In 2018, the government committed to legislate on the tests for employment and worker status in their Good Work Plan. However, in their recent response to their 2018 employment consultation, they rowed back on this commitment, stating that “the benefits of creating a new framework for employment status are currently outweighed by the risk associated with legislative reform”. They said so on the basis that (i) such might “create cost and uncertainty for businesses in the short term”, and (ii) they fear “unduly limit[ing] the ability of the courts to rule against unscrupulous employers engineering employment relationships that circumvent the law”.
This article agrees that legislative intervention in this area bears considerable risks, but notes that the law as it stands suffers from significant uncertainty. It identifies three areas in particular in which worker status under the Employment Rights Act 1996 (“ERA”) and its analogue under s.83(2)(a) Equality Act 2010 (“EqA”) have caused judicial strife in recent cases. Businesses and individuals alike will as a consequence struggle to predict where the scope of protection currently lies. We therefore conclude by making some tentative remarks about what the law is and how it should develop.
Mutuality of obligation and periods of non-work
To what extent, if any, should Tribunals look at a putative worker’s periods of non-work, in cases where that putative worker works periodically, or pursuant to a series of individual contracts rather than one overarching one?
In the recent case of Johnson v Transopco UK Ltd  EAT 6,  ICR 691, the Claimant was a black cab driver who carried out on average 1.5 daily trips via the Respondent’s app “Mytaxi”, from which he received around 15% of his total income. There was before the EAT no dispute that the Claimant was able to establish that he had contracted to perform work personally for the Respondent. However, the EAT upheld the ET’s finding that the Claimant was not a “worker” of the Respondent because he was a “client or customer” thereof. It declined to hold that the ET’s reliance on periods in which the Claimant’s activities outside of work when considering the nature of his business was impermissible (paras. 58, 75). Following the judgment of Underhill LJ in Windle and anor v Secretary of State for Justice  EWCA Civ 459,  ICR 721 paras. 22-24, the absence of “mutuality of obligation” in periods outside one’s work with a Respondent may influence the analysis as to whether one is acting as a client or customer of that Respondent when in work. In these cases, “mutuality of obligation” means an obligation on the putative employer to offer work and an obligation on the putative worker to accept work when it is offered: Johnson (para. 72).
The judgment in Sejpal v Rodericks Dental Limited  EAT 91,  I.C.R. 1339 also addressed the issue of periods of non-work but struck a contrasting tone. HHJ Tayler dismissed the idea that the concept of mutuality of obligation has much use in worker status cases. At para. 23 he stated that the concept has use in assessing whether the work is performed pursuant to a relevant employment (or worker) contract, since such bilateral contracts typically contain an ongoing obligation to keep the relationship in place until it is lawfully severed.
But outside of this limited function, the concept has little value. He added that there is no “irreducible minimum” of obligation that is required in periods of non-work: what is required is simply a focus on the work itself. These observations appear to have general application in worker status cases, and suggest a move away from the analysis of Johnson and rulings like Underhill LJ’s Pimlico Plumbers Ltd v Smith  EWCA Civ 51,  ICR 657 in which he suggested that mutuality in periods of non-work is essential in cases that depend “on the claimant's status during periods of non-work, either because he or she has to establish continuity of employment or because the claim itself relates to their treatment during that period” (para. 145).
So where are we now? If Sejpal stands then there is no requirement that a periodic worker must demonstrate some mutuality of obligation in periods outside of their actual working time. However, following Johnson, the fact that the putative worker spends significant periods of that time working for others as part of their business may militate against their argument that they are more than merely a client or customer of their putative employer. HHJ Auerbach in Johnson suggests therefore that mutuality of obligation has a part to play in both the contract and business exception stages of the analysis, which is one way in which the concept and its application are liable to confuse.
Although HHJ Auerbach says no more than “it’s not impermissible to look at periods outside of work”, in our view it is difficult to escape the conclusion that irregular workers, those who “multi-app” or those who need to work for multiple employers, will struggle under this analysis, despite being those towards whom the legislation should be most squarely oriented. Although this argument was rejected by HHJ Auerbach, Chris Milsom was right to suggest that where this leads to is a focus on the proportion of one’s work one does for the putative employer, and to a counterintuitive conclusion that those who work for only one employer will have stronger claims than those with several. A worker (X) who works 10 hours for an app on top of a 40-hour week under a contract of employment is therefore more likely to receive protection for that 10-hour period than a worker (Y) working 10 hours a week for five different apps. The idea, mooted in Johnson, that having several sources of income indicates that there is lesser “dependency” (para. 55) or greater equality of bargaining power (para. 82) in respect of an individual contract holds little water. Worker X is just as dependent on the 10 hours of work per week as worker Y is, and has just as little standing to negotiate on the terms and practicalities of their work with that app.
Because of the contrasting EAT judgments on mutuality of obligation, and because of these surprising and counterintuitive consequences, in our view judges on appeal should discourage this focus on non-working periods (though may hold that it is permissible in certain exceptional cases), and look more closely at the work that is actually done, taking a global view rather than individuating each individual working period. The Sejpal approach is therefore to be preferred. The primary question is who is the legislation designed to protect, not when do they get protection.
In a much-quoted passage, Sir Terence Etherton in Pimlico Plumbers Ltd v Smith  EWCA Civ 51,  ICR 657 set out some important principles relating to the statutory requirement of personal service at para. 84:
“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”.
It follows from the ruling in Uber v Aslam  UKSC 5 that assessing the conditionality of a right to substitute will not depend on looking primarily at the contract, and instead involves examining the reality of the arrangement between the parties. But the impact of Uber may go further. In obiter at para. 32 of Sejpal, HHJ Tayler suggested that:
“there could be a situation in which despite there being a contractual term that provides an unfettered right of substitution, the reality is that the predominant purpose of the agreement is personal service, so that the person is a worker. It might even be argued that personal service need not be the predominant purpose of the agreement, provided that the true agreement is for the provision of “any” personal service as required by the statute”.
In effect, a putative worker might succeed in their claim despite having a contract under which in practice only some work must be personally performed (and not performed by a substitute). The rest of the time, an unfettered right of substitution may take effect without debarring worker status. HHJ Tayler is here emphasising the importance of the reference in s.230(3)(b) to the requirement that the putative worker must “perform personally any work”. Employers will not be able to evade the legislation’s reach by permitting occasional periods in which substitution can take place without any oversight from them.
Some questions arise. If the focus of the statute is to protect those who undertake to do any work personally, does it even need to be shown that the predominant purpose of that contract is personal service? We could see a case in which a putative worker is largely free to perform the work via substitutes but a material aspect of the contract must be done by them personally, like a courier who can use substitutes but must perform some office work themselves. These office duties need not be a predominant purpose of the contract nor do they need to take up significant periods of the worker’s week. Does that worker then get protection for the periods in which that unfettered right was operative? And what about cases where the claims stretch over periods of time? Must a worker in a protected disclosure case demonstrate that the unfettered right to substitute was not operative at the moment that they made the disclosure, and again later when they suffered a detriment on the grounds of it?
It is clear that the law on substitution is evolving and is becoming less restrictive as it develops. As in relation to mutuality of obligation, judgments like that in Sejpal permit greater focus on the real question emphasised in Uber: who is the legislation seeking to protect? Although interesting questions arise to how this developing approach will work in practice, it is encouraging to see courts and Tribunals taking a more global view at a working relationship rather than allowing employment law’s protection to be evaded by occasional use of unfettered substitution powers or of performing some work for other employers.
In answering these tricky questions and developing the law, courts and Tribunals will need to apply the “purposive approach” approved in Autoclenz Ltd v Belcher  UKSC 41,  ICR 1157 para. 35 and later in Uber (paras. 75, 78). But how that manifests in practice is somewhat difficult to identify. In Johnson, the court asked itself (following Autoclenz and Uber) whether the Claimant was in a dependent or subordinate relationship. But how are we to understand cases to which the legislation is clearly intended to apply that lack subordination or dependency, such as the cases of CEOs and other executives? In the recent case of Catt v English Table Tennis Association  EAT 125, Eady J, president of the EAT, deprecated the first instance Tribunal’s focus on “issues of vulnerability, subordination and dependence” (para. 49). She continued: “while these were all matters of understandable emphasis in the Uber case, they were inevitably less relevant to the position of the claimant as a non-executive director”.
We know from Bates Van Winkelhof v Clyde & Co that the absence of subordination does not preclude worker status in every case. But if not subordination, to which principles are judges to turn when assessing the reach of the legislation? One question that remains is whether Tribunals will determine that the ERA and the EqA have different protective purposes. In addition, Lord Leggatt in Uber expressly declined to centre unequal bargaining power (para. 68), in line with the move away from a more contractual focus. Indeed, the concept of inequality of bargaining power can lead to some surprising analysis: the Tribunal in Johnson suggested that, because the Claimant brought in income from a range of app work, bargaining power as between him and Mytaxi was relatively even, despite the Claimant doubtless having close to no say in how his terms were formulated.
The answer may be that the legislation has a multitude of purposes, given that it seeks to grant protection to a huge range of individuals across the spectrum of working relationships. It seeks to protect those who lack bargaining power as well as those with plenty of that; it protects those in relationships of dependence and independence. But where borderline cases arise, at present it is difficult to predict which concepts courts and Tribunals will find most useful to help them in their determinations.
Despite the uncertainty, the dust is beginning to settle on this area of law. The extent to which well-advised employers can inhibit statutory protection by adding an unfettered right to substitute is diminishing. The focus of the analysis is being reoriented towards the statutory wording: workers can be protected having undertaken to perform personally any work, provided they are not a client or customer of the employer. In examining that latter aspect, post-Sejpal courts and Tribunals will likely focus less on periods of non-work and will analyse the claimant’s actual working time, though some latitude to look beyond that is preserved by Johnson in the appropriate case.
As to how a purposive analysis should operate in borderline cases, the following appears to be a sensible summary of the type of approach courts should take, at least in relation to ERA claims. (1) One should start by asking if there is a relationship of subordination or dependence. If there is, it is likely the legislation is intended to apply. (2) However, if there is not, one secondly asks: is this is the type of worker that the legislation appears to intend to cover in any event?