Uber B.V. ("UBV") & Ors v Aslam & Ors

In the words of the employment tribunal (“the ET”), from whose decision this appeal is brought, “Uber is a modern business phenomenon”. It was founded in the United States in 2009 and its smartphone app, the essential tool through which the enterprise operates (“the App”), was released the following year. At the time of the ET hearing in 2016 there were about 30,000 Uber drivers operating in the London area, and 40,000 in the UK as a whole. The organisation has some 2 million passengers registered to use its services in London.

The Claimants are current or former Uber drivers working in London.

The first Appellant, Uber BV (“UBV”), is a Dutch corporation and the parent company of the second and third Appellants. It holds the intellectual property rights in the App.

The second Appellant, Uber London Limited (“ULL”), is a UK registered company which, since May 2012, has held a Private Hire Vehicle (“PHV”) Operator Licence pursuant to the Private Hire Vehicles (London) Act 1998 and the regulations made under it. Its functions include making provision for the invitation and acceptance of PHV bookings and accepting such bookings.

The third Appellant, Uber Britannia Limited, holds or manages PHV Operator Licences issued by various local authorities outside London. It was named in the claim form in this case but its activities did not feature in the evidence in the ET nor in the argument before us.

The claims brought before the ET were under the Employment Rights Act 1996 (“ERA”), read with the National Minimum Wage Act 1998 (“NMWA”) and associated Regulations, for failure to pay the minimum wage and under the Working Time Regulations 1998 (“WTR”) for failure to provide paid leave. Two claimants, including Mr Aslam, also complained under Parts IVA and V of the ERA of detrimental treatment on “whistleblowing” grounds.

In their response form the Appellants, to whom we will refer collectively as “Uber”, denied that the Claimants were at any material time “workers” entitled to the protection of the legislation on which they relied. In addition, they raised jurisdictional defences based on applicable law and forum points.

The ET held a public preliminary hearing to determine the status and jurisdiction issues before Employment Judge Snelson and two lay members beginning on 19 July 2016. Thomas Linden QC appeared for the Claimants and David Reade QC for Uber. Oral evidence was heard from Mr Aslam and Mr Farrar, the First and Second Claimants, and, on behalf of Uber, from Ms Joanna Bertram, Uber’s Regional General Manager for the UK, Ireland and the Nordic Countries.

The ET summarised the principal issues before them at the preliminary hearing in terms taken from Mr Linden’s closing submissions:-

Uber now accepts that the Tribunal has jurisdiction in respect of all of the respondents, i.e. that it is competent (in the international jurisdiction sense) to adjudicate the claims against all of the respondents including UBV.

They also accept that the WTR apply to the claimants provided they are workers as defined;

They also accept that the ERA and the NMWA would apply to any claim against ULL provided they are workers.

But they say that the ERA and NMWA do not apply to any contract with UBV – Dutch law applies such that the claimants do not have any protection under UK employment legislation.”

The ET decided that:-

English law applied;

The Claimants were “employed” by ULL as “workers” within the meaning of section 230(3)(b) of the ERA 1996, the Working Time Regulations and the NMWA;

The working time of each of the Claimants started as soon as he was within his “territory” (London), had the App switched on and was ready and willing to accept trips, and ended as soon as any of those three conditions ceased to apply;

For the purposes of the National Minimum Wage Regulations 2015 the Claimants were engaged in “unmeasured work”.

The first and last of these rulings were not the subject of argument before us.

The driver is not made aware of the destination until he has collected the passenger (he learns it from the passenger directly or, where the passenger has stated the destination to Uber, from the app, when he presses the ‘Start Trip’ button). The App incorporates software linked to satellite navigation technology, providing detailed directions to the destination. The driver is not bound to follow the route proposed and will not do so if the passenger stipulates a different route. But an unbidden departure from the App route may have adverse consequences for the driver (see below).

On arrival at the destination, the driver presses or swipes the 'Complete Trip' button on his smartphone. Assuming he remains logged on to the App, he is then eligible to be allocated further trips.

At the end of any trip, the fare is calculated by the Uber servers, based on GPS data from the driver's smartphone. The calculation takes account of time spent and distance covered. In 'surge' areas, where supply and demand are not in harmony, a multiplier is applied to fares resulting in a charge above the standard level.

Strictly speaking, the figure stipulated by Uber is a recommended fare only and it is open to drivers to agree lesser (but not greater) sums with passengers. But this practice is not encouraged and if a lower fare is agreed by the driver, UBV remains entitled to its 'Service Fee' (see below) calculated on the basis of the recommended amount.

The passenger pays the fare in full to UBV, by credit or debit card, and receives a receipt by email. Separately, UBV generates paperwork which has the appearance of being an invoice addressed to the passenger by the driver. The 'invoice' document does not show the full name or contact details of the passenger, just his or her first name. Nor is it sent to the passenger. He or she would no doubt be vexed to receive it, having already paid the fare in full to Uber and received a receipt. The relevant driver has access to it electronically through the App. It serves as a record of the trip undertaken and the fare charged, but

The acceptance of PHV Bookings in accordance with paragraph 3 above, but without prejudice to Uber UK's rights at its sole and absolute discretion to decline any PHV

Allocating each accepted PHV Booking to a Transportation Provider via such means as Uber UK may choose;

Keeping a record of each accepted PHV Booking;

Remotely monitoring ... the performance of the PHV Booking by the Transportation Provider;

Receipt of and dealing with feedback, questions and complaints relating to PHV Bookings ... You are encouraged to provide your feedback if any of the transportation services provided by the Transportation Provider do not conform to your expectations; and

Managing any lost property queries relating to PHV Bookings.

Part 2 of the Rider Terms sets out detailed provisions purporting to govern the conditions on which the passenger is given access to the App. They avowedly characterise a contractual relationship between the passenger and UBV and are declared to be exclusively governed by the laws of the Netherlands. Para 2 includes these passages:

Para 4, entitled "Payment", includes the following:

Para 5 contains a lengthy disclaimer in respect of the use of the "Services" and an even longer clause purporting to exclude or limit UBV's liability for any loss or damage suffered by the passenger as a result of his or her use of the "Services".”

It is common ground that the vast majority of Uber drivers were and are sole operators such as Mr Aslam and Mr Farrar. Nonetheless, for the purposes of the Partner Terms, they provided "Driving Services" through their "Drivers" (ie in the ordinary case, themselves) to the "Customers".

A number of other features of the Partner Terms are worthy of note. By para 4.3.4 Partners were required to "support Uber in all communications", actively engage other Partners or Drivers if requested to do so and refrain from speaking negatively about Uber's business and business concept in public. Several provisions in para 9 imposed mutual duties of confidentiality.

The ET noted (at paragraph 40) that those interested in becoming Uber drivers can sign up online but must attend a specified location, produce certain documents and undergo a form of induction known as “onboarding”. They recorded that Ms Bertram appeared to suggest in evidence that there was no requirement for personal attendance by the putative driver and said “if that was her suggestion, we reject it”. They accepted “the general tenor of her evidence that Uber does not subject applicant drivers to close scrutiny”, adding: “that said, they must present themselves and their documents personally and they are, we find, subjected to what amounts to an interview, albeit not a searching one”.

The ET also recorded that the driver supplies the vehicle. Uber publishes a list of makes and models which it will accept. Vehicles have to be in good condition, manufactured in or after 2006 and preferably black or silver. The driver is responsible for all costs incidental to owning and running the vehicle, including fuel, repairs, maintenance, MOT inspections, road tax and insurance.

The Claimants’ case before the ET was that “in a host of different ways, Uber instructs, manages and controls the drivers”. The ET were shown a “Welcome Packet” containing materials used in the “onboarding” of new drivers. Under the heading “what Uber looks for” the following appeared:-

The ET made findings about acceptance and cancellation of trips as follows:-

Drivers who decline three trips in a row are liable to be forcibly logged off the App by Uber for 10 minutes. Ms Bertram denied that this amounted to a penalty but an Uber document called "Confirmation and Cancellation Rate Process" shows that the expression "Penalty Box warning" is current within the organisation. The third in a graduated series of standard form messages reads:

A similar system of warnings, culminating in the 10-minute log-off penalty, applies to cancellations by drivers after a trip has been accepted. As we have mentioned, the New Terms (and the Driver Addendum) provide that the right to cancel is subject to Uber's cancellation policy. There appears to be no document setting out the policy but the standard form warning messages state that cancellation amounts to a breach of the agreement between the driver and Uber unless there is a "good reason" for cancelling. A message from ULL to a driver dated 19 September 2014 reads:

The Claimants also relied on the ratings system as evidence of control. Passengers are asked by email at the end of every trip to rate drivers on a scale from zero (worst) to five (best). Ratings are monitored and drivers with average scores below 4.4, once they have undertaken their first 200 trips, become subject to a graduated series of “quality interventions” aimed at assisting them to improve. Experienced drivers, that is to say those who have undertaken 200 trips or more, whose figures do not improve to 4.4 or better, are “removed from the platform” and their accounts “deactivated”.

There is a rule prohibiting drivers from exchanging contact details with passengers or contacting them after the end of the trip, except for the purpose of returning lost property.

As well as undertaking work for or through Uber, drivers are expressly permitted by clause 2.4 of the 2015 New Terms to work for or through other organisations, including direct competitors operating through digital platforms. The drivers, as we have noted, must meet all expenses associated with running their vehicles. They must fund their own individual private hire licences. They treat themselves as self-employed for tax purposes. They are free (subject to being accepted by Uber) to elect which of the Uber “products” (Uber X, Uber Pool and various Uber Deluxe products) to operate. They are not provided with any clothing in the nature of an Uber uniform. In London they are discouraged from displaying Uber branding of any kind.

The Private Hire Vehicles (London) Act 1998 (“the PHVA 1998”) provides, so far as material:-

(1) Any person may apply to the licensing authority for a London PHV Operator Licence.

An application under this section shall state the address of any premises in London which the applicant proposes to use as an operating centre.

The licensing authority shall grant a London PHV Operator Licence to the applicant if the authority is satisfied that-

the applicant is a fit and proper person to hold a London PHV operator licence …..

Any further requirements that may be prescribed (which may be requirements relating to operating centres) are met.

(1) The holder of a London PHV Operator’s Licence (in this Act referred to as a “London PHV Operator”) shall not in London accept a private hire booking other than at an operating centre specified in his licence.

A London PHV operator shall secure that any vehicle which is provided by him for carrying out a private hire booking accepted by him in London is:

a vehicle for which a London PHV Licence is in force driven by a person holding a London PHV driver’s licence or;

a London cab driven by a person holding a London Cab driver’s licence.

(1) A London PHV operator (the first operator) who has in London accepted a private hire booking may not arrange for another operator to provide a vehicle to carry out that booking as sub-contractor unless;

It is immaterial for the purposes of subsection (1) whether or not subcontracting is permitted by the contract between the first operator and the person who made the booking.

For the avoidance of doubt (and subject to any relevant contract terms) a contract of hire between a person who made a private hire booking at an operating centre in London and a London PHV Operator who accepted the booking remains in force despite the making of arrangements by that operator to provide a vehicle to carry out that booking as sub-contractor.”

The Private Hire Vehicles (London) (Operators Licences) Regulations 2000 originally provided by regulation 9(3):-

agree the fare for the journey booked or,

provide an estimate of that fare.”

Agree the fare with the person making the private hire booking or;

Provide an accurate estimate of the fare to the person making the private hire booking.

Section 230 of the ERA 1996 provides:-

In this Act “employee” means an individual who has entered into or works under (or, where the employment has ceased, worked under) a contract of employment.

In this Act “contract of employment” means a contract of service or apprenticeship, whether express or implied, and (if it is express) whether oral or in writing.

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

any other contract, whether express or implied and (if it is express) whether oral or in writing, whereby the individual undertakes to do or perform personally any work or services for another party to the contract whose status is not by virtue of the contract that of a client or customer of any profession or business undertaking carried on by the individual; and any reference to a worker’s contract shall be construed accordingly.

In this Act “employer”, in relation to an employee or a worker, means the person by whom the employee or worker is (or, where the employment has ceased, was) employed.

In this Act “employment”—

in relation to an employee, means (except for the purposes of section 171) employment under a contract of employment, and

in relation to a worker, means employment under his contract; and “employed” shall be construed accordingly.”

The phrase “limb (b) worker” is now widely used to refer to individuals working under a contract within section 230(3)(b) above.

Section 43K of the ERA provides an extended definition of “worker” for the purposes of the legislation on protected disclosures. There are also extended definitions of “worker” under regulation 36 of the WTR and under Section 34 of the NMWA which provides:-

is supplied by a person “the agent” to do work for another (“the principal”) under a contract or arrangements made between the agent and the principal but;

is not as respects that work a worker, because of the absence of a worker’s contract between the individual and the agent or the principal and;

is not a party to a contract under which he undertakes to do the work for another party to the contract whose status is by virtue of the contract that of a client or customer of any professional or business undertaking carried on by the individual.”

The ET’s central conclusion (at paragraphs 85-86) was as follows:-

In view of the conclusion that ULL was the employer the conflict of laws issue became irrelevant, but for the avoidance of doubt the ET held that, on Rome I principles, English law would have been applicable in any event.

The Appellants appealed to the EAT. The case was heard by Judge Eady QC (sitting alone) on 27-28 September 2017. Uber’s grounds of appeal can be summarised as being that:-

The ET had erred in law in disregarding the written contractual documentation. There was no contract between the Claimants and ULL but there were written agreements between the drivers and UBV and riders, which were inconsistent with the existence of any worker relationship. As those agreements made clear, Uber drivers provided transportation services to riders; ULL (as was common within the mini-cab or private hire industry) provided its services to the drivers as their agent. In finding otherwise, the ET had disregarded the basic principles of agency law.

The ET had further erred in relying on regulatory requirements as evidence of worker status.

It had also made a number of “inconsistent” and “perverse” findings of fact in concluding that the Claimants were required to work for Uber.

It had further failed to take into account relevant matters relied on by Uber as inconsistent with worker status and as, on the contrary, strongly indicating that the Claimants were carrying on a business undertaking on their own account.

In her reserved judgment handed down on 10 November 2017 Judge Eady dismissed the appeal. She held that the ET had been entitled to reject the characterisation of the relationship between the drivers and Uber, specifically ULL, set out in the written contractual documents. Applying Autoclenz Ltd v Belcher [2011] UKSC 41, [2011] ICR 1157, the ET had to determine what was the true agreement between the drivers and ULL. In so doing it was important for the ET to have regard to the reality of the obligations and of the factual situation. The starting point must always be the statutory language, not the label used by the parties; simply because the parties have used the language of self-employment does not mean that the contract does not fall within section 230(1)(b). After referring to Secret Hotels2 Ltd (formerly Med Hotels Ltd) v Revenue and Customs Commissioners [2014] UKSC 16, [2014] 2 All ER 685, she continued:-

She therefore upheld the central finding of the ET that the drivers were “workers” providing their services to ULL. She held that the findings of the ET were consistent and that Uber had not met the high burden of showing that they were perverse.

On the issue of the time during which the drivers were to be treated as working, she found a “more difficult” issue the ET’s finding that the drivers were workers not only when they had accepted a trip request or were carrying passengers for Uber, but also in between accepting assignments. She held that, taking the ET’s findings in the round, “it permissibly found that Uber drivers assume an obligation when they are in the Territory, switch on the App and are available for work.” She added:-

She therefore dismissed the appeal. On a subsequent application by Uber, Judge Eady gave permission to appeal to this court but refused a certificate under Section 37ZA of the Employment Tribunals Act 1996 to enable Uber to make a leapfrog application to the Supreme Court.

Many reported cases have considered the distinction between a limb (b) worker and a self-employed contractor. In Cotswold Developments Construction Ltd v Williams [2006] UKEAT 0457, [2006] IRLR 181, Langstaff J suggested that:

In James v Redcats (Brands) Ltd [2007] UKEAT 0475, [2007] ICR 1006, Elias J said:

In the Supreme Court case of Bates van Winkelhof v Clyde & Co LLP and another [2014] UKSC 32, [2014] 1 WLR 2047, in which the central issue was whether a member of a limited liability partnership was a limb (b) worker, Lady Hale DPSC said:

Lady Hale also referred with approval to the previous observations of Langstaff J and Elias J which we have quoted.

The leading case is Autoclenz Ltd v Belcher. The claimants carried out car cleaning services on behalf of the appellant company. In order to obtain the work they were required to sign contracts which stated that they were sub-contractors and not employees, that they had to provide their own material, that they were not obliged to provide services to the company nor was the company obliged to provide work to them, and that they could provide suitably qualified substitutes to carry out the work on their behalf. They brought tribunal proceedings claiming that they were “workers” entitled to the national minimum wage (“the NMW”) and to statutory paid leave under the WTR. The ET found that the contractual documents did not reflect the true agreement between the parties and that the claimants came within both limbs of the definition of “worker” as (a) working under contracts of employment and as (b) working pursuant to contracts for services. The former finding was the subject of differing decisions on appeal but, since there is no suggestion in the present case that the Claimants have contracts of employment with any of the Uber companies, we need not consider it further. The finding that the claimants in Autoclenz were “workers” under contracts for services was upheld in the EAT, in this court and in the Supreme Court.

Lord Clarke of Stone-cum-Ebony JSC, with whom all the other members of the Supreme Court agreed, said (at paragraph 17) that the case:-

He said, at paragraph 20, that “the essential question in each case is what were the terms of the agreement.” He referred to Snook v London and West Riding Investments Ltd [1967] 2 QB 786 in which Diplock LJ had referred to the concept of a sham as being acts done or documents executed by the parties intended by both of them to give an appearance to third parties of creating legal rights and obligations different from the actual ones which the parties intended to create. In that type of case, Diplock LJ went on, all the parties must have a common intention that the acts or documents are not to create the legal rights and obligation which they give the appearance of creating.

Lord Clarke continued (at paragraph 23):-

He went on to approve these observations of Elias J in the EAT in Consistent Group Ltd v Kalwak [2007] UKEAT 0535, [2007] IRLR 560:

In other words, if the reality of the situation is that no one seriously expects that a worker will seek to provide a substitute, or refuse the work offered, the fact that the contract expressly provides for these unrealistic possibilities will not alter the true nature of the relationship. But if these clauses genuinely reflect what might realistically be expected to occur, the fact that the rights conferred have not in fact been exercised will not render the right meaningless.

… Tribunals should take a sensible and robust view of these matters in order to prevent form undermining substance…"

Lord Clarke continued:

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.

With characteristic clarity and brevity Sedley LJ described the factual position as follows:

The contracts began by spelling out that each worker was required to 'perform the services which he agrees to carry out for Autoclenz within a reasonable time and in a good and workmanlike manner' - an obligation entirely consistent with employment. Notwithstanding the repeated interpolation of the word 'sub-contractor' and the introduction of terms inconsistent with employment which, as the judge found, were unreal, there was ample evidence on which the judge could find, as he did, that this was in truth an employment relationship.

His finding did not seek to recast the contracts: it was a finding on the prior question of what the contracts were. Rightly, it was uninfluenced by the fiscal and other consequences of the relationship, which were by no means all one way."

There is no dispute that Autoclenz puts paid, at least in an employment context, to the idea that all that matters is the terms of any written contract, with the exception of a document intended by all parties executing it to be a sham. Clearly, however, the case goes a good deal further. We regard as particularly significant Lord Clarke’s endorsement of the advice of Aikens LJ to tribunals to be “realistic and worldly wise” in this type of case when considering whether the terms of a written contract reflect the real terms of the bargain between the parties; and of the similar advice of Elias J that tribunals should take a “sensible and robust view of these matters in order to prevent form undermining substance”.

We also attach importance to the approval given by Lord Clarke to the conclusions drawn by Sedley LJ in this court from what he (Lord Clarke) described as the “critical findings of fact” by Employment Judge Foxwell in the ET. Judge Foxwell noted that the claimants had no say in the terms on which they performed work; the contracts were devised entirely by Autoclenz; and the services they provided were subject to a detailed specification. The claimants had no control over the way in which they did their work. Judge Foxwell’s conclusion from the facts was that the “elaborate protestations in the contractual documents that the men were self-employed” bore no practical relation to the reality of the relationship. Consequently, Lord Clarke held, the documents did not reflect the true agreement between the parties. The ET had been entitled to “disregard” the terms of the written documents, insofar as they were inconsistent with the true terms agreed between the parties.

Ms Dinah Rose QC, for Uber, laid great emphasis on the later decision of the Supreme Court in Secret Hotels2. The appellant company, formerly known as Med Hotels, marketed hotel rooms and other holiday accommodation through a website. Any hotelier who wished his hotel to be marketed by the company had to enter into an accommodation agreement which began by identifying the hotelier as the “Principal” and the company as the “Agent”. The hotelier as Principal appointed the company as its selling agent and the company agreed to act as such. The company agreed to deal accurately with the clients’ requests for accommodation bookings and relay all money it received from the Principal’s clients which was due to the Principal. A potential customer (whether a travel agent or an individual holidaymaker who used the website) would be referred to booking conditions which stated that “reservations you make on this site will be directly with the company whose hotel services you are booking” and that the company was acting “as agent only for each of the hotels to provide you with information on the hotels and an online reservation service”. The customer had to pay

An issue arose under both domestic and EU law as to the treatment of these transactions for VAT purposes. The Revenue assessed the company for VAT on the gross sums it received from clients. The company challenged this on the basis that it was a travel agent acting solely as an “intermediary”. After decisions by the First Tier Tribunal, the Upper Tribunal and this court, the Supreme Court held that the effect of the contractual documentation was that the company marketed and sold accommodation to customers as agent of the hoteliers and that it was an “intermediary” for tax law purposes. It was in that context that Lord Neuberger (with whom all other members of the court agreed) said:

When interpreting an agreement, the court must have regard to the words used, to the provisions of the agreement as whole, to the surrounding circumstances in so far as they were known to both parties, and to commercial common sense. When deciding on the categorisation of a relationship governed by a written agreement, the label or labels which the parties have used to describe their relationship cannot be conclusive, and may often be of little weight. As Lewison J said in A1 Lofts Ltd v Revenue and Customs Commissioners [2010] STC 214, para 40, in a passage cited by Morgan J:

In English law it is not permissible to take into account the subsequent behaviour or statements of the parties as an aid to interpreting their written agreement – see FL Schuler AG v Wickman Machine Tool Sales Ltd [1974] AC 235. The subsequent behaviour or statements of the parties can, however, be relevant, for a number of other reasons. First, they may be invoked to support the contention that the written agreement was a sham – ie that it was not in fact intended to govern the parties' relationship at all. Secondly, they may be invoked in support of a claim for rectification of the written agreement. Thirdly, they may be relied on to support a claim that the written agreement was subsequently varied, or rescinded and replaced by a subsequent contract (agreed by words or conduct). Fourthly, they may be relied on to establish that the written agreement represented only part of the totality of the parties' contractual relationship.

In the present proceedings, it has never been suggested that the written agreements between Med and hoteliers, namely the Accommodation Agreements, were a sham or liable to rectification. Nor has it been suggested that the terms contained on the website ("the website terms"), which governed the relationship between Med and the customers, namely the Terms of Use and the Booking Conditions, were a sham or liable to rectification. In these circumstances, it appears to me that (i) the right starting point is to characterise the nature of the relationship between Med, the customer, and the hotel, in the light of the Accommodation Agreement and the website terms ("the contractual documentation"), (ii) one must next consider whether that characterisation can be said to represent the economic reality of the relationship in the light of any relevant facts, and (iii) if so, the final issue is the result of this characterisation so far as article 306 is concerned.

This is a slightly more sophisticated analysis than the single issue as it has been agreed between the parties, as set out in para 16 above, but, as will become apparent, at least in the circumstances of this case, it amounts to the same thing. In order to decide whether the FTT was entitled to reach the conclusion that it did, one must identify the nature of the relationship between Med, the hotelier, and the customer, and, in order to do that, one must first consider the effect of the contractual documentation, and then see whether any conclusion is vitiated by the facts relied on by either party.”

Autoclenz was not mentioned in the judgment, nor even apparently cited in argument, in Secret Hotels2. The latter is obviously not an employment case and there was no suggestion that the written terms misrepresented what was occurring on the ground. There was undoubtedly a contract between the company and each hotel, in contrast to the present case where Uber seek to argue that there is no contractual relationship between the drivers and ULL.

In the course of supplementary oral submissions Ms Rose argued that Autoclenz could not be used to disregard the Rider Terms, since these were a contract between passenger and driver, not an employment contract in any sense. Instead, she said, we should follow

Ms Rose cited two other decisions about minicab drivers. Mingeley v Pennock and others (trading as Amber Cars) [2004] EWCA Civ 328, [2004] ICR 727, was not about “worker” status but about whether the claimant was entitled to bring a claim of racial discrimination under section 78 of the Race Relations Act 1976, which he could only do if there was a contract of employment within the meaning of the section. There was a preliminary issue as to whether he was required personally to execute any work or labour. The issue was decided in favour of the respondents in the ET, the EAT and this court.

The essential facts were that the Claimant owned his own vehicle and paid the respondents minicab operators £75 per week for a radio and access to their company system, which allocated calls from customers to a fleet of drivers. He was required to wear a uniform and prohibited from working for any other operator, but was not required to work, nor (in contrast to the present case) to accept any fare allocated to him by the system. All the fare money was his to keep.

The judgments in this court were unreserved. Maurice Kay LJ, with whom Nourse LJ agreed, held that the absence from Mr Mingeley’s contract with the respondents of any obligation to work placed him “beyond the reach” of section 78. Buxton LJ said that:

The other minicab case to which we were referred at the oral hearing before us was Khan v Checkers Cars Ltd, an unreported decision of the EAT handed down on 16 December 2005. This also was an employment contract case. The claimant worked as a private hire car driver for the respondent company which operated a taxi service based at Gatwick Airport. The claimant owned and was responsible for his own vehicle. He paid his own income tax and national insurance. He was required to use set routes and charge set fares. He collected fares from customers, paying a commission to the respondent. He had complete flexibility over when he worked: he was not obliged to accept work and the respondent was not obliged to offer him work. Drivers were never required to attend work. The only issue (since the claim was one of unfair dismissal)

On 14 November 2018, after the oral hearing of this appeal, HHJ David Richardson gave the judgment of the EAT (himself and two lay members) in Addison Lee Ltd v Lange [2018] UKEAT 37. The claimants were drivers working for Addison Lee’s PHV business. Almost invariably they used a vehicle hired from Eventech Ltd, an associated company of the respondents (in contrast to the owner-drivers in the present case). The vehicles were in Addison Lee livery. Each driver was given a hand held computer known as an XDA. When ready to work the driver would use the XDA to log on to the respondent’s computer system which could locate the XDA and the vehicle. Allocation of jobs was automatic. When a job was notified to the driver he had to accept it forthwith or give an acceptable reason for not doing so. If the controller deemed the reason to be unacceptable, the matter was referred to a supervisor and a sanction might follow.

Each driver had a Driver Contract with the respondent. It provided (more than once) that the driver agreed he was an independent contractor and that nothing in the agreement rendered him an employee, worker, agent or partner of the respondent.

Clause 5, under the heading “Provision of Services”, stated:

For the avoidance of doubt, there is no obligation on you to provide the Services to Addison Lee or to any Customer at any time or for any minimum number of hours per day/ week/month. Similarly, there is no obligation on Addison Lee to provide you with a minimum amount of, or any, work at all.

You agree to perform promptly each Customer Contract in accordance with its terms and to indemnify us against any claims from Customers for your breach of the Customer Contract which are directed against us as a result either of having acted as your agent in concluding the Customer Contract or as principal where you have fulfilled the Customer Contract as a sub-contractor on our behalf.

By ticking the appropriate box at the start of this Driver Contract you select which of the "Anytime Circuit", the "Night Circuit", or the "Weekend Circuit" you wish to participate in.

As a Driver on the Anytime Circuit, you are indicating that, subject to Clause 5.2, you may be available to provide the Services whenever you wish under this Driver Contract at any time.

As a Driver on the Night Circuit, you are indicating that, subject to Clause 5.2, you may be available to provide the Services whenever you wish from 1700hrs each day until 0500hrs the following day.

As a Driver on the Weekend Circuit, you are indicating that, subject to Clause 5.2, you may be available to provide the Services whenever you wish from 1730hrs each Friday to

The ET held that: (a) there was an overarching contract between each claimant and the respondent; (b) but in any event, whether that was so or not, the claimants were workers within the meaning of the legislation; (c) whenever they logged on, they were undertaking to provide driving services personally; (d) even if they chose to park in a vehicle but remained logged on, they were no less at the disposal of Addison Lee.

The EAT held, at paragraph 55, that “applying Autoclenz principles the ET was..…entitled to reach the conclusion…..that the drivers, when they logged on, were undertaking to accept the driving jobs allocated to them”. They held that this conclusion was consistent with the finding that the driver had to accept a job allocated to him in the absence of an acceptable reason and that if he did not do so a sanction might be imposed. As to the terms of Clause 5.2 on which the respondents placed reliance, the EAT observed that:

The ET had disregarded some provisions of the Driver Agreement, particularly clause 5.2. The EAT found that the ET had been entitled to do so by application of the Autoclenz principles because the relevant provisions did not reflect the reality of the bargain made between the parties. After referring to other authorities, including the observations of Langstaff J in the Cotswold Developments case set out above in the citation from Autoclenz, they dismissed the appeal.

Although the facts of Addison Lee are not identical to those of the present case we consider it helpful, both in the summary by Judge Richardson of the relevant case law (which we will not repeat here) and because of the finding that the ET had been entitled to disregard clauses in the Driver Contract which did not reflect the reality of the bargain between the parties.

In their supplementary submissions on behalf of Uber, Ms Rose and Mr Campbell seek to distinguish the case on the grounds that “disregarding the written contract between Addison Lee and the drivers did not involve disregarding any contracts with third parties outside the employment field”. We do not accept that this is a significant distinction. The effect of Autoclenz in our view is that, in determining for the purposes of section 230 of the ERA 1996 what is the true nature of the relationship between the employer and the individual who alleges he is a worker or an employee, the court may disregard the terms of any documents generated by the employer which do not reflect the reality of what is occurring on the ground.

Ms Rose also cited two very different authorities. In Cheng Yuen-v-Royal Hong Kong Golf Club [1998] ICR 131 PC the question was whether the claimant was an employee or an independent contractor. It did not concern whether he was a “worker”. The claimant worked as a caddie for individual members of the respondent golf club. He was issued by the club with a number, a uniform and a locker. Caddying work was allocated to available caddies in strict rotation. They were not obliged to make themselves available for work and received no guarantee of work. The club was not obliged to give them work or to pay anything other than the amount of the fee per round owed by the individual golfer for whom they had caddied.

When told that his services were no longer required, the claimant brought claims against the club, for the purposes of which it was essential to show that he had been an employee of the club rather than an independent contractor. The majority of the Privy Council concluded that he had not been an employee. This is not a surprising conclusion since, as Lord Slynn emphasised in delivering the majority judgment, there was no mutuality of obligation. The case is of no assistance in deciding whether the Claimants in the present case are workers providing services to ULL.

Ms Rose also placed reliance on Stringfellow Restaurants Ltd v Quashie [2013] IRLR 99 CA; [2012] EWCA Civ 1735. That again was not a case about “worker” status but about whether the claimant was an employee or an independent contractor. The claimant was a lap dancer who performed for the entertainment of guests at the respondents' clubs. She paid the respondent a fee for each night worked. Doing so enabled her to earn substantial payments from the guests for whom she danced. She negotiated those payments with the guests. The respondent ended its working relationship with her and she complained of unfair dismissal. At a preliminary hearing, an ET held that there was no contract of employment. The EAT disagreed but the Court of Appeal restored the first instance decision. Elias LJ gave the only substantive judgment. After discussing the Cheng Yuen case, he said this:

Central to Elias LJ’s conclusion was the finding that the claimant took an economic risk in view of the fixed sums which she had to pay the club irrespective of the number of her customers. As with the golf club case and for similar reasons, we did not find this case of any assistance.

Whether or not there was a contract between each of the Claimants and ULL is a mixed question of fact and law. It has often been said in this court that an appellate court should not interfere with such a determination of the first instance court unless no reasonable tribunal, properly directing itself, could have reached the decision it did: see e.g. Stringfellow at [9]. What were the terms of any such contract, in the absence of a comprehensive written agreement, is a question of fact: Carmichael v National Power [1999] ICR 1226 at 1233B-C, Secret Hotels2 at [20].

As discussed above, Autoclenz shows that, in the context of alleged employment (whether as employee or worker), (taking into account the relative bargaining power of the parties) the written documentation may not reflect the reality of the relationship. The parties’ actual agreement must be determined by examining all the circumstances, of which the written agreement is only a part. This is particularly so where the issue is the insertion of clauses which are subsequently relied on by the inserting party to avoid statutory protection which would otherwise apply. In deciding whether someone comes within either limb of section 230(3) of the ERA 1996, the fact that he or she signed a document will be relevant evidence, but it is not conclusive where the terms are standard and non-negotiable and where the parties are in an unequal bargaining position. Tribunals should take a “realistic and worldly-wise”, “sensible and robust” approach to the determination of what the true position is.

An overarching argument of Ms Rose for Uber was that all the operational matters relied upon by the ET and put forward by the Claimants for characterising them as limb (b) workers are entirely consistent with them being simply conditions of the licence to use the App, and in that way entirely consistent with the written agreements between UBV and the drivers and between UBV and the passengers.

We suggest that the answer to that point is to look at the different stages in the process of carrying out a passenger’s request for a ride: (1) the request made to ULL by the passenger and its acceptance by the driver; (2) the picking up of the passenger by the driver; and (3) the completion of the journey and calculation of the fare, where estimated in advance.

At stage (1), acceptance of the request by the driver means that, subject to the right of the driver and the passenger to cancel, the driver is expected to proceed to collect the passenger from the notified location and to complete the journey. That is consistent with the language of 2.4 of the 2015 New Terms, which talks of the option “to cancel an accepted request”. In the language of Autoclenz, at paragraph 35, the “reality” is that at that stage there is an obligation on the driver to fulfil that expectation. The contractual documentation states that, at that stage, there is a contract between the driver and the passenger but that cannot be correct as vital elements of any such contract are missing.

It is also true that the driver does not know what the fare will be where ULL (as the PHV operator) has given an estimate, as the actual fare will be determined by Uber at the end of the ride. That, however, while relevant to the issue of the reality of the situation, may be seen as less decisive as a matter of strict contract law as it is arguably consistent with a contract that the fare will be as determined by Uber.

Our initial view was that, irrespective of the absence of agreement on an essential term (the destination), the passenger has not, at that stage, provided any consideration for an obligation on the driver to collect him or her. In supplementary submissions following circulation of draft judgments, which we heard in private at Ms Rose’s request as they concerned the draft judgment, Ms Rose argued that this point had not been argued below (nor expressly before us); that it was a mixed question of law and fact; and that it was unfair to her clients that it should be taken for the first time in this court by the court itself.

We do not think that there has been any unfairness to Uber: it has been clear from the start that the Claimants’ case is that there is in reality no contract between passenger and driver. However, the question of whether any consideration passes between driver and passenger is a minor point and, on reflection, we are content not to pursue it.

The passenger has no contract to compel the driver to pick up him or her. The contract at the point of acceptance of the request must be with ULL. The request is communicated to the driver by ULL and is accepted by the driver in responding to Uber. There is no basis for saying that it is with UBV, via the agency of ULL, as there is nothing in either version of the UBV agreement that says that ULL, in sending the request and receiving the driver’s acceptance, is acting as UBV’s agent. Clause 2.2 of the New Terms says that ULL is at that stage acting as the driver’s agent but, plainly, that cannot be correct if there is a contract between ULL and the driver.

Both the existence of any such contract and its terms must be established objectively.

The ET found that there is no contract between the driver and the passenger. That is not a necessary finding in order to support a contract between ULL and the driver at the point of acceptance by the driver. There is no analytical reason why there could not be two contracts subsisting at the same time, or rather from the time of pick up, the contract between ULL and the driver having commenced at an earlier point of time in accordance with the above analysis. But any contract was plainly not one under which

If that analysis is correct, two fundamental strands of Ms Rose’s submissions fall away. First, there is no general comparison with minicabs. There is no evidence about how minicabs generally operate. What is clear is that there is more than one business model for minicabs. That is apparent from the VAT Notices (BA2/ 54 AND 56). What is critical is that there is no evidence of contractual arrangements for minicabs which precisely mirror the contractual arrangements above, that is to say the contract between the driver and the operator (ULL) at a time when the driver does not know the intended destination of the intended passenger.

Secondly, in addition to the points we have already made distinguishing the cases relied upon by Ms Rose, the situation in the present case is highly fact specific and is not matched by that in any of those cases.

The minicab cases such as Mingeley and Khan considered above do not address the issue of the status of the minicab firm as statutory PHV operator, a regime which in any event is not the same outside London (Mr Mingeley worked in Leeds and Mr Khan was based at Gatwick Airport).

The analogy with black cab drivers drawn by Ms Rose is not helpful. Black cab drivers ply for hire, can advertise in their own right and can contract directly with passengers.

The Appellant’s submissions repeatedly referred to the regulatory regime as if it were irrelevant or of trivial importance. We disagree. In our view the statutory position strongly reinforces the correctness of the ET’s conclusion that the drivers were providing services to Uber (specifically to ULL), not the other way round.

ULL is the PHV operator for the purposes of the PHVA 1998 and the regulations made under it. It is ULL which has to satisfy the licensing authority for the purposes of section 3(3)(a) of the Act that it is a fit and proper person to hold a PHV licence. It is ULL which alone can accept bookings, and ULL which is required by the PHV Regulations to provide an estimate of the fare on request. For ULL to be stating to its statutory regulator that it is operating a private hire vehicle service in London, and is a fit and proper person to do so, while at the same time arguing in this litigation that it is merely an affiliate of a Dutch registered company which licenses tens of thousands of proprietors of small businesses to use its software, contributes to the air of contrivance and artificiality which pervades Uber’s case.

Consistently with what we have said about the reality being reinforced by the regulatory framework, it is of interest to note that section 56 of the Local Government (Miscellaneous Provisions) Act 1976 expressly provides for the hire of a licensed private hire vehicle to be deemed to be made with the operator who accepted the booking, whether or not he himself provided the vehicle. For this purpose, it is irrelevant that the Act only applies outside London.

There is a high degree of fiction in the wording (whether in the 2013 or the 2015 version) of the standard form agreement between UBV and each of the drivers:-

ULL, despite being the PHV operator in London, and therefore the only entity legally permitted to operate the business, is scarcely mentioned at all, even as an “Affiliate” of UBV;

The agreement refers to the party with whom UBV is contracting as the “Partner” (2013) or “Customer” (2015), as if it were a separate legal entity employing one or more drivers. Indeed, in the 2015 version, the

We agree with the submission of Mr Linden QC, for some of the

The omission of ULL from both versions of the standard terms is all the more striking because ULL enforces a high degree of control over the drivers and for the most part does so (quite understandably and properly) in order to protect its position as PHV operator in London. It is difficult to see on what basis ULL is entitled to act in this way

The ET found at paragraph 20 that after each ride has been completed UBV “generates paperwork which has the appearance of being an invoice addressed to the passenger by the driver. The invoice document does not show the full name or contact details of the passenger, just his or her first name. Nor is it sent to the passenger.” The ET described this standard document (at paragraph 87) as a “fiction”; as it clearly was.

There are, on the other hand, some points made by the ET with which we cannot agree. We do not find helpful, whether or not correct, that “Uber’s case has to be that if the organisation became insolvent the drivers would have enforceable rights directly against the passengers”. We also disagree with the ET’s proposition that, if the Rider Terms were worker contracts, the passenger would be exposed to potential liability as the driver’s employer under enactments such as the NMWA: this would indeed be absurd but it cannot be correct since, on this hypothesis, the “client or customer” exception would apply. We also do not attach significance to the possibility that Uber might reverse its policy, which was in evidence before the ET, that it will reimburse a driver who suffers loss because of fraud by the passenger. Even added together, however, all these points form only a small part of the ET’s reasoning and are certainly not essential to its conclusions.

Ms Bertram told us that Uber provides the drivers with "business opportunities", but strenuously denied that they had jobs with the organisation. However, in a submission to the GLA Transport Scrutiny Committee ULL boasted of "providing job opportunities" to people who had not considered driving work and potentially generating "tens of thousands of jobs in the UK."

On the subject of payment of drivers, we have referred above to the Partner Terms and New Terms, which provide for Uber to collect fares on behalf of drivers and deduct their 'Commission' or 'Service Fee'. But in its written evidence dated 3 October 2014 to the GLA Transport Scrutiny Committee, Ms Bertram on behalf of ULL stated:

We agree with the ET’s finding at paragraph 92 that “it is not real to regard Uber as working “for” the drivers and that the only sensible interpretation is that the relationship is the other way round. Uber runs a transportation business. The drivers provide the skilled labour through which the organisation delivers its services and earns its profits.”

We set out below the thirteen considerations (in para 92 of the ET’s decision) which the ET said led them to that conclusion in italics, with our comments in ordinary type:- (1) The contradiction in the Rider Terms between the fact that ULL purports to be the driver’s agent and its assertion of “sole and absolute discretion” to accept or decline bookings. Ms Rose criticised this on the grounds that it was necessary because under the regime of the PHVA 1998 only ULL can accept or decline bookings. In our view, the fact that this is a statutory requirement does not invalidate its significance: if anything it reinforces it.

The fact that Uber interviews and recruits drivers. We agree with the ET that this is significant.

The fact that Uber controls the key information (in particular the passenger’s surname, contact details and intended destination) and excludes the driver from it. Ms Rose argued that these were important and desirable measures in the interests of passenger safety. We agree that they are: but, as with the statutory requirement that only ULL may accept or decline bookings, this does not detract from the significance of what is stated.

The fact that Uber requires drivers to accept trips and/or not to cancel trips, and enforces the requirement by logging off drivers who breach those requirements. We agree that this is significant as showing a high degree of control.

The fact that Uber sets the (default) route and the driver departs from it at his peril. This is not as stringent an element of control as some others because the driver may depart from the route prescribed by the App and the peril is only financial: nevertheless, it does have some significance.

The fact that UBV fixes the fare and the driver cannot agree a higher sum with the passenger. (The supposed freedom to agree a lower fare is obviously nugatory). Ms Rose submits that this also is a regulatory requirement; again, in our view, that fact does not detract from its significance in supporting the ET’s conclusion that Uber runs a transportation business and the drivers provide the skilled labour through which its services are provided.

The fact that Uber imposes numerous conditions on drivers (such as the limited choice of acceptable vehicles) instructs drivers on how to do their work, and in numerous ways, controls them in the performance of their duties. Ms Rose submitted that these conditions are standard in the taxi and minicab industry. No doubt they are, but again they support the ET’s findings that the drivers are working for Uber, not the other way around.

The fact that Uber subjects drivers through the rating system to what amounts to a performance management/disciplinary procedure. This is a powerful point supporting the case that the drivers work for Uber.

The fact that Uber determines issues about rebates, sometimes without even involving the driver whose remuneration is liable to be affected. This is another similar point, though somewhat less powerful than the last one.

The guaranteed earning schemes (albeit now discontinued). As the words in parenthesis indicate, these had ceased by the time the case came before the ET. We did not hear argument from either side on whether this was in reality a significant point.

The fact that Uber accepts the risk of loss which, if the drivers were genuinely in business on their own account, would fall upon them. The ET may have overstated this point in summarising it. As their findings at paragraph 26 made clear, Uber’s general practice is to accept the loss in cases where the passenger has procured the ride by fraud, at least where, as Ms Bertram put it, Uber’s systems have failed. On those findings this does not seem to us a point of real significance in the Claimants’ favour.

The fact that Uber handles complaints by passengers, including complaints about the driver. This is another regulatory requirement, but again it supports the Claimants’ case and the ET’s conclusion.

The fact that Uber reserves the power to amend the driver’s terms unilaterally. We agree that this supports the ET’s conclusion.

Viewing paragraph 92 of the ET’s decision as a whole, and even if one discounts points

Before the ET it was submitted by leading counsel on behalf of Uber (David Reade QC) that “if the drivers had any limb (b) relationship with the organisation, it must be with UBV. There was no agreement of any sort with ULL, which only exists to satisfy a regulatory requirement”. This was not a prominent feature of the submissions of Ms Rose before this court. For the reasons we have given above, and for the avoidance of doubt, we agree with the following findings of the ET at paragraph 98:-

If, as the ET found and we accept, the drivers were workers providing their services to ULL, the final question (argued only briefly before us) is at what times they were to be classified as so working. Uber places great emphasis on the fact that its standard terms (whether in the 2013 or the 2015 versions) expressly permit drivers to use other competing apps and to have more than one switched on at the same time. There appears to have been very little evidence before the ET as to how often this occurs in practice.

It is common ground that a driver can only be described as providing services to Uber when he is in the Territory (i.e., for present purposes, in London) and has the Uber App switched on. The Claimants contended, and the ET found, that they were providing services to ULL throughout the time when they satisfied these requirements. Uber submitted that, if (contrary to its primary submissions) the drivers were providing services to ULL, it could only be during each ride, that is to say from the time the passenger is picked up until the time the car reaches the passenger’s destination. A middle course is to say that the driver is providing services to ULL from the moment

The ET (at paragraph 100) accepted the Claimants’ submissions for the following reasons:-

In paragraph 102 they held, in the alternative, that “at the very latest the driver is “working” for Uber from the moment he accepts any trip.

We agree with the ET that at the latest the driver is working for Uber from the moment when he accepts any trip. The point which we have found much more difficult, as did Judge Eady QC in the EAT, is whether the driver can be said to be working for Uber when he is in London with the App switched on but before he has accepted a trip. In the end, like Judge Eady, we take the view that the conclusion in paragraph 100 was one which the ET were entitled to reach. We bear in mind that appeal from an ET lies only on a question of law (Employment Tribunals Act 1996, section 21(1)).

Even if drivers are not obliged to accept all or even 80% of trip requests, the high level of acceptances required and the penalty of being logged off if three consecutive requests are not accepted within the ten second time frame justify the ET’s conclusion that the drivers waiting for a booking were available to ULL and at its disposal. If a particular driver had entered into an obligation of the same nature for another entity and also had the rival app switched on then, as a matter of evidence, Uber would be able to argue that that driver was not at Uber’s disposal. As Judge Eady observed:-

We would dismiss this appeal.

I have the misfortune to disagree with the Master of the Rolls and Bean LJ about the outcome of the appeal in this case. I shall have to give my reasons fairly fully, but I can gratefully adopt the introductory and background material at paras. 1-37 of their judgment and will use their abbreviations.

The Claimants’ primary case, which the ET accepted, is that an Uber driver is a worker, within the meaning of the relevant statutes/regulations (Footnote: ¹ ) , throughout any period when

It is Uber’s case, by contrast, that the only contract that drivers enter into to provide work or services is a contract which they make with the passenger at the moment that they accept a trip, with Uber acting only as the driver’s agent in making the booking and collecting payment.

The question of for whom, and under a contract with whom, drivers perform their services is the central issue in the case. It is worth noting that it is different from the issue in most of the reported cases on employee and worker status, and the familiar questions of whether the putative worker contracts to provide his or her services personally or whether they do so for the putative employer as a “client or customer” are not directly engaged. Having said that, the issue is not entirely novel. It was at the heart of the decisions of the Privy Council in Cheng Yuen v Royal Hong Kong Golf Club [1998] ICR 131 and of this Court in Stringfellow Restaurants Ltd v Quashie [2012] EWCA Civ 1735, [2013] IRLR 99, to which I return at para. 144 below.

I will deal first in this judgment with whether the drivers provide services for ULL, and under a contract with it, at all, which I will call “the main issue”. I will then deal with the secondary, but still potentially important, issue of the period covered by any such contract and with the closely related issues of whether such periods constitute working time for the purpose of the WTR or fall to be taken into account in calculating the national minimum wage under the National Minimum Wage Regulations 2015 (“the NMWR”).

If the main issue depended on the terms of the written contract to which the Claimants, like all Uber drivers, have agreed (“the Agreement” (Footnote: ⁴ ) ), there would be no room for argument. The Master of the Rolls and Bean LJ have set out passages from the ET’s Reasons containing extensive extracts from the Agreement, but I will repeat the key provisions. Para. 2.3 of the New Terms reads:

It is thus perfectly explicit in the Agreement that drivers provide their services to the passengers as principals, with Uber’s role being that of an intermediary. The contract between Uber and the passenger is to the same effect: see para. 3 of the Rider Terms set out at para. 28 of the ET’s Reasons (para. 13 in the judgment of the Master of the Rolls and Bean LJ.)

The ET draws attention to two points about the Agreement which I should address, though they were not at the centre of the argument before us and are not relied on in my Lords’ reasoning.

First, it observes that few if any Uber drivers would in practice read the Agreement and that even if they did not all would understand its effect. I am sure that that is so, but they signed up to them (electronically rather than in hard copy) and on ordinary principles, and, subject to the question of the effect of Autoclenz which I consider below, they are bound by them whether they read them or not: L'Estrange v F. Graucob Ltd. [1934] 2 KB 394.

As I have already observed, Uber’s terminology is idiosyncratic. The New Terms distinguish between “Customer” and “Driver”, to cater for the case where the App is licensed to a business which makes available the service of drivers, that business being the Customer; the equivalent under the Partner Terms is the Partner. In virtually all cases, however, the licensees are ownerdrivers, and Customer and Driver can be treated as equivalent.

See previous footnotes: here “Customer” means passenger (or “User” in the terminology of the New Terms), and “Partner” for all practical purposes means driver.

Secondly, the Agreement is made not with ULL, which is the putative employer, but with UBV. But in this context that does not matter. This is not a question of privity of contract but of identifying for whom, contractually, the Claimants perform their services. The fact that they have agreed with UBV that they do not do so under a contract with it or any affiliate is just as much an obstacle to their case as if they had agreed it with ULL itself.

On the face of it, therefore, the Claimants have clearly agreed that they perform their services for, and under a contract with, the passenger and not for, or under a contract with, Uber. But their case, which the ET accepted, and which my Lords also accept, is that the terms of the Agreement negativing any agreement to perform services for ULL can be disregarded in accordance with the principles established in Autoclenz Ltd v Belcher [2011] UKSC 41, [2011] ICR 1157.

The Master of the Rolls and Bean LJ have set out most of the relevant passages from the judgment of Lord Clarke in Autoclenz, but I should add that in the final substantive paragraph, para. 38, he summarised his decision and reasoning as follows (p. 1171B):

I believe that the principles emerging from those passages can be stated as follows:

It is open to an employment tribunal to disregard any terms of a written agreement between an employer and an employee (but also, it is clear, a worker) which are inconsistent with the true agreement between the parties. Such an agreement may be described as a “sham”, but it does not have to be a sham in the particular sense defined by Diplock LJ in Snook v London and West Riding Investments Ltd [1967] 2 QB 786.

What the true agreement is may be gleaned from all the circumstances of the case, of which the written agreement is part but only a part.

In ascertaining whether the written agreement does in fact represent the true agreement the relative bargaining power of the parties will be a relevant consideration, because employers will typically be in a position to dictate the terms of the paperwork to which an employee must sign up, including terms that do not reflect the true agreement. Tribunals should accordingly take a realistic and worldly-wise approach to deciding whether that is the case.

It is an essential element in that ratio that the terms of the written agreement should be inconsistent with the true agreement as established by the tribunal from all the circumstances. There is nothing in the reasoning of the Supreme Court that gives a tribunal a free hand to disregard written contractual terms which are consistent with how the parties worked in practice but which it regards as unfairly disadvantageous (whether because they create a relationship that does not attract employment protection or otherwise) and which might not have been agreed if the parties had been in an equal bargaining position. (Footnote: ⁵ ) In that connection it is worth noting that the facts in Autoclenz were very stark. The written agreements provided (a) that the putative employer was under no obligation to provide work to the claimants, nor they to accept it, so that they were engaged on a casual basis shift-by-shift, and (b) that they were entitled to provide substitutes. The reality, however, was that it was understood on both sides that the claimants would be available to work, and would be offered work, on a full-time basis, and that they should provide their services personally. There was thus a plain inconsistency between the contractual paperwork and the parties’ mutual understanding as appeared from how they worked in practice; and the tribunal was thus entitled to draw the conclusion that it was the latter and not the former that represented the real terms of the agreement.

The question therefore for the ET in the present case was whether, in all the circumstances of the case and taking a worldly-wise approach, the reality of the relationships between Uber, driver and passengers was inconsistent with that apparently created by the Agreement (and the Rider Terms). That is a question of fact: although the precise question is different, the approach required by Carmichael v National Power plc [1999] ICR 1226 plainly applies here also – see per Lord Hoffmann at p. 1233C.

So far as taxi drivers are concerned, when they are plying for hire they necessarily contract directly with the passengers who pick them up at a rank or flag them down: there is no intermediary. Passengers are very familiar with the idea that taxi drivers are in business on their own account and themselves either own or rent the cabs which they drive.

In addition to taxis plying for hire on the street, there have for many years been intermediary radio-cab services operating in London (and in other cities) for members of the public wanting to book a taxi. In their original form passengers phoned the service, which would then allocate bookings to drivers over the radio. We were not addressed about the legal analysis of such arrangements (Footnote: ⁶ ) . However, I do not believe that a member of the public would have found anything surprising in the proposition that the radio service acted as an intermediary only and that as regards the ride itself they were dealing directly with the driver, just as they would have had they hailed him on the street or picked him up at a rank.

The correct analysis of the contractual arrangements where taxi drivers (though not black cab drivers) operate under the aegis of a named operator was considered by the EAT in Khan v Checkers Cars Ltd [2005] UKEAT 0208/05/1612. The British Airports Authority gave Checkers Cars (“Checkers”) an exclusive licence to provide a taxi service at Gatwick airport. It had a fleet of over 200 drivers, of whom the claimant was one, who plied for hire at the airport taxi-rank. Checkers took a commission and imposed numerous conditions on its drivers, including requiring them to charge set fares, use fixed routes and wear a uniform. Drivers were entirely free as to whether and when they chose to work but they were not permitted to drive for anyone else. The issue was whether the claimant was an employee of Checkers within the terms of section 230 (3) and so could bring a claim of unfair dismissal. The EAT held that he was not because there was no mutuality of obligation between jobs. But Langstaff J expressed the view, obiter, at para. 32, that “the contract went no further than to amount to a licence by Checkers to permit the Claimant to offer himself as a private hire taxi driver to individual passengers on terms dictated by the administrative convenience of Checkers and BAA”, drawing an analogy with Cheng Yuen and Mingeley (as to the latter, see paras. 127-9 below).

The position of taxi drivers is different from that of Uber drivers, who do not ply for hire. A closer analogy is with minicab drivers, whose services have to be pre-booked. Traditional minicab operations have no doubt suffered some impact from the rise of Uber, at least in the largest cities, but they remain widespread and familiar. It is clear from the case law that a common structure for such operations is, or was, as follows:

The operator advertises minicab services to the public under its own name, typically in directories or online and by distribution of flyers and business cards.

The operator does not have a fleet of vehicles owned by it, or drivers employed by it, but instead has relationships with a number of individual drivers who own their own vehicles and have the appropriate private hire licences and insurance.

Customers obtain the services of a driver by phoning the operator, who contacts the nearest available driver by radio or telephone and offers them the job and, if they accept, gives them details of the passenger. (Latterly this element may have been to a greater or lesser extent computerised, so that customers can make bookings with the operator online and/or the operator may use software to allocate jobs efficiently.) Drivers are free whether to make themselves available to work and whether to accept particular jobs.

Fares are set by the operator (possibly, but not necessarily, in accordance with a regulatory requirement), who may also impose other conditions such as the use of uniforms, quality of vehicles to be used etc.

As regards payment, the procedure differs between cash and account customers. Cash customers pay the driver themselves at the end of the journey, whether by cash or card. In the case of account customers the driver notifies the amount of the fare to the operator, who debits the account accordingly and pays the driver within a specified period.

The operator either charges the driver a set fee or takes a commission.

The legal analysis of the operator-driver-passenger relationships in that model was considered by this Court in Mingeley v Pennock [2004] EWCA Civ 328, [2004] ICR 727. The facts incorporated essentially all the above features (save that there was no express finding about account customers). In particular, as appears from para. 3 of Maurice Kay LJ’s judgment (pp. 729-730):

the driver owned his own car and had his own licence from the Council;

he paid a flat weekly fee for access to what is described by Maurice Kay LJ as

he was under no obligation to work or even to notify Amber Cars of his ability to work;

there was a fixed scale of charges;

the driver was obliged to wear a uniform; and

Amber Cars had a procedure for dealing with complaints from passengers about the conduct of its drivers.

The claimant driver, who alleged racial discrimination, was found not to be “employed by” the operator, within the meaning of the Race Relations Act 1976. The Master of the Rolls and Bean LJ say at para. 55 of their judgment that the case “was not about ‘worker’ status”. But the definition of “employment” in the 1976 Act extends beyond employment under a contract of service to “employment under … a contract personally to execute any work or labour” (see section 78 (1)). That is substantially the same as the definition in the Equality Act 2010 (see section 83 (2) (a)), which has in turn been held to be to substantially the same effect as the more elaborate definition of “worker” in the legislation with which we are concerned: see Secretary of State for Justice v Windle [2016] EWCA Civ 459, [2016] ICR 721, at paras. 7-10 (pp. 723-5), discussing the judgment of Lady Hale in Bates van Winkelhof v Clyde & Co. [2014] UKSC 32, [2014] 1 WLR 2047.

Buxton LJ, at para. 23 of his concurring judgment (p. 735 C-D), explicitly rejected the argument of the driver’s counsel that he was an employee “because [he] had obligations to the passengers to whom he might be directed by [the operator] to execute work in respect of them [i.e. by driving them]”. Counsel had described those obligations as “a collateral contract”, and Buxton LJ turned that description against him, pointing out

The question whether minicab drivers contract directly with their passengers has also been considered in a series of first instance decisions of the High Court and the VAT and First-tier Tribunals in the context of whether the services supplied by private hire operators are subject to VAT. That depends on whether the services in question are in law provided by the drivers as principals, with the operator acting as a booking agent – what was referred to before us as “the intermediary model” – or by the operator. HMRC in its published guidance recognises that that question may have to be answered separately as regards cash and account customers. In the cases to which we were referred there has been no dispute that the services provided to cash customers were provided by the drivers as principals: the issue has been about the services to account customers. We were referred in counsel’s skeleton arguments to Carless v Customs and Excise Commissioners [1993] STC 632 (Hutchison J), Hussain v Customs and Excise Commissioners, VAT Tribunal case no. 19194 (1999), Argyle Park Taxis Ltd v

The legal position as appears from those authorities is summarised in HMRC’s VAT Notice 700/25 – How VAT Applies to Taxis and Private Hire Cars – as follows (para.

relay bookings to the drivers (usually on a rota basis) for an agreed fee;

collect fares on their behalf from account customers.

There are obviously differences between the arrangements under consideration in the minicab cases and Uber’s platform-based system. For one thing, the number of drivers whose services are potentially available through the Uber app, at least in London, is incomparably larger than in any minicab fleet and probably also much larger than the number of black cabs belonging to any one of the old radio taxi services. For another, the technology is much more sophisticated (though some aspects of it, such as use of web-based mapping services to plot a route, are not unique to Uber). But it does not necessarily follow that the essentials are different for the purpose of a legal analysis. Subject to some technological differences, all of the features enumerated in para. 126 above are present in Uber’s model. In oral argument, in answer to a question from the Master of the Rolls, Mr Galbraith-Marten accepted that some minicab businesses run on the intermediary model. He was asked to identify any differences in Uber’s arrangements which he said required a fundamentally different approach from that taken in such cases. The only difference to which he referred was that of scale, which he said was “not decisive but relevant”. I thus understood him to accept, as Mr Linden certainly did, that in principle Uber could operate on the intermediary model, though it was of course their case that on the Tribunal’s findings it did not do so. Indeed that was the view of the Tribunal itself: see para. 97 of the Reasons.

I should emphasise that I am not concerned to establish that the taxi and minicab cases reviewed above are on all fours with the present case or indeed necessarily that they are on their particular facts correctly decided. Rather, the significance of this body of law is that it demonstrates that one well-recognised means of operating a private hire business is for the operator to act as a booking agent for a group of self-employed drivers who contract with the passengers as principals. It is not decisive whether all or most passengers understand this to be the case, but I certainly do not think that they would regard it as outlandish. I have already observed that it is commonly understood that black cab drivers plying for hire are in business on their own account, and it is not a big step for passengers to appreciate that the same may be true of minicab drivers even if they are, of necessity, booked through an intermediary.

Very recently the EAT had to consider the worker status of private hire drivers in Addison Lee Ltd v Lange [2018] UKEAT 0037/18 (see paras. 59-65 of the judgment of the Master of the Rolls and Bean LJ). Nothing in that case casts doubt on what I have said in the previous paragraphs. It is not clear to what extent the issue in the present appeal – that is, whether the driver was providing services to Addison Lee rather than the passenger – arose at all, and it is certainly not directly addressed in the reasoning of the EAT. (I would add, though this is not the main point, that the arrangements between Addison Lee and the drivers were substantially different in any event from those in the present case.)

The ET’s reasoning on the primary issue is at paras. 87-97 of the Reasons. These are summarised by the Master of the Rolls and Bean LJ, at para. 96 of their judgment, but my comments on them require me to set them out in full, which I do in the annex to this judgment. The fact that I shall have to be critical of aspects of the Tribunal’s reasoning does not detract from the admiration that I feel for the thoughtfulness with which it undertook its task and the clarity with which it expressed itself.

As the Tribunal acknowledges, its eleven numbered points involve a degree of repetition, and some of them also in my view cover more than one point. That being so, I do not think it would be helpful to go through them one-by-one. The main thrust of the reasoning, through all its various iterations, is (a) that it is not realistic to treat Uber drivers as entering into a direct contractual relationship with their passengers, with ULL acting merely as the agent or broker (see in particular para. 91); and (b) – which is the corollary – that realistically the drivers contract with ULL to provide their services to it (see in particular para. 92). I have done my best to group thematically the various points made in support of that conclusion.

I start with a group of points which do not address the actual features of the relationships but appear to be intended to provide a context against which they should be considered. These are:

“The lady doth protest too much”. At para. 87 the Tribunal says that the very fact that Uber goes to such trouble to specify in its contractual paperwork the nature of the relationships created is cause for scepticism about whether the picture there painted is accurate. I do not accept that. There is nothing suspicious as such about Uber wanting to have full and careful paperwork setting out the terms of the relationships into which it enters: any prudent business of any size, would, or at least should, do the same. It would of course be different if the paperwork does not reflect what the parties otherwise understood or agreed; but that begs the very question that has to be answered in this case.

Idiosyncratic language. Also at para. 87 the Tribunal refers to the Agreement as resorting to “fictions, twisted language and … brand new terminology”. It gives examples in its footnotes, which do indeed show some egregiously ugly pieces of corporate-speak, tendentious definitions and lawyerisms. But, again, the question is whether these various offences against good English actually conceal a different reality.

“Transportation services”. The Tribunal attaches importance to the fact that Uber has from time to time described itself as providing “transportation services”: see paras. 88 and 93. I do not see that this has much significance, since it all depends what you mean by that term. In one sense Uber obviously provides transportation services. But the question is whether it does so by providing the services of the drivers itself or by providing a service for booking (and paying for) them. The same applies to the Tribunal’s reliance (para. 88) on the fact that Uber markets its “product range” in its own name: the question is what the products in question consist of. The fact that the service is branded “Uber” does not seem to me determinative: Checkers Cars and Amber Cars (see paras. 125 and 127 above) likewise advertised themselves in their own names, but that did

Uber’s references to its drivers. At para. 88 the Tribunal refers to Uber having acknowledged that it employs drivers for the purpose of the transportation services which it supplies. That is cross-referenced to paras. 67-69 (set out by my Lords at para. 94 of their judgment), but the examples there given are the use in marketing material of such phrases as “Uber drivers” and “our drivers” and attaching the label “Uber” to the ride as well as the booking. All of these are thoroughly equivocal: they could mean a driver or a ride provided through Uber just as much as a driver employed by Uber or a ride provided by it as principal.

“30,000 separate businesses”. The Tribunal says at para. 90 that it is “faintly ridiculous” to say that “Uber in London is a mosaic of 30,000 small businesses linked by a common ‘platform’”. I agree that in some contexts – though not all – it might seem rather unnatural to describe a driver with his own car and a private hire licence who gets all or most of his work through Uber as carrying on a “business”; and Uber’s references to drivers “growing” such businesses are unconvincing. But this seems to me to be another example of focusing on a label rather than on the underlying question. I see nothing inherently ridiculous in the notion that Uber provides access to 30,000 drivers who will offer their services as principals. The same, subject only to numbers, could be said of the old “radio taxi” services or, depending on its particular arrangements, a minicab service following the model summarised at para. 126 above.

The Tribunal’s points so far considered are, as I read it, by way of a preliminary barrage. Its consideration of the actual features of the relationships between Uber, the drivers and the passengers appears principally in paras. 91-92 of the Reasons, which are, as I have said, two sides of the same coin (though parts of para. 90 may be relevant also). The points there made can be grouped as follows:

Limited information available to the driver at the point of acceptance. The Tribunal found it “absurd” to believe that the driver enters into a contract with a person whose identity he does not know and who does not know his to drive him or her to a destination unknown at the time that he accepts the job: see paras. 91 and 92 (3). As to the passenger and driver not knowing each other’s names, I cannot see that this is inconsistent with the existence of a contract between them: that is the case not only whenever a passenger flags a taxi in the street but also whenever he or she books a minicab operating on the model described above where the driver is the principal. (In fact the passenger at least is not entirely in the dark, since Uber supplies the driver’s first name, and he or she will be able to ascertain his identity if necessary because he will have to display his private hire licence.) As for the driver not knowing the destination in advance, this is the case whether he contracts with the passenger or with Uber, and I do not see how it is relevant to that question. But I do not in any event see what the supposed absurdity consists in: the driver is in business to drive passengers where they want to go (Footnote: ⁹ ) , and it is not likely to be of importance to him (at least for any legitimate reason (Footnote: ¹⁰ ) ) to know the destination at the point of acceptance.

Driver’s lack of control over key terms. The Tribunal, again, found it absurd to treat the driver as entering into a contract with the passenger of which the key terms – specifically route and fare – are set by a non-party (i.e. Uber): see paras. 91 and 92 (5) and (6). As to the route, the Tribunal found at para. 54 that the driver was not required by Uber to follow the route shown on the App, but that if there was a departure from it and the passenger subsequently asked for a refund because the most efficient route was not followed the driver would have to justify the departure. I do not think that it is accurate to describe that, as the Tribunal does in this paragraph, as a finding that Uber “prescribes” the route; but, whether it is accurate or not, I cannot see that it is inconsistent with the passenger and driver contracting directly. Whenever a passenger hires a cab or minicab it must be an implicit term that the driver will make a reasonable judgement of the best route; the fact that on an Uber hire that judgement is normally, in effect, delegated to the App cannot make a fundamental difference. (Indeed, as already noted, it is increasingly usual for any private hire driver to employ satnav or similar apps.) As for the fare, though that is indeed set by the Uber software, with no opportunity for negotiation by the driver, I cannot see why that is inconsistent with the existence of a contract between driver and passenger. As set out above, it is very common for minicab operators to prescribe set fares, but the drivers may nonetheless contract as principals.

Payment arrangements. Another feature which the Tribunal believed rendered it “absurd” to treat the driver as entering into a contract with the passenger is that his or her payment is made to Uber: see para. 91. But it is not at all unusual for minicab operators (and booking services for taxis) to collect payment on behalf of their drivers: that will routinely happen in the case of account customers. It does not follow that the driver is not contracting with the passenger as a principal: the debt is owed to him, even though the passenger pays it through a third party. Again, that is apparent from the case law to which I refer above.

Invoice. The ET attaches importance to the fact that the payment mechanism generates an invoice from the driver to the passenger and says that this is clearly

Quality control. The Tribunal notes at para. 92 (7) that Uber “imposes numerous conditions on drivers”. The only specific example which it gives is the list of acceptable vehicles, but no doubt it had in mind earlier findings about what drivers are told about how to behave towards passengers. But this does not seem to me inconsistent with the existence of a contract between driver and passenger. Even if Uber acts only as an intermediary it plainly has an interest in maintaining the quality of the product from which it makes its profit. The same goes for the maintenance of the ratings and performance management system referred to at para. 92 (8) and more fully explained at paras. 55-56. Similar measures to ensure quality – including some more intrusive ones such as the requirement to wear uniform – are found in the taxi and minicab cases referred to above.

Recruitment. The Tribunal found at paras. 40-41 of its Reasons that would-be Uber drivers had to attend personally at its office to present the required documentation (Public Carriage Office licence, PHV licence, proof of insurance etc) and that they would be “assessed” in the very limited sense that if it was apparent that they could not speak English they would be excluded and that if they exhibited signs of mental illness they would be referred to TfL. At para. 91

I take separately a point made only briefly in para. 92 – see point (4) – but about which there was a fair amount of argument before us. At paras. 52-53 of its Reasons the ET finds that Uber drivers are liable to be logged off the system for ten minutes (more recently reduced to two) if they decline three offers in a row or too often cancel trips once accepted. At least the former practice is directly authorised by the Agreement. Para. 2.6.2 of the New Terms concludes:

Finally, the ET in para. 91 makes three points about what it regards as absurd consequences of Uber’s argument that the drivers provide their services for, and under a contract with, the passengers. These are:

It is said to be absurd that if Uber became insolvent and failed to pass on the payment the passenger should be liable to the driver. But if Uber – whether for insolvency or any other reason – failed to account to the driver for the fare paid in relation to a particular ride, the driver would have no claim against the passenger, since he or she would have made payment by the agreed mechanism (i.e. by authorising a debit to his or her card at the conclusion of the ride).

The Tribunal suggests that if the contract were between the passenger and the driver the passenger might have the obligations of an employer under the legislation protecting workers – e.g. to pay the national minimum wage. I agree with the Master of Rolls and Bean LJ that that is, with all respect to the Tribunal, obviously wrong: quite apart from anything else, the passenger is plainly a customer of the driver’s business so that the words of exception in section 230 (3) (b) would apply.

The Tribunal says that the parties cannot have contemplated that the driver, rather than Uber, would bear the risk of non-payment by the passenger as a result of a some failure in the card collection systems or of unauthorised use of the card by the passenger (fraud); and that that is illustrated by the fact that Uber in fact has a policy that it will pay the driver at least in cases of fraud. This too seems to me neutral. Even on Uber’s analysis it is its obligation to collect the fares and there is nothing surprising in it bearing the risk of a failure – innocent or dishonest – in the collection process.

The ET does not, at least as I read it, make any explicit point in para. 91 – which describes the notion of a direct contract between driver and passenger as “fictitious” – about the perception of the customer about who he or she is contracting with; but it may be that such a point is implicit, and I think it should be addressed. Of course in the real world few if any passengers would consider the question at all: the transaction is a simple one, with very little opportunity for disputes to arise (Footnote: ¹² ) . Even if they were forced to confront the question, I do not think it can be assumed that they would all say that they thought they were contracting with Uber as principal. It is, I believe, widely understood that Uber drivers own the cars which they drive, and are their own masters as regards how much they drive. They do not wear any kind of uniform and the cars

At paras. 94-95 the ET turns to the case law relied on respectively by the Claimants and by Uber. I take the two paragraphs in turn.

At para. 94 the ET refers, albeit rather obliquely, to two EAT authorities – Cotswold Developments Construction Ltd v Williams [2005] UKEAT 0457/05, [2006] IRLR 181, and James v Redcats (Brands) Ltd [2007] UKEAT 0475/06, [2007] ICR 1006 – which were cited with approval by Lady Hale in Bates van Winkelhof. In the former Langstaff J encouraged tribunals to focus on “whether the purported worker actively markets his services as an independent person to the world in general … or whether he is recruited by the principal to work for that principal as an integral part of the principal's operations” (para. 52). In the latter Elias J refers to the distinction between “dependent work relationships” and “[contracts] between two independent business undertakings”. The Tribunal regards the drivers’ relationship with Uber as “dependent” and finds that their services are marketed to the public as “an integral part of [Uber’s] operations”. Those are plainly – to put it no higher – legitimate conclusions. But they are not decisive of, or indeed directly relevant to, the issue on this appeal, which is whether the putative worker is providing the relevant services for, and under a contract with, a third party, namely the direct beneficiary of the services. That was not an issue in either Cotswold Development or James v Redcats. In the former the claimant was a “selfemployed” carpenter engaged by a building company, and in the latter she was a courier making deliveries for a delivery company. In neither was it, nor could it sensibly have been, argued by the putative employer that the claimant provided his or her services for, or under a contract with, the end-recipient of the services. Rather, the issues were of the more usual kind referred to at para. 109 above, and also about whether there was any mutuality of obligation when the claimant was not working. (Likewise in Bates van Winkelhof itself there was no question of the claimant, who was a partner in a firm of solicitors, providing her service under a contract with anyone save the firm itself.) I do not accordingly believe that these cases advance the argument.

At para. 95 the ET addresses the authorities relied on by Uber as illustrating relationships where the putative employer is held to be no more than an intermediary between the putative worker and a third party for whom the services are performed. These include Mingeley and Khan but also Cheng Yuen and Quashie. I do not agree with the Master of the Rolls and Bean LJ (see para. 69 of their judgment) that these cases are of no assistance. They confirm that there can be cases in which, on a proper legal analysis, A provides services to B’s customers under contracts with the customers themselves notwithstanding that the services in question are integral to B’s business

The upshot of that, I fear laborious, review is as follows. The essential proposition which the reasoning in paras. 87-97 of the ET’s judgment is deployed to support is that it is unrealistic to treat Uber drivers as performing their services for, and under a contract with, their passengers rather than for, and under a contract with, ULL; and, that being so, that the contractual paperwork can be ignored on Autoclenz principles. For the reasons which I have given, I do not believe that any of the points made by the Tribunal supports that proposition. In particular, the various features relied on in paras. 91 and 92 are in my view entirely consistent with the position as stated in the Agreement.

I have reminded myself that even if none of the individual points relied on by the ET might be inconsistent with the position set out in the contract the cumulative effect could be. But, standing back so as to be able to see the wood as well as the trees, it still seems to me that the relationship argued for by Uber is neither unrealistic nor artificial. On the contrary, it is in accordance with a well-recognised model for relationships in the private hire car business.

That being so, Autoclenz gives no warrant for disregarding the terms of the Agreement. Autoclenz is an important tool in tribunals’ armoury because it enables them to look to the reality of a relationship rather than a false characterisation imposed by the employer. But the premise is that the characterisation is indeed false. As I have said, Autoclenz does not permit the re-writing of agreements only because they are disadvantageous. Protecting against abuses of inequality of bargaining power is the role of legislation: I return to this below.

The Master of the Rolls and Bean LJ endorse much, though not all, of the ET’s reasoning as reviewed above. I will not repeat all the points on which I have already expressed my view. However, they also attach importance to the regulatory regime under which Uber operates: see para. 89 of their judgment. For myself, I see no inconsistency between Uber’s position as the operator of its service within the meaning of the 1998 Act and it being obliged to operate a system under which it makes all bookings and has to provide fare estimates on request. As Ms Rose pointed out, it used to be a regulatory rule that all barristers must deal with solicitors through a clerk; but that did not mean that the clerk was the principal. A minicab service operating on the intermediary model described above would be subject to the same regulatory obligations, but that would not mean that its drivers performed their services as its agents. In my view the focus must be on the arrangements between the parties themselves: the fact that they may be in order to comply with regulatory requirements is in itself neutral.

I am conscious that I have not addressed the reasoning of the EAT. Since ultimately the question for us is whether there was any error of law in the decision of the ET, and

I should pick up two points which do not feature in the ET’s reasoning but did feature, at least to some extent, in the arguments before us.

First, we were referred by Mr Galbraith-Marten to three decisions of the CJEU on the meaning of “worker” – Allonby v Accrington & Rossendale College (C-256/01) [2004]

Secondly, Ms Rose placed considerable emphasis on Secret Hotels2, to which I have already referred and which the Master of the Rolls and Bean LJ address in detail at paras. 51-53 of their judgment. As there appears, that was a case concerning VAT arising out of an internet platform-based service under which hotel rooms could be booked online. The issue was whether the intermediary who operated the website, Med Hotels, sold the rooms as principal or on behalf of the hoteliers. The contractual terms stated that Med Hotels acted only as an agent, but the FTT and this Court accepted

Mr Galbraith-Marten submitted that Secret Hotels2 was of no assistance because it was not a decision in the employment context, and Autoclenz was not cited; and I understand my Lords to take the same view. With respect, I do not agree that this disposes of the relevance of the decision. If the ET is right it is not only the Agreement which mischaracterises the relevant relationships but also the Rider Terms which apply between the passenger and Uber, which are a consumer contract and not in the employment field at all. In any event, although Lord Neuberger did not refer to Autoclenz itself the line of authorities which he made it clear that he was following is the same as that on which Lord Clarke’s analysis in that case was based: see para. 32 of his judgment and para. 23 of the judgment of Lord Clarke in Autoclenz, both of which, for example, refer to the seminal landlord-and-tenant case of Street v Mountford [1985] AC 809. Inequality of bargaining power is central to the analysis in both cases and was expressly referred to by Lord Neuberger: see para. 40 of his judgment. I accordingly think that Ms Rose is entitled to rely on Secret Hotels2 as confirming that the operator of an internet platform which puts together suppliers of services and customers of those services can effectively stipulate that it is acting only as an agent even if it has its own strong customer-facing brand and exercises a high degree of control over aspects of the transaction between supplier and customer. But it takes her no further than that: whether the contractual terms reflect the reality of the relationships in any particular case must depend on the circumstances of that case.

My Lords also make the point that there was in Secret Hotels2 a written contract between the platform and the hotelier, whereas there was no such contract between ULL and the driver. For the reason given at para. 116 above, I do not believe that that is a material difference. Drivers do have a contract with UBV, which provides in terms that its local affiliates – in this case ULL – act on its behalf in respect of specified matters, including the collecting of fares.

If, contrary to my view, Uber drivers do contract with ULL to provide services for it, the next question is over what period such a contract is in place. The Claimants have always accepted, given that they are under no obligation to switch the App on, that there is no “umbrella contract” creating rights and obligations between periods of work. On any view, therefore, they are only workers on a gig-by-gig basis: the question is what constitutes the gig.

What the ET held, and is the Claimants’ primary case before us, is, as I have said, that there is a contract in place between them and ULL throughout the period that the driver has the App switched on, is in the territory in which he is licensed to use it, and is ready and willing to accept trips: see paras. 86 and 100 of the Reasons. It held in the alternative (para. 102) that such a contract arises when the driver actually accepts a trip. The difference between the two alternatives is thus the period during which the driver satisfies the ET’s three requirements but has not accepted a trip: I will refer to this as “availability time”.

Uber’s case is that it has no relevant contract with the driver at all; but its fallback position is, as I understand it, that there is only a contract in place when the driver is actually carrying a passenger: see Mr Reade’s submissions in the ET, summarised at para. 100 of the Reasons. That differs from the Tribunal’s alternative conclusion since it does not cover the period between the driver’s initial acceptance on the App and the definitive acceptance that occurs when the trip actually starts, during which the driver retains the possibility of cancelling.

It is of course essential to all three heads of claim that the Claimants should be workers during the period to which their claim relates. But there are also two closely related issues relating to the claims under the WTR and the NMWR. Specifically:

Does availability time constitute “working time” for the purpose of the WTR – namely (regulation 2) “any period during which he is working, at his employer’s disposal and carrying out his activity or duties” ? We were not taken to any of the authorities about the effect of that definition.

Does availability time fall to be taken into account in calculating whether the driver has received the national minimum wage ? We were not taken through the NMWR, which are extremely complex, but the central element in the relevant provisions is the time during which the worker is “working”. The Claimants’ case, which the ET accepted, is that the during availability time they were doing “unmeasured work” within the meaning of Chapter 4 of Part 5 of the Regulations.

The submissions before us did not address the practical impact of a finding that availability time counts as working time or that it counted for national minimum wage purposes. So far as the latter is concerned, the impact would depend on the relationship between availability time and time spent actually carrying passengers: if drivers spent too high of a proportion of their time “available” but not carrying passengers (either because work was not offered or because it was offered but not accepted) the average of their earnings over the whole period when they had the App switched on would be liable to fall below the prescribed minimum.

In my view, if drivers provide services to, and under a contract with, ULL at all it is only during the period when they have accepted a trip. It is common ground that drivers are not obliged to accept any particular trip when offered. The only basis on which the ET held that they are nevertheless under a contractual obligation to Uber while the App is switched on is that they are liable to be disconnected for a specified period if they reject trips, or cancel them, too often. But, as I say at para. 138 above, I do not believe that that implies a positive contractual obligation on the part of drivers to accept (and not cancel thereafter) a minimum number of trips offered. I would add that if there were such an obligation it would be necessary to specify what the minimum obligation was. The ET did not in its actual reasoning rely on any finding as to that. The EAT, however, relied on a document quoted by the ET in its findings which referred to drivers being obliged to accept 80% of trips offered: see para. 51 of its judgment (quoted by my Lords at para. 21) and para. 89 of the judgment of the EAT. Ms Rose objected that the recitation of that document did not amount to a finding and that in fact the oral evidence had been that it was not a figure applied by Uber in the UK. There may be some force in that objection, but I do not in any event regard the point as central. If, contrary to my view, the right to disconnect drivers who declined offers or cancelled too often reflected a positive obligation on their part to accept most trips it would not be impossible to find an appropriate formulation for that obligation by reference to a criterion of reasonableness and/or evidence about what happened in practice.

My view on this issue is reinforced, at least as regards entitlement under the WTR and NMWR, by the consideration that under the Agreement drivers are explicitly entitled during availability time to be available also for other driving work, and specifically for platforms providing a similar app-based service to Uber (see para. 24 of my Lords’ judgment). It is well-known that such alternative providers exist in the United States. The ET makes no findings about whether they currently operate in London, or, therefore, about whether drivers do in fact “multi-app” in this way. Ms Rose told the Court that there are such services, albeit not on the same scale as Uber, but Mr Linden told us that Mr Farrar’s evidence had been that there were none at the period to which the claims relate. Be that as it may, what matters is that the right exists and cannot be regarded as merely theoretical. There is no conceptual difficulty about a worker being in a contractual relationship with two employers during the same period; but I find it much more difficult to see how they could be said to be at the disposal of two employers, and carrying out duties for both, during the same period, or how the same period could be taken into account twice (or indeed more) for the purpose of calculating the national minimum wage obligations of different employers. The position would be still more extraordinary if drivers could bring into account time when they were actually driving on a trip obtained through a different platform: I take the ET’s point that in such a case they would not satisfy the third of its criteria, but compliance would be very difficult to police. I will not explore this further, however, not least because, as I have said, we were not addressed on the details of either set of Regulations.

Those difficulties only apply up to the point that the driver accepts a trip – that is, presses “accept” on the App and is given details of the pick-up. It seems to me clear that at that point the driver comes (if I am wrong on the main issue) under an obligation to ULL to carry its passenger. That is subject to the right of cancellation, but the ET found, as one would expect, that cancellation could only be for a good reason: see para. 21 of its Reasons (quoted by my Lords at para. 21). Likewise I see no difficulty in treating the driver thereafter as working exclusively for Uber, for the purpose of the WTR and the NMWR, until the end of the trip. I would not, therefore, have accepted Uber’s case, as advanced by Mr Reade in the ET, that any obligation only arose at the moment that the passenger was picked up.

The question whether those who provide personal services through internet platforms similar to that operated by Uber (Footnote: ¹³ ) should enjoy some or all of the rights and protections that come with worker status is a very live one at present. There is a widespread view that they should, because of the degree to which they are economically dependent on the platform provider. My conclusion that the Claimants are not workers does not depend on any rejection of that view. It is based simply on what I believe to be the correct construction of the legislation currently in force. If on that basis the scope of protection does not go far enough the right answer is to amend the legislation. Courts are anxious so far as possible to adapt the common law to changing conditions, but the tools at their disposal are limited, particularly when dealing with statutory definitions. I have already explained why I do not think that Autoclenz can be treated as a tool to re-write any disadvantageous contractual provision that results from the disparity of bargaining power between (putative) employer and (putative) worker: in cases of the present kind the problem is not that the written terms mis-state the true relationship but that the relationship created by them is one that the law does not protect. Abuse of superior bargaining power by the imposition of unreasonable contractual terms is of course a classic area for legislative intervention, and not only in the employment field.

A similar point is made by Sir Patrick Elias in his recent article in the Oxford Journal of Legal Studies, Changes and Challenges to the Contract of Employment, in the context of the analogous question of zero-hours contracts. He says, at p. 16:

Even if it were open to the Courts to seek to fashion a common law route to affording protection to Uber drivers and others in the same position, I would be cautious about going down that road. The whole question of whether and how to adapt existing employment law protections to the development of the so-called gig economy, and in particular to the use of service-provision platforms such as Uber, is under active review by the Government at present. The Taylor Review (Good Work – The Taylor Review of Modern Working Practices) was published last year. It recommended the introduction of a new “dependent contractor” status, broadly but not wholly covering the same ground as the definitions of “worker”; and it also made recommendations on the very question raised by the secondary issues in this appeal – that is, how to calculate working time in the case of workers who obtain work through app-based services. In February this year the Treasury, BEIS and HMRC opened a consultation on a wide range of issues raised by the Review. Chapter 8 of the consultation is particularly apposite in the context of this appeal. Para. 8.5 observes that:

We were, perfectly properly, not addressed about this wider context, and it forms no part of my dispositive reasoning. I refer to it only because the issue is one of wide public concern, and I believe that it is important to spell out the different roles of the Courts and of Parliament in this context.

Second, our scepticism is not diminished when we are reminded of the many things said and written in the name of Uber in unguarded moments, which reinforce the Claimants' simple case that the organisation runs a transportation business and employs the drivers to that end. We have given some examples in our primary findings above. We are not at all persuaded by Ms Bertram's ambitious attempts to dismiss these as mere sloppiness of language.

Third, it is, in our opinion, unreal to deny that Uber is in business as a supplier of transportation services. Simple common sense argues to the contrary. The observations under our first point above are repeated. Moreover, the Respondents' case here is, we think, incompatible with the agreed fact that Uber markets a ‘product range’. One might ask: Whose product range is it if not Uber's? The ‘products' speak for themselves: they are a variety of driving services. Mr Aslam does not offer such a range. Nor does Mr Farrar, or any other solo driver. The marketing self-evidently is not done for the benefit of any individual driver. Equally self-evidently, it is done to promote Uber's name and ‘sell’ its transportation services. In recent proceedings under the title of Douglas O'Connor-v-Uber Technologies Inc the North California District Court resoundingly rejected the company's assertion that it was a technology company and not in the business of providing transportation services. The judgment included this:

Fourth, it seems to us that the Respondents' general case and the written terms on which they rely do not correspond with the practical reality. The notion that Uber in London is a mosaic of

Fifth, the logic of Uber's case becomes all the more difficult as it is developed. Since it is essential to that case that there is no contract for the provision of transportation services between the driver and any Uber entity, the Partner Terms and the New

Sixth, we agree with Mr Linden that it is not real to regard Uber as working ‘for’ the drivers and that the only sensible interpretation is that the relationship is the other way around. Uber runs a transportation business. The drivers provide the skilled labour through which the organisation delivers its services and earns its profits. We base our assessment on the facts and analysis already set out and in particular on the following considerations.

The contradiction in the Rider Terms between the fact that ULL purports to be the drivers' agent and its assertion of “sole and absolute discretion” to accept or decline bookings.

The fact that Uber interviews and recruits drivers.

The fact that Uber controls the key information (in particular the passenger's surname, contact details and intended destination) and excludes the driver from it.

The fact that Uber requires drivers to accept trips and/or not to cancel trips, and enforces the requirement by logging off drivers who breach those requirements.

The fact that Uber sets the (default) route and the driver departs from it at his peril.

The fact that UBV fixes the fare and the driver cannot agree a higher sum with the passenger. (The supposed freedom to agree a lower fare is obviously nugatory.)

The fact that Uber imposes numerous conditions on drivers (such as the limited choice of acceptable vehicles), instructs drivers as to how to do their work and, in numerous ways, controls them in the performance of their duties.

The fact that Uber subjects drivers through the rating system to what amounts to a performance management/disciplinary procedure.

The fact that Uber determines issues about rebates, sometimes without even involving the driver whose remuneration is liable to be affected.

The guaranteed earnings schemes (albeit now discontinued).

The fact that Uber accepts the risk of loss which, if the drivers were genuinely in business on their own account, would fall upon them.

The fact that Uber handles complaints by passengers, including complaints about the driver.

The fact that Uber reserves the power to amend the drivers' terms unilaterally.

Seventh, turning to the detail of the statutory language, we are satisfied, having regard to all the circumstances and, in particular, the points assembled above, that the drivers fall full square within the terms of the 1996 Act, s 230(3)(b). It is not in dispute that they undertake to provide their work personally. For the reasons already stated, we are clear that they provide their work ‘for’ Uber. We are equally clear that they do so pursuant to a contractual relationship. If, as we have found, there is no contract with the passenger, the finding of a contractual link with Uber is inevitable. But we do not need to base our reasoning on a process of elimination. We are entirely satisfied that the drivers are recruited and retained by Uber to enable it to operate its transportation business. The essential bargain between driver and organisation is that, for reward, the driver makes himself available to, and does, carry Uber passengers to their destinations. Just as in Autoclenz, the employer is precluded from relying upon its carefully crafted documentation because, we find, it bears no relation to reality. And if there is a contract with Uber, it is self-evidently not a contract under which Uber is a client or customer of a business carried on by the driver. We have already explained why we regard that notion as absurd.

Eighth, while it cannot be substituted for the plain words of the statute, the guidance in the principal authorities favours our conclusion. In particular, for the reasons already given, it is plain to us that the agreement between the parties is to be located in the field of dependent work relationships; it is not a contract at arm's length between two independent business undertakings. Moreover, the drivers do not market themselves to the world in general; rather, they are recruited by Uber to work as integral components of its organisation.

Ninth, we do not accept that the authorities relied upon by Mr Reade support the conclusion for which he argues. We have four main reasons.

None of the authorities actually turned on the limb (b) test.

They were concerned wholly or very largely with whether there was an ‘umbrella’ contract between the claimants and the respondents, an issue with which we are not concerned at all. Only one addressed (and then only in a single sentence) the question at the heart of our case of whether, in performing individual services (here driving trips), a claimant is working ‘for’ the putative employer pursuant to a contract.

Two of the cases arise out of facts which have little in common with the matter before us. Cheng Yuen and Quashie concern arrangements by which individuals were permitted to render to the golf club members and nightclub ‘clients' services ancillary to the principal service or facility offered by the proprietors. But there is nothing ‘ancillary’ about the Claimants' work. It seems to us that there are added difficulties for the putative employer with a defence modelled on Cheng Yuen and Quashie where the claimants perform the very service which the respondent exists to provide. In such a case it is (as Uber appears to recognise) essential to the defence for the Tribunal to find not only that the claimants contract personally with those who receive the services in question but also that they collectively, rather than the respondent, ‘are’ the business. In a proper case the evidence warrants such findings but on a careful review of all the material placed before us, our conclusions on both propositions are, for the reasons already stated, entirely adverse to Uber.

Although the facts of Mingeley and Khan are closer to those of the instant case, there was ample room in both for the finding that the arrangements between the parties were consistent with the claimant personally entering into a contract with each service user. As we have explained, there is no room for that interpretation to be placed upon the dealings (such as they are) between the Uber driver and his passenger.

Tenth, it follows from all of the above that the terms on which Uber rely do not correspond with the reality of the relationship between the organisation and the drivers. Accordingly, the Tribunal is free to disregard them. As is often the case, the problem stems at least in part from the unequal bargaining positions of the contracting parties, a factor specifically adverted to in Autoclenz. Many Uber drivers (a substantial proportion of whom, we understand, do not speak English as their first language) will not be accustomed to reading and interpreting dense legal documents couched in impenetrable prose. This is, we think, an excellent illustration of the phenomenon of which Elias J warned in the Kalwak case of “armies of lawyers” contriving documents in their clients' interests which simply misrepresent the true rights and obligations on both sides.

Eleventh, none of our reasoning should be taken as doubting that the Respondents could have devised a business model not involving them employing drivers. We find only that the model which they chose fails to achieve that aim.”