CHRISTOPHER JOHNSON v TRANSOPCO UK LTD

The respondent in the employment tribunal, Mytaxi Network Limited, operated the Mytaxi App (the “App”). The tribunal found that from 2014 the claimant worked full time in business on his own account as a black-cab (or Hackney Carriage) driver in London. In February 2017 he downloaded the driver version of the App. Apart from the odd trip in April he did not start to actively use it until around the end of July 2017. He made no trips using it after 18 April 2018 when he was removed from it. Throughout that period he continued to source work away from the App as a self-employed black-cab driver.

In August 2018 the claimant presented a claim to the employment tribunal. As of January 2020 his live complaints were of protected-disclosure detriment, for working-time holiday pay, of unlawful deduction from wages and of failure to pay the national minimum wage. To be entitled to pursue all of these complaints he had to be, in law, a worker of the respondent. Section 230(3) Employment Rights Act 1996 (“the 1996 Act”) defines a worker as follows.

“(3)

In this Act “worker” (except in the phrases “shop worker” and “betting worker”) means an individual who has entered into or works under (or, where the employment has ceased, worked under)—

a contract of employment, or

The definitions of “worker” for the purposes of the National Minimum Wage Act 1998 and the Working Time Regulations 1998 are materially the same. The claimant originally claimed also to have been an employee of the respondent pursuant to section 230(3)(a) of the 1996 Act; but, by January 2020 his case was solely that, at the relevant times, he was a worker of the respondent pursuant to section 230(3)(b) – a so-called “limb (b) worker”.

There was a preliminary hearing in the tribunal to determine the claimant’s status, before EJ Balogun. In a reserved decision promulgated in July 2020 the tribunal determined as follows:

The claimant appeals from the tribunal’s decision in respect of point (b). There is no appeal or cross-appeal in respect of points (a) or (c). On 1 July 2020 Transopco UK Limited acquired the assets and liabilities of Mytaxi Limited, including in respect of the claimant’s claim. Accordingly, by consent it had been substituted as the respondent to this appeal. In this decision, by “the respondent” we mean Mytaxi Limited, or Transopco UK Limited, as appropriate to the context.

In the employment tribunal Mr Milsom of counsel appeared for the claimant and Mr Linstead of counsel for the respondent. At the hearing of this appeal Mr Milsom appeared again. Mr Short QC now led Mr Linstead.

The tribunal’s decision should be referred to for its full content. Its findings of fact included, in summary, the following.

All London black-cab drivers must hold a Hackney Carriage licence and are, as such, regulated by Transport for London (“TfL”). Such a driver may only ply for hire within their licenced zone. The claimant had a Green Badge licence covering central London and nine suburban sectors. Under TfL regulations it is an offence for a driver to charge more than the metered fare for a trip within their licenced zone. A TfL-licenced driver may not refuse a fare where the destination is within 12 miles, or one hour, or 20 miles from Heathrow Airport.

In 2017 to register on the App the driver was required to input personal information, upload their Hackney Carriage licence and bank details, and agree to the respondent’s General Terms of Use and Framework Agreement. There was no interview. There was a voluntary induction and an opportunity to take “The Knowledge Plus”. However, the clamant did not take up either of these.

Passengers using the App can pre-book a journey or request a cab immediately. The latter is called an Instant Job. Two of the claimant’s fares were pre-booked. The rest were Instant Jobs.

Once logged on to the App the claimant had to click that he was “Busy” or “Go Free”. He might be “Busy” if he had, for example, been hailed on the street. The default setting was “Busy”.

When sent a passenger name and location the claimant had four seconds to accept or reject. Inaction would result in automatic rejection. Acceptance would lead to the claimant getting the passenger’s number, and the passenger getting his name, number and photo. They could then communicate directly with each other.

The fare was calculated via the taximeter applying the TfL rates. However, the passenger terms provided for a minimum £10 fare during core hours. Drivers picking up jobs from the App were not paid for the time spent travelling to collect the customer, and this was designed as an incentive to accept such “e-hails” over street hails. However, the tribunal noted that minimum fares did not appear in the drivers’ terms, and the claimant refused to charge them as he believed that to do so would breach his TfL licence. He was never compelled to do so. The respondent told him that it was up to him. However, where the actual fare was less than the minimum the claimant sometimes asked the respondent for a top up. That did not happen very often, as the average trip fare in London was £17. When the respondent did pay a top up it described it as a goodwill gesture.

Passengers could pay via the App if they had registered a card, or using the driver’s card terminal, or in cash to the driver. Around 50% of payments are via the App, and that money is received by the respondent. The respondent pays the driver twice-weekly the sums collected from the App less a 10% procurement fee in respect of both App payments and payments made in cash to the driver or using the driver’s credit card terminal.

Either the driver or the passenger can “scrub” the booking without reason. If the passenger does so more than 2 ½ minutes after the driver has accepted the job, or does not show up after 2 ½ minutes’ waiting time, the respondent pays the driver a £5 scrub payment. There is a £10 payment for pre-booked jobs cancelled less than 30 minutes before pick-up time. After five or more driver scrubs the driver is warned that their account may be suspended. On 13 November 2017 the respondent contacted the claimant because he had scrubbed 10 jobs the previous week. On 29 November he was suspended from the App for two weeks because he had scrubbed 23 jobs since 13 November. He was admonished for mostly not following the requirement to tell the passenger. Between 1 July 2017 and 30 April 2018 the claimant’s acceptance rate was 24.8% against a London average of 23.9%. His driver-scrub rate was 35.4%, against a London average of 7.9%.

Between April 2017 and April 2018 the claimant completed 282 trips via the App at a total value of £4560.48 after commission. In the same period he earned £30,472.45 through other sources. He generally worked a five-day week, eight hours a day. He therefore did less than 1½ App trips a day. The tribunal observed:

“The claimant contended that he signed up to the App as it was the only way to maintain a sustainable income. However, that does not chime with his level of trips overall and his high percentage of cancellations.”

The respondent has General Terms and Conditions of Use for Taxi Drivers (“Driver T & Cs”) and a Procurement Framework Agreement for Taxi Drivers (“Framework Agreement”). By logging on to the App the driver accepts the Driver T & Cs. There is no opportunity for negotiation. These require the driver to comply with any applicable local transportation laws and tariffs, and only to accept orders in so far as compatible with such regulations. The tribunal continued at [37]:

“Other clauses of note are:

i.

Clause 1.1 – this describes the service provided by the respondent to the driver as procuring potential passengers looking for a taxi via the mytaxi passenger app.

ii.

Clause 2.1 states that the user shall have no claim to the continuous and uninterrupted operability of the Driver App. In other words, there is no obligation upon the respondent to provide the driver with work through the App.

iii.

Clause 2.2 allows the respondent to discontinue the services offered via the app temporarily in whole or part without notice.

iv.

Clause 3.4 provides that drivers are obliged to accept open transport orders assigned to them. It is common ground that transport orders are assigned if they have been accepted by the driver.

v.

Clause 7 allows the respondent to block a driver’s use of the App if, inter alia, he does not carry out the assigned transport order without any particular reason and without furnishing proof thereof. This needs to be read in conjunction with the driver’s right to cancel orders. [239 & 240, 243]

vi.

Clause 9 provides that driver agrees to being rated by the passenger and for the respondent to publish the ratings (non anonymised) on its website.”

The Framework Agreement described the respondent as providing the service of procurement of potential passengers via the App for drivers registered with Mytaxi, and offering the possibility of collecting payment via the App. In such cases the driver assigns the fare to the respondent, and the respondent assumes the credit risk. Clause 2.6 asserted that a legally binding contract is entered into between the driver and the passenger when the driver presses the “Accept” button. Settlement is solely between driver and passenger. “For the avoidance of doubt, no contract will entered into [sic] between the passenger and Mytaxi regarding transportation services.” [40]

There was provision for a procurement fee. While there was reference to a limited right of substitution for Advance Reservations, the respondent conceded that in practice there was no right of substitution for jobs done via the App and drivers were expected to carry out the service personally.

There was provision that these two sets of terms and the respondent’s Privacy Policy together constituted the whole agreement between the parties.

The Passenger Terms and Conditions provided that the passenger contracts direct with the driver for transportation services.

The tribunal cited some extracts from the respondent’s marketing material, including references to the respondent’s fleet of 17,500 drivers, described as “Mytaxi drivers” and to the training provided through the Mytaxi academy to the community of drivers. The claimant relied on this as evidence that he was integrated into the respondent’s business, but the tribunal said at [49]:

“I do not read it in that way. It is not unusual for businesses to use exaggerated language and, dare I say, hyperbole in their advertising. I see this as no more than the respondent seeking to draw attention to and sell its product.”

Passengers can rate drivers. The claimant contended that this was to monitor performance, the respondent that it was a quality monitoring tool. The tribunal said: “To me that is a distinction without a difference.” The purpose was to incentivise drivers to maintain high standards. The claimant’s ratings were good, so it was never an issue. More detailed enquiries were made of him towards the end of his period using the App, but that was in the context of him posting critical comments about the respondent on social media. The tribunal observed: “In my view, that exercise was a one-off and not indicative of a general policy of performance monitoring.” [50]

The respondent had branded materials, but there was no requirement for drivers to use these.

Under the heading “Risk” the tribunal found that the respondent offers various discounts to incentivise passengers to use the App. When that happens the driver receives the full fare. When the passenger pays via the App the respondent bears the fraud risk. When passengers use the driver’s credit card terminal, the driver bears the fraud risk. Scrubs are a shared risk.

The tribunal concluded that, at the point when the driver accepted the job, his obligation was owed to the respondent, not the passenger. This remained the case at the point of pick up and the driver did not from that point also have a contract with the passenger. The respondent allowed passengers to rate drivers and dealt with complaints. This did not fit with the model of it being merely the claimant’s agent. It also bore certain risks. Taking matters in the round, the judge concluded that the passengers contracted with the respondent for the transportation services, which were then delivered by the claimant pursuant to a separate contract with the respondent. This part of the decision is not challenged before us.

The tribunal then turned to its conclusions on whether the claimant provided transportation services to the respondent in its capacity as client or customer of the claimant’s business undertaking. We should set them out in full.

“66.

In Bates van Winkelhof v Clyde & Co LLP [2014] UKSC 25, 2 types of self-employed people were identified. The first kind carries out a profession or a business undertaking on their own account and enters into contracts with clients or customers to provide work or services for them. The second kind provides their services as part of a profession or business undertaking carried on by someone else. Only the second is a limb (b) worker.

Regulatory regime

67.

The drivers that used the App held individual licences to drive black cabs within their designated areas. The claimant was licenced by Tfl to provide a black cab service long before his relationship with the respondent commenced and it was he who was subject to Tfl’s regulatory regime and who could be sanctioned or prosecuted for not complying with it. The regulatory regime in this case is a key factor distinguishing it from Uber as in that case, Uber was the holder of the licence and the party subject to regulation by Tfl, not the drivers. That makes a difference when it comes to the issue of control as in the present case, much of the control the drivers were subject to was dictated by Tfl rather than the respondent.

68.

In his submissions, Mr Milsom identifies a number of factors which he contends are relevant to the issue of status. I address some of these below.

Claimant’s mode of appointment

69.

This is dealt with at paragraph 17 of my findings and in my view is neutral. If anything, the minimal appointment process suggests a distance between the driver and the respondent less consistent with an integrated working relationship.

The contractual terms

70.

The claimant submits that the driver terms and conditions and framework agreement are at odds with the reality of the contractual arrangements. In Autoclenz Ltd v Belcher [2011] ICR 1157 (SC), it was held that the relative bargaining power of the parties had to be taken into account in deciding whether the terms of a written agreement represent the true agreement. The claimant points to the fact that the claimant had to sign up to the respondent’s non-negotiable terms as an indicator of the claimant’s weaker bargaining position. However, another feature of relative bargaining positions is how dependent a party is on the contract for work. Having signed up to the App in February 2017, the claimant only actively started using it in July 17’. From then until he was removed from the App, the claimant chose not to accept 75% of the trips offered and cancelled 35.4% of the jobs accepted. On average, he carried out 1.5 trips a day via the App. In money terms, this represented less than 15% of his overall income derived from driving his taxi. [paras 29 & 30]. It is clear from these statistics that the respondent was not the claimant’s main source of income and that he did not need to sign up to the App in order to work. This was not a dependent work relationship as envisaged in Cotswold Development Construction Ltd v Williams [2006] IRLR 181 at para 53.

Control

71.

It is common ground that the claimant was not a worker when the app was switched off. Even when the app was switched on and set to “Go Free” indicating that the claimant was available to accept jobs, he was under no obligation to accept any jobs and was still free to market his services to potential street hail passengers and pick them up in preference to jobs on the App. For the majority of the time, that is exactly what he did.

72.

In this case, as in Uber, there was no umbrella contract creating rights and obligations between periods of work as drivers are not obliged to switch on the App. In Windle and anor v Secretary of State for Justice [2016] ICR 721 CA, Lord Justice Underhill, held that although the ultimate issue for the purposes of s.83(2) of the Equality Act 2010 (EqA) is the nature of the relationship during the period when the work is being done, 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. In that case, the Court of Appeal restored the tribunal’s decision that the individuals were not employees. Although that case was to do with the extended definition of employment under section 83(2) Equality Act, the reasoning is as applicable to section 230(3)(b) ERA. In Windle the individuals were interpreters who provided services to the Courts and Tribunals Service (HMCTS) on a case-by-case basis. They also worked for other institutions. HMCTS was under no obligation to offer them work, and they were under no obligation to accept it when offered. There are clearly parallels to be drawn with the present case and as in Windle, I consider the absence of mutuality between the various transportation orders suggests a degree of independence that points away from worker status.

73.

Although clause 3.4 of the written contract provides that drivers are obliged to accept open transport orders assigned to them, that is the necessary quid pro quo for being allowed to use the App rather than a feature of control. Once a job was accepted, it was still open to the driver to cancel within certain parameters. This was not entirely without consequence as drivers could be suspended from the App for excess cancellations, as happened in the claimant’s case. However, I do not consider that sanction to be a significant measure of control when weighed against everything else. The respondent had a vested interest in protecting the reputation of its brand. Cancellations were detrimental to the overall customer experience and discouraging these speaks more to commercial expediency rather than control.

74.

The respondent exercised little if any control on how the claimant undertook the transportation services. The way in which the work as done was controlled by Tfl on the one part and the claimant on the other. That the claimant controlled the way he provided the service is not remarkable given that drivers were recruited for their skill and expertise as black cab drivers and would not be expected to be subject to close direction. Nevertheless, this is a point distinguishing this case from Uber, where Uber set the default route, fixed the fares and imposed other conditions, such as the choice of vehicle.

75.

Unlike Uber, the respondent did not exclude its drivers from receiving key passenger information. Not only were drivers sent the name, contact details and destination (if known) of the prospective passenger, they could also contact the passenger after the trip, something which was strongly discouraged by Uber. Whereas Uber fixed the fare charged, the respondent did not. The fare was set by the taximeter in accordance with Tfl regulations. The claimant relies on the minimum £10 fare as evidence of the respondent setting the fare. However as found above, drivers were under no obligation to charge it, and the claimant chose not to as he considered such charge to be in breach of Tfl rules. That approach was consistent with the written contract which provided that transport orders may be accepted only insofar as they are compatible with local regulations and rules the driver is subject to. The onus was on the driver to make that assessment, not the respondent.

76.

The claimant relies on the fact that the respondent asks passengers to rate drivers as a form of control. My findings on this are at paragraph 49 above. Performance monitoring is not necessarily the same as performance management. There were no consequences associated with poor ratings and it was a matter for individual drivers as to the steps they took to address them.

77.

The fact that the claimant could provide his services as infrequently or as often as he wanted, could dictate the timing of those services and was not subject to control by the respondent in the way in which those services were undertaken indicates a level of independence that is more consistent with an independent contractor in business on his own account than a self-employed person fully integrated into the respondent’s business.

78.

Taking the above reasons in the round, I find that the claimant and respondent contracted with each other as two independent businesses and that the respondent was a customer of the claimant’s taxi business. That the respondent was more powerful (in scale and financial terms) than the claimant is, when weighed against all the other factors, not material.

79.

I find that the claimant was not a limb b worker.”

Ground 1 of the appeal asserts that the tribunal erred in its conclusion that the respondent was the client or customer of the claimant’s business. This breaks down into a number of substantive sub-grounds, which may be summarised as follows:

Ground 2 raises, in the alternative, the contention that, even if it might have been open to the tribunal to reach the conclusion that it did, its given reasons for doing so were, in respect of one or more of the aspects raised by ground 1, not Meek compliant.

It is convenient at this point to have an initial look at the authorities. For our purposes, a useful route in is found in the judgment of Underhill LJ (Jackson and Lindblom LJJ concurring), in Secretary of State for Justice v Windle [2016] ICR 721. That case concerned claims brought under the Equality Act 2010, but, as the following passage explains, the extended definition of “employee” in section 83 of the 2010 Act has been interpreted as being subject, impliedly, to the same “client or customer” exception as expressly appears in the definition of “worker” with which we are concerned. The authorities on the two concepts therefore provide a single body of guidance.

At [8] to [14] Underhill LJ summarised the legal background as follows:

“8.

Section 83 (2) (a) identifies three kinds of contract. The first – "a contract of employment" – means a contract of service. The Claimants accept that they were not employed under such a contract. It is their case that they were employed under the third kind of contract listed, namely "a contract personally to do work". The best explanation of what that phrase refers to appears in Bates van Winkelhof v Clyde & Co. [2014] UKSC 32, [2014] 1 WLR 2047. In that case the Supreme Court was concerned with whether the claimant was a "worker" within the meaning of section 230 (3) of the Employment Rights Act 1996, but Lady Hale, who delivered the majority judgment, reviewed the field more widely. Limb (b) of section 230 (3) refers to employment under

"… any other contract … 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".

Lady Hale pointed out, at para. 25 (p. 2055 B-C), that that formulation distinguished between two 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] 1 WLR 1872 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 someone else. The general medical practitioner in Hospital Medical Group Ltd v Westwood [2012] EWCA Civ 1005; [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."

She then, at paras. 31-32, went on to observe that the same distinction was recognised for the purpose of discrimination law, even though section 83 (2) (a) of the 2010 Act does not contain anything equivalent to the elaborate words of exception in the second half of section 230 (3) (b). She said:

"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.

32.

In Allonby v Accrington and Rossendale College (Case C-256/01) [2004] ICR 1328: [2004] ECR–I873  the European Court of Justice was concerned with whether a college lecturer who was ostensibly self-employed could nevertheless be a 'worker' for the purpose of an equal pay claim. The Court held at para. 67, following Lawrie-Blum v Land Baden-Wurttemberg (Case C-66/85) [1987] ICR 483; [1986] ECR 2121: that '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'. However, such people were to be distinguished from 'independent providers of services who are not in a relationship of subordination with the person who receives the services' (para 68). The concept of subordination was there introduced in order to distinguish the intermediate category from people who were dealing with clients or customers on their own account. It was used for the same purpose in the discrimination case of Jivraj v Hashwani. [2011] 1 WLR 1872 "

9.

As Lady Hale there acknowledged, the qualification on the apparently broad scope of the phrase "a contract personally to do work" had in fact already been recognised in the decision of the Supreme Court in Hashwani v Jivraj [2011] UKSC 40, [2011] 1 WLR 1872, although the discussion is less explicit. In that case the issue was whether an arbitrator was an employee for the purpose of the Employment Equality (Religion or Belief) Regulations 2003, which had an identical definition. Lord Clarke, with whose judgment the other members of the Supreme Court agreed, emphasised that it was not enough that the putative employee should be a party to a contract personally to do work: he or she must be "employed under" such a contract (see para. 36, at p. 1887 B-C).

10.

It has become common to refer to persons employed under contracts falling within the terms of section 230 (3) (b) of the 1996 Act as "limb (b) workers". Because, inconveniently, the 2010 Act uses different language, it is inapt to refer to employees of the third kind listed under section 83 (2) (a) by the same label. I will refer to them as "employees in the extended sense".

11.

As to how the distinction is to be made between the two kinds of self-employment – that is, between employees in the extended sense and the "truly self-employed", as it is sometimes put – in Hashwani Lord Clarke said, at para. 34 (p. 1886 E-G):

"... The essential questions ... are ... those identified in paras 67 and 68 of Allonby [2004] ICR 1328, namely 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. Those are broad questions which depend upon the circumstances of the particular case. They depend upon a detailed consideration of the relationship between the parties ... . The answer will depend upon an analysis of the substance of the matter having regard to all the circumstances of the case."

12.

It will be seen that both Lady Hale in Bates van Winkelhof and Lord Clarke in Hashwani refer to the decision of the ECJ in Allonby v Accrington & Rossendale College (Case C-256/01) [2004] ICR 1328. This concerned an equal pay claim by part-time lecturers at a further education college, who had initially been employed by the college but had been made redundant and required to offer their services through an agency. One of the issues was whether the claimants were "workers" within the meaning of article 141 of the EU Treaty. At paras. 64-72 (pp. 1359-60) the Court said this:

"64.

The term 'worker' within the meaning of article 141(1) EC is not expressly defined in the EC Treaty. It is therefore necessary, in order to determine its meaning, to apply the generally recognised principles of interpretation, having regard to its context and to the objectives of the Treaty.

65.

According to article 2 EC, the Community is to have as its task to promote, among other things, equality between men and women. Article 141(1) EC constitutes a specific expression of the principle of equality for men and women, which forms part of the fundamental principles protected by the Community legal order: see, to that effect, Deutsche Post AG v Sievers (Cases C-270 and 271/97) [2000] ECR I-929, 952, para 57. As the court held in Defrenne v Sabena (Case 43/75) [1976] ICR 547 , 566, para 12, the principle of equal pay forms part of the foundations of the Community.

66.

Accordingly, the term "worker" used in article 141(1) EC cannot be defined by reference to the legislation of the member states but has a Community meaning. Moreover, it cannot be interpreted restrictively.

67.

For the purposes of that provision, 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: see, in relation to free movement of workers, in particular Lawrie-Blum v Land Baden-Württemberg (Case 66/85) [1987] ICR 483, 488, para 17, and Martínez Sala, para 32.

68.

Pursuant to the first paragraph of article 141(2) EC, for the purpose of that article, 'pay' means the ordinary basic or minimum wage or salary and any other consideration, whether in cash or in kind, which the worker receives directly or indirectly, in respect of his employment, from his employer. It is clear from that definition that the authors of the Treaty did not intend that the term 'worker', within the meaning of article 141(1) EC, should include independent providers of services who are not in a relationship of subordination with the person who receives the services (see also, in the context of free movement of workers, Meeusen v Hoofddirectie van de Informatie Beheer Groep (Case C-337/97) [1999] ECR I-3289, 3311, para 15).

69.

The question whether such a relationship exists must be answered in each particular case having regard to all the factors and circumstances by which the relationship between the parties is characterised.

70.

Provided that a person is a worker within the meaning of article 141(1) EC, the nature of his legal relationship with the other party to the employment relationship is of no consequence in regard to the application of that article: see, in the context of free movement of workers, Bettray v Staatssecretaris van Justitie (Case 344/87) [1989] ECR 1621, 1645, para 16, and Raulin v Minister van Onderwijs en Wetenschappen (Case C-357/89) [1992] ECR I-1027, 1059, para 10."

13.

Both Lord Clarke in Hashwani and the ECJ in Allonby refer to a "relationship of subordination". In Bates van Winkelhof Lady Hale warned against treating the presence or absence of "subordination" as the infallible touchstone for distinguishing between the two kinds of self-employed worker under section 230 (3): see para. 39 (pp. 2058-9). That term was, however, used by the ET in this case (loyally applying Hashwani) and neither party criticises it for doing so. I will occasionally use it myself, though bearing in mind Lady Hale's caveat.

14.

One other part of the legal background which it is necessary to refer to is the concept of mutuality of obligation. The position is most lucidly stated by Elias LJ in Quashie v Stringfellows Restaurant Ltd [2012] EWCA Civ 1735, [2013] IRLR 99, at paras. 10-12 (pp. 102-3), as follows:

"10.

An issue that arises in this case is the significance of mutuality of obligation in the employment contract. Every bilateral contract requires mutual obligations; they constitute the consideration from each party necessary to create the contract. Typically an employment contract will be for a fixed or indefinite duration, and one of the obligations will be to keep the relationship in place until it is lawfully severed, usually by termination on notice. But there are some circumstances where a worker works intermittently for the employer, perhaps as and when work is available. There is in principle no reason why the worker should not be employed under a contract of employment for each separate engagement, even if of short duration, as a number of authorities have confirmed: see the decisions of the Court of Appeal in Meechan v Secretary of State for Employment [1997] IRLR 353 and Cornwall County Council v Prater [2006] IRLR 362.

11.

Where the employee working on discrete separate engagements needs to establish a particular period of continuous employment in order to be entitled to certain rights, it will usually be necessary to show that the contract of employment continues between engagements. (Exceptionally the employee can establish continuity even during periods when no contract of employment is in place by relying on certain statutory rules found in section 212 of the Employment Rights Act.)

12.

In order for the contract to remain in force, it is necessary to show that there is at least what has been termed "an irreducible minimum of obligation", either express or implied, which continue during the breaks in work engagements: see the judgment of Stephenson LJ in Nethermere (St Neots) v Gardiner [1984] ICR 612, 623, approved by Lord Irvine of Lairg in Carmichael v National Power plc [1999] ICR 1226, 1230. Where this occurs, these contracts are often referred to as "global" or "umbrella" contracts because they are overarching contracts punctuated by periods of work. However, whilst the fact that there is no umbrella contract does not preclude the worker being employed under a contract of employment when actually carrying out an engagement, the fact that a worker only works casually and intermittently for an employer may, depending on the facts, justify an inference that when he or she does work it is to provide services as an independent contractor rather than as an employee [emphasis supplied]. This was the way in which the employment tribunal analysed the employment status of casual wine waiters in O'Kelly v Trusthouse Forte plc [1983] ICR 728, and the Court of Appeal held that it was a cogent analysis, consistent with the evidence, which the Employment Appeal Tribunal had been wrong to reverse." … ”

Windle concerned professional interpreters who worked for HMCTS, case by case. HMCTS had no obligation to offer them assignments, nor they to accept them. The tribunal held that they were not employees, in the Equality Act 2010 extended sense. It found that “the absence of mutuality of obligation between assignments points away from employee status under section 83(2) for the times when these claimants were engaged on assignments.” The EAT overturned the tribunal’s decision, but the Court of Appeal restored it. At [22] to [24] Underhill LJ said:

“22.

The principal submission of Mr Humphreys in seeking to uphold the decision of the EAT was that in determining whether a claimant is an employee in the extended sense the essential question is to what extent he or she is acting "under direction", or is in a "subordinate" position, while at work. As he put it in his skeleton argument:

"This will require an enquiry, founded on the contract, into the scope of that direction and the extent of any limitation on the putative employee's independence in that context. The absence of mutuality of obligation between engagement can add nothing to that enquiry … . "

23.

I do not accept that submission. 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.

24.That would be my view even without any reference to Quashie. But I do not in fact think that what Elias LJ said in the passage which I have italicised can properly be disregarded on the basis that the issue in that case was whether the claimant was employed under a contract of service. The underlying point is the same. The factors relevant in assessing whether a claimant is employed under a contract of service are not essentially different from those relevant in assessing whether he or she is an employee in the extended sense, though (if I may borrow the language of my own judgment in Byrne Bros (Formwork) Ltd v Baird [2002] ICR 667: see para. 17 (5), at p. 678H), in considering the latter question the boundary is pushed further in the putative employee's favour – or, to put it another way, the passmark is lower. I would add for completeness that I do not think that Judge Clark's point that continuity of employment is not an issue in Equality Act cases (see para. 19 above) affects the analysis. The question is whether the claimant is an employee at all; and it was that which was the issue in Quashie.”

A number of authorities have grappled with the more general question of how a tribunal might go about testing and determining, in the given case, whether the “client or customer” exception applies. On this issue, we can start with Byrne Bros (cited in the foregoing passage in Windle). In Byrne Bros, in the course of making a series of numbered points about this exception within [17], the EAT (Mr Recorder Underhill QC, as he then was, and members) said the following:

“(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-à-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.

(6)

What we are concerned with is the rights and obligations of the parties under the contract - not, as such, with what happened in practice. But what happened in practice may shed light on the contractual position: see Carmichael (above), esp. per Lord Hoffmann at pp 1234-5.”

In Cotswold Developments Construction Limited v Williams [2006] IRLR 181 the EAT (Langstaff J and members) said the following at [53]:

“53.

It is clear that the statute recognises that there will be workers who are not employees, but who do undertake to do work personally for another in circumstances in which that "other" is neither a client nor customer of theirs – and thus that the definition of who is a "client" or "customer" cannot depend upon the fact that the contract is being made with someone who provides personal services but not as an employee. The distinction is not that between employee and independent contractor. The paradigm case falling within the proviso to 2(b) is that of a person working within one of the established professions: solicitor and client, barrister and client, accountant, architect etc. The paradigm case of a customer and someone working in a business undertaking of his own will perhaps be that of the customer of a shop and the shop owner, or of the customer of a tradesman such as a domestic plumber, cabinet maker or portrait painter who commercially markets services as such. Thus viewed, it seems plain that a focus upon 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. It is not necessary for this decision to examine more closely the individual cases which may fall much closer to the dividing line, and the principles upon which those cases should be determined, because in the present case the Tribunal determined that Cotswold was not in the position of a client or customer of any profession or business undertaking carried on by the Claimant reason of "the nature of the Claimant's relationship with the Respondent" (paragraph 7.3). They did not elaborate further. However, it seems to us that they were entitled to draw that conclusion, in particular because no finding of fact suggests that the Claimant operated as an independent tradesman, and much of it is suggestive if not determinative of the fact that Cotswold recruited him to work for it.”

In James v Redcats (Brands) Limited [2007] ICR 1006 the EAT (Elias P), after citing the guidance in Byrne Bros and in Williams, observed of the latter’s reference to marketing to the world in general (at [50]): “I would agree that this will often assist in providing the answer, but the difficult cases are where, as in this case, the putative worker does not in fact market his services at all, nor act for any other customer even although Mrs James is not barred by her contract from so doing. In some cases the business is effectively created by the contract.”

In Autoclenz Limited v Belcher [2011] ICR 1157 (SC) Lord Clarke JSC (for the Court), citing the judgment of Aikens LJ in the Court of Appeal, said, of cases in the work field:

“34.

The critical difference between this type of case and the ordinary commercial dispute is identified by Aikens LJ in para 92 as follows:

‘92. I respectfully agree with the view, emphasised by both Smith and Sedley LJJ, that the circumstances in which contracts relating to work or services are concluded are often very different from those in which commercial contracts between parties of equal bargaining power are agreed. I accept that, frequently, organisations which are offering work or requiring services to be provided by individuals are in a position to dictate the written terms which the other party has to accept. In practice, in this area of the law, it may be more common for a court or tribunal to have to investigate allegations that the written contract does not represent the actual terms agreed and the court or tribunal must be realistic and worldly wise when it does so. ...’

35.

So the relative bargaining power of the parties must be taken into account in deciding whether the terms of any written agreement in truth represent what was agreed and the true agreement will often have to be gleaned from all the circumstances of the case, of which the written agreement is only a part. This may be described as a purposive approach to the problem. If so, I am content with that description.”

In The Hospital Medical Group Limited v Westwood [2013] ICR 415 Maurice Kay LJ (Longmore and Toulson LJJ concurring) observed at [18]:

“The striking thing about the judgments in Cotswold and Redcats is that neither propounds a test of universal application. Langstaff J's "integration" test was considered by him to be demonstrative "in most cases" and Elias J said that the "dominant purpose" test "may help" tribunals "in some cases" (paragraph 68). In my judgment, both were wise to eschew a more prescriptive approach which would gloss the words of the statute.”

He added at [20] that there was “no single key with which to unlock the words of the statute in every case”. In Bates van Winkelhof v Clyde & Co LLP [2014] ICR 730, which concerned a partner in a firm of solicitors, Lady Hale said at [39], after reviewing these authorities:

“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. As Elias J recognised in Redcats, a small business may be genuinely an independent business but be completely dependent upon and subordinate to the demands of a key customer (the position of those small factories making goods exclusively for the ‘St Michael’ brand in the past comes to mind). Equally, as Maurice Kay LJ recognised in Westwood, one may be a professional person with a high degree of autonomy as to how the work is performed and more than one string to one's bow, and still be so closely integrated into the other party's operation as to fall within the definition. As the case of the controlling shareholder in a company who is also employed as chief executive shows, one can effectively be one's own boss and still be a ‘worker’. 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 Uber BV v Aslam [2021] ICR 657 Lord Leggatt (the other Justices concurring) said, at [71], that the general purpose of the legislation in question is “to protect vulnerable workers” from being paid too little, required to work excessive hours or subjected to other forms of unfair treatment, such as for whistleblowing. He cited with approval the description of the purpose in Byrne Bros, referring to those who are in a “subordinate and dependent position.” Further on he observed, at [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.”

This ground contends that the tribunal placed “impermissible focus” on what the claimant was doing when he was not working for the respondent. It concerns, in particular, [72] of the tribunal’s decision, wherein it considered Windle, and stated that there “are clearly parallels to be drawn with the present case and … I consider the absence of mutuality between the various transportation orders suggests a degree of independence that points away from worker status.”

Mr Milsom cited Redcats at [93] where Elias J said that lack of mutual obligations when not working is “of little, if any, significance when determining the status of the individual when work is performed.” Absence of mutuality between assignments does not preclude worker status or even employment status during an assignment. In Mingeley v Pennock [2004] ICR 727 a private hire taxi driver was held not to be an employee (in the extended sense) of a cab firm that provided jobs via a radio and then a computer link; but in Uber Lord Leggatt said that the fact that he was free to work when he chose was not a sufficient reason for holding that, when he was working, he was not employed under a contract to do work for the firm. If that conclusion was justified, it would have to be on the basis that he was not working for the firm when transporting passengers.

Further, the reach of Windle was, submitted Mr Milsom, “strongly curtailed” by Underhill LJ himself in Pimlico Plumbers v Smith [2017] ICR 657 in which he said, at [145]:

“Second, although the argument before the ET and before us was couched in terms of whether Mr Smith was subject to a legal obligation to work (or be available) for a minimum number of hours, it should not be assumed that if there had been no such obligation the evidence about what hours he worked in practice would have been irrelevant. It is necessary to distinguish two separate circumstances in which the issue of whether a putative employee/worker is engaged on a casual basis might arise. The first is where the substantive claim directly depends on their enjoying employee/worker status in respect of their periods of work (e.g. because the claim concerns their pay or some discriminatory treatment in the workplace). In such a case the question whether the engagement is casual is indeed relevant, but only on the basis that it may shed light on the nature of the relationship while the work in question is being done (see Quashie v Stringfellow Restaurants Ltd [2002] EWCA Civ 1735, [2013] IRLR 99, at paras. 10-13, and Windle at paras. 22-25). But it is not only legal obligations that may shed light of that kind. If the position were that in practice the putative employee/worker was regularly offered and regularly accepted work from the same employer, so that he or she worked pretty well continuously, that might weigh in favour of a conclusion that while working he or she had (at least) worker status, even if the contract clearly (and genuinely) provided that there was no legal obligation either way in between the periods of work. The second situation is where the claim directly depends on the claimant's status during periods of non-work, either because he or she has to establish continuity of employment or because the claim itself relates to their treatment during that period: in such a case mutuality of legal obligations is essential.”

Further, he argued, in Pimlico Plumbers in the Supreme Court [2018] ICR 1511 Lord Wilson (for the Court) asked, at [36]: “Is it necessary, or even relevant, to ask whether Mr Smith’s contract with Pimlico cast obligations on him during the periods between his work on its assignments?” and had observed cautiously, at [41], that Windle “must await appraisal on another occasion.” Multi-apping does not preclude worker status. The cycle courier in Addison Lee Limited v Gascoigne [2018] ICR 1826 and the private hire drivers in Addison Lee Limited v Lange [2019] ICR 637, both cases involving the use of an App, were in both cases workers.

In relation to the EU-derived working time complaint (which now forms part of retained EU law), Mr Milsom referred to Fenoll v Cenre d’Aide par le Travail [2016] IRLR 67 in which it was said that “any person who pursues real, genuine, activities, to the exclusion of activities on such a small scale as to be regarded as purely marginal and ancillary, must be regarded as a ‘worker’”.

He submitted that the tribunal had erred by failing to have regard to these authorities. In Fenoll terms, the proportion of the claimant’s work accounted for by jobs from the respondent showed this work was more than “marginal”; the finding of 1.5 trips for the respondent per day was sufficient to meet the “consistent” and “regular” benchmark suggested by Underhill LJ in Pimlico. The absence of mutuality could not assist without considering the “fundamental differences” between the work for the respondent and the claimant’s “truly self-employed black cab work”.

Mr Milsom submitted that, if the tribunal’s approach was correct, access to workplace rights risked becoming a “numbers game”, turning on the number of journeys performed, and fraught with uncertainty, with some drivers who work for the respondent not being workers, and others being workers, on the basis of a case by case assessment. This approach also risked individuals who multi-app, as is common in the private hire industry, losing out on workplace protections altogether. But this was rightly not seen as a bar to worker status in the Addison Lee, Uber, and other cases.

Mr Short submitted that the tribunal had properly relied upon Windle, which remains the leading authority on this point. In Uber at [91] Lord Leggatt agreed with Elias P in Redcats that lack of worker status in the gaps does not preclude worker status during a period of work:

“… subject only to the qualification that, where an individual only works intermittently or on a casual basis for another person, that may, depending on the facts, tend to indicate a degree of independence, or lack of subordination, in the relationship while at work which is incompatible with worker status: see Windle … para 23.”

The present tribunal was entitled to rely upon its findings about the proportion of the claimant’s time and income, and his rate of decline of jobs offered, and cancellation of jobs accepted, when considering whether his work for the respondent formed a part of his own business, and as pointing towards its conclusion at [70] that this was not a dependent work relationship. It had plainly considered the position when the claimant was doing jobs for the respondent.

The tribunal’s approach was also consistent with EU law, as stated in Allonby. Fenoll did not signify a different approach. Lord Leggatt in Uber at [72] regarded these, and other recent authorities of the CJEU, as all taking the same approach. In B v Yodel Delivery Network [2020] IRLR 550, the CJEU gave an Art 99 reasoned order (so it must have considered the position to be clear from existing case law) including stating at [28] that what its case law, including Fenoll, required was “an overall assessment of all the circumstances of the case brought before it, having regard both to the nature of the activities concerned and the relationship of the parties concerned.”

Our conclusion is that the statement of the law on this point in Windle remains authoritative and, of course, binding on us, and that none of the other authorities points to a different approach.

In Pimlico Plumbers in the Court of Appeal Underhill LJ did not qualify what he said in Windle. He restated it and cited Windle as authority for it. Rather, he made an additional point which is not in conflict. This is that, while a lack of legal obligation between assignments may be regarded as a pointer away from worker status during assignments, if, in fact, assignments are offered and accepted so frequently that the complainant works for the putative employer “pretty well continuously” that might be treated as pointing the other way. In the Supreme Court at [36] Lord Wilson posed a question. He did not give the answer. At [45] he described Windle as the “leading authority” on the point, and said that it was a submission challenging it that “must await appraisal on another occasion.” Lord Leggatt also, at [23], specifically referred to, and effectively endorsed, what Underhill LJ said in Windle.

We also agree with Mr Short that the CJEU authorities do not signify that a different approach is required. The passage in Fenoll relied upon by Mr Milsom needs to be understood within the context of the particular issue in that case, which was whether the nature of the activities undertaken by a disabled person when he visited a work rehabilitation centre, which offered various kinds of support, enabled him to claim worker status. The answer, at [27], was affirmative, if they were “real, genuine activities, to the exclusion of activities on such a small scale as to be regarded as purely marginal and ancillary”. As is clear from this, and later paragraphs, the CJEU was concerned there with whether, in this particular context, the nature of the activities and tasks carried out might mean the individual could not be treated as a worker. It was not saying that anyone who provides services on anything more than such a small scale as to be marginal must be a worker.

We agree with Mr Short that the position in EU law remains as stated in Allonby v Accrington and Rossendale College [2004] ICR 1328. There, in considering the concept of “worker” in EU law at [67] and [68] the CJEU stated that it was not intended to “include independent providers of services who are not in a relationship of subordination with the person who receives the services”. In Uber at [72] Lord Leggatt referred to this passage in Allonby and to later CJEU authorities, including Fenoll,as identifying the concept in the same terms.

We turn to Mr Milsom’s “numbers game” point. He submits that it cannot be right that there might be different answers to whether two drivers carrying out a job offered and accepted through the same app on the same terms are or are not workers, depending on what proportion of their overall work such jobs accounted for, by reference to numbers of jobs or remuneration.

The short answer to this is that the nature of the legal test, containing as it does an element of factual subordination or dependency, means that different outcomes in different cases involving the same app cannot be ruled out. The point was specifically addressed in Windle at [26]:

“Mr Humphreys also submitted that it was wrong in principle that a person who would otherwise satisfy the criteria to be treated as an employee during a particular engagement should fall out of protection only because there was no "umbrella contract"; and, by the same token, that where there were two people who were in substantially the same position at work but one of them was working on a casual basis and one was not it was wrong in principle that they should not enjoy the same protection. But that is an Aunt Sally. The absence of an umbrella contract is relevant only if and to the extent that it contributes to the conclusion that the claimant is not in fact in a "subordinate" relationship characteristic of an employee – in which case he or she will not be in the same position as their comparator. Whether that is so in any particular case will depend on the circumstances of that case.”

We would add the following. First, the metaphor of a numbers game is inapposite. The fact that the respective scale or size of the claimant’s activities for the respondent, and overall, could be quantified precisely numerically, in terms of jobs or money, does not mean that the significance of these should have been, or was, judged by some mathematical formula. It was for the tribunal to assess the significance and weight to be attached to these respective features, as part of its overall evaluative judgment. The numbers just helped the tribunal to get a handle on the matter. Mr Milsom submitted that the judge had effectively introduced a minimum hours threshold for worker status, but we find no sign that she did. Her conclusions at [34] and [70] speak to the evaluative question of whether the claimant was in a dependent or vulnerable relationship with the respondent.

Secondly, the additional observation made by Underhill LJ in Pimlico Plumbers that we have already discussed is consistent with the foregoing approach. If, in a given case, notwithstanding the absence of an umbrella contract, the practical reality is that a given individual works for the putative employer “pretty well continuously” then that might be found to support the conclusion that there was a relationship of subordination or dependency in that particular case. But that, too, depends upon all the facts of the given case. Further, cases will differ as to whether the claimant is, apart from what they do for the respondent, in business on their own account (though as noted in Redcats, they can, potentially, be in business, even if the respondent is their only client).

All of that said, it bears repeating that it is not in every case that what happens when the claimant is not working for the respondent will make the difference to the tribunal’s conclusion about whether they are a worker when actually doing a job for the respondent. All the authorities say is that it is not necessarily wrong to regard it as a relevant consideration. In some cases, however, features of the arrangements when they are performing a job may be decisive; and in such cases the outcome may therefore, in fact, be the same for all those who work for that operator under those same arrangements. The fact that in some other recent cases multi-apping was not a bar to worker status when working for a particular operator, does not mean that in this case this claimant must in law have been a worker of this respondent whenever he took a job through its App.

There was a further strand to Mr Milsom’s argument. This drew on a point highlighted in Westwood, being that a finding that a claimant runs his own business will not necessarily, in every case, preclude the conclusion that he was a worker, as the further question that the wording of limb (b) requires the tribunal to consider in such a case, is whether the putative employer was a client or customer of that business. So, here, said Mr Milsom, the finding that the claimant ran his own business was not enough. He submitted that the tribunal erred by not properly considering whether the nature of what he did when he worked for the respondent was materially different, such that it could not be regarded as work which was under the umbrella of that same business.

There is overlap in the way that this point was advanced, with sub-strand (b) of ground 1(5). It is therefore convenient also to pick up that sub-ground at this point. The relevant factual context is this. As the tribunal described at [13] and following, a licensed London black-cab driver may ply for hire, but only within the geographical zone of their license; and, for trips within their licensed zone, they may only charge the (TfL-regulated) metered fare. The claimant had sought clarification from the respondent as to whether using its App involved plying for hire, but was told he must take his own view. TfL told him that the answer would depend on the circumstances of each case. The tribunal declined to make a finding on the plying-for-hire point, saying, at [16] that it was not “within my remit” and: “Also, I don’t believe it will assist my determination of status in this case.”

The second strand of ground 1(5) contends that it was not open to the tribunal to decline to determine the plying-for-hire issue; and that there was also a tension between that passage of its decision, and the tribunal at [67] identifying the regulatory regime in the present case as a key point of distinction from the facts of Uber. Mr Milsom submitted that the tribunal had failed to take on board the full significance of the regulatory regime. Had it properly engaged with the point, it would have been bound to conclude that the claimant was not plying for hire when taking jobs through the respondent’s App. Having regard to that, and to the fact that regulation of fares applies only to black cabs plying for hire, the only “lawful lens” through which to view the relationship was one in which the respondent contracts with the passenger to provide a private hire service and the driver acts as its agent. That was “fundamentally at odds” with the respondent being regarded as the client or customer of the claimant’s black-cab business.

Subsequent to the hearing of this appeal, the High Court handed down its decision in R (United Trade Action Group Limited) v TfL and Transopco (UK) Limited (trading as Free Now) joined with Uber London Limited v TfL and others [2021] EWHC 3290 (Admin), 6 December 2021 (which we will call UTAG). Mr Milsom thereafter emailed in a short supplemental written submission by reference to it, and Messrs Short and Linstead a short written reply.

In UTAG the court held that (a) as contemplated by Lord Leggatt in Uber, the private hire licensing regime requires the operator to contract with the passenger as principal; and (b) such drivers who use the respondent’s app (now called Free Now) are not plying for hire. Mr Milsom submitted that the latter conclusion meant that, when accepting jobs from the respondent’s App, the claimant was not plying for hire. That fed into his Westwood point, that the claimant’s work for the respondent was of a substantively different nature to that of an ordinary Hackney Carriage driver, so that it could not properly be viewed as performed in the course of the claimant’s business.

Mr Short submitted that the present case was not one where it was said (or could have been) that there was an interrelationship between whether the claimant was a worker and whether the respondent was acting in compliance with the applicable licensing regime. As the tribunal found, the parties recognised at the time that there was some uncertainty around the plying-for-hire issue. So the respondent did not require the claimant to do anything that would be contrary to his license if he was to be treated as plying for hire; and he, for his part, took care not to do so. The tribunal found at [21] that the algorithm which offered out jobs was supposed to identify a driver’s licensed zone. Occasional glitches with this were largely resolved by July 2017. The claimant was also not required to charge a minimum fare above the metered fare; and he consistently declined to do so.

The tribunal also rightly noted at [67] that the regulatory regime was materially different in Uber. That case concerned private-hire vehicles (not Hackney Carriages) in respect of which the applicable regime means that the operator has to have a private hire license; and in that case, in fact, the operator controlled the fares. In the present case the respondent (properly) relied upon the fact that the claimant had a Hackney-Carriage license. It was common ground before the tribunal that in any event a licensed Hackney-Carriage driver can lawfully do private hire jobs (such as a pre-booking). The decision in UTAG did not show that the present tribunal was wrong to regard the plying-for-hire issue as irrelevant to what it had to decide. It had still properly concluded that the licensing regime had no bearing on the essential character of the claimant’s business activities, being conveying passengers for reward, regardless of the way in which they were obtained.

We turn to our conclusions on this further strand of the ground 1(1) argument and on this part of ground 1(5). First, we need to have a further look at Westwood. There, the claimant was a partner in a GPs’ practice. The respondent ran cosmetic surgery clinics and the claimant carried out hair restoration procedures for it. He also did work for another entirely unconnected clinic giving advice on transgender issues. The tribunal found that he was a limb (b) worker of the respondent, a decision upheld by the Court of Appeal.

At [11] and following, Maurice Kay LJ (for the court), having noted that the tribunal had found that the claimant was engaged in business on his own account, observed: “The next question is: what was that business?” He rejected the submission that he had one business performing three activities. “They were three distinct businesses or outlets for his professional skills, quite unrelated to each other.” Previous cases relied upon which went the other way each turned on their own particular facts. There was no universal principle which will condition the answer in every case. Maurice Kay LJ’s more general discussion of possible tests postulated in previous authorities, to which we have already referred, followed. It was then observed that, in that case, the particular terms on which the claimant provided his services to the respondent (including a non-competition provision) supported the conclusion that he was “clearly an integral part of its undertaking when providing services in respect of hair restoration”, even though he was also separately in business on his own account. Maurice Kay LJ concluded with his “no single key” observation.

Accordingly, we conclude that where, as here, the tribunal finds that the claimant carries on a profession or business undertaking, whether the work done for the respondent is done in the course of such profession or business, or is a distinct activity, is a matter of fact and impression for the tribunal. This is to be determined having regard to all the relevant facts and circumstances of the particular case, including the tribunal’s findings about the respective nature of the activities performed otherwise in the course of the profession or business, and performed for the respondent, and other relevant facts relating to the relationship between the provider and the respondent. There is no single key to unlock that sub-strand of the definition any more than any other part of it. The EAT can only interfere with the tribunal’s conclusion on the usual perversity grounds.

The present tribunal found as a fact (at [12]) that from 2014 the claimant worked full-time “in business on his own account as a black-cab driver in London” and, throughout the period when he used the respondent’s App, continued to source “other work as a self-employed black cab driver.” At [60] it said: “The services we are concerned with here can loosely be described as transportation services – picking up passengers and driving them to their desired destinations.” Its ultimate conclusion was that the respondent was a customer of “the claimant’s taxi business.” The tribunal was plainly of the view that the essence of the claimant’s business activity was that it involved picking up passengers and driving them to where they wanted to go, however they were obtained. It was entitled to take that view, including that whether a particular passenger was obtained by plying for hire, for regulatory purposes, by street hail or via an App, was not relevant to whether the work of conveying them for reward was done in the course of that same business.

Neither Uber nor UTAG indicates otherwise. Both cases concerned the specific licensing regime for private-hire vehicles, which was said to cast light on whether the operator contracted with the passenger as principal or as agent for the driver. In Uber, it was contended by Uber that the passenger’s contract was, only, with the driver, and that Uber acted as the driver’s agent. Lord Leggatt was unconvinced that this could be compatible with the private hire licensing regime. He opined that, as only Uber held the requisite license, it must contract with the passenger as principal. But he said that it was unnecessary to express a concluded view on the point. (See Uber at [45] to [57].) However, his putative analysis was adopted and confirmed in the first part of UTAG. In the present case the licensing regime was different, the tribunal in any event found that the respondent contracted as principal with the passenger, a finding not challenged on appeal, and this part of UTAG was not relied upon by Mr Milsom as adding anything material to the present appeal.

The second question considered in UTAG, as to whether a private hire driver using the respondent’s Free Now app was plying for hire, was important in that case, because private hire drivers cannot lawfully ply for hire at all. The court noted at [13] that it was not concerned with the use of the Free Now app to order a Hackney Carriage; and at [53] that it assumed that the facts as to how the driver operated were materially the same as in Reading Borough Council v Ali [2019] 1 WLR 2635. The court in UTAG therefore (at [54] and [55]) followed and applied the analysis of the concept of plying for hire in Ali.

As Mr Short noted, the proposition that a licensed Hackney Carriage can accept a private booking (and that the private hire licensing regime does not apply when it does) was established in Stockton-on-Tees BC v Fidler [2020] EWHC 2430 (Admin). The employment judge’s remarks in the course of [22] – that, once they had each other’s details, the passenger could have contacted the claimant directly for a pre-booking – and at [24] – that a (black-cab) driver can hire his car to a private operator who can pursue its own terms – suggest that she had this point on board.

Ali concerned a private hire Uber driver. The facts, and the reasoning in Ali, respectively assumed as materially the same, and followed, in UTAG, are set out in UTAG at [48] and [49]. We do not need to set it out. That is because, even if (which we do not have to decide) the application of the legal concept of plying for hire, as analysed in Ali and UTAG, to the facts of the present case, would point to the conclusion that the claimant was not plying for hire in respect of jobs obtained from the respondent’s App, the tribunal was still entitled to take the view that this was irrelevant to the question of the essential nature of the claimant’s business activities, and whether those jobs were carried out as part of that business. At the risk of repetition, this task required an evaluative determination by the tribunal of the overall factual picture, and the EAT can only intervene on perversity grounds. The tribunal was properly entitled to make the findings that it did about the essential nature and character of the claimant’s business activities, and hence that jobs obtained from the respondent’s App involved materially the same activities, even if, on a correct legal analysis, the latter did not involve plying for hire. This is so having regard also to the tribunal’s factual findings and conclusions on the simultaneous nature of the activities, subordination, dependency, control and integration.

Nor was the tribunal factually wrong to note at [67] that this case differed from Uber both insofar as in the former case it was the operator that held the license, and it was the operator that exerted control over such matters as fares. The material finding here was, clearly, as the tribunal explained at [74] and [75], that the present respondent did not exert control over such matters (or some other things). Even if, on analysis, the reach of TfL’s regulatory power was not as wide as the parties thought, or assumed, it was or might be, the tribunal was still entitled to regard the absence of such control being exerted or asserted by the present respondent as a relevant consideration.

Pausing there, we conclude that the tribunal did not err by placing an impermissible focus on the claimant’s activities whilst not working for the respondent. Neither domestic nor EU law precluded it from having regard to this as part of the overall picture. It did not err by taking the approach indicated in Windle, nor by failing to conclude that the claimant must be a worker of the respondent because his activities for it were “more than marginal or ancillary”. It was entitled to take into account the found facts relating to the overall scale of the claimant’s activity and of his work for the respondent, both in terms of money and jobs. It was entitled to conclude that whether the claimant was plying for hire when he used the respondent’s App was irrelevant to what it had to decide. Ground 1(1) and the sub-thread of ground 1(5) pertaining to this point therefore fail.

Ground 1(2) contends that the tribunal erred in concluding that the level of financial risk was consistent with a client/customer relationship. The ground, and Mr Milsom’s submissions, together referred to the payment by the respondent of scrub fees, airport parking expenses, provision of additional incentives to drivers via a loyalty scheme or the absorption of passenger discounts. In the majority of instances the risk of fraud or non-payment was borne by the respondent. He also referred to the minimum fee scheme. He submitted that these features of the allocation of risk, and how the claimant was remunerated, were irreconcilable with the conclusion that the respondent was a customer of the claimant’s business.

Mr Short noted that all of the features were noted and considered by the tribunal at various points in the course of its decision, save for the loyalty scheme, in respect of which the evidence was that there was none in London. (Mr Milsom did not tell us that was wrong.) Mr Short submitted that, in reality, this sub-ground amounted to the claimant contending that the tribunal should have attached more weight to facts relating to this aspect than it did. But that was a matter for the tribunal. The facts found showed that the risks, including of fraud, and scrubs, were shared. The evidence was that minimum fares were rarely applied in practice and airport parking fees were only applicable to a small proportion of jobs. In any event, some assumption of cost and risk by the respondent was not consistent only with the claimant being a worker. As the tribunal discussed, these features reflected practicalities, such as the fact that the driver was not paid for travelling to the passenger’s location, and the logistics of different payment methods, and were designed to ensure that taking e-hails remained financially attractive when compared with taking street hails.

We can deal with this sub-ground shortly. We do not think the tribunal failed to consider these aspects, or sufficiently to consider them, either separately or in the round. They are not all mentioned in the closing section but they are all mentioned in the course of the decision. The decision must be read as a whole, and it shows a firm command of the material and the rival cases. It is clear that the tribunal did not consider that these features tipped the balance in favour of worker status. The fact that some incentives and risk sharing were offered by the respondent to reflect the risks associated with using their platform, or generally to enhance its financial attractiveness as an option, does not mean that the tribunal erred by not considering this aspect as pointing to the work done for the respondent as not forming part of the claimant’s general business, or otherwise bespeaking a subordination and dependency such as must point to worker status.

This sub-ground therefore fails.

We went to ground 1(3) which contends that the tribunal erred in its approach to the question of control. Mr Milsom submitted that, as discussed in Bates, the absence of subordination did not preclude worker status in every case. Greater attention should have been given to the question of control. Citing Wright v Aegis Defence Services (BVI) Ltd, UKEAT/0173/17, the issue was not whether, or how often, the respondent in fact exercised control over the claimant, but one of the extent of its right to do so. Here, the tribunal “failed to pay any or adequate account” to: red-flag checks carried out at appointment stage; detailed instructions in the Driver Guides; the fact that the respondent set non-negotiable terms; the monitoring of performance, including scrubbing; the right to suspend and terminate the claimant’s services; the ability of passengers to rate the driver, which could lead to adverse action; and the enquiries made of 11 of the claimant’s former passengers. The tribunal had wrongly taken account of the motivation for wanting to exercise control of these aspects, rather than the fact that the respondent could do so.

Mr Short submitted that, whilst it was correct that subordination was not a necessary prerequisite of worker status, Lord Leggatt in Uber regarded subordination or dependency, where present, as the thing that required protection of worker status to be accorded, and spoke of the significance of the concept of control as being that it pertained to this issue of subordination. The tribunal did not err in its approach to the law in this regard. Once again, he submitted, the essence of this sub-ground was that the tribunal should have attached more weight to the highlighted features than it did – but this was not something that could be challenged on appeal.

In any event, as reflected in the tribunal’s findings at [18], there were no “red flag checks” at the time of the claimant’s use of the App. The other features were all considered by the tribunal in the course of its decision. None of them were such as to make it perverse for the tribunal not to conclude that they together necessarily pointed to the conclusion that the claimant was a worker. It specifically considered the feature of standard terms, and its contribution to the overall picture regarding relative bargaining power, at [70], the significance of driver scrubs for its assessment of control at [73], and the significance of passenger ratings and performance monitoring at [76].

That the tribunal understood the subtleties of the significance of control in the context of skilled work was apparent from [74]. Once again, none of these features was necessarily consistent only with worker status. The terms on which independent sub-contractors do work may often be wholly determined by the client. Any business has an interest in protecting its brand, and the adverse effects which a poor customer experience delivered by a provider may have on it.

Once again, we essentially agree with Mr Short’s submissions on this point. The tribunal did not err in principle in its approach to the law in relation to this aspect. Although Uber had yet to reach the Supreme Court, it plainly understood, and paid attention to, the import of Autoclenz (discussed at [60] and [70]) and Uber in the Court of Appeal. As Lord Leggatt’s approach was to highlight, specific attention was given to whether this was a dependent relationship, the tribunal finding at [70] that it was not, and to the various facets said to touch on the control issue at [71] to [77]. Its conclusions on each of these features are soundly and thoughtfully reasoned, including as to the significance of sanctions for excessive scrubs [73], the features discussed at [74], the more detailed enquiries made at the end of the relationship being a one-off [50] and the purposes of customer ratings at [50] and [76]. The findings of fact did not point to the only possible conclusion being that the claimant was a worker. There is no basis on which the EAT can properly interfere.

Ground 1(4) contends that the tribunal erred in its approach to the level of integration of the claimant into the respondent’s activities when working for it. This is directed to the findings under the heading “marketing”, in particular as to the statements about the respondent’s “fleet of 17,500 drivers”. It was submitted that these and other statements in that material show that the respondent’s “product” is the drivers. Mr Milsom also submitted that the tribunal had impermissibly focussed on the respondent’s motivations at [49]. What mattered was the factual image presented, and whether this was compatible with the respondent being a client or customer of the claimant. In cases such as Westwood, Pimlico Plumbers and Uber, the descriptions of integration given by the operators to customers was a pointer to worker status. This tribunal erred in concluding otherwise.

Once again Mr Short submitted that the tribunal considered this aspect and reached a permissible conclusion about it, at [49] and there was no error of law. We agree. This was marketing or advertising material, therefore self-evidently aimed at differentiating the respondent from competitors and persuading prospective passengers to use its App rather than some other method of getting a ride. Consideration of its purpose, as well as its content, was properly part of the process of considering what light this material did or did not cast on the degree of integration of the claimant into the respondent’s operations when doing work obtained from its App, alongside all of the other facts found that touched on that aspect. The tribunal was entitled to take a view that it did at [49]. It cannot be said that the only permissible conclusion, taking account of this material, was that the claimant was so integrated into the respondent’s business when doing jobs for it, that he was a worker.

We turn to ground 1(5). This contends that the tribunal erred in its treatment of Uber. We have already considered the second sub-strand. The first is that the tribunal erred by, at [10], characterising the claimant’s contentions as being that his case was “on all fours” with Uber, and by then engaging in a spot-the-difference exercise, treating Uber as setting some sort of minimum factual benchmark or pass mark that must be reached or exceeded for worker status to be achieved. In submissions Mr Milsom highlighted various passages in which the tribunal referred to factual features present in Uber but not in this case, as symptomatic of this erroneous tick-box approach.

Mr Short submitted that it was not surprising that the tribunal’s decision contained many references to Uber, not least because of the many references that had been made to it in submissions from the claimant to the tribunal. There was no basis for saying that the tribunal had erred in considering any of the factual features that were in fact different from those of the Uber case, as such. It reached proper conclusions about the significance of each of them, in this case, for the issues of subordination, dependency, control and integration. Its opening remarks showed that it had not taken a mechanistic approach, or treated the facts of Uber as a minimum threshold.

As to this, in the course of its introductory overview of the facts, the tribunal said this:

“9.

The way the App works is that rather than hailing a Black Cab in the street by sticking their hand out, people can book one via the the App downloaded onto their mobile phones. Taxihailing apps have become widely used in the UK, and indeed worldwide, over recent years. The most well-known of these is of course the Uber app, used to book private hire vehicles such as minicabs. The company operating the Uber app in the UK (“Uber”) has had its own issues over the status of drivers using its app, culminating in litigation at ET, EAT and the Court of Appeal - Uber BV v Aslam [2019] ICR 845 (“Uber”).

10.

Both parties have quoted Uber extensively, albeit from different standpoints. The claimant contends the present case is on all fours with Uber and should be decided in the same way. The respondent on the other hand contends that there are significant distinguishing features which justify a departure from Uber.

11.

As is often said and is well worth repeating; issues of worker status are a mixture of fact and law. Many of the Uber conclusions turned on the specific findings of fact in that case and should therefore not be seen as a panacea for all driver-status questions relating to app based taxi hailing operations. On that point, it was recognised in Uber that Uber could have devised a business model that did not involve them employing drivers [ para 97 ET Uber].”

Uber concerned a car-ride-hailing app. At the time of the tribunal’s hearing it had reached the Court of Appeal. It is hardly surprising that there was much reference made in submissions to its facts and reasoning. The reference to “on all fours” is plainly shorthand for the fact that the claimant, unsurprisingly, placed strong reliance on it as being (on his case) materially similar factually. The judge however correctly noted that the decision was not a “panacea” for all driver-status questions, and that every case does turn on its own facts. There is no sign in this passage, or in the decision as a whole, that the judge engaged in a mechanistic or tick-box exercise, treated the facts of Uber, or any of them, as a minimum requirement, or misunderstood the substance of the claimant’s case as to why the particular factual features of this case pointed to the conclusion that he was a worker. Every factual feature of this particular case is considered and evaluated with thought and insight, and as part of a careful process of weighing and evaluation.

Ground 1 as a whole therefore fails.

Ground 2 contends, in the alternative, that the tribunal’s reasons were, in respect of one or more of the matters raised by ground 1, or the overall question of whether the respondent was a client or customer of the claimant’s business, insufficient. Mr Milsom’s submission highlighted the point made in some authorities that, while an appellate court should not be hypercritical in its approach to a tribunal’s reasoning, nor is it acceptable for a tribunal simply to assert a conclusion without setting out some reasoning explaining how it has been reached. But, he contended, the tribunal had simply, at [78], made the bare unreasoned assertion that the respondent was the claimant’s customer. He also referred to the failure to decide the plying-for-hire point.

Mr Short submitted that the decision is Meek-compliant (Meek v City of Birmingham District Council [1987] IRLR 250). He referred to the helpful summary of the guiding principles recently given by Cavanagh J in Frame v The Governing Body of the Llangiwg Primary School, UKEAT/0320/19/AT at [47]. The present tribunal’s decision tells the claimant why he lost. There was no inadequacy in the reasoning.

This ground has really largely been answered by what we have already said. In the course of considering the various strands of ground 1 we have noted a number of aspects on which it was submitted that, if the tribunal’s decision was not based on an incorrect understanding of the law, or perverse, then it failed adequately to address the point; and we have explained why we disagree. The tribunal explained that it regarded the plying-for-hire issue as irrelevant. We have found that this involved no error. We do not accept that the tribunal’s conclusion at [78] amounted to a bare unsupported statement. It opens with the words: “[t]aking the above reasons in the round”. It is plainly a statement of the tribunal’s conclusion, drawing on all of the reasoning that has come before. The tribunal is here standing back and looking at the overall picture in the round before coming to its final conclusion, which was precisely the right thing to do.

More generally, in our shared view, this decision is a model of focussed and articulate reasoning. Concision should not be mistaken for lack of depth. The judge at every stage homes in on the salient evidence, facts and points of law. She was not obliged to analyse the submissions at great length. She deals with that at [56]. Her introduction and conclusions show that she was plainly on top of them. The concluding section covers all the key issues, propounds the salient points of law from the authorities, and takes the reader through how the judge has applied them to the facts that she has found. They convey why the judge concluded that, though the respondent was the more powerful contracting party, in scale and financial terms, the claimant was not in fact in a relationship of dependency or subordination to it; and why she concluded that jobs done for the respondent formed part of his taxi business. The decision properly conveys why the claimant lost.

The appeal is dismissed.