The Employment Status of

Individuals in Non-standard
Employment

Brendan Burchell,* Simon Deakin** and Sheila Honey***

* Faculty of Social and Political Sciences and ESRC Centre for

Business Research, University of Cambridge

** Faculty of Law and ESRC Centre for Business Research,

University of Cambridge

*** Department of Trade and Industry

March 1999

Acknowledgments

We wish to acknowledge with thanks the assistance and advice of a number of

colleagues in the completion of the work for this report. For the DTI, Mark Cully and
Paul Teasdale, provided much helpful advice and comment thoughout the project.
For the Cambridge team, Roy Mankelow, Ines Wichert and Frank Wilkinson each
played a major part in planning the qualitative stage of the research and in carrying
out the interviews and focus groups. Laura Scamponi also carried out a number of the
interviews and, with Stephanie Auge, assisted in the analysis of the interview
transcripts. Hilary Taylor took part in organising and running the focus groups.
Catherine Barnard gave advice on the legal assessments of employment status. We
would also like to thank Bernard Casey of the LSE for his comments on an earlier
draft of the report.
Contents
Executive Summary i

Chapter 1. Introduction 1

1.1 Employment status and 'non-standard' work 1

1.2 Extending employment legislation: the concept of 'worker' 2
1.3 Practical implications of employment status 2
1.4 Aims and objectives of the research 3
1.5 Methods 3

Chapter 2. The Law Governing the Classification of Employment 5

Relationships

2.1 The four common law tests of employee status 5

2.2 Factors taken into account by the courts 10
2.3 ‘False’ self-employment 12
2.4 Mutuality of obligation and continuity of employment 13
2.5 Fixed term employment, employment for a specific task, and contracting 15
out of employment protection legislation
2.6 Legislative extensions to the categories of protected workers 16
2.7 A note on terminology 18
2.8 Conclusion 20

Chapter 3. The Quantitative Wave: Aims, Methods and Main Findings 20

3.1 Introduction 20
3.2 Background 20
3.3 Methods 24
3.4 Economic Activity of Sample 26
3.5 Comparisons with the Labour Force Survey 27
3.6 Perceptions of status 27
3.7 Self-employment 30
3.8 In non-standard work? 30
3.9 Conclusion 35

Chapter 4. Clarifying Employment Status: Quantitative Analysis 36

4.1 Clarifying employment status 36

4.2 Common law tests of employment status 38
4.3 Estimating Numbers of ‘Workers’ 43
4.4 Qualifying for employment protection rights on the basis of continuity of 47
employment
4.5 Reasons for working in non-standard arrangements 48
4.6 Conclusions 48
Chapter 5. The Qualitative Wave: Aims and Methods 50

5.1 Aims of the in-depth interviews 50

5.2 Achieving the sample for in-depth interviews 51
5.3 Data Analysis 52
5.4 Characteristics of the sample 53
5.5 Filters for ‘non-standard’ employment 55
5.6 Test-retest reliability of the dependency questions 57
5.7 Conclusions 60

Chapter 6. Employment Status in Practice: Findings of the ‘Qualitative 62

Wave’

6.1 Assessments of employment status 62

6.2 Cases classed as ‘unclear’ by the quantitative analysis but classified 68
clearly by the qualitative analysis
6.3 Cases classed as ‘borderline’ by the qualitative analysis 70
6.4 ‘Clear’ discrepancies 74
6.5 Identifying the sources of uncertainty in the application of the legal tests 75
of status
6.6 Conclusion 76

Chapter 7. Employment Status and the Experience of Non-standard Work 78

7.1 Reasons for choosing particular forms of employment 78

7.2 Advantages and disadvantages of forms of employment: perceptions of 79
autonomy and control
7.3 Advantages and disadvantages of forms of employment: perceptions of 81
insecurity and risk
7.4 Making explicit ‘trade-offs’ between employment and self-employment 83
7.5 Conclusion 84

Chapter 8. Conclusions 86

8.1 The coverage of employment protection legislation 86

8.2 Estimating numbers employed as ‘employees’ and ‘workers’ in Great 86
Britain
8.3 Verifying the survey 87
8.4 Identifying sources of uncertainty in the application of legal tests of 88
status
8.5 Employment status and the experience of non-standard work 88

Bibliography 90

Appendix 1 Glossary of Terms 91

Appendix 2 Questionnaire used in Quantitative Wave 93

Appendix 3 Aide mémoire for individual interviews 98

Appendix 4 Codes for the categorisation of information from the interviewee 103
transcripts
Executive Summary

Background

This paper reports the findings of an empirical study into the operation of laws
governing the classification of employment relationships. Access to employment
rights depends to a large extent on whether an individual is employed as an employee.
The self-employed and a number of other groups (such as agency workers) tend to
find themselves excluded from employment protection law. The research was aimed
at discovering how laws relating to the 'status' of employees and the self-employed
work in practice.

The research was prompted by a number of concerns relating to the present law.
Firstly, the criteria by which the courts and employment tribunals distinguish
employees from the self-employed are perceived to be uncertain in their effects.
Secondly, there is concern that the existing classifications fail to reflect the growth of
certain flexible or non-standard forms of employment, in particular casual work, zero
hours contracts, fixed term and task employment and freelancing.

One possible policy response to these concerns would be to reform the definitions
currently in use in employment law. In particular, greater use might be made of the
legal concept of 'worker', which is wider than that of 'employee'. Some self-employed
individuals who are not in business on their own account and who contract to provide
their personal services to another would come under the definition of 'worker'. These
'dependent self-employed' are among the principal groups whose employment status
is currently in doubt.

Objectives

The principal objectives of the research were: to identify the numbers of individuals
in the employed labour force who are employed under various working arrangements
(employees, 'workers', and the self-employed); to estimate of the number of
individuals who might be affected by the wider adoption of the concept of 'worker' in
employment law; to identify the sources of uncertainty in the application of the legal
tests of employment status; to examine the reasons individuals work in non-standard
forms of employment, the nature of the choices made and the constraints on those
choices, and the perceived costs and benefits to individuals of working under
particular arrangements with uncertain legal status; and to compare the survey results
with those of the Labour Force Survey (LFS) and to attempt to develop a set of
questions which could be included in the LFS in future for the purpose of obtaining
more reliable information on the operation, in practice, of the law relating to
employment status.

Methods

A two-fold methodology was adopted. Firstly, a quantitative study was undertaken

based on a representative sample of just over 4,000 members of the workforce in
Great Britain (the 'quantitative wave'). After an initial screening to exclude those who
appeared to be unambiguously employees or self-employed, approximately 1,200
individuals were asked a range of questions about the nature of their employment
relationships.

i
Secondly, a small sub-sample of respondents were revisited in order to gather more
in-depth information in less structured discussions, and to analyse their written
contracts of employment or terms and conditions of employment. Qualitative case-
studies were compiled using a mix of focus groups and individual semi-structured
interviews, with the sub-sample of respondents, 36 of whom were interviewed face-
to-face and 24 by telephone (the 'qualitative wave').

The Quantitative Wave

The quantitative analysis provided two forms of evidence on employment status:

respondents' assessment of their own status, and secondly an assessment based on
their answers to questions which were designed to reproduce the types of factors
which were relevant to the classification of employment relationships by courts and
employment tribunals.

It was found that 86 per cent of those surveyed perceived themselves to be employees
and 13 per cent saw themselves as self-employed. An initial set of questions asked
respondents about the nature of their work in such a way to identify individuals who
were likely to have ambiguous status. On the basis of replies to these questions, it was
estimated that 30 per cent of those in employment had an employment status that, on
first inspection, had elements of uncertainty and was not completely clear. They were
made up of two groups: those defining themselves as self-employed, but who were
not directors or partners in their own business, and who did not employ others; and
those defining themselves as employees who had some type of non-standard working
pattern or classified their jobs as non-permanent.

An estimate of the number of 'workers' in the workforce was then obtained by using
answers to a combination of questions based on the legal tests underlying the 'worker'
concept. It was found that the proportion of all those in employment who could be
regarded as 'workers' varied from 80 per cent to 92 per cent depending on how those
individuals whose status was particularly unclear would be categorised at a tribunal.
Similarly, the group of 'independent self-employed' could account for between around
7 per cent and 19 per cent of those in employment.

This suggests that by using the 'worker' definition, the number covered by
employment rights might be higher than the 87 per cent who believe that they are
employees according to the LFS. The use of the worker definition might protect up to
a further 5 per cent of all those in employment (or about 16 per cent of those whose
employment status is unclear because they are employed in non-standard work of
some kind). It is not possible to be more precise about the numbers who would be
affected by this change because of the difficulties of attempting to assess employment
status without examining each case individually.

The Qualitative Wave

The qualitative wave was used initially to assess the validity of the findings from the
quantitative wave and to evaluate the effectiveness of the questions used. It was found
that a number of respondents had not followed the exact definitions given during the
quantitative wave. It is possible that better survey items might be developed which
would be more reliable in testing for employment status. However, it seems that a

ii
large part of the unreliability of these questions is caused by the inherently complex
nature of the economic relationships which they are attempting to describe.

Overall, findings from both waves of the research suggest that, given the normal LFS
definitions of atypical or non-standard work and the current routing of the questions,
the numbers employed in some groups may be overestimated - for example seasonal
workers - while others may be greatly underestimated - for example the number
employed on fixed term contracts.

Evidence from the qualitative analysis shows that the one of the main causes of
uncertainty in the application of the tests for identifying an 'employee' is the legal test
of 'mutuality of obligation'. The test causes particular uncertainty for casual workers
and workers on zero-hours contracts.

Further uncertainty derives from the widespread use by employers of contractual

documentation containing 'waiver' or 'relabelling' clauses which purport to determine
employment status. Contractual documentation, in particular for agency workers, also
reflects attempts to draft terms and conditions in such a way as to minimise the
likelihood that a court would find that the individual was employed as an employee
('status-denying clauses').

The case studies suggest that while individuals in non-standard employment exercise
a significant degree of choice over the form of their employment, several factors are
widely perceived as conditioning that choice: these include the need to fit in with
family arrangements; the costs of retraining following time spent out of full-time
work; the time and complexity of setting up a business; discrimination by reference to
age, gender and disability; and the lack of availability of regular work.

The research also suggests that the legal division between employment and self-
employment does not correspond to perceptions of a clear divide between these
different forms of work on the part of many individuals in non-standard employment.

iii
Chapter 1. Introduction

1.1 Employment status and 'non-standard' work

The issue of 'employment status' is centrally important to the administration and

implementation of employment law. The classification of workers according to
whether they are employees, who are employed under a contract of employment, or
independent or self-employed workers, who may work under a contract for services or
one of a number of other arrangements, determines, at an initial level, the application
of rights and obligations under legislation concerning employment protection, social
security, and taxation (Deakin and Morris, 1998: ch. 3). However, a high degree of
uncertainty attaches to the legal criteria by which workers are classified. This means
that the effectiveness of the law, in practice, may be called into question.

Moreover, there is a perception that the existing classifications have become too rigid
to deal effectively with the growth of non-standard forms of employment, that is to
say, those forms of work which depart from the model of the 'permanent' or
indeterminate employment relationship constructed around a full-time, continuous
working week. The result is said to be the exclusion of certain groups of workers,
whose status is unclear, from the protection of employment legislation (Hepple, 1986;
Leighton, 1986; Deakin, 1986, 1988; Collins, 1990).

These questions have been the subject of numerous legal and policy-orientated
analyses over the past 20 years. From sociological and economic perspectives, a large
number of studies of labour-use strategies have examined employers' reasons for
using flexible forms of work (e.g. Atkinson, 1985; Pollert 1988; Wood, 1989, Hakim,
1990; McGregor & Sproull, 1991, Hunter et al., 1993; Beatson, 1995; Rubery and
Wilkinson, 1995; Casey et al., 1997). Aside from the work of Leighton (1983, 1984),
very few empirical studies have directly addressed the impact of legal classifications
on the form in which labour is contracted. This report presents findings from a
quantitative and qualitative study of the operation in practice of the rules relating to
employment status, with the focus on the perceptions and experiences of individuals
whose employment is, in some respect, 'flexible' or 'non-standard'.

1.2 Extending employment legislation: the concept of 'worker'

One possible means of increasing the number of individuals covered by employment

law is to use the broader definition of 'worker' in preference to that of 'employee' as a
basis for determining the scope of protective legislation. This would have the effect
of including certain individuals who do not have a contract of employment, but who
nevertheless contract to supply their own personal services to the employer and who
to some degree are economically dependent on the employer's business, that is to say,
they derive a high proportion of their income from this particular employment. This
wider definition of dependent labour is used for the purposes of equal treatment
legislation and legislation concerning the payment of wages, as well as for certain
limited purposes related to employment protection legislation; it has also recently
been adopted under the National Minimum Wage Act 1998 and under regulations
implementing the Working Time Directive.

1
Notwithstanding the growing adoption in legislation of the concept of the 'worker',
many aspects of its use remain unclear. In contrast to the voluminous body of case-
law which exists on the question of the distinction between an 'employee' and a self-
employed worker, there are very few decisions on the distinction between a dependent
'worker' and a genuinely independent self-employed individual. It is not entirely clear
which criteria the courts will apply in determining where this line is to be drawn, nor
how they will weigh competing factors in the balance (as they currently have to do
when applying the tests of 'employee' status). Partly for this reason, and partly
because of a lack of empirical research on this question, it is not known how many
individuals would be affected by an extension of employment protection rights to
'workers' as opposed to 'employees'.

1.3 Practical implications of employment status

The difficulties associated with defining employees, workers and the self-employed,
and the problems of assigning status to those in non-standard work, have a number of
practical implications for the operation of the law in practice.

· While some degree of uncertainty in the operation of the law in this area is
probably unavoidable, a situation in which a substantial proportion of the workforce is
unsure as to its legal position would give rise to concern.

· In particular, the ambiguities surrounding the distinction between employees

and self-employed mean that there may be many individuals who have employment
rights without being aware of this. Conversely, individuals who may think that they
are employees, and therefore have access to employment rights, may turn out not to
be on closer legal inspection.

· Given uncertainty in the law, some employers may avoid their obligations by a
number of means including exploiting their workers' lack of awareness about their
entitlements.

· The official source of information on the labour market is the Labour Force
Survey (LFS). This gives a measure of self-employment but it is based on self-
reporting - that is the individual's perception of their status. However, self-reporting
may not correspond to the legal classification. Using the LFS alone, then, it is
difficult to assess whether reported changes in the labour market amount to changes in
the number of workers who are protected by the law.

· In addition the LFS does not provide enough information to estimate the
numbers of individuals in the economy who might be 'workers' and so makes
evaluation of the effects of legislation which is applied to workers highly problematic.

2
1.4 Aims and objectives of the research

As a result of these practical implications arising from the issue of employment status,
it was clear further research was needed. This project aimed to address a number of
specific issues arising from the points just discussed. The principal objectives were:

· To identify the numbers of individuals in the employed labour force who are
employed under various working arrangements (employees, 'workers', and the self-
employed).

· To estimate of the number of individuals who might be affected by the wider

adoption of the concept of 'worker' in employment law.

· To identify the sources of uncertainty in the application of the current legal

tests of employment status.

· To examine the reasons individuals work in non-standard forms of

employment, the nature of the choices made and the constraints on those choices, and
the perceived costs and benefits to individuals of working under particular
arrangements with uncertain legal status.

· To attempt to develop a set of questions which could be included in the LFS

for the purpose of obtaining more reliable information in future on the operation in
practice of the law relating to employment status.

1.5 Methods

To tackle these research questions required an approach which combined a large

sample (to address issues of prevalence) and also highly detailed analyses of the
working lives of a (necessarily) much smaller sample of individuals. This was
achieved by a multi-method approach.

The first wave of data was collected from a representative sample of 4,006 members
of the British workforce, as part of an omnibus survey. After an initial screening to
exclude those who were unambiguously employees or self-employed, 1,182 were
asked a range of questions about the nature of their employment relationships. The
second wave of data collection involved re-visiting a small proportion of those 1,182
respondents to gather more in-depth information in less structured discussions, and to
analyse their written contracts of employment or terms and conditions of employment.
These qualitative case-studies were based on a combination of focus groups and
individual semi-structured interviews, involving altogether 36 respondents.
Telephone interviews were carried out with a further 24 individuals to illuminate
some non-standard working patterns of particular interest which were still not clear
from the face-to-face interviews.

The quantitative data analysis provided two kinds of evidence on the employment
status of individuals. Firstly, it offered evidence of individuals' own assessments of
whether they were employees or self-employed, or had some other status such as that
of a trainee on a government training scheme. Secondly, individuals' answers to a

3
series of questions concerning the nature of their work provided some indication of
whether they satisfied the tests and factors by which the courts and tribunals judge
employment status. As we shall see in chapters 3 and 4 below, combining these types
of evidence enables us to make an assessment of the numbers of employees, 'workers'
and the self-employed in the employed population as a whole.

The second, qualitative wave of data analysis provided evidence of a third kind,
namely very detailed accounts of individuals' experiences of employment which made
it possible for more complete assessments of their legal status to be made.
Documentary evidence, including individuals' contracts of employment and written
statements of terms and conditions, could be take into account at this stage. The
assessments made involved an element of judgement on the part of the legal members
of the project team, and do not purport to be definitive since it is of course possible
for lawyers to disagree on application of the law to individual cases: so much is clear
from the voluminous case law on this subject. However, the qualitative wave of
research fulfilled a number of important purposes: testing the reliability of the
quantitative analysis; identifying the sources of uncertainty in the application of the
law; and examining perceptions of the law's operation in practice.

Chapter 2 below explains in more detail the nature of the legal tests which are used by
the courts and tribunals to determine issues of status in employment cases. Chapters 3
and 4 then describe the quantitative wave of the research and provide estimates of the
numbers covered by the legal concepts 'employee' and 'worker'. Chapter 5 provides
an account of the methods used during the qualitative wave and chapters 6 and 7
describe its findings. These suggest that uncertainty in classifying non-standard
employment relationships derives to a large extent from difficulties in applying the
test of mutuality of obligation. The contractual and related documentation issued by
certain employers to non-standard workers is also a source of uncertainty, in
particular where employers insert contractual clauses which purport to 'label' the
relationship as one of self-employment, or which have the effect of excluding implied
terms which would assist in establishing employee status. Chapter 8 concludes.

4
Chapter 2. The Law Governing the Classification of Employment Relationships

In this chapter we briefly examine the law relating to the classification of employment
relationships. The law contains various ‘tests’ which can be broken down into a
number of ‘factors’. In practice, courts and tribunals have considerable discretion in
how they apply these tests and factors to individual cases.

2.1 The four common law tests of employee status

The criteria by which legal classifications are determined are not laid down in
legislation, but have been very largely developed through case law. Four tests, in
particular, are widely relied on: ‘control’, ‘integration’, ‘business reality’, and
‘mutuality of obligation’ (see Deakin and Morris, 1998: ch. 3.4). Behind these tests
lies a set of ‘factors’, such as the method of payment chosen by the parties, the length
and stability of the employment relationship, and the degree of coverage of
disciplinary and grievance procedures. The weight which courts attach to any
particular factor appears to be very much a matter of discretion, in part because lower
courts are only subject to review if they commit errors of law in their identification of
the relevant tests or in their application.

2.1.1 Control

Of the common law tests, the control test is the most traditional, with roots going back
to at least the nineteenth century (for an account of the evolution of the control test,
see Deakin, 1998). Its use has by no means died out. As recently as 1995, the Court
of Appeal suggested that the test to apply in determining whether an individual was an
employee or self-employed was ‘who lays down what is to be done, the way in which
it is to be done, the means by which it is to be done, and the time when it is done?’1

In its classic form, then, the control test represents the idea that an independent
contractor has greater discretion and autonomy over the way in which they carry out
their work than is the case with an employee. However, in so far as the ‘control’ test
is read as meaning that only those workers who are subjected to close supervision and
monitoring can be classified as employees, it has been subjected to trenchant criticism
(in particular by Kahn-Freund, 1951). Some commentators have suggested that ‘the
right of control fails to distinguish employment from self-employment because its
presence is entirely consistent with either type of contract’ (Brodie, 1998: 140). For
these reasons, the test has come to be relied upon less frequently.

2.1.2 Integration

In cases dating from the late 1940s, the integration test was relied on as an alternative
to control which was used to explain, for example, why skilled and professional
workers, with a large degree of autonomy over how and when they did their work,
were nevertheless employees if they worked within large, bureaucratic organisations.
The basis of the integration test is that under a contract of service or employment, ‘a

1
Lane v Shire Roofing Co. (Oxford) Ltd. [1995] IRLR 493, 495 (Henry LJ).

5
man [sic] is employed as part of the business and his work is done as an integral part
of the business; whereas under a contract for services his work, although done for the
business, is not integrated into it but is only accessory to it’.2 Similarly, it has been
suggested that ‘the greater the skill required for an employee’s work, the less
significant is control in determining whether the employee is under a contract of
service’.3

Integration, then, places less emphasis upon the personal ‘subordination’ of the
employee to the employer, and more upon the way in which their work is organised.
The test could be said to be appropriate to situations in which managerial authority is
exercised in a de-personalised way, and subjected to bureaucratic rules and
procedures. The test is arguably of less use in situations where the boundaries of the
organisation are diffuse or unclear, as in the cases of sub-contract or agency labour.

2.1.3 Economic reality

The courts have rarely addressed directly the question of what precisely the common
rules on employment status are trying to achieve. In decisions which would now be
regarded as anachronistic, courts at the turn of the century saw the purpose of the
‘control’ test as distinguishing between groups of workers on the basis of their social
class.4 From the more modern perspective of the economic and social legislation
which has adopted the common law tests of classification, the legal tests can be seen
as determining the incidence of statutory rules whose purpose is broadly protective of
the individual employee. The division between employment and self-employment
therefore involves an assessment of how far certain social and economic risks are to
be shared between employers, workers, and the state. For example, employees have
access to certain social security benefits (claims against the state) and to the
protection of certain expectations of continuing security of income and employment
(claims against the employer), from which the self-employed are excluded. At the
same time, employees do not have the same opportunities as the self-employed to set
off their work-related expenditures against income tax.

Conversely, the imposition of certain legal liabilities upon employers can be seen as
based on the view that the employer is better placed than the employee either to avoid
the risk in question by taking steps to contain or neutralise it (the least-cost avoider
rationale), or to spread the risk through insurance or pricing policies (the best insurer
rationale), possibly in conjunction with the state through the social insurance and
taxation systems. When, therefore, a court or tribunal makes a ruling on the
employment status of a particular individual, it is in effect deciding where the burden
of taking precautions against the risk of a certain type of loss (such as loss of

2
Stevenson, Jordan & Harrison v MacDonald & Evans [1952] 1 TLR 101 (Denning
LJ).
3
Beloff v Pressdram Ltd. [1973] 1 All ER 241, 250; see also Cassidy v Minister of
Health [1951] 2 KB 343.
4
Simpson v Ebbw Vale Steel, Iron and Coal Co. [1905] 1 KB 453, discussed by
Deakin, 1998.

6
employment or income, or interference with health of safety at work) should be
allocated.

Occasionally, the courts articulate the process of classifying employment relationships

in precisely this functional sense of ‘risk allocation’. The ‘economic reality’ test,
which was adopted by the US Supreme Court in a decision of 19465 and which has
had some influence on the English courts since the late 1960s, comes closest to
expressing this idea.6 This test involves ‘looking to see where the financial risk lies,
and whether and how far [the worker] has an opportunity of profiting from sound
management in the performance of his task’.7 It therefore implies a test of economic
dependence, in the sense that employee status is the result of ‘the extent to which the
individual is dependent or independent of a particular paymaster for the financial
exploitation of his talent’.8

The use of the economic reality test may also depend upon the context being
considered. In decisions concerning health and safety the courts tend to take the view
that a broad interpretation should be given to the concept of ‘employee’ in order to
ensure that employers who are assumed to bear the principal responsibility for
maintaining occupational safety do not escape their statutory obligations.9 In other
contexts, however, such as unfair dismissal, it seems that reliance on ‘economic
reality’ is less likely to be decisive in favour of a finding of ‘employee’ status.10

2.1.4 Mutuality of obligation

A more serious difficulty in applying the common law criteria of employee status
derives from the growing use since the late 1970s of the ‘mutuality of obligation’ test.
This has had the effect of excluding from protection workers in casual employment
relationships, where the existence of mutual obligations to provide work (in the case
of the employer) and to accept any work which is offered (in the case of the worker)
is in doubt. Homeworkers,11 agency workers,12 zero-hours contract workers13 and
5
United States v Silk 331 US 704 (1946).
6
Market Investigations v. Minister of Social Security [1969] 2 QB 173.
7
Lane v. Shire Roofing Co. (Oxford) Ltd. [1995] IRLR 493, 496 (Henry LJ).
8
Hall v. Lorimer [1994] IRLR 171, 174 (Nolan LJ).
9
In particular in Ferguson v. John Dawson & Partners (Contractors) Ltd. [1976]
IRLR 376; Lane v. Shire Roofing Co. (Oxford) Ltd. [1995] IRLR 493.
10
See, for example, the unfair dismissal case of Wickens v. Champion Employment
Agency Ltd. [1984] ICR 365.
11
Airfix Footwear Ltd v Cope [1978] ICR 1210; Nethermere (St Neots) Ltd v Taverna
and Gardiner [1984] IRLR 240.
12
Wickens v Champion Employment Agency [1984] ICR 365; Ironmonger v
Movefield Ltd [1988] IRLR 461; Pertemps Group plc v. Nixon, 1 July 1993,
unreported, EAT/496/91.

7
workers in casualised trades or occupations14 have been held to be outside the scope
of protective legislation on these grounds.

The ‘mutuality’ test differs from the ‘economic reality’ test in focusing on the terms
of the individual hiring, rather than on evidence of economic dependence. The
mutuality test looks for formal evidence of subordination in the contract terms
themselves. Since the contract terms in question are often expressed neither verbally
nor in writing,15 but may be implied terms whose existence and effect must be derived
from a process of legal ‘construction’ (or interpretation) with which the parties
themselves (and in particular the individual worker) are unlikely to be completely
familiar, it may be difficult to determine their meaning in advance of a decision by a
court or employment tribunal.

The difference between a bare expectation of continuing work, and a contractual right
to receive it, may be a fine one, as indicated by the leading case of O’Kelly v.
Trusthouse Forte plc. The applicants, who were ‘regular casual’ [sic] wine waiters,16
in the view of the industrial tribunal ‘entered into their relationship with the company
in the expectation that they would be provided with any work which was currently
available.’ However, the tribunal concluded that this ‘was a purely commercial
transaction for the supply and purchase of services for specific events, because there
was no obligation for the company to provide work and no obligation for the
applicants to offer their further services.’17

The application of the mutuality test is particularly significant for workers employed
in non-standard forms of work, since it may mean that individuals who do not have a
business of their own and hence are not genuinely in business on their own account,
but who lack a regular and stable employment relationship with a particular employer,
are effectively left in a ‘grey zone’ between employment and self-employment. Such
individuals are often not in a position to take advantage of the benefits, from the point
of view of the tax and national insurance system, of self-employment, since they may
have few or no business assets and their earnings may be too low or too insecure for

13
Clark v Oxfordshire Health Authority [1998] IRLR 125.
14
O’Kelly v. Trusthouse Forte plc [1983] IRLR 369; Carmichael v National Power
plc [1998] IRLR 301.
15
It may be noted that the statutory obligation upon the employer to supply a written
statement of written particulars of employment applies only to employees
(Employment Rights Act 1996, section 1; Deakin and Morris, 1998: ch. 4.3).
Therefore, there can be no breach of this obligation if a court decides that the effect of
unspoken or unwritten terms is that the individual concerned was not an employee in
the first place.
16
The ‘regular casual’ status of the applicants meant that rather being required to
reapply for work from scratch after each job of work had been completed, they were
kept on a list by the company and given some preference in hiring for jobs as they
came up.
17
See [1983] IRLR 369.

8
them to make significant gains from setting off work-related expenditures against
income from employment. The ‘economic reality’ test would often (although not
always) classify such individuals as employees, but, as we just noted, this test is not
applied as a matter of course; on the contrary, the courts have explicitly rejected the
proposition that casual workers who have no business of their own are, for that
reason, to be regarded as employees of the employers upon whom they are, in
practice, dependent.18

2.1.5 The importance of the statutory context

We have already noted that the economic reality test is more likely to be applied in
favour of employee status in cases involving health and safety. However, the
potentially vulnerable position of workers ‘in between’ employment and self-
employment is exacerbated by a lack of consistency in the application of the tests of
status in the different contexts of tax, social security and employment law. It is
generally recognised that for the purposes of income tax and liability for social
security contributions, the tests for the existence of a contract of employment are
more easily satisfied than in cases falling under employment protection legislation.19
In addition, legislation in the fields of tax and national insurance occasionally deems
individuals to be dependent workers for tax and social security purposes even though
such persons would otherwise be classified as self-employed. This is the case, for
example, with certain categories of construction workers,20 and with agency
workers.21

In this report, we are not directly concerned with the scope of tax and social security
law, nor with the administration of the tax and national insurance systems by the
Inland Revenue and Contributions Agency respectively. Our focus is on the
classification of employment relationships for the purposes of employment law.
However, the classification of workers for tax and national insurance purposes bears
indirectly on the approach taken by the courts and tribunals in employment cases, and,

18
See Wickens v. Champion Employment Agency Ltd. [1984] ICR 365, 371: the law
‘does not include as a necessary element the question whether the individual carries
on a separate business. If it did, then it would follow that a casual worker must
always be employed under a contract of service unless he has his own business and
that, plainly [sic], cannot be the law’. This was a case of an agency worker. See also
O’Kelly v. Trusthouse Forte plc [1983] IRLR 369, in which the Court of Appeal
described as a ‘misdirection of law’ the view that ‘every independent contractor who
is content or able only to attract one client [should] be held to work under a contract
of employment’.
19
On this point, see McKendrick, 1990.
20
Construction workers are subject to a system of classification which strictly
regulates whether they are entitled to be treated as self-employed for the purposes of
income tax. See Harvey, 1995.
21
Income and Corporation Taxes Act 1988, s. 134 (on which see Brady v Hart
(trading as Jaclyn Model Agency) [1985] STC 498; Bhadra v Ellan [1988] STC 239);
Social Security (Categorisation of Earners) Regulations, SI 1978/1689.

9
as we shall see below, also affects the perceptions of employment status which are
held by individuals in non-standard work.

In particular, some difficulties arise from the lack of a precise fit between the status of
individuals for employment purposes and their position under income tax and social
security legislation. For example, the fact that the legislature has intervened to deem
agency workers to be akin to employees for tax and national insurance purposes is not
seen as relevant one way for the other for the general determination of individuals’
status in employment law.22 This means that individuals who are classified as self-
employed under employment law but as employees (or equivalent to employees)
under tax and social security law have none of the potential tax advantages of self-
employed status. Their income tax is deducted at source, there is no chance to set off
work-related expenditures against earnings, and both the individual and their
employer must pay Class 1 national insurance contributions. Nor do they qualify for
protected employee status under legislation governing unfair dismissal, redundancy
compensation, guaranteed pay, statutory sick pay and statutory maternity pay.

2.2 Factors taken into account by the courts

A high degree of uncertainty is introduced into the law by the courts’ practice of
taking a very large range of factors into account when applying the tests of status. For
example, in O’Kelly v. Trusthouse Forte plc,23 the following factors (among others)
were identified by the industrial tribunal as being consistent with the existence of
employee status; in other words, they would support a claim of employee status
without necessarily being decisive on their own. These included: the lack of any
financial investment by the applicants in the company’s business; the payment by the
company of holiday pay and an incentive bonus based on past service; the presence of
control by the company when the waiters were at work; and the fact that they were
paid weekly in arrears with tax and national insurance deducted at source. A second
category contained factors which were regarded as ‘not inconsistent’ with a contract
of employment – in other words, their presence would not necessarily be fatal to a
claim of employee status. These were that the applicants were only paid for work
actually performed; the lack of regular weekly working hours; and their exclusion
from the company’s sick pay and pension schemes. Finally, the industrial tribunal
recognised a category of factors which were inconsistent with employee status, that is
to say, which would count against the applicant being an employee. These were: the
contracts were terminable without notice on either side; the applicants had the right to
refuse work; the company was under no obligation to provide it; the parties
themselves took the view that the applicants, as casual workers, were independent
contractors; and it was also the recognised custom and practice in the industry to treat
casual workers as employed under contracts for services.

The very final factor listed - the custom in the industry - turned out to be a particularly
important influence on the tribunal, which decided that the applicants were self-
employed (and this judgement was ultimately upheld by the Court of Appeal). The
22
See McMeechan v Secretary of State for Employment [1997] IRLR 353, 358.
23
[1984] QB 90.

10
precise weight which a tribunal may give to a particular factor is not easily
determined in advance, since a decision to give greater priority to one such
consideration over another is unlikely, in itself, to give rise to a question of law on
which there is a right and wrong answer; the determination of employment status is
said to be a question of ‘mixed law and fact’ on which tribunals faced with the same
or very similar facts could, legitimately, disagree. It is only if the tribunal makes an
error of law - in the sense of applying the completely wrong test, or arriving at a
conclusion on the facts which is ‘perverse’, in the sense of being a conclusion which
no tribunal could reasonably reach - that an appellate court has the right to intervene
and reverse the judgement.24

Table 2.1 lists the principal factors which the courts take into account when deciding
the issue of employment status, and matches them to the four tests referred to in the
discussion above. Since the courts have repeatedly stressed that they have a wide
discretion in deciding which factors to take into account, the list cannot be regarded as
exhaustive, although it represents an attempt to include those factors which feature
most frequently in the recent case law.

TABLE 2:1. THE RELATIONSHIP BETWEEN FACTORS AND TESTS FOR

CLASSIFYING EMPLOYMENT RELATIONSHIPS

Control duty to obey orders

Integration
Economic reality
Mutuality of obligation
discretion on hours of work
supervision of mode of working
disciplinary/grievance procedure
inclusion in occupational benefit schemes
method of payment
freedom to hire others
providing own equipment
investing in own business
method of payment of tax and NI
coverage of sick pay, holiday pay
duration of employment
regularity of employment
right to refuse work
custom in the trade

24
This approach was confirmed in the O’Kelly case itself, although it can also be
found in numerous earlier decisions.

11
2.3 ‘False’ self-employment

The courts have consistently taken the view that the parties to the employment
relationship cannot decide to ‘opt in’ or ‘opt out’ of the coverage of legislation,
simply by choosing to describe that relationship in a particular way: a ‘label’ will be
disregarded, or at least given very little weight, by comparison with other factors.
Hence, if other considerations (of the kind considered above) clearly point towards
employee status, an agreement between the parties to the effect that the individual is
self-employed will have no legal effect.

However, this is not to say that the parties’ own view of the nature of their
relationship is completely irrelevant. It may be a factor in clarifying an otherwise
unclear situation.25 However, it is only if the other factors are unclear that the
parties’ own view of their relationship will carry much weight, and, even then, the
courts will be aware of the need not to take at face value arrangements which only
benefit one side.26

Nevertheless, there is nothing to prevent the parties voluntarily accepting an

arrangement which, objectively speaking, is one of self-employment.27 Moreover, as
we have just seen, the courts focus closely upon the contract terms themselves as
indicators of employment status, a focus which, as we have seen, has intensified as a
result of the growth in importance of the ‘mutuality’ test. The interpretation of the
terms of the contract is, in the end, a matter for the court and not for the parties, and
hence the court can always exercise a degree of control over attempts illegitimately to
evade the reach of statutory provisions. However, if contracts are drafted clearly
enough, they can be used to increase the likelihood that a particular relationship is
classified one way or another. For example, the employer may seek to avoid a finding
of employee status by stipulating that the worker is not entitled to regular
employment, but only to such employment as the employer chooses, in its discretion,
to provide. This may be used as evidence of a lack of mutuality of obligation, and,
therefore, that the relationship is not one of employer and employee. In practice, then,

25
As in Pertemps Group plc v. Nixon, unreported, 1 July 1993, EAT/496/91.
26
See e.g. Cataraman Cruisers Ltd v Williams [1994] IRLR 386. Any agreement by
the individual to waive his or her protective rights under the Employment Rights Act
1996 or the Trade Union and Labour Relations (Consolidation) Act 1992 is void (see
Employment Rights Act 1996, s 203; Trade Union and Labour Relations
(Consolidation) Act 1992, s 288), but this is little help if the effect of the ‘agreement’
is truly to take the individual outside the scope of this protective legislation in the first
place. See Deakin and Morris, 1998: ch. 3.4.3. See also Brodie, 1998, for a
suggestion that ‘judges will steadily become more willing to respect the parties’ own
categorisation’ at least in cases where the parties perceive the relationship as one of
self-employment or where self-employment is the custom for the trade in question.
27
‘A man [sic] is without question free under the law to contract to carry out certain
work for another without entering into a contract of service. Public policy has
nothing to say either way.’ Calder v H Kitson Vickers & Sons (Engineers) Ltd. [1988]
ICR 232, 250 (Ralph Gibson LJ).

12
the line between (illegitimate) ‘evasion’ and (legitimate) ‘avoidance’ of protective
legislation may be a very fine one.

Nevertheless, if the court finds evidence of what it takes to be ‘evasion’, the

consequences for the employee can be very far-reaching. If a court or employment
tribunal is convinced that the parties have ‘colluded’ in adopting a particular working
arrangement in order to evade tax or national insurance contributions, it is likely to
regard the contract (of whatever variety) as void on the grounds of illegality; since the
individual will not then be employed under a legally enforceable contract of any kind,
he or she cannot qualify for any employment protection rights.28 This places potential
employees in an invidious position, particularly where the employer offers the terms
of employment in question on a ‘take or leave it’ basis. In some cases, it has been
enough for the individual merely to have known that the purpose of the arrangement
was illegal, in order for the court to have regarded the contract as unenforceable.29

In this paper, we will use the term false self-employed to refer to an individual who,
objectively speaking, is an employee, but who, for reasons connected to the evasion of
regulatory legislation, is described as self-employed by themselves and/or by their
employer. The terms borderline employee and borderline self-employed refer to
individuals whose status is, objectively speaking, so unclear that they cannot be easily
classified as being in one group or another.

2.4 Mutuality of obligation and continuity of employment

It should be noted that many employees who are employed in non-standard

employment do not have access to social security or employment protection rights, for
example because their weekly earnings are below the threshold for national insurance
contributions, or because of gaps in their continuous service. The rules on continuity
of employment may interact with the ‘mutuality’ test to deny protection to employees
with irregular working arrangements. Even if an individual can show that he or she
was employed as an employee for the purposes of a particular task or job, or a
particular period of time, they may be unable to show that they have a ‘global’ or
‘umbrella’ contract of employment which spans the gaps between periods of work.30
Where this is the case, they may be unable to accumulate the length of service needed
to claim certain employment protection rights. For example, two years of continuous
service are needed to claim a statutory redundancy payment or to benefit from the
general protection against unfair dismissal, and regular employment of several
months’ duration is normally required in order to acquire rights to statutory sick pay,
statutory maternity pay, and guaranteed pay.31
28
Salveson v Simons [1994] IRLR 52.
29
Davidson v Pillay [1979] IRLR 275; Corby v Morrison [1980] IRLR 218.
30
See Nethermere (St Neots) Ltd v Taverna and Gardiner [1984] IRLR 240. There
seems no reason in principle why there cannot be a contract of employment for the
period of individual hirings, even if, in between hirings, there is no ‘global’ contract
of employment. See Clark v Oxfordshire Health Authority [1998] IRLR 125.
31
See Deakin and Morris, 1998: ch. 3.7, for a full account of these rules.

13
A full consideration of the rules relating to continuity of employment is outside the
scope of this report. However, the existence of the continuity rules is significant for
the wider issue of employment status, precisely because they mean that the granting
of employee status does not, in itself, necessarily have the effect of conferring
employment protection rights upon a given individual. A second level of ‘thresholds’
must be overcome before many of the protective rights which depend upon length and
stability of employment are established.

Given that this is the case, it may be noted that there is currently a degree of
duplication between the initial ‘threshold’ test of mutuality of obligation, which
excludes workers from protection on the grounds that they lack either length or
stability of employment, and the additional threshold of continuity of employment.
The justification for excluding workers at the initial stage on the grounds of the
instability of their employment is unclear. It could be said, for example, that the
calculation of rights which depend on length of service (such as rights in respect of
sickness, lay-off, maternity, holiday pay and notice of dismissal) becomes more
problematic if the employment is highly irregular to begin with.32 However, it must
also be borne in mind that there are many employment protection rights which do not
depend upon length of service at all. These include the various rights not to be
dismissed for an ‘inadmissible’ reason.33 However, these policy factors appear not to
have swayed the courts either way; they have generally taken the view that the
existence of the continuity rules has no bearing on the application of the common law
test of mutuality of obligation at the initial threshold stage.34

32
The assumption that an individual ‘earns’ rights to employment protection in return
for a formal commitment to his or her employer underlies much of the courts’
discussion of mutuality of obligation, although it is rarely articulated precisely in
these terms. If this is the purpose of the mutuality test, then it should be remembered
that the test, in its present form, goes well beyond excluding cases of short-term
hirings; workers with decades of service can lose employment protection rights to
which they would otherwise have been entitled if, before the end of their employment,
their relationship with the employer becomes irregular as a result of a falling-off of
demand. See e.g. Hellyer Bros. Ltd. v McLeod [1987] ICR 526.
33
Employment Rights Act 1996, ss. 100-104; see Deakin and Morris, 1998: ch. 5.5.3.
34
See, in particular, O’Kelly v. Trusthouse Forte plc. [1983] IRLR 369, where the
applicants claimed that they had been dismissed for the ‘inadmissible’ reason of their
membership of an independent trade union.

14
2.5 Fixed term employment, employment for a specific task, and contracting
out of employment protection legislation

The employer’s failure to renew a fixed-term contract of employment upon its expiry
is, by law, a dismissal, which may in the circumstances be unfair and/or give rise to
an obligation to pay redundancy compensation.35 However, if the contract of
employment is for a fixed term of one year or more, the employee may validly agree
to waive any statutory protection against unfair dismissal which would otherwise arise
upon the failure of the employer to renew the contract when the term expires. If a
contract for a fixed term of two years or more is agreed, the employee may, in
addition, validly agree to waive their statutory rights to redundancy compensation
upon the expiry of the term. Certain conditions are attached to the form of the waiver;
in particular, it must be in writing.36 In most other circumstances, waivers of
employment protection rights are legally ineffective.37

The important point to make for present purposes is that these legal provisions do not
take fixed-term employees completely outside the scope of legal protection. Their
status remains that of employees (assuming that this can be established on the normal
grounds), and their period of service is capable of counting towards continuity of
employment. The waiver may only enable the employer to avoid statutory liabilities
in one type of situation, namely where a fixed-term contract comes to an end without
being renewed. The employee cannot validly agree to waive their statutory rights to
unfair dismissal protection and redundancy compensation in other situations, such as a
dismissal prior to the end of the agreed term. However, it should be borne in mind
that an employee must accumulate two years of continuous employment before he or
she can bring a claim of unfair dismissal or redundancy compensation,38 unless they
can show that they were dismissed for an ‘inadmissible’ reason, such as their
membership or non-membership of an independent trade union.39

The position of an employee employed under a so-called task contract – a contract to

complete a specific job or task – is different. In law, the completion of the task in
question discharges the contract automatically – in other words, without a dismissal.40
As a result, the employee concerned has no claim for unfair dismissal or redundancy
compensation merely because the employer fails to offer him or her further work.

35
Employment Rights Act 1996, s. 95(1)(b).
36
Provision for waivers in respect of both unfair dismissal and redundancy
compensation is made by Employment Rights Act 1996, s. 197. The conditions for
the effectiveness of a waiver are considered in detail by Deakin and Morris, 1998:
455-458, 512.
37
Employment Rights Act 1996, s. 203(1).
38
Employment Rights Act 1996, ss. 108 and 155 respectively.
39
On inadmissible and automatically unfair reasons, see Deakin and Morris, 1998:
470-477.
40
Wiltshire CC v. NATFHE [1980] ICR 455.

15
There is no need, then, for the employer to insist upon a waiver clause. However, for
all other purposes the employee is treated as any other employee would be – they
accumulate continuity of employment by virtue of their service and may qualify for a
range of other employment rights.

2.6 Legislative extensions to the categories of protected workers

An extension to the scope of employment legislation has been achieved in certain

contexts through the use of the concept of worker. Under the Employment Rights Act
1996, section 230(3), a worker is defined as ‘an individual who has entered into or
works under (or, where the employment has ceased, worked under) - (a) a contract of
employment, or (b) any other contract, whether express or implied and (if it express)
whether oral or in writing, whereby the individual undertakes to do or perform
personally any work or services for another party to the contract whose status is not
by virtue of the contract that of a client or customer of any profession or business
carried on by the individual’.41 Similar definitions have been used in the context of
recent legislation on the national minimum wage42 and the organisation of working
time.43

The concept of ‘worker’ includes ‘employees’ - those employed under a contract of

employment - but also certain independent contractors, who contract personally to
supply their work or labour to the employer. Impliedly, then, this group includes
workers who have a relationship of dependence with the employer in question, but
without meeting the requirements of employee status; hence they may conveniently
be referred to as ‘dependent self-employed’, although this is not a legal term of art.
This category potentially includes freelance workers, sole traders, homeworkers and
casual workers of various kinds.

Certain other groups have a sui generis status. Agency workers fall into this category.
At common law, it is sometimes said that their contractual position is unique, and falls
outside the normal classifications; in other cases, they have been classified as
employed by their agency under contracts for services, on the grounds that they lack
mutuality of obligation. Under legislation they are regarded as akin to employees for
the purposes of tax and national insurance contributions; they are also covered by
minimum wage protection and by laws governing health and safety and
discrimination. Relevant legislation does not however stipulate what their status
should be for other purposes, including employment protection.44

The growing use of the ‘worker’ concept and the specialised definitions of agency
employment which are used in legislation on minimum wages, working time and

41
See also Trade Union and Labour Relations (Consolidation) Act 1992, s. 296(1).
42
National Minimum Wage Act 1998, s. 54(3).
43
The Working Time Regulations 1998, SI 1998/1833, reg. 2(1).
44
The complex legislation relating to agency workers is summarised in Deakin and
Morris, 1998: ch. 3.5.4. On the specific rules relating to trainees, see ibid., ch. 3.5.1.

16
discrimination demonstrate that scope exists for statutory intervention aimed at
overcoming the rigidity of the common law tests of status.

This is not to suggest that the problem of classifying non-standard employment

relationships has thereby been completely solved. In particular, the criteria by which
‘dependent self-employed’ workers are to be distinguished from those who are
genuinely in business on their own account are not clear.

The extended definition of ‘worker’ offers the possibility of reversing the tendency
towards the exclusion of workers in the above categories. In one sense, it reinstates
the ‘economic reality’ test in place of mutuality of obligation. It does this by re-
focusing attention on whether the worker is in business on his or her own account, as
opposed to being economically dependent on the business of the employer. On this
basis, individuals are less likely to be ‘workers’ in the statutory sense of ‘dependent
self-employed’ the greater the extent to which they have an identifiable business of
their own and hire others to work with them.

From a reading of the aims and form of the ‘worker’ definition, two particular factors
would seem to stand out. The first is whether the individual undertook to supply his
or her personal services, as opposed to the services of a third party or, at a further
remove, a finished product. This is central to the definition of the contract for
services in this context. Although the individual worker does not have to show that
there is mutuality of obligation with the employer, he or she does have to demonstrate
that they had a contract for the supply of their personal services to the other party.
Therefore, if they had the right to substitute another person in their place as the
supplier of labour, there could be scope to argue that they were genuinely in business
on their own account. This is likely to be a cause of uncertainty in the application of
the ‘worker’ concept. As with mutuality of obligation, any test which depends, in the
end, upon the reading of the contract terms will open up the scope for drafting aimed
at widening any potential exclusion.

The second factor is whether there was a sufficient degree of economic dependence
between the two parties. The issue here is how far the applicant was dependent upon a
particular employer or client for the bulk of the work they do, as opposed to having
numerous clients or buyers of their services. For example, in the tax case of Hall v.
Lorimer,45 a key finding was that the respondent, who was a freelance film mixer,
worked in the course of a year for over twenty clients, and had over one hundred
separate engagements. In the view of the Court, it could not be said that he was
dependent on any single one of them for a significant proportion of his income. The
Court of Appeal therefore held that he was self-employed and so classified under
Schedule D for the purposes of income tax. It seems likely that, on this basis, he
would have not have been classified as a ‘worker’ for the purposes of employment
legislation.

Notwithstanding the clarity of the judgment in Hall’s case, an inherent difficulty in

the application of the ‘worker’ concept is that ‘the courts do not have a coherent
vision, let alone definition, of what constitutes a business’ (Brodie, 1998: 143). This

45
[1994] IRLR 171.

17
means that the dividing line between the ‘worker’ who is dependent on the business
others on the one hand, and the independent entrepreneur on the other, may be just as
unclear as the line between employees and the self-employed.

2.7 A note on terminology

The legal concepts of ‘employee’, ‘self-employed’ and ‘worker’ were explained

above. In this report we use a number of other terms which are not legal terms of art,
but which are useful for referring to workers in particular groups. Hence we
sometimes use the term ‘dependent labour’ to refer to individuals who are ‘workers’
in the legal sense just explained. We also use two further expressions to clarify the
distinction between those self-employed who are ‘workers’ and those who are not.

The former group are termed ‘dependent self-employed’ and the latter are termed
‘independent self-employed’.

The relationship between the different categories is illustrated in figure 2.1. This
shows while the two categories of ‘employee’ and ‘self-employed’ are mutually
exclusive, those of ‘worker’ and ‘self-employed’ overlap. The overlapping group, for
whom there is no precise legal term, are those referred to in the report as ‘dependent
self-employed’.

A Glossary of Terms used in the report is contained in Appendix 1.

18
Figure 2.1

Set A - Workers Set B - Self-employed

Employees

Dependent Self-employed

Independent Self-employed

2.8 Conclusion

It has been suggested that although the judges often seem to have a strong ‘intuitive
understanding of what amounts to employment and what does not’, the underlying
reasoning behind their judgments is often ‘difficult to supply’ (Brodie, 1998: 139).
Perhaps, then, the law offers a kind of ‘elephant test’: the contract of employment is
‘an animal too difficult to define, but easy to recognise when you see it’ (Wedderburn,
1986 116). If the courts were to take this approach to its extreme, it would become
impossible to predict how they would decide the status of a particular individual. In
practice, it is only in certain cases – generally those arising out of non-standard
employment relationships – that serious uncertainty arises. What we seek to do in the
remainder of this Research Paper is to use empirical methods to help identify more
precisely the sources of uncertainty in the application of the law. To that end, we turn
next to an analysis of the data derived from the quantitative wave of survey analysis.

19
Chapter 3. The Quantitative Wave: Aims, Methods and Main Findings

3.1 Introduction

As we saw in chapter 2, under employment law the relationship between a worker and
an employer or work provider is generally either one of employment under a contract
of employment (sometimes referred to as a contract of service) or one of self-
employment under a contract for services. This is not a clear cut distinction, however.
The growth of non-standard forms of work has meant that number of people fall
between employment and self-employment has grown. We also saw that the scope of
employment legislation has also been extended in certain contexts through the
concept of ‘worker’, as defined in section 230(3) of the Employment Rights Act 1996.
As explained above, the project aimed to examine the contractual arrangements of
workers in ‘non-standard’ forms of employment, with a view to examining how the
legal tests for employment and self-employment operated in practice. This usage is
not necessarily the same as in some other contexts where the term non-standard
includes part-time employees. Working part-time does not in itself normally detract
from employee status, if other factors clearly point towards it.

The initial survey stage (or quantitative wave) aimed to produce a representative
picture of the prevalence of these non-standard forms of employment relationships
among the working population. The survey used questions in the first instance, to be
compatible with the Labour Force Survey (LFS). This chapter looks at the available
data from the LFS and some of its limitations. The chapter then discusses the survey
methodology and the basic characteristics of the sample interviewed. It looks in detail
at respondents’ perception of their status, perceptions of the permanence of their jobs
and the types, if any, of non-standard employment arrangements they work under.

The survey went on to explore respondents’ employment arrangements in more detail,

using a series of questions to replicate the types of tests that have been used by courts
and employment tribunals in deciding employment status. This enabled us to assess to
some degree whether the perceptions of individuals about their status were likely to
be accurate according to the relevant legal tests. This is reported in Chapter 4.

3.2 Background

3.2.1 Groups affected

There are a number of working arrangements that are particularly prone to difficulties
of establishing status. These are predominantly the forms of working referred to as
‘non-standard’ working arrangements. These include casual work, zero-hours
contracts, homeworking, agency work and so-called ‘borderline’ self-employment
(that is, where an individual is not genuinely in business on their own account but
nevertheless self-employed in common law, and so excluded from most of the rights).
Certain other groups, such as fixed-term contract workers, may have employee status
but be subject to waiver clauses and rules on continuity of employment which exclude
them from employment protection. We look at the groups included in our definition
of

20
non-standard employment in more detail below. The definitions used here are for the
purposes of data analysis and are not necessarily legal definitions (see Chapter 2 for
detailed legal terms and Appendix 1 for a Glossary of Terms).

Self-employment

For many self-employed, it is clear that they are running a business on their own
account since they bear the responsibility for profit or loss, work for many different
clients and are not dependent on another business. However, there are also
individuals who do not have an identifiable business but who work as professionals or
for a number of separate employers or clients. These include freelancers. Their
employment status is difficult to determine and they are referred to here as ‘borderline
self-employed’.

There are also cases of ‘false self-employment’ where either the employer regards the
individual as self-employed to avoid employment protection rights and/or the
employee knowingly declares themselves self-employed for tax and National
Insurance purposes.

Agency workers

With agency work, the agency will normally engage or employ the worker whose
labour is supplied to a third party. In this case the traditional functions of the
employer are split between the agency, which has a number of functions including
payment, and the client or user who benefits from the labour.

It could be argued that the individual is an employee of the agency and, therefore, it is
the agency that should be responsible for ensuring minimum employment standards.
However, the employment relationship between the agency and worker is not always
clear, and depends upon the nature of the terms of the overall contract with the agency
and the specific terms and conditions set out for each assignment. As we shall see
below, it is common for agencies to regard workers as self-employed and to issue
statements of particulars stating that to be the case.

Casual workers, homeworkers and zero-hours contracts workers

Difficulties also arise with workers who have no fixed hours of work, but who may be
called upon by the employer as and when work is available. Many of these workers do
not have contracts of employment and are often regarded as self-employed by their
employer. In addition, qualifying periods of continuous employment exclude many
casual workers since they tend to work for an employer intermittently over a number
of years, without any single engagement exceeding the minimum qualifying period.

Zero-hours contracts are found where workers are not guaranteed hours of work (or,
therefore, earnings) though they may be required to be available for work at any time
or at specified times. Recent research on zero-hours contracts (Cave, 1997) has found
that some employers believe that zero hours workers are not employees and many do
not provide such employees with the same benefits as permanent workers. In terms of
the legal position, until more cases are decided, status of workers on zero hours
contracts is still unclear.

21
Fixed-term contract workers

There are no legal obstacles to employing workers on fixed-term contracts. Many

fixed-term contract workers will be employees. However, as explained in chapter 2
above, a feature of fixed-term employment is that the employee can contract out of
protection on unfair dismissal protection and redundancy compensation which he or
she would otherwise have upon the completion of the contract. Moreover, general
claims for unfair dismissal or redundancy cannot be brought unless the employee has
two years of continuous employment. This means that employers are provided with a
means of avoiding key provisions of employment protection legislation.

Task workers and seasonal workers

As we saw in chapter 2 (Section 2.5), a worker employed for a specific task may be
an employee, but will have no claim for unfair dismissal or redundancy compensation
upon the expiry of the contract since this takes place automatically, upon the
completion of the task in question, without there being a dismissal. Seasonal workers
may be affected by this rule; it is not always clear whether they have a fixed-term
contract or a task contract according to the definitions just given. Seasonal workers
will also most likely lack the necessary two years’ continuity of employment to bring
claims in respect of general unfair dismissal and redundancy. Although certain gaps
between jobs can be counted towards continuity, the gap must be relatively short in
relation to the periods of work on either side.

Other groups

There are a number of other groups that probably fall into the grey area between
employment and self-employment. These include trainees, sub-contractors and those
working in partnerships. However, this project confined itself to the main groups
listed above.

3.2.2 Available Data

The most useful source of information on the groups discussed above, is the Labour
Force Survey (LFS). The LFS is the largest regular household survey in the UK. In
any three month period, a nationally representative sample of approximately 120,000
people aged 16 or over, in around 61,000 households, are interviewed.

The LFS quarter that most closely matches the timing of the interviews for our own
survey is that of Winter 1997/98 when interviewing ran from December through
February. For this quarter 87 per cent of respondents reported that they were
employees whilst 12 per cent said they were self-employed. The LFS also asks all
employees whether, leaving aside their own personal intentions, their main job is non-
permanent in some way. The data show that for the same quarter, 1.7 million people
were in temporary employment, representing 6.5 per cent of all those in employment.
The survey goes on to ask in which way their jobs are not permanent. Details of these
are given in Table 3:1. The largest group by far, are those on fixed term contracts.
They make up half of those who say their jobs are temporary.

22
Data on homeworkers and zero-hours contracts is also available from the LFS. This
indicates that, for the same quarter, 621,000 individuals were working in their own
home as opposed to those employed in the same grounds and buildings as their home.
The number of individuals working on zero hours contracts was fairly small - around
200,000 or 0.7 per cent of those in employment.

TABLE 3:1. LFS DATA ON NON-STANDARD WORKING ARRANGEMENTS

(WINTER 1997/98)

Proportion of all Numbers in non-

those in standard work
employment and self-
employment
%

LFS Temporary Workers 6.5 1.7 million

Fixed term contracts 3.3 868,000

casual employment 1.2 324,000

agency-supplied temp 0.9 222,000

workers
seasonal work 0.3 88,000
others 0.7 187,000
Other ‘Non-standard’ working
zero hours* 0.7 180,000
homeworking 2.4 621,000
Self-employed 12.4 3.2 million
with employees 3.2 835,000
without employees 9.2 2.4 million

* Figures relate to GB for Spring 1998

It is difficult to estimate the size of other groups, such as borderline self-employed,

sub-contractors and people working in partnerships since accurate data are not
available. Rough estimates of the borderline self-employed can be obtained by
looking at all self-employed, with no employees, by the individual’s occupation. For
example, it is unlikely that someone working in a secretarial occupation, categorised
as self-employed with no employees is truly self-employed. The Winter 97/98 LFS
data show that, of the self-employed without employees which number around 2.4
million, 35,000 are secretaries and clerks, 50,000 sales representatives, 24,000
unskilled building workers, 35,000 drivers, 8,000 clergy. This is however, a crude
measure of borderline self-employment since it is impossible to establish whether
these people are really running a business.

23
3.2.3 Limitations of the LFS Data

The main difficulty with the LFS data is that they are based on self-reporting - that is
the individual’s own perception of their status. The responses given may or may not
correspond to the legal division between those working under a contract of service
(employee) and those working under a contract for services (self-employed).

Added to this is that the questions on the temporary nature of jobs are only asked of
those people who believe themselves to be employees. Individuals who think they are
self-employed but might be classed as an employee at an employment tribunal will
not be asked these questions. The LFS therefore, misses a potentially important group
in terms of those with unclear status.

Finally, the LFS does not provide the information necessary to distinguish between
those employed under a contract of employment, the further group of those who
would be included by the concept of ‘worker’ – the ‘dependent self-employed’ - and
the remaining group of ‘independent’ self-employed. It can not therefore, give an
accurate measure of the number of workers who are covered by discrimination
legislation, National Minimum Wage and Working Time Directive.

3.3 Methods

The quantitative wave of the research had a number of objectives. It aimed to produce
an accurate assessment of the prevalence of non-standard working arrangements and
the proportion of those in employment who may be covered by the ‘worker’
definition. The survey data was also to be compared with the LFS and from this,
suggestions made as to how the survey could be made more useful for identifying true
status. Finally, the survey was to generate a sample for further qualitative study.

3.3.1 Surveying Employers?

A number of approaches were considered in the design of the project. One option was
to undertake an employer survey. Although a representative sample of employers is
easy to obtain, there are a number of drawbacks.

• It is likely that employers will have more than one form of non-standard
employment in operation at their establishment. Thus, only a limited amount of
detail could be obtained about each, without using an overly complex and time
consuming questionnaire.

• Many employers may not fully understand the legal implications of the contractual
arrangements under which staff are employed. This is particularly the case with
small employers.

• If employers are fully conversant with the employment contracts in use, it is

possible that they will not be candid about practices that may be disadvantageous
to the employee. In other words, they may not freely admit that they regard as self-
employed, workers who should objectively be treated as employees.

24
It may have been possible to use an employer survey to identify employees in non-
standard forms of employment who could then be interviewed separately. However,
this posed problems of self selection by employers. Overall, this approach was not
considered the most suitable methodology.

3.3.2 Survey of individuals

It was felt that the most appropriate approach was through individual workers. This
has the advantage that the individual will only be in one form of non-standard
employment for each job held. It is therefore more practical to gather information
about each case in some detail. It would also be possible to ask respondents to provide
copies of their written contracts for later examination. There are, however, two major
disadvantages with this approach:

• there is no available ‘register’ of people in non-standard forms of employment.

Therefore, a screening exercise would be necessary to identify individuals who fell
into the scope of the project. This could lead to a complex and costly methodology.

• many individuals are unlikely to fully understand the nature of their employment
status. In the absence of a copy of their contract or other documentation, a
considerable part of any survey or interview would have to be devoted to
establishing their true employment status.

3.3.3 Survey

It was not possible to conduct a bespoke, large scale, random survey of individuals in
the time available. The alternative adopted was to buy space in an omnibus. Since
these run on a weekly basis, the time and costs involved in setting up a large survey of
individuals is minimised. Thus, this option provided the possibility of contacting a
large number of individuals over a relatively short period of time.

The weekly RSGB Household Omnibus Survey was used as the vehicle through
which to run the questions. The omnibus uses a random location sampling method.
This utilises 1991 UK Census small area statistics and the Post Office Address File to
divide the country into 600 areas of equal population within standard regions. These
are then stratified by population density and socio-economic grade. To maximise the
statistical accuracy of the sampling, sequential waves of fieldwork are allocated
systematically across the sampling frame to ensure maximum geographical
dispersion.

For each wave of the omnibus one post code from a selection of the 600 sampling
points is chosen. Each interviewer is provided with 100 addresses from each post
code. Addresses are contacted systematically, with three doors being left after each
successful interview. Interviewing is restricted to afternoons in the weekdays. To
ensure a balance of responses, a quota is set by sex, economic activity and presence of
children. Questions are administered face-to-face using computer-aided interviewing.

As a form of pilot, the questions were placed in the survey for one week and results
evaluated before committing to further weeks of the survey. This allowed a check to
be made as to whether the questions were working properly and whether the spread of

25
responses was appropriate. The questionnaire had three broad sections: the first
looking at the individual’s perceived status; the second looking at temporary working
arrangements; and a final section containing a series of questions attempting to
replicate the tests used by Employment Tribunals or Courts.

Once satisfied that the questionnaire was suitable, the questions were repeated for
three weeks mid-January to early February 1998. A copy of the questionnaire is
included in Appendix 2. Across the four weeks, 7651 people aged 16 or older were
interviewed. All results are weighted to produce population estimates for Great
Britain. In presenting the data, each table includes unweighted bases. This base
number varies because those cases where the respondent could not answer the
question have been excluded.

3.4 Economic Activity of Sample

The survey was only interested in following up people who were in employment at
the time of the interview. At an early stage in the omnibus questionnaire, individuals
are asked about their current situation. Respondents fell into the categories set out in
table 3:2.

Of the 7651 people aged 16 or older that were interviewed over the four weeks, 4,006
were in employment, that is working full-time or part-time. The rest were split
between being retired or unemployed, either seeking work or inactive, and were
therefore not asked the employment status questions. All of the analysis in the
remainder of the report, therefore, refers to those in employment.

TABLE 3:2. STATUS OF WHOLE OMNIBUS SAMPLE

column percentages

Numbers of Proportion of
respondents all
respondents

Working full-time 2882 37.7

Working part-time 1124 14.7
Unemployed and seeking work 1242 16.2
Unemployed and not seeking work 347 4.5
Still at school 73 1.0
In full-time higher education 293 3.8
Retired 1690 22.1

(Unweighted N=7651)

3.5 Comparisons with the Labour Force Survey

The first stage of the data analysis was to establish whether there was any bias in the
omnibus survey respondents. The survey data are compared with those of the Winter
1997/98 quarter of the LFS. As indicated in the previous section, interviewing for this
quarter most closely coincides with the timing of the omnibus survey. Table 3:3

26
makes two points of direct comparison with the LFS, to vouchsafe the
representativeness of the employment status survey.

As can be seen from the table, the results from the two sources are very similar - what
differences there are would mostly fall within the normal range of survey error. We
can, therefore, rely with some confidence on the employment status survey results.

TABLE 3:3. MAIN POINTS OF COMPARISON WITH THE LABOUR FORCE

SURVEY
column percentages
LFS - Employment status
Winter 97/98 survey
In employment
Males
full-time 50.6 50.8
part-time 4.8 5.0
Females
full-time 24.5 23.3
part-time 20.1 21.0

Employment status
Employee 86.9 85.9
Self-employed 12.3 12.7
Govt training programme 0.5 0.6
Unpaid family workers 0.3 0.8
(Unweighted N=4006)
Base: All in Employment

3.6 Perceptions of status

Employment status, as defined in Table 3:2, is based on the standard LFS questions
that was replicated in the omnibus survey.46 The respondents were asked whether they
were working as an employee, self-employed or employee of your own business, on a
government scheme or as an unpaid family worker. The responses are self-reported -
in other words, 85.9 per cent of those in employment perceive themselves to be
employees, while 12.7 per cent categorise themselves as self-employed. There was
some difference by sex with a higher proportion of males, 17 per cent, being self-
employed compared with only 7.5 per cent of females. Part-time workers were
slightly less likely to be self-employed than those working full-time. Details are given
in Table 3:4.

46
Throughout the survey, respondents who had more than one job were asked to
answer on the basis of their main job.

27
TABLE 3:4. PERCEIVED STATUS BY SEX AND PART-TIME/FULL-TIME
row percentages
Employee Self- Government Unpaid Family
employed Training Worker

Males 82.0 16.9 0.5 0.6

Females 90.9 7.5 0.7 1.0

Full-time 85.1 14.0 0.7 0.3

Part-time 88.3 9.2 0.3 2.2

(Unweighted N = 4006)
Base: All in Employment

At this stage those who answered that they were on a government training scheme or
worked as an unpaid family worker, were filtered out of the analysis since their status
is not at issue, and so were no longer in scope of the survey.

Most surveys of employment status rely on the individual’s own assessment of their
current position. It is likely however, given the complexity of establishing status and
the range of factors tribunals have to take into account, that some individuals may not
be certain of their status or they maybe mistaken. All remaining respondents were
asked further a question about their working arrangements to provide more insight as
to the likelihood of their being employees, genuinely self-employed or falling
somewhere in between. Respondents were presented with a series of statements and
were asked which best described their situation. Details are given in Table 3:5, broken
down by whether they perceived themselves as employed or self-employed.

As to be expected, the majority of respondents indicated that they were being paid a
wage or salary by an employer. This is consistent with the fact that 86 per cent of
respondents thought that they were employees. What also appears consistent is the
split of responses between the employees and self-employed. Nearly all those who
had said they were employees were paid by a wage or salary. Equally, nearly all of
those indicating that they were self-employed, worked for themselves in some way or
were sole directors of their own businesses, running partnerships or involved in
professional practices.

28
TABLE 3:5. FURTHER ANALYSIS OF STATUS BY PERCEIVED STATUS
column percentages
Which best describes your current situation? Perceived
Status
Employee Self- All
employed
% % %

Paid a salary or wage by an employer 97.3 4.1 85.3

Paid a salary or wage by an agency 1.5 0.9 1.4
Sole director of own limited business 0.0 7.2 1.0
Running or a partner in a business or professional 0.1 15.8 2.1
practice
Working for yourself 0.1 52.8 6.9
A sub-contractor 0.2 8.5 1.3
Doing freelance work 0.3 9.6 1.5
Refused/Don’t Know 0.4 1.2 0.5

(Unweighted N = 3955)
Base: All in employment except 51 respondent who are on Government Training Schemes or are
Unpaid Family Workers

Of interest were those being paid by an agency, the majority of whom thought that
they were employees despite the growing trend of agencies to treat temps as self-
employed. Only a very small number of employees said they were contractors and
freelancers and a very small number of the self-employed said that they were paid a
salary by an employer.

This question was also used to identify individuals who were in partnerships or sole
directors of their own limited companies. Although there is some question as to the
real status of individuals in such circumstances, particularly the latter group from a
taxation point of view, it was not felt that this survey was the appropriate tool to
explore their status. Indeed, for these people their employment rights are unlikely to
be in dispute since they are the directors of their own companies. These respondents
were filtered out of the questionnaire from this point.

On initial view of status, most individuals would appear to categorise themselves

fairly neatly as employees or self-employed and this seems to be consistent with the
means by which they are paid or a superficial view of the circumstances under which
they work. Only a very small proportion, 1 per cent of all respondents in employment,
seemed to provide conflicting answers to the two initial questions examining status.
However, as we saw earlier, an individuals’ perceptions may not coincide with their
legal status and so the survey went on to explore their working arrangements in more
detail to establish if such discrepancies existed. This firstly looked at the self-
employed and then examined the temporary nature of jobs.

29
3.7 Self-employment

When identifying genuine self-employment, one of the more reliable indicators is

whether the individual works alone or has employees. The Inland Revenue and DSS
guidance suggest individuals ask themselves, when assessing their own status,
whether they are able to hire people on their own terms and whether they pay them
out of their own pocket.47 Those respondents who indicated that they were self-
employed and/or said they were working for themselves, a sub-contractor or doing
freelance work, were asked whether they employed anyone.

The majority of those who said they were self-employed (76%), worked on their own.
Only a fifth had workers or employees who they paid whilst 3.2 per cent employed
unpaid family workers. This is consistent with findings from both the DTI and the
LFS. The former shows that, of the 3.7 million active enterprises at the start of 1996,
2.5 million were run by self-employed people without employees.48 The LFS
indicates that 74 per cent of self-employed have no employees.

From these responses it was assumed that those who had paid employees were
genuinely running a business on their own account. Their status was thought to be
fairly clearly self-employed so these individuals were filtered out of the survey. This
left the remaining three-quarters of self-employed with no paid employees to
participate in the rest of the survey. Those employing unpaid family workers or who
answered ‘Don’t know’ were also included.

3.8 In non-standard work?

Having eliminated a number of individuals whose status was not an issue for this
survey49, the questionnaire went on to establish whether any of the respondents were
working under ‘non-standard’ arrangements. This would then be used to identify
those individuals whose status was potentially ambiguous.

3.8.1 Permanent or temporary?

In order that comparisons could be made with the LFS, similar questions were asked
about whether an individual’s job was temporary and in what way it was temporary.
Respondents were asked, ‘leaving aside your own personal intentions and
circumstances, is your job a permanent job or is there some way it is not permanent’.
Unlike the LFS, self-employed without employees were also asked about the
permanency of their work. For the genuinely self-employed, this question could
47
Employed or Self-employed: A Guide for Tax and National Insurance Inland
Revenue, May 1995.
48
Small and medium enterprise - Statistics for the United Kingdom, 1997, SME
Statistics Unit, DTI, July 1998.
49
Those excluded were: self-employed with employees (75 respondents), sole
directors of a limited business or partner in a professional practice (113 respondents),
those on government training schemes and unpaid family workers (51 respondents).

30
appear irrelevant because either answer would be accurate. They could regard
themselves as permanently self-employed, whilst each job that they undertake is
temporary. It was felt that by excluding them, however, there was a danger the
borderline self-employed would be eliminated from the survey.

Of those still in scope, just under 90 per cent said their job was permanent. Given that
there is a degree of insecurity associated with self-employment, it would be expected
that more of the self-employed would see their job as temporary compared with
employees, and this was the case.

There was also variation depending on whether the individual worked full-time or
part-time. Nearly a fifth of those working part-time indicated that their jobs were
temporary, compared with only 8 per cent of those working full-time. Of those
working part-time, a much larger proportion of males had temporary jobs compared
with females (34% compared with 16%) and, in the main, these were respondents
who perceived themselves to be employees. Male part-time employees, therefore,
were more likely to see themselves in temporary jobs than female part-time workers.
This accounts for why there was little variation in job permanence by sex, despite a
larger proportion of females working part-time. Details are given in Table 3:6.

TABLE 3:6. FURTHER ANALYSIS OF STATUS BY PERCEIVED STATUS

row percentages
Is your job a permanent job or is Permanent Not Permanent
there some way that it is NOT % %
permanent

All 89.2 10.8

Sex
Males 89.8 10.2
Females 88.6 11.4

Perceived status
Employee 91.0 9.0
Self-employed 70.5 29.5

Working:
Full-time 92.2 7.8
Part-time 80.9 19.1

(Unweighted N = 3753)
Note: Table excludes 14 respondents who answered ‘Don’t know’ to question about permanency of
their job.

31
3.8.2 Types of non-standard work

The questionnaire went on to explore non-standard working patterns. Rather than

filtering out all those who said their job was permanent, as in the LFS, all respondents
were asked whether their work could be described by a number of non-standard
working arrangements. This was done because some respondents could be working
under non-standard arrangements but not view their jobs as being temporary in any
way. The question also included options for homeworking, working under a zero-
hours contract and working as an apprentice. Finally, an option of ‘none of these’ was
included to ensure that those who were not working under any non-standard
arrangements were able to answer the question.

Overall, three-quarters of respondents said that their job was not described by any of
non-standard arrangements listed. Table 3:7 shows the breakdown by whether
respondents said their work was permanent or not in the previous question. As to be
expected, the majority of those who believed their jobs to be permanent jobs, were not
working under any of the non-standard forms listed. Of this group 95 per cent
perceived themselves to be employees. We can assume from this that they are indeed
employees and their status is probably clear.

TABLE 3:7. NON-STANDARD WORK UNDERTAKEN BY PERMANENCE OF

JOB
column percentages
Is your work best described by any of In a permanent In a temporary All
the following? job job
% % %
(Numbers) (Numbers) (Numbers)

Not in non-standard work 81.0 82.1 74.2

In non-standard work 19.0 17.9 25.8

(Unweighted N = 3732)
Note: Table excludes 35 respondents who answered ‘Don’t know’ to questions on types of non-
standard working and permanency of job

More interesting were the sizeable minority of individuals, just under a fifth, who had
said their job was permanent and went on to describe it as being non-standard in some
way. In the case of homeworkers there may be no ambiguity over status. Some
homeworkers, particularly in service sector jobs and non-manual occupations, may
clearly be employees and have an arrangement with their employer to work
permanently at home. For others however, it is likely that their status is unclear.

The qualitative wave of the research found that despite the question explicitly asking
respondents to leave aside their own personal intentions and circumstances, the term
‘permanent’ was clearly open to various interpretations. For instance, in a subsequent
telephone interview with a student in her gap year, the interviewee reported that her
contract was ‘permanent’, but that in the initial survey she had responded that it was
not permanent ‘because it’s not a career job’, meaning that she was not intending to
do the job for ever. The term “permanent” was also used to refer to a ‘way of life’
rather than a particular contract by some self-employed individuals. One face-to-face

32
and two telephone respondents described themselves as permanently self-employed,
although for both of them this meant alternating between contracts and periods
without work – ‘chronic insecurity’ might be a more accurate way to describe their
situation. This response could reflect the fact that they regard themselves as
permanently self-employed, whilst each job that they undertake is temporary. It may
also reflect long term relationships with particular clients producing regular, although
not continuous, work.

There was also a group of respondents who indicated their job was temporary but said
that it was not described by any of the forms of non-standard working listed. For these
individuals, this answer may well reflect that they are working in a form of non-
standard employment other than those listed in the questions. These respondents will
also have ambiguous status.

It is also possible to look at the split between standard and non-standard working by
perceived status, sex, and full-time and part-time work. Non-standard working was
more prevalent amongst part-time than full-time workers. The greatest variation
however, was between employees and self-employed. Over three-quarters of those
perceiving their status to be employed said they were not working under any of the
arrangements listed. For the self-employed, 55 per cent indicated some form of non-
standard working. These self-employed would be excluded from the LFS questions on
temporary work. Thus, non-standard working appears to be much more prevalent
amongst the self-employed and slightly more prevalent with those working part-time.

TABLE 3:8. TYPES OF NON-STANDARD WORK UNDERTAKEN BY

PERMANENCE OF JOB
column percentages
Is your work best described by any of In a permanent In a temporary All
the following? job job
% % %
Of those in some form of non-standard
work:
Under contract for fixed period/task 70.6 49.9 63.5
Casual work 10.3 22.6 14.5
Homeworking 9.0 3.5 7.1
Agency temping 1.0 12.3 4.9
Seasonal work 2.1 8.0 4.1
As an apprentice or trainee 4.7 2.1 3.8
Under a zero-hours contract 2.9 3.0 3.0

(Unweighted N =) (655) (343) (998)

Note: Table excludes 35 respondents who answered ‘Don’t know’ to questions on types of non-
standard working and permanency of job

Table 3:8 also looks at the types of non-standard working arrangements. The most
common form was working under a contract for a fixed period or task, accounting for
nearly two-thirds of respondents. This was followed by casual working and
homeworking. It is possible to make some comparison with the LFS if analysis is
confined to employees whose job is temporary. By matching the LFS filtering, the
spread of respondents by type of temporary work was similar, being at most, only one

33
or two percentage points different. This shows that our sample is reasonably in line
with LFS results.

Table 3.8 shows that for those who said their job was permanent and specified some
form of non-standard working, the majority were working under a fixed term contract,
with a small proportion in casual work and homeworking. For those whose job was
temporary, nearly half were working in fixed term contracts, just over a fifth were
doing casual work and a small proportion working through agencies.

Looking at this another way, it was only for casual work, agency temping and
seasonal work that the majority said their job was not permanent in some way. For
those working under fixed term contracts, nearly three quarters thought they had
permanent jobs and for causal workers, just under half believed they had a permanent
job. Under the current routing for the LFS questions, these people would be excluded
from the figures for temporary workers. These data reproduced in Figure 3.1 in
graphical form.

FIGURE 3:1. PERMANENCY OF JOB WITHIN TYPES OF NON-STANDARD

WORKING (NUMBERS)

500

400

300

200

100

0
Zero-hours
Homeworking
Fixed period/task

Apprenticeship
Seasonal
Agency
Casual

Permanent Not permanent

Examining those in non-standard work further, we found that half perceived

themselves to be employees and were working full-time. A third also perceived
themselves to be employees but worked part-time. The remainder were self-
employed, the majority working full-time. This reflects the dominance of non-
standard work by people on fixed term contracts. Outside these groups, self-
employment and part-time work were more prevalent.

34
3.9 Conclusion

This analysis has shown that at an initial stage of questioning, respondents fairly
neatly categorised themselves as employed and self-employed, with few appearing to
provide conflicting information about their status and working arrangements. When
the permanency of their jobs and existence of non-standard arrangements is examined,
more inconsistencies appear. So far we have identified a number of groups whose
status could be seen as ambiguous. These include:

• the small number of self-employed who say they are paid a salary or wage by an
employer

• a number of contractors and freelance workers who believe they are employees

• three-quarters of the self-employed with no employees, of whom just over half say
they are working in non-standard forms of employment

• nine per cent of employees who say their job is not permanent

• the quarter of employees who work under some form of non-standard arrangement;
and

• agency temps

Combining these groups together, suggests that there is a fairly large proportion of
people in work with potentially ambiguous status. The survey also shows that simply
asking individuals whether their jobs are permanent or not, is not adequate to fully
capture those with potentially temporary jobs or ambiguous status. In particular, the
current structure of questions in the LFS may significantly understate the true level of
non-standard working arrangements, especially for estimating numbers on fixed term
contracts.

35
Chapter 4. Clarifying Employment Status: Quantitative Analysis

4.1 Clarifying employment status

The previous chapter provided evidence of the ambiguity in employment status of a

proportion of respondents to the survey. Although not the majority of respondents,
these ambiguous groups formed a sizeable minority. This self reported status may or
may not correspond to the legal division between those working under a contract of
employment or service (employees), those included under the definition of ‘workers’,
and the independently self-employed.

The purpose of the remainder of the survey was to ask questions applying the types of
tests that have been used by courts and employment tribunals in deciding this
question. This would allow estimates to be generated of the proportion of people in
employment who might fall within the wider definition of ‘worker’ as defined in
Section 230 of the Employment Rights Act 1996.

4.1.1 Identifying those with ambiguous status

We saw in the previous chapter that the survey asked a series of question to attempt to
identify individuals who were likely to have ambiguous status. The way this was
approached was to identify those groups for whom we thought there was no
ambiguity whatsoever about their employment status. Figure 4.1 shows how we put
this into effect.

FIGURE 4.1: IDENTIFYING THOSE WITH CLEAR EMPLOYMENT STATUS

Clearly employees Clearly self-employed

define themselves as an employee are a director or partner in own

business
&
&/or
are paid a salary or wage
employ others
&

hold a permanent job

&

have no ‘non-standard’ working

patterns

Just under two-thirds (64 per cent) of respondents were clearly employees since they
defined themselves as employees, were paid a salary or wage, held a permanent job

36
and had no non-standard working patterns. There were no characteristics whatsoever
to suggest any ambiguity. Five per cent of respondents were clearly self-employed
since they were a director or partner in their own business and/or employed others.
There seems little doubt that a court or employment tribunal would have any
difficulty in these cases.

4.1.2 Unclear Status

In total, therefore, 30 per cent of all those in employment have an employment status
that, on first inspection, has elements of uncertainty and is not completely clear. This
residual group, for whom employment status may be unclear, is made up of two sub-
groups:

(i) those defining themselves as self-employed, who are not directors or partners
in their own business, and who do not employ others; and,

(ii) those defining themselves as employees who have some type of non-standard
working pattern or classified their jobs as non-permanent.

The first sub-group may include a proportion of people who could be borderline self-
employed, while the non-standard workers may include many who are clearly
employees, but their atypicality raises queries about their employment status in law.

Table 4:1 shows how the original classification of employment status, based in LFS
questions, is affected by this categorisation. In the first instance, 86 per cent of
respondents in work saw themselves as employees. Of these however, a quarter
proved to have uncertain status (22 per cent of respondents in employment). Under
the LFS classification, 13 per cent of respondents in work said they were self-
employed but on closer examination, employment status was not certain for nearly
two-thirds of respondents (or 8 per cent of all respondents in employment).

TABLE 4:1 EMPLOYMENT STATUS ON FIRST INSPECTION

column percentages
Employment status
Employee 85.9
Clearly employees 63.6
Unclear status 22.3
Self-employed 12.7
Clearly self-employed 5.0
Unclear status 7.7
Govt training programme 0.6
Unpaid family workers 0.8

(Unweighted N = 4006)

A slightly higher proportion of females appear to be clearly employees and a smaller

proportion are clearly self-employed. For those with ambiguous status, the
proportions are almost identical. In terms of part-time and full-time workers, part-time
workers are more likely to have unclear status compared with those working full-time.

37
4.2 Common law tests of employment status

As has been noted, the ultimate arbiter of employment status is a court or employment
tribunal. Over time the type of test and its application has varied (as have the tests and
applications used by administrative arms of government such as the Inland Revenue
and Benefits Agency). We saw in Chapter 2 that courts and tribunals use a
combination of four separate tests, none of which in themselves can determine
employment status, but examined as a whole allow for a conclusion to be drawn. In
the survey these tests were tackled in the following ways:

• mutuality of obligation: this refers, amongst other things, to the right to refuse to
work on the part of the worker. Respondents were asked whether, when working
for a particular client, organisation or agency, they were allowed to take on work
from other clients, organisations or agencies. They were also asked whether they
could refuse work offered to them by the clients, organisations or agencies they
worked for.
A further indication of mutuality of obligation is the regularity and duration of
employment. If an individual works for a large number of employers or clients
during the course of a year they are more likely to be regarded as self-employed. If
they only work for one organisation for a substantial period of time, they are more
likely to be regarded as employees. The survey asked respondents with unclear
status how many organisations, agencies or clients they had worked for in the last
six months.
• the extent of control: this is interpreted as the employer’s right to specify the way
in which work is done. The survey asked respondents whether, when they are at
work, someone has the right to tell them, at any time, what to do or when and how
to do it, even if this rarely happened in practice.
• level of integration into the organisation: this looks at whether the worker is ‘part
and parcel’ of the organisation. The survey asked respondents whether they were
covered by the grievance or disciplinary procedures of their clients, organisations
or agencies they work for.
• economic reality: this attempts to establish whether the worker is in business on his
or her own account or works for another who takes the ultimate risk of loss or
chance of profit. An indicator of economic reality is the ability to sub-contract. The
survey asked respondents whether they were free to hire other people, who answer
to and are paid by them, to do the work the respondent has personally taken on.
The survey also established whether they had to provide the main items of
equipment they needed to do their job, not just the small items many employees
provide for themselves.
Another important indicator of economic reality is how people are actually paid,
the assumption being that those paid a monthly salary or weekly wage are likely to
be employees. The survey also asked individuals whether they paid their own
National Insurance contributions or income tax or whether this was usually
deducted by the organisation they worked for. For those people who are
employees, it is the employer’s responsibility to deduct tax and NI contributions
from their pay. If this is being done, it may be an indication of employee status.
Finally, respondents were asked whether they were entitled to receive sick pay or

38
 paid holidays from the organisation they worked for. Again, receiving such
benefits may be an indication of employee status.
The final issue explored was that of written contracts of employment and other
documentation. Respondents were asked whether they had a written contract of
employment or another document, from their employer or agency, which set out their
terms and conditions in some way. If an employer provides statutory statements of
particulars, whether or not they believe the worker to be an employee, a tribunal
might decide that the individual is an employee. However, an actual contract or
statement of particulars does not settle status. An absence of documentation certainly
does not preclude employment under a contract of employment.

These questions are only crude attempts to address the sorts of questions an
employment tribunal would examine. In reality the tribunals do not take any particular
test on its own. In addition, these are often not black and white situations and a
tribunal will assess the merits of each individual case. In designing the survey, we
were aware that some of these questions could be misinterpreted by respondents and
answers given could be inadvertently misleading. For example, the control test is
largely irrelevant to those workers who have a high level of professional training.
They are likely to have a great deal of autonomy over their work but still be
employees. They are likely to answer that their employer does not dictate the way the
work is done. These problems are explored further in the qualitative stage of the
project and are reported in the following chapters.

4.2.1 Tests by perceived status

The rest of the survey, and thus the remainder of the analysis in this chapter, includes
only those individuals identified as having unclear status. Those who were clearly
employees or clearly self-employed (see section 4.1.1) were excluded. Table 4:2
presents results of the questions we asked to apply the legal tests, broken down by
self-reported employment status. The percentages reflect the proportion of
respondents who gave an answer consistent with being dependent labour, that is
consistent with being workers or employees and not independently self-employed.
The remainder not shown either indicated independence or answered ‘Don’t know’.

The Table shows that of all those with unclear status, the test with the largest
proportion indicating dependency was that of control. Seventy-eight per cent said that
there was someone who could direct how their work was undertaken. The questions
with the lowest levels of dependency were two of the three for mutuality of
obligation. These were not as clear cut as would be expected and this suggests that
this line of questioning was perhaps the weakest or that mutuality of obligation is a
difficult concept to pin down. In all, just under half provided answers suggesting
dependency.

All the questions for the mutuality of obligation test had a fairly high proportion of
respondents answering ‘Don’t know’ - 8 per cent for ability to refuse work and for the
number of employers worked for in the last six months and 9 per cent for whether
they were allowed to work for more than one work provider. The test for integration,
that is whether the individual was covered by the clients’ or employers’ grievance
procedure also produced a fairly large number of don’t knows - 14 per cent. This may

39
suggest either confusion over the meaning of the question or it may show that
respondents genuinely do not know whether they are covered.

TABLE 4:2 TESTS OF DEPENDENCY BY SELF-REPORTED STATUS

Employees Self- All with

with employed unclear
unclear with unclear status
status status
Control test
can be told what to do 90 45 78

Integration test
Covered by grievance/disciplinary 69 21 57
procedures

Economic reality test

Cannot hire people/sub-contract 91 36 76
Employer provides equipment* 85 20 68
Paid a weekly wage or monthly salary 82 21 66
Employer pays NI/tax 87 7 66
Entitled to sick pay or paid holidays 72 7 54

Mutuality of obligation test

Cannot work for more than one work 56 23 47
provider
Cannot refuse work 52 19 43
Worked for only one employer in last 6 81 45 71
months

Other information
Have a written statement of particulars 75 19 60

(Unweighted N = 1182)
Base: all those with potentially unclear status
* Excludes 10 per cent of respondents who did not need any equipment.

Of those who perceived themselves to be employees, the tests with the largest
proportion indicating dependency were ability to sub-contact their work (91%) and
the test of control (90%). The tests with the lowest proportion indicating dependency,
as with all respondents, were those relating to mutuality of obligation.

For the self-reported self-employed, the tests with the highest incidence of
dependency (45%) were again that of control and the economic reality test of working
for a single employer in the last six months. In the case of the control test, this could
reflect respondents interpreting the question as following instructions of clients. Just
over a third of self-employed also indicated that they could not hire or sub-contract
their work to other people. In the case of professionals who are self-employed this
could be a legitimate response since they are often employed for their personal skills
and would not be able to pass the work onto others.

40
The lowest indicator of dependency was whether tax and NI were deducted at origin
and whether they were entitled to paid leave or sick pay. Only 7 per cent of self-
reported self-employed provided answers consistent with dependency. Again, in the
case of agency workers, for example, they may be working as self-employed but still
have their tax and national insurance deducted by the agency, as required by law.

Other key findings include:

• Seventy per cent of those with unclear status had worked for only one employer in
the last six months. Of the self-employed, just under half said this was the case. It
could be that the individual had not been working and had only just started with the
current employer but this is unlikely to be the case for the majority. Duration of
employment is examined in section 4.4.

• Sixty per cent had a written statement of particulars, three-quarters of employees

and a fifth of self-employed. Part-time workers were also less likely to have a
contract of employment.

• Two-thirds were paid a monthly salary or a weekly wage. The majority of those
who said they were employees were paid in this way compared with a fifth of self-
employed.

• The majority of those perceiving themselves to be employees said their employer

paid their NI and tax. However, 8 per cent of employees said they paid their tax
and NI and the remainder were unsure.

• Only just over half of respondents with unclear status said they were entitled to
paid leave. Just under three quarters of employees said they were entitled to paid
leave whilst only 7 per cent of self-employed said this was the case. Of the part-
time workers, over half said they were not entitled to paid holidays or sick pay
compared with 39 per cent of full-time workers.

As can be seen from the data, for the most part the answers correspond to what we
would expect to find. Of those with unclear status, self-reported employees are much
more likely than self-reported self-employed to have the characteristics of
dependence. For example, 90 per cent of employees said that while at work someone
had the right to tell them what to do or when and how to do it, compared with 45 per
cent of those calling themselves self-employed.

Looking at any single test in isolation, the proportion of all those with unclear status
who would be identified as clearly self-employed ranges from 22 per cent (under the
control test) to 57 per cent (under the mutuality of obligation test of refusing work). In
other words, 22 per cent of those with unclear status said there was no one who had
the right to tell them how the work should be done or replied don’t know.

The closest predictor of self-reported status is whether the individual pays their own
NI contributions and income tax, but as we saw in chapter 2 above, the courts and
employment tribunals do not regard this is as key indicator of actual employment
status. Among the self-employed with unclear status 89 per cent pay their own NI/tax,

41
but fewer than 89 per cent meet the self-employment criteria for the four tests. For
example, there is a 25 percentage point difference between the proportion defined as
self-employed under the economic test of sub-contracting, and those paying their own
NI/tax (i.e. 36 - 11 per cent).

4.2.2 Applying the tests to non-standard workers

It was also possible to look at the characteristics of the main groups of non-standard
workers, that is those on fixed term contracts, casual workers, homeworkers and
agency workers.

Fixed Term Contracts

The majority of those on fixed term contracts (86 per cent) said they could not hire
other people to do the work they had been contracted to do. Three-quarters said they
were covered by the grievance and disciplinary procedure of the organisation they
worked for. Again, the majority (over 80 per cent) said their employer provided the
equipment they had to use, that there was some one who could direct the work they
were doing, that they were paid a salary or weekly wage, had tax and NI deducted by
the employing organisation and were given a contract of employment. Three-quarters
received sick pay and paid holidays and just over three-quarters had worked for only
one organisation in the last 6 months.

Thus, in the most of the tests, the majority on fixed term contracts provided answers
consistent with being an employee. It was only the tests of being able to refuse work
and working for more than one client at a time, were dependency in any doubt.

Casual workers

Nearly two-thirds of casual workers said they could work for other employers, could
refuse work offered to them and were not covered by their employers' grievance or
disciplinary procedures. Only half were paid a weekly wage or salary, two-thirds
received no sick leave or holiday pay and had no contract of employment.

However, the majority (82 per cent) could not hire other people and a similar
proportion said they could be directed in their work. Seventy per cent said their
employer provides the equipment they used whilst two-thirds said their employing
organisation deducted tax and NI. Just over two-thirds had worked for only one
employer for over a year.

With casual workers the picture is more mixed than with those on fixed term
contracts. Some of the tests suggest dependency for the majority, whilst other clearly
suggest that they are independent labour.

Agency Workers

Ninety-three per cent of agency workers said that someone had the right to tell them
what to do at work and the majority were not able to hire other people. Just under
three-quarters had equipment provided by their employer and three-quarters had their
tax and NI deducted by their employing organisation.

42
Despite these indications of dependency, only half of agency temps said they were
paid a salary or weekly wage by their employer. Half were covered by their
employer’s or client's grievance procedure and half of the agency temps said they had
a contract of employment or statement of particulars. Just under half said they could
not work for more than one work provider at a time. Only a quarter were entitled to
paid holidays and sick pay and only 14 per cent said they could not refuse work the
work offered to them.

More agency temps display characteristics of dependency than homeworkers or casual

workers. However, a sizeable number, around half, are clearly demonstrating
characteristics of independence.

Homeworkers

Just under two-thirds of homeworkers said that they could work for more than one
work provider and the majority said they could refuse work. Just under two-thirds
paid their own tax and NI and most were not were entitled to paid holidays and sick
pay. Only a quarter said they were covered by their client’s or employer’s disciplinary
procedures and just over a quarter had a contract of employment or statement of
particulars. Only a third said their employer provided the equipment they needed.

Half of the homeworkers said they could not hire other people and just under 42 per
cent said they could be told what to do in their work. Also, half of the homeworkers
said they were paid a weekly wage or monthly salary.

Homeworkers present a stronger picture of independence than any of the other

groups. In any of the tests the largest proportion indicating dependence was around
half and this was only for the ability to sub-contract, how they were paid and the test
of control.

4.3 Estimating Numbers of ‘Workers’

We can now combine these data in different ways to estimate the proportion of people
in employment who might be categorised as clearly working on their own account
(i.e. the group excluded from the ERA definition of ‘workers’).

Firstly, we looked at the data by the number of tests of dependency individuals

passed. The results are presented in Table 4:3. This excludes the question on
provision of equipment because ten per cent of respondents answered they did not
need equipment and could not be classified dependent or independent. Those who
passed all remaining ten questions would be deemed as dependent labour or workers
whilst those passing none would be self-employed.

As can be seen from this, 12 per cent of all those with unclear status have no
characteristics of being an independent contractor, and on all ten questions are shown
to be dependent labour or workers. At the other end of the spectrum, just 2 per cent
satisfy all the different tests of independence. This range of results suggests that
attempting to divide those with unclear status in this way would still be difficult for a
large proportion of workers.

43
For employees with unclear status, 17 per cent gave answers consistent with
dependency to all ten questions. A further 21 per cent provided an answer consistent
with independence for only one of the ten questions. There were however a small
minority who provided answers consistent with independence for five or more of the
questions. For the self-employed only 9 per cent provided answers consistent with
independence for all ten questions. A further quarter gave a consistent answers for all
but one and another 26 per cent did so for all but two. A small proportion provided
answers consistent with dependency for five or more of the questions. These
accounted for ten per cent of the self-employed with unclear status. Again this still
leaves a large group for whom it would be difficult to assess their status.

TABLE 4:3 MULTIPLE TESTS OF DEPENDENCY BY SELF-REPORTED

STATUS
column percentages
Self- Self- All with
reported reported unclear
employee self- status
employed
Multiple tests of dependency
10 tests 16.6 0.2 12.3
9 tests 20.7 0.3 15.3
8 tests 19.2 0.7 14.3
7 tests 17.9 2.2 13.8
6 tests 11.9 2.7 9.4
5 tests 6.4 4.1 5.8
4 tests 4.2 13.8 6.7
3 tests 2.0 17.3 6.0
2 tests 0.8 25.5 7.4
1 test 0.3 24.5 6.7
none of the tests 0.0 8.5 2.2

(Unweighted N = 1182)

The information provided in Table 4:3, suggests we need an alternative approach. We

saw in Chapter 2 above that two factors are particularly important in the definition of
a worker. These are whether the individual undertook to supply his or her personal
services, and whether there was a sufficient degree of economic dependence between
the two parties. We saw that on this basis, individuals are less likely to be dependent
labour or ‘workers’ the greater the extent to which they have an identifiable business
of their own and hire others to work with them. What is also important is how far they
are dependent upon a particular employer or client for the bulk of the work they do, as
opposed to having numerous clients.

We therefore sought to gain an estimate of the number of ‘workers’ by using a

combination of questions based, firstly, on the number of organisations the respondent
has worked for in the last 6 months, and, secondly the group of questions associated
with economic reality (these were whether the respondent could hire others, whether
they were paid a wage or salary, whether they paid their own tax and national
insurance, and whether they were entitled to sick pay or to paid leave). As with Table

44
4:3 we excluded the question relating to the provision of equipment. These questions
were, we considered, most appropriate for differentiating between dependent and
independent self-employment as defined above.

On this basis, those respondents with unclear status who had worked for more than
one organisation in the last six months and passed all the tests of economic reality
were regarded as independently self-employed. In other words this group had:
• worked for more than one employer in the last six months,
• were able to sub-contract,
• were not paid a weekly wage,
• paid their own tax and NI, and
• were not entitled to sick pay or paid holidays.
These accounted for only 8 per cent of respondents with unclear status, leaving the
rest (92 per cent) classified as ‘workers’.

However, if the tests are applied strictly in reverse, in other words, to be a ‘worker’
respondents must have worked for only one employer in the last six months and must
fail all the tests of economic reality, in that they did not pay their own tax/NI, were
entitled to paid holidays and so on, then only 35 per cent would be regarded as clearly
‘workers’ on this basis.

A further 18 per cent had worked for only one employer and passed only one of the
tests of economic reality. We looked at the characteristics of this group and it would
seem that they are also likely to be ‘workers’ or dependent labour. For example, 98
per cent said they were employees, 80 per cent were paid a salary or wage by an
employer and had NI contributions and tax deducted at source. Two-thirds said they
had a permanent job and a similar proportion had a contract of employment. In terms
of the questions relating to the tests, three-quarters were unable to sub-contract and
had the equipment needed provided by their employer. Just under two-thirds were
covered by their employer’s or client’s disciplinary procedures and 90 per cent had
someone who could direct their work. Just over half were working on fixed term
contracts, the remainder spread across the other forms of non-standard working.

On this basis we have inferred that just over half of the unclear group are clearly
‘workers’ or dependent labour. Thus we arrive at a proportion of individuals, 39 per
cent, whose status is still unclear. We also looked at the characteristics of this
remaining unclear group to try to identify any pattern which could give a better
indication of their status. This group were evenly split between those who believed
themselves to be employees and those who thought they were self-employed, those
who believed their jobs were permanent and those who thought they were temporary,
and those who had a contract of employment or similar document and those who did
not. Just under half were paid a salary by an employer and half paid their own tax and
NI contributions. The majority did not get paid leave or sick pay. Three-quarters
could refuse work from clients and two-thirds could work for more than one client at
a time. Just under two thirds said that they were not covered by their employer’s or
client’s disciplinary procedures. However, just under two-thirds could not sub-
contract, a similar proportion said that there was someone at work who could tell
them what to do and 60 per cent had equipment provided by their employer. In other

45
words, the picture is unclear. Short of examining each case individually, it would be
difficult to establish whether these individuals should be regarded as ‘workers’ or not.

TABLE 4:4 COMPARISON OF SELF-REPORTED STATUS AND REASSESSED

STATUS
column percentages
Self- Self- All in
reported reported employment
employee self-
employed
On first inspection
Employee 74.0 63.6
Self-employed 37.6 5.0
Unclear 25.7 62.4 30.0
Gov’t training/Unpaid family worker 1.3

After investigation, adopting economic

reality tests and number of clients worked
for in 6 months
Dependent workers 92.2 1.4 79.5
Own-account workers 0.3 56.1 7.4
Still unclear 7.4 42.4 11.8
Gov’t training/Unpaid family worker 1.3

TOTAL 100 100 100.0

The cases that we could identify as clearly employed or self-employed were then
combined with the original group of respondents who were filtered out at an early
stage because their status was clear from answers to early questions. The results are
shown in the second half of Table 4.4. This shows that 12 per cent of the total
respondents still have unclear status. In other words, the proportion of all those in
employment who could be regarded as ‘workers’ or dependent labour varies from 80
per cent to 92 per cent depending on how those individuals who were ‘still unclear’
would be categorised at a tribunal. Similarly, the group of independent self-employed
could account for between 7.4 per cent and 19 per cent of those in employment.

4.4 Qualifying for employment protection rights on the basis of continuity of

employment

At present to qualify for many employment protection rights, not only must it be
demonstrated that a contract of employment exists, but the employee must have a
continuous period of employment with the same employer of two years or more. This
is particularly relevant for the right to claim unfair dismissal and eligibility for
maternity rights. The DTI’s Fairness at Work White Paper included the proposal that
this qualifying period be cut to one year.

We asked all those with whose status had some ambiguity on first inspection, how
many organisations they had worked for in the last six months. We saw in Chapter 4
that 71 per cent had worked for only one organisation in the last six months. These
respondents were then asked how long they had been working for the same work

46
provider. A quarter had been working for that one organisation for less that a year, 17
per cent had worked for the one organisation for between one and two years, and 56
per cent had worked for that organisation for over two years.

Although far fewer self-employed had worked for only one organisation in the last six
months, 45 per cent compared with 81 per cent of employees, for those that had, there
was little difference in the pattern of duration compared with employees. This
suggests that if someone is self-employed and has been working for only one
organisation in the last six months, they are likely work for that organisation a similar
length of time as employees.

We can combine the answers to these two questions to look at duration of

employment with the same employer across all those with ambiguous status. Of all
those with unclear status, 40 per cent had worked for one employer continuously for
two years or more and 68 per cent had done so for one year or more.

These figures varied by status. Of those who believed they were employees, 45 per
cent had worked for only one employer for more than two years and 60 per cent had
worked for the same employer for more than one year. Only 34 per cent of self-
employed had worked for the same employer for more than a year. Assuming that
these respondents who believe they are employees can demonstrate that a contract of
employment exists, lowering the threshold from two to one year, will involve a
further 15 per cent being covered by employment rights.

If we look at the duration of employment using the re-assigned status and assume that
all those who are in the final unclear group are classified as dependent labour or
‘workers’, those qualifying for employment rights would be 44 per cent with a two
year threshold and 58 per cent with a two year qualifying period. This is slightly
lower than the proportions under self-reported status.

47
4.5 Reasons for working in non-standard arrangements

In the final part of the survey, we asked respondents whose status was ambiguous,
why they worked under their current arrangements i.e. in temporary job, under some
sort of non-standard working and/or self-employed without employees. Details are
given in Table 4:5. Overall, just under two-thirds indicated that the way they worked
had been their preference, with a fifth saying that this was the only basis on which
work was available. The self-employed, were more likely to be working under that
arrangement by their own choice compared with employees. Those who were
employees were slightly more likely to say that this was their employers' preference
or the only basis on which they could get work. There was very little variation by sex
or whether the respondent worked full or part-time.

TABLE 4:5 REASONS FOR NON-STANDARD WORKING BY STATUS

column percentages
Self- Self- All in
reported reported employment
employee self-
employed

Own preference 57 78 64
Employers’/Clients’ preference 14 6 11
Only basis on which the work was available 22 14 19
Some other reason 6 1.5 4
Don’t know 2 0.4 1

(Unweighted N = 1182)

4.6 Conclusions

The analysis of the data has shown that a significant proportion of those in
employment have some ambiguity about their status on first inspection. We examined
those with unclear status using the tests of economic reality and by examining the
number of clients worked for in the last six months. This allowed us to divide some of
the remaining respondents into dependent labour, or ‘workers’, and independent self-
employed. However, there still remained a group, accounting for 12 per cent of all
those in employment, who could not be assigned either way because they displayed
characteristics of both dependence or independence. As a result, the proportion of
those in employment who could be classified as ‘workers’ could range from 80 per
cent to 92 depending on how the remaining unclear individuals would be categorised
by a court or employment tribunal.

This suggests that by using the 'worker' definition, the number covered by
employment rights might be higher than the 87 per cent who believe that they are
employees according to the LFS. The use of the worker definition might protect up to
a further 5 per cent of all those in employment (or about 16 per cent of those whose
employment status is unclear because they are employed in non-standard work of

48
some kind). It is not possible to be more precise about the numbers who would be
affected by this change because of the difficulties of attempting to assess employment
status without examining each case individually.

The second half of the project was designed to approach the assessment of status on
just such an individual, case-by-case basis. It was hoped that this would provide
greater insight as to how the survey questions were interpreted by respondents, and to
make it possible to make a more informed assessment of the status of a subset of the
overall sample. The remainder of this report focuses on this stage of the work.

49
Chapter 5. The Qualitative Wave: Aims and Methods

5.1 Aims of the in-depth interviews

We now turn to an examination of individuals’ perceptions of non-standard work, as

indicated by the evidence obtained from in-depth interviews with the sub-sample of
respondents, 36 of whom were interviewed face-to-face and 24 by telephone. These
interviews were designed to achieve two broad goals. Firstly, they aimed to uncover
richer and more detailed descriptions about the respondents’ employment situation, to
better understand the difficulties that might arise in assigning non-standard workers to
the status of employee, worker or self-employed. Secondly, the interviews were to be
used to assess quality of the data obtained in the initial survey, in terms of the
reliability and validity of the responses.

It was anticipated that the material to be collected in this part of the project would be
of a complex nature, raising issues with which respondents might not be familiar. We
decided that the fieldwork would therefore need to be qualitative and to permit
respondents to explore their responses in an unhurried and supportive environment.
In order to overcome these problems a three-pronged approach was adopted, using a
combination of focus groups, semi-structured face-to-face interviews and brief
telephone interviews.

The interviews and focus groups were conducted on the basis of an aide-mémoire
consisting of a series of open-ended questions, the purpose of which was to structure
the interview in a loose and flexible way (the aide-mémoire is reproduced in
Appendix 3). Questions related to the following general areas: the nature of the
respondent’s job; flexibility; working with others; pay, the employment relationship;
being one’s own boss; setting up a business; security; choice; documentation; and
perceptions of employment status.

Focus groups proved to be highly useful for the generation of very rich data,
permitting individuals within the groups to contribute more than they might in an
interview.50 One was conducted in a Cambridge College and two others in hotel
conference rooms, in the centre of Coventry and Cardiff respectively. Participants
were greeted individually, and the initial survey questionnaire (see Appendix 2) was
re-administered to them. The groups were moderated by a specialist in group
counselling and focus groups, and the discussion was led by one of the principal
researchers in the project team. Although the participants were guided through the
topics using the aide mémoire, they were also given considerable freedom to develop
the conversation in the directions that they saw fit. The discussions lasted for
approximately 100 minutes, including a short coffee break.

Semi-structured interviews, conducted in respondent’s own homes, then pursued these

lines of enquiry further, but under greater control of the interviewer. Finally, brief
telephone interviews were used to clarify some of those issues where the numbers of

50
For a recent review of the methodological basis for the use of focus groups, see
Gibbs, 1997.

50
focus group and interview respondents were not sufficient, or the issues were seen to
be particularly important. These three methods, combined with the results of the
representative sample survey, were thought likely to achieve a good combination of
in-depth exploration, theory testing, and theory generation.

5.2 Achieving the sample for in-depth interviews

It was considered that a total of 36 interviewees, split between focus groups and
individual interviews, and a further 24 telephone interviews would provide an
adequate basis for the work. 546 eligible respondents from the initial survey agreed
to be available for re-interviewing. 11 respondents participated in three focus
groups51 and a further 26 respondents were then be interviewed individually. The total
sample was designed, as far as possible, to contain a range of respondents in fixed
term/task contracts, seasonal work, casual work, zero-hours contracts, homeworking,
agency work and borderline self-employment, and to be heterogeneous in terms of
class, gender and age within each of these categories.

Because of the highly geographically dispersed yet locally clustered nature of the
sample, the focus groups had to be in population centres with a population of at least
one million within approximately 30km of a city centre. An inspection of the sample
suggested the three best possibilities to be Cambridge, Coventry and Cardiff. Even
after refining our approach technique (a letter followed by a telephone call and the
offer of a £15 payment, travelling expenses and childcare), the response rate was very
low. For example in our last and largest focus group in Cardiff, only five of the 16
respondents resident in or near Cardiff who had agreed to be re-interviewed were
included. Five of the telephone numbers listed in the printout of responses were
incorrect, and only three of these could be corrected via Directory Inquiries or by
attempting variations on the number given, thus two potential respondents were lost
due to this52. Of the remaining 14 respondents one refused to take part, five reported
that they were unavailable on the nominated Saturday morning, and three of those
who had agreed to attend the focus group did not turn up.

The response rate in all stages of the fieldwork was particularly low among those
listed as employed under zero-hours contracts. For instance, of the 30 who were
identified as zero hours contracts in the quantitative wave, only 13 agreed to be re-
interviewed, and an interview was only achieved with four of these, two face-to-face
and two via telephone. It was difficult for them to make appointments more than a day
or two ahead if they did not know when they might be working, and in one case
leaving the house for several hours carried with it the risk that they might miss a
telephone call offering work.
51
One of the participants in the first focus group had been recruited from a local
employment agency, and had not participated in the initial interview. Her account of
her employment was not used in subsequent analyses.
52
Throughout all of the fieldwork approximately 30% of the telephone numbers were
found to be inaccurate. The reasons for this are not clear, but in some cases trailing
zeros were omitted (suggesting a programming problem) in other cases a middle digit
was omitted or incorrect.

51
The response rate for individual interviews was slightly higher than for the focus
groups, as the interviewers could be more flexible in arranging interviews to suit the
interviewees and they could be conducted in interviewee’s own homes. Yet among
those identified in the initial survey as being in the most irregular types of
employment there was still a high refusal rate. We suspect that some respondents
were concerned about the confidentiality of the process, particularly as the research
was linked to a government department and had asked questions about pay and tax.

The focus groups took place in late June and early July 1998. Individuals were
interviewed in late July and early August 1998. The sample for the interviews was
drawn from the East Midlands, South Wales, East Anglia, Bristol, Herefordshire,
Merseyside, Essex and Birmingham.

5.3 Data Analysis

Three distinct types of data were employed in this project. Firstly, the interview
schedule was re-administered to all 36 face-to-face respondents, and this data was
added to the initial survey dataset.

Secondly, the recordings of the three focus groups and the 26 face-to-face interviews
were transcribed and analysed using ATLAS.ti software. After several readings of the
transcripts, a coding system of around 30 codes, as shown in Appendix 4, was used to
categorise the data.

Thirdly, the contracts and interview transcripts were copied and analysed separately
and blind to the other’s judgement by the two legally qualified members of the project
team.53 The aim of this process was not to provide a definitive legal answer to the
question of any one individual’s employment status; given the uncertainty that
attaches to employment status, this is simply not possible in the case of non-standard
employment relationships. However, it was expected that a useful working
assessment of status could be made by researchers who were both legally trained and
also experienced in socio-legal aspects of interview work and in the interpretation of
documentary contractual material. Such an assessment could be regarded as an
approximation of the assessment of status which would be made by the relevant
judicial or administrative body, should the occasion arise in any individual case. As
just noted, in order to minimise the risk of subjectivity in the process of assessment,
each of the two members of the project team concerned carried out an initial
assessment independently of the other; the final assessments were then agreed
between them.

53
Simon Deakin and Catherine Barnard.

52
5.4 Characteristics of the sample

Characteristics of the sample of respondents are indicated in the Table 5.1, broken
down by the type of non-standard work that they reported themselves as performing
in the first wave of the survey. Also reported is the perceived permanence of their
employment.

TABLE 5.1 NUMBERS OF RESPONDENTS IN THE INTERVIEW SAMPLE, BY

TYPE OF CONTRACT AND PERMANENCE OF RELATIONSHIP54

Type of Contract Permanent Non-permanent Don’t

Know
Seasonal work 1 1 0
under contract for a fixed period or a 9 5 0
fixed task
agency temping 1 3 0
casual type of work 2 2 0
Homeworking 3 0 0
under a zero-hours contract 1 1 0
as an apprentice or trainee 1 0 0
none of the above 5 1 1
Don’t Know 0 0 0
Total 23 12 1

Table 5.2 breaks down the sample by individuals, indicating their (self-reported) type
of work, occupation, age and gender. The questionnaire was re-administered during
the second wave of interviewing, and in some cases different answers were given to
this question; these are also recorded in Table 5.2. The Case-study codes are those
used for the qualitative analysis of the interviews (see below, chapters 6 and 7).

54
This table includes the 10 focus group and 26 individual respondents. One
respondent was both a homeworker and on fixed contracts, hence the total amounts to
37.

53
TABLE 5:2. COMPOSITION OF THE SAMPLE BY TYPE OF WORK, OCCUPATION,
AGE AND GENDER
Case Type of work, Type of work, Wave Occupation Age Sex
study Wave 1 2 (if different from
code Wave 1)
6 Agency work Fixed term/task and Clerical worker 59 F
agency work
8 Agency work None* Clerical 34 F
Child minder in previous
job
27 Agency work Accounts assistant 45 M
35 Agency work None Clerical worker 26 F
33 Apprenticeship or Casual Nurse 24 M
traineeship
3 Casual Canteen worker 47 F
13 Casual Heavy goods mechanic 24 M
31 Casual None Stock controller 36 F
Market researcher
11 Casual Viewing assistant 59 M
1 Fixed term/task Freelance copy-editor 30 F
2 Fixed term/task Manager of charity 50 F
4 Fixed term/task Book keeper 37 F
10 Fixed term/task Child minder 49 F
14 Fixed term/task None Administrator 28 F
20 Fixed term/task None Family care worker 40 F
21 Fixed term/task Hospital assistant 35 F
23 Fixed term/task Teacher 42 F
25 Fixed term/task School inspector 57 M
26 Fixed term/task Clerical assistant 39 F
28 Fixed term/task Cleaner 28 F
29 Fixed term/task Hospital doctor 26 M
36 Fixed-term/task None Warehouse operative 37 F
17 Home working Mortgage consultant 35 M
32 Home working None Educational consultant 50 F
5 Home working & Freelance author and 40 M
fixed term/task editor
7 None Agency work Carer 50 F
9 None Health promotion 49 F
practitioner
16 None Construction worker 56 M
19 None Taxi driver 63 M
22 None Casual Cleaner 43 M
24 None Builder 37 M
30 None Fixed term/task Electronics engineer 31 M
Company director
15 Seasonal Mechanic 43 M
34 Seasonal Fixed term/task Construction worker 35 M
12 Zero hours Coach driver 51 M
18 Zero hours Process worker 26 M
* None refers to those who said that their job was not described by any of non-standard arrangements
listed in the question.

54
As Tables 5.1 and 5.2 indicate, the sample was highly heterogeneous in terms of the
types of work covered and the nature of the occupations. It also covered a wide range
of respondents by age, and consisted of approximately equal numbers of male and
female respondents.

Just over a third of those interviewed said that they were on fixed-term contracts. The
majority of these believed they were in permanent jobs, although around a third said
their jobs were temporary. They worked in a range of occupations. Professional
workers in the sample included a freelance copy-editor, a hospital doctor, and a
manager of a charity. Less highly skilled workers included a cleaner and a warehouse
operative.

The agency workers interviewed were predominantly in clerical jobs except for one
who was working as a carer. Most thought of their jobs as temporary. Four casual
workers were interviewed, two of whom thought their jobs were permanent. Their
occupations included both low skilled work such as a canteen worker and skilled jobs
such as that of a heavy goods mechanic. The three homeworkers interviewed were all
in professional jobs that they said were permanent.

Two interviewees were on zero-hours contracts, one as a coach driver and the other as
a process worker. One thought that their job was permanent while the other did not.
Two seasonal workers were interviewed, one a mechanic and one a construction
worker. Again one thought that their job was permanent while the other said that it
was temporary. The one apprentice was a male nurse who said his job was permanent.

The remaining interviewees did not state in the survey in what respect their job was
temporary. These worked in a range of occupations from a director of their own
printing company to a taxi driver, a cleaner and a builder.

5.5 Filters for ‘non-standard’ employment

As described in Chapter 3, in the quantitative survey once the self-employed with

employees, sole traders and directors of their own businesses had been filtered out of
the sample, respondents were asked about the permanence of their job and whether
they were working under any non-standard forms. As Tables 3.7 and 3.8
demonstrate, there were many apparently contradictory responses. Some respondents
reported that they were both permanently employed and working under types of
contract usually assumed to be non-permanent, such as fixed-term contracts, casual,
seasonal or zero-hours.

A small proportion of these apparent inconsistencies appear to have arisen because of

ways in which respondents interpreted the term ‘permanent’ (see section 3.8.2,
above). Four other groups of respondents were targeted for further exploration by
telephone to boost the sample of face-to-face interviews and explore how respondents
were understanding the terms used in the questionnaire. The four groups were: all
seasonal workers; all zero-hours contract workers; respondents who were both ‘under
contract for a fixed period or a fixed task’ and permanent; and respondents who were
not permanent but reported that they were ‘none of these’ to the list of non-standard
types of work.

55
5.5.1 Seasonal employment

Among those reporting to be seasonal workers, several were clearly using a very
different meaning for the term to that assumed in the literature. These were people for
whom the quality or quantity of work varied over an annual cycle, but who were paid
a fixed monthly salary. These included, for instance, an employee in a garage who
was busiest in August with sales of new cars, and outdoor workers who were
restricted by daylight in the hours that they could work in the winter. There were also
several cases of individuals in agriculture or tourism who were laid off for several
months of the year, but expected to be re-hired each year, and therefore reported their
employment to be permanent.

5.5.2 Zero-hours contracts

An attempt was made to contact all 13 respondents who reported being on zero-hours
contracts and had consented to a re-interview, but only four could be re-interviewed,
two by telephone and two face-to-face. Neither of the two who were interviewed by
telephone were correctly classified as being on zero-hours contracts. One worked
part-time, with a contract specifying 22 hours per week, but she often worked more
hours (for which she was paid) to cover for colleagues. The other was a charity
worker who received an annual honorarium of £1,000 to provide a service. Both
respondents reported that it was the interviewer who, after hearing their job
descriptions, decided that they were best categorised as employed on zero-hours
contracts. Other reports from respondents also suggested that, despite the ‘none of
these’ option, in some cases interviewers seemed to encourage respondents to select at
least one item in the list for Q5, even if none was applicable to them. While the
numbers of errors caused by this might have been trivial in the total sample of 4,006,
they would have formed a higher proportion of the sub-sample of 1,182 selected as
potentially non-standard.

5.5.3 Permanent and fixed-term contracts

Nine respondents from the initial survey who claimed to be both permanent and fixed-
term were re-contacted for a brief telephone interview. One claimed that, although
she signed a new contract every three years, this was entirely routine and the job was,
to all intensive purposes, permanent. Two were self-employed, and obtained work a
series of fixed-term contracts, interspersed with periods without work. One was not
sure whether his own contract was fixed-term or permanent, but the firm he worked at
had recently been set up to take over local council gardening services on a
competitive tender basis. The contracts were awarded to the successful tenderer for
periods of five years, and he thought it extremely unlikely that he would be retained if
their contract were not renewed. One was in a gap year between school and
university, so felt that she was only taking on the job for a fixed period, even though
the contract was apparently permanent. The other four seemed not to be on fixed-
term contracts at all. They were all in long-standing jobs of between 11 and 34 years’
tenure to date, all had permanent contracts and all felt confident that they could, if
they so wished, continue until retirement. As far as can be determined, they chose the
‘fixed term’ category as they thought that one came closest to their case (‘fixed’ and
‘permanent’ being semantically similar).

56
5.5.4 Not permanent and ‘none of these’

66 respondents out of the total sample in the quantitative wave responded that they
were not permanently employed, but also said that none of the types of non-standard
work applied to them. Seven of them were interviewed by telephone to clarify their
status. One was working as a nurse on an NHS ‘nursing bank’ which meant being on
call to meet the employers’ requirements as they arose (presumably on a zero hours
contract). Two attributed their lack of permanence to being self-employed. One
carpet fitter had thought himself non-permanent because he was ‘only’ working part-
time, but now (at the time of the telephone follow-up interview) considered himself
full-time and permanent. One woman was employed to sell houses on a new estate
where there was a clear understanding that she would no longer be employed when all
of the houses were sold (which would take years, not months). Yet she expected to
be re-employed when the builders had built another estate in her area. One woman
had been working in a new telephone sales branch, which was set up on an
experimental basis to see whether there was enough work to justify its existence.
Although no contractual or other documentation was provided to her, there was a
clear understanding that the branch might not remain open. Finally, one machine tool
operator had a permanent contract, but reported that his company had a policy of
preferentially choosing permanent staff over fixed-term staff for redundancy. He
explained that management believed that the legal minimum redundancy payments
were set to increase substantially as a consequence of pending European Union
legislation, and thus wanted to reduce their liability before this happened.

5.6 Test-retest reliability of the dependency questions

For all respondents re-interviewed or involved in the focus groups, the dependency
questions asked of them in the first, screening interview were re-administered. For
the five respondents who had changed jobs between the two interviews, this revealed
little about the usefulness of the questions, but for the 31 respondents who were in the
same job as before, the ‘test-retest’ data permitted us to evaluate whether respondents
were consistent in their answers. Closer inspection of the results revealed that three
of the respondents frequently responded differently on the two occasions. Two of
these respondents held multiple jobs, and it seems as if they were taking different jobs
to be their ‘main’ job on the two occasions. The other one had been a trainee and may
now have been upgraded. Table 5.3 shows the consistency of answers between the 28
respondents for whom we are confident that they are referring to the same job.
Respondents are divided into three groups: those giving totally consistent answers;
those who gave completely contradictory answers (for example, ‘no’ then ‘yes’, or
‘yes’ then ‘no’); and those who changed between ‘yes’ or ‘no’ on one occasion to
‘don’t know’ on the other.

TABLE 5.4 TEST-RETEST RELIABILITY FOR DEPENDENCY QUESTIONS

(NUMBERS OF CASES)

Question Consistent Partly Fully

Inconsistent Inconsistent
11 When you are working for a particular 17 4 7
client, organisation or agency, are you

57
allowed to take on work from other
clients, organisations or agencies or not?
12 Can you refuse work which is offered 23 2 3
to you by the clients, organisations or
agencies which you work for?
13 In your current work are you free to 23 1 4
hire other people, who answer to you and
are paid by you, to do work you have
personally taken on, irrespective of
whether you need to or not?
14 At work, are you covered by the 20 4 4
grievance and disciplinary procedures of
your clients, organisations or agencies
you work for?
15 In your current work, do you have to 24 4 0
provide the main items of equipment you
need to do your job, not just the small
items many employees provide for
themselves such as pens and so on?
16 When you are at work, does someone 27 0 1
have the right to tell you, at any time,
what to do or when and how to do it,
even if this rarely happens in practice?
Q17 Which of these best describes how 17 1 10
you are paid?
Q18 Do you pay your own national 24 0 4
insurance or tax, or is this usually
deducted by the organisation you work
for?
Q19 Are you entitled to receive sick pay 26 1 1
or paid holidays from the organisation or
agency you work for.

Question 16, concerning control, and Q19, concerning holiday and sick entitlements,
clearly performed better than the others in respect of consistency. At the other end of
the spectrum, only 17 out of the 28 (61%) were consistent in their answers to Q11
concerning mutuality of obligation, and Q1755 concerning method of payment. And
only 20 out of 28 (71%) were consistent on Q14 (integration). The others all achieved
consistency in the 82-86% band.

The in-depth interviews provided some details of the ways in which respondents
found these questions difficult to answer unambiguously. Respondents were not sure

55
On Q17 some of the inconsistencies were, for present purposes, trivial in as much
as they reflected movements within the forms of payment associated with
employment (monthly and weekly payments), or within forms of payment associated
with self-employment (that is, fees agreed in advance and piece rates). However,
there were still at least three respondents who shifted between these categories. This
seems to be largely caused by confusion between how and how often people are paid.

58
whether Q11 included second jobs or just referred to similar sorts of work. For Q12
several employees recounted anecdotes about when they had refused to carry out an
unreasonable order. Q13 was sometimes taken to include employees hiring other
employees as subordinates within their own department. Even the issue of the
provision of paid holidays is not clear; one employee had 8% of his income withheld
by the employer, to be paid when he was on leave.

Although the respondents’ interpretations of the questions might not reflect a careful
or accurate reading or understanding of them, the complex nature of the issues
involved means that the questions are longer than would be normal for standard
questionnaire items. While the issues they raise may (or may not) be more accurately
resolved by careful investigation through qualitative research (or in a court or
employment tribunal), in this form of wording they do not provide simple and
unproblematic information for the purposes of a questionnaire.

There is another way in which to investigate the reliability of the answers to these
questions, that is, by examining the individual cases where there is an apparent
inconsistency between the responses to these questions and an individual’s self-
reported employment status. Two questions were selected for this more detailed
scrutiny on the basis that they were perhaps most direct in differentiating between the
self-employed and employees, namely Q13 (hiring) and Q16 (taking orders).

Of the 31 re-interviewed respondents who had not changed job between the two
waves, there were 12 who had reported that they were self-employed and 19 who
described themselves as employees at the first wave.

The employees, with only two exceptions, were consistent over the two waves in
saying that, in keeping with their status, they could not hire others but that someone
else had the right to tell them what to do. Further questioning in the qualitative stage
revealed that one of these employees meant that he had a lot of freedom to recruit
other employees to his department where necessary (but hadn’t fully taken on board
the clause ‘who are paid by you’). The other also considered himself to be running an
autonomous group or department, and as such could not think of anyone who ever
might give him an order, but was in all other respects of clear employee status.

The self-employed were more problematic. When first interviewed, four of the 12
interviewees said that they could not hire others, but on the second interview only one
said that they could not hire others, and one replied ‘don’t know; (both of these had
replied in the affirmative on the first interview). The confusion and ambiguities were
primarily a distinction between what they thought was practical or possible, rather
than what they had the right or freedom to do. In some cases they said that they had
had bad experiences with taking on others, or that they just did not want to, or that it
was not a possibility because they just could not get someone with the rare skills
necessary (for instance translations into Welsh).

The question concerning taking orders gave even more mixed results; six of the 12
self-employed respondents said that someone did have the right to tell them what to
do or how to do it on the both of the waves; a further two answered in the affirmative
on one but not the other wave. Clearly for many of them their normal modus
operandi was to be told what needed doing (whether it was where to hang a door, or

59
what to write an article about, or what to proof-read, or what colour paint to use) and
when to do it. Whether these types of instructions constitute a right to give orders to
some-one or simply an offer of an exchange is arguable, but this distinction was not
apparent in the responses to these questions or in the qualitative interviews.

Nor is the question necessarily clear-cut in differentiating the dependent from the
independent self-employed. While some dependent self-employed indicated that they
were regularly told what to do by one other individual, at least one of the independent
self-employed respondents (such as a domestic builder) also said that they were
subject to orders, meaning from the customers themselves.

5.7 Conclusions

The questions commonly used to define non-standard work are themselves

problematic. A number of respondents did not follow the exact definitions used in the
questions for a variety of reasons. Some did not differentiate between their own
personal intentions and the nature of the contract or arrangement which they had. In
the case of interpreting the meaning of permanence, some related this to the nature of
contract signed but others related it to the viability of the firm they worked for.
Finally, in some cases confusion arose because people answered in terms of what they
thought was practical or possible, rather than what they what they had the right to do.

This does not necessarily have direct implications for the main aims of the present
research project. However, it does suggest that, given the normal LFS definitions of
atypical or non-standard work, some groups may be overestimated - for example
seasonal workers - while others may be greatly underestimated. For example, the
number on fixed term contracts may well be double that detected in the LFS, where
they would be excluded from the count of non-standard work as soon as they
answered that they were permanently employed. Other categories of individuals,
particularly those employed on zero-hours contracts, clearly give rise to problems
associated with measurement using the survey items here, but it is not clear whether
these would lead to an over- or under-estimate.

There were sufficient reliability problems associated with the questions used to test
dependency to suggest that respondents were having problems relating them to
practices in their own jobs. It is possible that better survey items might be developed
that would be more reliable – many of the questions broke several of the conventions
of questionnaire wording, being long (over 30 words), containing several sub-clauses,
and being ‘double-barrelled’ (for example asking about both grievance and
disciplinary procedures together). However, this was to some degree unavoidable;
while there is scope for improving the questions in future surveys, it seems as if a
large part of the unreliability of these questions is caused by the inherently complex
nature of the economic relationships that they are attempting to describe. Hence those
questions which gave rise to the greatest discrepancies were those concerning the
mutuality of obligation test which, as we shall see below, is difficult to apply in
practice and a source of considerable uncertainty in the classification of employment
relationships.

60
Chapter 6. Employment Status in Practice: Findings of the ‘Qualitative Wave’

6.1 Assessments of employment status

We are now in a position to evaluate the evidence gained from the qualitative wave of
data collection for our understanding of how the rules governing employment status
operate in practice. Altogether, there are three separate sources of information
concerning the assessment of the employment status of the 36 individuals who were
interviewed face-to-face in the qualitative wave:
• the self-assessments made by each individual when they were first interviewed for
the purposes of the survey (the quantitative wave) and again at the time of the
second interview (the qualitative wave), when the first questionnaire was re-
administered;
• the assessment of status made by the statistical analysis of the survey data derived
from the quantitative wave (see chapter 4, above); and
• the legal assessment made by members of the project team in the light of the data
obtained from the interviews carried out in the qualitative wave (see chapter 5
above for an account of the methods used here).

6.1.1 The nature of the assessments made

The nature of the assessments made differs in each of the three cases. The short
questionnaire used for the quantitative wave offered interviewees the following
choices for their self-assessment: employee, self-employed, working on a government
training scheme, and unpaid family worker. The second assessment, as we saw in
chapter 4 above, was based on a statistical analysis of respondents’ answers to the
short questionnaire. As Table 4.4 above shows, the data were analysed in two ways:
firstly according to the categories of employee, self-employed, government training
scheme/unpaid family worker and unclear, and secondly using the categories of
dependent labour, self-employed/independent labour, government training
scheme/unpaid family worker, and ‘still unclear’. The second analysis enabled us to
make an estimate of the number of ‘dependent workers’, that is, employees and the
‘dependent self-employed’, in the overall sample.

For the third assessment, the categories used were: employee, dependent self-
employed, and independent self-employed. Again, this classification was used in
order to divide the self-employed into those who would probably be ‘workers’ under
relevant employment legislation (such as section 230(3) of the Employment Rights
Act) and those who probably would not (for the meaning of the term ‘worker’ see
chapter 2, above). At this stage of the analysis, the former group were termed
‘dependent self-employed’ while the latter – those who were not workers – were
termed ‘independent self-employed’. The legal assessment also used the category
‘borderline employee/dependent self-employed’ to describe individuals who were
neither clearly employees nor self-employed.

Table 6.1 summarises the assessments made under each of the three headings for the
interviewed sample.

61
TABLE 6:1. ASSESSMENTS OF THE EMPLOYMENT STATUS OF RESPONDENTS’ INTERVIEWED FACE-TO-FACE IN THE
QUALITATIVE WAVE

Notes:

Under ‘type of self-reported contract’, entries in italics indicate a different type of contract being reported by the respondent on the occasion of
the second interview.

Under ‘assessed status’, dependent self-employed = a self-employed individual who is also a ‘worker’; independent self-employed = a self-
employed individual who is not a ‘worker’.

'In some rows, responses for more than one job are entered. The first set of entries refers to the job described by the respondent at the time of the
first interview (the 'quantitative wave'). Subsequent rows refer to a further job described at the second interview (the 'qualitative wave').

Case Self-reported type Occupation Age Sex self-reported status Assessment based on Assessment based on
study of work the quantitative wave the qualitative wave
code
1 Fixed term/task Freelance copy-editor 30 F self-employed Independent self- dependent self-
employed employed
2 Fixed term/task manager of charity 50 F employee Dependent labour employee

3 Casual canteen worker 47 F employee Dependent labour employee

4 Fixed term/task book keeper 37 F employee Dependent labour employee
5 Home work Freelance author and 40 M self-employed Independent self- independent self-
editor employed employed
6 Agency work clerical worker 59 F employee Dependent labour borderline employee or
Fixed-term/task dependent self-
employed

62
Case Self-reported type Occupation Age Sex self-reported status Assessment based on Assessment based on
study of work the quantitative wave the qualitative wave
code
7 None Carer 50 F self-employed Dependent labour borderline employee or
Agency work dependent self-
employed
8 Agency work clerical 34 F employee Still unclear borderline employee or
None dependent self-
child minder self-employed employed (agency)
independent self-
employed (child
minder)
9 None Health promotion 49 F self-employed Still unclear independent self-
practitioner employed
10 Fixed term/task child minder 49 F self-employed Still unclear independent self-
employed
11 Casual work viewing assistant 59 M employee Still unclear employee
12 Zero hours contract coach driver 51 M employee Dependent labour borderline employee or
dependent self-
employed
13 Casual heavy goods mechanic 24 M employee Dependent labour employee
14 Fixed term/task Administrator 28 F employee Dependent labour employee
None
15 Seasonal Mechanic 43 M employee Dependent labour employee
16 None Construction worker 56 M self-employed Still unclear borderline employee or
dependent self-
employed

63
Case Self-reported type Occupation Age Sex self-reported status Assessment based on Assessment based on
study of work the quantitative wave the qualitative wave
code
17 home working Mortgage consultant 35 M employee Dependent labour employee
18 zero hours contract process worker 26 M employee Dependent labour employee
19 none taxi driver 63 M self-employed Still unclear independent self-
employed
20 fixed term/task family care worker 40 F employee Still unclear employee
none
21 fixed term/task hospital assistant 35 F employee Dependent labour employee
22 none Cleaner 43 M self-employed Independent self- independent self-
casual employed employed
23 fixed term/task Teacher 42 F employee Dependent labour employee
24 none Builder 37 M self-employed Still unclear independent self-
employed
25 fixed term/task school inspector 57 M self-employed Independent self- independent self-
employed employed
26 fixed term/task clerical assistant 39 F employee Dependent labour employee
27 agency work accounts assistant 45 M employee Dependent labour borderline employee or
dependent self-
employed
28 fixed term/task Cleaner 28 F employee Still unclear employee
29 fixed term/task hospital doctor 26 M employee Still unclear employee
30 none Electronics engineer 31 M employee Dependent labour employee
fixed-term/task
company director independent self-
(printing company) employed

64
Case Self-reported type Occupation Age Sex self-reported status Assessment based on Assessment based on
study of work the quantitative wave the qualitative wave
code
31 Casual stock controller 36 F employee Still unclear employee
None
market researcher independent self-
employed
32 home working Educational consultant 50 F self-employed Independent self- independent self-
employed employed
33 apprentice or trainee Nurse 24 M employee Dependent labour borderline employee or
casual dependent self-
employed
34 Seasonal Construction worker 35 M self-employed Still unclear independent self-
fixed-term employed
35 agency work clerical worker 26 F employee Still unclear borderline employee or
none dependent self-
employed
36 fixed-term/task Warehouse operative 37 F employee Dependent labour employee
none

65
Table 6:1 shows that there were some variations between the respondents self-
reported status and the assessment based on the qualitative wave. Of the 24
individuals who said they were employees at the quantitative wave, 18 were finally
assessed to be employees and 6 were thought to be borderline employees or
dependent self-employed. Of the 12 who thought they were self-employed, 9 were
finally assessed as independent self-employed, one was thought to be dependent self-
employed, and two were classed as borderline employees or dependent self-employed.

Taking the assessments as a whole, we can identify a number of cases of uncertainty –

that is, cases in which it was not possible, at one or another stage of the analysis, to
make a clear assessment of legal status. There were also cases of discrepancy, that is,
cases in which the assessments made for the individual concerned, although clear in
themselves, differed between the quantitative and qualitative waves of analysis.
Summarising Table 6.1, these were as follows:

Uncertainties in the assessments made by the analysis of the quantitative wave: cases
classed as 'still unclear'

The analysis of the quantitative wave (see chapters 3 and 4 above) concluded that the
status of 13 individuals out of the sub-sample of 36 was ‘still unclear’ (36% of the
sub-sample). These were case studies 8, 9, 10, 11, 16, 19, 20, 24, 28, 29, 31, 34 and
35.

Uncertainties in the assessments made by the analysis of the qualitative wave: cases
classed as 'borderline'

In eight cases (Case Studies 6, 7, 8, 12, 16, 27, 33 and 35), equivalent to 22% of the
sub-sample of 36, a clear assessment could not be made after the legal assessment of
the qualitative stage. These were the cases classified as ‘borderline employee or
dependent self-employed’.

Discrepancies in the assessments made

To some extent, discrepancies are also a sign of uncertainty in the assessment process,
and so the division between ‘discrepancies’ and ‘uncertainties’ is not always clear cut;
there are also considerable overlaps between the two categories.

There were no differences or discrepancies in the self-assessments made by

respondents on the occasion of the two interviews. This was so even though some of
them changed jobs between the first and second interview, and some gave different
answers on the two occasions to the question of which type of non-standard work or
contract best described their situation. This suggests that respondents have a clear idea
of their legal status which remains the same whatever the particular type of non-
standard work which they are doing, and whether or not they frequently change jobs.

Discrepancies arising from the analysis of the quantitative wave were nearly all cases
where no clear classification could be made by these means, in other words, the case
was categorised as ‘still unclear’. For the purposes of our analysis, these cases are
better thought of as cases of uncertainty rather than clear discrepancies. The only
case of a clear discrepancy arising from the quantitative wave was Case Study 7, in

66
which the respondent assessed herself as self-employed but the analysis of the survey
data classified her as an employee. At the final stage she was classified as borderline
employee/dependent self-employed.

In the case of the final assessment which was based on all the evidence including the
qualitative data and documentary materials derived from the second interview, we
find a clear discrepancy only in Case study 1: here the respondent saw herself as self-
employed and the analysis of the quantitative data classed her as independently self-
employed, but the legal assessment was that she was dependent self-employed, that is
to say, a ‘worker’ within the meaning of section 230 of the Employment Rights Act.

We now examine these cases of uncertainty and discrepancy in more detail.

6.2 Uncertainties in the assessments made by the analysis of the quantitative

wave: cases classed as 'still unclear'

Of those classified as unclear by the quantitative analysis, five were classified in the
final assessment as cases of employees, five as independently self-employed, and
three as borderline employees/dependent self-employed. These last three cases are
considered in the next subsection together with the other illustrations of respondents
who were classified as ‘borderline employee/dependent self-employed’.

Of the five cases finally classified as employees, three were employed on fixed-term
or task contracts (Case Studies 20, 28 and 29). In each case, the reason they had been
classified as ‘still unclear’ at the quantitative stage was that at the first interview, they
reported having worked for between two and five employers in the past six months.
In other respects, their answers to questions relating to economic reality were more in
line with being dependent labour, that is, ‘workers’. At the second interview, when
the short questionnaire which had been used for the survey analysis was re-
administered, respondents in Case Studies 28 and 29 changed their answers to this
question, indicating now that they had only worked for one employer in the previous
six months. Thus the element of uncertainty here stems from the attempt to determine
employment status on the basis of the number of employers the respondent has had in
the recent past. While this can be a reliable guide in some cases, in the case of fixed-
term contract or task workers who move around frequently, it is not a definitive guide
since it is quite possible for them to be employed on a series of contracts some of
which may be contracts of employment and some of which are not.

Two others who were finally classified as employees but whose status was ‘still
unclear’ at the quantitative stage were casual workers (Case Studies 11 and 31). Case
Study 31 was classed as unclear largely because he reported at the first interview that
he paid his own income tax and national insurance contributions. At the second
interview, it was established that this was incorrect, and that his income tax and
national insurance contributions were deducted at source. The classification was
therefore largely caused by an error or misunderstanding at the stage of the first
interview. Case Study 31 was classified as ‘still unclear’ because she had worked for
more than one employer in the previous six months and stated that she was able to
take on work from clients and/or organisations other than her own employer.

67
Notwithstanding this answer, the more detailed analysis carried out at the qualitative
stage revealed that she was in essence employed as a full-time employee.

In case studies 9, 10, 19, 24 and 34 the final assessment was that the individuals
concerned were independently self-employed. Case Study 9 had worked for several
employers in the previous six months but failed one of the tests of independence,
namely being able to hire others who worked for her. More detailed analysis showed
that she was effectively running her own business even though the business had few
assets and she had refused to consider mortgaging the house to get a bank loan which
might have enabled her to expand. Although she reiterated at the second interview
that those with whom she worked were her co-workers and not employed by her, it is
also clear from her responses that she was responsible for directing their work.
Although she had one principal client, which pointed in the direction of employee or
at least ‘worker status’, she also carried out work for several others. On this basis, she
was classed as independently self-employed in the final assessment.

Case Study 10, a childminder, stated at the second interview that she had seven or
eight clients on average at any one time. At the stage of the initial interview she
stated she could not hire others to work with her, and also that she was paid on a time
rate. These answers were taken to indicate dependent workers status although, on a
more detailed analysis, she was clearly an independent worker. It is possible that she
said she was unable to hire others because of the practical (for example, financial)
difficulty of doing so.

Case Studies 24 and 34 were both construction workers who stated at the first
interview that they had worked for just one employer for the preceding six months.
From the data collected at the interview, it is not clear why Case Study 24 answered
this way; the reply does not tally with other details given of his work. The respondent
for Case Study 34 changed his answer at the second interview. In other respects, their
answers at both stages clearly indicated independent status: they could choose
whether to take work offered to them by an employer or client, they could hire others
to work with them, and they arranged their own payments of income tax and national
insurance contributions.

Case Study 19 was a taxi driver who, again, reported that he had only worked for one
employer in the previous six months. It appears from the data obtained at the
qualitative stage that this must have been a reference to the taxi company with which
the respondent was associated. However, it was considered at the final stage that the
taxi company could not be described as the employer of the respondent in a legal
sense. The company’s role was essentially to provide certain facilities (such as a
centralised booking service) for which the respondent paid a fee. It did not guarantee
to provide either work or income to the respondent. On this basis, it was concluded
that the respondent was a genuinely independent self-employed individual.

Our examination of the above cases illustrates some of the limitations which are
inherent in the use of quantitative analysis to make a determination of employment
status. The quantitative stage of analysis involves a trade-off between these
limitations and the advantages of obtaining a large representative sample of
respondents.

68
Firstly, some of the questions posed at the quantitative stage were unavoidably
ambiguous as means of identifying whether respondents satisfied one or more of the
various factors which determine employment status. As we saw in chapter 5 above,
some ambiguity is unavoidable because those factors themselves are, to a high degree,
ambiguous.

Secondly, the quantitative analysis did not allow for the answers to the individual
questions to be weighted. In practice, though, a court or tribunal is highly unlikely to
place equal weight on each of the factors which could be taken into account.
Similarly, at the final stage of assessment, a judgement could be made of the weight
likely to be placed on such factors, based on the existing body of case law. This
process of weighting is inevitably impressionistic to some degree; the idea of an
‘elephant’ test (see section 2.8, above) cannot be entirely avoided. This accounts for
some of the divergencies between the assessments made at the quantitative and
qualitative stages.

6.3 Uncertainties in the assessments made by the analysis of the qualitative

wave: cases classed as 'borderline'

Eight cases in total could not be clearly classified in the final assessment; they were
considered to fall on the borderline between the status of employee and that of a
‘dependent self-employed’ person, that it to say, a self-employed ‘worker’. Five of
these were agency workers (one of whom was also a trainee); the other two were
employed on arrangements containing elements of false self-employment.

The difficulty in classifying the status of agency workers stemmed largely from the
application of the mutuality of obligation test. Since the agencies56 concerned did not
necessarily guarantee to find work for their workers between jobs, the working
arrangements could be said to give rise only to a contract for services. The
documentation given to these agency workers appeared to have been issued with the
aim of deflecting employee status. Under these circumstances, it became highly
problematic to predict whether a court might imply a term requiring the agency to
supply work if it was available (in return for the worker undertaking to accept it), and
hence establish the necessary conditions for mutuality of obligation.

In Case Study 6, several specimen contracts with agencies were obtained. The
specimen contracts all contained clauses which appeared to be aiming at the exclusion
of employee status by a variety of means. One purported to deny that the arrangement
constituted a legally binding contract of employment between the worker and either
the agency or the user, by stating instead that the arrangement was a contract for
services (in other words, an arrangement for self-employment). This is an example of

56
Under the Conduct of Employment Agencies and Employment Businesses
Regulations 1976 (SI 1976/715), the correct term for an agency which hires out
workers for particular jobs is an ‘employment business’; an ‘employment agency’ is
an organisation concerned with the recruitment and placement of workers with other
employers. In the text, however, we use the term ‘employment agency’ to refer to the
former kind of business, since this term is in wide use.

69
what may be called the ‘relabelling’ technique. Another stipulated that the worker
was not required to accept any assignment, nor was the agency required to find an
assignment for the worker. This is an example of the use of express terms to deny the
existence of mutuality of obligation (the ‘mutuality’ technique).

The terms used in these contracts were very similar to the contractual documentation
analysed by the Employment Appeal Tribunal in the Pertemps case.57 That contract
contained the following clauses:

• Temporary workers are engaged under a contract for services…

• The Employment Business agrees to offer to the Temporary Worker

opportunities to work as a … where there is a suitable assignment with a
hirer… requiring such a Worker. The Employment Business reserves the
right to offer any assignment to such Temporary Workers as it may elect
where that assignment is suitable for several workers…

• There is no obligation by the Employment Business to provide or the

Temporary Worker to serve any normal number of hours in any day or
week. In the event of the Temporary Worker declining the accept any offer
of work, or failing to attend work for any reason, for any period the
contract shall terminate…

In Pertemps, the EAT considered that the terms of the written contract, taken as a
whole, were an important indicator that the agency worker concerned was not an
employee.

The contractual documents in Case Study 6 also indicated that once each assignment
was accepted by the worker, she came under a number of stated obligations, for
example to cooperate with the user’s staff and to follow the work rules of the user.
Making such obligations explicit could be seen as evidence that the relationship with
the agency was not based on a contract of employment, for the very reason that such
obligations are normally implied into such a contract; if the worker were truly an
employee, it may be argued, there would be no need to spell them out.58 On the other
hand, it could be said that there is nothing to prevent an employer (or any other party)
seeking to make explicit what would otherwise be implicit. On this basis it could be
said that spelling out the obligations of obedience and care does not help to settle the
issue of status either way.59

Leaving this point to one side, the presence in the contract of obligations of obedience
and care could be seen as pointing towards the presence of ‘control’. However, it is
inherently ambiguous, in the context of an agency relationship, whether such control
is to be exercised by the agency or by the user. It is possible to regard to agency as
57
Unreported, 1 July 1993, Appeal No. EAT/496/91.
58
This was the view of the EAT in the Pertemps case, ibid.
59
This was the approach taken by the Court of Appeal in McMeechan v.
Secretary of State for Employment [1997] IRLR 353.

70
ultimately having a sufficient degree of control over the way in which the worker is to
work during the assignment.60

Moreover, in the case of an agency worker it is unlikely, on balance, that the ‘global’
arrangement would constitute a contract of employment. This does not rule out the
possibility that there is a contract of employment during each individual assignment
but again the point is not clear.

In Case Study 6, the respondent herself saw her relationship with the agency as
flexible on the issue of stability of work:

‘ So [X] if you wanted to have six months off.’

‘I could.’

‘You could, the agency wouldn’t have any hold over you at all, apart from
threatening that they wouldn’t employ you. Does that ever happen, or perhaps
you wouldn’t get such good jobs?’

‘They just don’t call on you. Like where I turned down the twelve-month
contract, I said is this going to affect me? They said no and in fact they rang
up the other day so it obviously hasn’t.’

‘Does that worry you a bit?’

‘Really with an agency, you have to keep in touch with them anyway,
otherwise they could forget.’

Case Studies 7, 8, and 35 raised similar issues; in each case, the respondents referred
to the flexibility and open-endedness of their arrangement with the agency as offering
them certain benefits, as well as disadvantages in terms of insecurity. In Case Study
27, by contrast, the respondent was employed for long periods on a job which
involved working alongside permanent staff doing very much the same job (he was an
accounts assistant) in a large organisation. He would have preferred to have a more
stable, ‘permanent’ job:

‘Would you rather have a job with a firm or are you quite happy working for
an agency?’

‘I would prefer a permanent job, yeah, if [the client] said OK we’ll take you
on permanently, I would say yes.’

‘What would the advantages be?’

‘A more stable sort of environment and you would also have a pension scheme
and … the benefits that go with a company.’...

60
See Mersey Docks and Harbours Board v. Coggins and Griffiths (Liverpool)
Ltd. [1947] AC 1.

71
He also saw himself as under the control of the agency:

‘Do you look upon yourself as being basically your own boss, that you can
decide whether you want to stay with them or not?’

‘No, I wouldn’t say I am my own boss because you know you are working for
a firm - if it wasn’t them, it would be another agency.’

In this case, the greater stability of the working relationship could be seen as pointing
towards employee status, although, again, there can be no certainty over the
implication of a term under which the agency was required to offer work and the
respondent required to accept it.

Case Study 33 concerned a trainee nurse who was also an agency worker. Although
still at a probationary stage, he reported that he was employed by an agency and sent
to work for a given period for a client or ‘user’. Under his terms of employment, he
was not guaranteed work from the agency; this pointed away from him being an
employee.

Case Study 12 combined elements of zero-hours working and false self-employment.

The respondent, a coach driver, had initially been employed on call, but with a
contract of employment. Since the initial interview he had been put on a ‘self-
employment’ arrangement and now invoiced his employer/client for each job he
carried out. He continued to work exclusively for that company but at the time of the
second interview was in the process of setting up his own business and was applying
for a licence to do so. At this time, it could have been argued that, given the element
of exclusive employment and the past history of employee status, the new
arrangement would have been classified by a court or employment tribunal as a
‘sham’. The initial survey classification of ‘employee’ may be justifiable on the
answers given to the omnibus questionnaire, but the detailed case study evidence
shows that it is impossible to make a classification here with any degree of certainty.

Case Study 16 involved a construction worker whose position resembled that of an

employee. He had long service for one employer, saw himself as under a duty to obey
orders, worked more or less exclusively for one organisation, was employed on a
regular and continuous basis, and was under a duty to accept work when offered. He
also regarded himself as an employee in all but name, even though he had reported his
status as ‘self-employed’ at the interview for the initial survey. The initial survey
assessment of ‘still unclear’ seems justified in this case: a high degree of uncertainty
must attach to the classification of a case of this kind.

72
6.4 Discrepancies in the assessments made

Case Study 1 represents a case of what may be called a clear but ‘partial’ discrepancy.
The respondent regarded herself as self-employed and the initial survey analysis
classified her as ‘independent self-employed’. The final legal assessment was that she
was a dependent self-employed worker; in other words, she was more likely to be a
protected worker within section 230(3) than a genuinely independent self-employed
individual.

This respondent worked as a freelance copy-editor for a publishing house. Under the
contract issued by the publishing house, a firm distinction was drawn between
freelancers and employees; the documentation referred to the importance of this
distinction from the point of view of complying with Inland Revenue and
Contributions Agency rules on tax and National Insurance contributions respectively.
The documentation spelt out the employer’s view of the relationship, which was that
the distinction between freelances and employees was part of what made freelance
work attractive to both sides, and specified reasons for the employer’s understanding
that the respondent was not employed as its employee: as a freelancer she would not
normally work on the employer’s premises; she controlled when the work was done,
within agreed completion dates; the contractual relationship between the employer
and the respondent only began when an order was placed and ended when the
employer paid for the completed work; the respondent did not have an email,
telephone number or desk at the employer’s place of work; and she was paid only on
presentation of an invoice for work done.

The documentation contains elements which could have the effect of excluding the
possibility that there was ‘mutuality of obligation’ between the parties. The
respondent’s account of her work in the interview was ambiguous from this point of
view. When asked whether the employer had a duty to provide her with work she
replied ‘absolutely’, but she did not think that she had an overriding contractual
relationship with the employer.

The fact that the respondent worked consistently for a single employer in this way
could weigh in favour of employee status; it could be argued that she had employee
status during each assignment, but on balance it is unlikely that this was the case. It is
extremely doubtful that she had a global contract of employment in between jobs. She
also stated in her initial survey interview that she could if necessary hire others to
work with her, which would count strongly in favour of her being independently self-
employed although she did not appear to exercise this right. On the other hand, her
dependence on a single client or employer, when coupled with her agreement to
provide her own personal services, indicate that although self-employed, she would
very probably be covered by the extended definition of ‘worker’ and so qualify as
dependent labour.

73
6.5 Identifying the sources of uncertainty in the application of the legal tests
of status

On the basis of our analysis of the preceding cases, the following factors can be
identified as principal sources of uncertainty in the application of the legal tests of
status:

• attempts by employers to shift the status of workers from that of employees to

independent contractors through methods ranging from ‘relabelling’ to changes in
working arrangements and in methods of payment;

• the use by employers of contractual clauses (which we may usefully call 'status-
denying clauses') to point away from employee status, when other factors might
have pointed towards employee status or were ambiguous; and

• the difficulty in knowing whether a court would be prepared to imply a term for
mutuality of obligation.

These factors tended to interact in particular cases, above all those involving agency
workers and workers employed under highly casualised contractual arrangements.

Clear cases of false self-employment in which the employer attempted to change the
status of the employee by attaching a different ‘label’ to the relationship were less
common than situations in which employers seem to have been trying to take
advantage of the case-law on mutuality, by altering the contract terms in such a way
as to lead a court or tribunal to regard the relationship in terms of a contract for
services. Hence the evidence from the qualitative wave suggests that there may be a
degree of iteration between the development of the mutuality test by the courts, and
attempts by drafters, acting on behalf of employers, to exclude employee status by
denying any obligation upon the employer to provide work - in other words, the use of
‘status-denying’ clauses.

For example, a contract entered into by an agency worker with her agency contained
both types of terms – a ‘relabelling’ clause and one designed to exclude mutuality of
obligation - a ‘status-denying’ clause. The first of these two clauses stated that ‘for the
avoidance of doubt, these terms shall not give rise to a contract of employment... and
therefore the [worker] will not have the statutory rights accorded to employees’. This
was unlikely to be effective in excluding employee status since it simply represented
the view of one of the parties as to the agreement’s legal effect.

The relevant other term stated:

‘The [worker] acknowledges that it is the nature of temporary work that there
may be periods when no suitable work is available and agrees (a) that
suitability shall be determined solely by the [agency and user] and (b) that the
[agency and user] shall incur no liability towards the [worker] should they fail
to offer opportunities to work...’.

This would be a more substantial obstacle to employee status, since it raised a

presumption that there was no mutuality of obligation. As we saw in chapter 2, a

74
court or employment tribunal could conclude, on evidence of a regular and stable
working pattern, that there was an implied obligation to provide work, but this
becomes highly problematic when an express, written term of the contract clearly
states otherwise. Even if the individual can show that he or she regarded themselves
as under a de facto obligation - which, on the basis of evidence from the interviews, is
common - this may or may not translate into a legal obligation in the eyes of the court.
Moreover, it is often only through costly litigation that the issue can be settled.

The use of contractual documentation to tilt the balance away from employee or
dependent worker status also leads to a divergence between the self-perception of the
individuals concerned and the assessment which a court or tribunal is likely to make.
As Table 6.1 shows, there was, on the whole, a substantial degree of correspondence
between the self-assessed status of the respondents and the final assessments made in
the light of the evidence from the qualitative wave of analysis. However, there were
significant discrepancies in the case of agency workers. In Case Studies 6, 8, 33 and
35, agency workers reported themselves to be employees although, as we have just
seen, because of the application of the mutuality test, their status would not be clear.
Only one of the agency workers reported herself to be self-employed.

Most of the agency workers interviewed, then, appeared to regard themselves as

employees at the same time as appreciating the insecure nature of their relationship
with the agency. The respondents themselves did not see their insecure contractual
position as affecting the question of whether they were employees or self-employed.
Nor did they appear to see employee status as incompatible with the degree of
flexibility which agency work offered to them.

6.6 Conclusion

Analysis of the data from the qualitative wave showed a substantial degree of
correspondence between the self-assessed status of the respondents and the final
assessments made in the light of the evidence from the qualitative wave of analysis.
However, for a quarter of the sub-sample there were discrepancies but in most cases
these were due to uncertainty in the final assessment. For a substantial proportion of
the sub-sample (22% of cases) it was not possible to make a definitive judgement on
whether the respondents were employees or self-employed.

The principal causes of uncertainty in these cases derived from attempts by employers
to use contractual devices of various kinds to tilt the balance away from ‘employee’
status. These included ‘relabelling’ clauses which described the working relationship
as one of self-employment and ‘mutuality’ clauses which specified that the employer
was under no obligation to provide work on a continuing basis to the individual
concerned. Clear cases of relabelling were less common than situations in which
employers had used status denying clauses. This suggests that there may be a degree
of iteration between the development of the mutuality test by the courts, and the use
of ‘status denying’ clauses.

Agency workers and workers employed on highly casual working arrangements

(including zero-hours contract workers) were the groups most likely to be subject to
this degree of uncertainty. Agency workers on the whole perceived themselves to be

75
employees of their agencies, notwithstanding the absence of a formal legal
commitment to find them work. They did not see their insecure contractual position as
affecting the question of whether they were employees or self-employed.

All the cases classified as ‘borderline employees’ at the qualitative stage would
probably have come under the extended definition of ‘worker’ which is that used in
section 230(3) of the Employment Rights Act 1996. In other words, although self-
employed, they contracted to provide their own personal services and had a
sufficiently high degree of economic dependence on one employer to come under the
extended definition. By contrast, the key characteristics of individuals who were
genuinely self-employed were the large number of clients for whom they worked, and
the ability to hire others to work with them (or in substitution for them).

76
Chapter 7. Employment Status and the Experience of Non-standard Work

The interview data enable us to build up a picture of how the individuals in the sample
perceived the process of choosing to work in a particular way, and to what extent they
saw self-employment, ‘non-standard’ work and ‘regular’ employment as distinct
categories with advantages and disadvantages. We consider in turn the reasons
respondents gave for choosing to work in a particular way; the perceived advantages
and disadvantages of different forms of work; and whether they saw their choice of
working arrangements in terms of a trade-off between these advantages and
disadvantages.

7.1 Reasons for choosing particular forms of employment

Respondents gave a range of answers concerning the nature and degree of choice
which they had had in deciding on their present form of work. Factors conditioning
choice included the need to fit in with family arrangements; the cost of retraining
following time spent out of full-time work; the time and complexity of setting up a
business; discrimination against older workers; and the unavailability of regular work.

Achieving greater control over working time so as to fit in with commitments outside
work was a major reason given for adopting non-standard working arrangements: a
typical comment was, ‘what I like best about the job is the flexibility it gives me to do
my work anytime I like’ (Case Study 1, freelance editor). At the same time, sex,
family responsibilities and the age of individuals were important factors influencing
the decision to take up non-standard work. The need to fit working arrangements
around childcare requirements was one of the principal factors which motivated those
who had taken up childminding:

‘I had a good job once! I started doing this when I had my own children, you
see, and it’s difficult when you’ve got children. I didn’t want to put mine with
a childminder, so you incorporate an extra one into the home and when your
children are at school, you need a job that will put you here when they’re here
and in school holidays.’ (Case Study 10, childminder).

The same point was made by Case Study 8, also a child minder.

Age discrimination was also cited as a factor in the context of a return to full-time
employment after a break for child-rearing. The difficulty of getting a regular job
made agency work attractive for one individual in this position:

‘Initially I did hope to get another permanent job, either full-time or part-time,
but I knew my age was against me. I really think equal opportunities is a
name only .... that’s with age, disablement. Yes, they’ll call you for an
interview because there are equal opportunities. But you know full well that
you are not going to get it. So I am quite content with now, the way I am
working and making a life for myself and paid of course.’ (Case Study 6,
agency worker, clerical.)

77
The respondent was in her late 50s and had returned to work after a break of almost
20 years raising her children. She was also registered as disabled.

Many respondents saw non-standard work as inferior to regular working

arrangements. Agency workers and fixed-term contract workers, in some cases, took
these forms of work because permanent work was not made available to them (Case
Studies 27 and 26 respectively). One respondent had been dismissed by his employer
and re-employed on a self-employed basis. Once this happened he decided to set up
his own business but had not previously considered doing so (Case Study 12).

7.2 Advantages and disadvantages of forms of employment: perceptions of

autonomy and control

Respondents tended to associate autonomy and flexibility with self-employment, on

the one hand, and control and commitment with employment as an employee on the
other. However, the nature of these associations was complex and the divide between
employment and self-employment was not always clear-cut.

Some freelance workers and self-employed valued the right to organise their work as
they saw fit. One stated: ‘I like to be in control, so the advantage of being self-
employed [is] there - I was in control’ (Case Study 8, child minder). Another said:

‘I like the fact that I’m my own boss. To a certain extent, you set your own
hours because if people come and they want you to start at 6 in the morning if
they’re nurses or they do shift work, you can always say no, you don’t work
those hours. It’s nice to be at home in the summer. It’s nice to call the tune,
basically’ (Case Study 10, child minder).

Freelance workers appeared to value the autonomy to arrange a pattern of working

which suited their needs:

‘I find that I have developed a way of working, over a series of fifty or so

contracts, that I feel very happy with. I wasn’t given guidance, but I didn’t
find that a problem.. I could work out the best way to do things and interact
with the office...I don’t even have to tell them [clients] really. I just take the
product that they give me, and then return it to them and hopefully that’s what
is required, and then I’m done.’ (Case Study 1, freelance copy editor).

These perceptions have something in common with the idea, implicit in the ‘control’
test, that the degree of freedom which an individual has over the way in which they
work is an important indicator of employment status. However, a drawback of the
control test - namely that many employees, and not just the self-employed, see
themselves as enjoying a degree of autonomy - also came out in the interviews.
Hence, for a senior manager of a charity, the control exercised by her board of
trustees was a matter of degree:

‘It is not they who control me, it is a partnership, I go to them with proposals
and papers and discuss things with them and then they say, carry on... I don’t
see myself as needing to be autonomous or needing to do my own thing

78
 because I ought to be accountable to the people we are there to serve who are
all those voluntary organisations.’ (Case Study 2)

Similarly, a viewing assistant, who was clearly employed as an employee (among

other things he had a written contract of employment) commented:

‘I feel I’m my own boss, yes. I’m obviously not my own boss because I’ve
got a manager, but the manager is so flexible with me I feel as if I’m my own
boss’ (Case Study 11).

In practice, then, perceptions of control and accountability are much more complex
than a simple division between employment and self-employment would imply. The
self-employed may come under pressure to accept work from particular clients and
may have to operate to very tight deadlines. Similarly, employees in non-standard
work, such as zero-hours or on-call contracts, often see themselves as being required
to respond to the employer’s demands, even if their contracts suggest that they may
have the right to turn down work offered to them. Freelance workers and others who
had no formal commitment to a regular client, and had no formal expectation of
receiving continuous work, also said that in practice it was often difficult to refuse
work:

‘So what you’re basically saying is you are free to decide when you work and
for whom’

‘Well, I’m free inasmuch as I’m not free - well, to be honest, no - I’m free to
decide when I work, I’m not free to decide who I work for at the moment
because I have to take whatever I can get. I always pretty much have had to
take whatever I can get. So I’m not free to decide who I work for. (Case
Study 5, freelance author and editor.)

Working time flexibility could also cut both ways. Although, as we have seen, many
respondents saw such flexibility as an advantage of non-standard employment, very
long hours, unsocial hours working and variability of working hours were problems in
practice both for employees and for the self-employed. For certain employees,
including managerial and professional workers, hours were perceived as being
flexible in the employer’s favour. A charity manager said that while her contracted
hours were 37 a week, she normally worked 60 or more hours, with adverse
consequences for her health:

‘Yes, its 37 hours in theory and flexible. If I do evening and weekends, in

theory I have time off in lieu. One of the reasons I am not well now is that I
have been overdoing it over 2 years and I look around the voluntary sector and
see a lot of people under stress and getting ill because they are trying so hard
to make ends meet and do the job with limited resources and be everything to
everybody and fulfil bottomless demands – so you end up doing excess hours.’
(Case Study 2.)

Similarly, a schoolteacher on a fixed-term contract commented on the lack of fixed

hours as follows:

79
 ‘Well, I do like it but sometimes, after I’ve had a holiday in particular, but like
I was saying… you can never be a normal person. For example, if friends call
unexpectedly, it would be very difficult - my grandmother was ill last year and
because I had to go and see her in the hospital, I had to stay up till 3 and 4 in
the morning to do my marking! Because you couldn’t go into class and say to
GCSE children, I’m not marking your work because - you just can’t do it. So
that is a source of irritation sometimes.’ (Case Study 23.)

Variability of hours could also be a problem for part-time workers. A part-time

employee, initially taken on for fifteen hours per week, found herself taking a cut in
hours when the firm was ‘going through a fairly rough patch’ (Case Study 4).

7.3 Advantages and disadvantages of forms of employment: perceptions of

insecurity and risk

With respect to insecurity, on the balance of evidence of the interviews there is again
no straightforward division between the perceptions of employees, who might be
expected to have feelings of security, and those of the self-employed, who might be
expected to accept insecurity in return for the prospect of greater reward.

For example, some self-employed workers were not necessarily more insecure than
employees since they were often able to meet concerns about insecurity on the basis
of their reputations:

‘How important is stability for you in having an income, a career structure, a

pension, working conditions?’

‘Well, I’d like them all but I haven’t got any of them at the moment! I guess
my stability in that job was based purely on my reputation, which was always
good and I never, ever had available places.’ (Case Study 8, child minder.)

By contrast, many employees had concerns about the inherently insecure nature of
their jobs, concerns which were largely outside their control. This was particularly the
case with fixed-term contract workers. One commented on

‘the anxiety and insecurity of it all and I can keep telling myself, well, in the
commercial sector this is the fact of life, if you don’t sell what you do, you
don’t survive… We have to sell what we are doing to fund us and if we don’t
do that successfully then we don’t deserve to carry on’ (Case Study 2, charity
manager).

A hospital doctor employed on a fixed-term contract as part of his training referred to

a growing lack of security of employment:

‘medicine’s always been looked on as having a lot of job security, but I think
as things are changing and the market environment’s creeping in then that
really is going’ (Case Study 29).

80
Being employed on a series of fixed-term contracts without a clear understanding of
the legal position61 was also a source of concern for employees:

‘I was a bit concerned but I didn’t want to mention it because I didn’t know
whether my at the time my boss knew about the two year rule and I thought
well, if he keeps me on longer, I’ll keep quiet! But he knew - he obviously
knew about it - because I saw a note from him asking the admin. person there -
when is [X’s] second anniversary, you know.’ (Case Study 26, clerical
assistant.)

A hospital doctor employed on a three-year fixed-term contract commented:

‘It gives you a bit of stability but even three years isn’t that long and after that,
I may not be able to get a job at all. It’s not common in medicine, but it does
happen’ (Case Study 29).

Agency workers, likewise, had concerns with insecurity and also with difficulties of
working alongside permanent staff who were paid more than they were:

‘Those sorts of people - the ones that don’t like agency workers - have they
indicated why they don’t like them?’

‘I think it was because they felt threatened first of all, because they were
actually training us to do what they were doing and human nature being what
it is, you’re going to think - I’m going to train these, they’re paying them less
than me, I’m going to be out! I think once - they have had redundancies but it
was voluntary and the people who wanted to go went. The other people are
still there. And I think over the time of being there, their fears gradually eased
off.’ (Case Study 26, accounts assistant.)

For agency workers, resentment against the permanent staff was tempered by the
feeling ‘if you weren’t doing this, you could be on the dole’ (Case Study 26). Even
then, there was the possibility that the employment relationship could be brought to an
end at very short notice; the same respondent described how

‘one Monday morning we were working in the city centre and as we went in
there were two of our [agency] supervisors there and they had a list and saw
who was coming in… They were just literally stopping people at the door,

61
As explained in chapter 2 above, an employee employed under a fixed-term
contract may lawfully waive his or her entitlement to claim unfair dismissal or
redundancy compensation upon the expiry of the contract (Employment Rights Act
1996, section 197); in addition, an employee, whether employed under a fixed term or
not, may not acquire general protection against unfair dismissal, or the right to
redundancy compensation, until they have acquired two years of continuous
employment. In practice, the application of these principles to the position of fixed-
term employees who work under a series of renewable contracts is highly complex.
See Deakin and Morris, 1998: sections 3.7.1, 3.7.2, and 5.4.7.

81
 sending them over, and I think in one go they got rid of about 150 people that
morning.’

Respondents on both sides of the employee/self-employed divide expressed concerns

about health and safety. An agency worker reported difficulty in obtaining
compensation after suffering an injury at work for which neither the agency nor the
user would take responsibility. A self-employed construction worker commented on
the high health and safety risks facing subcontractors on building sites:

‘… a lot of companies do this now. They put a scaffold up for brickwork or

for people to put the roof on and then you want to go on the scaffolding after
to put the window in, as soon as they’ve finished the roof… because they’ve
put the scaffolding up for the roofers, they’ve put a sign on it saying “not
suitable for anything else”. So if you go on after them and you fall off that
scaffolding, it clears them of liability for your injuries. Which is common
practice now.’ (Case Study 34)

A majority of the respondents also associated the work they were doing with an
absence of long-term financial security. Several agency workers, casual and zero-
hours contract workers, fixed-term contract workers and self-employed construction
workers had no access to an employer’s pension scheme and had made no provision
for themselves.

7.4 Making explicit ‘trade-offs’ between employment and self-employment

Some respondents explicitly identified with the idea of a ‘trade off’ between the
stability and security of employee status, on the one hand, and the greater autonomy
of self-employment. Freelance workers and the self-employed contrasted the
‘control’ to which they had had to submit when they were employees, with the greater
freedom but also the responsibility which being self-employed entailed:

‘The advantages of being employed over self-employed are obvious in that

you’ve got a regular income coming in and the buck doesn’t stop with you!
Although I’ve always had fairly senior positions, there was always somebody
who I could offload on to, if you like. And so there were advantages in that. I
was made redundant once when the company closed down… it was
completely out of my hands and I don’t like that. I like to be in control, and
that’s the advantage of being self-employed - I was in control. And of course,
I’ve worked for large companies and small companies.’ (Case Study 8, child
minder.)

‘The only good thing about working for an employer is that you don’t take any
of the strain, you start work at 8.30 and you finish at 5.30 - you don’t have to
worry about what happens overnight. Here, I sort of think “Oh, you know,
who’s got the keys to the next venue”, because we swap around, and like
today, I was worried about [XXX] because she was very quiet this morning
and I thought “Oh, I’ve upset her”.’ (Case Study 9, health promotion worker.)

82
The downside of being self-employed was, according to one respondent, ‘being poor’
but the advantages included: ‘you do the job and as long as you’re doing the job,
people don’t worry about what you’re doing at that precise moment they walk past
your desk’ (Case Study 5, freelance author and editor).

For employees, the advantages included stability and predictability of income, and
also the fact that the employer is prepared to take on certain responsibilities. The
following comment was made by an agency worker (who considered herself to be an
employee of the agency although her status was probably unclear, see chapter 6
above):

‘The agency is responsible for you and they are employing you in fact, not the
person they send you to. So if there is a problem, the agency is duty bound to
sort it out… Presumably if something happened to you at a place of work, you
fell down the stairs and broke your leg, it would be the agency’s problem.
Their responsibility is to me as mine is to them.’ (Case Study 6.)

The downside of being an employee, however, was not simply a degree of

inflexibility over working arrangements, but also, in many cases, the lack of any
compensating security:

‘As an employee you have set times and you are tied down to that job and in
all honesty employers mess you about, because they know that if they were to
get rid of you within a matter of days they could get someone else doing the
same job. So they don’t care for the worker like they should because they
know there is someone there to replace them and that is one of the reasons
why I set up on my own. I used to always get laid off at Christmas time.’
(Case Study 24, self-employed builder.)

7.5 Conclusion

The qualitative wave provided evidence on individuals’ reasons for choosing

particular forms of work and their perceptions of flexibility and autonomy, on the one
hand, and of insecurity and risk, on the other. Some respondents saw the advantages
and disadvantages of particular forms of work in terms of trade-offs between
flexibility and security, suggesting that they exercised a degree of choice in weighing
up which form of work to adopt. In numerous cases, however, the choice of non-
standard work was seen as influenced and constrained by external pressures, the most
important of which were family commitments, retraining costs, age and disability
discrimination, and the lack of availability of alternative work. In particular, for
those with family obligations, it was a matter of necessity to find employment which
offered them the opportunity to arrange their work around domestic commitments.
Those returning to employment after a period of unemployment or after family
commitments chose non-standard work because of the costs of acquiring or re-
acquiring skills of the kind needed for a more stable and permanent position. There
was a perception that it was easier for older workers to get employment with an
agency than with an employer looking for a longer-term commitment.

83
Although many respondents clearly identified particular advantages and
disadvantages with the form of work in which they were engaged, there was a
blurring of the division between standard and non-standard work, and between
employment and self-employment. Hence self-employment could result in a
considerable restriction of personal autonomy in practice and to long and intense
working hours, thanks to the need to meet tight deadlines and maintain reputation
with clients. Employees in non-standard employment, conversely, commented on
growing insecurity and stress caused by uncertainty over their future job prospects.
Both employees and the self-employed reported being affected in different ways by
financial insecurity. Agency workers and the self-employed often had no access to
pension schemes, and this was also a problem for employees working on fixed-term
or task contracts.

In short, because the choice of non-standard working arrangements is heavily

conditioned by external constraints, individuals are rarely in a position to make an
explicit trade-off between the advantages and disadvantages of forms of work. Nor
does the legal division between employment and self-employment correspond in a
clear-cut way to the perceptions which individuals in non-standard work hold of
forms of employment. Many of the social and economic risks for which employment
legislation makes provision are perceived as being common to both employment and
self-employment; these include low pay, insecurity of work, health and safety risks
and absence of long-term financial security.

84
Chapter 8. Conclusions

This research paper has provided an account of an empirical study into the
employment status of workers in non-standard employment. The principal aims of
the work were to estimate the numbers of individuals employed in different categories
of employment (in particular 'employees', 'workers' and the independent self-
employed), to identify the sources of uncertainty in the application of legal tests
relating to employee status, and to investigate individuals' perceptions of the
operation in practice of these legal classifications. The study proceeded in two
phases: a 'quantitative wave' which consisted of a survey of just over 4,000
individuals in employment, which was carried out in January-February 1998; and a
'qualitative wave' based on a mix of focus groups and individual semi-structured
interviews, with the sub-sample of respondents, 36 of whom were interviewed face-
to-face and 24 by telephone. These were carried out in the summer of 1998.

8.1 The coverage of employment protection legislation

The concept of the 'worker' in principle allows an extension of the categories of

protected labour beyond 'employees', to include economically dependent workers who
do not have a regular or stable relationship with an employer. Therefore it could have
a potentially significant impact on the position of casual workers, agency workers and
zero-hours contract workers. Although the concept is already found in some areas of
employment law (including the legislation implementing the national minimum wage)
there is comparatively little case law on the meaning of the term 'worker'. However,
on the basis of the existing case law it seems likely that the two criteria which will be
most important are whether the individual contracts to supply his or her personal
services, and whether they are economically dependent to a significant degree upon
the employer.

8.2 Estimating numbers employed as 'employees' and 'workers' in Great

Britain

The quantitative survey estimated that around 86 per cent of all those in employment
in Great Britain consider themselves to be employees, while around 13 per cent
categorise themselves as self-employed. It also found that non-standard working -
defined to include employment under a contract for a fixed-term or task, casual work,
homeworking, agency temping, seasonal work, work as an apprentice or trainee, and
work under zero-hours contract - is much more prevalent among the self-employed
than among employees, and is slightly more prevalent among those working part-time
than among full-time workers.

It was also found that estimates derived from the Labour Force Survey may
significantly understate the numbers employed on non-standard work, in particular
fixed-term employment. This is because the LFS is not picking up a large number of
individuals who regard themselves as being permanently employed even though they
have a fixed-term contract.

An initial set of questions asked respondents about the nature of their work in such a

85
way as to identify individuals who were likely to have ambiguous status. On the basis
of replies to these questions, it was estimated that around 64 per cent of respondents
were clearly employees and 5 per cent were clearly self-employed. This left 30 per
cent who had an employment status that, on first inspection, had elements of
uncertainty and was not completely clear. They were made up of two groups: those
defining themselves as self-employed, but who were not directors or partners in their
own business, and who did not employ others; and those defining themselves as
employees who had some type of non-standard working pattern or classified their jobs
as non-permanent.

The quantitative stage went on to ask the respondents with some ambiguity in their
status a series of questions designed to replicate the type of issues which a court or
tribunal would focus on when making an assessment of employment status. In order
to analyse precisely the status of individuals in this 'unclear' group, the focus was
placed on answers to two sets of questions. The first set of questions concerned the
number of employers for whom they had worked in the previous six months (a test
giving some indication of the degree of economic dependence). The second set aimed
to determine whether individuals were in business on their own account (the
'economic reality' test). This second category of questions asked whether the
individual could hire others to work with or for them, whether they were paid a wage
or salary or through some other form of remuneration, whether they paid their own
income tax and National Insurance contributions or whether they were deducted at
source, and whether they were entitled to sick leave and/or annual paid holiday leave.

On this basis, it was estimated that 80 per cent of all individuals in employment were
clearly either employees or dependent 'workers'; 7 per cent were clearly independently
self-employed; 1.3 per cent were in government training schemes or were unpaid
family workers; and 12 per cent had a status which was still unclear.

In other words, the numbers of 'workers' who would be covered by the extended
definition of dependent labour - the definition used, for example, in section 230(3) of
the Employment Rights Act 1996 and also under the National Minimum Wage Act
1998 - ranges from between 80 per cent to 92 per cent of the working population.
Those in independent self-employment who are outside the scope of this extended
definition range from between 7 per cent to 19 per cent of the working population.
The range of numbers in each category is caused by the uncertainty in knowing
precisely how courts and tribunals will apply the relevant legal tests in practice.

This suggests that by using the 'worker' definition, the number covered by
employment rights might be higher than the 87 per cent who believe that they are
employees according to the LFS. Thus, the use of the worker definition might protect
up to a further 5 per cent of all those in employment (or about 16 per cent of those
whose employment status is unclear because they are employed in non-standard work
of some kind).

86
8.3 Verifying the survey

The qualitative wave was used initially to assess the validity of the findings from the
quantitative wave and to assess the effectiveness of the questions used. It was found
that a number of respondents had not followed the exact definitions given during the
quantitative wave. For example, rather than basing their answers to the question of
whether they were in a permanent employment relationship on the nature of the
contract which they had with their employer, some answered on the basis of their own
personal intentions with regard to their present employment, and some in terms of
whether they thought continuing employment was practical or possible.

Overall, findings from both waves of the research suggest that, given the normal LFS
definitions of atypical or non-standard work and the current routing of the questions,
the numbers employed in some groups may be overestimated - for example seasonal
workers - while others may be greatly underestimated - for example the number
employed on fixed term contracts.

It is possible that better survey items might be developed which would be more
reliable in testing for employment status. However, it seems that a large part of the
unreliability of these questions is caused by the inherently complex nature of the
economic relationships which they are attempting to describe.

8.4 Identifying sources of uncertainty in the application of legal tests of status

Evidence from the qualitative wave of the research suggests that there are a number of
sources of uncertainty in the application of the rules governing employment status. A
principal difficulty lies applying the 'mutuality of obligation' test, which depends on
making inferences of contractual intention from a wide body of evidence. Further
uncertainty derives from the widespread use by employers of contractual
documentation containing 'waiver' or 'relabelling' clauses purporting to determine
status. A further problem here is the use of terms denying that the employer has an
obligation to provide work to the worker ('status-denying clauses'). Contractual
documentation, in particular for agency workers, reflects attempts to draft terms and
conditions in such a way as to minimise the likelihood that a court will find that there
were mutual commitments to accept work and to provide it.

A problem is also associated with the growth of certain forms of work which are
particularly subject to uncertainty as a consequence of the mutuality of obligation test,
in particular agency work and zero hours working. Discrepancies between self-
perceptions of employment status and assessments made on the basis of the tests
applied in the courts were greatest in the case of agency and zero-hours contract
workers.

8.5 Employment status and the experience of non-standard work

The evidence from the qualitative wave suggests that while individuals in non-
standard employment exercise a significant degree of choice over the form of their
employment, several factors are widely seen as conditioning that choice: these include

87
the need to fit in with family arrangements; the costs of retraining following time
spent out of full-time work; the time and complexity of setting up a business;
discrimination by reference to age, gender and disability; and the lack of availability
of regular work.

Many individuals employed in non-standard work see autonomy and flexibility as

desirable aspects of self-employment, on the one hand, and regard some degree of
control and commitment as a part of employment as an employee on the other.
However, the nature of these associations is complex and the divide between
employment and self-employment is not always clear-cut. The independently self-
employed as well as freelance workers can be subject to considerable control from a
client or user, in the sense of being unable for economic reasons to refuse work;
agency workers and zero-hours contract workers rarely exercise any legal right they
might have to turn work down.

Perceptions of insecurity are common to both employees and the self-employed, and
are not consistently associated with one form of work as opposed to another.
Insecurity is seen as increasingly significant by employees in professions or jobs
where employment had previously been secure. Employees in agency work and zero-
hours contracting, in particular, see themselves as vulnerable to summary dismissal.
Some self-employed workers have the possibility of building a reputation among
clients which will counter uncertainty over the demand for their services, but for
others self-employment is associated with chronic insecurity and low incomes.

More generally, evidence on the individuals' experience of non-standard work

suggests that the legal division between employment and self-employment does not
correspond to perceptions of a clear divide between these different forms of work. In
the context of non-standard work, there is considerable ambiguity in the notions of
control, autonomy and mutuality of obligation which have been used as guidelines by
the courts in the assessment of employment status.

88
Bibliography

Atkinson, John 1985. Flexibility, Uncertainty and Manpower Management. Brighton:

Institute of Manpower Studies.
Beatson, Mark 1995. Labour Market Flexibility, Employment Department Research
Series No. 48. London: HMSO.
Brodie, Douglas 1998. ‘The contract for work.’ Scottish Law and Practice Quarterly, 2,
138-148.
Casey, Bernard, Hilary Metcalf and Neil Millward 1997. Employers’ Use of
Flexible Labour. London: Policy Studies Institute.
Cave, Katherine. 1997. Zero Hours Contracts - A report into the incidence and
implications of such contracts. University of Huddersfield.
Collins, Hugh. 1990. Independent Contractors and the Challenge of Vertical
Disintegration to Employment Protection Laws Oxford Journal of Legal Studies 10,
353-380.
Deakin, Simon. 1998. ‘The Evolution of the Contract of Employment 1900-1950: the
Influence of the Welfare State.’ In N. Whiteside and R. Salais (eds.) Governance,
Industry and Labour Markets in Britain and France. The Modernising State in the
Mid-Twentieth Century. London: Routledge.
Deakin, Simon and Gillian S. Morris 1998. Labour Law. 2nd. ed. London:
Butterworths.
Department of Trade and Industry 1998. Fairness at Work. Cm. 3968. London: TSO.
Kahn-Freund, Otto 1951 ‘Servants and Independent Contractors’ Modern Law Review,
14: 504-509.
Gibbs, Anita. 1997. ‘Focus Groups.’ Social Research Update, 19.
Leighton, Patricia. 1983. Contractual Arrangements in Selected Industries. Department
of Employment Research Paper no. 39. London: Department of Employment.
Leighton, Patricia. 1984. ‘Understanding Employment Contracts.’ Industrial Law
Journal, 13: 86-106.
Leighton, Patricia. 1986. ‘Marginal Workers’. In Roy Lewis (ed.) Labour Law in
Britain. Oxford: Blackwell.
Hakim, Catherine. 1990. ‘Core and Periphery in Employers’ Workforce Strategies:
Evidence from the 1987 ELUS Survey.’ Employment, Work and Society, 4, 157-188.
Hunter, L., A. McGregor, J. McInnes and A. Sproull 1993. ‘The “Flexible Firm”:
Strategy and Segmentation’, British Journal of Industrial Relations, 31, 383-407.
Hunter L C & McInnes J, Employer Labour Use Strategies - Case Studies, 1991, ED
Research Paper, No 87.
McGregor Alan and Sproull Alan, Employer Labour Use Strategies, Analysis of a
National Survey, 1991, ED Research Paper, No 83.
Pollert, Anna (ed.) 1988. Farewell to Flexibility? Oxford: Blackwell.
Rubery, Jill, and Frank Wilkinson (eds.) 1993. Employer Strategy and the Labour
Market. Oxford: Oxford University Press.
Wedderburn, Lord. 1986. The Worker and the Law. 3rd. ed. Harmondsworth:
Penguin.
Wood Douglas, Employer Labour Use Strategies - First Report on the 1987 survey,
1989, ED Research Paper ;No 63.

89
Appendix 1 Glossary of Terms

Legal definitions

Employee: an individual who is employed by another under a contract of employment.

Self-employed: an individual who provides labour to another, normally under a

contract for services, and who may or may not be in business on his or her own
account (also referred to as an independent contractor).

Worker: an individual who (1) is employed by another either as an employee or as an

independent contractor, and (2) is economically dependent to some degree on the
business of the other, in other words, is not in business on his or her own account.

Other terms

Agency worker: an individual whose labour is supplied to another on a temporary

basis through an employment agency.

Borderline self-employed: an individual whose legal status (employee or self-

employed) is unclear.

Casual worker: an individual employed by or supplying labour to another under an

irregular or informal working arrangement.

Dependent self-employed: a self-employed individual who is not in business on his or

own account (and is therefore, in legal terms, a ‘worker’).

False self-employed: an individual whose working arrangements, from an objective

point of view, are those of an employee, but who maintains (or whose employer
maintains) that he or she is self-employed.

Fixed-term contract worker: an individual employed by another under a contract for a

fixed term of weeks, months or years.

Independent self-employed: a self-employed individual who is in business on his or

her own account (and is therefore not, in legal terms, a ‘worker’).

Non-standard employment. Employment which in some respect departs from the so-
called 'standard' model of regular employment under a contract of employment for an
indeterminate or indefinite duration. A non-exhaustive list of non-standard
employment includes agency employment, casual employment, fixed-term
employment, seasonal employment, self-employment, task employment, employment
as a trainee or on a government-funded employment programme, and zero-hours
contract employment.

Part-time worker: an individual who is employed to work for less than the normal
working week in the establishment in question.

90
Seasonal worker: an individual employed to work for a particular part of the year,
normally in the context of a specific task or job or for a fixed period of time.

Task worker: an individual employed by another for the duration of a particular task.

Trainee: an individual undergoing training with an employer and/or on a government-

sponsored training programme.

Zero-hours contract worker: an individual who supplies labour or services to another

when called upon to do so, but who has no guarantee of work and no regular working
hours.

91
Appendix 2

Questionnaire used for the quantitative survey

The survey will have established economic activity/working status in an earlier

question with the following options:

1. Full-time
2. Part-time
3. Retired
4. Still at school
5. In full-time higher education
6. Unemployed and not seeking work
7. Unemployed and seeking work

All those who say they are not working (i.e. 3 to 7) are filtered out.

Introduction

We are doing some research on behalf of the Department of Trade and Industry into
the different arrangements under which people work, whether for clients, employers
or agencies. The aim is to look at whether people in different circumstances are
covered by employment protection legislation, such as the right to redundancy
payments. I would like to ask a few confidential questions about your employment
status in the work you are currently doing. If you have more than one job at the
moment, we are interested in your main job.

All in Work (i.e. answered 1 and 2 to activity question above)

1. Which of these describes your work:

1 I am an employee
2 I am self-employed or employee of your own business
3 I am on a government scheme
4 I am an unpaid family worker

All those who answered 3 or 4 are filtered out of survey.

2. Can I just check, which of these best describes your current situation? (Please
remember I am interested in your main job if you have more than one):

1 I am paid a salary or wage by an employer

2 I am paid a salary or wage by an agency
3 I am a sole director of my own limited business
4 I am running or am a partner in a business or professional practice
5 I work for myself
6 I work as a sub-contractor
7 I do freelance work

92
 8 DK
9 Refused

All those who answered 3 or 4 to question 2 are filtered out of survey.

Self-employed ( answered 1 to question 1 and 5 to 7 to question 2) OR (answered 2

to question 1)

3. Do you have people working for you?

1 Yes, paid workers or employees

2 Yes, unpaid family members or friends
3 No, I work on my own
4 Don’t know

All those who answered 1 to question 3 are filtered out of survey.

4. Leaving aside your own personal intentions and circumstance, is your job:

1 A permanent job
2 There is some way that it is NOT permanent

5. Is your work best described by any of the following:

1 seasonal work
2 under contract for a fixed period or a fixed task
3 agency temping
4 casual type of work
5 homeworking (i.e. you home is you main place of work)
6 under a zero-hours contract
7 as an apprentice or trainee
8 none of the above

6. Do you have a written contract of employment or another document, from

your employer or agency, which sets out your terms and conditions in some
way?

1 yes
2 no
3 don’t know

Those answering 1 to question 4 and 8 to question 5 are filtered out of the survey.

7. Which of these best describes the sort of work you do? (If you are not sure,
please choose the category which you think comes closest and I will show you
some more information to help you decide)

code to Standard Occupational Classification

8. More detailed coding if necessary

93
9. Approximately, how many individual organisations/agencies have you worked
for in the last 6 months?
(for agency workers this means the number of agencies)

1 one go to question 9
2 between two and five go to question 10
3 over five go to question 10
4 don’t know go to question 10

10. And about how long have you been working for this organisation or agency
without a break apart from holidays or sickness?

1 Less than one year

2 One to two years
3 Over two years

11. When you are working for a particular client, organisation or agency, are you
allowed to take on work from other clients, organisations or agencies, or not?

1 yes, am allowed
2 no, not allowed
3 don’t know

12. Can you refuse the work offered to you by the clients, organisations or
agencies you work for?

1 yes
2 no
3 don’t know

13. In your current work, are you free to hire other people, who answer to you and
are paid by you, to do the work you personally have taken on, irrespective of
whether you need to?

1 yes
2 no
3 don’t know

14. At work, are you covered by the grievance and disciplinary procedures of your
clients, organisations or agencies you work for?

1 yes
2 no
3 don’t know

94
15. In your current work, do you have to provide the main items of equipment you
need to do your job, not just the small items many employees provide for
themselves such as pens and so on?

1 yes
2 no
3 do not need any equipment
4 don’t know

16. When you are at work, does someone have the right to tell you, at any time,
what to do or when and how to do it, even if this rarely happens in practice?

1 yes
2 no
3 don’t know

17. Which of the following best describes how you are paid?

1 by an monthly salary
2 a weekly wage
3 commission or profit sharing only
4 a fee agreed in advance
5 by the piece (i.e. piece rates)
6 a time rate (by the hour or by the day)
7 other

18. Do you pay your own NI or tax or is this usually deducted by the
organisation(s) you work for i.e. your client, employer, agency etc.?

1 pay own NI and tax

2 NI and Tax is deducted by organisation
3 don’t know

19. Are you entitled to receive sick pay or paid holidays from the organisation(s)
you work for i.e. your client, employer, agency etc.?

1 yes
2 no
3 don’t know

20. What is the main reason for you working in the way that you do?

1 my own preference
2 my employer’s\clients preference
3 it was the only basis on which work was available
4 some other reason

21. The DTI would like to do some more research into the way people are
employed and whether they are covered by employment protection legislation.

95
Could we give your details to another researcher who would like to ask you a
little more about the work that you do?

1 yes
2 no
3 don’t know

96
Appendix 3

Aide mémoire for individual interviews

1. Nature of the job

Please describe the kind of job you do.

How long have you been doing this job?

How many employers have you worked for over the past three years?

Do you currently have more than one job? If so please describe your other jobs.

Which of the jobs you do, if any, is your main job?

What do you most like/dislike about your job/jobs?

2. Flexibility

How free are you to decide:

when you work

the number of hours or days a week/month/year you work

how fast you work

where you work

who you work for

how many people you work for

when to change jobs

Is it important to you to be able to decide these things?

3. The people you work with

Do you work as part of a team?

Or on your own? Please give details.

If you work with others, are they employed in the same way as you? Or differently?

97
[For example: if respondent is temporary, are the others permanent? If casual, are
the others regular? If part-time, are the others full-time? etc.]

Do you hire other people?

Do they work for you or for somebody else in the organisation or company?

4. Pay

How are you paid?

Who is responsible for paying you?

Who, in your view, should be responsible for paying you?

[prompt for agency workers: who pays them: the agency or the client/user? Who
does the worker think should pay him or her?]

Do you receive: holiday pay, sick pay, maternity pay, lay-off pay?

Are you paid a fixed wage per week/month/year?

Are you paid by the hour? or on piece rates? or by some other means...

Are you paid in the form of a fee? or profits?

Do you bill for your services using e.g. an invoice?

Can you choose how you are paid?

Would you rather be paid differently from how you are paid now? If so, why?

Is your tax paid for you by your employer or another person? Or do you organise
your own tax? Are you content with the present arrangement or would you like to
change it?

5. The employment relationship

Do you work mainly for one person/organisation or for more than one? Please
explain.

Are you under an obligation to work exclusively for one person or organisation?

Is it your responsibility to find work or does your employer find it for you?

Can you refuse to work if the employer asks you to?

98
What happen if you do refuse work?

Are you on call? If so please give details. [e.g.: are you contacted by letter or by
phone?]

Are you an agency worker? If so please give details.

Are you a franchisee? If so please give details.

6. Are you your own boss?

Do you see yourself as self-employed?

or as your own boss?

Do you own your own tools or equipment?

If so can you use them for more than one job or for more than client or employer? Or
are there any constraints on how you may use them?

Have you invested your own money in your trade, profession or business?

Do you arrange your own tax and national insurance?

Can you set expenses off against your income or profits?

Do you think that you would be better off if you worked for somebody else as their
employee?

7. Setting up a business

Have you considered setting up in business?

If so, how difficult was it to obtain funds or capital from a bank or other funder?

How did you deal with the tax and financial aspects of setting up your own business?

Were the following major obstacles: complexity of regulations; financial risk; threat
of bankruptcy?

8. Security

How important to you is security or stability in your work, in the sense of having:

a secure and regular income

99
a pension

safe and good working conditions

a career structure?

Whose responsibility is it to look after a regular income, pension, health and safety: is
it you or your employer?

9. Choice

Did you choose to become an employee, or a self-employed person?

[prompt for security, stability, flexibility, autonomy, tax/NI]

If you are an employee now, have you ever been self-employed? [or vice versa?]

What were the reasons for this?

Were your parents/anybody else in your family/close friends self-employed?

If you are an employee now, would you rather be self-employed? [or vice versa?]

Why?

[prompt for role of gender, family responsibilities, ethnic group, age]

10. Documentation

Has your employer given you written documentation about your job?

If so, is it:

Contract of employment

Written statement of particulars of employment

Wage slip

Invoice

other [please specify]

If no, has your employer ever given you any information about your job verbally?

Do you think he should have done?

If yes, what is this document for?

100
Does the document help to tell you what your employment rights and responsibilities
are? Or what your employer’s rights and responsibilities are?

Does it make reference to a grievance and/or disciplinary procedure?

Or to a collective agreement?

other [please specify]

Does it help to tell you whether you are employed as an employee or if you are self-
employed?

Would it help you, do you think, in the event of a legal dispute?

11. Status

Have you ever been concerned that your employment status is not clear?

or that your rights in relation to employment are not clear?

if so why?

in what context: employment rights; social security rights; tax; health and safety; etc.?

have you ever consulted any one of the following about your employment rights:
trade union [prompt for information about whether they are or have been a member of
any trade union]; company or staff association; citizens’ advice bureau; solicitor or
legal advice centre; any other person (please give details)?

101
Appendix 4

Codes for the categorisation of information from the interviewee transcripts

Code Description
age how far did the respondent see their age
as affecting their choice or situation

choice how far being an employee or self-

employed was calculative choice
discrepancies how far there were discrepancies in the
answers given at different interviews
dismissal whether the interviewee had experience
of dismissal; whether they had protection
against dismissal
documents and contracts any mention made of documentation
family how far family matters affected the
weighing up of pros and cons
flexibility how far being an employee or self-
employed was associated with flexibility
of hours, working arrangements, etc.
gender whether the interviewee saw their gender
as affecting their choices or situation
getting work whether the interviewee or the employer
was responsible for finding work
grievance and disciplinary procedures whether the interviewee was covered by
the employer’s procedures
health and safety who took responsibility for health and
safety
holidays did the interviewee have the right to paid
holidays
hours of work could the interviewee set their own hours
of work or starting and finishing times
insurance and cover did the interviewee take out their own
work-related insurance or did the
employer provide it
maternity leave was the interviewee entitled to maternity
leave
method of payment how was the respondent paid
number of employers how many employers/clients/users did the
interviewee have
pension did the employer provide a pension
scheme
pros and cons how far individuals weighed up the pros
and cons of being an employee or self-
employed
race how far the interviewee saw their race as
affecting their choices or situation

102
relationships at work
right to refuse work
risk
security
tax and national insurance
training
uncertainty
working as part of a team
103
the nature of relationships with fellow
workers
whether the interviewee could refuse
work offered by an employer or client
whether the interviewee associated being
an employee or self-employed with a
greater or lesser degree of economic or
physical risk
whether the interviewee associated being
an employee or self-employed with a
greater or lesser degree of security
did the interviewee arrange their own tax
and NI payments or was this done by the
employer
did the interviewee receive training from
the employer
uncertainty in whether the interviewee
saw themselves as an employee or self-
employed; or uncertainty as to their legal
status
whether the interviewee worked
alongside other workers