‘The Simple Things You See Are All Complicated’: Thoughts on Deliveroo¹

1. INTRODUCTION

Is gigging in the digital economy the ultimate freewheeling, laidback working lifestyle or the worst kind of exploitative casualisation?

It would seem, from their many attempts to gain access to a better regulated regime, well documented in the case law, that many gig workers tend to the latter rather than the former view. There have been important advances in the common law: in Pimlico Plumbers v Smith² the Supreme Court held that a worker’s right under the contract to use a substitute to do their work was not inconsistent with an obligation to provide personal service where the employer imposed stringent limitations on who the substitute could be. Even more significantly, in Uber v Aslam³ the Supreme Court held that the question of whether someone was a ‘worker’ for the purpose of the statutory definition was primarily a matter of statutory interpretation rather than contractual interpretation and that the statute should be construed purposively, bearing in mind that its purpose was to protect vulnerable workers. But it transpires that celebrations of this more enlightened approach were premature. In November 2023, the Supreme Court held that Deliveroo delivery riders are not ‘workers’ in an employment relationship with the company, which means that they can only be regarded as self-employed.⁴ Thus, their union, the Independent Workers Union of Great Britain (IWUGB) was not able to invoke the statutory recognition procedure on their behalf, and it was held that this did not breach their right to freedom of association under Article 11 of the European Convention on Human Rights (ECHR).

This piece is mainly concerned with the employment status aspects of that decision. It argues for a reappraisal of the indicia of employee and worker status on the basis that traditional approaches have failed to keep pace with evolving business models.

2. HISTORY

Eight years ago, in November 2016, the IWUGB sought recognition from Deliveroo for the purposes of collective bargaining in respect of Deliveroo riders working in the Camden and Kentish Town districts of London. When the company rejected the approach, the union applied to the CAC to set in motion the statutory recognition procedure under TULRCA Schedule A1. However, a union can only apply for recognition in respect of groups of workers, and the term ‘worker’ is defined in TULRCA section 296 as someone with (1) a contract of employment, or (2) ‘any other contract … to do or perform personally any work or services for another party who is not a professional client of his’. The scope of the so-called ‘limb (b)’ definition, to be found in similar terms in the Employment Rights Act 1996 section 230(3), the Equality Act 2010 section 83(2), the Working Time Regulations 1998 regulation 2 and the National Minimum Wage Act 1998 section 54, was the issue in this case, as in so many others involving people working in the gig economy. Employment in such industries is characterised by a high degree of autonomy for these workers in relation to when and for how long they do the work coupled with a high degree of insecurity as to the availability of work and generally low wages. Work frequently, but not invariably, becomes available via an online platform, hence the EU prefers to use the term platform workers in its attempt to regulate for minimum standards.⁵

Contracts for gig workers are invariably in a standard form drawn up by the employing company with no opportunity for variation through negotiation, and the Deliveroo contract was no exception. It described itself as a Supplier Agreement, according to which riders would render services as a ‘supplier in business on your own account’,⁶ and contained a clause negativing mutual obligation, ie, the company had no obligation to offer work, and the riders could choose whether or not to log on to the app to make themselves available to work without restriction. Even when logged on, they were under no obligation to accept any job that was offered. They were paid a fee per delivery (usually £3.75 per job), with the prices set by Deliveroo.

So far, the Deliveroo standard form contract is substantially similar to other platform work contracts. The major difference, and the one on which the union’s case foundered at every level in the legal process, was that Deliveroo gave the riders carte blanche to use any substitute they wanted, at any time, to do the work in their place. Actually, the carte was slightly grise, since although they did not need to seek permission in advance, there were a few limitations, namely: (1) that they should not use anyone who had had an agreement with Deliveroo terminated for breach; (2) that they should inform Deliveroo if the substitute used a different kind of vehicle from the contracted rider and (3) that the originally contracted rider remained responsible in every way for the substitute’s performance. Interestingly, there is no mention of age, although a BBC investigation recently alleged that there was widespread evidence of under-18s acting as riders.⁷

Although the substitution clause was in a sense completely unnecessary, since a rider unwilling to act personally needed only to log off the app, the CAC found that it was not a sham clause⁸ because there was evidence of riders actually using substitutes. This was ‘the central and insuperable difficulty for the union’ as the CAC commented.⁹ While a limited right of substitution has been held by the Supreme Court not to be inconsistent with a contract to render personal service,¹⁰ the virtually unfettered and unpoliced right to substitute in the Deliveroo supplier agreement was found at every level to be inconsistent with an obligation to render personal service. The CAC therefore held that the delivery riders were not ‘workers’ within the meaning of TULRCA section 296 and therefore the union could not invoke the recognition procedure under TULRCA Schedule A1 in respect of them. Furthermore, the CAC considered that ECHR Article 11, guaranteeing freedom of association, including the right to form and join trade unions, made no difference to this. The union sought judicial review of the CAC decision on several grounds but was allowed to proceed only on the Article 11 issue.¹¹

3. INTERNATIONAL NORMS

The appeal to the Supreme Court, therefore, was purely on whether the Deliveroo riders’ rights under ECHR Article 11 had been infringed by not allowing their union to seek recognition and collective bargaining rights on their behalf. Since the European Court of Human Rights’ (ECtHR) seminal decision in Demir v Turkey,¹² the right to form and join trade unions entails in addition the right to engage in collective bargaining. The argument on behalf of the union, therefore, was that excluding these workers from the opportunity to invoke the statutory procedure which might compel the company to recognise their union for the purposes of collective bargaining was an infringement of their rights under ECHR Article 11. Furthermore, the union argued that Article 11 rights were not confined to people who counted as workers under domestic law.

Four important points emerged from the Supreme Court’s review of the jurisprudence of the ECtHR, in particular the Court’s decision in The Good Shepherd.¹³ First, the right to form and join trade unions, as a sub-set of the right to freedom of association, is confined to people who are in an employment relationship with an employer. Secondly, the concept of an employment relationship for the purpose of ECHR Article 11 is autonomous, specific to the Convention, and applies to all its member states regardless of domestic law. Thirdly, regard should be had to the criteria identified by ILO Recommendation 198 on the Employment Relationship (2006) in deciding whether there is an employment relationship. Fourthly, it is important for national authorities to be alert to the possibilities of abuse of legal forms to disguise the reality of an employment relationship and to ensure that workers’ protection is not undermined by such manipulation.

The reference to ILO Recommendation 198 is a positive step since its mainspring is the need to ensure that workers are not deprived of their employment rights. As its Preamble states: ‘situations exist where contractual arrangements can have the effect of depriving workers of the protection they are due’. Paragraph 13 of the Recommendation suggests that member states should define specific indicators of an employment relationship and goes on to suggest 14 relevant factors. The list contains no surprises: control, integration and personal service are all there, along with provision of tools and equipment by the employer and absence of financial risk to the worker. It may be surprising that a Recommendation drawn up in 2006 should take little account of newer patterns of working: in its terms, it could as well have been drafted a generation earlier. However, it should be remembered that strictly, the instrument is a recommendation to member states, and the roster of indicators in paragraph 13 is prefaced by the words, ‘The list of factors might include’—thus the enthusiastic adoption of these indicators as a test by the ECtHR, and in its turn, the Supreme Court, is in that sense problematic.

Nonetheless, the emphasis on focusing on the substance of the relationship rather than just its legal form is welcome. Specifically, paragraph 9 of the Recommendation states:

For the purposes of the national policy of protection for workers in an employment relationship, the determination of the existence of such a relationship should be guided primarily by the facts relating to the performance of work and the remuneration of the worker, notwithstanding how the relationship is characterised in any contrary arrangement, contractual or otherwise, that may have been agreed between the parties.

As the Supreme Court’s decision notes, this is also mandated by domestic law¹⁴ and EU law.¹⁵

Since the ECtHR expressly endorsed the relevance of the ILO criteria in The Good Shepherd, they must be considered as part of the test of an employment relationship for the purposes of ECHR Article 11, as the Supreme Court recognised.¹⁶ However, like the Court of Appeal, the Supreme Court adopted an unduly broad brush approach in concluding that the multifactorial approach advocated by Recommendation 198 ‘broadly parallels’¹⁷ the approach of domestic law. Having stated that ‘The correct approach requires the application of a multifactorial test, focusing on the practicalities of the relationship and how it operates in reality’,¹⁸ the Supreme Court actually goes on to give absolute primacy to the personal service requirement which it holds to be defeated by contractual provision for a nearly limitless substitution right rather than weighing it as one factor among the many suggested in Recommendation 198. While paying lip service to the Recommendation, it actually ignores the injunction in paragraph 9 that they should be ‘guided primarily by the facts relating to the performance of work …’.

It is true that domestic legislation focuses on personal service in defining both who is an employee as well as who is a limb (b) worker, but in considering whether an employment relationship existed for the purposes of ECHR Article 11 there was scope for a different approach to be taken. Before considering what that approach might look like, let us consider the personal service requirement a little further, anticipating the objection that personal service has to be the crucial distinguishing feature of employment relationship.

The main reason for arguing that this is not the case is that although personal service was famously stated to be one of three conditions of a contract of employment by MacKenna J in Ready Mixed Concrete (South East) Ltd v MPNI,¹⁹ even he considered that a limited or occasional power of delegation would not be inconsistent with such a contract.²⁰ A right to substitute has been held to be consistent with worker status where it applied only if the worker was unable (as opposed to unwilling) to work,²¹ and where the worker could substitute another worker who also worked for the same employer.²² As a matter of logic, once it is admitted that some delegation or substitution is not inconsistent with a limb (b) worker contract (or, for that matter, a contract of employment), it should follow that the possibility of substitution of any kind cannot be a complete answer to whether or not such a contract exists. It is not a prior defining characteristic. Furthermore, the Supreme Court’s decision focuses on the contractual provision for substitution rather than looking closely at the practicalities or reality of the situation. There seems to be no reason in principle why it should not be possible for a court or tribunal to look at how substitution actually works in practice, and on that basis to decide whether it is extensive enough to tip the balance (alongside other indicators) against finding an employment relationship. An analogy might be drawn with the approach of the House of Lords in construing ‘temporary cessation of work’,²³ holding that the decision as to whether or not an absence from work was temporary should be judged with hindsight, looking back over the whole duration of the relationship.²⁴

It may also be noted that it is an equivocal test in that it is perfectly possible to envisage situations where a worker is clearly self-employed, yet his or her personal service is regarded as essential to the contract. Commissioning a portrait painter is an obvious example, but it might well extend also to a trusted plumber or electrician.

4. A BUSINESS REALITY APPROACH

A drawback of the ILO Recommendation 198 approach is that it focuses principally on the worker in the relationship. This is apparent also in the EU’s Platform Work Directive.²⁵ This can lead to relevant facts being either completely overlooked or at least substantially under-appreciated. What is needed is a wider holistic approach that takes account of the dimension currently missing: the nature of the business, the business model and how the business represents itself through its branding and marketing.

This is not a totally new idea: in Montreal v Montreal Locomotive Works Ltd²⁶ Lord Wright, in the Privy Council, said ‘…it is in some cases possible to decide the issue by raising as the crucial question whose business is it’, and in USA v Silk²⁷ the American Supreme Court proposed an ‘economic reality’ test,²⁸ but that broad analytical approach has got somewhat lost in recent years. In addition, the old business reality test was looking at the issue mainly from the point of view of deciding who owns or profits from the business, rather than looking at those characteristics of the business referred to above.

Deliveroo describes itself as

…a hyperlocal three-sided marketplace, connecting local consumers, restaurants and grocers, and riders to fulfill [sic] a mission critical, emotional purchase in under 30 minutes. By offering fast and reliable delivery which consumers can track online, Deliveroo has grown rapidly.²⁹

It is an online platform-based company that makes food deliveries, with some 50,000 people making those deliveries, and it obviously could not exist without those people. Indeed, the CAC quoted from Deliveroo’s own recruitment literature describing the riders as: ‘The very lifeblood of our company. Without them Deliveroo wouldn’t exist—a fact at the very heart of how we operate as a business’.³⁰ It is quite unlike a company seeking customers for its goods via a web platform or any enterprise that needs a delivery service for its goods. Its business is delivering food and the business model cannot work without riders making those deliveries. But yet it employs none of these people? How realistic is that?

In terms of branding and marketing, the brand is clearly Deliveroo and the service marketed—speedy food delivery—stands or falls according to the Deliveroo reputation. Customers order via the Deliveroo website and expect a Deliveroo delivery. They would undoubtedly be very surprised to hear that they were dealing with a third-party entrepreneur.

Finally, consider the nature of the contract between the parties. It is invariably (and not just in this case) a standard form contract drawn up by the employing party with no opportunity for negotiation or variation by the party who does the work. That again should be recognised as a clear indicator of a hierarchical relationship in which the worker is fairly obviously subordinate to the employer.

Surely any tribunal or court asking itself the question—whose business is this?—could only come up with one answer: that the business is Deliveroo’s and the riders are working for that business.

5. THE PURPOSIVE APPROACH

After the CAC made its initial decision in 2016, and after the union’s appeal was allowed on the ECHR Article 11 point only in 2018, a differently constituted Supreme Court delivered its groundbreaking decision in Uber BV v Aslam³¹ emphasising the importance of a purposive approach to interpreting legislation designed to protect vulnerable workers. That case also concerned the statutory definition of a worker, principally in relation to working time and the national minimum wage. In fact, the Supreme Court’s decision was handed down after the Court of Appeal hearing in the Deliveroo case, but before judgment had been given, which meant that the court at least had written submissions from counsel as to its impact.

The timing was unlucky: it was too late to re-open the domestic law issue, and the Court of Appeal held that it was not relevant to the ECHR Article 11 point because in Uber there were no issues about personal service nor about Article 11.³² The court was not persuaded that the purposive approach adopted in Uber was relevant in considering the ECtHR’s approach to Article 11.³³ The Supreme Court (none of whom, incidentally, sat on Uber v Aslam) was equally dismissive.³⁴

Had the timing been better, and the purposive approach used when this case first arose, it is conceivable that the result might have been different. Conceivable, but by no means definite, given that the very definition of ‘worker’ in the legislation refers to personal service and given the breadth of the substitution clause devised by Deliveroo.

6. CONSEQUENCES

In May 2022, while the IWUGB’s claim was making its way through the court system, Deliveroo concluded a ‘Voluntary Partnership Agreement’ with the GMB union. Recognition of another union, regardless of whether it represents most of their workers, is, of course, a classic strategy for a company to use to avoid having to recognise a union it dislikes, and is a loophole permitted under the statutory recognition procedure.³⁵ RIP, Bridlington and union solidarity.³⁶ In the event, the precaution proved to be unnecessary. Thus, Deliveroo remains free to impose its terms on anyone prepared to accept them.

Consider the following (imaginary) scenarios:

1. Deliveroo riders Anne and Ben are identified as the leaders of the movement to get riders to join the IWUGB. The company therefore terminates its Supplier Agreements with them, stating expressly that this is the reason for their termination.

2. Deliveroo writes to all the other riders in Camden and Kentish Town, warning them that they will get the same treatment if they encourage union activity.

3. Deliveroo decides to discriminate in future against trade union members when offering Supplier Agreements.

If the riders are not ‘workers’ then it would appear that Deliveroo would have no liability in these situations. Such results would seem extraordinary, and there must be some question as to whether the ECtHR would hold that there was no breach of ECHR Article 11 in these cases. But for the time being, this is apparently the position of English law.

Consider one more scenario:

1. Fed up with the situation, Deliveroo riders form themselves into the Deliveroo Riders Action Association (DRAA). They decide to all make themselves unavailable at pre-arranged times, for example, Friday and Saturday evenings. Storms erupt on social media and Deliveroo loses substantial numbers of customers to other delivery services. This scenario is not imaginary: a grouping calling itself Delivery Job UK, organised through social media, did indeed call a strike on Valentine’s Day 2024 in which some 3,000 riders for Deliveroo and other platforms took part.

Despite the best efforts of successive Conservative administrations to make lawful industrial action nearly impossible, it is hard to see that there could be any legal redress for the employer in these circumstances. First, the DRAA would not constitute a trade union, since it would not be an association of ‘workers’ as required by TULRCA section 1. Secondly, the action would not constitute a breach of contract, so DRAA would not be liable for inducing a breach of contract.³⁷ Thirdly, while it is the case that ballots have to be held before industrial action, even where it does not involve a breach of contract,³⁸ the people who have to be balloted are employees or workers with a personal service contract³⁹—which Deliveroo riders are not. Perhaps this is the way forward for them!

7. CONCLUSIONS

This note makes no apology for approaching the Supreme Court’s decision in Independent Workers Union of Great Britain v CAC and Roofoods Ltd, t/a Deliveroo from the standpoint that it must be wrong. Describing Deliveroo riders as self-employed people in business on their own account and the Deliveroo company as operating virtually without staff despite a business model depending completely on people is a distortion of reality.

Some may feel that the Deliveroo decision is less than cataclysmic since very few enterprises would be prepared to allow such an unfettered right of substitution as set out by Deliveroo. However, it is devastating for the 50,000 plus people working for Deliveroo in the UK, and if this company has found the benefits of avoiding responsibility as an employer to outweigh the risks, it may be that it will become the model for the proliferating private delivery industry.

Most damagingly, it shows a simple method by which to undermine employment protection rights and, it is submitted, a method against which neither the provisions of ILO Recommendation 198 nor the EU Platform Work Directive would necessarily be proof.

As noted already, the issue in the Supreme Court was purely on whether the Deliveroo workers could invoke ECHR Article 11. However, the reason that they were not allowed to appeal on the domestic law point was that it was regarded as beyond argument that they would lose. Even if a purposive approach to interpretation is adopted, it is not at all certain that it would produce a different answer, given the wording of the legislation. The doctrine that personal service is a sine qua non for a contract of employment, or for a ‘limb (b)’ worker’s contract, has achieved such a stranglehold in the case law that it would seem that the only possible solution would be via legislation.

The newly elected Labour Government will shortly unveil details of its promised Employment Bill to improve the rights of UK workers. Its policy paper before the election, A new deal for working people, indicated an intention to give ‘workers’ the same rights as ‘employees’. However, as things stand, this would not assist workers such as the Deliveroo riders, because of the current interpretation of the requirement for personal service. It will be remembered that the Taylor Review recommended codification of the factors for judging employment status back in 2017.⁴⁰ It now appears to be essential in order to deal with this problem. It is true that a new legislative definition risks causing yet more litigation, although this could be minimised by providing that the question is one of fact for the initial tribunal, provided it has directed itself correctly. However, any list of relevant factors would need to be chosen with care. They should include a clear statement that the existence of a substitution clause does not preclude the finding of a contract for personal service. Furthermore, the fact that terms are imposed via a standard form contract without negotiation should be expressly stated as a relevant factor, and consideration of the business model should also be enjoined. Reforms of this kind might go some way to reducing the current levels of complication.

Footnotes