for licences, and it was unclear how a licensing system for owners would be enforced and would operate in practice.

On 27 August 2018 the Secretary of State published a press release and a document entitled ‘Electronic training collars for cats and dogs in England: Summary of Responses and Government Response – August 2018’, that is, the response document.

The Judge quoted from the press release in paragraph 81 of the judgment. Its headline was ‘Cruel electric shock collars for pets to be banned’. The press release described how e-collars work. It said, ‘As well as being misused to inflict unnecessary harm and suffering, there is also evidence e-collars can re-direct aggression or generate anxietybased behaviour in pets – making underlying behaviour and health problems worse’. It added that, ‘…after listening closely to the views of pet owners and respondents, the Government will not extend the ban to invisible fencing systems which can keep pets away from roads and potential traffic accidents’. The press release then quoted the Secretary of State’s statement in Parliament.

The Judge summarised the response document at paragraphs 82-86 of the judgment. The Judge said that the response document did not analyse the content of the responses, but provided a quantitative summary of the various types of response. For example, it said that 10.5% of respondents pointed out that unintended consequences of a ban could be more dogs having to be put down and an increase in attacks by dogs. The document also summarised the position of various stakeholders. The document recorded ‘points made by consultees favouring regulation, as opposed to a ban’.

In paragraph 87, the Judge noted that ‘The Decision itself is then recorded in the final section, under the heading “Government response”’. He quoted the whole of that part of the response document, apart from the last paragraph. This part of the document announced the proposed ban. It recorded that the Government accepted that ‘where this is necessary as a last resort, to prevent other serious risk of harm, there is an argument for retaining the ability to use invisible fencing containment systems for cats and dogs subject to them being set up and used properly…’ The Decision was said to be based on ‘the concern that hand-held remote controlled devices can be all too easily open to abuse and can be harmful for animal welfare’. The response document referred to the Lincoln research which was said to show that ‘many users of hand-held devices were not using then properly in compliance with the manufacturer’s instructions leading to welfare problems for the dogs’. It said that ‘In many cases alternative positive reward training can be used to encourage and correct a dog’s behaviour’. By contrast, it was possible to minimise the ‘adverse animal welfare impact’ of containment systems, if they were installed and set up properly, and if proper training was provided; and ‘at the same time, these systems can avert other risks to animal welfare’, such as preventing them from straying into potential danger. With proper training, animals would not be exposed to regular electric shocks.

I would add that the response document begins with a background section, which explains what the devices are and how they work. There is also a section headed ‘Welfare effects of e-collars’. This refers to the Lincoln research, which is said to show that ‘e-collars can have a detrimental welfare effect on dogs and can cause harm and suffering’. The research also showed that many owners did not read the manufacturer’s instructions before using e-collars. Many oppose e-collars both because they cause harm and because they are negative forms of training; their view is

Mr Casale explained, in paragraphs 48-59 of his first witness statement, that once officials had analysed the consultation response, they drafted the response document. They then sent it to Ministers with the submission. Mr Casale summarised the submission. In paragraph 58, he said ‘Ministers responded to the submission deciding that Option 2 should be implemented.’. The Judge does not refer to this evidence in the judgment, but I infer that he must have had it in mind, because the Judge’s description of the analysis of the consultation responses in paragraph 69 of the judgment seems to be based on paragraphs 31-37 of Mr Casale’s first witness statement.

The Judge carefully considered, and rejected, the As’ first three grounds of challenge in paragraphs 109-184 of the judgment. As I have already indicated, there is no appeal from the Judge’s decision to dismiss those grounds of challenge. This is a recognition, no doubt, that there is no arguable error of approach in any of the reasoning in this substantial part of the judgment. Wednesbury unreasonableness

The Judge accepted that a challenge to a decision could succeed under this head, either if a decision was outside the range of reasonable responses open to a decision maker, or if there was ‘a demonstrable flaw’ in the reasoning. He said that he had not found the issue ‘straightforward, most particularly in the light of the Secretary of State’s change of position’ (judgment, paragraph 194). He said it was necessary, first, to concentrate on ‘the underlying rationality of the Decision, rather than on the change of position’ (judgment, paragraph 194). The Decision, not the decision to consult, was the subject of the challenge. By the time of the Decision, the Secretary of State had received, and considered, ‘a substantial number of consultation responses’. There had to be some evidence of harm to animal welfare, as it would not have been lawful to impose a ban for moral reasons only (judgment, paragraph 195).

The Judge considered that the Decision gave two reasons for the ban: the ‘harm to animal welfare’ from the shock itself, and the risk of abuse.

He said that there was positive evidence that the use of e-collars could harm animal welfare, which he described in paragraph 196 of the judgment, and, as he noted, the As recognised that ‘the use of e-collars does raise animal welfare issues which need to be addressed’. In paragraph 197, he acknowledged that both e-collars and containment systems administered an electric shock. He said that there was a rational basis for distinguishing between them ‘as set out in the Decision’. Containment systems prevented other risks to animal welfare (eg straying into a road). Physical fencing was not ‘a real or practical and/or affordable alternative’. A pet learnt not to approach a fence after receiving only one or two shocks. He also referred to the Lincoln research about cats (see paragraph 25, above).

The Judge said, in paragraph 198 of the judgment, that the risk of misuse or of use otherwise than in accordance with the manufacturer’s instructions was inherent in the difference between a device which was operated manually, and one which was triggered automatically. The scope for harm to animal welfare from the first type of device was greater. There was evidence of concerns about the use of e-collars otherwise than in accordance with the manufacturer’s instructions in the consultation responses and in Lincoln 1 (which was referred to in the Decision). The As did not dispute that there was a potential for abuse; indeed that was one of their main reasons for supporting ‘stringent regulation’ of e-collars and their use. The Judge said that, as Beatson J had pointed out in Petsafe, the fact that other devices could have been banned did not make it irrational to ban e-collars.

Regulations under section 12 of the Act could address issues which were not covered by sections 4 and 9. The power conferred on the Secretary of State involved the exercise of ‘a broad discretionary judgment’. The Secretary of State considered the full breadth of the material, including the risks of dogs being destroyed (judgment, paragraph 199).

The Judge concluded (judgment, paragraph 200) ‘on the basis of this evidence and reasoning’, the Decision was not outside the range of reasonable responses open to the Secretary of State. Nor was there any ‘demonstrable flaw in the reasoning which led up to it’.

He then asked whether the Secretary of State’s change of position in February 2018 should lead to a different conclusion (judgment, paragraphs 201-202). ‘…stark and sudden’ though the change was, the question was whether that change could make an otherwise rational decision irrational. There was evidence of harm to animal welfare. Whether that evidence was enough to justify a ban was a question for the assessment of the decision maker. The Judge referred to the approach of Beatson J in Petsafe. Different decision makers could reach different views on that question, particularly where there were ‘strongly held divergent opinions on the evidence’. The authorities which had considered whether to ban e-collars had reached different views about that (admittedly, on evidence which was not the same). In a similar way, it was open to the same decision maker to reach different views on the same evidence at different times. The Secretary of State’s decision to propose a ban, on the same evidence, was a policy judgment. The Decision, four months later, was based on the evidence and informed by the consultation responses. A1P1

The Judge started by considering whether A1P1 could be engaged by a decision to ban e-collars in the future. He recorded the As’ submissions that the Decision was a

The Judge also summarised the As’ submissions on proportionality. The Decision was a deprivation or expropriation. There was a greater burden of justification in the case of an expropriation. There was no evidence of a proportionality assessment by the Secretary of State. A ban which increased the number of otherwise avoidable animal deaths and/or the use of other aversive products could not be shown to be rationally connected with the legitimate aim of animal welfare. The aim could have been achieved by less intrusive measures, that is, effective regulation. This has worked in other countries. The Secretary of State seems to have accepted initially that the evidence was not strong enough to support a ban. Concerns about abuse could be addressed by ‘proper regulation’. The Secretary of State could not show, on a fair balance, that the benefits of achieving the aim by a ban outweighed the disbenefits resulting from restricting the As’ rights.

The Judge then summarised the Secretary of State’s submission that the claim was premature, because there was no actual interference with the As’ rights, the scope of the impact had not crystallised, and there was no evidence of loss of present marketable goodwill because any impact was on future sales and there was no claim for the loss of any actual contracts. There was no evidence of a decisive effect on the As’ businesses, and no deprivation for the purposes of A1P1.

The case could not be distinguished from Petsafe on justification. There were no new factors militating against a ban. The broad rationale for the Decision was the same (the shock causes discomfort, e-collars can cause pain and distress in untrained hands, and e-collars conflict with reward-based training). The Secretary of State was plainly pursuing a legitimate aim. There was a wide margin of appreciation. The respective advantages and disadvantages of regulation and of a ban had been considered in the submission. There was a proportionality assessment in the RTA. The costs to business were low. The alternative of regulation would not necessarily be less intrusive.

The Judge’s reasoning about A1P1 and article 34 TFEU is in paragraphs 211-219 of the judgment. He summarised the relevant principles in paragraphs 211 and 212. The Judge said, that, as Beatson J had done, he would ‘proceed on the basis that’ a ban on e-collars would interfere with the As’ possessions, and that the Decision was also an interference. He accepted that a loss of future sales contracts was not enough. There was, nevertheless, evidence of ‘an actual downturn in sales’ since the announcement of the proposed ban, and, in the light of the evidence from Ms Critchley which he had summarised in paragraph 94 of the judgment (and which I refer to in paragraph 29, above) he considered it ‘likely that the present capital value of the goodwill of a member’s business (based on past sales of e-collars) is adversely affected by the knowledge that, in future, e-collar sales will be prohibited. To that extent, the Decision itself has already had an impact and amounts to a relevant interference’.

The promotion of animal welfare was a legitimate aim. The proposed ban was rationally connected with that aim. As he had already decided (under what was then ground 4), in the current circumstances, the proposed ban was a rational response to concerns about animal welfare. The three concerns identified by Beatson J (see paragraph 77, above) applied here. Just because the ban could have gone further (and applied, for example, to choke collars), did not mean that a narrower ban was not justified (the Judge referred to paragraph 65 of Petsafe).

In paragraph 217, the Judge considered whether the aim could have been achieved by a less intrusive measure, ‘namely effective regulation’. Various different forms of regulation had been suggested at different times. He said that ‘It was far from clear what measures are proposed and how and whether they would work in practice’. As Beatson J had observed (in Petsafe), regulation ‘would address only the concern about potential misuse and not, in particular, the harm arising from the electric shock itself. On any view, regulatory measures are likely to be complex, and, what is more, to impose a cost and time on business and/or a regulator and the consumer’. The Judge agreed with the analysis of option 4 in the submission; ‘the alternative of “regulation” has not been shown to be less intrusive or less costly’.

He then considered proportionality. In the same paragraph, he said that he gave ‘some weight’ to the assessments in the submission and in the RTA. His conclusion, ‘In any event’ was that ‘the burden of a ban (but excluding containment systems) upon the Claimants’ members is not so substantial nor so excessive as to be disproportionate when set against the benefits of promoting animal welfare’. Submissions

Mr Wise explained at the start of his oral submissions, in answering a question from the Court, that there is a dial on a hand-held device which enables the user to control the strength of shock which an e-collar gives a dog, and frankly accepted that that did open the way to abuse. His clients were very aware of this, and that had led them, in their representations to the Secretary of State, to stress the need for regulation. Handheld devices could be bought on the internet, and that, too, was an opportunity for abuse.

He submitted that e-collars had advantages which were not, or were not sufficiently, taken into account by the Secretary of State. For instance, they could be used to reduce some risks of harm, such as sheep worrying and attacks on other dogs, and could prevent the putting down of aggressive dogs for which no other training or control method worked. This advantage, reducing some risks of harm, was the same advantage as that of containment systems. He did accept that an animal had the choice and opportunity to avoid the shock delivered by a containment system, but no choice to avoid a shock of an e-collar. He suggested that both e-collars and containment systems gave rise to the same risk that the owner could set the shock level too high. In both systems the risks could be minimised by regulation. The Judge was wrong to describe containment systems as ‘automatic’.

Initially he argued that the decision to ban e-collars ‘may well not have been outside the range of reasonable responses’. Later in his submissions he said that it was not the As’ case that there was only one reasonable response, a point, which he accepted, Davis LJ had picked up when he gave permission to appeal. He therefore accepted that it was open to the Judge to hold that the Decision was not outside the range of reasonable responses (judgment, paragraph 200).

He also accepted that it was open to the Secretary of State to change his mind, but only if there was ‘good reason to do so’, or there were ‘cogent reasons’ and only if the Secretary of State explained the change. He accepted that it was likely that the February letter did no more than to express current policy, as opposed to a recent decision to maintain the policy. If there was no explanation ‘[w]e are in totally arbitrary territory’. The explanation for the change given in Mr Casale’s third witness statement was not supported by any evidence at all of changes in public attitudes, such as evidence from charities or opinion polls. Consistency is an aspect of rationality. The decision to ban e-collars did not treat e-collars and containment systems in a consistent way. He referred to R (Gallaher Group Limited) v Competition and Markets Authority [2018] UKSC 25; [2019] AC 96.

He supported the view of the Judge that A1P1 was engaged, referring to Breyer v

The As’ representations identified schemes of regulation which worked in practice.

In his reply Mr Wise argued that the submission could not be relied on as evidence of the Decision or of the process by which it was reached. It was drafted by officials. The only evidence of the Decision was one page in the published consultation response. That was a freestanding decision and it did not refer to the submission.

Mr Turney accepted that a Wednesbury challenge could succeed if there was a fundamental error in the process by which the Decision was made. He pointed out, however, that the Secretary of State was exercising a broad discretionary power in a ‘policy-heavy context’. He submitted that containment systems and e-collars are materially different. The Secretary of State was rationally entitled to distinguish between them. There was evidence, and it was not disputed, that e-collars can cause harm and suffering. Many owners did not read the instructions. E-collars could be abused. Many animal welfare organisations opposed them on good grounds. By contrast, containment systems had a limited impact on the welfare of animals, and could minimise other risks to the welfare of animals. The crucial difference was that e-collars allowed greater scope for human intervention, and thus for abuse. The Judge had considered the evidence at great length, and reached a conclusion which was open to him.

The Secretary of State was also entitled to change his mind. The possibility of a lawful change of mind is inherent in policy making. Mr Turney submitted that the Judge had gone too far in expressing surprise about this aspect of the case. The change of mind led to the consultation. The decision to consult was not challenged, as it should have been if the change of mind was the real target of challenge. There will always be a ‘switch point’ in the development of policy. It would be absurd if a decision maker had to produce reasons at every stage.

The Decision was framed as a response to a consultation, and preceded by a summary of the consultation responses. The Decision could be taken to have been informed by the consultation. The submission was also taken into account by the Secretary of State (and by the Judge). There was evidence both of negative effects from e-collars, and that cats did not suffer negative effects from containment systems.

Mr Turney challenged, by a Respondent’s Notice, the Judge’s decision that A1P1 was engaged. The A’s evidence about loss of profits was not strong. The only evidence concerned loss of future sales. Breyer shows that neither a right to acquire possessions in the future nor a prospective loss of future income (unless it is based on an enforceable obligation, such as a contract), is a possession for the purposes of A1P1. Marketable goodwill, on the other hand, is a possession, as is an existing contract. The As’ evidence failed to establish that the ban had any effect on their goodwill, so understood.

Bank Mellat did not show that an authority could ban one type of harmful device while not banning another. That case concerned the arbitrary singling out of one bank for intrusive measures, when other banks were in exactly the same position but the measure was not applied to them. It was clear from the submission that the Secretary of State had considered the option of regulation. The point was that no regulatory regime, however well run, could obviate all the risks to animal welfare posed by ecollars. The evidence about regulation elsewhere did not show whether it was effective or ineffective.

If A1P1 was engaged at all on the facts, it was only engaged tangentially. The statutory provision gave the Secretary of State a broad power concerning animal welfare. The Judge’s reasoning had to be read against his extensive rehearsal of the facts and his reasoning on rationality. There was not much more, if anything else, for the Judge to say. The Judge was entitled to give weight to the way in which the Secretary of State had struck the balance between the limited interference with the As’ possessions and the objective of protecting the welfare of animals. Discussion

I start with Mr Wise’s argument that the submission is not relevant to an assessment of the Decision and the reasons for the Decision. It is not unusual for ministerial decision makers to be provided by officials with a written submission which summarises the considerations which are thought to be relevant to a decision, and which makes recommendations to them. The evidence of Mr Casale about how the Decision was made (see paragraph 65, above) does not seem to have been challenged before the Judge. I consider that, in these circumstances, the Judge was entitled, and right, to take into account the submission and the response document in reaching his conclusions about the Decision. Irrationality

There are two starting points. First, the premise of the grant of permission to appeal, which Mr Wise accepted in his oral submissions, was that the Decision was within a range of reasonable responses. Second, the Secretary of State was exercising a broad power for the purpose of promoting the welfare of animals. The exercise of this power will often depend on one or more evaluative judgments. Some of those judgments may be more intuitive, and less amenable to scientific proof, than others. I consider that the comparison of, and differentiation between, e-collars and containment systems involve at least two evaluative assessments.

One judgment (whether either system harms the welfare of animals) can be informed by research evidence on that subject. The research evidence before the Secretary of State suggested that e-collars do harm some dogs, whereas containment systems did not harm cats. A further question is whether, and if so, how, to regulate the two systems. There is no scientific answer, or part-answer, to that question. A range of factors is relevant to it; such as the extent to which an owner can manipulate the system and use it to inflict unnecessary suffering on an animal. The evidence before the Secretary of State overwhelmingly suggested that containment systems are set up by professionals who provide training and ensure that the system is appropriate to the animal and its environment. There was no such evidence about e-collars. I accept Mr Turney’s submissions that the Secretary of State was rationally entitled to decide to treat the two systems differently. I do not consider that, in distinguishing between the two, the Decision was reached by a flawed process of reasoning. The submission that both systems give an animal an electric shock and should therefore be treated in the same way is, in my judgment, far too crude.

I consider that the Judge somewhat overstated the apparent change of position by the Secretary of State between 2013, the letter of 5 February 2018, and the decision to consult on ban.

First, the Secretary of State’s position in 2013 was that Lincoln 2 did not provide evidence of a ‘significant risk of harm to animal welfare’ and that, for a ban to be justified, it was necessary to show that e-collars were ‘harmful to the long-term welfare of dogs’. This position does not address harm to the short-term welfare of dogs, and is based on a value judgment about what amounts to ‘a significant risk to dog welfare’ (cf paragraph 33, above).

Second, the position in the letter of 5 February 2018 was that:

Once it is accepted, as Mr Wise had to, that a decision to ban e-collars, and a decision not to ban them, are both within the range of reasonable responses, it is very difficult to see how it can be irrational for a decision maker to change his mind about whether a ban is appropriate, or that he should be required to give reasons, let alone ‘cogent reasons’ for changing his mind, even if the underlying evidence is the same. In the case of broad evaluative judgments such as these, a decision maker can decide, without error, that his view of the fine balance of the relevant factors has tipped from one side to another. The letter of 5 February 2018, in any event, was not evidence of a recent re-evaluation of the relevant factors, but simply a re-statement of what had been the Department’s policy for several years. It did not, in my judgment, make it irrational for the Secretary of State to ask officials for advice about banning e-collars, or irrational to initiate the consultation which informed the Decision.

A silent premise of Mr Wise’s submissions on less intrusive means seemed to me to be that, to succeed on this issue, the As only had to point to other jurisdictions in which e-collars were regulated, rather than banned. That is not correct, as the less intrusive means must be capable of achieving the legitimate aim which is at issue. I accept Mr Turney’s submission that there was no evidence, one way or the other, about whether systems of regulation in other jurisdictions worked, or not. It might be thought that it was unlikely that there would have been much evidence, as it seems that, at the date of the hearing before the Judge, the Scottish system was still a proposal, and the system in the Netherlands had not yet come into force. The only system which had been in force for some years was the system in the state of Victoria (see paragraph 46, above). Moreover, I do not consider that it is possible to prove, by any method of quantification, whether a system of regulation ‘works’. Finally, the concern about e-collars is that, if they are not banned, no system of regulation will prevent a cruel owner from using one to inflict unnecessary suffering on an animal. These factors mean that whether regulation is preferable to a ban is essentially a policy judgment.

I consider that these were the points which the Judge had in mind in paragraph 217 of the judgment. I do not consider that it is arguable that he was wrong to take those points into account, or to reach the conclusion which he did on this issue. Indeed, if it were necessary for me to do so, I would say that I agree with that conclusion.

I will return to the extent of the interference, if any, with the As’ rights in the next paragraph. It is relevant to the issue of proportionality. Even if I assume, for the moment, that there is such an interference, the As’ evidence about its nature, which I have described in paragraph 29, above, was thin. The proportionality balance in this case involved a policy judgment about two incommensurables. The Secretary of State, and the Judge, were fully informed by the RTA (and in a way which was very generous to the As, because it went further than their evidence did) about the potential economic effect of a ban on the As. That went into one side of the balance. The Secretary of State and the Judge were fully informed, by the submission and the consultation responses, about the animal welfare effects of e-collars, about the advantages and disadvantages of the four alternatives to a ban, and about the range of views about a ban and its likely effects. Those went into the other side of the balance. The Judge expressed himself laconically in paragraph 217 of the judgment, but against the background of his detailed review of the evidence and the arguments in the preceding paragraphs of his judgment, I do not consider that he was required to say more. I do not consider that he reached the wrong conclusion, and, again, if it is necessary for me to say so, I agree with it.

I said that I would return to the question whether the Judge was right to hold that there was an interference with the A1P1 rights of the As. In the light of the views I have expressed in the previous two paragraphs, I do not consider that it is necessary for me to say very much about this. Breyer dealt with two questions:

The Judge was right to hold that (if the effect of the proposed ban was an effect on ‘possessions’ for the purposes of A1P1), then, Breyer shows that a practical rather than a legal effect is sufficient. ‘Possessions’ for the purposes of A1P1 is an autonomous concept. ‘Goodwill’, as referred to in the decisions of the ECtHR which are quoted in Breyer, does not equate with ‘goodwill’ as accountants understand it, however. I consider that there is much force in Mr Turney’s argument that the evidence in this case did not establish an interference with the As’ possessions for the purposes of A1P1. At its highest, the evidence showed no more than a potential effect on future income. That is not enough.

For those reasons, I would dismiss this appeal.

I agree.