England and Wales Court of Appeal (Civil Division) Decisions

IN THE COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE HIGH COURT OF JUSTICE
QUEEN'S BENCH DIVISION
THE ADMINISTRATIVE COURT
The Hon. Mr Justice Mitting
[2017] EWHC 435 (Admin)

		Royal Courts of Justice Strand, London, WC2A 2LL
		25/05/2018

Lady Justice Gloster :

Introduction

1. This is an appeal by Transport for London ("TfL" or "the appellant") against the order of Mitting J ("the judge") dated 3 March 2017 allowing the judicial review application of Uber London Limited ("Uber") and others and quashing Regulation 9(11) of the Private Hire Vehicles (London) (Operators' Licences) Regulations 2000 ("the Regulations").
2. TfL regulates the taxi and private hire trade in London. Its responsibilities include the licensing of private hire vehicles ("PHVs"), drivers and operators.
3. Uber is a PHV operator in London that utilises the Uber app. The second to fourth respondents are PHV drivers who provide passengers with PHV services using the Uber app (but who, in contrast to Uber, are only indirectly affected by this appeal).
4. This appeal concerns whether the judge was correct to conclude that the imposition by TfL of a requirement on PHV operators in London to provide a "listening" service to the passenger for whom a booking had been made, which was to be available at all times during the operator's hours of business (and at all times during a journey), and in respect of any matter (referred to in argument and in this judgment as "the Voice Contact Requirement") constituted a disproportionate interference with the rights to freedom of establishment of PHV operators, contrary to Articles 49 and 54 of the Treaty on the Functioning of the European Union ("TFEU").
5. The substantive hearing took place before Mitting J from 28 February 2017 to 2 March 2017 and he gave an oral judgment on 3 March 2017. An order dated 3 March 2017 was drawn up reflecting the judge's decision to allow Uber's and the other claimants' claim for judicial review of the Voice Contact Requirement and to quash Regulation 9(11). The order also dismissed Uber's and the other claimants' claim to challenge an English language requirement ("the English Language Requirement") imposed by Regulations 3A and 3B of the Private Hire Vehicles (London PHV Driver's Licences) Regulations ("the 2003 Regulations").
6. On 12 February 2018, Uber's appeal against the part of the judge's order that upheld the English Language Requirement was withdrawn by consent.

Uber's "click-to-call" Function

7. On 16 February 2018, shortly before the hearing of the appeal, Uber emailed its customers informing them that it would be introducing a dedicated support line using the following language:

"Sometimes there's no substitute for talking to a real, live person. That's why we're introducing telephone support. Starting later this year, riders and drivers will be able to access trained support staff 24 hours a day, 7 days a week through a dedicated line."

8. This court was presented with further evidence in the form of a fifth witness statement from Mr Thomas Elvidge (Uber's General Manager) about Uber's new "click-to-call" function, which will operate in the app for passengers and drivers. Uber has not yet finalised the full details and it will take, in its estimation, approximately six months to implement. It expects to use this time to put in place technology, infrastructure and resources needed to carry this out in a manner that is consistent with its internal target of a live voice response to safety-related issues within two minutes. Mr Elvidge provided further detail at [9f]:

"[w]hile this phone support will not be just for emergencies or safety-related issues, it is likely that for some non-urgent and low-impact issues, the automated voice-based system may redirect callers to our existing support channels of in-app messaging."

9. As accepted by Uber, while the "click-to-call" function goes some way to fulfilling the Voice Contact Requirement, it is not fully compliant.

Statutory framework governing PHVs

10. I am grateful to counsel for TfL, Mr Martin Chamberlain QC, Mr Tim Johnston and Mr David Heaton, for their clear summary of the relevant law in this area.
11. TfL was established by the Greater London Authority Act 1999 ("GLAA") (s. 154). Parliament conferred on TfL wide statutory powers to secure the provision, promotion and encouragement of safe, integrated, efficient and economic transport facilities and services to, from and within Greater London (ss. 141, 173 and 181). It can use these powers to impose conditions subject to which licences for operators, vehicles and drivers in the private hire trade can be granted.
12. TfL is obliged to secure the implementation of the Mayor's Transport Strategy ("the Strategy"). The Strategy must "contain the Mayor's proposals for the provision of transport which is accessible to persons with mobility problems" (s. 142(2)). Its goals include:

i) improving the safety and security of all Londoners, including "ensur[ing] that journeys by taxi and private hire vehicle (PHV) are as safe as possible for passengers and drivers";

ii) improving transport opportunities for all Londoners, which includes: "improving accessibility" and in particular "overcoming the barriers that exist for some users", especially disabled Londoners, older and young people; "providing better information and communications"; and "improving the actual and perceived safety and security of transport services and travel"; and

iii) enhancing quality of life for all Londoners, including "improving journey experience".

13. The Private Hire Vehicles (London) Act 1998 ("the 1998 Act") establishes a three-part licensing regime that governs private hire operators (ss. 2-5), vehicles (ss. 6-11) and PHV drivers (ss. 12-14) in London. TfL's powers include suspending or revoking a licence (s. 16).
14. Section 2(1) of the 1998 Act provides that PHV bookings have to come through an operator:

"No person shall in London make provision for the invitation or acceptance of, or accept, private hire bookings unless he is the holder of a private hire vehicle operator's licence for London."

15. Section 2(2) of the 1998 Act makes it an offence to breach that provision.
16. Section 32 of the 1998 Act provides the mechanism by which TfL may prescribe conditions subject to which some licences may be granted:

"(1) [TfL] may make regulations for any purpose for which regulations may be made under this Act (other than section 37) or for prescribing anything which falls to be prescribed under any provision of this Act (other than section 37).

(2) Regulations under this Act may –
(a) make different provision for different cases;
(b) provide for exemptions from any provision of the regulations; and
(c) contain incidental, consequential, transitional and supplemental provision."

The consultation process

17. From March 2015 to February 2016, TfL conducted a comprehensive three-stage review and consultation on the licensing regime that governs the private hire sector. Mr Chamberlain QC, on behalf of TfL, took the court through this process in some detail but I do not consider it is necessary to repeat that process here. It is clear that TfL engaged in a wide, detailed consultation and gathered responses from a range of customers. It used these results to refine its proposals.
18. Through this consultation, TfL identified public safety, equality and passenger convenience benefits in relation to the Voice Contact Requirement. During the process, it made one significant change to this particular proposal: the omission of the requirement that an operator must have a fixed landline telephone available for passengers to ring at all times, including for bookings. After the measure was first made, TfL also revised it to omit the requirement for contact to be at the London operating centre.
19. It is clear that TfL considered the costs of compliance with the Voice Contact Requirement and the potentially disproportionate impact on small operators including single driver-operators.
20. Following the consultation, TfL identified 19 reforms that would improve customer safety, equality of access to PHV services and customer care (amongst other objectives). It prescribed a number of new conditions of an operator's licence in exercise of its powers under s. 32 and under s. 3(4). Three of these new conditions were the subject of the judicial review proceedings before the judge.
21. The first was the English Language Requirement. This is a requirement that drivers demonstrate proficiency in the English Language to level B1 on the Common European Framework of Reference in speaking, listening, reading and writing. As I have already said, on 12 February 2018, Uber withdrew its appeal in relation to the judge's dismissal of its challenge to this requirement.
22. The second was the Voice Contact Requirement. This is a requirement that:

"at all times during the operator's hours of business and at all times during a journey, the operator shall ensure that the passenger for whom the booking was made is able to speak to a person at the operating centre or other premises with a fixed address in London or elsewhere (whether inside or outside the United Kingdom) which has been notified to the licensing authority in writing if the passenger wants to make a complaint or discuss any other matter about the carrying out of the booking with the operator."

23. I use Mr Chamberlain's reformulated label of "Voice Contact Requirement" rather than the label "Telephone Requirement" used below. It is more accurate since the voice contact service TfL requires can be by telephone or over the internet (or otherwise).
24. The third was the Insurance Requirement. This is a requirement concerning insurance for PHVs. This requirement is not the subject of this appeal and will not be discussed further.

The customer benefits before the judge

25. The evidence before the Judge showed that there was a wide variety of circumstances where the customer would derive a real benefit from being able to speak to a real person, whether the issue being complained about could properly be described as an emergency or not. The evidence before this Court also showed that there was a real difficulty in distinguishing emergency situations from non-emergency ones.

Case law on proportionality

26. There is no issue between the parties as to the relevant legal principles but rather as to their application. Again, I am grateful to Mr Chamberlain and his team for their clear presentation of these principles.

Article 49 and 54 TFEU

27. It is not disputed that the Voice Contact Requirement engages Articles 49 and 54 of the TFEU. Article 49, which guarantees the freedom of establishment, provides that:

"[w]ithin the framework of the provisions set out below, restrictions on the freedom of establishment of nationals of a Member State in the territory of another Member State shall be prohibited. Such prohibition shall also apply to restrictions on the setting-up of agencies, branches or subsidiaries by nationals of any Member State established in the territory of any Member State.

Freedom of establishment shall include the right to take up and pursue activities as self-employed persons and to set up and manage undertakings, in particular companies or firms within the meaning of the second paragraph of Article 54, under the conditions laid down for its own nationals by the law of the country where such establishment is effected, subject to the provisions of the Chapter relating to capital."

28. Article 54 makes it clear that Article 49 applies to Uber. It provides that:

"[c]ompanies or firms formed in accordance with the law of a Member State and having their registered office, central administration or principal place of business within the Union shall, for the purposes of this Chapter, be treated in the same way as natural persons who are nationals of Member States.

Companies or firms means companies or firms constituted under civil or commercial law, including cooperative societies, and other legal persons governed by public or private law, save for those which are non-profit-making."

Application of R (Lumsdon) v Legal Services Board

29. Consequently, the judge had to determine whether the Voice Contact Requirement was proportionate. The relevant approach is set out in R (Lumsdon) v Legal Services Board [2016] AC 697 and relevant EU case-law. At [33], in the judgment of Lord Reed and Lord Toulson, with which the other members of the court agreed, it was held:

"[p]roportionality as a general principle of EU law involves a consideration of two questions: first, whether the measure in question is suitable or appropriate to achieve the objective pursued; and secondly, whether the measure is necessary to achieve that objective, or whether it could be attained by a less onerous method. There is some debate as to whether there is a third question, sometimes referred to as proportionality stricto sensu: namely, whether the burden imposed by the measure is disproportionate to the benefits secured. In practice, the court usually omits this question from its formulation of the proportionality principle. Where the question has been argued, however, the court has often included it in its formulation and addressed it separately." [emphasis added]

30. Mr Chamberlain addressed the question of proportionality stricto sensu separately. However, Mr Thomas de la Mare QC, on behalf of the Respondents, abandoned this aspect of his case. Accordingly, it is no longer an issue.
31. At [37], Lumsdon deals with proportionality as a ground of review of national measures. It states that in that context:

"[it] has been applied most frequently to measures interfering with the fundamental freedoms guaranteed by the EU Treaties. Although private interests may be engaged, the court is there concerned first and foremost with the question whether a member state can justify an interference with a freedom guaranteed in the interests of promoting the integration of the internal market, and the related social values, which lie at the heart of the EU project. In circumstances of that kind, the principle of proportionality generally functions as a means of preventing disguised discrimination and unnecessary barriers to market integration. In that context, the court, seeing itself as the guardian of the Treaties and of the uniform application of EU law, generally applies the principle more strictly."

32. In the same paragraph, Lord Reed and Lord Toulson continue:

"[w]here, however, a national measure does not threaten the integration of the internal market, for example because the subject matter lies within an area of national rather than EU competence, a less strict approach is generally adopted. That also tends to be the case in contexts where an unregulated economic activity would be harmful to consumers, particularly where national regulatory measures are influenced by national traditions and culture."

33. In Lumsdon, Lord Reed and Lord Toulson then set out the general approach to proportionality for a national measure derogating from fundamental freedoms, by reference to Case C-55/94 Gebhard [1995] ECR I-4165 at [52]:

"'national measures liable to hinder of make less attractive the exercise of fundamental freedoms guaranteed by the Treaty must fulfil four conditions: they must be applied in a non-discriminatory manner; they must be justified by imperative requirements in the general interest; they must be suitable for securing the attainment of the objective which they pursue; and they must not go beyond what is necessary in order to attain it.' (para 37)" [emphasis added]

34. At [55], the Supreme Court provided guidance on how to apply the proportionality criteria in relation to suitability. Lord Reed and Lord Toulson endorsed Advocate-General Sharpston's statement that the member state does not have to

"establish that the restriction is the most appropriate of all possible measures to ensure achievement of the aim pursued, but simply that it is not inappropriate for that purpose."

35. The Advocate-General went on to state that it is not enough to show that the public authority could have adopted a measure that is less restrictive if that measure would also have been less effective. In the same paragraph, she stated:

"[a]s regards proportionality, however, it is necessary to establish that no other measures could have been equally effective but less restrictive of the freedom in question."

36. I accept Mr Chamberlain's submission that the fact the Voice Contact Requirement is subject to review, is a feature which goes to the breadth of the margin that TfL has as a regulator in determining its proportionality. Authority for this point comes from [60] of Lumsdon:

"[p]articularly in situations where a measure is introduced on a precautionary basis, with correspondingly less by way of an evidential base to support the particular restrictions imposed, it may well be relevant to its proportionality to consider whether it is subject to review in the light of experience."

37. At [63] in Lumsdon, the Supreme Court considered the less restrictive alternative test and made it clear it is not to be applied mechanically. I agree with Mr Chamberlain that TfL was not required in its evidence to consider every possible alternative, including those that were never suggested by consultees. (For example, Uber never suggested an alternative voice contact service just for emergencies.) Authority for this point comes from the citation of Commission of the European Communities v Italian Republic (Case C-518/06) [2009] ECR I-3491, which stated:

"[the] burden of proof cannot be so extensive as to require the member state to prove, positively, that no other conceivable measure could enable that objective to be attained under the same conditions."

38. TfL has a discretion about the appropriate level of consumer protection and how to go about achieving it. At [64] the Supreme Court noted that:

"where a relevant public interest is engaged in an area where EU law has not imposed complete harmonisation, the member state possesses 'discretion' (or, as it has sometimes said, a 'margin of appreciation') not only in choosing an appropriate measure but also in deciding on the level of protection to be given to the public interest in question."

39. Further clarification on the scope of the margin of appreciation is provided at [66]:

"[t]his margin of appreciation applies to the member state's decision as to the level of protection of the public interest in question which it considers appropriate, and to its selection of an appropriate means by which that protection can be provided. Having exercised its discretion, however, the member state must act proportionately within the confines of its choice. A national measure will not, therefore, be proportionate if it is clear that the desired level of protection could be attained equally well by measures which were less restrictive of a fundamental freedom."

Ex-post facto reasoning

40. In determining whether a measure is proportionate, the court may consider information, evidence or other material available at the time it gives its ruling and is not confined to considering the material available when the decision to implement the national measure was taken: Scotch Whisky Association v Lord Advocate 2017 SLT 1261 at [65].
41. However, a public authority should not be afforded the same margin of appreciation in relation to justifications and material supporting them which it did not take into account when imposing the relevant restriction, and which it only developed in response to litigation. This point was made by the Supreme Court in Re Brewster [2017] 1 WLR 519 at [50]:

"But the margin of discretion may, of course, take on a rather different hue when, as here, it becomes clear that a particular measure is sought to be defended (at least in part) on grounds that were not present to the mind of the decision-maker at the time the decision was taken. In such circumstances, the court's role in conducting a scrupulous examination of the objective justification of the impugned measure becomes more pronounced."

42. The Supreme Court went on to state that, if the justifications were real and within the decision maker's competence, a reviewing court must afford him deference. It continued at [52]:

"[o]bviously, if reasons are proffered in defence of a decision which were not present to the mind of the decision-maker at the time that it was made, this will call for greater scrutiny than would be appropriate if they could be shown to have influenced the decision-maker when the particular scheme was devised. Even retrospective judgments, however, if made within the sphere of expertise of the decision-maker, are worthy of respect, provided that they are made bona fide."

It was then held at [55], for the purpose of that appeal:

"the test to be applied is that of 'manifestly without reasonable foundation'. Whether that test requires adjustment to cater for the situation where the proffered reasons are the result of deliberation after the decision under challenge has been made may call for future debate. Where the state authorities are seen to be applying 'their direct knowledge of their society and its needs' on an ex post facto basis, a rather more inquiring eye may need to be cast on the soundness of the decision."

The judgment below

43. Although, the judge dealt with a judicial review application in relation to the Voice Contact Requirement, the English Language Requirement and the Insurance Requirement, I will only focus on the parts of the judgment that concern the first of these.
44. The judge accepted the parties' submission that the Voice Contact Requirement engaged Article 49 TFEU. At [16] – [18] of the judgment, he set out the passages that he considered relevant from Lumsdon in relation to the proportionality test. Both parties were satisfied that the judge referred to the correct law. However, there is a dispute about the judge's application of it.
45. The judge held that the cost that the Voice Contact Requirement imposed on Uber was significant and was likely also to be significant for "operator drivers": operators who run two or fewer PHVs. He did not, however, put this into context.
46. He held that the justifications given for the Voice Contact Requirement in TfL's evidence established that there was no less intrusive means of securing the high level of protection of public safety and consumer protection that TfL had selected, at least in emergency situations. It became common ground during the trial – and, if not, he would have concluded – that no alternative measure would give sufficient assurance and speed of response to customers in emergencies.
47. The judge went on to reject TfL's case. He held that the Voice Contact Requirement was unlawful "in its full width", because the system operated by Uber – for interacting with its passengers via its online app, initiated by a written complaint from the customer – was adequate for "non-emergency issues". The judge did not address why the benefits he identified in emergency situations did not also count as benefits in non-emergency situations.
48. The judge considered it important that Uber's clientele was, by self-selection, willing to make use of the app-based facility it offered to book PHVs. In his judgment, most of them could, accordingly, be expected to be content to use the same facility, at least initially, to resolve non-emergency issues.
49. The judge noted that TfL conducted no survey of users of app-based platforms alone, only of PHV and taxi users generally. He rejected TfL's reliance on Uber's estimate that 25% of its complaints or enquiries would migrate from app to telephone under the Voice Contact Requirement as "unsound".
50. He asserted that a less restrictive measure existed whereby customers were provided with "a 'hot line' for emergencies broadly defined to include situations in which immediate action is required to remedy a breach of the law" ("the Emergency Telephone Alternative"). In his view, that would secure the same level of passenger convenience for PHV passengers.
51. I accept Mr Chamberlain's submission that it was relevant that the Emergency Telephone Alternative was: (a) not suggested by any respondent (including Uber) during the entire consultation process; (b) formed no part of the Respondents' pleaded case on the evidence; and (c) was, consequently, not addressed by TfL in evidence in the initial hearing. The judge failed to consider the practicability or benefits of an "emergency" rather than a "general" contact facility.
52. The judge concluded that TfL had failed to show that other less restrictive measures would not meet the level of protection required for the only public interest - passenger convenience - which he identified was at stake. Consequently, he quashed the Voice Contact Requirement and left it to the TfL Commissioner to reconsider the Emergency Telephone Alternative.

The issues on the appeal

53. The issues that arise on this appeal are as follows:

i) Whether the judge erred in relying on the distinction between an emergency and a non-emergency contact facility?

ii) Whether the judge accorded TfL a proper margin of appreciation in this area?

iii) Whether the Voice Contact Requirement achieves a higher level of protection than the Emergency Telephone Alternative and achieves legitimate objectives beyond safety, such that the latter is not a less restrictive alternative?

iv) Whether the Voice Contact Requirement was a proportionate and lawful interference with Uber's freedom of establishment in any event and the judge erred in concluding otherwise?

v) Whether it was unlawful that TfL had not imposed a similar requirement on taxis?

Discussion and determination

54. Broadly speaking, Mr Chamberlain supported his grounds in his oral submissions. Mr de la Mare supported the conclusions of the judge and offered additional reasons for upholding his decision. The basis of his submission was that the Voice Contact Requirement, in its current form, was too broad and would not achieve any practical benefits where an effective app system of dealing with complaints already existed.

Issue 1 - Whether the judge erred in relying on the distinction between an emergency and a non-emergency contact facility?

55. For the reasons given by the appellant in argument, I consider the judge erred in relation to this issue and it is appropriate for this court to revisit it. Mr Chamberlain submitted that there was no evidential basis for the Emergency Telephone Alternative. It was not suggested by any Respondent, formed no part of the Respondents' pleaded case or evidence and was therefore not something which TfL addressed in evidence before the judge. The consultation documents further indicate that it was never suggested, by Uber or anyone else, during the extensive consultation.
56. I agree with TfL that the judge had no evidence before him as to whether the Emergency Telephone Alternative was actually practicable. I consider he erred by substituting his view for that of the regulator and suggesting a "less intrusive alternative" that was not fully considered. As such, Uber's introduction of a voice contact facility that Mr de la Mare submitted was in the spirit of the judge's conclusion, does not, in my judgment, go far enough.

Practicability

57. There are three other difficulties raised by the Emergency Telephone Alternative, as described in the evidence put forward on appeal by TfL. The first is practicability. It is unclear how a distinction between an emergency and a non-emergency contact facility would work in practice. It is difficult to define an "emergency" because one man's emergency is another man's inconvenience. An oft-quoted example of this predicament is the lost mobile phone. For some, this is a simple lost property issue and for others, it is viewed as an emergency. As a result, I consider TfL has a legitimate concern that it would be difficult to frame and communicate such a requirement to passengers so as to ensure they only use it in the type of broad emergency situations envisaged by the judge.
58. Further, there would be a real difficulty for operators to understand which calls the Emergency Telephone Alternative required them to take. It follows that I accept that TfL would face problems with enforcing such a requirement. This is because in determining whether a particular voice contact facility was properly being made available to passengers, TfL would need to consider both the nature of the voice contact that the operator received and whether this fell within the broad definition of emergencies.
59. I am also persuaded that there is a public safety concern in that the restriction of the Emergency Telephone Alternative to broadly defined "emergencies" might give the impression that it was an alternative to calling 999 in situations where an immediate referral to the police, ambulance or other emergency services would be more appropriate.
60. Given these difficulties, I agree with Mr Chamberlain that the judge's Emergency Telephone Alternative constituted a regulatory requirement, which TfL, as the regulator, would have had some difficulty applying and monitoring. TfL is entitled to conclude that it is not practicable.

The additional benefits

61. The second is the Emergency Telephone Alternative's failure to achieve many of the benefits of the Voice Contact Requirement. I accept Mr Chamberlain's submission that the latter achieved a number of public safety, equality and customer convenience benefits, which the former did not achieve to the same extent, and in some cases, at all.

Less burdensome

62. The third is the Emergency Telephone Alternative is not actually less burdensome than the Voice Contact Requirement. To begin with, I am persuaded that once passengers are provided with a number on which it is possible to contact an operator, they are likely to use it regardless of whether they consider their issue to be a broad "emergency" or whether this is objectively true. Following on, in order to determine whether an operator is obliged by the Voice Contact Requirement to take a call, it is necessary that he answers and screens all calls. While an automated system may work to filter out some calls, it is difficult at this stage to determine how effective that would be. In order to comply, the facility would have to allow for the caller to reach a live person relatively quickly (especially where it is limited to emergencies).
63. Given all of this, I agree with TfL that it is unclear that the burden of providing the Voice Contact Facility, in the form of costs (both the capital costs of a physical activity, if they are necessary, and ongoing staffing costs of providing the voice contact service), imposed on operators would be meaningfully less than for the Emergency Telephone Alternative. The Court has not seen any evidence from Uber to support this point and, indeed, I am supported in this view by the fact that, by the time of the hearing, Uber had indeed decided to introduce its own version of the Voice Contact Requirement in the form of the "click-to-call" function.
64. In my judgment, having accepted TfL's case that it was necessary and proportionate to require operators to provide a voice contact facility in the case of genuine emergencies, the judge should not have quashed the Voice Contact Requirement. He erred in relying on the distinction between an emergency and a non-emergency contact facility. I would thus allow the appeal on this point.

Issue 2 - Whether the judge accorded TfL a proper margin of appreciation in this area? / Issue 3 - Whether the Voice Contact Requirement achieves a higher level of protection than the Emergency Telephone Alternative and achieves legitimate objectives beyond safety, such that the latter is not a less restrictive alternative?

65. Issues 2 and 3 are linked and accordingly I can take them together. In my judgment, the appellant's arguments in relation to both of these issues are to be preferred and the judge was wrong to decide otherwise.

Margin of Appreciation

66. There is an issue as to whether this case engages with Article 49 at all. I find it difficult to accept that the Voice Contact Requirement actually interferes with the right of Uber and other PHV operators to establish themselves as companies in the UK or London. Both parties agreed this case was towards the de minimis end of the spectrum in terms of its effect on freedom of establishment. This is not a disguised discrimination case. It is this background against which any proportionality analysis must be measured.
67. Mr de la Mare submitted that as the regulator's legitimate interest deviates from core public safety issues, any claim to deference and, in some cases, any claim to vires to regulate for those kinds of matters decreases. However, I do not consider this to be the correct application of Lumsdon. I am persuaded by Mr Chamberlain's submission that the regulation of PHVs is not an area of EU competence and it is an economic activity where a lack of regulation would be harmful to consumers, such that the court must adopt a less strict approach to proportionality (Lumsdon at [37]).
68. As mentioned earlier, it is also relevant that the Voice Contact Requirement is subject to review. This is a feature which goes to the breadth of the margin that TfL has, as a regulator, in determining its proportionality. I am persuaded that it supports the conclusion that the margin of appreciation afforded to TfL is a wide one.
69. It is accepted by both sides that TfL's justification for the Voice Contact Requirement expanded over time from public safety to include reassurance and speed of response. Mr de la Mare submitted that a number of cases, culminating in Re Brewster, suggested that, where a public authority's reasoning is to an extent ex post-facto, it should not be afforded the same margin of appreciation in relation to these additional justifications. In my judgment, when one considers the material before TfL, in the form of the consultations, there was consideration of issues beyond safety and including customer convenience. Further, this evidence is admissible in this context when considering the question of proportionality under EU law. Mr de la Mare did not challenge this point. These additional justifications are bona fide and within TfL's sphere of expertise. As such, Re Brewster does not change my analysis on the deference that is to be shown to it in this area.
70. Lumsdon makes it clear that TfL has a discretion about the appropriate level of consumer protection and how to go about achieving it (at [66]). Mr Chamberlain submitted that the judge erred by substituting his own view for that of TfL, as to the appropriate level of protection for customers in a non-emergency situation. I agree. In doing so, he failed to accord TfL a proper margin of appreciation in this area.

Less restrictive alternative

71. I found force in TfL's submission that that the ability to contact the operator by phone would confer on customers of those operators a real and identifiable benefit over and above what they currently enjoyed. This was the case for Uber as well as other PHV operators, some of whom may not have the same extensive app complaint system as Uber. As such, the true comparison was between the level of protection enjoyed across the market as a whole and the higher level which TfL was entitled to seek to achieve rather than between Uber and other PHV operators, as the judge's conclusion suggests.
72. The judge accepted that the voice contact facility provided real benefits, in the form of reassurance and speed of response to customers in a genuine emergency. Mr de la Mare accepted this conclusion but submitted the Voice Contact Requirement was not justified in its full width. I disagree. In my judgment, when determining the lawfulness of the Voice Contact Requirement, it is not enough that the judge went on to find that Uber's current system was adequate for dealing with non-emergency issues. He erred by failing to explain why the benefits he identified as existent in the emergency situation, namely, reassurance and a speedy response, were not also legitimate objectives in a non-emergency situation.
73. Whilst it is true that some Uber users are, by self-selection, willing to make use of the app-based facility it offered to book PHVs, it does not follow that they are necessarily content to use the same facility, at least initially, to resolve non-emergency issues. On this point, Mr de la Mare referred to the Competition and Markets Authority's ("CMA") response to the consultation process. In the latter's view, the introduction of the Voice Contact Requirement should be avoided because it will create barriers to entry and stifle innovation. Mr de la Mare submitted this was a quality of service issue and customers who did not like using an app-based operator should simply not use one. I did not find the reference to the CMA's response helpful as the present case is not about competition.
74. The legislation empowers TfL to impose obligations for the benefit of consumers in this area, providing that those obligations are proportionate. The judge considered the evidence and it was clear, on Uber's own evidence, that even after booking using an app, a certain substantial number of customers still preferred voice contact over other methods. I consider he was wrong to describe Mr Chamberlain's reliance on this as "unsound."
75. Further, on Uber's own evidence at the hearing of the appeal, it is planning to introduce a voice contact facility that is not limited to emergencies. I find it hard to believe that Uber would be doing this if some customers who make contact via the app do not value the opportunity to speak to somebody by voice. It must also be remembered that the Voice Contact Requirement is not limited to Uber. Accordingly, I consider the judge was wrong to find that "no useful purpose" would be served by a voice contact facility for non-emergency cases, even for Uber's own customers.
76. In my judgment, the Emergency Telephone Alternative was not a less restrictive alternative. The system proposed by TfL achieved a higher level of protection and also achieved the further legitimate objectives outlined at paragraph 69 above, beyond safety or consumer protection in a non-emergency situation, which the judge did not properly consider. He put the Emergency Telephone Alternative forward as a less intrusive means of achieving the same objectives. For it to be less intrusive, it had to be just as effective. It was not.

Proportionality

77. The next question was whether, once TfL had exercised its discretion to introduce the Voice Contact Requirement, it was proportionate when considering TfL's desired level of protection. As part of this analysis, I consider it important, as the judge did, to deal with the costs imposed on Uber and other PHV operators of complying with the Voice Contact Requirement. Mr Chamberlain addressed this as part of his submissions on proportionality stricto sensu; however, I consider, as the judge did, that it arises as part of the first part of the proportionality test in Lumsdon.
78. I agree with TfL's submission that there is a problem with the judge's failure to provide context for his finding that the costs imposed by the Voice Contact Requirement were significant. I accept Mr Chamberlain's submissions on this point. First, on the evidence before the judge, the minimum cost to Uber was £700,000 if carried out from India and £3.4 million if from Uber's call centre in Limerick, as Mr Byrne, Uber's head of public policy for the UK, explained in his witness statement of 12 August 2016. On the lower of these, the price per journey would be less than 1p per trip or 0.1% to 0.2% of the average fare. Uber indicated it would pass these costs onto its customers. The burden imposed on Uber was thus nil and on its customers was very low.
79. Mr de la Mare's submitted the costs were substantial when considering the upfront capital costs in setting up and maintaining the facility. Further, he argued that to most businesses, one single cost accounting for 0.1% of their margin was more significant than Mr Chamberlain had suggested.
80. I reject Mr de la Mare's submissions under this head for the following reasons:

i) First, neither of these submissions can be maintained in the context where Uber is planning to provide a "click-to-call" facility anyway.

ii) Second, the judge failed to recognise or consider that the scale of the costs imposed on operators correlates directly with the extent to which customers use the service. If Uber is correct that its technologically inclined smartphone users would not use the voice contact facility to contact their operator, the costs imposed by the Voice Contact Requirement would reduce as a result. I am persuaded that any realised costs indicate use of the facility and a real customer benefit, for which TfL was entitled to regulate.

iii) Third, following on from this, Mr de la Mare submitted that it was not justified to impose the extra cost of the telephone call on Uber or its customers when certain categories of complaint are dealt with more simply, speedily and cheaply electronically. I cannot accept this submission. It fails to recognise that TfL has identified real benefits to all customers, even in non-emergency situations, and ignores that the Voice Contact Requirement also applies to PHV operators other than Uber.

Other operators

81. I consider it is necessary separately to consider the costs imposed on operators other than Uber, which at the time of the trial, was the only solely app-based operator. There are a number of other operators who use apps but also accept bookings by telephone. TfL's view was the impact on these operators would be lower. This is because they already have a telephone and employees answering phones to accept bookings. As such, the incremental cost for them is going to be much lower. I find no problem with this analysis.
82. The other group the judge considered was the approximately 700 driver-operators (of circa 117,000 drivers) to whom no more than two PHVs were registered. The judge held correctly that they too would be affected by the Voice Contact Requirement. Mr Chamberlain sought to argue that the judge overestimated the costs imposed on at least some of these operators as they would already have such a service in place to enable them to receive bookings while driving. However, I consider there is some force to the submission that the Voice Contact Requirement may have a potentially disproportionate impact on this particular group of operators in two ways. First, they are unlikely to be able to set up call centres abroad. Second, it requires somebody to be available at all times while the driver is driving, which is likely to be particularly burdensome.
83. Despite all of this, I do not consider the costs imposed on these operators as a result of compliance with Voice Contact Requirement will be disproportionate. It is important to provide some context for this conclusion. These operators must already comply with the 1998 Act. Under this Act, they are subject to the requirements that their bookings have to be accepted at the London operating centre (section 4(1)). This is a requirement not of TfL, but of the legislation. This means even where the operator is the driver and there is another individual, in another premises, accepting the calls, they are still required to have somebody at the London operating centre. This person must be at the centre licensed by TfL, where TfL can come and inspect the record of accepted bookings. In my view, it is not too much of a stretch to expect this same person, who accepts bookings, to also take the complaints.
84. Accordingly, I accept Mr Chamberlain's submission that the impact may not be as significant for this population of operators as the judge considered. I also find it relevant that they only account for 0.68% (at the time of trial) of active drivers. In my view, the judge held correctly that TfL mitigated any impact on this group by allowing them to club together. This would enable them to share the telephone line where complaints can be made (because there is no longer any requirement that calls to be taken at the London operating centre).
85. From the evidence, it is clear that TfL considered the potentially disproportionate impact on small operators including these single driver-operators. In my judgment, TfL came to a conclusion to which it was entitled, namely that "despite these impacts, the need for passengers to be able to speak to an operator at all times during their journey is an important public safety issue" (TfL Board Paper, March 2016).
86. In my judgment, the costs imposed on all PHV operators were fully considered by TfL and are not disproportionate to the public safety, equality and customer convenience benefits. For the reasons given above, I reject the Respondents' arguments in relation to issues 2 and 3 and conclude the Voice Contact Requirement satisfies the proportionality test in its full width. I would allow the appeal on both of these points.

Issue 4 - Whether the Voice Contact Requirement was a proportionate and lawful interference with Uber's freedom of establishment in any event and the judge erred in concluding otherwise?

87. This is proportionality stricto sensu, which was abandoned in the hearing. I do not consider that there would be any utility in expressing what would be obiter views on whether Lumsdon requires this question to be addressed separately.

Issue 5 - Whether it is unlawful that TfL has not imposed a similar requirement on taxis?

88. This was a further or alternative reason for dismissing TfL's appeal proposed in Uber's respondent's notice. It was addressed before the judge but is not discussed in his judgment. In my view, TfL's submissions on this issue are to be preferred.
89. To consider this point, one needs to understand the difference between a taxi and a PHV. Taxis are hackney carriages and are licensed to pick people up from the roadside, i.e. they can be hailed in the street. They are colloquially known as black cabs. PHVs are only permitted to pick up pre-arranged bookings (made with an operator) and are not permitted to pick people without a booking up from the side of the road. Colloquially, PHVs are sometimes known as minicabs. The point of dispute is that the Voice Contact Requirement is only imposed on PHVs and not on taxis.
90. I accept Mr Chamberlain's argument that taxis are governed by a completely different regulatory regime; a much older one that dates in part from the 19^th century. There is authority for this point in Eventech Ltd v Parking Adjudicator [2012] EWHC 1903 (Admin). This was a case brought by a different PHV operator, who alleged that the rule permitting taxis but not PHVs to use the bus lane in London was in breach of EU law. In this case, Burton J summarised the differences between the two regulatory regimes at [12]:

"It is important at this stage to set out the material differences between mini-cabs and black cabs. A Law Commission Consultation Paper issued earlier this year (No 203) described the 'two-tier licensing system' justified by 'the very different characteristics' of the pre-booked market and the market for hailing and picking up at ranks:

i) As set out in paragraph 9 above, only black cabs can 'ply for hire' without pre-booking.

ii) Black cabs are subject to 'compellability', dating from the London Hackney Carriage Acts 1831 and 1853, which requires that where a black cab at a rank or in the street accepts a passenger, the taxi must take the passenger anywhere that he wishes to go, within a prescribed distance or up to a prescribed journey time. There is no such 'cab rank' obligation on a minicab.

iii) Black cabs are instantly recognised by reason of their shape and size and the illuminated TAXI sign. This is because they must comply with the Conditions of Fitness ("CoF"), which contain a number of standards (including the requirement for the illuminated sign). Currently only two vehicle makes comply with the CoF. Minicabs can be of any colour and any design: there are some 700 different makes and models of vehicles presently licensed.

iv) The fares of black cabs are strictly regulated and can only be charged by reference to a taxi meter. Minicabs are free to charge their own fares and are not metered. According to Mr Griffin, the founder and chairman of Addison Lee, Addison Lee's fares are on average 35% cheaper than black cabs: the fare to be paid is quoted when the minicab is booked, irrespective of the duration of the journey, while black cab fares will of course vary depending upon the length of time that the journey takes.

v) Black cabs are required to be adapted for wheelchair access. There are no accessibility requirements for minicabs.

vi) Before being licensed, black cab drivers must undertake the 'Knowledge of London' examination process, which can take two to four years to prepare for ("the Knowledge"). Minicab drivers must before licensing undertake a topographical test, which generally takes a day. Addison Lee voluntarily imposes more extensive training on their drivers, by a short attendance at their driver training school. Black cab drivers must pass the Driving Standards Agency Advanced Driving Assessment: there is no similar requirement for minicab drivers."

91. The case went via this court, which did not give judgment but instead made a reference to the ECJ on a point about state aid. In Case C-518/13 R (Eventech Ltd) v Parking Adjudicator [2015] 1 WLR 3881, one of the issues the ECJ had to consider in relation to state aid was whether the selectivity requirement was met in that case, and that meant addressing the question as to whether black cabs and minicabs are comparable. It is Advocate-General Wahl's conclusion in his Opinion, which is set out at [59] – [61], that I consider to be the most relevant for present purposes:

"Black cabs and PHVs both compete on the market for pre-bookings, and it is common ground that competition is affected by the bus lane policy on that market. If that market were the only one relevant, they would clearly be comparable and, consequently, the bus lane policy would be selective.

However, one cannot simply extract a part of the business model of an undertaking and then limit the comparison with another undertaking to the segment thus extracted. Certainly that is true for the taxi business as well, certain specificities of which make it unjustified to limit the assessment of comparability to the market for pre-bookings alone.

In brief, taxis provide a service which supplements the existing methods of public transportation and which, in some ways, can arguably be assimilated to a universal public service. At a time when methods of communication were less developed, being able to hail a taxi from the street or to pick one up from a cab rank was an essential alternative to the other methods of transportation available. This is the reason why Black Cabs traditionally have a monopoly on 'ply for hire' journeys, and the same reason why taxis in many cities across Europe enjoy similar privileges, including the right to use bus lanes."

92. The ECJ then essentially adopted this reasoning again at [59] – [61] of its own judgment:

"In that regard, it must be stated, first, that the identification of the factual and legal situation of Black Cabs and minicabs cannot be confined to that prevailing in the market sector in which those two categories of conveyors of passengers are in direct competition, namely the pre-booking sector. It cannot seriously be doubted that all the journeys made by Black Cabs and minicabs are liable to affect the safety and efficiency of the transport system on all the road traffic routes in London.

Secondly, it must be taken into consideration that, by virtue of their legal status, only Black Cabs can ply for hire; they are subject to the rule of 'compellability'; they must be recognisable and capable of conveying persons in wheelchairs, and their drivers must set the fares for their services by means of a taxi meter and have a particularly thorough knowledge of the city of London.

It follows that Black Cabs and minicabs are in factual and legal situations which are sufficiently distinct to permit the view that they are not comparable and that the bus lanes policy therefore does not confer a selective economic advantage on Black Cabs."

93. While I found force in Mr de la Mare's argument that taxis and PHVs are in a materially identical position, at least when providing pre-booked journeys, I am not persuaded that this is enough. This was the answer provided by the ECJ to this court, which was the referring court in Eventech. It is true that Eventech was a state aid case and concerned a different measure, namely, the bus lane policy. However, in my judgment, it is applicable in this context of the freedom of establishment.
94. There are two main problems with Uber's argument that it is unlawful that TfL has not imposed a similar requirement on taxis that mean TfL's position on this issue is to be preferred.
95. First, a key distinction between PHVs and taxis is the lack of a person in the position of an operator in respect of the latter. Mr de la Mare submitted that TfL could impose the Voice Contact Requirement as a condition on taxi drivers' licences for permitting them to provide pre-booked services through the agency of a dispatcher. I cannot accept this argument. TfL cannot impose conditions on taxi licences in order "indirectly" to regulate taxi booking agents. I accept Mr Chamberlain's submission that this would be an improper use of that licensing power, where TfL does not have the power to regulate the booking agent directly.
96. Second, even if I were wrong on this and TfL could impose such a condition on the licences of taxi drivers, I agree with Mr Chamberlain's submission that it would be too difficult for TfL to monitor and enforce such a condition. Unlike with PHVs, it would not have the power to inspect the records of these booking agents because they are not licensed individuals. In a situation where TfL had evidence of a booking agent's non-compliance with the Voice Contact Requirement, I am not persuaded that there would be an effective method of enforcing compliance. With a PHV operator, such non-compliance could be enforced via threat of suspension or the removal of the licence. With a taxi booking agent, there is no licence to suspend or revoke. An option for TfL would be to threaten to revoke the individual licence of the taxi driver. In my judgment, this would be punishing a taxi driver for the non-compliance of a booking agent, over whom he has no control. I consider this to be unreasonable.

Disposition

97. For the reasons set out above, I would allow TfL's appeal.

Lord Justice Patten:

98. I agree.

Lord Justice Floyd:

99. I also agree.