The Changing Concept of the 'Employer' in British Labour Law
Simon Deakin (University of Cambridge)
It is a striking feature of modern labour law that the volume of material devoted by courts and commentators to refining the concept of the 'employee' (and, now, the 'worker') completely overshadows the few attempts which have been made to address the nature of the 'employer'. In part this is because the legal or juridical nature of the employer can be inferred from an examination of the case law on the definition of the employee. The 'control', 'integration' and 'economic reality' tests all touch on issues of managerial coordination and the assumption of risk which can be seen as identifying the enterprise in legal terms. However, the legal nature of the employer is increasingly emerging as an issue in its own right. Judging by the number of reported cases alone, this is now one of the major issues facing courts and employment tribunals. Many of these cases stem from the growing practice of supplying labour services to an end user through intermediaries, such as personal service companies or employment agencies. The problem of identifying the employer also arises in other contexts, including corporate groups and the outsourcing of labour services. To resolve the problems which arise here, it is necessary to go beyond an analysis which focuses on the familiar issue of whether the applicant or claimant is an employee. What is required is an understanding of how particular rights and liabilities are to be allocated when the traditional functions of the employer - in particular coordination, in the sense of managerial decision-making, and the assumption of certain social and economic risks - are divided among a number of different entities (1).
Over time, a variety of techniques have been used to address this issue. These include the adoption of special rules to govern agency work, the partial lifting of the corporate veil to allow legal duties to be shared among 'associated employers' within a corporate group, and, best known of all, the transfer of employment rights and obligations from one employer to another under TUPE (2). These techniques, however, deal with a series of specific problems in isolation from each other. The purpose of this note is to see to whether, in the course of a review of the recent case law, it is possible to develop a more integrated approach to the definition of the employer in labour law.
As a first step, the employer may be defined for labour law purposes as simply the 'other party to the contract of employment' (or, in the case of certain 'workers', the contract for services). This definition is nowhere clearly stated either in legislation or in the common law, but it may be inferred from those provisions of the principal labour statutes in which the terms 'employee', 'worker' and 'contract of employment' are defined (3). This contractual definition does not take us very far, but it does have one virtue, which is to remind us what, in legal terms, the employer is not. The legal meaning of the employer is not synonymous with the sociological or economic idea of the 'enterprise' or 'organisation', nor with the workplace, that is, the physical site on which work is carried out. The identification of the employer may well be determined in part by considerations of organisational and workplace boundaries. However, these are just some of the factors which a court or tribunal may take into account under particular circumstances, and they may be discounted altogether if a contractual analysis points in the opposite direction.
The need to establish a contractual nexus between employer and employee (or worker) can have the effect of completely dis-embedding the employment relationship from the organisational context in which the work is performed. The employer is not necessarily the person to whom service or services are supplied. This is starkly illustrated by the outcome of Nokes v. Doncaster Amalgamated Collieries (4). Despite continuing to work in the enterprise for a period of several weeks after the ownership of the business concerned was transferred from one company to another, the appellant at no stage gave his consent to the supposed novation of his contract and so, at the time of his alleged misconduct, was not an employee of the transferee. By these means, he avoided being subject to the disciplinary powers of his would-be employer - it is not often remembered that Nokes was a case in which the respondent was attempting to invoke the quasi-criminal jurisdiction of the magistrates in cases of breach of the contract of service under the Employers and Workmen Act 1875. This emphasis upon the workers' consent as the basis for novation was recently reasserted in Bolwell v. Redcliffe Homes and O'Connor (5), a case arising from the not unusual problem of identifying which of several employers on a construction site should be made liable for personal injury arising from a breach of health and safety regulations.
Nokes is of course better known as the case which, prior to the passage of TUPE in 1981 (6), enabled the purchaser of a business to acquire its physical and commercial assets without coming under any obligation towards its employees. Here, then, freedom of contract meant that the employees' expectations of continuing employment could be defeated by the will of the transferee. Under regulation 5 of TUPE, assuming that the Regulations can be shown to apply, the contract of employment is automatically and compulsorily novated (unless the employee objects), thereby overcoming some of the limitations of a contractual analysis.
No such statutory constraint operates in the case of the supply of labour through intermediaries. One of the consequences is that workers may frequently find themselves without an effective claim to dismissal protection or compensation following a reorganisation of the workplace in which they have been working. In Secretary of State for Education and Employment v. Bearman (7), two disabled workers had worked for several years in a government-run jobcentre when their employment there was ended following a reorganisation. Their employment was administered under a 'Sheltered Protection Scheme' under which their 'sponsor' organisation, Royal British Legion Industries (RBLI), was responsible for finding them a placement with their 'host', the Employment Service. Under this arrangement, the Employment Service was responsible for paying each of the applicants a proportion of their salary (which was calculated by reference to how far their disability prevented them from completing a 'standard' work load), with RBLI paying the balance of the salary under the terms of a government grant paid for that purpose. Although for most purposes treated in the same way as the civil service employees alongside whom they were working, the applicants did not qualify for a civil service pension, nor for other occupational benefits.
The contractual documentation issued to the applicants stated that they were employees of RBLI, and this was also stated in the contract agreed between RBLI and the Employment Service. The tribunal chairman nevertheless held that the 'true nature' of their relationship was that they were employed by the Employment Service: '[t]he length of their service, the way in which they were managed, the way they regarded themselves and must have been regarded by an external observer, the way in which their employment came to an end are factors which outweigh the distinctions between them and other Employment Service employees and the involvement of RBLI in their welfare and in the arrangements for their payment' (8). This ruling was overturned on appeal to the EAT, which decided that the tribunal chairman had misdirected himself: 'the correct approach would have been to start with the written contractual arrangements and to have inquired whether they truly reflected the intention of the parties' (9). In the absence of any indication that they did not, and no evidence of any subsequent novation of the applicants' contracts, they could not be regarded as employees of the Employment Service.
According to Morison J., 'it would be unfortunate if an industrial tribunal were to discourage such schemes [as the present] by treating the person's integration as a reason for disturbing very carefully structured contractual arrangements'. The purpose of these arrangements was precisely to release the user of labour from those obligations which it would normally owe to its employees, on the grounds that '[t]here may well be a number of employers who would be reluctant to commit themselves to an employment relationship with, but would be quite happy to take on the services of, a disabled person and to integrate him or her properly into their organisation' (10). This is another way of saying that the arrangement arrived at by the parties involved subsidising the user (in the sense of providing labour at less than its normal cost) in order to overcome its reluctance to employ disabled persons. The spirit of the scheme was quite different from the assumption now made in the Disability Discrimination Act 1995, namely that much of the discriminatory treatment suffered by disabled workers has no objective basis. For present purposes, what is particularly striking about the decision of the EAT is the pre-eminence accorded to contractual analysis in its approach to identifying the employer.
A purely contractual analysis provides employers with strong incentives to adopt arrangements whose sole or principal purpose is to avoid the application of protective legislation. Regulatory legislation must have ways of dealing with the more egregious attempts at evasion of this kind, or face the prospect of apparently mandatory provisions becoming waivable at the employer's option. The deeper issue, however, is how far the law is prepared to go in upsetting, after the event, the contractual allocation of risks which the parties have made. This may involve going around or overriding arrangements which were not set up as 'shams' in the sense of being nothing more than attempts at evasion.
In Bearman, the court was reluctant to undermine a scheme under which the 'host' and 'sponsor' divided the normal responsibilities of the employer between them and, in the process, effectively left the workers concerned unprotected in a number of significant respects. The decision is questionable since the workers had had no effective say in this division of risks and responsibilities. As individuals facing the possibility of exclusion from the labour market because of stereotypical views of their ability to work 'normally', they were in a particularly vulnerable bargaining position. It could be argued that they received a benefit, in terms of access to employment, as a result of the scheme under which they were placed with the Employment Service. It does not follow, however, that this benefit should only have been made available at the expense of employment rights which they would otherwise have enjoyed.
At the other extreme to Bearman, it is possible to envisage situations in which the employee or worker is an active participant in setting up arrangements which split the functions of the employer among two or more parties. It has become common for workers in the computing and information technology sectors to offer their services to employers through their own 'personal service companies'. The reason for doing so is described in the following terms in an explanatory note prepared by HM Treasury on the Finance Bill 2000:
'Prior to 6 April 2000, it was possible for workers to work through intermediaries such as personal service companies to provide services to clients in circumstances where, if it were not for the service company, the worker would be an employee of the client. The use of intermediaries in this way allowed the client to make payments to the personal service company rather than the individual, without deducting PAYE or NICs [national insurance contributions]. The worker would then take the money out of the service company in the form of dividends instead of salary. Dividends are not liable to NICs so that the worker paid less in NICs than either a conventional employee or self-employed person. There were also tax advantages in these arrangements' (11).
The government's response to this has been to put forward legislation which deems a worker in this position to be directly employed by the client for tax and national insurance purposes. In effect, the presence of the intermediary personal service company is ignored, and the workers is treated as an employee of the client if, in all other respects, the nature of their employment points to this conclusion. Here, then, the contractual arrangement made by the parties is treated by the legislation as little more than a sham. The assumption behind the statutory intervention is that an arrangement of this kind amounts to an illegitimate fiscal subsidy, which is achieved at the direct expense of those workers and employers who conform to the normal model of payments of income tax and national insurance contributions. If there is to be a subsidy, the Treasury insists on the right to direct it: 'by tackling avoidance activity such as this, the Government will be able to target its support for small businesses more effectively towards those who are creating wealth and employment' (12).
Where fiscal considerations arise, legislation which overrides a contractual allocation of risks can be justified on the grounds that the interests of third parties (other taxpayers) are directly at stake. Where the focus is on labour law, it may be less clear that this is the case, and the argument for non-interference may therefore be more compelling. In Abbey Life Assurance Co. Ltd. v. Tansell (13), the complainant had set up his own personal service company for the supply of his services as a computer consultant. On this occasion, his personal service company (Intelligents) contracted with an employment agency (MHC) to supply his services to the end user, Abbey Life (14). The contract between Intelligents and MHC had the effect of placing the complainant 'under the control of Abbey Life' as part of a team working on the impact of the 'millenium bug' on Abbey Life's computer systems. About five months into this arrangement, Abbey Life terminated its use of the complainant's services, shortly after he had been diagnosed as suffering from diabetes. He brought claims against both MHC and Abbey Life for disability discrimination.
The EAT and the Court of Appeal held that Abbey Life, but not MHC, owed the complainant obligations of non-discrimination under section 12 of the Disability Discrimination Act. This provision, which mirrors similar provisions in the Sex Discrimination Act 1975 and the Race Relations Act 1976, imposes the equal treatment obligation on a 'principal' to whom the labour of a 'contract worker' is supplied under the terms of a contract entered into between the principal and another person. The normal case which the Act appears to have in mind is that in which the contract is made between the user and the entity which supplies the worker: a 'principal' is defined as 'a person ("A") who makes work available for doing by individuals who are employed by another person who supplies them under a contract made with A' (15). Here, there was a contract between Abbey Life and MHC, on the one hand, and between MHC and Intelligents, on the other; however, there was no contract between Abbey Life and Intelligents. This did not deter the Court of Appeal. According to Mummery LJ, 'the statutory definition only requires the supply of the individual to be "under a contract made with A". It does not expressly stipulate who is to be the party who contracts with A'. This interpretation, he said, was 'achieved by a conventional process of judicial construction of legislation', while at the same time enabling the court to give 'the wide ranging provisions of the discrimination legislation a generous interpretation' (16). Putting the matter more forcefully, Morison J., in the EAT, noted the argument of counsel that a contrary interpretation would 'make it easy to evade the wide coverage which the Act was intended to achieve. It would, as [counsel] submitted, be all too simple to insert another contract somewhere along the line'. The correct approach was to have regard to 'the general principle which applies in social legislation of this kind, namely that the statute should be construed purposively, and with a bias towards conferring statutory protection rather than excluding it' (17).
This is a frank acknowledgement of the limits of a contractual analysis. The remarkable feature of section 12 of the 1995 Act (and the parallel provisions governing sex and race discrimination) is that it treats the user as if it were the employer, even though there is no contract of any kind between the user and the worker whose labour is being supplied. Tansell follows a line of earlier decision, s, including BP Chemicals Ltd. v. Gillick (18), Harrods Ltd. v. Remick (19) and, most recently, Patefield v. Belfast City Council (20), in which versions of the equal treatment principle were applied to the end user of labour. The justification for doing so must lie in the priority accorded to the principle of equality of treatment: the legislation rests on the assumption that this principle is so fundamental as to override considerations derived from the contractual allocation of the rights and duties of the parties. Thus one implication of Tansell is that the applicants in Bearman may well have had rights under the 1995 Act to be treated equally with their co-workers, rights which were not tested in that litigation.
The result in Tansell is all the more impressive because of the clear evidence that the complainant entered into this arrangement with full awareness of its implications: '[t]he employment tribunal found that Intelligents was not a sham company but formed under advice before the applicant embarked on selling his computer knowledge and skills through agencies' (21). The EAT suggested that computing consultants are 'required' to contract through personal service companies 'because of the risks that they will incur if they were to make a negligent mistake in the course of carrying out their duties. The consequences of such a mistake could be enormous and sufficient to cause the individual to become bankrupt' (22). It is not clear whether the fiscal advantages of personal service companies were also make known to the courts. Nevertheless, there is an unavoidable sense in which the complaint was, in this case, able to obtain some of the advantages of employee/worker status, while avoiding some of the normal disadvantages.
The exceptional nature of the Tansell judgment is also apparent if it is compared with other decisions testing the scope of the Disability Discrimination Act. The Act contains a significant exclusion, in that it does not apply to any employment in which the employer has fewer than 15 employees. In Colt Group Ltd. v. Couchman (23) the complainant was employed by the parent company of a group which had 40 subsidiary companies operating worldwide. The parent company, however, employed only seven persons. The EAT ruled out the claim on these grounds (at that time, the relevant threshold was twenty employees), noting that, as a matter of company law, a corporate group has no legal identity as such: [t]he legal persons or bodies are the individual companies that make up that group' (24). Likewise, in Hardie v. C.D. Northern (25) the EAT ruled that the term 'employer' in the 1995 Act could not be extended to refer to 'associated employers' in the absence of a clear statutory intention.
Tansell is also exceptional from the point of view of other cases concerning agency labour. In Costain Building & Civil Engineering Ltd. v. Smith (26) the EAT ruled that an agency-supplied worker who was appointed by his trade union to be a safety representative on the site at which he worked had no basis for claiming unfair dismissal under section 100(1) of the Employment Rights Act when his employment was terminated at the behest of the user, since his only contract was with his agency.
From the wider perspective of this body of recent case law, then, Tansell seems to be the exception which proves the rule. The default position taken by the courts is one of respect for the pre-existing arrangements made by the parties, whether these relate to supply through intermediaries or the use of corporate group structures. It is only in those few instances where a specific statutory provision extends employment rights in across contractual or capital boundaries - as in the case of the statutory provision relied on in Tansell - that they will depart from this approach.
One factor in the reluctance of the courts to move beyond their current approach is, undoubtedly, the prevailing attitude towards statutory interpretation, which regards reasoning by analogy from one statutory context to another as illegitimate. Thus, as Couchman shows, the broad approach to defining the rights of 'contract workers' taken under equal treatment legislation cannot be extended to the parallel problem of defining corporate group liabilities, while Costain v. Smith tells us that what is good for disability discrimination has no application to employee representation. However, just as the reluctance of the courts to reason by analogy from statute to the common law is now decreasing (27), so it is possible to envisage a greater willingness on their part to entertain analogical reasoning between different statutory contexts. If this impression is correct, it becomes important to consider whether a unifying set of underlying principles can be stated in such a way as to lend coherence to the court's task of attaching the liabilities of the employer to a particular entity or entities, so that, if only in cases where legislation is currently unclear, a purposive approach of the kind adopted in Tansell can be taken further. More generally, the identification of such an approach could also lend legitimacy to certain types of legislative intervention.
From this point of view, three criteria for identifying the employer suggest themselves, which can be characterised in terms of 'coordination', 'risk' and 'equity'. The criterion of coordination associates the concept of the employer with the exercise of powers of centralised management. On this basis, the scope of employer liability would be determined by reference to the presence of managerial control. In an economic sense, the enterprise is defined by reference to the existence of an implied 'authority relation' which grants the employer a certain discretion to direct the factors of production, including labour, without the need for express recontracting. 'Outside the firm, price determines the allocation of resources, and their use is coordinated through a series of exchange transactions on the market. Within the firm, these market transactions are eliminated, and the allocation of resources becomes the result of an administrative decision' (28). Both the 'control' test, with its emphasis on the worker's personal subordination to the discretionary powers of the employer, and the common law implied terms of obedience and fidelity, can be understood as expressing in legal terms this sense of the enterprise as a space within which managerial control is exercised: 'it is the fact of direction which is the essence of the legal concept of "employer" and "employee", just as it [is] in the economic concept' of the enterprise or firm (29).
However, it is not simply for the employer's benefit that labour law associates the employer with the exercise of managerial power in this way. The reach of health and safety laws, and of laws governing employee representation, are also to a certain extent defined by this idea. Thus in both these contexts, the employer's obligations have from time to time been extended beyond the immediate contractual nexus, by reference to notions of associated employers and group-level liabilities (30). Here, then, the notion of the employer is extended as widely as possible in order to ensure that the scope of employee protection is coterminous with the exercise of centralised managerial coordination.
The criterion of 'risk' forms a complementary basis for identifying the enterprise. The underlying idea is that the enterprise operates as a mechanism for absorbing and spreading certain economic and social risks, including the risks of unemployment, interruption to income, and work-related injury and disease. The mechanisms for spreading risks in this way include the income taxation and national insurance contribution systems. These focus on the employment relationship as the conduit through which payments are channelled and risks are pooled. Relationships which fall outside the scope of the employment relationship are subject to different tax and social insurance regimes.
In the traditional labour law conception of the employer, the 'coordination' and 'risk' functions were united. The law granted the employer wide implied powers of managerial coordination, in return for imposing upon it the 'social' obligations of risk shifting and pooling. As Alain Supiot has put it, '[u]nder the model of the welfare state, the work relationship became the site on which a fundamental trade-off between economic dependence and social protection took place. While it was of course the case that the employee was subjected to the power of another, it was understood that, in return, there was a guarantee of the basic conditions for participation in society.' He goes on: 'it is the very foundations of this compact which are now being called into question: economic pressures are stronger than ever (both for those who are in work and for those who are not), but they are no longer compensated for by security of existence’ (31).
The fragmentation of the concept of the employer is a particular symptom of this process. The problem with agency work and with the supply of labour through intermediaries is precisely that the 'coordination' and 'risk' functions of the employer are now split between different entities: the 'coordination' function vests in the end user of labour, while the residual 'risk' function is left with the agency or with the individual worker. Most obviously, this affects the worker, who is unable to assert normal claims for the protection of income and employment against the user. In British labour law, at least, the worker's rights against the agency are also likely to be insubstantial, thanks to the apparent presumption that agency workers lack the necessary mutuality of obligation to establish a contract of employment (32). Legislation deems the worker to be an employee of the agency for tax and social security purposes (33), but this is a device aimed above all protecting the integrity of the tax base, rather than ensuring that agency workers have access to social security rights.
What is less often noticed is that, from the user's point of view, the absence of a clear contractual nexus may make it difficult to establish the existence of duties of implied obedience and fidelity (34). This makes agency work unsuitable for the delivery of certain 'core' services which depend upon the worker making an open-ended commitment of loyalty. Contracts for the hiring of labour through intermediaries attempt to bridge this gap by making express provision for the worker to come under the 'control' of the user, and for the worker to undertake obligations of fidelity and confidentiality. This was the case in Tansell. However, such arrangements are of limited utility. Precisely because there is normally no contract between the user and the worker, these terms have to be expressed in the contract of hiring between the user and the agency; as a result, they are unlikely to bind the worker directly, thanks to the doctrine of privity of contract. Even if such terms can bind the worker, express terms of this kind are unlikely to provide the user with quite the same residual power that the common law implied terms, which are necessarily more diffuse and imprecise, can supply.
One route which can be taken by labour law in response to the phenomenon of labour supplied through intermediaries is to insist on the unity of the employer as a legal concept by banning the use of agency labour completely, or very severely restricting its use, if necessary through penal sanctions. Until very recently, this was an option adopted in several mainland European labour law systems (35). This option is potentially unattractive if only because it rules out certain uses of the agency form which might well be legitimate in the sense of being beneficial to all the parties concerned. If, as has always been the case in the UK, agency labour is not, per se, to be regarded as illegal, what options remain? One route is for the legislator to ensure that, if the 'risk' and 'coordination' functions of the employer are to be split in this way between the agency and the user, the obligations which would normally attach to the exercise of these functions are to be imposed upon the relevant parties in each case. On this basis, the agency should be under an obligation to provide continuing employment to the worker and to provide access to pension and occupational employment benefits, as is the case in certain mainland European systems. While UK legislation has not yet gone this far, proposals to adopt regulations adopted under the Employment Relations Act 1999 (36) mark a limited step in this direction by regularising certain aspects of the agency labour relationship. It remains to be seen whether the courts will take this approach further by relaxing the presumption that, in the absence of clear contractual intent, the agency relationship lacks mutuality. Conversely, the user should clearly assume the normal protective functions of the employer with regard to the coordination of work, in particular the maintenance of health and safety.
It is relevant that, in other contexts, labour law is highly intolerant of attempts to break the unity of 'risk' and coordination'. This, after all, is what lies behind the compulsory novation of employment contracts under TUPE. The Litster decision (37) is a particularly strong affirmation of this idea: the transferee, when it assumes the managerial function which goes with the ownership of the business as a going concern, cannot at the same time displace the economic risks of the resulting reorganisation on to the workforce by leaving them with limited claims against an insolvent transferor. The limits of TUPE are set by a further juridical attempt to define the boundaries of managerial coordination. The idea, now contained in the revised version of the Acquired Rights Directive, of 'an economic entity which retains its identity' following the transfer could be said to represent the smallest unit within which it is feasible for centralised managerial control to be exercised. In other contexts, as we have seen, the law may need to define the largest feasible unit of managerial control: the transnational corporate group which forms the basis for the application of the European Works Councils Directive (38) is a particularly clear illustration of this.
A third way of defining the employer which goes beyond a purely contractual approach is by reference to the idea of 'equity'. This refers to the identification of the enterprise with a space within which the principle of equal treatment must be observed. In the same way as the employer assumes the responsibility for pooling and spreading economic risks, the obligation to respect the equality principle is imposed as a counterpart to, and is co-terminous with, the power of managerial coordination. This supplies the justification for imposing the equal treatment requirement on the end user in the case of agency labour. In other words, it is the organisational unity of the enterprise which makes it legitimate, in the Gillick, Remick, Tansell and Patefield line of cases, for contract workers to be accorded equal treatment rights with the permanent workers alongside whom they work. A similar justification can be offered to extending the equal pay principle across contractual boundaries by invoking, again, the notion of associated employers.
This conception of the enterprise has been tested to breaking point in the context of transfers of employment. North Yorkshire CC v. Ratcliffe (39) establishes that workers whose terms and conditions are reduced in order to ensure that their employer beats off a lower-paying rival, in order to win a round of competitive tendering, can then seek to restore their pay by bringing an equal pay claim using fellow workers in the same employment as comparators. In Lawrence v. Regent Office Care Ltd. (40), the EAT held that where it is the external competitor which succeeds in the tendering process, the workers, once transferred to the employment of the competitor as a result of TUPE, lose the right to equality with their former colleagues in the transferor: 'there is nothing about this case which would distinguish it from any other case where an applicant claimed equal pay with a comparator employed by another company' (41). But this apparently clear cut conclusion may not withstand the effects of the equality principle as it is articulated in EC law, under Article 141 of the EC Treaty. On appeal in Lawrence (42), the Court of Appeal decided that it could not rule out the application of Article 141 to the transfer of employment in that case, and referred the matter on a preliminary reference to the ECJ. If the ECJ were to rule in favour of the applicants in this case, it could perhaps do so only by going beyond the idea that the equality principle is confined to the boundaries of the enterprise. Another way of looking at this case, though, is that certain organisational ties remain after the transfer. This is not simply the effect of the employment history of the applicants, but also because the applicants' previous employer, the council, retained a residual responsibility for ensuring the delivery of the services in question.
This commentary has examined recent case-law in which courts and tribunals have grappled with the issues which arise from the so-called 'fragmentation' of the enterprise through sub-contracting and outsourcing and the use of intermediate forms of labour supply. While the courts are very far from reaching a unified approach to the different contexts in which this problem arises, there are nevertheless signs that techniques have been developed to counter some of the more negative implications, for employment protection and representation, of these developments. In the process, a more distinct legal conception, or, more accurately, series of conceptions, of the employer is emerging. Three overlapping and complementary criteria for identifying the employer are suggested: coordination, which associates the employment unit (or parts of it) with the presence of centralised managerial control; risk, which treats the employer as a juridical form for absorbing, processing and spreading social and economic risks which are associated with the enterprise; and equity, which sees the employment unit as a space within which the equal treatment principle, in its various forms, should be observed. Depending on the particular context, one of these criteria tends to predominate over the others. The problems of classification for the law are greatest in those cases, agency work being the most significant, where the separate functions of the enterprise are split among different entities. Even here, the law can limit the negative effects of fragmentation by seeking to attach certain obligations to one or other of the entities concerned. If this principle were pushed to its limit, the law would attach to the user those obligations whose rationale operates in relation to the function of coordination (such as health and safety and employee representation at the workplace) while the agency would bear the residual risks of insecurity of employment and interruption to earnings. This is an option which may increasingly commend itself to policy makers, as well as to courts which are aware of the dangers of contractual constructions which deflect the protective purpose of social legislation.
Notes:See H. Collins, 'Independent contractors and the challenge of vertical disintegration to employment protection laws' (1990) 10 OJLS 353-80, and 'Ascription of legal responsibility to groups in complex patterns of economic integration' (1990) 53 MLR 731-44.
2For an overview of the relevant law and a discussion of the notion of the employer with reference to the concepts of 'undertaking' and 'establishment', see S. Deakin and G. Morris, Labour Law (London: Butterworths, 2nd. ed., 1998), ch. 3, in particular at pp. 208-231.
3See Employment Rights Act 1996, s. 230(1), (2); Trade Union and Labour Relations (Consolidation) Act 1992, ss. 295-96.
4 AC 1014.
5 IRLR 485.
6The Transfer of Undertakings (Protection of Employment) Regulations, SI 1981/1794.
7 IRLR 431.
8 IRLR 431, 433.
9 IRLR 431, 434.
11HM Treasury, Explanatory Notes to the Finance Bill 2000, 'Clause 59 and Schedule 12: Provision of Services through Intermediary', at p. 1 (http://www.treasury.gov.uk). Section 59 and Schedule 12, as enacted, introduce the so-called IR 35 scheme of deduction at source for services supplied through intermediary companies.
13 IRLR 387; see also the judgment of the EAT sub. nom. MHC Consulting Ltd. v. Tansell  IRLR 677.
14 IRLR 387, 388.
15Disability Discrimination Act 1995, s. 12(6). The parallel provisions to s. 12 in the case of sex and race discrimination are, respectively, Sex Discrimination Act 1975, s. 9 and Race Relations Act 1976, s. 7.
16 IRLR 387, 390.
17 IRLR 677, 679.
18 IRLR 128.
19 IRLR 9 (EAT), 583 (CA).
20 IRLR 664.
21 IRLR 677, 677.
22 IRLR 677, 678.
23 ICR 327.
24 IRLR 327, 331.
25 ICR 207.
26 ICR 215.
27See J. Beatson, 'Public law influences in contract law', in J. Beatson and D. Friedmann (eds.) Good Faith and Fault in Contract Law (Oxford: OUP, 1995), and, for a discussion specific to labour law, S. Deakin, 'Private law, economic rationality and the regulatory state', in P.B.H. Birks (ed.) The Classification of Obligations (Oxford: OUP, 1997), at pp. 294-303.
28R.H. Coase, 'Industrial organisation: a proposal for research', reprinted in Coase, The Firm, the Market and the Law (Chicago: University of Chicago Press, 1988), at p. 62.
29Coase, 'The nature of the firm', reprinted in The Firm, the Market and the Law at p. 54. Coase cites here to the textbook by F.R. Batt, The Law of Master and Servant (London: Pitman, 1929). It is interesting that, in formulating his (highly influential) economic theory of the firm, Coase should have drawn on a legal formulation of the master-servant (or employment ) relationship.
30For discussion of the concept of associated employers and its application to various contexts within labour law, see Collins, 'Ascription of legal responsibility', op. cit.; Deakin and Morris, Labour Law, at pp. 213-217.
31A. Supiot, ‘Préface’, in A. Supiot (ed.) Au-delà de l’emploi: Transformations du travail et l’avenir du droit du travail en Europe. Rapport pour la Commission Europeénne (Paris: Flammarion), at pp. 7-15, at p. 10.
32See e.g. Wickens v. Champion Employment Agency  ICR 365, although cf. McMeechan v. Secretary of State for Employment  IRLR 353.
33See Income and Corporate Taxes Act 1988, s. 134; SI 1978/1689.
34J. Rubery, J. Earnshaw, M. Marchington, F. L. Cooke and S. Vincent, 'Changing organisational forms and the employment relationship', paper presented to the 'Future of Work' seminar, UMIST, 16 November 2000.
35For an account of the Italian law of 1997 which lifted the previous restrictions on agency work but which also imposed a number of protective obligations which go far beyond those which currently operate in the UK, see F. Liso and U. Carabelli (eds.) Il lavoro temporaneo. Commento alla legge n. 196/1997 (Milan: FrancoAngeli, 1999).
36See DTI, Regulation of the Private Recruitment Industry: A Consultation Document, September 1999 (http://www.dti.gov.uk).
37Litster v. Forth Dry Dock & Engineering Co. Ltd.  IRLR 161.
38Council Directive 94/45, transposed into UK law by SI 1999/3323.
39 ICR 839.
40 IRLR 149.
41 IRLR 149, 154.
42 IRLR 608.