A Manning v Walker Crips Investment Management Limited

This is an appeal brought by Mr Manning, who was the claimant before the employment tribunal (“ET”), against a judgment of Employment Judge Stout (the “EJ”) following a preliminary hearing on 25 March. In reasons sent to the parties on 29 March 2022, the EJ decided that Mr Manning was not a worker within the meaning of section 230(3)(b) of the Employment Rights Act 1996 (“ERA”), and nor was he a worker for the purpose of regulation 2 of the Working Time Regulations 1998 (“WTR”).

I shall refer to parties as the Claimant and Respondent, as they were before the ET.

The Claimant was represented by Mr Carr KC and Mr Cordrey. The Respondent was represented by Mr Halliday. Both Mr Cordrey and Mr Halliday appeared before the ET. I am grateful to all counsel for the high quality of the written and oral submissions.

The Claimant worked for the Respondent as an investment manager from 7 April 2015 until 18 January 2021. In a claim form received on 14 May 2021, he claimed that he had been subject to detriments for making protected disclosures contrary to section 47B of ERA. He also claimed that he had not been paid in respect of his statutory entitlement to annual leave due under WTR. In its response, the Respondent denied the claims and also denied that the Claimant was a worker.

The claim was listed for a preliminary hearing on the question whether the Claimant was a worker within the meaning of section 230(3)(b) of ERA or regulation 2(1)(b) of the WTR. It was not contended for the Claimant that he was an employee of the Respondent. The relevant facts are set out fully in the EJ’s reasons at paras 10-39 and are only summarised below.

The Respondent had two categories of investment manager: those engaged as employees and those, such as the Claimant, who were engaged as “self-employed” investment managers or “associates”. Both appear identical to the outside world: for example, both are allocated offices, equipment, e-mail addresses and business cards and represent the company at external events. But there were differences between the two groups, including that employed investment managers alone are subject to restrictive covenants and are paid a salary rather than commission and fees (paras 11-12).

The Claimant was engaged on terms set out in a letter dated 31 March 2015 and in a contract signed on 7 April 2015 (the “Contract”). Some of the terms were negotiated with him individually and departed from the Respondent’s standard terms (para. 13). The letter referred to the Claimant’s payment by commission and fee sharing and said he would be provided with a “desk, telephone, PC, internet/email/webmail, the first set of business cards”; said he was permitted to work no later than eight pm and that the Respondent would pay half of the cost of a dedicated assistant; provided for the potential payment of a sum on voluntary retirement on condition that the Claimant agreed to an “orderly handover plan” for his clients; and set out the expectations of how he was to behave towards colleagues, staff and management.

The terms of the Contract are summarised by the EJ at para. 14 of her reasons. The Claimant was described in the Contract as the “Associate” and the Respondent was referred to as “WCSB”. According to clause 2, the Claimant was engaged to act for the Respondent as its agent for two purposes (i) procuring persons with whom the Claimant dealt to enter into “Designated Investment Business” with the Respondent or its subsidiaries (but not with anyone else); and (ii) giving advice to such persons, referred to as the “Customer”, about entering into Designated Investment Business with the Respondent (but not with anyone else).

Two clauses are of particular importance to the appeal. Clause 1.3 said as follows:

“In this agreement references to the ‘Associate’ shall, subject to clause 2.5, be deemed to include references to any employee or agent of the Associate or to any person acting on the Associate’s behalf”.

Clause 2.5, referred to in clause 1.3, stated:

“All employees or agents of the Associate providing services on behalf of the Associate under this agreement must, prior to the provision of any such services, be approved by a director of WCSB Group in writing on behalf of WCSB and the Associate must have provided to the WCSB Group Compliance Department such information and documentation as requested by the Compliance Department. The approval of such employee or agent shall be at the sole discretion of WCSB and such approval may be withdrawn by WCSB at any time. Upon withdrawal of approval by WCSB the employee or agent of the Associate may not provide any services under this agreement.”

Various clauses listed in section 4 under the title “Covenants” required the Associate to comply with rules of the Financial Conduct Authority (“FCA”) and internal rules, including the Respondent’s “Personal Account Dealing notice” (see clause 4.1). The Associate covenanted to comply with any “procedures, manuals, instructions or requests issued form time to time” by the Respondent (clause 4.3). In consideration of the Associate introducing business to the Respondent under clause 2, he was paid a share of net commission (clause 5.1). Either party could terminate the agreement on one month’s notice (clause 6.1) and the Respondent had a right to terminate with immediate effect “in the event that the Associate conducts him or herself in a way deemed unbecoming or detrimental to the good reputation of the company” or if the Associate breached the agreement (clause 6.2).

Clause 11.1 said that the “Associate is an independent contractor and the agreement shall not be construed to create a partnership, joint venture or employee/employer relationship between the parties”.

The EJ went on to make findings about what happened in practice during the period the Claimant was working for the Respondent. She described how he was “integrated into the Respondent’s workforce in many practical respects” (para. 18); he worked solely for the Respondent during his engagement (para. 19); after the Covid pandemic he worked at home much of the time (para. 20); he paid for his own professional memberships and professional indemnity insurance (para. 21); he decided “when and whether to take holiday” and was “free to work when and where he wanted”, though he informed his personal assistant when he would be away (para. 22); and he was paid commission based on a fixed percentage of the fees charged, so that if he generated no fees from his clients he would earn nothing (para. 29).

The EJ noted that the Claimant was subject to oversight and control by the Respondent in various ways through its policies, rules and procedures (para. 29). Owing to its responsibility to certify him as a “fit and proper” person for the Financial Conduct Authority or FCA, it had a “very significant degree of control” over his ability to operate as a professional broker. He was also subject to the Respondent’s appraisals and its complaints process, though not its disciplinary policy (para. 35).

Three paragraphs of the factual findings are of particular relevance to this appeal. At paras 36-38 the EJ said this:

“36.

The Claimant worked closely with another nominally self-employed investment broker engaged by the Respondent, Ian Amiee. The Respondent’s position is that the Claimant was using Mr Amiee as a substitute and that he worked on the Claimant’s clients, but was never paid for that work by the Respondent. Indeed, from October 2020 the Respondent had required each broker to nominate an alternate broker as a formal cover during periods of absence and Stephen Simper suggested that the Claimant nominate Mr Amiee, and vice versa. The Claimant’s position is that this arrangement was just what it would have been if the two of them were employees of the Respondent: they provided cover for each other.

37.

There are print-outs in the bundle showing Mr Amiee as having ‘dealt with’ trades for clients of the Claimant. The Claimant said that the vast majority, if not all, of these were not ‘normal’ on the market trades, but ‘placings’ in response to cash calls by companies where the ‘dealt by’ person would simply be the person who took the call from the company or entered the placing on the system. I accept the Claimant’s evidence in this respect, but it is nonetheless the case that in those trades Mr Amiee was formally the nominated deal broker on the transaction, but as the transaction was for the Claimant’s clients he would not have been paid. Moreover, the Claimant accepted that in theory Mr Amiee could have placed on the market trades for his clients, and vice versa, but he was unable to think of an example where that had happened in six years.

38.

So far as the possibility of any broader use of a substitute was concerned, the Claimant said “I don’t think anyone in the office thought it remotely likely” and “it would have been totally impractical”. Mr Darbyshire gave no example of a wider type of substitute being used, although he did say that some of the assistants were qualified to do trades and did carry out trades on behalf of individual associates. He added that the authorisation process need not take long at all for someone already qualified, especially if they were known to the Respondent in which case it would take only a few hours.”

At paras 40-57 the EJ set out a thorough exposition of the legal principles relevant to deciding whether an individual is a “worker”. After citing the statutory provisions, she referred to Autoclenz Ltd v Belcher [2011] ICR 1157 and Uber BV v Aslam [2021] ICR 657, in which Lord Leggatt elucidated the theoretical justification for applying the Autoclenz doctrine when examining worker status.

Next, as regards the issue of “personal service”, the EJ cited the summary of the principles set out in the judgment of Sir Terence Etherton MR in Pimlico Plumbers v Smith [2017] ICR 657 at para. 84, and the subsequent clarification of his judgment by the Court of Appeal (“CA”) in Stuart Delivery v Warren Augustine [2022] ICR 511. At para. 47 she stated:

“In each case, the Tribunal must focus on the nature and design of any fetter on the right or ability to appoint a substitute, to determine whether that was inconsistent with any obligation of personal service.”

The EJ also summarised the judgment of Lord Wilson in the Supreme Court (“SC”) in Pimlico Plumbers v Smith [2018] ICR 1511 in which he said it was helpful in some cases to focus on the “dominant purpose” of the contract, while at the same time emphasising that sole test was the statutory wording.

Turning to the “client or customer” element of the statutory definition, at para. 51 of her reasons the EJ cited the judgment of Mr Recorder Underhill QC (as he then was) in Byrne Brothers (Formwork) Ltd v Baird [2002] ICR 667. At para. 54 of her reasons she listed factors which, according to Uber, were relevant to answering the statutory question, such as the level of subordination, the degree of control and dependence, the extent to which the putative worker is free to develop their own business and the degree of integration.

Finally, the EJ referred to the CA judgment in Paragon Finance v Nash [2002] 1 WLR 685, relied on by Mr Halliday for the Respondent to contend that clause 2.5 in the Contract was subject to an implied term that consent to what was described as a “substitute” would not be unreasonably withheld. I consider that case in more detail below.

The EJ then set out in her conclusions how she had applied these principles. I will only summarise them here, because they are subject to specific challenge under the grounds of appeal. The first issue she dealt with was whether the Claimant met the statutory test of personal service for the purpose of section 230(3)(b) of ERA or regulation 2(1) WTR. There are, it seems to me, five critical steps in her reasoning on this matter:

The EJ then turned to consider whether the Respondent was a client or customer of a profession or business undertaking carried on by the Claimant. She listed twelve factors to which she had regard at para. 68. Her conclusion, taking into account of all those factors, was that the Respondent was a client or customer of the Claimant, who was a person in business on his own account (para. 69). On this alternative basis, therefore, she found that the Claimant was not a worker for the purpose of s.230 of ERA or regulation 2 of the WTR.

Many judges have emphasised that there is no substitute for applying the statutory language in deciding whether someone is a worker, though they have referred to principles, indicative factors or previous authorities which may offer guidance. The distinguished list includes Maurice Kay LJ in The Hospital Medical Group Limited v Westwood [2013] ICR 415 at paras 19-20, Baroness Hale in Bates Van Winkelhof v Clyde & Co [2014] ICR 730 at para. 39, Lord Wilson in Pimlico Plumbers at paras 32 and 43, and Lord Leggatt in Uber at para. 87.

Given those edicts, I had better start with the statutory definition in section 230(3) of ERA, which states:

It follows that there are two express categories included in the statutory definitions and a residual, excluded third category: (i) employees at common law; (ii) workers who are self-employed but who meet the statutory definition, usually referred to as “limb (b) workers”; and (iii) the remaining group of self-employed individuals who are in business on their own account, undertaking work for clients or customers of that business: see Baroness Hale in Bates at para. 31. The third category is not defined and occupies the space left by the other two. However, as Lord Wilson noted in Pimlico at para.35, the second category is partly (and clumsily) defined by the absence of a particular status - the other party to the contract is not “by virtue of the contract” a client or customer of a profession or business carried on by the putative worker.

It is now established that determining whether a person is a worker is primarily an exercise in statutory interpretation rather than an analysis of what the parties agreed to in the written contract. At para. 70 of his judgment in Uber, Lord Leggatt cited the canonical formulation of Ribeiro PJ in Collector of Stamp Revenue v Arrowtown Assets (2003) 6 ITLR 454 as summarising the modern approach to statutory interpretation: “The ultimate question is whether the statutory provisions, construed purposively, were intended to apply to the transaction, viewed realistically”. It is that overarching principle which informed how Lord Leggatt said tribunals should approach and apply the statutory definition of worker (para 87):

“In determining whether an individual is a “worker”, there can, as Baroness Hale said in Bates van Winkelhof case [2014] ICR 730, §39, ‘be no substitute for applying the words of the statute to the facts of individual cases’. At the same time, in applying the statutory language, it is necessary both to view the facts realistically and to keep in mind the purpose of the legislation. As noted, earlier, the vulnerabilities of workers which create the need for statutory protection are subordination to and dependence upon another person in relation to work done.”

This citation echoes what Lord Leggatt had earlier said was the purpose of extending legislative protection to limb (b) workers: that they are in a position of subordination or economic dependence similar to that of employees, the “paradigm case”, and so they need similar protection: see para. 71 of Uber, citing para.17(4) of Byrne Bros. The words of the statutory definition must therefore be approached in light of a purpose to extend protection beyond the common law category, even if the broader group of beneficiaries exhibits similar characteristics.

Mr Carr KC submitted that a wide definition should be given to “worker” in the context of whistleblowing because of the public interest at stake and the specific vulnerability of whistleblowers. However, at para. 71 of Uber Lord Leggatt specifically referred to the general purposes of employment legislation as including the protection of workers against “being victimised for whistleblowing”, as well as ensuring they were not paid too little or did not work excessive hours. In that light, and in the absence of any argument based on the Human Rights Act 1998, I consider Uber indicates that the same, inclusive concept of “worker” should apply in the context of whistleblowing as applies under, for example, WTR or in the National Minimum Wage Act 1998.

The element of personal service in the statutory definition has spawned its own large and growing body of case-law. The authorities originated in cases concerned with whether an individual was an employee at common law. In Byrne Bros at para. 13 the EAT cited in full the passage from Atiyah’s Vicarious Liability in the Law of Torts (1967) which had earlier been relied upon by MacKenna J in Ready Mixed Concrete (South East) v Minister of Pensions [1968] 2 QB 497 at 515 and which was also later referred to by Lord Wilson in Pimlico Plumbers (para. 22). P.S. Atiyah wrote:

“it seems reasonably clear that an essential feature of a contract of service is the performance of at least part of the work by the servant himself. If, therefore, the person in question is entitled to delegate the entire performance of the work to another it is thought that this would be conclusive against the contract being a contract of service”

According to Atiyah, it was only if an individual could delegate the entire work to someone else that he or she was not an employee. Express Echo Publications Ltd v Tanton [1999] ICR 693, described by Lord Wilson in Pimlico (para. 21) as a “clear case”, is most commonly cited as the illustration of how a wide substitution clause could produce that result.

The authorities relevant to employee status have since been applied to cases addressing whether an individual undertakes to perform personally “any work or services” for the purpose of the definitions of “worker” under, e.g., section 230 of ERA. The underlying rationale for this is clear. For if the individual has a contractual right to delegate, per Atiyah, the entire work to another, he or she owes no obligation to perform “any” of the work personally. Hence the individual is not an employee, nor a worker. But it is important that tribunals do not lose sight of the actual statutory question on worker status when it comes to examining a substitution clause: does the worker owe an obligation to perform personally “any work or services” (Footnote: ¹) for the other party? I suspect that sometimes a microscopic focus on the meaning of the substitution clause can obscure an examination of the bigger picture.

A trilogy of recent cases has given guidance on answering that question in the context of substitution clauses. The first authority is Pimlico Plumbers v Smith in the CA, cited by the EJ in her judgment. After referring to the existing authorities on substitution clauses, Sir Terence Etherton MR stated at para. 84:

“Some of those cases are decisions of the Court of Appeal, which are binding on us. Some of them are decisions of the EAT, which are not. In the light of the cases and the language and objects of the relevant legislation, I would summarise as follows the applicable principles as to the requirement for personal performance. Firstly, an unfettered right to substitute another person to do the work or perform the services is inconsistent with an undertaking to do so personally. Secondly, a conditional right to substitute another person may or may not be inconsistent with personal performance depending upon the conditionality. It will depend on the precise contractual arrangements and, in particular, the nature and degree of any fetter on a right of substitution or, using different language, the extent to which the right of substitution is limited or occasional. Thirdly, by way of example, a right of substitution only when the contractor is unable to carry out the work will, subject to any exceptional facts, be consistent with personal performance. Fourthly, again by way of example, a right of substitution limited only by the need to show that the substitute is as qualified as the contractor to do the work, whether or not that entails a particular procedure, will, subject to any exceptional facts, be inconsistent with personal performance. Fifthly, again by way of example, a right to substitute only with the consent of another person who has an absolute and unqualified discretion to withhold consent will be consistent with personal performance”

In dismissing the appeal against the tribunal finding that Mr Smith was a worker, Sir Terence Etherton noted there was no express contractual right of substitution (para. 87) and rejected an argument that an implied right of unfettered substitution was necessary to give the contract commercial or practical coherence (para. 89).

The reasons of Underhill LJ in Pimlico were “essentially the same” as those of Sir Terence Etherton (para. 121). Once more citing Atiyah’s work, Underhill LJ pointed out that the fact an individual has help in doing work does not mean that the individual is not also working (para. 130). It was only a “right to pass on an entire job” which was inconsistent with personal performance (para. 130). Underhill LJ agreed with Sir Terence Etherton MR that there was no basis for implying a term (para. 131), agreed with the summary of the principles given by the Master of the Rolls (para. 132) and highlighted the possibility that a written substitution clause might be disregarded in accordance with the Autoclenz principles (para. 132). Davis LJ agreed with both judgments (para. 120).

The second case in the trilogy is Pimlico Plumbers in the Supreme Court. Lord Wilson referred to the tribunal’s finding that Mr Smith could arrange for the work to be done by another Pimlico operative if he found more lucrative work elsewhere (para. 25). In common with the CA below, he also considered that Mr Smith had no contractual right to appoint a substitute, though for the purpose of his judgment he was prepared to assume that the factual arrangement reflected a contractual right (para. 25). On that premise, he offered this guidance at para 32:

“The sole test is, of course, the obligation of personal service; any other so-called sole test would be an inappropriate usurpation of the sole test. But there are cases, of which the present case is one, in which it is helpful to assess the significance of Mr Smith’s right to substitute another Pimlico operative by reference to whether the dominant feature of the contract remained personal service on his part.”

Noting that it was not usual to use a substitute and that various clauses in the contract were directed to Mr Smith personally, Lord Wilson concluded that the tribunal was entitled to find that the dominant feature of Mr Smith’s contract was an obligation of personal performance (paras 33-34). Here, he drew attention to how the presumed contractual right of substitution was restricted to existing Pimlico operatives, all of whom were bound by the same obligations as applied to Mr Smith (para. 34).

Lord Wilson’s focus on the “dominant feature” in Pimlico and on the attributes of the substitutes is not identical to the principles expressed by the CA below, though both courts reached the same result on the facts of Pimlico. In the third case in the trilogy, Stuart Delivery, the CA considered both judgments. Lewis LJ, with whom Snowden LJ and Moylan LJ agreed, began his analysis of the case-law by explaining that the five points made by Sir Terence Etherton at para. 84 of Pimlico were not statutory categories into which particular facts had to be slotted (para. 34). Lewis LJ returned to this theme at paras 40-41, again referring to para. 84 of Pimlico:

“40.

That is the paragraph on which the respondent places great reliance in this case. In considering that paragraph, however, it is important to bear in mind the following. First, the actual issue for a tribunal is whether a claimant is under an obligation personally to perform the work or provide the services. Secondly, Sir Terence Etherton MR was seeking to summarise the principles to be drawn from existing case law: he was not seeking to establish a rigid classification or lay down strict rules as to what did or did not amount to personal performance or when a right of substitution did or did not negate the existence of an obligation to do work personally. Thirdly, on analysis of paragraph 84, there are only two principles summarised. The first is that if the claimant has what is described as an unfettered right to substitute another person to do the work or perform the services that is inconsistent with an undertaking to do so personally. The second principle is that a conditional right "may or may not be inconsistent" with personal performance depending on the precise contractual arrangements and, in particular "the nature and degree of any fetter on a right of substitution". The third to fifth points made in paragraph 84 are provided, expressly, "by way example", of situations where a contractual right on the part of the claimant may be one indicator that the obligation is or is not one to do the work or perform the services personally. The points made are, in effect, a summary of the earlier decisions (which each involved particular facts) which had been analysed by Sir Terence Etherton MR at paragraphs 76 to 83 of his judgment.

41.

Against that background, it would be wrong to seek to treat those five points as setting out definitive categories of what situations do, or do not, involve a right for a claimant to substitute another person to carry out the work sufficient to displace any contractual obligation to perform the work personally. It will usually be unhelpful to try and shoehorn the particular facts of a case into one of the "categories" listed (they are not in fact categories at all) and then to treat that as dispositive of the issue of whether the claimant is contractually obliged to perform the work personally”

After noting that the CA in Pimlico did not apply those principles, and referring to Lord Wilson’s judgment in the Supreme Court - two factors which serve further to undermine any fixed, rigid categories - Lewis summarised the position in these terms at para. 55:

“It is unhelpful to attempt to force the facts of this case into the language used in point 4 [of para. 84 of Pimlico in the CA] or to try and analyse it by reference to the language used in point 4. It is more appropriate to focus on the real issue, that is whether the nature and degree of any fetter of the right or ability to substitute to determine whether that was inconsistent with personal performance.”

In that light, he rejected the appeal against the tribunal’s conclusion that a limited right or ability to be released from a slot if another courier took it up was not a sufficient right of substitution to defeat worker status (para. 56). That the tribunal was entitled to reach the decision it did was reinforced by the analogous facts of Stuart Delivery to those in Pimlico (para. 57).

The second legal issue on this appeal concerns the analysis of the “client or customer” element of the statutory definition of worker. It, too, is regarded as a question of fact for the tribunal, in which the sole test is the statutory wording itself. But the courts have again drawn on various sources of guidance which assist in answering the question. Three in particular stand out.

One source is the Court of Justice’s interpretation of “worker” in Directives in the social field, which draws a distinction between workers who perform services for and under the direction of another for remuneration and those who are independent providers of services: see especially Allonby v Accrington and Rossendale College [2004] ICR 1328 (Case C-256/01), referred to in Bates at para. 32 and in Uber at para. 72; and Kunsten Informatie en Media v Staat der Nederlanden [2015] 4 CMLR 1 (Case C-413/13), referred to by Lord Wilson in Pimlico, SC, at para. 46.

A second source of guidance is the judgment of Langstaff J in Cotswold Developments [2006] IRLR 181 at para. 53, which the EJ cited at para.54 of her judgment, where he said:

“a focus on whether the purported worker actively markets his services as an independent person to the world in general (a person who will thus have a client or customer) on the one hand, or whether he is recruited by the principal to work as an integral part of the principal’s operations, will in most cases demonstrate on which side of the line a given person falls”.

That paragraph was cited by Baroness Hale in Batesat para. 34, was said to provide “some assistance” by Lord Wilson in Pimlico (paras 43-44), and finds an echo in para. 74 of Lord Leggatt’s judgment in Uber, where he recognised integration could be a hallmark of dependency.

In Bates at para. 38 Baroness Hale referred to a third source of guidance: the “dominant purpose” test which originally arose in respect of the differently worded concept of a “contract personally to execute any work or labour” in the previous discrimination legislation: see Mirror Group v Gunning [1986] ICR 546. In Westwood, drawing on the judgment of Elias J in James v Redcats (Brands) Ltd [2007] ICR 1006, the CA considered it was another useful tool which could assist in drawing the boundary between limb (b) workers and those who were in business dealing with customers and clients: see Maurice Kay LJ at para. 17.

This guidance has sometimes been formulated differently in terms of whether the “dominant feature” of the contract is personal service. In James itself, Elias P explained that the search for a “dominant purpose” might prove to be elusive and it might be better for tribunals to focus on whether the “dominant feature” is personal service (paras 65-67). That aspect of his judgment was approved by Lord Wilson at para. 32 of Pimlico Plumbers, albeit in the context of the personal service element of the statutory definition of “worker”. Given the problems of identifying a single, shared purpose to a contract in the work sphere (where the parties may have different aims) and the emphasis, after Uber, on practice rather than contractual form, it may be that “dominant feature” is a more useful formulation to apply in this context.

It is against that background that I consider the individual grounds of appeal.

I approach the grounds in light of the orthodoxy on the restricted role of an appellate court on questions of worker or employee status. As Lord Leggatt explained in Uber, the assessment of whether someone is a worker involves the evaluation of the factual circumstances in which work is performed and is regarded as a question of fact for the employment tribunal (para. 119). Unless there is a misdirection of law, “the tribunal’s finding on this question can only be impugned by an appellate court (or the appeal tribunal) if it is shown that the tribunal could not reasonably have reached the conclusion under appeal”. Those words must apply equally to the assessment of the individual elements of the worker definition to the extent they involve considering and weighing a range of factual matters: see, e.g., Stuart Delivery at paras 34, 46.

I also bear in mind the now very familiar words of Popplewell LJ in DPP Law Ltd v Greenberg [2021] IRLR 106, summarising much of the earlier case law, that a tribunal judgment must be read fairly and as a whole, without being hypercritical (para. 57). He also said at para. 58 that:

“where a tribunal has correctly stated the legal principles to be applied, an appellate tribunal should, in my view, be slow to conclude that it has not applied those principles, and should generally only do so where it is clear from the language used that a different principle has been applied to the facts found.”

Those cautionary words are especially pertinent here because no criticism is made of the EJ’s clear and comprehensive self-direction and the focus of the appeal has been on the application of the principles.

The grounds of appeal fall into two categories. Grounds (1A) to (1E) challenge the EJ’s reasoning about the substitution clause, whereas grounds (2A) to (2E) challenge her decision on the “client or customer” element of the statutory definitions.

Ground 1 is a challenge to the EJ’s reasoning and conclusion that the Claimant owed no obligation of personal service. It features five sub-grounds , numbered 1A to 1E.

Ground 1A - implication of term. In this ground the Appellant contends that the EJ erred in implying a term into clause 2.5 of the Contract, to the effect that consent to the use of a substitute was not to be unreasonably withheld.

In implying a term, the EJ relied upon Paragon Finance, in which the CA held that the power of a mortgagee to set interest rates was not completely unfettered but was subject to an implied term. Dyson LJ (as he then was) implied a term that the clause would not be exercised dishonestly, for an improper purpose, capriciously or arbitrarily because it met the reasonable expectation of the parties and the “oh of course” test for implying terms in fact: see para. 36. It seems for the same reasons he considered there was an implied term that the lender would not act in a way no reasonable lender would act, drawing an analogy with Wednesbury irrationality in public law (paras 41-42; see Associated Provincial Picture Houses v Wednesbury [1948] 1 KB 223). But he accepted that “the scope of the implied term will depend on the circumstances of the particular contract” and declined to imply a broader implied term that the lender would not charge unreasonable rates (para. 41).

Mr Halliday contends that the EJ’s limitation of the discretion in clause 2.5 of the Contract by means of an implied term was a routine exercise which has been applied in a wide range of contexts where contractual clauses give discretion to one party. He referred, in particular, to Braganza v BP Shipping [2015] ICR 449 in which the Supreme Court held that a contractual term giving a discretion to one party to form an opinion - on the facts, as to whether a death was suicide or not for the purpose of death in service benefits - was subject to an implied term that the decision-maker would not exercise the power arbitrarily, capriciously or irrationally in the Wednesbury sense. Baroness Hale (with whom Lord Kerr agreed) explained the rationale for the implied term in this way at para. 18:

“Contractual terms in which one party to the contract is given the power to exercise a discretion, or to form an opinion as to relevant facts, are extremely common. It is not for the courts to re-write the parties' bargain for them, still less to substitute themselves for the contractually agreed decision-maker. Nevertheless, the party who is charged with making decisions which affect the rights of both parties to the contract has a clear conflict of interest. That conflict is heightened where there is a significant imbalance of power between the contracting parties as there often will be in an employment contract. The courts have therefore sought to ensure that such contractual powers are not abused. They have done so by implying a term as to the manner in which such powers may be exercised, a term which may vary according to the terms of the contract and the context in which the decision-making power is given.”

Baroness Hale went on, at paras 21-27, to refer to Paragon Finance and other authorities expressing the view that a decision-maker’s contractual discretion will ordinarily be subject to an implied term similar in effect to the judicial review principles which apply to public decision making: see, for example, Rix LJ in Socimer International Bank v Standard Bank London Ltd [2008] Bus LR 1304, cited at para. 22, and the statement of Lord Sumption in British Telecommunications v Telefónica O2 UK Ltd [2014] Bus LR 765, para.37, cited at para. 27 of Braganza.

Lord Hodge (with whom Lord Kerr also agreed) appeared to agree with paras 18-31 of Baroness Hale’s judgment (para. 52) and considered the personal relationship in employment might justify a more intense scrutiny of decision-making than in some commercial contracts. Lord Neuberger (with whom Lord Wilson agreed) also cited Rix LJ in Socimer, and thought there was no difference between any of the judgments on the Wednesbury question (paras 102-103).

I consider that care needs to be taken not to read Braganza (or Nash) as an invitation to subject any discretion in an employment contract to an implied term that it will not be exercised unreasonably, however attractive that appears at first blush. Both parties accept that the relevant term implied here was a term implied in fact not in law, and which therefore depends upon the terms of the particular contract. This is consistent with Braganza itself, in which Baroness Hale said the term to be implied “will depend on the terms and the context of the particular contract” (para. 31) and Lord Hodge’s cautionary note that “scope for such scrutiny differs according to the nature of the decision which the employer makes” (para. 56). (Footnote: ²)

In that light, first, it is important not to lose sight of why the implied term is being relied on by the Respondent here. It is not, as in Braganza, being relied upon to protect the putative worker or employee against abuse of power in an unequal relationship or in circumstances of a conflict of interest. On the contrary: it is being relied upon to take the individual outside the scope of the statutory rights which protect against the imbalance of power. For the logic of Mr Halliday’s argument is that many express substitution clauses giving the employer an apparently untrammelled discretion to refuse consent to a substitute should equally be subject to the implied term based on Wednesbury rationality, restricting the scope of the employer’s power to give consent but with the fortuitous consequence that it is much more likely that such individuals are not workers or employees at all. The implied term turns out to be Janus-faced in its consequences.

In this context, the implied term cuts against the courts’ recognition that where a putative employer has an absolute and unqualified express discretion to refuse consent to a substitute, such a clause is not inconsistent with personal performance: see, for instance, the example given by Sir Terence Etherton MR in his fifth point in para. 84 of Pimlico. The introduction of an implied term may also be in tension with the approach in Uber. In Autoclenz a written substitution clause was framed in wide terms but Lord Clarke upheld the ruling of the employment judge that it did not reflect the reality (see para. 37, citing para. 37 of the tribunal judgment). Reformulating that approach in light of Uber, the facts “viewed realistically” may further reduce the space for introducing an implied term into the written contract which has the effect of taking an individual outside the statutory definition.

Second, subsequent to Braganza, in Marks & Spencer Plc v BNP Paribas Securities [2016] AC 742, the Supreme Court emphasised the strict constraints on the power to imply terms in fact: see Lord Neuberger at paras 16-24. The relevant principles are now conveniently summarised in the judgment of Carr LJ in Yoo Design Services Limited v Iliv Realty PTE Limited [2021] EWCA Civ 560 where, after referring to Marks & Spencer and other cases emphasising that the concept of necessity must not be watered down, she said this at para. 51:

“In summary, the relevant principles can be drawn together as follows:

(i)

A term will not be implied unless, on an objective assessment of the terms of the contract, it is necessary to give business efficacy to the contract and/or on the basis of the obviousness test;

(ii)

The business efficacy and the obviousness tests are alternative tests. However, it will be a rare (or unusual) case where one, but not the other, is satisfied;

(iii)

The business efficacy test will only be satisfied if, without the term, the contract would lack commercial or practical coherence. Its application involves a value judgment;

(iv)

The obviousness test will only be met when the implied term is so obvious that it goes without saying. It needs to be obvious not only that a term is to be implied, but precisely what that term (which must be capable of clear expression) is. It is vital to formulate the question to be posed by the officious bystander with the utmost care;

(v)

A term will not be implied if it is inconsistent with an express term of the contract;

(vi)

The implication of a term is not critically dependent on proof of an actual intention of the parties. If one is approaching the question by reference to what the parties would have agreed, one is not strictly concerned with the hypothetical answer of the actual parties, but with that of notional reasonable people in the position of the parties at the time;

(vii)

The question is to be assessed at the time that the contract was made: it is wrong to approach the question with the benefit of hindsight in the light of the particular issue that has in fact arisen. Nor is it enough to show that, had the parties foreseen the eventuality which in fact occurred, they would have wished to make provision for it, unless it can also be shown either that there was only one contractual solution or that one of several possible solutions would without doubt have been preferred;

(viii)

The equity of a suggested implied term is an essential but not sufficient pre-condition for inclusion. A term should not be implied into a detailed commercial contract merely because it appears fair or merely because the court considers the parties would have agreed it if it had been suggested to them. The test is one of necessity, not reasonableness. That is a stringent test.”

It should be noted that Etherton MR relied upon Marks & Spencer at para. 89 of his judgment in Pimlico Plumbers, in deciding that there was no scope for an implied term conferring an unfettered right to substitute. Such a right was not so obvious that it went without saying and nor was it necessary to give the contract commercial or practical coherence.

Third, and perhaps most importantly, the context and language of the contractual clause here are very different from the contractual terms considered in Paragon Finance or Braganza. As Mr Carr KC pointed out, in Braganza there was a sharp conflict of interest between the employer and employee, to which Baroness Hale referred at para. 18, because the employer had an obvious economic incentive to exercise its discretion in a way disadvantageous to the deceased employee. That acute conflict meant the contract lacked coherence without an implied term regulating the scope of the discretion. A similar sharp conflict arose in Paragon Finance if the lender could capriciously or irrationally set high interest rates. I am not persuaded by Mr Halliday’s argument that there was a similar conflict here, between the Claimant’s wish to engage a substitute and the Respondent’s desire that the Claimant did the work personally. The contract and clause was perfectly coherent without any limitation on the express term.

A further contextual reason for not subjecting clause 2.5 to the implied term was relied upon by Mr Carr. Just as at common law an employer may decide to employ someone for any or no reason, so he argued that clause 2.5 gave the Respondent complete freedom to object to proposed employees or agents of the Claimant who might otherwise do work for the Respondent. That context, it seems to me, is a further powerful factor counting against any necessity to limit the discretion in the clause.

Finally, there is the language of clause 2.5. It is detailed and specific. It makes clear that, prior to providing services on behalf of the Associate, employees or agents must (i) be approved by a director of the Respondent in writing and (ii) have provided information and documentation to the Respondent’s Compliance Department. It emphasises the breadth of the discretion because approval is at the “sole discretion” of the Respondent and it may be “withdrawn...at any time”: if anything, the term is framed so as to give a wider discretion than the right to give notice in clause 6.1. The existence of an unrestrained discretion also avoids the uncertainty to which the implied term, with its incorporation of the protean standard of reasonableness, gives rise. What if, for example, the Respondent said it would prefer that the Claimant did not use person X?

The implication of a term is a question of law: See Chitty on Contracts, 34^th Edition, para. 16-005. The EJ decided in favour of implying the term because it was required to give clause 2.5 “efficacy and not leave that clause as being ‘writ in water’”. I do not accept these are sufficient reasons for implying the relevant term. The clause and contract had business efficacy without that restriction, just like other express substitution clauses giving the employer an untrammelled discretion to refuse consent to a substitute. The phrase “writ in water” comes from the judgment of Asquith LJ in Howard v Pickford Truck Co [1951] 1 KB 417, 421 (“an unaccepted repudiation is a thing writ in water and of no value to anybody”). But the power given to the employer in clause 2.5 was operable, effective and of value to the Respondent without the implied term - it gave them a power to reject someone who was acting on behalf of the Claimant without having to give any reason or justification for doing so.

Nor do I accept that the term should have been implied on the alternative test of business efficacy, that such a term is so obvious that it goes without saying. A notional reasonable employer might want to be able to reject someone for any number of reasons, good or bad, without having to meet the uncertain requirements of an implied term of the kind introduced in Braganza. Or the notional reasonable employee might object to such a term if aware that its effect would be to defeat all his statutory employment rights. (Footnote: ³)

For completeness, I should deal with three other points:

For these reasons, I allow ground 1A of the appeal. The consequence is that the Respondent had an unqualified discretion to withhold its consent under clause 2.5. It follows that, subject to the Respondent’s answer (see below), the Claimant owed an obligation of personal performance, just as the EJ indicated would be the consequence without the implied term at para. 61 of her judgment. Such a result also accords with the final example given in the judgment of Sir Terence Etherton MR at para. 84 of Pimlico Plumbers, CA.

In that light, I deal more briefly with the other grounds falling under Ground 1.

Grounds 1B-1E: Considering the “True Agreement”. These four grounds overlap with each other and all draw upon what is said to be the “true agreement”. In summary:

I have already set out some of the Tribunal’s key finding of facts about what in fact happened. It was common ground, and accepted by Mr Halliday, that no formal approval for the use of any of the so-called substitutes was ever given as required by clause 2.5: he accepted there was no evidence before the EJ that the process in that clause had ever been followed.

In addition, I was shown the e-mail exchanges before the EJ about the practice which arose involving the Claimant and Mr Amiee. On 5 October 2020 Mr Simper, the Respondent’s director of human resources, e-mailed both to say that HR wanted to have an “out of office” saying which Associates would cover for each other. He said “Can I take it that you will cover each other and if so I will arrange for you to sign the out of office form”. The Claimant replied shortly afterwards saying that “Yes Ian [Amiee] and I cover each other as associates”.

The EJ considered the factual evidence in two stages. She appeared to decide, first, that the factual practice alone was not inconsistent with an obligation of personal performance (para. 60). However, second, she considered the existence of the express right in clause 2.5 of the Contract and for the reasons she gave at para. 66 decided that “it would not be proper on the facts of this case to regard that substitution clause as a fiction”.

On balance, I do not consider ground 1B of the appeal is made out as a free-standing ground of challenge. The EJ directed herself in accordance with Autoclenz and Uber at paras 42-44 and she considered the factual practice at paras 60 and 66. As Mr Halliday reminded me, the authorities recognise that a substitution clause may be perfectly valid even if never used: see Lord Clarke in Autoclenz at para. 19. In that light, I do not consider the law has reached the stage where the exclusive focus of a tribunal must be on the practice alone and the written terms should simply be ignored altogether, even if there is no presumption or rule it represents the true agreement: on this, see Lord Leggatt in Uber at para. 85.

However, the effect of Uber is to require close scrutiny of the practice to see if the facts are harmonious with an express written term of substitution. That, I consider, is consistent with the judgment of Lewis LJ in Stuart Delivery at para. 58 and HHJ Tayler in Sejpal v Rodericks Dental Ltd [2022] ICR 1339 at para. 32. It is only if the facts “viewed realistically” were consistent with clause 2.5 that the clause should be upheld as genuine.

While the fact that a written substitution clause is never exercised or used does not logically entail it is not genuine, such a fact must at least be highly relevant to whether it reflects the true agreement in the Autoclenz and Uber sense. However, in her assessment of whether clause 2.5 was genuine at para. 66, the EJ did not have regard to this factor. She referred, for example, to the fact that “to a certain extent” Mr Amiee and the Claimant acted for each other’s clients, that “occasionally” Mr Amiee dealt with the Claimant’s clients, and that some other associates made use of assistants, echoing her earlier findings of fact at, e.g., paras 37 and 38. But none of these practices appears to have been based on clause 2.5 at all; none went through the approval process in that clause. Any extra-contractual practice of occasional substitution, operating independently of clause 2.5, was of little consequence in answering the question of whether clause 2.5 itself was, when the facts were viewed realistically, a genuine clause which gave the Claimant an unfettered right to use substitutes.

Mr Halliday nonetheless sought to uphold the EJ’s reasoning on the basis that at para. 66 the EJ was addressing “compendiously” clauses 1.3 and 2.5 of the contract. However, I do not consider that is a fair reading of the reasons: all the discussion in those paragraphs is directed to clause 2.5, the very clause which she referred to as the “substitution clause” in para. 60, and in any event clause 1.3 was expressly subject to clause 2.5.

In that light, and always bearing in mind the limits on an appeal court’s powers to intervene in this sphere, I allow the appeal on ground 1C. The failure to follow the process required by clause 2.5 on any single occasion must, at the very least, have been a highly relevant factor to which the EJ should have had regard in deciding whether the facts “viewed realistically” meant clause 2.5 gave the Claimant a right never to do any work.

Grounds 1D and 1E overlap to a large extent, and ground 1E is closely related to ground 1C. In essence, they both contend that the matters to which the EJ referred at para. 66 did not show that the “true agreement” was that the Claimant would never have to perform any work personally and her conclusion to the contrary was perverse. In his response to ground 1D in his skeleton argument, Mr Halliday says that the Tribunal was only looking at practice to see “whether the substitution clause was genuine; it was not asserting that such a practice in itself constituted substitution which precluded a duty of personal service” (para. 32). He makes a similar point in response to ground 1E.

The EJ clearly directed herself in accordance with the case-law and referred, in particular, to the task of a tribunal to determine whether a right to appoint a substitute is “inconsistent with any obligation of personal service” (para. 47). In examining the facts, she was no doubt entitled and right to examine matters such as the extent to which in practice Mr Amiee did work for the Claimant and other associates used assistants, regardless of whether they were called “substitutes” or something else. With hesitation, I do not consider it can be said that the only conclusion open to her on the evidence was that clause 2.5, even if it had never been used in practice and had never been used by the Claimant throughout his six years of service, was not genuine (though, as set out above, this was perhaps the most relevant consideration to address in considering whether it was).

For these reasons I dismiss the appeal on grounds 1D and 1E, although these grounds of appeal have been overtaken by my conclusions on grounds 1A and 1C.

The Respondent’s answer. It is convenient to address at this stage the first of the two arguments raised in the Respondent’s answer for upholding the EJ’s reasons. Mr Halliday submits that (i) the evidence and findings of the ET showed that Mr Amiee could cover for the Claimant and (ii) because the Claimant never owed an obligation to work, Mr Amiee could always “cover” for the Claimant, meaning the Claimant never owed any obligation to do any work. In my judgement, this argument allows contractual form to triumph over the facts “viewed realistically” and departs from the Uber approach to the statutory question.

The Claimant may have owed no obligation to work but when he was working he was doing so under his contract. I did not understand Mr Halliday to dispute this and, as the EJ recorded, it was not in dispute before her that there was sufficient “mutuality of obligation” (see paras 41, 49). Moreover, even if the EJ found that the Claimant was free to work when he wanted (para. 22), she found that in fact during his engagement he worked solely for the Respondent, noting his evidence that it was not practicable for him to work for anyone else because of the degree to which he was integrated into the Respondent (para. 19). While the EJ did not actually find how many hours the Claimant worked each week, the obvious implication of these paragraphs is that in fact the Claimant provided extensive and exclusive personal service for the Respondent during the time of his engagement. Reinforcing this is para. 61, where the EJ considered that the “facts alone”, viewed without clause 2.5 in the picture, pointed towards personal service.

The practice, or the “facts viewed realistically”, thus fell far short of showing that the Claimant never turned up for work or that Mr Amiee did all the Claimant’s work: all the facts showed was that “occasionally” Mr Amiee dealt with the Claimant’s clients, which was a “very small amount of work” (see ET reasons, paras 66(a) and (b)). The practice, as reflected in the e-mails, also implied that the Claimant would have to be present when Mr Amiee was absent, inconsistent with any obligation on his part never to do any work.

I suspect that, stripped to its essentials, Mr Halliday’s argument does not depend on “cover” by Mr Amiee at all. Its real premise is that the Claimant owed no obligation to do anything under his contract because he could take every day as a holiday. For him, every day was St Monday’s day. This reflects an argument Mr Halliday raised in the ET, based on James v Redcats, to which the EJ referred at paras 48-49 of her reasons. However, just like a zero-hours contract worker, or the valeteers in Autoclenz, the Claimant may not have been obliged to work on any particular occasion or at all under the terms of his written contract. But this was not a reason the EJ relied on for finding he owed no obligation to work personally. On the contrary, she was inclined to the view that, in the absence of clause 2.5 in his Contract and the implied term to which it was subject, the “facts alone” indicated he did owe an obligation of personal service (para. 60). There was material to support that finding, including the fact that the Claimant worked solely for the Respondent and notified his personal assistant when he was away. It was a finding which reflects the Uber approach and I do not consider it is open to me, on appeal, to take a different view.

The second group of grounds of appeal challenges the EJ’s decision, in the alternative, that the Respondent was not a client or customer of the Claimant, based on the 12 factors which she listed at para. 68 of her reasons. There are five specific challenges to the reasons of the ET. In considering each of them, I bear in mind the caution that must be exercised by the EAT on appeals of evaluative questions of this kind, where a tribunal should consider all the circumstances of the particular case and where no single touchstone provides the answer: see, among many others, Windle v Secretary of State for Justice [2016] ICR 721 per Underhill LJ at paras 11-12 and paras 42-3 above.

Ground 2A. This ground is that the EJ, in examining control, wrongly looked at the bargaining power at the time the contract was made and the Claimant’s ability to take clients with him after he left, rather than at control during the currency of the contractual relationship.

The EJ took into account the relative equality of bargaining power at the outset of the relationship (para. 68(a)), as demonstrated by the adjustment the Claimant obtained to the Respondent’s standard terms on fees and the loan granted to him (para. 68(b)). The Claimant’s ability to influence his terms, rather than the Respondent being able to dictate them, was relevant to answering whether he fell within the broad class of persons whom it is the purpose of the legislation to protect: see Uber at paras 75-6. Whereas in Uber the driver’s inability to influence the price paid by passengers and the fact that they had no say in the standard terms were important factors pointing towards worker status, here the same relevant factors pointed the other way.

The Claimant’s ability to take clients with him at the end of the relationship, to which the EJ referred at paragraph 68(c), was also potentially relevant to the extent to which he was in a position of subordination during the currency of the relationship or had a continuing degree of equality of bargaining power throughout the relationship. I do not consider it can be said this factor is irrelevant in making an informed impression of whether the Claimant was a worker.

Finally, the EJ at para. 68(l) addressed the extent of control during the relationship. Her conclusion was that the Respondent had “very significant elements of control” but it was control that was not exercised in the ordinary course of events and did not negate the parties’ relative equality. I accept Mr Halliday’s submission that this aspect of the appeal is, in essence, that the EJ gave insufficient weight to this matter, which is not a sufficient basis to overturn the EJ’s reasons.

For these reasons, I dismiss the appeal on ground 2A.

Ground 2B. This ground of appeal is that the EJ did not focus on or apply her findings about (i) the integration of the Claimant into the Respondent’s business and (ii) the uncontested evidence and finding that during his employment with the Respondent he never marketed his services to anyone else. These were highly relevant considerations to the “client or customer” question, as reflected in para. 53 of the Cotswold judgment and para. 47 of Pimlico in the Supreme Court, to which the ET referred at para. 54 of her reasons. Had it correctly applied those principles, the Claimant argues, the EJ would inevitably have concluded that the Claimant was a worker. The particular focus of this ground was paragraphs 68(e) to (g) of the EJ’s reasons.

The EJ accepted at para. 68(e) that the Claimant was “for the most part fully integrated into the Respondent”, consistent with her earlier findings at paras 11 and 18 that to the outside observer he was indistinguishable from an ordinary employee, though she then went on to refer to two factors pointing the other way. I accept Mr Halliday’s submission that, read fairly and in context, in paragraph 68(e) the EJ was weighing factors which counted for and against worker status.

However, at para. 68(f) the EJ seemed to discount the importance of integration because, she said, the degree of integration of the Claimant was “to a significant extent a matter of choice”. She adopted a similar approach to marketing to third parties, saying this too was “largely a matter of choice” (para. 68(g)). Mr Carr was critical of this approach because, he submitted, it demonstrated that the EJ wrongly looked at hypothetical circumstances rather than what happened in fact. That was not, he argued, what is envisaged by the authorities.

I can see the force in this argument. Anyone who works part-time or on a zero-hours contract has a significant degree of contractual freedom over how much they, in terms of their time, are integrated into the employer’s business. The drivers in Uber were entirely free to work for other companies, but this did not preclude a finding that they were workers during the times they were working: see Lord Leggatt at para. 91. Similarly, throughout the period when he worked for the Hospital Medical Group, of which he was found to be a worker, Mr Westwood also carried on his main profession as a GP and it seems he owed no obligation to operate on patients for hair restoration or work any set hours for HMG, though in practice he worked for them two days a week (see the tribunal’s findings of fact at paras 15-17, cited by Maurice Kay LJ at para. 3 of his judgment in Westwood). But the CA upheld a finding that Mr Westwood was a worker of HMG’s during the periods he was working for it and, on the findings of the tribunal, was “clearly an integral part of its undertaking when providing services in respect of hair restoration” (Maurice Kay LJ at para. 19).

Mr Carr’s argument also finds support in Uber. At para. 91 Lord Leggatt said it is “well established” that “the fact an individual is entirely free to work or not, and owes no contractual obligation when not working, does not preclude a finding that the individual is a worker, or indeed an employee, at the times when he or she is working”. He approved a statement to that effect from Elias J in James v Redcats, adding this was:

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

However, on reflection, I consider this ground of appeal crosses the boundary into the terrain reserved for the ET on the overall evaluative question. On balance, I consider that in the particular circumstances of this case, in which the parties had relative equality of bargaining power and in which the Claimant was free to choose his own hours, the EJ was entitled to have regard to the factors to which she referred at paras 68(f) and (g) in considering the broad question of whether the Claimant was a worker. Viewed in the context of the larger canvass, the freedoms of the Claimant to work for other platforms, and to market himself independently might - I put it no higher than that - be relevant to whether he was in a relationship of dependency on the Respondent or subject to sufficient control by it while working.

I appreciate that at first blush the EJ’s conclusions appear surprising in light of the guidance in Cotswold, the result in cases such as Westwood, and the absence in fact of the intermittent or casual working in practice which, according to Lord Leggatt in Uber, could take an individual outside the worker category. On the facts found by the EJ, the Claimant was fully integrated into the Respondent’s business, worked only for it, and did not market his services to third parties at all. But the passage in Cotswold is no more than an “analytical tool” which may provide assistance and does not itself provide the key to unlock the answer to statutory question (per Maurice Kay LJ in Westwood at paras 16, 20). Given the broad range of matters potentially relevant to that fact-specific exercise, and the limited scope to interfere with a tribunal’s decision on appeal, I do not consider I can say that the EJ was not entitled to have regard to the Claimant’s ability, if he had chosen, to work for third parties. She did not fail to have regard to the level of integration and the absence of independent marketing but gave those matters less weight in paras 68(f) and (g); and weight is quintessentially a matter for the tribunal. Nor, in light of the particular facts and all the matters she considered globally in para. 68, can I say that her decision was perverse: the facts of full integration and no outside marketing fell to be balanced against all the matters to which she referred at para. 68.

For this reason, I do not allow the appeal ground 2B.

Ground 2C. In this ground of appeal, the Claimant contends that the EJ erred in para. 68(c) when she attributed weight to the fact that the parties characterised their relationship, and arranged their tax affairs, on the basis that the Claimant was self-employed. This, it is said, is a misdirection when it comes to considering limb (b) worker status.

The older cases on whether someone was an employee recognise that the label adopted by the parties cannot alter the “true relationship” but it may nonetheless be relevant: see Ackner LJ in Young and Woods Ltd v West [1980] IRLR 201 at para. 30. Ackner LJ gave as an example Massey v Crown Life Insurance Co Ltd [1978] ICR 590, where it was said that a label might resolve an ambiguity in the relationship. If the label often had no or little weight under the older cases about employees, it has even less significance following Autoclenz and Uber. As Lord Leggatt explained in Uber at para. 76 (emphasis added):

“It is the very fact that an employer is often in a position to dictate such contract terms and that an individual has little or no ability to influence those terms that give rise to the need for statutory protection in the first place. The efficacy of such protection would be seriously undermined if the putative employer could by the way in which the putative employer could by the way in which the relationship is characterised in the written contract determine, even prima facie, whether or not the other party is to be classified as a worker”.

Little wonder that in both Autoclenz and Uber the characterisation of the relationship in the written documents was disregarded. The prominence both cases give to the facts on the ground also reduces the room for any ambiguity which a label is needed to resolve: the facts will usually provide a full answer.

However, in contrast to Autoclenz and Uber, here the terms of the Claimant’s written contract were not entirely dictated by the Respondent, as the ET found at para. 13. In principle, that might give more scope for a label to have some weight; but the question needs to be approached in light of the statutory question and what the relevant clause said.

It is now crystal clear on the authorities that the business undertaking exclusion in the worker definition is not co-extensive with those who are self-employed or carrying on business on their own account. In Westwood, the CA specifically rejected an argument to that effect. There, the tribunal had found that Mr Westwood was a “self-employed independent contractor”, in business on his own account, and thus not an employee (see para. 30 of the ET decision, cited at para. 4 of Westwood). The submission to the Court of Appeal for the employer, summarised at para. 9, was that this finding entailed Mr Westwood could not be a worker. It was rejected at para. 19 of the judgment of Maurice Kay LJ, principally because it was inconsistent with the express words of section 230(1)(b) of ERA. As a result, the CA upheld the finding of the tribunal that, applying the more “nuanced” words of that provision, Mr Westwood was a worker despite his also having the status of self-employed independent contractor.

Subsequent cases have affirmed Westwood and recognised that a “limb (b)” worker is a sub-category of the self-employed and so individuals should properly pay tax as such: see Bates at paras 25, 31; Pimlico, SC, at paragraph 42.

In that light, it is hard to see why the ET should have given any weight to the parties’ characterisation of the relationship “on the basis that the Claimant was self-employed” or to the fact that he was, qua self-employed person, responsible for his own taxes (ET reasons, para. 68(d)), when those factors or labels were perfectly consistent with his having worker status. Indeed, the lack of an express exclusion of a “worker” relationship in clause 11.1, set out above, might well support the opposite inference.

Faced with these difficulties, Mr Halliday submitted that the ET was “almost certainly” using “self-employed” in para. 68(d) as a reference to the sub-group of the self-employed who are genuinely carrying on an independent business undertaking for clients or customers of that business.

I do not accept that submission. First, it does not fit with the EJ’s reference to “self-employed rather than employed”, and nor with her reference to the arrangement of tax affairs, both of which are directed to the division between employees and the self-employed.

Second, although the EJ was no doubt referring back to clause 11 in para. 68(d), the clause itself did not purport to exclude worker status, and nor did it state or suggest that the Respondent was a client or customer of the Claimant’s business undertaking. The most obvious purpose of the clause was to distinguish associates, such as the Claimant, from those investment managers who were direct employees, and the reference to “independent contractor” should be read in that context. It therefore provided little support for an argument that the parties viewed the relationship as being between the Claimant’s business undertaking and one of its clients. (Footnote: ⁴)

In Mr Halliday’s oral submissions he relied on para. 18 of Byrne Bros, in which Mr Recorder Underhill, in addressing whether labour-only subcontractors in the construction industry should be considered limb (b) workers, said that the “fact that such a contractor may be regarded by the Inland Revenue as self-employed, and hold certificates to prove it, is relevant but not decisive”. Mr Halliday drew a parallel with what the ET did here: it treated the label of self-employed as relevant to whether the Claimant was a worker.

The comments in Byrne must be read in their context and in light of later authorities. I note that in Byrne Bros Mr Recorder Underhill was unclear about the effect of the words “client or customer” in that definition and wondered if they were simply neutral terms, adding nothing to the term “business undertaking”: see para. 17(3). Westwood, which was decided subsequently to Byrne Bros, has now answered his query and made clear that they are not neutral terms. The “client or customer” element is an additional and necessary ingredient of the statutory question, so that it is not sufficient that someone is genuinely self-employed or carries on a business undertaking: see Westwood at para. 19 and Sejpal at para. 35. Byrne, in other words, may have assumed that the status of genuine self-employment had rather more significance for worker status than the later, binding cases. The comment in para. 18 of Byrne should be read in that light.

As a result, after Westwood, I respectfully consider tribunals should treat with caution the comments in Byrne that answering the question of worker status will involve “almost all or most of the same considerations” as arise in deciding whether someone is an employee but with the boundary pushed further in favour of the worker (para 17(5)). While similar considerations such as subordination, control and integration may apply to both, as Uber explains, it must not be forgotten that the limb (b) element has its own specific elements, including the third, negative hurdle of whether the putative employer is a client or customer of the putative employee. Focussing on whether someone is employed or self-employed, even with the boundary nudged in the individual’s favour, risks overlooking every element of the specific statutory language - which the authorities remind tribunals is the ultimate and sole test.

In that light, I do not consider a decision of the Revenue on the question of whether someone is self-employed for tax purposes can displace the primary task of the ET to find the facts and decide for itself whether those facts “viewed realistically” mean that someone meets the specific terms of the statutory definition of worker. I am inclined to consider that a determination by the Revenue that someone is self-employed is likely to be of little assistance in that exercise, though it is not an issue that arises here. By the same token, I consider that the characterisation of the relationship in clause 11 as not being one of employer and employee, and the accompanying obligation on the part of the Claimant to pay his own taxes, were not matters which should have been given any weight. Either the facts “viewed realistically” showed that the Claimant was a worker within the terms of the statutory definition or they did not, and the label in clause 11 was neutral on that question and had no weight.

I therefore consider the EJ misdirected herself at para. 68(c) in saying that weight should be attributed to clause 11 and I allow the appeal on ground 2C, subject to the point raised in the Respondent’s answer considered below.

Ground 2D. In this ground the Claimant contends that the ET erred in considering, at paras 68(e) and (l), that the fact a number of elements of control were exerted for regulatory reasons weakened the Claimant’s case on worker status.

In Uber Lord Leggatt said this at para. 102:

“I would add that the fact that some aspects of the way in which Uber operates its business are required in order to comply with the regulatory regime - although many features are not - cannot logically be, as Uber has sought to argue, any reason to disregard or attach less weight to those matters in determining whether drivers are workers.”

See, to similar effect, Sejpal at para. 21.

The EJ expressly cited this passage at para. 55 of her reasons, so considerable caution must be exercised before deciding she misapplied it. Her conclusion at para. 68(e), that the Claimant was appraised “only from a regulatory perspective rather than more broadly”, reflected the fact she found at para. 33. I do not consider it can be said that she was giving any less weight to the nature of the appraisals because they were focussed on regulatory compliance. Rather, she was stating what in fact was the nature of those appraisals, and contrasting them with the wider appraisal process applicable to investment managers who were employees.

Nor do I consider she discounted the extent of the factual control exercised by the Respondent because it had, in part, a regulatory source in para. 68(l) of her reasons. The focus of that paragraph was on the extent of the factual control exercised by the Respondent. Her statement that the Claimant enjoyed great freedom so long as “he stayed within the regulatory framework as it was understood and implemented by the Respondent” was directed to the control that the Respondent in fact might exercise.

In those circumstances, I reject the appeal on ground 2D.

Ground 2E. There are two aspects to this ground of appeal:

The second element of this ground of appeal has largely been addressed already under ground 2B. The EJ specifically addressed the Claimant’s evidence that it was not practical for him to work for other investment platforms, to which she referred at para.19, and decided there was “no reason in principle” this could not have been done (para. 68(f)). This ground of appeal is, in essence, a perversity challenge to her factual conclusion and I do not consider it can surmount the high hurdle of appeals on such a ground.

As to the first element of this ground of appeal, at para. 68(i) the EJ held that the Claimant “conducted his own business as well during normal business hours, investing on his own personal account to a substantial extent”. She drew an analogy with Wolstenholme v Post Office Ltd [2003] ICR 546, in which the EAT, in upholding a finding that sub-postmasters were not employees or workers of the Post Office, held it was a client or customer of the business undertaking of sub-postmasters. That business undertaking comprised (i) the retail business of selling goods and services and (ii) acting as the Post Office’s agent in providing goods and services to the public (see EAT at para. 45).

But in undertaking personal investment activities, Mr Carr submits, the Claimant was doing no more than what other staff could do under the Respondent’s “Personal Account Dealing Rules”, which applied both to employed investment managers and associates, such as the Claimant. Those rules allowed trading on personal investments, provided it was “not undertaken with such frequency” that it took staff away from their priority of putting the clients first (see para. 15). In doing such trading, Mr Carr submitted, the Claimant was not conducting a business undertaking but simply making investments for himself personally. Hence, he submitted, this factor was irrelevant to whether the Respondent was a client or customer of his business undertaking and the EJ was wrong to draw any comparison with Wolstenholme.

The response of Mr Halliday was that the EJ at para. 68(i) was simply looking at the autonomy of the Claimant while working for the Respondent, relevant to the degree of control it exercised over him and hence the statutory question.

I do not consider Mr Halliday’s interpretation fits with what the EJ said at para. 68(i). At para. 53 she cited Wolstenholme as authority for the proposition that an agency relationship was not inconsistent with the principal being a client, and she returned to the same point at para. 69, again in addressing whether the putative employer, the Respondent, was the Claimant’s client. In that context, I do not consider para. 68(i) was addressing the degree of the Claimant’s autonomy. I consider that, rather, the EJ was relying on the Wolstenholme case to demonstrate that, just like the sub-postmasters in that case, the Claimant was conducting his own business undertaking both for the Respondent and in other respects during working hours. It therefore supported her conclusion that the Respondent was a client of the Claimant’s business undertaking because, on that assumption, his business was not restricted to his work for the Respondent but had other clients.

In my view, that reasoning discloses an error of law. In Wolstenholme the EAT held that the sub-postmasters’ business undertaking comprised the retail business and their acting as agents for the Post Office, and the two parts of that business conducted from a single premises could not be severed (para. 46). That supported a finding that the Post Office was a client of a sub-postmaster because acting for it was part of a larger business, in which the sub-postmaster had multiple clients/customers - the Post Office, any other principals for whom they acted as agents in selling goods or services, and the public to whom they provided goods or services.

But that was not the case here. In trading in his personal investments, the Claimant was not conducting or carrying on part of his business undertaking at all. He was not acting for anyone else, nor providing a service to any client or customer. He was doing exactly what was done by directly-employed investment managers conducting trades for their personal benefit. It follows, in my judgement, that the analogy with Wolstenholme was incorrect: the personal trading of the Claimant, on his own behalf, did not support a conclusion that the Respondent was a client or customer of a larger business undertaking carried on by the Claimant.

I therefore uphold ground 2E of the appeal, but only to the extent it relates to para. 68(i) of the ET’s judgment. This is subject to the Respondent’s answer, which I now consider.

Respondent’s answer. In the Respondent’s answer, an alternative basis is given for upholding the ET’s decision is that personal service was not the “dominant feature” or “dominant purpose” of the Contract. Rather, it is submitted, its dominant feature was to generate fees and commissions from clients, and to share those fees between the Claimant and Respondent. Reliance is placed on paras 52-68 of James v Redcats.

I do not accept that argument. First, James preceded the more detailed guidance to the statutory question provided by the CA in Westwood and subsequent cases. Those cases emphasise that “dominant purpose” or “dominant feature” is not the legal test but just one potential source of guidance, and that tribunals must apply all elements of the statutory test, including the “client or customer” part. Even assuming that the dominant feature of this Contract was not the provision of personal service, I do not consider it could be said that the only conclusion open to the ET here was that the Respondent was a client or customer of the Claimant’s business undertaking.

Second, underlining this point is that at para. 50 of James Elias J acknowledged that the guidance in Cotswold would often provide the answer to the statutory question. Here, as set out above, the facts of the Claimant’s integration, and the absence of marketing to the world in general, far from suggesting he was not a worker tended to suggest he was. Although I do not consider it was perverse of the EJ to decide he was not a worker in light of those matters, they undermine any contention that the only permissible conclusion from the EJ’s findings was that the Claimant was not a worker.

Third, as mentioned in para. 40 above, in James Elias J highlighted that the search for a dominant purpose may be elusive and capable of more than one answer. While the “dominant feature” may be a better test, the Contract here was silent as to its dominant feature or purpose and the EJ made no finding on this matter. As Mr Carr submitted, the present case is a long way distant from Mirror Group v Gunning [1986] ICR 546, to which Elias J referred in James, where it seems the only concern of the respondent was that the newspapers were efficiently delivered. In light of clause 2.1 of the Contract, requiring the Claimant to act as agent for the Respondent for the purposes set out therein, and the need for employees or agents to be approved by the Respondent, it is perhaps more plausible that its dominant feature was personal service by the Claimant. But this only serves to emphasise that the assertion in the Answer is not made out.

In light of the above, in which I have had the benefit of extensive submissions and no doubt more time than the EJ, my conclusions are:

When a copy of this judgment in draft was sent to the parties, I asked for their submissions on disposal. Both provided short written submissions and Mr Halliday made a short response to the Claimant’s submissions in an e-mail. Following Jafri v Lincoln College [2014] ICR 920, and based upon the findings of fact of the ET, I do not consider that there is only one possible answer to the question whether the Respondent was a client or customer of a business undertaking carried on by the Claimant. As set out in para. 91 above, I have rejected the Claimant’s ground of appeal that the EJ’s conclusion on this element of the statutory test was perverse. The parties do not agree to the EAT determining the matter, and I therefore consider the matter should be remitted to an employment tribunal.

The remaining issue is: to which employment tribunal? The Claimant contends it should to a different tribunal; the Respondent argues it should be remitted to EJ Stout. I have considered the relevant factors in Sinclair Roche & Temperley v Heard [2004] IRLR 763 at para. 46, which both parties agree apply. First, as to proportionality, there is likely to be some saving in the parties’ time and costs if the matter is remitted to EJ Stout, though I accept that this is only likely to be a small saving in a preliminary hearing which only took two days to be heard. Second, the hearing took place a little more than a year ago and much of the evidence is documentary. In those circumstances, loss of recollection is not a factor counting against remission to EJ Stout. Third, there is no suggestion of any bias or partiality. Fourth, the EJ’s decision was not wholly flawed: I have allowed the appeal only on the grounds specified above, in an area which continues to trouble employment tribunals (and EAT and appellate judges). Fifth, I have confidence the EJ can fairly re-examine the issues relevant to worker status in light of this judgment. I note the Claimant’s concern that the EJ was a member of the same chambers as the Respondent’s counsel. But this is not ordinarily a factor giving rise to a real danger of bias (and it was not raised as a ground of appeal before me) - see Locabail (UK) Ltd v Bayfield Properties Ltd [2000] QB 451 at para. 25 – and I consider there is nothing to displace the assumption that the EJ will deal with the matter professionally and fairly on remission.

Weighing those factors, I consider the balance is in favour of remission to EJ Stout. The sole issue for that hearing will be the “client or customer” element of the statutory definition in section 230 of ERA and regulation 2 of WTR, to be addressed in light of this judgment. Any case management directions for that hearing are a matter for the employment tribunal.