Mr Z Bandi and Others v Bolt Operations OÜ and Bolt Services UK Ltd: 2206953/2021 and others
EMPLOYMENT TRIBUNALS
Case No 2206953/2021
Between
Mr Z Bandi and OthersClaimantBolt Operations OÜ and Bolt Services UK LtdRespondent
Before
Employment Judge A M SnelsonDate 7 November 2024
REASONS
[1]The Claimants in these consolidated proceedings are some 10,000 current or former private hire drivers who provide (or provided) transportation services to passengers in various locations around the UK in the name of Bolt.[2]The First Respondent (‘BOO’) provided payment processing and collection services between 11 June 2019 and 31 July 2022 but had no function relevant for present purposes after 1 August 2022, when the Bolt business model changed.2[3]The Second Respondent (‘BSUL’) holds licences to operate private hire businesses in London and various other cities and towns within the UK and is the entity through which those businesses are run.[4]We will refer to BOO and BSUL individually by name where appropriate. When speaking about the entire organisation or brand, it will be more convenient to refer simply to ‘Bolt’.[5]The Claimants bring claims under the Employment Rights Act 1996 (‘ERA 1996’), Part II, read with the National Minimum Wage Act 1998 (‘NMWA 1998’) and the National Minimum Wage Regulations 2015 (‘NMWR 2015’), for failure to pay the minimum wage, and the Working Time Regulations 1998 (‘WTR 1998’), for failure to provide paid leave. 2 The business model changed on 1 August 2022 in London and later in other parts of the UK.[6]The Respondents’ first line of defence is their denial that the Claimants were at any material time ‘workers’ entitled to the protection of the legislation on which they rely. Rather, they contend that they were and are independent contractors in business on their own account.[7]At a preliminary hearing for case management 13 March 2023 a preliminary hearing was listed to be held in public over 15 days commencing on 11 September 2024 to determine three issues (agreed between the parties) namely:a. Whether the Claimants are or were workers within the meaning of section 54(3)(b) NMWA 1998 and/or regulation 2 of WTR 1998; orb. Whether the Claimants are or were agency workers within the meaning of section 34(1) NMWA 1998 and/or regulation 36 of WTR 1998; andc. If the drivers are/were workers or agency workers, during what periods the drivers are/were working for the purposes of NMWA 1998 and/or WTR 1998.[8]Shortly before the public preliminary hearing the claims of many Claimants were settled. The result was that the hearing was limited to the claims of eight Sample Claimants (‘the Sample Claimants’) (all of whom were represented by the same legal team), namely Mr Abasse Abdou-Salami, Mr Baqer Ali, Mr Abdellatif Albab, Mr Abdul Kasim Wuraola, Mr Huseyin Akgul, Mr Shuhel Ahmed, Mr Alperen Nayir, and Mr Sofiane Belahcene.[9]In the course of case management, the parties sensibly agreed that directions should be given permitting the Claimants to make ‘rolling amendments’ to their claims on a quarterly basis without the need for formal application. A consequence of this is that events post-dating the commencement of proceedings (in late 2021 and early 2022) are not merely of evidential interest but strictly relevant to the claims.[10]The preliminary hearing came before us on 11 September this year. Mr Caspar Glyn KC and Ms Olivia-Faith Dobbie appeared for the Sample Claimants and Mr Davide Reade KC and Mr Rajiv Bhatt for the Respondents. We are very grateful to the advocates and the legal teams behind them for the careful preparation of the evidence and the persuasive presentation of the arguments. We also pay tribute to the good-natured spirit in which this fiercely contested litigation has been conducted.[11]We heard evidence from the Sample Claimants and, on behalf of the Respondents, Mr Gareth Taylor, Bolt’s Regional General Manager for Western Europe since 2023 but latterly on garden leave following service of notice of resignation, and Mr Bilal Nasir, a registered driver (or ‘Account Holder’ in Bolt’s parlance) and a user of its ‘Bolt Link’ facility.[12]The documents were voluminous, consisting of 11 lever arch files and separate digital materials exceeding 50,000 pages in length.[13]We also had the benefit of written openings and closing submissions from both legal teams.[14]Following three days of reading in, we heard evidence over seven days and then adjourned to allow time for the preparation of closing submissions, which were delivered on 27 September. Our private deliberations in chambers occupied two days, 1 October and 4 November. THE LEGAL FRAMEWORK Worker status
THE LEGAL FRAMEWORK
[15]The definition of a worker is to be found in ERA 1996, s230, which includes: (3) In this Act “worker” … means an individual who has entered into or works under (or, where the employment has ceased, worked under)–(a) a contract of employment, or(b) any other contract, whether express or implied and (if it is express) whether oral or in writing, whereby the individual undertakes to do or perform personally any work or services for another party to the contract whose status is not by virtue of the contract that of a client or customer of any profession or business undertaking carried on by the individual; and any reference to a worker’s contract shall be construed accordingly. We will refer to a contract within s230(3)(b) as a ‘limb (b) contract’ or a ‘worker contract’.[16]The same definitions apply under WTR 1998 and NMWA 1998.3[17]In Byrne Brothers (Formwork) Ltd v Baird & others [2002] ICR 667, Mr Recorder Underhill QC (as he then was), sitting in the EAT, offered this guidance on the proper interpretation of the definition of the limb (b) worker (para 17):(1) We focus on the terms “[carrying on a] business undertaking” and “customer” rather than “[carrying on a] profession” or “client” …(2) “[Carrying on a] business undertaking” is plainly capable of having a very wide meaning. In one sense every “self-employed” person carries on a business. But the term cannot be intended to have so wide a meaning here, because if it did the exception would wholly swallow up the substantive provision and limb (b) would be no wider than limb (a). The intention behind the regulation is plainly to create an intermediate class of protected worker, who is on the one hand not an employee but on the other hand cannot in some narrower sense be regarded as carrying on a business. (Possibly this explains the use of the rather odd formulation “business undertaking” rather than “business” tout court; but if so, the hint from the draftsman is distinctly subtle.) It is sometimes said that the effect of the exception is that the Regulations do not extend to “the genuinely self- employed”; but that is not a particularly helpful formulation since it is unclear how “genuine” self-employment is to be defined.(3) The remaining wording of limb (b) gives no real help on what are the criteria for carrying on a business undertaking in the sense intended by the Regulations — given that they cannot be the same as the criteria for distinguishing employment from self-employment. Possibly the term “customer” gives some slight indication of an arm’s-length commercial relationship — see below — but it is not clear whether it was deliberately chosen as a key word in the definition or 3 Reg 2 and s54(3) respectively. simply as a neutral term to denote the other party to a contract with a business undertaking.(4) It seems to us that the best guidance is to be found by considering the policy behind the inclusion of limb (b). That can only have been to extend the benefits of protection to workers who are in the same need of that type of protection as employees stricto sensu — workers, that is, who are viewed as liable, whatever their formal employment status, to be required to work excessive hours (or, in the cases of Part II of the Employment Rights Act 1996 or the National Minimum Wage Act 1998, to suffer unlawful deductions from their earnings or to be paid too little). The reason why employees are thought to need such protection is that they are in a subordinate and dependent position vis-a-vis their employers: the purpose of the Regulations is to extend protection to workers who are, substantively and economically, in the same position. Thus the essence of the intended distinction must be between, on the one hand, workers whose degree of dependence is essentially the same as that of employees and, on the other, contractors who have a sufficiently arm’s-length and independent position to be treated as being able to look after themselves in the relevant respects.(5) Drawing that distinction in any particular case will involve all or most of the same considerations as arise in drawing the distinction between a contract of service and a contract for services — but with the boundary pushed further in the putative worker’s favour. It may, for example, be relevant to assess the degree of control exercised by the putative employer, the exclusivity of the engagement and its typical duration, the method of payment, what equipment the putative worker supplies, the level of risk undertaken etc. The basic effect of limb (b) is, so to speak, to lower the pass-mark, so that cases which failed to reach the mark necessary to qualify for protection as employees might nevertheless do so as workers.[18]In the Supreme Court case of Bates van Winkelhof v Clyde & Co LLP and another [20114] 1 WLR 2047, in which the central issue was whether a member of a limited liability partnership was a limb (b) worker, Lady Hale DPSC offered these comments:[24]First, the natural and ordinary meaning of "employed by" is employed under a contract of service. Our law draws a clear distinction between those who are so employed and those who are self-employed but enter into contracts to perform work or services for others.[25]Second, within the latter class, the law now draws a distinction between two different kinds of self-employed people. One kind are people who carry on a profession or a business undertaking on their own account and enter into contracts with clients or customers to provide work or services for them. The arbitrators in Hashwani v Jivraj (London Court of International Arbitration intervening) [2011] UKSC 40 were people of that kind. The other kind are self-employed people who provide their services as part of a profession or business undertaking carried on by some-one else. The general medical practitioner in Hospital Medical Group Ltd v Westwood [2013] ICR 415, who also provided his services as a hair restoration surgeon to a company offering hair restoration services to the public, was a person of that kind and thus a "worker" within the meaning of section 230(3)(b) of the 1996 Act. Had Parliament wished to include this "worker" class of self-employed people within the meaning of section 4(4), it could have done so expressly but it did not. …[31]As already seen, employment law distinguishes between three types of people: those employed under a contract of employment; those self-employed people who are in business on their own account and undertake work for their clients or customers; and an intermediate class of workers who are self- employed but do not fall within the second class. Discrimination law, on the other hand, while it includes a contract "personally to do work" within its definition of employment (see, now, Equality Act 2010, s 83(2)) does not include an express exception for those in business on their account who work for their clients or customers. But a similar qualification has been introduced by a different route. …[34]In Cotswold Developments Construction Ltd v Williams [2006] IRLR 181 EAT, Langstaff J suggested, at para 53, that ". . . a focus on whether the purported worker actively markets his services as an independent person to the world in general (a person who will thus have a client or customer) on the one hand, or whether he is recruited by the principal to work for that principal as an integral part of the principal's operations, will in most cases demonstrate on which side of the line a given person falls".[35]In James v Redcats (Brands) Ltd [2007] ICR 100, Elias J agreed that this would "often assist in providing the answer" but the difficult cases were those where the putative worker did not market her services at all (para 50). He also accepted, at para 48, that ". . . in a general sense the degree of dependence is in large part what one is seeking to identify – if employees are integrated into the business, workers may be described as semi-detached and those conducting a business undertaking as detached – but that must be assessed by a careful analysis of the contract itself. The fact that the individual may be in a subordinate position, both economically and substantively, is of itself of little assistance in defining the relevant boundary because a small business operation may be as economically dependent on the other contracting party, as is the self- employed worker, particularly if it is a key or the only customer."[36]After looking at how the distinction had been introduced into the sex discrimination legislation, which contained a similarly wide definition of worker but without the reference to clients and customers, by reference to a "dominant purpose" test in Mirror Group Newspapers Ltd v Gunning [1986] ICR 145, he concluded, at para 59: ". . . the dominant purpose test is really an attempt to identify the essential nature of the contract. Is it in essence to be located in the field of dependent work relationships, or is it in essence a contract between two independent business undertakings? . . . Its purpose is to distinguish between the concept of worker and the independent contractor who is on business in his own account, even if only in a small way."[37]The issue came before the Court of Appeal in Hospital Medical Group Ltd v Westwood [2013] ICR 415, a case which was understandably not referred to in the Court of Appeal in this case; it was argued shortly before the hearing in this case, but judgment was delivered a few days afterwards. The Hospital Medical Group argued that Dr Westwood was in business on his own account as a doctor, in which he had three customers, the NHS for his services as a general practitioner, the Albany Clinic for whom he did transgender work, and the Hospital Medical Group for whom he performed hair restoration surgery. The Court of Appeal considered that these were three separate businesses, quite unrelated to one another, and that he was a class (b) worker in relation to the Hospital Medical Group.[38]Maurice Kay LJ pointed out (at para 18) that neither the Cotswold "integration" test nor the Redcats "dominant purpose" test purported to lay down a test of general application. In his view they were wise "not to lay down a more prescriptive approach which would gloss the words of the statute". Judge Peter Clark in the EAT had taken the view that Dr Westwood was a limb (b) worker because he had agreed to provide his services as a hair restoration surgeon exclusively to HMG, he did not offer that service to the world in general, and he was recruited by HMG to work as an integral part of its operations. That was the right approach. The fact that Dr Westwood was in business on his own account was not conclusive because the definition also required that the other party to the contract was not his client or customer and HMG was neither. Maurice Kay LJ concluded, at para 19, by declining the suggestion that the Court might give some guidance as to a more uniform approach …[39]I agree with Maurice Kay LJ that there is "not a single key to unlock the words of the statute in every case". There can be no substitute for applying the words of the statute to the facts of the individual case. There will be cases where that is not easy to do. But in my view they are not solved by adding some mystery ingredient of "subordination" to the concept of employee and worker. The experienced employment judges who have considered this problem have all recognised that there is no magic test other than the words of the statute themselves. … 19. In Autoclenz Ltd v Belcher and others [2011] ICR 1157 SC , the Supreme Court upheld the decision of the Employment Tribunal (‘ET’) that the claimant car valeters were, notwithstanding the express terms under which they worked, employed by the respondent company as ‘workers’ for the purposes of, inter alia, the Working Time Regulations 1998. Those terms, which were drafted on behalf of the company and the claimants were required to sign, declared that they were subcontractors, that they had to provide their own materials, that there was no obligation on them to provide any services or on the company to give them work, and that they were free to provide substitutes (suitably qualified) to carry out the work on their behalf. The ET found that the terms did not reflect the true agreement between the parties since, inter alia, the claimants were required to perform defined services under the direction of the company and were required to carry out the work offered and to do so personally (despite the substitution clause). Moreover, they would not have been offered the work if they had not signed the terms. In his judgment, with which all other members of the Court agreed, Lord Clarke resolved a conflict in the authorities as to whether the freedom of a court to disregard terms apparently agreed between contracting parties depended on whether or not those terms were a ‘sham’ in the sense that both parties intended to misrepresent the true nature of their obligations to one another. The learned Justice emphatically rejected that view, stating (para 29): The question in every case is … what was the true agreement between the parties. Lord Clarke also cited with approval (paras 25-26) these remarks of Elias J (as he then was) in Consistent Group Ltd v Kalwak [2007] IRLR 560 at para 57: The concern to which tribunals must be alive is that armies of lawyers will simply place substitution clauses, or clauses denying any obligation to provide or accept work, in employment contracts, as a matter of form, even where such terms do not begin to reflect the real relationship. At paras 33-35 Lord Clarke also stressed the need, particularly in the context of employment relationships, to take into account the relative bargaining power of the parties when considering whether the written contract represents the true agreement between them. 20. The requirement of personal performance is an essential characteristic of any contract of employment or worker’s contract. A genuine substitution clause may negate such a requirement. The key principles are set out in the judgment of Sir Terence Etherton MR in Pimlico Plumbers Ltd v Smith [2017] ICR 675 CA, at para 84: I would summarise as follows the applicable principles as to the requirement for personal performance. Firstly, an unfettered right to substitute another person to do the work or perform the services is inconsistent with an undertaking to do so personally. Secondly, a conditional right to substitute another person may or may not be inconsistent with personal performance depending upon the conditionality. It will depend on the precise contractual arrangements and, in particular, the nature and degree of any fetter on a right of substitution or, using different language, the extent to which the right of substitution is limited or occasional. Thirdly, by way of example, a right of substitution only when the contractor is unable to carry out the work will, subject to any exceptional facts, be consistent with personal performance. Fourthly, again by way of example, a right of substitution limited only by the need to show that the substitute is as qualified as the contractor to do the work, whether or not that entails a particular procedure, will, subject to any exceptional facts, be inconsistent with personal performance. Fifthly, again by way of example, a right to substitute only with the consent of another person who has an absolute and unqualified discretion to withhold consent will be consistent with personal performance. 21. In 2015 Uber drivers brought claims similar to those of the Bolt drivers against three Uber companies. At the time Uber operated an agency model similar in many respects to Bolt’s Agency Model, which we will describe in our narrative below. A preliminary hearing was held before the ET in 2016 which resulted in a decision that the Sample Claimants had been employed under worker contracts and had been ‘working’ under such contracts for the purposes of WTR 1998 and NMWA 1998 when they were within the ‘territory’ in which they were authorised to operate, had the Uber app switched on and were able and willing to accept trips. Appeals followed to the EAT, Court of Appeal and Supreme Court. All failed, that in the Court of Appeal notwithstanding a notable dissenting judgment by Underhill LJ. The judgments at all four levels are reported under the title of Uber BV and others v Aslam and Farrer at [2017] IRLR 4, [2018] IRLR 97, [2019] 3 All ER 489 and [2021] UKSC 5 respectively. Unless otherwise stated, any reference below to Uber is to the judgment of Lord Leggatt in the Supreme Court, with which the five other members of the Court agreed.4 22. In the following passage, the significance of which has been widely acknowledged, Lord Leggatt developed the reasoning in Autoclenz:65. Uber submits that what the Autoclenz case decided is that, for the purposes of applying a statutory classification, a court or tribunal may disregard terms of a written agreement if it is shown that the terms in question do not represent the “true agreement” or what was “actually agreed” between the parties, as ascertained by considering all the circumstances of the case including how the parties conducted 4 The Court originally sat in a constitution of seven Justices but one was not able to participate in the decision owing to ill-health. themselves in practice. If, however, there is no inconsistency between the terms of the written agreement and how the relationship operated in reality, there is no basis for departing from the written agreement.66. Uber further submits that there is no inconsistency in the present case between the written agreements between Uber, drivers and passengers and how that tripartite relationship actually operated in practice. In particular, Uber argues that the facts found by the employment tribunal (or alternatively, which the tribunal should have found) are consistent with the written terms stipulating that the drivers were performing their services under contracts made with passengers through the agency of Uber London and not for or under any contract with any Uber company. Uber submits that there is in these circumstances no legal basis for finding that the terms of the written agreements did not reflect the true agreements between the parties and hence for departing from the classification of the parties’ relationships set out in the contractual documentation.67. This argument was accepted by Underhill LJ in his dissenting judgment in the Court of Appeal. In his view (stated at para 120): “It is an essential element in that ratio [ie of the Autoclenz case] that the terms of the written agreement should be inconsistent with the true agreement as established by the tribunal from all the circumstances. There is nothing in the reasoning of the Supreme Court that gives a tribunal a free hand to disregard written contractual terms which are consistent with how the parties worked in practice but which it regards as unfairly disadvantageous (whether because they create a relationship that does not attract employment protection or otherwise) and which might not have been agreed if the parties had been in an equal bargaining position.” Interpreting the statutory provisions68. The judgment of this court in the Autoclenz case made it clear that whether a contract is a “worker’s contract” within the meaning of the legislation designed to protect employees and other “workers” is not to be determined by applying ordinary principles of contract law such as the parol evidence rule, the signature rule and the principles that govern the rectification of contractual documents on grounds of mistake. Not only was this expressly stated by Lord Clarke but, had ordinary principles of contract law been applied, there would have been no warrant in the Autoclenz case for disregarding terms of the written documents which were inconsistent with an employment relationship, as the court held that the employment tribunal had been entitled to do. What was not, however, fully spelt out in the judgment was the theoretical justification for this approach. It was emphasised that in an employment context the parties are frequently of very unequal bargaining power. But the same may also be true in other contexts and inequality of bargaining power is not generally treated as a reason for disapplying or disregarding ordinary principles of contract law, except in so far as Parliament has made the relative bargaining power of the parties a relevant factor under legislation such as the Unfair Contract Terms Act 1977.69. Critical to understanding the Autoclenz case, as I see it, is that the rights asserted by the claimants were not contractual rights but were created by legislation. Thus, the task for the tribunals and the courts was not, unless the legislation required it, to identify whether, under the terms of their contracts, Autoclenz had agreed that the claimants should be paid at least the national minimum wage or receive paid annual leave. It was to determine whether the claimants fell within the definition of a “worker” in the relevant statutory provisions so as to qualify for these rights irrespective of what had been contractually agreed. In short, the primary question was one of statutory interpretation, not contractual interpretation.70. The modern approach to statutory interpretation is to have regard to the purpose of a particular provision and to interpret its language, so far as possible, in the way which best gives effect to that purpose. In UBS AG v Revenue and Customs Comrs [2016] UKSC 13; [2016] 1 WLR 1005, paras 61-68, Lord Reed (with whom the other Justices of the Supreme Court agreed) explained how this approach requires the facts to be analysed in the light of the statutory provision being applied so that if, for example, a fact is of no relevance to the application of the statute construed in the light of its purpose, it can be disregarded. Lord Reed cited the pithy statement of Ribeiro PJ in Collector of Stamp Revenue v Arrowtown Assets Ltd (2003) 6 ITLR 454, para 35: “The ultimate question is whether the relevant statutory provisions, construed purposively, were intended to apply to the transaction, viewed realistically.” The purpose of protecting workers71. The general purpose of the employment legislation invoked by the claimants in the Autoclenz case, and by the claimants in the present case, is not in doubt. It is to protect vulnerable workers from being paid too little for the work they do, required to work excessive hours or subjected to other forms of unfair treatment (such as being victimised for whistleblowing). The paradigm case of a worker whom the legislation is designed to protect is an employee, defined as an individual who works under a contract of employment. In addition, however, the statutory definition of a “worker” includes in limb (b) a further category of individuals who are not employees. The purpose of including such individuals within the scope of the legislation was clearly elucidated by Mr Recorder Underhill QC giving the judgment of the Employment Appeal Tribunal in Byrne Bros (Formwork) Ltd v Baird [2002] ICR 667, para 17(4)5 …72. The Regulations referred to in this passage are the Working Time Regulations 1998 which implemented Directive 93/104/EC (“the Working Time Directive”); and a similar explanation of the concept of a worker has been given in EU law. Although there is no single definition of the term “worker”, which appears in a number of different contexts in the Treaties and EU legislation, there has been a degree of convergence in the approach adopted. In Allonby v Accrington and Rossendale College (Case C-256/01) [2004] ICR 1328; [2004] ECR I-873 the European Court of Justice held, at para 67, that in the Treaty provision which guarantees male and female workers equal pay for equal work (at that time, article 141 of the EC Treaty): “… there must be considered as a worker a person who, for a certain period of time, performs services for and under the direction of another person in return for which he receives remuneration …” The court added (at para 68) that the authors of the Treaty clearly did not intend that the term “worker” should include “independent providers of services who are not in a relationship of subordination with the person who receives the services”. In the EU case law which is specifically concerned with the meaning of the term “worker” in the Working Time Directive, the essential feature of the relationship between employer and worker is identified in the same terms as in para 67 of the Allonby judgment: Union Syndicale Solidaires Isere v Premier Ministre (Case C- 428/09) EU:C:2010:612; [2010] ECR I-9961, para 28; Fenoll v Centre d’Aide par le Travail “La Jouvene” (Case C-316/13) EU:C:2015:2000; [2016] IRLR 67, para 29; and Syndicatul Familia Constanta v Directia Generala de Asistenta Sociala si Protectia Copilului Constanta (Case C-147/17) EU:C:2018:926; [2019] ICR 211, para[41]As stated by the Court of Justice of the European Union (CJEU) in the latter case, 5 Cited above “[i]t follows that an employment relationship [ie between employer and worker] implies the existence of a hierarchical relationship between the worker and his employer” (para 42).73. In Hashwani v Jivraj [2011] UKSC 40; [2011] 1 WLR 1872 the Supreme Court followed this approach in holding that an arbitrator was not a person employed under “a contract personally to do any work” for the purpose of legislation prohibiting discrimination on the grounds of religion or belief. Lord Clarke, with whom the other members of the court agreed, identified (at para 34) the essential questions underlying the distinction between workers and independent contractors outside the scope of the legislation as being: “whether, on the one hand, the person concerned performs services for and under the direction of another person in return for which he or she receives remuneration or, on the other hand, he or she is an independent provider of services who is not in a relationship of subordination with the person who receives the services.”74. In the Bates van Winkelhof case at para 39, Baroness Hale cautioned that, while “subordination may sometimes be an aid to distinguishing workers from other self-employed people, it is not a freestanding and universal characteristic of being a worker.” In that case the Supreme Court held that a solicitor who was a member of a limited liability partnership was a worker essentially for the reasons that she could not market her services as a solicitor to anyone other than the LLP and was an integral part of their business. While not necessarily connoting subordination, integration into the business of the person to whom personal services are provided and the inability to market those services to anyone else give rise to dependency on a particular relationship which may also render an individual vulnerable to exploitation.75. The correlative of the subordination and/or dependency of employees and workers in a similar position to employees is control exercised by the employer over their working conditions and remuneration. As the Supreme Court of Canada observed in McCormick v Fasken Martineau DuMoulin LLP 2014 SCC 39; [2014] 2 SCR 108, para 23: “Deciding who is in an employment relationship … means, in essence, examining how two synergetic aspects function in an employment relationship: control exercised by an employer over working conditions and remuneration, and corresponding dependency on the part of a worker. … The more the work life of individuals is controlled, the greater their dependency and, consequently, their economic, social and psychological vulnerability in the workplace …” See also the illuminating discussion in G Davidov, “A Purposive Approach to Labour Law” (2016), Chapters 3 and 6. It is these features of work relations which give rise to a situation in which such relations cannot safely be left to contractual regulation and are considered to require statutory regulation. This point applies in relation to all the legislative regimes relied on in the present case and no distinction is to be drawn between the interpretation of the relevant provision as it appears in the Working Time Regulations 1998 (which implement the Working Time Directive), the National Minimum Wage Act 1998 and the Employment Rights Act 1996.76. Once this is recognised, it can immediately be seen that it would be inconsistent with the purpose of this legislation to treat the terms of a written contract as the starting point in determining whether an individual falls within the definition of a “worker”. To do so would reinstate the mischief which the legislation was enacted to prevent. It is the very fact that an employer is often in a position to dictate such contract terms and that the individual performing the work has little or no ability to influence those terms that gives rise to the need for statutory protection in the first place. The efficacy of such protection would be seriously undermined if the putative employer could by the way in which the relationship is characterised in the written contract determine, even prima facie, whether or not the other party is to be classified as a worker. Laws such as the National Minimum Wage Act were manifestly enacted to protect those whom Parliament considers to be in need of protection and not just those who are designated by their employer as qualifying for it.77. This point can be illustrated by the facts of the present case. The Services Agreement (like the Partner Terms before it) was drafted by Uber’s lawyers and presented to drivers as containing terms which they had to accept in order to use, or continue to use, the Uber app. It is unlikely that many drivers ever read these terms or, even if they did, understood their intended legal significance. In any case there was no practical possibility of negotiating any different terms. In these circumstances to treat the way in which the relationships between Uber, drivers and passengers are characterised by the terms of the Services Agreement as the starting point in classifying the parties’ relationship, and as conclusive if the facts are consistent with more than one possible legal classification, would in effect be to accord Uber power to determine for itself whether or not the legislation designed to protect workers will apply to its drivers.78. This is, as I see it, the relevance of the emphasis placed in the Autoclenz case (at para 35) on the relative bargaining power of the parties in the employment context and the reason why Lord Clarke described the approach endorsed in that case of looking beyond the terms of any written agreement to the parties’ “true agreement” as “a purposive approach to the problem”. 23. It will be necessary to return to Uber in our conclusions below. 24. In Commissioners for His Majesty’s Revenue & Customs v Professional Game Match Official Ltd [2024] UKSC 29 (‘PGMOL’) the Supreme Court considered an appeal concerning the employment status of referees and other match officials who provided matchday services at Level 1 football matches. It was common ground that there was no ‘umbrella’ contract6 and the only question was whether the officials were employed under service contracts during individual assignments. Lord Richards, with whom the other four Justices agreed, prefaced his discussion of the main issues with these observations: 30. First, there has been a tendency in some judgments, and still more in the submissions made in some cases, to focus unduly on the issues of mutuality of obligation and control and to treat all other terms of the contract and the surrounding circumstances of the parties' relationship as of less significance, or even as being relevant only if they negative the existence of an employment relationship. However, not only did MacKenna J himself [in Ready Mixed Concrete (South East) Ltd v Minister of Pensions and National Insurance [1968] 2 QB 497] make clear that mutuality of obligation and control were necessary, but not necessarily sufficient, conditions of a contract of employment, but there are decisions of high authority which emphasise the need to address "the cumulative effect of the totality of the provisions [of the contract] and all the circumstances of the relationship created by it" and to view "in the round, the relationship between the parties recorded in the agreement in the setting of the surrounding circumstances": White v Troutbeck SA [2013] EWCA Civ 1171, [2013] IRLR 949, per Sir John Mummery at paras 38 and 41, and see also O'Kelly v Trusthouse Forte plc [1984] QB 90, Lee Ting Sang v Chung Chi-Keung [1990] 2 AC 374 (PC), Hall v Lorimer [1992] 1 WLR 939 (Mummery J) and [1994] 1 WLR 209 (CA). 6 A contract governing the parties’ relationship between periods of work. 31. This broader perspective can work both ways. While an impermissibly narrow focus on mutuality of obligation and control might lead to a finding of employment, notwithstanding the importance of the surrounding circumstances in assessing the effect of the contract pointing the other way, the decision in White v Troutbeck illustrates the opposite: the claimant caretakers were employees, notwithstanding that they had day-to-day control over their own work. 32. Second, while as a pre-condition to a finding of employment there must be, under the contract, a sufficient degree of control by the putative employer over the putative employee, the extent of that control in any particular case remains a relevant factor in the overall determination of whether there exists an employment relationship. It is not the case that once the pre-conditions of mutuality of obligation and control are satisfied, they drop out of the picture as relevant factors in the overall assessment of whether a contract of employment exists: see Atholl House at para 76. 33. These are relevant factors for the correct approach in the present case. If mutuality of obligation and control are regarded as largely determinative, with only a minor role for other considerations, courts may be led to apply an unduly restrictive interpretation of control in order to prevent relationships which overall are not suggestive of employment from being characterised as such. Conversely, by according a real significance to the "totality of the provisions ... and all the circumstances of the relationship created by" the contract, a realistic approach can be taken to the issue of control. In other words, the bar to the existence of control need not be set at an unduly high level. 34. Flexibility in approach to deciding whether a sufficient level of control exists is critically important, given the ways in which employment practices have evolved and continue to evolve. Turning to mutuality of obligation, Lord Richards reviewed the authorities and concluded:[49]None of these authorities establishes that, where there is a single engagement (such as officiating at a particular match), there must be mutual obligations in existence before the engagement commences, for example before the referee arrives at the ground on the day of the match. On the contrary, there are authorities that establish the contrary. In Clark v Oxfordshire Health Authority, immediately following the passage quoted above, Sir Christopher Slade said, "I can find no such mutuality subsisting during the periods when the applicant was not occupied in a 'single engagement'".[50]The point is made in clear and direct terms in a number of authorities that a contract of employment may exist covering only the period while the employee carries out work for which he or she is paid. …[57]In my judgment, it is clear that the individual engagements of referees to officiate at matches satisfied the test of mutuality of obligation, which is a necessary but not sufficient condition to the existence of a contract of employment, and that the Court of Appeal was correct so to hold.[58]In developing his submissions on behalf of PGMOL, Mr Peacock KC said that it was not their case that the right to terminate an engagement without penalty necessarily negated mutuality of obligation. While not determinative, it was a relevant factor to be weighed when addressing the nature of the obligations owed. PGMOL's primary case was that it was relevant at the initial stage of deciding whether there existed the mutuality of obligation necessary for a contract of employment but, if that was not right, it was a relevant factor at the third stage of assessing overall whether the contract was one of employment. Mr Peacock said that it did not greatly matter to PGMOL's case whether it came at the first or the third stage.[59]In my judgment, the right to terminate is irrelevant at the first stage of determining whether there exists the mutuality of obligation required for a contract of employment. Where there exist the necessary mutual obligations under the contract, as was the case with each engagement to officiate at a match, and the contract remains in place, it satisfies the condition of mutuality. Mr Peacock’s submission that it did not greatly matter whether this point came in at the first or third stage overlooks that, if it did come in at the first stage and was held to be decisive on the facts of the particular case, the contract in question could not be one of employment. By contrast, if it is a relevant factor at the third stage, it is just one of many factors that may be relevant to determining the nature of the contract. On the subject of control, Lord Richards stressed (para 65) that the question at the first stage of analysis was whether, to adopt the formulation of Buckley J in Montgomery v Johnson Underwood Ltd [2001] ICR 819 CA, there is a ‘sufficient framework of control’ in the putative employer over the provision by the putative of his or her services. The framework of control needs to be assessed separately in respect of each contract (para 73). Sufficient control may take many forms (para 76). 25. The outcome of the appeal in PGMOL is instructive. The Supreme Court was satisfied that the first-stage tests of mutuality and control were met, but that meant only that, on a proper analysis, the officials might be employees. It therefore remitted the matter to the First-tier Tribunal to determine, on a broad assessment at the final stage of analysis of all material facts taken in the round, whether the officials were, or were not, employed for individual assignments under contracts of employment. 26. Mr Reade placed great reliance upon Secretary of State for Justice v Windle & another [2017] ICR 83 CA, in which the Court of Appeal was concerned with whether the absence of an umbrella contract was a factor of potential relevance to the assessment of the putative employee’s status when working. Reversing the EAT and restoring the decision of the ET, the Court of Appeal held that it was. Underhill LJ commented (para 23): I accept of course that the ultimate question must be the nature of the relationship during the period that the work is being done. But it does not follow that the absence of mutuality of obligation outside that period may not influence, or shed light on, the character of the relationship within it. It seems to me a matter of common sense and common experience that the fact that a person supplying services is only doing so on an assignment-by-assignment basis may tend to indicate a degree of independence, or lack of subordination, in the relationship while at work which is incompatible with employee status even in the extended sense. Of course it will not always do so, nor did the ET so suggest. Its relevance will depend on the particular facts of the case; but to exclude consideration of it in limine runs counter to the repeated message of the authorities that it is necessary to consider all the circumstances. 27. Mr Reade also drew our attention to Johnson v Transopco UK Ltd [2022] ICR 691, in which the EAT (HH Judge Auerbach and members) considered an appeal by the claimant, a self-employed black cab driver, against a decision of an ET that he and the respondent, an app-based transportation business, had contracted with each other as independent enterprises, the latter having the status of a customer of his taxi-driving business. The ET had attached significance to the fact that there was no umbrella contract between the parties and noted parallels with Windle. It also attached importance to the scant use which the claimant had made of the respondent’s app during the parties’ association, which accounted for only 15% of his overall income during the relevant period (the balance of his work coming from street hails). This, the ET reasoned, argued against the existence of a dependent work relationship. The ET further found that the respondent exercised little practical control over the way in which the claimant worked, noting that this was unremarkable given the skill and expertise which a black cab driver will be expected to bring to his work. Rejecting the claimant’s arguments based on Uber, the ET found important differences. Unlike Uber drivers, the claimant, to the extent that he was controlled by anyone, was controlled not by the respondent but by TfL, the regulator. The control which Uber exercised over its drivers, in relation to routes, remuneration and many other matters was not replicated in the claimant’s relationship with the respondent. Nor did the respondent seek to restrict contact between the claimant and his passengers, in marked contrast to the strictures which Uber imposed on its drivers. The EAT dismissed the claimant’s appeal. It noted that the assessment of employment status was a matter of fact and impression for the ET, and that it had reached permissible conclusions on the basis of facts it had found. Protection of wages – ERA 1998, s27A and Windle 28. Mr Glyn addressed a novel argument to us based on ERA 1996, s27A. That section renders any exclusivity clause in a zero hours contract unlawful and of no effect. The material provisions read as follows:(1) In this section “zero hours contract” means a contract of employment or other worker’s contract under which – (a) the undertaking to do or perform work or services is an undertaking to do so conditionally on the employer making work or services available to the worker, and (b) there is no certainty that such work or services will be made available to the worker.(2) …(3) Any provision in a zero hours contract which – (a) prohibits the worker from doing work or performing services under another contract or under any other arrangement, or (b) prohibits the worker from doing so without the employer’s consent, is unenforceable against the worker.(4) Subsection (3) is to be disregarded for the purposes of determining any question whether a contract is a contract of employment or other worker’s contract. 29. Mr Glyn submitted that these provisions precluded the Respondents from relying upon Windle in resisting the Claimants’ case on worker status. In our view there is no merit in this argument. Windle decides only that, depending on the circumstances, the absence of an umbrella contract may (or may not) point away from the existence of an employment relationship during individual assignments. We are quite unable to see how a provision neutralising exclusivity clauses in zero hours contracts can have any bearing on the applicability of the Windle principle. Mr Glyn contended (closing submissions, para 63) that ‘it cannot have been Parliament’s intention that when considering the claims of two zero-hours workers, Worker A and Worker B, where Worker A is made the subject of a void exclusivity clause, Worker B should have a less good claim to limb (b) status’. We do not see how such a consequence can be spelled out of the statutory language. The effect of s27A(4) is simply that the fact that any exclusivity clause is void cannot assist the putative employer in resisting the putative worker’s claim to hold worker status. In other words, whether there is an exclusivity clause in place or not does not affect the analysis on status. We can find nothing in s27A suggesting a parliamentary intention to affect the reasoning in Windle or restrict its application. Working time 30. By WTR 1998, reg 2(1), ‘working time’ means ‘any period during which [a worker] is working, at his employer’s disposal and carrying out his activity or duties.’ 31. NMWA 1998, s1(1) provides that a person who qualifies for the national minimum wage (a worker of above compulsory school age working, ordinarily working in the UK under his contract (see subsection (2)) ‘is entitled to be remunerated by his employer in respect of his work in any pay reference period at a rate which is not less than the national minimum wage.’ 32. Under NMWR 2015 ‘work’ is divided into four categories for the purposes of the national minimum wage scheme. We will identify and discuss the applicability of the categories in our analysis and conclusions below. THE FACTS 33. The population of Bolt drivers appears to be almost exclusively male. Purely for the sake of brevity, we will refer to them using masculine pronouns only. For the same reason we will refer to ‘trips’ rather than Bolt’s term, ‘journeys’. 34. Our narrative covers a period of some years in the course of which certain significant changes have occurred. We will use the present tense when referring to acts or events or states of affairs which have happened or subsisted throughout the period and continue to date. Where material changes have arisen, we will attempt to identify and date those changes. Setting the scene 35. The Bolt Group was founded in Estonia in 2013. Mr Taylor describes it as a leading mobility platform operating in 45 countries across Europe, Africa, Western Asia and Latin America. 36. In August this year, 100,494 drivers were registered with Bolt in the UK. Of these, all but 636 were solo drivers operating through conventional accounts with the Bolt App (‘the App’). Of the 636, some 583 were ‘Black Cab Drivers’ licensed to provide services in London through the App, and 53 were designated ‘Link Drivers’ under the ‘Bolt Link’ scheme introduced in October 2022. We are not concerned with Black Cab Drivers; the subject of ‘Bolt Link’ and ‘Link Drivers’ will be considered in due course. 37. Bolt is one of a number of private hire operators (‘PHOs’) in the UK. Mr Taylor told us without challenge that its market share over the last year fluctuated in London between 15% and 23% and across the rest of the UK between 8.27% and 9.88%. The market is highly competitive. Uber is comfortably the largest player. 38. Mr Taylor passed some comments on the phenomenon of ‘multi-apping’.7 This is the practice to which many drivers in the ‘gig economy’ resort of keeping more than one app open, or switching from one app to another, when looking for work with a view to taking the opportunities which seem most advantageous to them. As many witnesses testified, the practice is widespread. It is also entirely unobjectionable since (as we understand it) all PHOs now work, like Bolt, on the premise that drivers are entirely free to accept or reject any trip offered to them as they see fit. We accept that multi-apping is likely to have grown in popularity in recent years8, but we are mindful that the practice was debated in the Uber litigation and featured in the reasoning of the appellate courts in that case. 39. We heard much evidence and argument about the freedom of Bolt drivers to reject trips offered to them. Mr Taylor gave evidence about average rejection rates. The figure in respect of all drivers for the year ending on 31 July 2024 was 82%. The Sample Claimants’ average percentage for the same period was lower (56.2%).9 Mr Taylor also gave rejection rates for the Claimants and the Sample Claimants over the entire period of their usage of the App of 63% and 54.7% respectively. For a reason not explained to us, Mr Taylor did not supply the corresponding percentage for all drivers. The statistics which were provided to us were not challenged and we accept them as accurate in themselves. But they are of very limited utility because it emerged in evidence that they do not differentiate between trips within a driver’s chosen ‘radius’ (the area in which he has elected to make himself available to work) and trips which would take him outside it. This clearly matters: not surprisingly, Bolt does not expect drivers to accept out-ofradius offers. One of the Sample Claimants, Mr Ahmed, told us without challenge that he rejects only out-of-radius offers. We have no idea what proportion out-ofradius rejections bear to total rejections. Moreover, even if it can be said that there appears to be a trend of increasing rejection rates in respect of within-radius offers over recent years, we are in no position to interpret any such trend or spell out its 7 The term (and its cognates) may not yet have reached the Oxford English Dictionary but we hope to be forgiven for using it as a convenient shorthand. 8 Mr Taylor told us that, in a survey in 2021, ‘46.8% of drivers’ (presumably the percentage refers to those who responded) told Bolt that they worked with multiple ride-hailing companies. 9 Surprisingly, Mr Reade complained (submissions, para 41c) that this percentage was ‘skewed’ because of a significant minority having particularly low rejection rates. That, we would have thought, is how averages work: a significant proportion of relatively low (or high) values will push the mean down (or up). causes. The evidence provides us with no empirical foundation on which to attempt to do so. 40. It was not in issue before us that driving under the Bolt banner (as under any other app-based PHO) is not well paid. Many drivers, including some of the Sample Claimants, combine driving with other forms of paid work in order to make ends meet. A very high proportion of Claimants are not native English speakers and some have a poor grasp of spoken and/or written English. (Several of the Sample Claimants who gave evidence before us told us that they struggled particularly with written English.) No doubt some Bolt drivers are sophisticated and articulate, but we think it safe to assume that many will not have the benefit of a high standard of education. Bolt’s two business models 41. Between 11 June 2019 and 31 July 2022 Bolt operated under what it styled an agency arrangement, by which, on its case, BSUL served as an intermediary between the drivers and the passengers, providing booking and payment processing services through the App and taking a commission from the drivers in return. It strenuously denied providing transportation services (or any other service) to passengers. It denied being party to any contract for the conveyance of passengers or even being in the business of transport at all. Rather, it claimed (see Mr Taylor’s witness statement, para 55): (a) [BSUL] accepted the booking request from the Passenger, for regulatory purposes, and identified an Account Holder [Bolt’s term for driver] willing to undertake the Journey. (b) The Passengers agreed to pay a fare to the Account Holder who contracted with the Passenger to complete the Journey. (c) [BOO] provided payment collection and processing services to the Account Holders in collecting the fare from the Passengers, handled invoicing arrangements and retained a commission payment accordingly. 42. With effect from 1 August 2022 Bolt’s trading arrangements in London underwent a significant change. As of that date, as Mr Taylor explained (para 56): (a) [BSUL] accepts the booking request from or on behalf of the Passenger. (b) The Passenger agrees to pay a fare to [BSUL] (who under this model is responsible for providing transport services to the Passengers). (c) [BSUL] contracts with and pays a journey fee to Account Holders (who provide transport services – directly or via a Link Driver). As can be seen, the new arrangements excluded BOO entirely and BSUL became the only relevant Bolt entity. 43. Mr Taylor told us without challenge that the change of business model was imposed upon Bolt in late 2021 by Transport for London (‘TfL’) as a consequence of a judgment of the High Court.10 Mr Reade submitted that a later decision of the Court of Appeal11 called that judgment into question, but, for reasons given below, 10 R (United Trade Action Group Ltd) v Transport for London [2021] EWHC 3290 (admin) (cited as UTAG) 11 DELTA Merseyside Ltd & another v Uber Britannia Ltd [2024] EWCA Civ 902 (cited as Sefton) we do not think that it is appropriate for us to enter into that controversy. Whether TfL was right or not, Bolt not only implemented its direction in London but also made a later, parallel change to its business model across the rest of the UK. 44. For convenience, we will adopt Mr Taylor’s terminology, referring to the two models as the Agency Model and the Principal Model. 45. One consequence of the move to the Principal Model is that VAT is charged on the ‘fare’ which Bolt collects from the passenger. The commercial and financial implications of this (for the passenger, driver and Bolt) have not been explained to us. Bolt’s five sets of standard driver terms12 46. In its trading period of a little over five years to date, Bolt has introduced five sets of driver terms as well as making some modifications between these changes. There was no consultation with the drivers: Bolt merely notified them of the changes and made it clear that they must accept them if they wished to remain Bolt drivers. We will borrow the labels which Mr Taylor proposed. 47. The ‘Launch Terms’ sought to give effect to the Agency Model and were in place for less than three months (11 June 2019 to 21 August 2019). 48. The ‘Initial Terms’ applied for the remainder of the period of the Agency Model (22 August 2019 to 31 July 2022). They included (cl 2.3): Bolt does not provide Journeys. The Journeys are solely provided by You, as an independent business undertaking carried on by You to the Passenger. And (cl 6.2): Each Journey booking You choose in your absolute and sole discretion to fulfil through the Driver App will constitute a separate and distinct agreement in respect of the provision of that Journey by You to the Passenger, with each such agreement coming into force from the time You agree to undertake the specific allocated Journey booking made available to You by Bolt through the Driver App. You acknowledge that Bolt provides the Software Service to facilitate the Journey bookings and allocation thereof to Drivers. 49. The ‘Principal Model Terms’ (1 August 2022 to 24 April 2023) were the product of a radical reformulation of the Initial Terms. Bolt now candidly acknowledged a contract between it and the driver for the provision of driving services, but sought to exclude any notion of employment: Your engagement by Bolt as a Driver will be as a self-employed and independent contractor. As will be seen, although some measures to control or influence drivers’ conduct had been withdrawn by August 2022, many were continued under the Principal 12 We were told about a sixth set of terms, prepared in draft in 2021 but never brought into effect. We do not think it necessary to discuss them here. Model Terms. 50. One innovation during the currency of the Principal Model Terms was Bolt Link, which Bolt launched in October 2022. Our findings on this are set out under a separate heading below. 51. A further change, implemented between August 2022 and February 2023 was the introduction of ‘Custom Pricing’ (or the ‘Minimum Expected Journey Fee’). We deal with this under Pricing and payment - Principal Model below. 52. The ‘Early Cashout Terms’ (25 April 2023 to 31 July 2024) made minor changes to the Principal Model terms, including the introduction of the ‘Early Cashout’ scheme. This too is addressed under Pricing and payment - Principal Model below. 53. Finally, the ‘Current Terms’ (1 August 2024 onwards) incorporate minor changes to the Early Cashout Terms. New driver benefits: pension, holiday pay and minimum wage 54. On 1 May 2024 Bolt introduced what Mr Taylor called an ‘opt-in’ contributory pension scheme for drivers. Presumably his choice of language is to acknowledge the point that the scheme differs from those made for the benefit of persons admitted to be employees or workers, where automatic enrolment is mandatory, subject to the worker’s right to opt out. We were told that only 6.5% of drivers have signed up for the pension scheme. In cross-examination, Mr Taylor was shown documents demonstrating that Bolt reported to TfL in March 2024 that it planned to fund its 3% contributions (the drivers would pay 5%) by increasing its ‘Take Rate’ (the difference between the passenger’s ‘Fare’ and the ‘Journey Fee’ paid to the driver, as more fully explained below) and that it did not intend to share this information with the drivers. Mr Taylor did not appear to dispute that, at least for the 93.5% of drivers not electing to sign up for the pension scheme, the increase in the ‘Take Rate’ could only leave them worse off overall. 55. With effect from 1 August 2024, ‘Bolt Boost’ was launched. This provides for a ‘holiday pay supplement’ calculated at 12.07% of their Journey Fee income and a ‘journey fee supplement’ designed, according to Bolt, to ensure that drivers’ remuneration does not fall below the level of the National Living Wage.13 Mr Taylor told us (para 129) that there was a ‘strong competitive rationale’ underlying this innovation. He did not accept that it reflected any weakness (or perception of weakness) in Bolt’s position on worker status. Conditions applicable to drivers throughout: those relied on by Bolt 56. Mr Reade for the Respondents prayed in aid particularly the following terms, obligations and practices which have applied to drivers throughout. (1) Drivers are completely free to use, or not use, the App and as to when, and 13 The Claimants say that if this was Bolt’s purpose, the scheme falls well short of achieving it. how frequently, they use it. (2) Drivers are free to provide their services to other providers and organisations, including Bolt’s competitors and/or to undertake any other work or employment. (3) It is for drivers to provide their own plant and equipment (most notably, a motor vehicle). (4) Drivers are free at any time to terminate their relationship with Bolt or simply cease using the App altogether.(5) Drivers treat themselves as self-employed for tax and national insurance purposes and HMRC treats them as such. Conditions applicable to drivers throughout: those relied on by the drivers 57. In addition to those considered under separate headings in this narrative, Bolt’s driver terms have always featured the following stipulations. (1) The right to use the App is personal to the driver and non-delegable. (2) Bolt has sole and unfettered discretion as to whether or not to accept any request for a ride by a proposed passenger. (3) The driver is obliged to provide his services personally and may not do so through a company or any form of incorporated association. (4) Bolt has unfettered freedom to change the driver terms. (5) Drivers’ entitlement to continue to use the App following any change of terms by Bolt is conditional upon their acceptance of the new terms and their continued use of the App is treated as manifesting such acceptance. Terms between Bolt and passengers 58. Bolt’s passenger terms in operation during the period from June 2019 to August 2022 are consistent with Bolt’s case on the legal effect of the Agency Model. So, for example, cl 2.3 stipulated: Once an individual Driver confirms that they (sic) will accept Your Journey request, You will enter into a separate agreement with the Driver for the provision of the Journey on such terms and conditions as You agree with the Driver. Bolt does not provide Journeys and is not a party to Your agreement with the relevant Driver. 59. Bolt’s General Platform Terms: Passengers, which seem to have come into effect on 24 September 2022 and to have been updated at least once since, are consistent with the Driver Terms issued on and after the introduction of the Principal Model. They specify the terms on which passengers may use the App, set out payment arrangements, incorporate an ‘Acceptable Use Policy’ (breach of which may result in the passenger’s suspension), make provision for the handling of complaints and resolution of disputes, reserve to Bolt an unfettered right to amend its terms and deal with sundry other matters. The regulatory regime
THE FACTS
[60]The private hire industry is the subject of close regulation under primary and secondary legislation.14 The licensing authority in London is Transport for London (‘TfL’); elsewhere, the responsibility falls upon the applicable local government body.[61]Mr Reade rightly pointed out (without challenge) that many driver terms shown to us are mandated by regulatory obligations to which Bolt is subject. Mr Glyn rightly replied (without challenge) that(a) many are not and(b) to the extent that they are, that cannot stand as a factor arguing against worker status (Uber, para 102). The Bolt system A Bolt trip in outline[62]Any first-time passenger (who must be not less than 18 years old) begins by downloading the Bolt Rider App, signing in to enter payment details and agreeing to Bolt’s ‘Rider Terms’. To use the App, the passenger must enter a destination and specify the desired trip category or categories. Bolt then issues an estimate (or more than one if multiple trip categories are specified). Where the passenger elects to proceed Bolt will ordinarily accept the booking (although it reserves to itself an absolute discretion to refuse) and proceed to offer the trip to drivers in accordance with its allocation criteria (largely based on proximity to the pick-up point, but see further below). In order to be eligible to receive an offer of a trip, a driver must have logged on to the App in the usual way, specifying the trip category or categories which he is able to fulfil and the ‘radius’ within which he wishes to work. The offer will contain scant information: the initial or first name of the passenger, the passenger’s rating, the first four characters of the postcode of the destination and the estimated ‘journey fee’ (the sum which the driver can expect to receive). (Disclosure of the (partial) destination information was brought in by Bolt in early 2020: prior to that the driver was told nothing of the destination until after he had accepted the trip.) The driver has a short period (usually, it seems, 10-15 seconds) in which to accept the offer. If he does not respond in that time his silence is treated as a rejection. Where the first driver does not take the trip, the offer goes to the next in line according to the allocation criteria until the trip is accepted. On acceptance the driver and passenger are placed into contact with one another, but this is done exclusively through the App and precludes either from obtaining the telephone number of the other. By this means they are able to communicate, for example to agree precise arrangements and/or location for pickup. The passenger is provided with the driver’s first name only.[63]Bolt treats the trip as starting when the passenger enters the car. By means of the App, the driver signals that the trip has begun. Bolt monitors the journey throughout. At the destination, the driver signals through the App that the trip has been completed. He is also expected to give a rating to the passenger. Likewise, the passenger is requested to rate the driver. On completion of the trip, the driver is free to seek further driving opportunities. 14 In London, the Private Hire Vehicles (London) Act 1998; elsewhere in England & Wales, the Local Government (Miscellaneous Provisions) Act 1976; in Scotland, the Civic Government (Scotland) Act 1982.[64]Bolt immediately processes the passenger’s payment (into its bank account) and, in due course, pays the driver his fee. Bolt also handles the (electronic) ‘paperwork’. These aspects are considered more fully below. Trip categories[65]By way of the App, Bolt markets a range of driver services, designated ‘trip categories’, intended to cater for particular requirements or preferences of service users. In addition to the standard ‘Bolt’, there are, for example, ‘XL’ (larger vehicles with capacity for six or more passengers), ‘Executive’ (premium vehicles), ‘Pet’ (vehicles with capacity for up to two small to medium-sized pets) and ‘Electric’ (fully electric vehicles). Larger or more prestigious vehicles attract higher fares.[66]While many vehicles can be considered only for ‘Bolt’ trips, any driver whose vehicle satisfies Bolt’s definitions of more than one of the various categories is (subject to the allocation criteria to which we next turn) eligible to have driving opportunities in any and every applicable category offered to him. Allocation of trips – Bolt’s criteria[67]The Respondents’ Statement of Facts sets out (at para 7) the system by which Bolt identifies drivers to whom to offer any driving opportunity arising through the App. We accept the account given as correct and can do no better than to reproduce it here. Drivers are offered Journeys when they are logged into the Bolt App based on their trip category, and their proximity to the Passenger who has requested the relevant Journey (and the relevant trip category). Proximity is determined by reference to the Driver’s estimated time of arrival (ETA) to the Passenger’s pick-up location. ETAs are calculated on a standard basis, so do not take into account personal characteristics of the particular Driver (i.e. the speed or any other similar characteristic or criteria). If a Driver is in close proximity to the Passenger but already engaged on another Journey via the Bolt App, they would not normally be within the relevant ETA range to be offered the Journey unless the drop-off location for their current Passenger was near the potential Passenger and the relevant Driver was nearing the end of that Journey such that their ETA to the potential Passenger, after the drop-off, was within the relevant ETA range. The ETA range applicable to each offer of a Journey will vary based on market conditions. For example, if a significant number of potentially eligible Drivers are in close proximity to the relevant Passenger, the relevant ETA range could be quite small (i.e. Drivers within 1-2 minutes of the Passenger). On the other hand, if there were only a limited number of potentially eligible Drivers in close proximity, the ETA range could be large (i.e. Drivers within 10 minutes of the Passenger). Once the Bolt App has identified the Drivers who are in the relevant ETA range, it will offer the Journey to the Driver with the lowest ETA within the ETA range and, if not accepted by that Driver, to the Driver with the next lowest ETA and so forth. With effect from the launch of the custom pricing feature … The Journey Fee payable to the potentially eligible drivers within the ETA window is also taken into account. The Bolt App will consider the pricing structures in place for all potentially eligible Drivers within the relevant ETA window and determine the median price (this may be the same for all Drivers if the relevant Drivers in the ETA window have elected to use the dynamic pricing model; but may be different if one or more potentially eligible Drivers in the ETA range has adopted custom pricing). Having determined the median price, any Driver with a custom price higher than the median price would no longer be considered potentially eligible for the journey. That median price, if higher than the dynamic price otherwise applicable, will then be the new dynamic price (reflecting the fact that dynamic pricing changes to reflect market conditions). All potentially eligible Drivers at that new dynamic price (being the median price of potentially eligible Drivers, and being at or above the custom price of all potentially eligible Drivers) would then be offered the Journey based on their ETA as described above. There are no other instances in which a Driver would be given lower priority for offers of Journeys, offered fewer Journeys or offered lower payment for Journeys (of the same trip category) than another Driver.[68]For completeness, we should add one further point. As we understand it, the system for allocating work does not take account of the driver’s specified radius, with the consequence that, if he finds himself outside his radius but close to the passenger’s proposed pick-up point, his physical location will not preclude him from receiving an offer from the App. It seems that in such circumstances the App marks (or marked) the offer as ‘optional’, which may well reflect the former state of affairs in which (as we record below), drivers were at risk of suffering adverse consequences if their rate of rejection of offers of work within their radius exceeded a specified percentage. (Although, as already noted, the rejection statistics put before us by Bolt did not differentiate between ‘optional’ and other rejections.) Bolt Link[69]As already mentioned, Bolt launched the Bolt Link scheme in October 2022. We heard a lot of evidence and argument about it, no doubt because Bolt’s case before us was that it had the effect of ‘negating an obligation of personal service’ and thus was ‘fatal’ to the Claimants’ case on worker status (opening skeleton, para 101). Mr Taylor told us (para 87) that the aim of Bolt Link was to seize the opportunity of the change in business model to create a new ‘entrepreneurial’ feature which would serve the interests of drivers as well as Bolt. He explained that it involved building on existing Bolt technology called the Bolt Fleet Manager Portal to assist ‘fleet owners’ in the management of their fleets, by providing them with access to essential data. A ‘fleet owner’ is ‘a person who owns a number of vehicles and arranges for them to provide services via the Bolt App – including directly to the fleet owner’s clients’ (para 87). We did not gain from Mr Taylor’s evidence an understanding of(a) why or how the existing technology needed to be modified or(b) why he thought that the innovation would give Bolt an advantage over the competition (para 89), given the fact that Uber was operating a fleet scheme before the Employment Tribunal hearing in 2016 (see its Reasons in that case, para 90, f/n 44).[70]The Bolt Link driver terms to which we were referred can be found in the Current Terms. These include: 14.3 Bolt UK recognises your freedom to engage other individuals licensed to provide private hire services to fulfil Journeys by associating them with your account as a Driver in accordance with these Private Hire Terms. However, this may not include any individual licensed to provide private hire services who has previously had their account terminated by Bolt UK for a repeated, serious or material breach of contract, or who has engaged in conduct which would have provided grounds for such termination had they been a direct party to these Terms. … 14.4 When associating any other individuals licensed to provide private hire services with your account, you continue to bear full responsibility for ensuring that your obligations under these Terms are met. All acts and omissions of any individuals licensed to provide private hire services associated with your account will be treated as though those acts and/or omissions were your own. You must ensure that any individual licensed to provide private hire services associated with your account does not put you in breach of these Terms. You are wholly responsible for the remuneration of any such individuals licensed to provide private hire services associated with your account and the invoicing arrangements as set out in these Terms between you and Bolt UK will continue to apply.[71]As careful reading of the terms confirms and as was established in evidence at rather greater length than seemed to us necessary, Bolt Link as ‘rolled out’ manifestly did not have the effect of entitling a Bolt account holder (‘Driver A’) to allocate any trip offered to him to another driver ‘associated with’ his account (‘Driver B’). As in any other case, Driver A had three choices when offered a trip on the App: accept the offer, reject it or do nothing (tantamount to a rejection). In the standard way, if Driver A did not accept the offer, it fell to Bolt to allocate the trip to the next driver in line, applying the selection criteria already explained. Only in the unlikely event of that driver being Driver B,15 would the opportunity pass to him and even then it would do so not by Driver A substituting him but by Bolt offering it to him direct by virtue of his being the driver closest to the pick-up point. No published term of Bolt’s said otherwise (in particular, cl 14.3 did not say that any Driver B could be ‘engaged’ by Driver A to fulfil a Journey offered to Driver A). Neither of its witnesses said otherwise. No communication shown to us between Bolt and the drivers or between Bolt and TfL said otherwise. Mr Reade in oral closing argument did not say otherwise.16[72]The fare payable in respect of the services of a Bolt Link driver is the same as that payable in any other case. What changes is that payment to Driver B is the responsibility of Driver A and Bolt takes no interest in the dealings between the two (although it cautions every Driver A not to infringe the Modern Slavery legislation). It seems that Bolt’s main selling point to putative Driver As was (and perhaps still is) that they could make money by paying their Driver Bs less than the prescribed fare, and retain the difference. As for any Driver B, he operates like any conventional Bolt driver (save that he will almost certainly have to accept less than the standard fare). He is bound by Bolt’s terms in the usual way. He makes himself available in the usual way, by using his log-in credentials. As already mentioned, he gets offers of driving work only by being at the top of the pecking order according to Bolt’s allocation criteria. It seems that Bolt Link’s only selling point for him is that, if he does not possess a car, he will be able to use a car from Driver A’s fleet and so will be spared the expense of funding one himself.17[73]Drivers were puzzled by the Bolt Link idea and struggled to understand how it might benefit them. As already noted, and despite vigorous promotion by Bolt, the take-up was pitifully low: 53 out of around 100,000 registered drivers signed up as ‘Driver B’ drivers. 15 Or Driver C, D E etc (ie any other driver ‘associated with’ Driver A’s account). 16 As we will explain in our analysis below, Mr Reade argued instead that to focus on whether Driver A could pass a driving opportunity to Driver B was to miss the point. 17 We cannot think of any reason why someone with a satisfactory car might wish to be a Driver B.[74]In his witness statement, Mr Taylor described Bolt Link as ‘an exciting new feature which showcased the entrepreneurial benefits provided by [Bolt’s] business in that it further facilitated the development and growth of micro-businesses’ (para 89), but under cross-examination and by reference to numerous documents, he eventually accepted that ‘a primary reason’ for its introduction was to defeat the drivers’ case on worker status. Pricing and payment – Agency Model[75]Under the Agency Model, Bolt determined the ‘Fare’ to be paid by the passenger.18 This might take the form of an ‘actual Fare’ calculated by reference to distance, journey time and other factors or an ‘Upfront Fare’, based on an estimate. The Initial Terms provided (cl 10.4) that an ‘actual Fare’ would be applied in place of an ‘Upfront Fare’ where the characteristics of the trip as envisaged at the time of the estimate are ‘impacted’ by ‘unexpected circumstances.’ It was for Bolt alone to decide in any case whether to substitute the ‘actual Fare’ for the ‘Upfront Fare’ and, if so, to determine what the ‘actual Fare’ should be.[76]Bolt stipulated the ‘commission’ payable to it by the driver.[77]Bolt could alter the level of its ‘commission’ at will and did so, merely informing drivers of the change it had made.19[78]Under the Agency Model, Bolt managed the (electronic) paperwork associated with each trip. It took payment immediately from the passenger through the App and issued the passenger with a receipt. That receipt was not shared with the driver. It then issued what was presented as an ‘invoice’ from the driver to the passenger, the latter being identified only by a first name or initial. This document was made available to the driver through the App but was not sent to the passenger (who, of course, had already paid for the trip and received a receipt from Bolt). There was a further system of ‘invoices’ between Bolt and the driver, by which Bolt purported to demand payment of its ‘commission’. But Bolt had no reason to invoice the driver, who was never in Bolt’s debt. Bolt already held the passenger’s payment in full, which included its ‘commission’. The only outstanding liability was upon Bolt, to pay the driver for his driving service. That payment was not made at once but weekly, unless ‘Early Cashout’ applied.20 Accordingly, Bolt had the advantage of immediate receipt of the full sum paid by the passenger and the driver had the disadvantage (subject to ‘Early Cashout’) of having to wait for his payment. Pricing and payment - Principal Model[79]Under the Principal Model, Bolt takes a ‘Fare’ from the passenger at once 18 Necessarily, the freedom to set the overall ‘Fare’ applied equally to each of the many elements by reference to which it was calculated (starting rate, rate per mile, rate per minute etc). 19 In July 2021 Bolt increased the commission rate in London from 15% to 20%. There was no prior negotiation or even discussion. Drivers who complained were simply told that they must accept the change or end their association with Bolt. 20 As to ‘Early Cashout’ see further below. In any week when Bolt’s liability to a driver stood at less than £15, the weekly payment would be stood over to the following week. and pays accrued ‘Journey Fees’ to the driver at a later point, usually on a weekly basis. The driver terms, with which Bolt’s practice corresponds, preclude the driver from seeing the ‘Fare’ and the passenger from seeing the ‘Journey Fee’. The difference between the two is retained by Bolt and is known as the ‘Take Rate’. This may (or may not) incorporate a ‘Service Fee’, apparently something in the nature of a booking fee payable by the passenger. The evidence on this was less than clear but for practical purposes it seems to have little significance. Whether it includes a ‘Service Fee’ or not, the Take Rate is Bolt’s overall income from the relevant driving service.[80]At the start of the hearing, we expressed surprise at the Respondents’ stated position that the Take Rate was irrelevant and, for reasons of commercial sensitivity, had not been divulged. Mr Reade supported that stance on grounds which appeared to us less than persuasive but since Mr Glyn seemed content to comment on the suppression of this information and did not pursue any application before us, we left the matter there. Much later in the hearing, Mr Reade argued that the Tribunal could calculate the Take Rate for itself on the basis of the voluminous data which had been disclosed. We could not safely embark on that task without the assistance of the Respondents, which was not offered to us. In any event, it is for parties to put disclosable evidence before the Tribunal not to supply it with a mass of impenetrable material and leave it to its own devices in seeking to interpret it. We are left with the simple fact that Bolt has elected not to disclose the Take Rate.[81]We have already mentioned the ‘Early Cashout’ scheme. Eligible drivers (eligibility terms are set by Bolt and include a requirement to have completed at least 25 trips) are allowed to request payment for trips at once. as already noted, Bolt receives payment from passengers at once, but normally pays the drivers their ‘Journey Fees’ weekly. Under the scheme, the driver gets the benefit of immediate payment but it comes at a cost: every Journey Fee paid early is discounted by an amount determined by Bolt.[82]Under the Principal Model, Bolt operates a system of receipts and invoices similar in some respects to that in place under the Agency Model. In particular, it addresses to itself an invoice in the name of the driver in respect of each trip undertaken. This document is shared with the driver and so reflects the sparse information which he is allowed (for example, the passenger’s name is not shown because the driver is never allowed to know it). It also specifies the ‘Journey Fee’, which Bolt has determined without reference to the driver. As under the Agency Model, the (electronic) paperwork is generated and managed exclusively by Bolt, with no involvement on the part of the drivers.[83]We have already mentioned Custom Pricing, an innovation which Bolt brought in between August 2022 and February 2023. Mr Taylor told us that this was another scheme to reflect the ‘entrepreneurial nature of the Bolt App’ by giving drivers the chance to influence their pay. They do this by stipulating on the App, as an alternative to the Bolt standard (‘Dynamic’) rate, a minimum rate per mile (the rate per mile is one of up to 11 variable elements which determine the driver’s pay for each trip), within parameters set by Bolt. All potentially eligible drivers (according to location and trip category or categories) are then considered. If all rates entered are the same, the usual selection criteria are applied without more. If not, the median rate is calculated and any driver who has named a higher rate falls out of contention. Offers are then made in the normal sequence to all who remain. An adjustment is also made where any driver’s chosen minimum rate per mile falls below the Dynamic rate (which changes constantly according to market conditions), automatically replacing it with the Dynamic rate. Bolt’s case is that the purpose of the parameters and this extra adjustment is to promote entrepreneurship while avoiding the twin evils of drivers pricing themselves out of work on the one hand or working at below cost on the other.[84]In messages promoting Custom Pricing which must have been sent between August and December 2022, Bolt promised to provide drivers with ‘indications for how your price fits with others in the marketplace’ and (apparently for the same purpose) an ‘Available Driver List’ showing some details of up to 11 other drivers nearby, including their ‘journey price’, a term which was not defined. These promises were not honoured. No ‘Available Driver List’ has been published and no relevant information has been shared with drivers concerning the price per mile bids put in by other drivers in the locality.[85]Mr Taylor told us that Custom Pricing had been ‘very successful’ and that a little over one million journeys had been completed using it in the 12 months to 31 July 2024. By our arithmetic, that seems to equate to approximately 10 trips per registered driver per annum. Further facts relevant to ‘control’ issues Recruitment or ‘onboarding’ of drivers[86]Bolt advertises to attract new drivers. Those interested apply by means of an application form, which Bolt reviews. It invites those in whom it is interested to attend ‘activation sessions’. Personal documentation (some requiring verification on regulatory grounds, some not) is checked. Some training is given. Precise recruitment or ‘onboarding’ processes appear to have changed from time to time and may well vary by location. Vehicle checks[87]It is for all drivers to provide their own vehicles, which must comply with Bolt’s requirements concerning age,21 appearance, seating capacity, cleanliness and mechanical condition. Identity checks[88]Personal identity documents produced to Bolt and verified on recruitment/’onboarding’ are checked and re-checked at frequent intervals. Some of these checks are regulatory requirements, many are not.[89]Drivers are required to provide ‘selfie’ photographs of themselves for the 21 It seems that a stipulation requiring all vehicles to be under 10 years old was withdrawn in 2022. purposes of routine biometric checks carried out by Bolt. These include, or are supplemented by, a ‘Superman’ test which is designed to cross-refer the biometric verification to the spot where Bolt, by means of geolocation technology, locates the App. If any discrepancy is found, the driver is liable to be ‘blocked’ (suspended) while the latter is investigated. The ‘Superman’ test is not a regulatory requirement. Discretion to accept requests from passengers[90]As already mentioned, Bolt reserves to itself an untrammelled discretion to accept or reject any passenger’s request for a Bolt trip. The discretion is exercised without reference to drivers. Communication between drivers and passengers[91]There is no question of any communication between a passenger and driver until after the driver has accepted the trip. It is at that point that Bolt puts the two into contact with one another by telephone, but only through the App and in such a way as to preclude either from learning the telephone number of the other.[92]Drivers are explicitly prohibited from contacting passengers after trips. The only exception is where the purpose is to return an item of lost property and that facility is exercisable only through the App and only for a period of 24 hours immediately following the trip. The prohibition is liable to be enforced under Bolt’s graduated system of enforcement (‘education’, warning, ‘block’ (ie suspension), as to which see further below).[93]The prohibition of contact between drivers and passengers is obviously very important to Bolt. Mr Albab told us of an occasion when he had been assaulted in a ‘road rage’ incident, which his passenger had caught on camera. His request to Bolt for permission to contact the passenger to obtain evidence about the assault was refused, as was his alternative request for Bolt to contact the passenger on his behalf. Service standards[94]Under the Initial Terms, drivers were mandated to ‘commit to each passenger’ that they would provide a ‘high level of service’. The same terms noted that Bolt monitored drivers’ activity ‘[i]n order to maintain Bolt’s integrity and reputation’ and obliged drivers to ‘act in a professional manner in accordance with the business ethics applicable to providing passenger transportation …’ The Initial Terms also recorded that Bolt was entitled to terminate its relationship with any driver and ‘block’ his access to the App if (inter alia) he provided ‘poor service (as evidenced by Passenger complaints and/or ratings) and such poor service standards would, in Bolt’s reasonable opinion, bring Bolt … into disrepute.’[95]Under the Principal Model Terms, Bolt requires drivers to adhere to ‘Minimum Service Standards’ and sets out several kinds of conduct (explicitly stated to be a non-exhaustive list) which Bolt regards as not consistent with those standards. These include, ‘actions giving rise to concerns of safety, health or conduct from Passengers’.[96]Bolt also seeks to promote and uphold service standards by means other than driver terms. These include generalised messaging and interventions directed at individual drivers. An example of the former is an email to all drivers of 8 July 2020 which included: ‘… we are placing a greater emphasis on the quality of conduct from both drivers and passengers. Adapting to the changing benchmark of quality standards prompted by the emphasis of safety from the pandemic is paramount to maintaining Bolt’s high quality of service.’ Mr Abdou-Salami gave two examples of the latter. The first occurred in February 2021 when, as a consequence of a complaint by a passenger, he was ‘blocked’ from access to the App for 24 hours because of an allegedly offensive smell in his car. Bolt’s messages to him included: ‘As you know, your car is expected to be in impeccable condition to provide a safe and comfortable service to the passengers’ and ‘We sincerely hope that you will take this warning seriously and that in future we will not be receiving any similar reports or we may be required to take further actions.’ The second occurred in December 2023, when he was ‘blocked’ for almost a month on the strength of a complaint by a passenger, which he denied, that he had had ‘another passenger’ in his car. Bolt’s messages to him included: ‘please note that any future reports of this kind may lead to further action on your account. This is not to assume guilt, but just for safety precautions.’ Routes[97]As we have noted, Bolt relies on the fact that, under its terms, drivers are free to complete journeys by the routes of their choice. But we agree with Mr Glyn that this freedom is not quite what it first appears to be. Under the Initial Terms, drivers were declared free to complete journeys by ‘any reasonable route’ but they were warned that in the event of any journey not being completed in accordance with the terms generally (which, as noted above, included the obligation to provide a ‘high level of service’) any refund or compensation paid to a passenger might be set off against the fare. And under the Principal Model Terms the Minimum Service Standards (breach of which may result in penal ‘blocking’) oblige drivers to convey passengers to their destinations in a ‘reasonably direct and efficient manner’.[98]In practice, Bolt does not respect the freedom of choice in respect of route which its standard documentation proclaims. In September 2021, Mr Wuraola’s pay was docked following a passenger’s complaint about the route he had taken and he was subjected to a penal ‘block’ of three days’ duration. A contemporary message from Bolt included: ‘Please be mindful to always take the route suggested by the app …’ On 4 September 2022, Mr Ali completed a 29-mile journey for which he was paid £30. His request for a review was rejected by Bolt with the following explanation: ‘… it was noted that during part of the journey you took a different route other than the one suggested on the App, please always use the suggested route unless the passenger suggests otherwise.’ There was nothing before us to suggest that these adjudications were anything out of the ordinary. Cross-border working[99]Bolt seeks to control where drivers work with a view to ensuring that their activities are limited to the areas in which they are licensed. Not unnaturally, it points out that this intervention is merely intended to ensure compliance with regulatory obligations.22 Bolt exercises territorial control through a graduated system of disciplinary measures, from warnings to ‘7-day block’ to ‘hard block’ (tantamount to dismissal). In its application to TfL for its 2024 licence Bolt stated: ‘We have an active monitoring and warning process for managing drivers so that the majority of journeys in London are carried out by TfL licensed drivers.’ Bolt’s use of drivers’ data[100]Bolt collects large volumes of data about drivers. Such data relate to identification, geolocation, passenger feedback and ratings, usage of the App, journeys fulfilled, ‘driver records’ (dealing with, for example, ‘guidance’, ‘blocks’ and other interventions by Bolt and any associated reviews or appeals) and communication and correspondence records between Bolt and drivers. In ‘Privacy Notices’ issued after July 2022, Bolt justifies collecting and processing these data as assisting it to perform its services and provide ‘a quality and enjoyable service to everyone’. The amount and range of data handled does not appear to have changed materially when the Principal Model replaced the Agency Model. ‘Blocking’/suspension[101]Bolt operates several types of ‘block’ including ‘temporary blocks’, ‘permanent blocks’ and ‘hard blocks’. The temporary block requires no explanation. A ‘permanent block’ remains in place indefinitely unless and until lifted by Bolt. The affected driver is barred from using the App to earn his living but retains access to it for other purposes (for example to receive messages from Bolt or to review information stored on it relating to earlier work performed by him). A ‘hard block’ is similar to a ‘permanent block’ save that the driver concerned is excluded entirely from the App and the block can only be lifted by specially authorised Bolt managers.102. ‘Blockers’ may apply ‘generic’ or ‘custom’ blocks. The former are routinely applied in response to certain categories of act or omission on the part of a driver or where a particular circumstance or state of affairs arises (such as required documentation expiring). For these, standard templates are used. The latter attract bespoke treatment, where Bolt considers that special circumstances warrant ‘blocking’.103. ‘Generic blocks’ are frequently imposed where drivers have failed to update relevant documentation or to satisfy Bolt that they have done so. Similarly, drivers are not infrequently ‘blocked’ as a consequence of failing identification and/or authentication checks. ‘Blocks’ of these kinds are typically lifted quite quickly, once Bolt is satisfied that the relevant concern has been met. No doubt many ‘blocks’ relate to documentation, but we were shown evidence dating from September 2020 suggesting that, in the cases of almost half of the 935 drivers then suspended, the reason was something other than a failed documents check. 22 Although, as already noted, such motivation does not make it any the less a material factor (Uber, para 102). ‘Statistical blocks’[104]‘Statistical blocks’ featured large in the evidence and argument before us. These were applied where, according to statistics collected by Bolt, a driver’s performance fell short of a prescribed minimum standard. They divided into four main categories, which we will address in turn. 105. ‘Acceptance rate blocks’ were applied by Bolt from the outset (June 2019) until, it seems, May 2021. Initially the minimum acceptance rate was set at 70%. It was soon reduced to 60% and then, in November 2019, 50%, in February 2020, 40%,23 and in June 2020, 20%. Bolt communications to drivers warned of the risk of incurring ‘blocks’ if the minimum rates were not achieved. Some of these messages were confusing, adopting the language of ‘confirmation rates’ rather than ‘acceptance rates’. This change of terminology was explained by internal communications (to which drivers were never privy) adverting to the ‘legal risk of mentioning acceptance‘. It seems that, for the purposes of enforcement of applicable acceptance rates, rejection of any driving offer outside they driver’s nominated radius would not be counted, although (as already noted) the rejection figures put before us by Mr Taylor did not exclude offers of out-of-radius trips.[106]Although we accept that Bolt did not apply acceptance rate ‘blocks’ after May 2021, internal communications demonstrate that this was because of a perceived ‘reclassification’ (ie worker status) risk. Bolt remained alive to the commercial benefits of maintaining high acceptance rates and staff were encouraged to remind drivers to keep their rates high24. A message to Mr Wuraola on 10 September 2021 offered ‘useful tips to prevent getting blocked in future’, which included, ‘Maintain a low rejection/cancellation rate’. On the evidence put before us, we are not persuaded by Bolt’s undocumented assertion that it circulated messages to London drivers on two occasions in May 2021 communicating its decision to end acceptance rate ‘blocks’.25 That information may or may not have been generally shared with new drivers or drivers re-joining the organisation at some point after May 2021 but we have no clear evidence on this one way or the other. Some training material generated in August 2022 appears to omit any reference to acceptance rates. A training document which seems to date from September 2023 does state that there is no minimum acceptance rate in London, which might be interpreted as suggesting that the situation outside the capital was different.[107]We make the following findings on the Sample Claimants’ knowledge of Bolt’s decision not to apply minimum acceptance rates with effect from about May 2021.(1) Mr Abdou-Salami We find that Mr Abdou-Salami presented as a sincere and frank witness. He did not have a strong command of detail. We are not 23 This percentage may have been increased to 60% for a few months (see the Respondents’ statement of facts). 24 An internal Bolt communication in July 2021 stated: ‘… we are in the clear using acceptance rate as a carrot …’ 25 As far as we are aware, Bolt has not even claimed to have directed any such written communication to any drivers working outside London. persuaded that Bolt sent written notification to him, in May 2021 or at any other time, that the 20% minimum acceptance rate then in force had ceased to apply. He routinely used Bolt’s ‘auto-accept’ facility, at least in part because he believed (no less after than before May 2021) that failing to maintain a very high acceptance rate might result in adverse consequences for him.(2) Mr Ali was a satisfactory witness. He joined Bolt in December 2021. We find that he was made aware at that time that no minimum acceptance rate applied to him.(3) Mr Albab We find that Mr Albab was a satisfactory witness. We accept his evidence that he did not receive any written notification from Bolt, in May 2021 or at any other time, that the current acceptance rate had been discontinued. He may have formed the impression, perhaps through conversations with his peers, that Bolt was no longer enforcing a minimum acceptance rate, but he never received any confirmation from Bolt to that effect.(4) Mr Kasim Wuraola was a satisfactory witness. We find that he received no written notification, in May 2021 or at any other time, that the acceptance rate then in force had ceased to apply. As already noted above, by way of example only, he was warned in September 2021 he would do well to keep his rejection and cancellation rates low in order to avoid being ‘blocked’ in the future.(5) Mr Akgul was a satisfactory witness. He was a driver working through the Bolt App from 2019 until the summer of 2023. Some six months before his departure, as a consequence of a conversation with a friend, he was given to understand that Bolt was no longer applying ‘blocks’ based on rejection rates. He never received that information from Bolt.(6) Mr Ahmed was a satisfactory witness. We find that Bolt did not communicate to him, in May 2021 or at any other time, that it was no longer enforcing a minimum acceptance rate.(7) Mr Nayir was a satisfactory witness. He was suspended by Bolt in or about May 2021 but re-engaged in or about May 2022. We find that he was made aware on his return that no minimum acceptance rate was being applied.(8) Mr Belahcene was a satisfactory witness. We find that he was not made aware, in May 2021 or at any other time, that Bolt would no longer enforce minimum acceptance rates.[108]Drivers have throughout been nominally free under Bolt’s terms to cancel any trip which they have accepted. But cancellation rate ‘blocks’ were applied between June 2019 and June 2022 when the proportion of trips accepted and then cancelled by the driver crossed the relevant Bolt threshold.[109]Moreover, there is no evidence that Bolt ever communicated to drivers that it had ceased to operate a system of cancellation rate ‘blocks’, and communications to drivers after June 2022 appeared to carry the strong implication that such a system continued. In particular, the ‘Minimum Service Standards’ introduced under the Principal Model Terms listed among actions inconsistent with the Minimum Service Standards (and thus capable of being visited with ‘blocking’), ‘repeated and consistent cancellation of Journeys by Driver prior to arriving at the pickup location but after they have elected to fulfil a Journey’. The words ‘repeated’ and ‘consistent’ are not defined and are for Bolt to interpret and apply as it sees fit. 110. ‘Driver rating blocks’ were applied by Bolt up to June 2022. Minimum ratings varied as follows: June to July 2019: 3.8/5; July to October 2019: 4/5; October 2019 to May 2020: 4.2/5; June to December 2020: 4.5/5; December 2020 to June 2022: 4/5. For some trip categories, special driver rating requirements applied.[111]Again, Bolt did not inform drivers, in June 2022 or at any other time, that the driver rating ‘block’ scheme was revoked. The training materials produced in September 2023 state: ‘The current minimum driver rating is 4.0’. As before, the Bolt driver’s current rating appears on the home screen of his App when he logs on. Conduct ‘blocks’[112]The subject of ‘blocking’ on grounds of drivers’ conduct has already been touched on in the context of Bolt’s terms under the Agency Model requiring ‘high standards of service’ and the Minimum Service Standards stipulated under the Principal Model. Internal documents list numerous ‘blocking scenarios’. Many offences are listed as instances of ‘fraud’, although, as Mr Taylor accepted in evidence, they cannot reasonably have that label attached to them.26 Bolt routinely applies ‘blocks’ on the strength of allegations which have not been tested or substantiated, often simply because a complaint is similar to one which has been made about the same driver before.27 In September 2022, Bolt introduced a policy of ‘blocking’ any driver who had not responded to a complaint or allegation within six hours of being notified of it, but it seems that a very similar policy was applied under the Agency Model. Working time ‘blocks’[113]Any driver who works for an uninterrupted period of 12 hours is automatically blocked for six hours. Bolt justifies this practice, which is not a regulatory requirement, as a means of ensuring a service of ‘high quality’. Blocks generally – impact on drivers and their inability to raise challenges[114]Bolt exclusively administers and controls the application and lifting of ‘blocks’. No provision is made for any right of appeal or any other mechanism by which an affected driver may question or challenge any ‘blocking’ decision.[115]In evidence which was wholly or very largely unchallenged, the Sample Claimants provided many illustrations of the ‘blocking’ process in operation, the powerlessness of drivers to resist or influence it and the serious impact of being ‘blocked’ on their capacity to earn a living. 26 One example is having a warning light showing on the dashboard. 27 A ‘Global Blocking Matrix’ applicable with certain modifications in London and around the UK sets out a graduated system of penalties based on allegations. It advises staff to apply penalties ‘… depending on the amount (sic) of previous offences and whether we have proof or not. We do not contact the reported user for more information …’ The bundle includes a version of the Matrix dated July 2024. Activity score[116]Bolt operates a metric called ‘activity score’, which is calculated by reference to a driver’s response to offers of trips within his ‘radius’. Where a driver accepts and completes a trip, the score goes up. Where he rejects or cancels a trip or fails to respond to an offer of a trip, the school goes down. The activity score and the current driver rating both appear on the App whenever the driver logs on. Activity score is designed to encourage drivers to accept and complete more trips. No target score is set and drivers are not at risk of penalties related to activity score. Driver score[117]‘Driver score’ is mentioned in numerous documents disclosed in these proceedings. The Respondents’ position before us is that the subject does not arise because Bolt has never implemented a ‘driver score’ system in the UK. Accordingly, it has not given disclosure on this aspect. Bolt told TfL on 29 October 2021 that it planned to introduce: ‘a new driver scoring system which aggregates different metrics (eg cancellation rate and passenger’s post-trip ratings) into an easy to understand percentage score so that drivers can monitor their own performance. In the case of declining performance, the system would result in the usual enforcement measures, culminating in ‘blocking’. The following month, Bolt told TfL that it was not pursuing the ‘driver score’ plan. But numerous documents shown to us appear to demonstrate that, between November 2021 and April 2023 at least, Bolt did apply a ‘driver score metric’ and that its purposes in doing so included developing a prioritised system of work allocation to favour higher-scoring drivers. Such a system, called ‘Driver Preferential Allocation’ (‘DPAL’), was referred to in the Principal Model Terms and linked Privacy Notice. Bolt’s position on disclosure in regard to driver score and DPAL (as to which disclosure was refused on the same grounds) was, in our view, most unsatisfactory. Campaigns and incentives[118]Drivers are frequently exposed to special ‘campaigns’ designed to encourage them by means of financial inducements to accept more Bolt trips. These are time-limited and location-specific, often targeting specific groups of drivers selected by reference to various classes of data collected through the App. Some campaigns are run by local BSUL staff; others are the province of a team of growth analysts based at the Bolt headquarters in Tallin. 119. ‘Surge pricing’ is a quite different form of inducement. As already explained, drivers’ pay is made up of many different elements. One, pay-per-mile, is constantly adjusted by Bolt according to current market conditions. The greater the demand, the higher the rate climbs. Naturally enough, drivers generally wish to be operating in areas where the best rates apply. But Bolt’s interests may not always coincide with those of the drivers. Under the Agency Model, Bolt would identify ‘surge’ areas and the pay rates which they attracted. But that information is now withheld from drivers under the Principal Model. Instead, Bolt uses a colour coding system to encourage drivers to move into, or closer to, areas where it wants them to be available, but suppresses detailed information which would enable them to make an informed judgment as to whether it would serve their interests to do so. Challenges and disputes[120]As we have already explained, Bolt retains for itself a complete discretion in dealing with complaints and concerns raised by drivers in relation to matters of payment and ‘blocking’. This state of affairs merely exemplifies the wider practice: in any dispute the decision rests with Bolt. In a document dating from June 2022, Bolt explains its system for internal complaint-handling. This provides for investigation of complaints not excluded in limine as being (among other things) expressions of dissatisfaction with Bolt's general policies or with the exercise of its discretion where no professional or other misconduct, mistake, error, lack of care, unreasonable delay, bias or lack of integrity is alleged, or where the complaint is ‘clearly unfounded’ or a renewal of a complaint already made. No provision is made for any independent adjudication or any formal appeal mechanism.28 Risk[121]Although Bolt requires drivers to provide and fund their own vehicles and keep them in good order and condition, it volunteers to provide them with cover for, or at least a contribution towards the cost of, vehicle cleaning resulting from soiling by any Bolt passenger. This is subject to terms of eligibility specified by Bolt. The level of the protection is also in Bolt’s discretion. ‘Soiling charges’ have been set, variously and without reference to the drivers, at £60 (2022), £80 (2023) and £40 (2024).[122]Despite Bolt maintaining that drivers are and always have been acting as independent contractors in business on their own account, it appears to be common ground that, in the event of payment for any trip not being recovered from the passenger, Bolt accepts the loss and pays the driver nonetheless, and that this practice has obtained throughout, under the Agency Model as well as under the Principal Model. THE RIVAL CASES[123]The prodigious written submissions on both sides (which are testament to the skill and stamina of all four counsel and the legal teams behind them) defy summarising and we will leave them to speak for themselves. For present purposes, it is sufficient to identify here what we take to be the central planks of the cases advanced on each side, in so far as it is necessary to engage with them. In our analysis and conclusions below, it will be necessary for us to address in closer detail some parts of those arguments.29 28 A dissatisfied complainant may refer the matter to ‘one of the sworn advocates who has been entered to the ‘list of mediators’ by the Estonian Bar Association. According to the Claimants’ Opening Case, the attached hyperlink takes the reader to a website which appears to be written in Estonian, and to a ‘‘404’ Internet error’. It seems that any challenge to Bolt’s decision on a complaint, if one could be mounted at all, would be at the complainant's expense. 29 When we do, purely for brevity, we will refer to them as the submissions of leading counsel only. Bolt’s case[124]Mr Reade’s main points in relation to the Agency Model were the following.(1) Bolt was ‘not Uber’ and the relevant circumstances were different in many important respects.(2) In particular, the environment in which digital platforms have been operating since the time with which the Uber litigation was concerned had changed dramatically with the arrival of many new competitors, increasing rejection rates and the ubiquity of multi-apping.(3) There had at the same time also been parallel, significant changes in the terms under which drivers in the ‘gig economy’, including Bolt drivers, operate, most notably with the phasing out of penalties for, among other things, high rejection and cancellation rates.(4) The driver terms are of great importance and cannot be disregarded: neither Uber nor Autoclenz (nor any other authority) says otherwise.(5) The focus must be on the effect of the driver terms and not on the purpose for which Bolt drafted them.(6) The driver terms correspond with the reality of the relationship between Bolt and the drivers.(7) The purposive interpretation for which the Claimants contend ignores the reality that, certainly in the context of the courier driver market since June 2019, drivers using the App are not in a dependant working relationship with Bolt but are free to seek work through numerous other platforms.(8) The changes in market conditions make the observations of Underhill LJ in Windle, para 23 all the more compelling.(9) Other authorities such as Transopco also lend significant support to Bolt’s case on worker status.[125]Turning to the Principal Model, Mr Reade’s main submissions may be summarised in this way.(1) Bolt Link is not a sham. It is a real scheme which forms part of drivers’ contractual entitlements under the Principal Model and drivers have engaged with it. The motivation behind it is irrelevant.(2) Bolt Link negates an obligation of personal service by entitling any Driver A to engage any number of drivers to provide trips by associating them with his Bolt account.(3) The existence of Bolt Link is what matters: that the Sample Claimants, or any other drivers, have not chosen to take up the chance to join the scheme is neither here nor there.(4) The employment status of associated drivers of Driver A (Drivers B, C, D etc) is not in issue in this preliminary hearing.[126]Addressing both Models, Mr Reade submitted that, even if, contrary to Bolt’s case, a contract for personal service or services was shown, it was one under which, having regard to the key factors, drivers were and are in business on their own account and Bolt is properly seen as their customer.[127]As to the question when the drivers, if ‘workers’, are ‘working’ for the purposes of WTR 1998 and NMWA 1998, Mr Reade submitted:(1) As the parties agree, the drivers are ‘working’ when driving on an accepted trip (ie travelling to collect the passenger and carrying the passenger to his or her destination).(2) The Claimants’ further claim that, provided that they are not multi-apping, they are ‘working’ when waiting for the chance of a trip is ‘absurd’.(3) A claim based on that view of the law would be impossible for Bolt to contest since it would not be privy to key information (such as whether, and if so for how long, a driver had been multi-apping).(4) When waiting for the chance of a trip, a driver is not, for the purposes of WTR 1998, reg 2(1) ‘working, at [Bolt’s] disposal or carrying out his activity or duties’. Indeed, the mere ability to sign up with another app-based PHO (whether exercised or not) would preclude a driver from satisfying any of the three requirements of reg 2(1).(5) By the same token, when waiting for the chance of a trip, or travelling with a view to getting the chance of a trip, a driver is not performing ‘work’ for the purposes of NMWA 1998 and NMWR 2015.(6) If, contrary to Bolt’s case, classification of time for the purposes of NMWR 2015 is required, drivers are engaged on Time Work, alternatively Output Work, alternatively Unmeasured Work. The Claimants’ case[128]Mr Glyn submitted that it was obvious that the Claimants were employed by Bolt throughout under agreements indistinguishable from those examined by the ET, the EAT, the Court of Appeal and the Supreme Court in Uber and found at every stage to be ‘worker’ contracts. The driver terms on which Bolt relied were tactical, lawyer-driven devices contrived to misrepresent the simple reality of the relationship between the parties.[129]In relation to the Agency Model, all or most of the key factors in Uber applied no less to Bolt.[130]In relation to the Principal Model, given that here there was no dispute about a direct contractual relationship between Bolt and the drivers, the organisation’s position on personal service was hopeless. It was left with no case at all (on personal service) in respect of the period from August to October 2022 and its argument that the introduction of Bolt Link in that month made any difference was obviously untenable. Quite simply, Bolt Link did not, and could not, ‘negate’ the requirement for personal service by every Driver A. It did not, and could not, operate to create a power substitution.[131]In relation to both Models, the contention that, if any contract with Bolt requiring personal service by the drivers was shown, it was one under which Bolt had the status of a customer of the drivers was untenable as having no correspondence with reality. The reasoning in Uber was applicable.[132]As to when the drivers were properly to be seen as ‘working’ for the purposes of WTR 1998 and NMWA 1998/NMWR 2015, Mr Glyn made it clear that the drivers would not press claims in respect of any period in which they were multi-apping. But subject to that qualification, he submitted that they were working when in the territory for which they were licensed, had the App switched on and were ready and able to accept trips. Again, the core reasoning in Uber applied. ANALYSIS AND CONCLUSIONS Preliminary observations[133]We agree with the parties that the questions concerning employment status must be addressed by reference to the Agency Model and Principal Model separately. The changes brought in on 1 August 2022 were substantive and not merely presentational. The fact that, from the perspectives of passengers and drivers, little seemed to change is nothing to the point.30[134]Mr Reade went further. He submitted that, since the Tribunal must have regard to all relevant circumstances when considering the question of employment status, it cannot leave out of account material developments during the periods of each of the two business models. As a matter of principle, we think that must be right. An important change in the terms under which the drivers worked (at some point within the period of either model) might dictate the conclusion that it was accompanied by a change in the legal status of those drivers. This said, the point has its limits. We do not think that it falls to us to examine the relationship between Bolt and the drivers on a week-to-week or month-to-month basis. Nor did Mr Reade urge us to do so. He did submit that we should have particular regard to the relaxation by Bolt during the period of the Agency Model of certain constraints upon the freedoms of drivers. We agree and we have done so. Worker Status[135]We have returned to the statutory language and reminded ourselves of the repeated message of the authorities that there can be no substitute for applying the words which Parliament has enacted. The Agency Model[136]This part of the debate began with a disagreement about the significance of remarks in the judgment of Lord Leggatt in Uber concerning the lawfulness of Uber’s agency model in London. The learned Justice strongly doubted whether, as a matter of law, the company’s arrangements were compatible with the Private Hire Vehicles (London) Act 1988 but expressed no concluded view on that point and held in any event that the agency argument failed for want of any factual foundation for the argument that Uber London Ltd, the PHO in London, acted as agent for drivers when accepting bookings (paras 45-56). Mr Reade submitted that Lord Leggatt’s analysis on the licensing aspect was obiter and that in any event Bolt’s case is distinguishable from Uber’s on the facts. As already noted, he further contended that the UTAG decision, relied on by TfL as necessitating the switch to the Principal Model, had been called into question by the judgment of the Court of 30 At the time, Bolt went to some trouble to reassure drivers that the change in model would have little practical effect upon them. Appeal in Sefton (both cases already cited). Mr Glyn submitted that Sefton was of no assistance, having been decided on facts quite different from those under consideration here and by reference to legislation drawn in different terms to that applicable to private hire licensing in London.[137]We prefer to leave these submissions where they are. Rightly, we have heard limited argument on the licensing point and it is not an area of law with which can claim to be familiar. In addition, Lord Leggatt’s remarks stopped short of expressing a concluded view and so must be treated as both provisional and obiter. Moreover, they address only the London regulatory regime, whereas the Claimants (including at least one Test Claimant) work not only in the Capital but also in many locations around the UK. And in any event, as will be explained, it is not at all necessary for us to address the licensing argument in order to determine the worker status issue.[138]Bolt’s case on the Agency Model requires the Tribunal to accept the following propositions.(a) Bolt did not operate a transportation business.(b) Drivers appointed Bolt as their agent for the purposes of booking work, managing payments and associated services.(c) The drivers did not work ‘for’ Bolt; rather, Bolt worked ‘for’ them, providing services to them as their client or customer.(d) The drivers entered into individual contracts with passengers to carry them for reward. Each one forms an essential element of its case. Together, they are presented as an interlocking framework of reasoning which is said to sustain Bolt’s position on worker status.[139]Each of the four propositions is, to our minds, unreal. In our judgment the simple, practical reality is as follows.(a) Bolt did operate a transportation business.(b) The drivers did not knowingly appoint Bolt as their agent and the great majority would have had no idea of the significance of the concept of agency; it was the work of Bolt (and in particular its lawyers) to craft bogus documentation designed to set up and protect the agency-based construct.(c) The drivers did work ‘for’ Bolt: they were the skilled labour which the company required in order to run its transportation business and earn its profits.(d) The drivers did not enter into individual contracts with passengers: Bolt’s documentation to that effect was a device which did not correspond with reality. These points overlap but it may help to attempt to address them individually.[140]As to (a), we fear that Bolt’s stance parts company with reason and common sense. Bolt is self-evidently a company in the transportation business (whatever the status of those who drive in its name). It promotes and advertises its transportation services. It is Bolt, not individual drivers, that markets a range of ‘trip categories’. That is Bolt’s range. Most drivers qualify only for the standard, ‘Bolt’ category, but those who own (or more probably hire) a vehicle which qualifies for more than one category are not in a position to offer a range of trip categories to potential passengers. The offer to the public comes from Bolt. Investing in a larger or more prestigious car will not enable a Bolt driver to expand what Bolt calls his business: at best it may, perhaps, improve his earnings as a Bolt driver (although he will have to set against any improved income the costs of that investment). Bolt drivers cannot market their services at all. They are not permitted to approach or engage with putative Bolt passengers. It is only through the Bolt App that they can receive work carrying Bolt passengers. And work comes to them only if Bolt (in its absolute discretion) elects to accept the putative passenger’s request and they find themselves sufficiently close to the pick-up point to receive an offer in accordance with Bolt’s allocation criteria. Nor can the Bolt driver, however good a service he provides, hope to develop a following from among Bolt’s clientele. Sharing contact information is prohibited as is making contact with a passenger following a trip. Self-evidently, these restrictions are devised in order to protect Bolt’s commercial interest in retaining the customers of its transportation business. Likewise, the many measures which Bolt employs to ensure that its drivers provide a quality service are designed to preserve the reputation of its transportation enterprise and enhance its prospects of securing repeat business.[141]As to (b), it is, in our view, unreal to treat individual drivers as principals and Bolt as their agent. We have two main reasons. First, there was nothing approximating to a genuine agreement between the parties to create an agency relationship. The drivers did not set out to appoint Bolt as their agent. Most (like the witnesses before us) would not have read Bolt’s terms at all, even in outline, and of those who did precious few would have been able to grasp, even in a very general way, the legal relationship for which Bolt’s draughtsmanship was intended to provide. In any event, there was no genuine meeting of minds about agency. Bolt did not, and does not, discuss its standard terms with drivers. It notifies drivers of its terms and drivers understand that they have no choice but to accept or cease to work with Bolt. There was nothing remotely resembling an arms’ length agreement between business entities. Second, in countless ways the dealings between Bolt and the drivers were incompatible with Bolt’s case that the drivers were the principals and Bolt the agent. Our points under (a) above and (c) and (d) below all argue strongly against the agency analysis, which conjures up a fantastical vision of thousands of very small dogs, each being wagged by an enormous tail. Overwhelmingly, the power lies with Bolt. There is nothing in the relationship which demands, or even suggests, agency. The agency notion is posited simply to defeat the obvious interpretation which the facts invite, that Bolt employs the drivers to provide their labour in furtherance of its transportation business.[142]As to (c), our points under (a) and (b) above and (d) below are repeated. The judgment of the majority of the Court of Appeal in Uber included a summary of 13 particular points on which the ET had placed reliance at first instance. The Court discounted two as insubstantial or worse and we exclude them accordingly. The remaining 11, with original numbering retained, were as follows (findings in italics with the Court of Appeal’s comments in ordinary font):(1) The contradiction in the Rider Terms between the fact that ULL purports to be the driver's agent and its assertion of "sole and absolute discretion" to accept or decline bookings. Ms Rose [Leading counsel for Uber] criticised this on the grounds that it was necessary because under the regime of the PHVA 1998 only ULL can accept or decline bookings. In our view, the fact that this is a statutory requirement does not invalidate its significance: if anything it reinforces it.(2) The fact that Uber interviews and recruits drivers. We agree with the ET that this is significant.(3) The fact that Uber controls the key information (in particular the passenger's surname, contact details and intended destination) and excludes the driver from it. Ms Rose argued that these were important and desirable measures in the interests of passenger safety. We agree that they are: but, as with the statutory requirement that only ULL may accept or decline bookings, this does not detract from the significance of what is stated.(4) The fact that Uber requires drivers to accept trips and/or not to cancel trips, and enforces the requirement by logging off drivers who breach those requirements. We agree that this is significant as showing a high degree of control.(5) The fact that Uber sets the (default) route and the driver departs from it at his peril. This is not as stringent an element of control as some others because the driver may depart from the route prescribed by the App and the peril is only financial: nevertheless, it does have some significance.(6) The fact that UBV fixes the fare and the driver cannot agree a higher sum with the passenger. (The supposed freedom to agree a lower fare is obviously nugatory). Ms Rose submits that this also is a regulatory requirement; again, in our view, that fact does not detract from its significance in supporting the ET's conclusion that Uber runs a transportation business and the drivers provide the skilled labour through which its services are provided.(7) The fact that Uber imposes numerous conditions on drivers (such as the limited choice of acceptable vehicles) instructs drivers on how to do their work, and in numerous ways, controls them in the performance of their duties. Ms Rose submitted that these conditions are standard in the taxi and minicab industry. No doubt they are, but again they support the ET's findings that the drivers are working for Uber, not the other way around.(8) The fact that Uber subjects drivers through the rating system to what amounts to a performance management/disciplinary procedure. This is a powerful point supporting the case that the drivers work for Uber.(9) The fact that Uber determines issues about rebates, sometimes without even involving the driver whose remuneration is liable to be affected. This is another similar point, though somewhat less powerful than the last one.(10) …(11) …(12) The fact that Uber handles complaints by passengers, including complaints about the driver. This is another regulatory requirement, but again it supports the Claimants' case and the ET's conclusion.(13) The fact that Uber reserves the power to amend the driver's terms unilaterally. We agree that this supports the ET's conclusion. In the Supreme Court, Lord Leggatt agreed (para 97) that the Tribunal’s findings, even if sub-paragraphs (10) and (11) were discounted, amply justified the finding that Uber drivers worked ‘for’ Uber.[143]We entirely agree with Mr Reade that this case is ‘not Uber’, but it is instructive to consider how much the operations of Uber and Bolt had in common. Viewing the Bolt enterprise as it functioned at the outset, in June 2019, we can see very little to separate its key features from those of the Uber system.[144]In his judgment in Uber, Lord Leggatt picked out (at paras 94-101) five aspects of the Tribunal’s findings which he judged to be worth emphasising:(1) Uber’s right to set the fare and determine its own ‘service fee’;(2) Uber’s right to set the drivers’ contractual terms and their obligation to accept them;(3) Uber’s constraints upon the drivers’ choice whether or not to accept trips, in particular (a) suppression of information about the passenger’s destination until pick-up and (b) the application of penal measures based on acceptance and cancellation rates;(4) Uber’s imposition upon drivers of quality standards (eg in relation to (a) choice of car; (b) choice of routes; and (c) use of ratings as a performance management tool);(5) Uber’s control of information, in particular (a) passengers not being offered a choice between available drivers; (b) communication between driver and passenger being prohibited following trips ; (c) management of payment and any complaints being controlled by Uber in such a way as to preclude any contact between driver and passenger (management of the payment arrangements being graphically illustrated by Uber’s contrivance of the ‘invoice’ ostensibly addressed by the driver to the passenger, which is never sent).31[145]Lord Leggatt continued (para 101): Taking these factors together, it can be seen that the transportation service performed by drivers and offered to passengers through the Uber app is very tightly defined and controlled by Uber. Furthermore, it is designed and organised in such a way as to provide a standardised service to passengers in which drivers are perceived as substantially interchangeable and from which Uber, rather than individual drivers, obtains the benefit of customer loyalty and goodwill. From the drivers’ point of view, the same factors - in particular, the inability to offer a distinctive service or to set their own prices and Uber’s control over all aspects of their interaction with passengers - mean that they have little or no ability to improve their economic position through professional or entrepreneurial skill. In practice the only way in which they can increase their earnings is by working longer hours while constantly meeting Uber’s measures of performance.[146]In our judgment, the five special considerations identified by Lord Leggatt (which overlap materially with the 11 (of the original 13) factors listed by the Tribunal and approved by the Court of Appeal) applied in almost equal measure to the Bolt regime in force from 2019.[147]Our overall assessment under (c) is that it is clear that, in 2019, drivers operating under the Bolt banner were working ‘for’ Bolt.[148]We turn to (d). We find that the supposed contract between the Bolt driver and the passenger is a fiction designed by Bolt (and in particular its lawyers) to defeat the argument that it has an employer/worker relationship with the driver. The device is necessary since, without it, the inevitable inference is that such a relationship exists and explains the dealings between the parties. And the fiction becomes all the more obvious when one considers the inconsistency between the driver and passenger terms from which we have already quoted. The former (Initial Terms, cl 6.2), told the driver that he had entered into an agreement to carry the passenger effective at the moment when he accepted the trip. But the latter (cl 2.3) told the rare passenger interested enough to read the terms to prepare to enter into a separate agreement with the driver on pick-up ‘on such terms and conditions as the passenger agrees with the driver’. The confusion and consternation which would have been generated by any attempt by a passenger to engage the driver in a conversation of the sort which the clause envisaged can only be imagined. The reality did not correspond with Bolt’s documents. As Bolt, the driver and the passenger all knew, the terms on which the passenger was to be carried had all been sewn up well before the passenger stepped into the driver’s car. The problem for Bolt is that the law requires the Tribunal to focus in a ‘worldly wise’ manner on the reality before it. It must hold itself free to disregard written terms which do not correspond with that reality (Autoclenz). Indeed, it must start with the reality and not with the written terms (Uber). In a conventional contract for carriage each party will know, at the very least, who the other party is; what the essence of the carrier’s 31 And which, if it were sent, the passenger would be less than pleased to receive, having already paid the fare directly to Bolt. obligation amounts to (typically, carriage of one or more persons from identifiable point A to identifiable point B); and what fare the party carried is to pay. Here, at the point when, according to Bolt’s driver terms, the contract was concluded, the driver was not privy to any of these three crucial pieces of information. Throughout, Bolt has prevented drivers from knowing the name (ie forename and surname) of the passenger. Throughout, Bolt has prevented the driver from ascertaining the precise destination until after he has accepted the trip (no details were released to drivers from June 2019 to January 2020 and since then only the first four characters of the postcode). And throughout, Bolt has prevented drivers from learning what fares passengers pay for journeys in which they are involved. Of course, a contract for carriage may be entered into by an agent on behalf of a carrier. But in that case one would expect the agent to be acting for the benefit of the carrier. One would not expect the agent to impose terms precluding its principal from having access to key information to do with the contract which he is obliged to perform. At this point in the analysis, the supposed driver/passenger contract already has a very peculiar look about it which provokes the question whether there is a more straightforward and rational explanation for the true relationships and obligations between Bolt, drivers and passengers. In our judgment Bolt’s position becomes all the more obviously untenable when one reflects on the numerous further ways in which Bolt controls drivers (see especially under (c) above).[149]Our reasoning above is arrived at notwithstanding the heroic submissions of Mr Reade on behalf of Bolt. We will deal with his main arguments. It would not be practicable to deal with every point advanced but we have given most careful consideration to everything put forward in resistance to the Sample Claimants’ submissions.[150]In the first place, we have not made the mistake of treating analysis of the Agency Model as a simple repeat of Uber. As Mr Reade says, this is not Uber. There are material differences between Uber’s system and Bolt’s. We have applied our minds strictly to the evidence put before us in these proceedings.[151]Second, we have already acknowledged that the so-called ‘gig economy’ has moved on since the Uber litigation. The first-instance hearing was in the summer of 2016 and evidence was heard about events and practices dating back to 2014 or thereabouts. The app-based private hire and courier market has become much more congested and competitive since those days and although the legal implications of drivers using more than one app was debated at various levels in Uber, we have already noted that the practice of multi-apping must be taken to have increased as the years have passed. Accordingly, we agree with Mr Reade that the Tribunal’s broad and multi-faceted analysis of all the relevant circumstances must take account of the change in the landscape since Uber. But we do not accept that the point carries the weight which Mr Reade attaches to it. The environment is material, but far from decisive. We agree with Mr Glyn that our principal focus is necessarily on the driver and his relationship with Bolt when at work. The point is made in numerous authorities starting with McMeechan v Secretary of State for Employment [1997] ICR 549 CA, which are collected by Lord Leggatt in Uber, at para 92.[152]Third, building on his second point, Mr Reade also relied on the changed landscape since Uber and in particular the prevalence of multi-apping, the removal of fetters on the freedom to reject trips and the high incidence of drivers rejecting trips, as support for the closely related submission that current conditions are inconsistent with the proposition, central to the analysis in Uber, that drivers operate under a ‘dependent‘ work relationship with Bolt. The argument appeared to be that, viewed in the new context, the drivers should not be regarded as falling into a class of people seen by Parliament as needing the protection of worker status. We reject this submission. It calls to mind the story (apocryphal perhaps) of a conversation between two American workers. One sings the praises of the President, saying, ‘He created six million jobs!’, which draws the other’s jaundiced reply, ‘I know, I’ve got four of them!’ The fact that many drivers multi-app and/or reject a high proportion of trips offered to them does not, in our judgment, justify the conclusion that they are not to be seen as dependent or vulnerable in light of the statutory purpose behind Parliament’s decision to recognise ‘workers’ as meriting a degree of protection not enjoyed by the independent contractor. In a fragmented labour market many people earn a small living from multiple sources. If a Bolt driver also earns money through other PHO apps, or combines driving with stacking supermarket shelves, there is no obvious reason to consider him any more independent or secure that the single-app driver. Taken in the round, the rewards from other apps are likely to be similar to what Bolt pays. Likewise, the fact that a Bolt driver whose account is ‘blocked’ might be able to find driving opportunities through another app does not warrant the conclusion that he is not in a dependent relationship vis-à-vis Bolt. The environment does not alter the reality that Bolt drivers, poorly-paid individuals often facing financial difficulty, often without native English or high levels of education, carry out their work wholly and exclusively on Bolt’s terms. They can only work on those terms or cease working for Bolt. They are subject to control in numerous forms. They are in no position to build a customer base through Bolt or otherwise develop their careers as drivers. They can increase their earnings only by working longer hours and/or responding to Bolt incentives and campaigns.[153]Fourth, Mr Reade also made much of the changes in driver terms introduced by Bolt between June 2019 and July 2022. We accept that certain controls were relaxed to an extent during that period and that these changes make Bolt’s case marginally less difficult in relation to the period after May 2021, when acceptance rate ‘blocks’ were discontinued. But even here the Respondents’ case is greatly overstated. The overwhelming weight of relevant ‘control’ factors in play after May 2021 and to date favours the Claimants. Moreover, we have not been persuaded that Bolt communicated its change of stance on acceptance rates to the drivers at the time, which had the inevitable practical consequence of a substantial degree of control in relation to acceptance levels being retained after May 2021. As to cancellation and driver rating ‘blocks’, we have found that these were ended in June 2022 but again, owing to a failure to communicate the change (which, in the absence of explanation, we cannot see as accidental), driver behaviour was influenced (one might say manipulated) well beyond that date. In our view, even if the changes relied on by Mr Reade had been properly and fairly implemented with clear notice to the drivers, they would not have come close to making a material difference to our analysis on the status issue. And any small forensic advantage derivable from the changes in terms is gravely diminished owing to the manner in which they were implemented. In all the circumstances, we are satisfied that the Bolt driver of July 2022 (ie at the very end of the Agency Model period) fell full square within the category of persons whom the ‘limb (b)’ definition is intended to capture.[154]Fifth, Mr Reade relied heavily and repeatedly on the remarks of Underhill LJ in Windle from which we have quoted above. Again, we accept that he identifies a factor to be placed in the balance but find that he greatly overstates its value. The ratio of the Court of Appeal’s decision was simply that the EAT had erred in law in holding that the absence of any ‘umbrella contract’ governing relations between interpreters and the Ministry of Justice between assignments was not relevant to the question of their legal status when working. That circumstance is indeed a factor which, depending on the circumstances, may bear on a proper assessment of their status during assignments. But it is as well to bear in mind two points. First, the passage from the judgment of Underhill LJ on which Mr Reade places such heavy reliance begins with the remark that the ‘ultimate question’ must be the nature of the relationship during the period of working. Second, Windle was explicitly approved in Uber (at para 92) but there is nothing in Lord Leggatt’s judgment to suggest that he saw the absence of an ‘umbrella contract’ as a factor of any great significance to the issues before him.[155]Sixth, we note Mr Reade’s submissions concerning the Transopco case, but are not at all persuaded by them. There, the EAT simply found that the firstinstance decision that the claimant had not been employed under a limb (b) contract involved no error of law. It did not hold that the ET had arrived at the only conclusion open to it. In any event, the facts on which the decision rested were, as we have noted, quite different from those with which we are concerned. There, the claimant was an established, skilled black cab driver, working for most of his time in that (avowedly independent contractor) capacity. And his work with the respondent represented a small supplement to his income (about 15%). This is nothing like the profile of the average Bolt driver. Among countless other points of distinction, the respondent did not prescribe routes, did not determine the fare and did not control or restrict communication between driver and passenger. Nor did it resort to fictions such as the bogus ‘invoice’ from the driver to the passenger. The ET could and did permissibly hold that the written terms between the parties reflected the true agreement between them. We cannot say the same here. For the reasons we have given, we are clear that the terms which Bolt operated places its drivers in a position of vulnerability and subordination and that it is not in keeping with reality to characterise Bolt as their customer.[156]We have reminded ourselves that our first and last guide must be the statutory language itself (Bates van Winkelhof, para 39) and so we have returned to it. For the reasons we have given, we are satisfied to a high standard that, in respect of the entirety of the period of the Agency Model, the Claimants were ‘workers’ working for Bolt in furtherance of its transportation business. That the drivers undertook work personally was not in issue: they were not permitted to transfer the App or to pass any task to any substitute. The freedom to reject or cancel trips, while certainly a factor, does not militate against the obligation of personal service in respect of trips accepted and undertaken. Nor does it defeat the requirement for mutuality of obligation (see PGMOL). The drivers worked ‘for’ Bolt: the reverse analysis is untenable. And it would offend language and common sense to attribute to Bolt in its dealings with drivers the status of a client or customer of a business or profession carried on by the drivers. It is very hard to characterise the driver as a business. He was simply a man with a car looking to earn a living from it. It is even harder to characterise Bolt as the driver’s customer. There was no arm’s length transaction. There was never any question of the two parties negotiating or striking a bargain. The business was Bolt’s and the transaction between it and the driver was its purchase of the driver’s labour, strictly on its ‘take it or leave it’ terms.[157]Although our focus is on the statute, we have had regard to the valuable guidance of the superior courts. We bear in mind that in Uber the conclusion of the Supreme Court was that, on the facts found, the only reasonable outcome was a finding that the drivers were ‘workers’ (para 119). Accepting as we do that this case is ‘not Uber’, we are satisfied that on the facts which we have found here there is little of real substance to separate Bolt’s Agency Model from the Uber (pre- 2017) model and that such differences as there are do not begin to justify the dramatic difference in the applicable legal analysis for which Mr Reade contends. We also take comfort from the fact that our conclusion appears to fit with the guidance contained in other authorities which we have cited above. We find that an approach based on Cotswold Developments would confirm our conclusion: the Bolt driver was not in a position to market his services as an independent person to the world in general, but was recruited by Bolt to work (for so long as he was working) as an integral part of its transportation business. An analysis modelled on James v Redcats would lead to the same result: the proper view is that the Bolt driver found himself in a subordinate and dependent work relationship with Bolt.[158]Finally, we have measured our conclusions against the very recent guidance of the Supreme Court in PGMOL. It is clear that the ‘irreducible minimum’ of mutuality of obligation and ‘a sufficient framework of control’ are amply demonstrated.32 Reminding ourselves that mutuality of obligation and control are not the sole determinants to be applied when considering employment status, we have stepped back to survey the entirety of the relevant evidence and our primary and secondary findings thereon. These, again, strongly confirm the analysis already offered based on the statutory language itself and the earlier authorities. The Principal Model[159]It was not, and could not be, in issue that under the Principal Model as launched on 1 August 2022, drivers worked subject to terms which obliged them to provide their services personally. The App did not become transferable on that date (or on any later date). And the Principal Model Terms did not provide for any right of substitution. In the circumstances, we can only interpret Mr Reade’s arguments as implicitly accepting that, in respect of the period between 1 August 2022 and the introduction of Bolt Link in October 2022, the drivers must succeed on the issue of personal service.[160]Mr Reade submitted that everything changed with Bolt Link. We will not 32 PGMOL was, as noted, a case in which contracts of employment properly so called were in issue. Where the question is of worker status, the bar is set lower. repeat our findings above save to record that Bolt Link did not create any right of substitution. That simple fact was not, ultimately, in issue before us. When we pressed Mr Reade to explain how Bolt Link helped his case on personal service, he appeared to say that the question of freedom to provide a substitute was something of a red herring: the key point was that the scheme made good Bolt’s argument about the absence of any obligation of personal service. A Bolt fleet owner (‘Driver A’) was never under an obligation to accept any trip offered to him. He was free simply to run his fleet and derive an income from drivers ‘associated with’ his Bolt account (his Drivers B, C, D etc). But this prompted the Tribunal’s obvious next questions: since Driver A (like any other Bolt account holder) had never been under any obligation to accept any particular trip offered to him, why was Bolt Link seen as necessary at all and what difference did its introduction make to the legal analysis? We got no satisfactory answer to these questions. Asked whether there was any decided case to support his submissions on the legal effect of Bolt Link, Mr Reade acknowledged that he could not cite any ‘direct’ authority.[161]In our judgment Mr Reade was faced with a near-impossible task in seeking to make anything of value out of Bolt Link. We cannot disagree with the Claimants’ contention that it was conceived by Bolt and its lawyers for the sole, or at least predominant, purpose of protecting Bolt against claims by drivers seeking to establish worker status. For our purposes in determining the issue of status, it has no legal significance whatsoever. We are in no doubt that, under the Principal Model, drivers are engaged under contracts to provide their services personally to Bolt.[162]That leaves the question whether, under the Principal Model, Bolt’s status under the contract between it and the driver is that of a client or customer of any profession or business undertaking carried on by the driver. Here we adopt and repeat our findings above in relation to the Agency Model. In our view, those findings are applicable no less to the Principal Model than to the Agency Model. If anything, the change on 1 August 2022 serves to accentuate the imbalance in the relationships between the company and the drivers and to make the argument that Bolt was the client or customer of the drivers all the more unreal.[163]Mr Reade submitted that the ‘Custom Pricing’ arrangements brought in between August 2022 February 2023 (see above) favoured Bolt’s characterisation of itself as the client or customer of the driver. We disagree. A person running a professional practice or business is able to dictate the terms on which his or her goods or services are offered to the market. The drivers have no such freedom. Their remuneration was and is strictly on Bolt’s terms. Under ‘Custom Pricing’ a driver is at liberty to take a ‘punt’ on a higher rate per mile than the ‘Dynamic’ rate.33 Doing so may have a minor, potentially beneficial effect in pushing up the ‘Dynamic’ rate, but only if other like-minded drivers in the vicinity do likewise. On the other hand, if other drivers do not follow suit it may have the negative consequence of pricing him out of contention for any offer of work which may follow, leaving him in a worse position financially than he would have been in had he not taken his ‘punt’. We regard ‘Custom Pricing’ as another Bolt device aimed 33 As we have noted, the rate per mile is one of numerous elements which Bolt combines in order to determine the driver’s fee. primarily at boosting its case on worker status. The innovation is, in our view, without legal significance. If that is wrong, it amounts at best to a factor offering minimal support to Bolt’s case that it was a client or customer of the drivers. The case the other way is overwhelming. If they are ‘workers’, when are Bolt drivers working?
Introduction
[164]It is not in dispute that, if they are properly classified as ‘workers’, drivers are ‘working’ from the moment when they accept any Bolt trip to its conclusion or cancellation.[165]The main disagreement between the parties concerns the time spent by drivers who are online and within the territory in which they are licensed to work. Such drivers might, for example, be stationary and taking a refreshment break, or circling in a busy area in the hope of picking up another trip, or driving towards home on the lookout for the last trip of the day. In his dissenting judgment in the Court of Appeal in Uber, Underhill LJ called such time ‘availability time’, and we will gratefully adopt his expression. The Claimants’ case is that all drivers(a) with the App switched on, and(b) in the territory in which they are licensed to operate, and(c) not ‘multi-apping’ are ‘working’ when each of these three conditions is satisfied. Bolt’s position is that drivers are ‘working’ only in the circumstances referred to in our last paragraph.[166]A similar controversy arose in Uber. The ET took the view that a driver was ‘working’ under a limb (b) contract when he had the App switched on, was in the territory in which he was licensed to use the App and was ready and willing to accept trips. Leading counsel for the Respondents had argued that, even if there was a limb (b) contract in existence, the driver was not ‘working’ under it unless and until he was performing the function for which the contract existed, namely the carriage of passengers. The ET rejected that submission, stating (para 100): … it confuses the service which the passenger desires with the work which Uber requires of its drivers in order to deliver that service. It is essential to Uber’s business to maintain a pool of drivers who can be called upon as and when a demand for driving services arises. … To be confident of satisfying demand, it must, at any one time, have some of its drivers carrying passengers and some waiting for the opportunity to do so. Being available is an essential part of the service which the driver renders to Uber …[167]This analysis was, as we understand it, upheld by the EAT, the majority of the Court of Appeal and the Supreme Court. In our view, it is no less applicable to Bolt than to Uber.[168]In the Supreme Court, Lord Leggatt observed (para 122) that, before ‘working time’ is addressed, attention must be given to identifying the periods during which the individual concerned ‘is a “worker” employed under a “worker’s contract”’. The arguments before us did not always appear to acknowledge the need to separate the two stages of analysis. The period of the contract[169]Addressing the first question, Lord Leggatt held (paras 123-130) that the ET had permissibly found (by implication) that a contract came into existence when a driver, within his territory, switched on the App. (The third condition identified by the ET, that the driver should be ready and willing to accept trips, was material to ‘working time’ but not to the period of the contract.) Lord Leggatt found the ET’s analysis permissible having regard to its findings of fact, which he judged sufficient to establish the ‘irreducible minimum of obligation’ required for a worker contract. Those findings related in particular to control measures (such as forcibly logging drivers off for short periods), which were held by the ET to operate ‘coercively’ as a penalty if drivers failed to maintain a prescribed rate of trip acceptances.[170]By reference to our findings of fact in relation to the Sample Claimants, we conclude that the requisite minimum of obligation is not established in the cases of Mr Ali (on and after his arrival as a Bolt driver in December 2021) and Mr Nayir on his re-appointment in May 2022.34 In the cases of each, Bolt communicated the critical fact that no minimum acceptance rate applied. No obligation to accept any offer of a trip arose.[171]We see the cases of the other six Sample Claimants differently. In our view, it was for Bolt to convey to its drivers in clear terms any decision to relax its acceptance rate terms. It has failed to establish on evidence that it did so. It is not entitled to look to the Tribunal to credit it with acts which it has failed to prove. Nor has it established in relation to any Sample Claimant a date by when the Tribunal can safely assume that the abolition of minimum acceptance rates must have been conveyed to him. Absent any means of identifying on evidence a cut-off date, we are driven to hold that the irreducible minimum of obligation is shown throughout and to date.[172]If we are wrong in holding on the evidence that the ‘irreducible minimum of obligation’ is established in the case of the six Sample Claimants throughout, we would hold in the alternative that it is shown at the very least up to May 2021, the last month in which, as we have found, Bolt applied acceptance rate ‘blocks’.[173]And if we are wrong altogether in holding that a worker contract could, in any circumstances, come into existence before a driver accepted an offer of a trip, we would hold that the period of any contract began at the moment of acceptance. Working time – generally[174]As we have noted, the Claimants’ position is that drivers are working for Bolt during availability time and so, in principle and subject to the proper application of the relevant legislation, entitled to be paid, save where they are multi-apping. Mr Reade derides this argument as absurd. We do not see that it is. In our view, it simply follows the Uber line of reasoning (by which we are bound) but pragmatically acknowledges the evident difficulty, in the case of a driver who is multi-apping, of showing that he is working for Bolt (or any other PHO whose app 34 There is no difficulty about mutuality of obligation in relation to Mr Nayir’s period of working with Bolt prior to May 2021. he holds). We do not see how the legal analysis can be affected by Mr Reade’s subsidiary points about the practical difficulties for Bolt arising from the fact that it cannot know whether, during any period of availability time for which a claim is made, the driver was, or was not, multi-apping. No doubt, the issue in any particular case could ultimately be decided by the ET on evidence, although common sense would surely propel well-represented parties towards a negotiated resolution.[175]For the avoidance of doubt, given the way in which the Sample Claimants put their case, it does not fall to us to decide any point of principle concerning the availability time of a multi-apping Bolt driver. Our Judgment above neither makes nor implies any finding on the subject. We simply place on record that no claim is raised on behalf of the Sample Claimants in respect of such time. WTR 1998[176]By WTR 1998, reg 2(1), ‘working time’ is defined as ‘any period during which [a worker] is working, at his employer’s disposal and carrying out his activity or duties.’[177]In Uber Lord Leggatt had no difficulty in upholding the ET’s finding that availability time was ‘working time’ under reg 2(1) (para 134), subject only to the qualification (paras 135-6) that the case of the multi-apping driver35 would fall to be addressed as a matter of fact and degree, on the basis of evidence.[178]Mr Reade’s challenge in this part of the case appears to consist of an overnice reading of selected words in Lord Leggatt’s judgment (already addressed in our last footnote) and/or a frontal assault on the central reasoning in Uber, seemingly on the basis that it should no longer be regarded as applicable in the ‘gig economy’ context. Both lines are impermissible. As to the former, there is no need to say anything further. As to the latter, there is, in our judgment, nothing to differentiate our case from Uber on the subject of working time.[179]For the reasons given, the availability time in respect of which the Sample Claimants’ claim (ie excluding any such time spent multi-apping) qualifies as working time under WTR 1998, reg 2(1). NMWA 1998 and NMWR 2015[180]Under the national minimum wage legislation work is divided into four categories for the purposes of classification. It is common ground that the first of these, Salaried Work, is inapplicable.[181]Mr Reade submitted that the case fell within the definition of Time Work under NMWR 2015, reg 30(b). This applies to work ‘in respect of which a worker is entitled under their [sic] contract to be paid … by reference to a measure of output 35 The jargon was not used but Lord Leggatt’s meaning in para 135 is not open to doubt. Moreover, contrary to Mr Reade’s suggestion, he did not say or imply (in paras 136-137 or anywhere else) that the Uber driver who could have multi-apped but did not elect to do so was also outside, or arguably outside, the definition of working time for so long as that possibility subsisted. in a period of time where the worker is required to work for the whole of that period.’ In our view, this definition is clearly inapplicable. The Bolt driver is not paid by reference to any ‘measure of output’ and is not required to work for any specified period of time, much less the whole of any such period. As the commentary in Harvey on Industrial Relations and Employment Law, Div B explains [210], reg 30(b) captures the case of a worker working set hours who is paid by reference to a measure of output.[182]Notwithstanding Mr Reade’s submissions, the third category, Output Work, is also inapplicable. This applies to the worker who has no set hours and is paid simply by reference to ‘a measure of output, including a number of pieces made or processed, or a number of tasks performed’ (see NMWR 2015, reg 36 and Harvey, op cit, Div B, [211]).[183]The fourth category, Unmeasured Work, is a residual category which captures all classes of work not falling into the first three. Accordingly, we find that, for the purposes of the national minimum wage legislation, the Bolt driver performs Unmeasured Work.[184]The question of classification may be relevant for the purposes of the deeming provisions under the NMWR 2015. By reg 47 hours spent by a worker ‘travelling for the purposes unmeasured work’ are to be treated as hours of unmeasured work. But we have found that the Bolt driver travelling to pick up a passenger, carrying a passenger or travelling during availability time is ‘working’. So no deeming in respect of such time is required. No doubt there might be some instances where the deeming provisions are engaged (such as, perhaps, where a trip has taken a driver outside his licensed area, or even his chosen radius) and he is returning to the area (or radius) in order to seek his next trip. But this sort of small detail is surely for consideration at the remedies stage, particularly as we have not heard argument directed to it.[185]Likewise, although some reference was made to time spent training, it seems to us that this too is better considered at the remedies stage, on the strength of clear legal argument (if there is anything about which the parties disagree). OUTCOME AND FURTHER CONDUCT[186]For the reasons given, the Sample Claimants succeed to the extent stated in our Judgment and these Reasons.[187]Our decision applies strictly to the Sample Claimants only but, subject to any appeal, we hope that the parties will approach disposal of the stayed claims of all Claimants in a pragmatic and realistic way.[188]On 4 and 7 October 2024 the Tribunal received correspondence from the parties concerning further conduct of the case. We think that the best course is to allow both sides time to reflect on our adjudication and its implications for the wider litigation before conferring and, we hope, reaching as much common ground as possible concerning the way forward. We would envisage a preliminary hearing for case management being held (before the judge alone) very early in the New Year, and would welcome prompt notification from the parties of preferred dates.