SERIES

IZA DP No. 10042

PAPER

What Role Did Management Practices Play

in SME Growth Post-Recession?

Alex Bryson
DISCUSSION

John Forth

July 2016

Forschungsinstitut
zur Zukunft der Arbeit
Institute for the Study
of Labor
 What Role Did Management Practices
Play in SME Growth Post-Recession?

Alex Bryson
University College London
and IZA

John Forth
National Institute of Economic and Social Research

Discussion Paper No. 10042

July 2016

IZA

P.O. Box 7240

53072 Bonn
Germany

Phone: +49-228-3894-0
Fax: +49-228-3894-180
E-mail: iza@[Link]

Any opinions expressed here are those of the author(s) and not those of IZA. Research published in
this series may include views on policy, but the institute itself takes no institutional policy positions.
The IZA research network is committed to the IZA Guiding Principles of Research Integrity.

The Institute for the Study of Labor (IZA) in Bonn is a local and virtual international research center
and a place of communication between science, politics and business. IZA is an independent nonprofit
organization supported by Deutsche Post Foundation. The center is associated with the University of
Bonn and offers a stimulating research environment through its international network, workshops and
conferences, data service, project support, research visits and doctoral program. IZA engages in (i)
original and internationally competitive research in all fields of labor economics, (ii) development of
policy concepts, and (iii) dissemination of research results and concepts to the interested public.

IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion.
Citation of such a paper should account for its provisional character. A revised version may be
available directly from the author.
IZA Discussion Paper No. 10042
July 2016

ABSTRACT

What Role Did Management Practices

Play in SME Growth Post-Recession?1

Small and medium-sized enterprises (SMEs) are known to contribute significantly to

aggregate economic growth. However, little is known about the role played by management
practices in SME growth since recession. We contribute to the literature on SME growth by
analysing longitudinal administrative data on firms’ employment and turnover, taken from the
UK’s Business Structure Database (BSD), with data on management practices collected in
face-to-face interviews from the HR Managers and employees who were surveyed as part of
the 2011 British Workplace Employment Relations Survey (WERS). We find off-the-job
training is the only management practice that is robustly and significantly associated with
higher employment growth, increased turnover, and a decline in closure probabilities, over
the period 2011-2014. The findings suggest SME investment in off-the-job training is sub-
optimal in Britain such that firms could benefit economically from increasing the amount of
off-the-job training they offer to their non-managerial employees.

JEL Classification: L25, M12, M50, M53

Keywords: SMEs, small and medium-sized enterprises, employment growth, sales,

workplace closure, HRM, training, recession

Corresponding author:

Alex Bryson
Department of Social Science
University College London
Gower Street
London WC1E 6BT
United Kingdom
E-mail: [Link]@[Link]

1
We gratefully acknowledge funding from Acas who funded this work under their programme on
workplace productivity and growth. We also gratefully acknowledge the advice and support received
from Gill Dix and George Boyce, who managed the research project within Acas. This report presents
research based on data from the 2011 Workplace Employment Relations Survey (WERS). We
acknowledge the Department for Business, Innovation and Skills, Acas, the UK Commission for
Employment and Skills, the ESRC and the National Institute of Economic and Social Research as the
sponsors of the Workplace Employment Relations Study, and the National Centre for Social Research
as the data collectors. The report also presents research based on data from the Business Structure
Database (BSD), produced by the Office for National Statistics (ONS). The data are Crown Copyright
and reproduced with the permission of the controller of HMSO and Queen’s Printer for Scotland. The
WERS and BSD were supplied by the Secure Data Service at the UK Data Archive. The use of the
WERS and BSD data does not imply the endorsement of the WERS sponsors, the ONS or the Secure
Data Service in relation to the interpretation or analysis of the data. This work uses research datasets
which may not exactly reproduce National Statistics aggregates. The syntax used in our analyses are
available on request from the authors.
1. INTRODUCTION

In recent years, researchers and policy makers have paid great attention to the
contribution that small and medium-sized enterprises (SMEs) make to economic
growth. One of the reasons for this is that numerous empirical studies have
shown the importance of SMEs in job creation. Definitions vary as to what
constitutes a SME: some definitions are based on employment size, others on
turnover and still others on the value of assets.2 However, it is most common to
define a SME as any private sector firm with fewer than 250 employees. On this
basis, in 2015, SMEs accounted for 60 per cent of all private sector employment
in the UK (15.6 million jobs) and 47 per cent of all private sector turnover (£1.8
trillion) (Department for Business Innovation and Skills, 2015).

The role of SMEs in the creation of jobs and output has been highlighted in a
variety of domestic and international studies over the past thirty years (e.g.
Birch, 1981; OECD, 2002; Hijzen et al, 2010; Criscuolo et al, 2014). In recent
firm-level evidence for the UK, Hijzen et al (ibid.) showed that SMEs accounted
for between 64 per cent and 77 per cent of all private sector job creation in the
UK over the period 1997-2005 – and between 48 per cent and 67 per cent of
private sector job destruction – with the precise figures dependent on the
methodology used to assign jobs to size classes. Research which has sought to
identify ‘high-growth firms’ has shown that many of the firms which register high
rates of growth in employment or turnover are relatively small. For instance,
Anyadike-Danes et al (2009, 2014) have shown that, over the period 1998-2013,
around six per cent of all UK private sector firms with 10 or more employees met
the Eurostat-OECD’s (2007) definition of ‘high growth’: that is, average
employment growth of at least 20 per cent per annum over three years. Although
this share of high-growth firms does not differ greatly by firm size, the numerical
importance of SMEs in the business population means that they are significant
contributors to overall job creation.

Whilst the recent empirical literature has focused on quantifying the contribution
of different types of firm to job creation or output, there has been less of a focus
on identifying the factors which may help a firm to grow, despite evidence that
SMEs find it more difficult than larger firms to identify and adopt innovative
technologies and working methods due to their weaker internal resources (Roper
and Hart, 2013). In part, this focus is a natural consequence of the data available
to researchers. Studies such as that by Anyadike-Danes et al (2009), Hijzen et al
(2010) and Criscuolo et al (2014) have relied on the analysis of firm-level
administrative data, such as that provided in the Office for National Statistics’
Business Structure Database (BSD) (Welpton, 2009). As explained in Section
Three, the BSD provides a comprehensive longitudinal database of all firms in the
UK; it therefore forms a robust basis for quantifying growth at firm-level and for
identifying some of the basic demographic features of high-growth firms.
However, it lacks any detailed information about the internal organisation of the
firm – its management practices, competitive strategy or use of technology. The
broader management literature suggests these are potential determinants of
growth (see Section 2), and so there remains a great deal to learn about which
SME workplaces are high-growth, and why.

2
The 2006 Companies Act uses all three, defining an SME as an enterprise that
meets at least two of the following criteria: turnover of no more than £25.9m; a
balance sheet total of no more than £12.9m; and no more than 250 employees.

3
Our contribution to this literature is two-fold. First we use data from the ONS
Business Structure Database (BSD) to provide an up-to-date picture of patterns
of growth among SMEs over the ten years from 2004 to 2014. This decade
includes the years of economic growth from 2004-8, the period of recession from
2008-9, and the period of slow recovery from 2009-14.3 Our analysis thus
complements that of Anyadike-Danes and Hart (2014) by updating some of the
earlier evidence on firm growth (e.g. Anyadike-Danes et al, 2009; Hijzen et al,
2010) which focused on the relatively buoyant period of the late 1990s and early
2000s.

Second, we link information on the growth of individual firms, taken from the
BSD, to data on management practices and other firm characteristics that were
collected in the 2011 Workplace Employment Relations Survey (WERS)
(Department for Business, Innovation and Skills et al, 2014). This linked dataset
allows us to examine the determinants of firm growth – and survival – in a
sample of around 500 SMEs over the period 2011-2014. This is the first time that
the BSD and WERS have been linked for such an investigation.

We find off-the-job training is the only management practice that is robustly and
significantly associated with higher employment growth, increased turnover, and
a decline in closure probabilities, over the period 2011-2014. The findings
suggest SME investment in off-the-job training is sub-optimal in Britain such that
firms could benefit economically from increasing the amount of off-the-job
training they offer to their non-managerial employees.

Section Two provides an overview of the existing literature on firm growth,

ranging across economics, management sciences and HRM, and identifies the
hypotheses we test in our empirical investigation. Section Three introduces the
BSD and WERS data in more detail, and outlines the methods used in our
quantitative analysis. Section Four presents our analysis of patterns of growth
among SMEs based on the BSD, followed by the results of our analysis of the
determinants of growth based on the BSD-WERS linked dataset. Section Five
concludes and discusses some of the implications of the analysis.

2. EXISTING THEORY AND EMPIRICAL EVIDENCE ON FIRM GROWTH

Firm growth is defined in a number of ways in the literature, with some analysts
focusing on employment, others on sales and still others on value-added. We take
a similarly broad view, noting only where the evidence differs strongly according
to the measure used. We do focus, however, on measures of scale, rather than
productivity, but it should be noted that there may even be a trade-off between
the two in the short term, if actions to improve process efficiency (for example)
lead to a reduced demand for labour. However, one would expect improvements
in total factor productivity to drive growth in the long-run at the firm level, just as
it does at the macro-economic level, and there is much in common between the
two micro-economic literatures (Syverson, 2011). Whilst empirical studies have
not tended to find a strong relationship between firm-level productivity and
growth – beyond its role in reducing the probability of firm closure (see Coad,
2007: 22-3, 25-6) – recent evidence for the UK does point to a mutually-

3
The output of the UK economy finally returned to its pre-recession level in 2014,
whereas employment did so one year earlier.

4
reinforcing relationship (Du and Temouri, 2015). In this section we focus on
research that is of particular relevance to SMEs.4

The starting point for much of the economics literature on firm growth is Gibrat’s
‘Law’ which proposes that the expected growth rate of a given firm is independent
of its size at the beginning of the period being examined (Gibrat, 1931). Studies
focusing on larger firms do indeed find no significant relationship between firm
growth and size (see Caves, 1998). As newer data sources have become available
which include smaller firms, the weight of evidence has tended to shift such that
most studies now find a negative relationship between firm size and growth: that
is, smaller firms tend to grow faster than larger ones. Variance in the growth
rates decreases with size, consistent with the observation that small firms are
much more likely to fail than larger firms. In short, the likelihood of survival for
small firms tends to be low; however, those that do survive tend to experience
higher growth rates than their larger counterparts.

These observations are consistent with models which emphasise the learning
process in start-up firms. Jovanovic (1982) proposes a model in which firms have
limited information on their productivity at birth, but discover their ability over
time, which naturally implies that some will expand and others fail. Ericson and
Pakes (1995) propose an ‘active’ learning model in which small firms grow by
learning and innovating, implying that the defining feature that is associated with
firm growth may, in fact, be age rather than size. Whilst it is often difficult to
precisely disentangle the two, since size and age are typically highly correlated, a
number of empirical studies provide evidence that it is, in fact, young start-ups
(rather than small firms per se) which tend to have the highest growth rates on
average (Dunne and Hughes, 1994; Haltiwanger et al, 2013). This has led to a
great deal of attention among policy makers on ‘gazelles’: high-growth firms that
are less than five years old (OECD, 2015).

There are a variety of reasons to expect that the rate of firm growth will vary
across industry sectors. First, one can expect firms in relatively young industries
to have higher growth rates than those in more mature sectors, because of the
higher level of new opportunities that are available in the former. This is
particularly evident in high-tech industries, where rapid technological progress
can provide growth opportunities to innovators and early adopters. The empirical
evidence does indicate, however, that high-growth firms are found to some
extent in all industries and sectors (Anyadike-Danes et al, 2009) and so sector of
activity is not, by itself, a strong predictor of growth.

The co-location of firms from related sectors can be important for their growth,
however. One reason is that local competition encourages them to innovate;
another is that smaller firms can use ‘regional clusters’ to increase their visibility
and for reputational advantage which they could not easily secure by their own
means (Gilbert et al, 2008). Operating on international markets is also found to
be positively associated with growth in empirical studies. For example,
Greenaway and Kneller (2008) find that entry into export markets is associated
with an increase in firm size, wages and productivity. One reason is that the firm
accesses a larger market, but another is that, by exposing itself to international

4
Those interested to read further are encouraged to consult the more detailed
overviews provided by Coad (2007) and Audretsch (2012).

5
competition, the firm learns about new technologies and work processes
(‘learning by exporting’).

Competition does not have universally positive implications for firm growth,
however, especially for small firms. Whilst higher levels of competition typically
encourage improvement and innovation, and stimulate higher levels of
productivity, it is also the case that highly competitive markets tend to be
characterised by high levels of exit – a process of ‘creative destruction’ in which
the least productive firms are likely to be forced out by the more productive
(Bravo-Biosca, 2011). As noted above, small firms can be particularly vulnerable
early in their life whilst they are learning how to operate efficiently.

Access to finance is important for growing firms because it provides the resource
base from which investments can be made to facilitate future growth, and
research finds that financial constraints can prevent firms with growth potential
from exploiting new opportunities (Bottazzi et al, 2011). A prime cause of such
constraints is that potential lenders tend to have less information on the
capabilities of smaller or younger firms, which makes it more difficult to judge
credit risks (Audretsch, 2012). Credit is therefore typically rationed, in a way that
tends to constrain many small and young firms from accessing the finance that
they would like in order to grow. In this environment, the social capital and
experience of the business owners can be crucial. Indeed, there is evidence that
new firms benefit from being founded by teams rather than individuals because of
the wider set of networks and relationships that the team brings to the business,
as well as the greater diversity of skills and experience (Department for Business
Enterprise and Regulatory Reform, 2008). However, the type of ownership can
also affect growth in other ways. For instance, ‘managerial’ theories of the firm
(e.g. Marris, 1963) propose that firm size is a more important component of the
managerial utility function than profit, and so managers will have greater
incentives than owners to maximise growth. This argues in favour of the
separation of ownership and control. Indeed, there is some evidence that
management-controlled SMEs have stronger preferences for growth than owner-
controlled firms (Hay and Kamshad, 1994).

External conditions also affect firms’ growth rates. One important set of factors
relates to the regulatory and institutional environment. In their 10-country study,
Bravo-Biosca et al (2014) find that a more developed financial sector, greater
banking competition and better contract enforcement are all associated with a
more dynamic firm-growth distribution, which includes a higher share of growing
firms. Stringent employment protection legislation, on the other hand, is
associated with a lower share of growing firms. But greater stability (less
dynamism) does have the benefit of reducing the number of shrinking firms. The
state of the economy is, of course, also influential. One effect of economic
downturns is to shift the average growth rate downwards (shifting the central
mass in the firm growth distribution to the left). Another effect is to reduce the
spread or variation between firms, consistent with a hypothesis that firms have
more discretion over their growth rates when the economy is in good health, but
face stricter discipline in recessions (Coad, 2007). Nonetheless, the recent
evidence for the UK suggests that, even in the recent economic crisis, many firms
were still able to register ‘high growth’ (Anyadike-Danes and Hart, 2014),
indicating that some firms are able to prosper even in the most difficult of
aggregate conditions.

6
It is implicit in some of the preceding discussion that knowledge, skills and
innovation are important determinants of firm growth. These are the issues we
focus on in our empirical investigation having controlled for many of the other
factors identified as important above.

In Penrose’s (1959) theory of firm growth, knowledge and skills are ‘resources’
which allow firms to create a competitive advantage. This can occur through the
development of new products and processes, and also by increasing the firm’s
ability to absorb knowledge generated elsewhere. In addition, knowledge and
skills create ‘dynamic capabilities’ which allow firms to reconfigure their
operations in fast-changing markets and grow at above-average rates (Teece et
al, 1997). Indeed, young start-ups are argued to be particularly dependent on
skilled labour because of the particular challenges of establishing a new business
under conditions of uncertainty (Cardon, 2003). Cardon (2003) emphasises the
role that contingent labour can play in allowing firms to acquire appropriate skills
and capabilities during periods of expansion. But training and reward practices
are also likely to be important in SME growth.

The literature on how the effectiveness of different HR practices might vary

across a firm’s life-cycle is rather under-developed, and has tended to focus on
models in which firms develop through a very-stylised set of stages (birth, rapid
growth, maturity and decline), whereas the reality is much more varied (Phelps et
al, 2007). There is nonetheless a recognition that the range of HR practices that
are most effective for a firm will need to alter depending on the organisation’s
stage of development. Baird and Meshoulam (1988), for example, argue that
employee training and developing appropriate pay systems are of particular
importance in the high-growth stage, whereas issues to do with communication,
co-ordination and employment relations become more important in later stages,
when the firm is larger and its activities and workforce are more diverse.
Certainly, there is plentiful evidence that the HR practices of small, medium and
large firms differ in some notable respects (Forth et al, 2006). A small empirical
literature supports the contention that training and employee development are
important for SME growth. Barringer’s (2005) study of 50 high-growth and 50
slow-growth firms showed that high-growth firms were more likely to engage in
training and employee development, and were also more likely to use financial
incentives to motivate and reward employees. Barratt and Mayson’s (2007) study
of 600 small firms came to similar conclusions as to the relevance of training and
incentives for growth. The empirical literature on employee involvement indicates
that small firm financial performance is positively associated with the least
bureaucratic, more informal methods of direct communication between
employees and management (Bryson, 1999). Given the theoretical and emergent
empirical evidence to date we test three hypotheses:

Hypothesis One: Off-the-job training will be positively associated with SME

growth.

Hypothesis Two: the provision of financial incentives to employees through

performance pay (PP) will be positively associated with SME growth.

Hypothesis Three: more formal mechanisms for employee engagement will not be
associated with SME growth.

7
Firm growth is linked in the theoretical literature to innovation (Geroski, 2000;
Aghion and Howitt, 1992). One important way in which innovation can contribute
to growth is by expanding the firm’s market through the development of new
products. Another is innovation in processes that result in improved productivity.
However it is not obvious how much time is required for innovative activities
(such as R&D) to be converted into economic output. Some studies find a positive
association between innovation and sales growth (e.g. Geroski and Toker, 1996;
Roper, 1997). Others have found that it can be negatively associated with growth
among poorly-performing firms (Coad and Rao, 2008). Process innovation can
also result in reductions in employment. For instance, the introduction of new
technology may reduce the demand for unskilled labour within the host firm
(although it may increase the demand for skilled labour in the host firm and in
supplying firms). Innovative forms of work organisation can also bring about
efficiencies which reduce the firm’s demand for labour. The impact of process
innovation on employment growth is thus ambiguous a priori. Nevertheless, much
of the firm-level evidence does tend to find a positive relationship (e.g. Doms et
al, 1995; Mason et al, 2009). We test two related hypotheses:

Hypothesis Four: product innovation will be positively associated with SME

growth.

Hypothesis Five: process innovation will be positively associated with SME sales
growth but negatively or non-significantly associated with SME employment
growth.

Managerial capabilities are developed through knowledge acquisition and

experience, such that educational background, prior experience in the industry
and prior experience in a high-growth firm are all consistently found to have a
positive impact on firm growth (Agarwal et al, 2004). One strand of this literature
argues that family-owned firms may particularly lack managerial expertise
because the practice of handing over control to the next generation limits the
talent pool (Bloom and Van Reenen, 2010). This may explain the negative
relationship between family ownership and growth that is evident in the literature
on firm growth (Hart, 2011). Irrespective of managerial experience and
capability, however, it is also important to acknowledge that some firms do not
wish to pursue growth, even if the opportunity presents itself; in other words, the
attitudes of the owner or senior manager are also important. One reason that
owners or managers may not seek to grow is that it typically entails less control
of the firm as the number of activities grows and the hierarchy increases
(Williamson, 1967). Evidence of a lack of desire for growth has been put forward
by Tether (1997) among others. The desire to retain control may be particularly
strong among family owners. We therefore test the following hypothesis:

Hypothesis Six: family ownership is negatively associated with SME growth.

3. DATA and METHODS

Our analyses are based on data from two sources: the ONS Business Structure
Database (BSD) and the 2011 Workplace Employment Relations Survey (WERS).
The BSD is an administrative database compiled by the Office for National
Statistics from the UK’s official business register. The strength of the database
lies in its comprehensive coverage of all but the smallest firms in the UK
economy. Firms may be traced over time, and so the database is able to provide

8
a dynamic view of firm entry, exit and growth. The database comprises annual
‘snapshots’ taken from the Inter-Departmental Business Register – the UK’s
official register of businesses and the sampling frame for surveys of business
activity conducted by the Office for National Statistics. The IDBR contains records
on all firms with turnover above the VAT threshold and all firms with PAYE
schemes (a total of over 2 million firms). Firm-level records in the BSD may be
linked over time via a unique ‘enterprise reference number’ (the ‘Entref’). The
number of variables found in the BSD is small relative to other data sources.
However, the BSD has the virtue of providing extensive coverage of the
population of firms in the UK. It therefore forms a robust basis for quantifying
levels of growth at firm-level and for identifying some of the basic demographic
features of high-growth firms. The BSD has provided the basis for much of the
recent evidence on firm growth in the UK (e.g. Anyadike-Danes et al, 2009;
Hijzen et al, 2010; Criscuolo et al, 2014).5

WERS is a linked employer-employee survey which provides nationally

representative data on workplaces in Britain with five or more employees. The
strength of the survey lies in the richness of data collected on workplace policies
and practices.

The WERS survey was undertaken between February 2011 and June 2012 via
face-to-face interviews with managers responsible for employment relations at
the workplace. The 2011 WERS carried forward some workplaces from the
preceding WERS (carried out in 2004) to construct a panel sample of around
1,000 workplaces. These were supplemented by a “refreshment sample” of
around 1,700 workplaces sampled from the 2010 Inter-Departmental Business
Register (IDBR). Once sampling weights have been applied, WERS is
representative of all workplaces in Britain with five or more employees, with the
exception of those in agriculture and mining which are excluded from the survey.
The full 2011 WERS sample contains data on 1,691 private sector workplaces
(1,193 in the refreshment sample and 498 in the panel).

WERS does not itself provide data on the recent growth trajectory of the
workplace or firm. However, we are able to analyse growth in the years following
the 2011 WERS by matching the refreshment sample workplaces to their
appropriate firm-level records in the BSD via the IDBR Entref. This is possible for
92 per cent of the refreshment sample workplaces – those where the HR manager
who responded to the WERS survey has given permission for data matching to
take place.6 It is then possible to use the longitudinal administrative data
contained within the BSD to measure the growth trajectory of the firm over the
period 2011-2014, whilst also controlling for its rate of growth prior to 2011.

The linked BSD-WERS sample allows us to measure firm growth using the same
metrics that are well established in the existing literature (e.g. Anyadike-Danes et
al, 2009), and then to go on to explore the characteristics associated with that
growth in a way that has not previously been possible in Britain. We are thus able

5
Welpton (2009) provides further details.
6
It is not possible to link WERS panel workplaces to the BSD in large numbers, as
the panel workplaces possess Entrefs from 2003 (the point at which they were
first sampled from the IDBR). Consequently, we focus solely on the refreshment
sample.

9
to generate unique and valuable insights into the factors associated with growth
among SMEs.7

We arrive at a linked WERS-BSD sample of 1,060 private sector workplaces, of

which 515 belong to SMEs (private sector firms with fewer than 250 employees in
the 2011 BSD). Of the 513 firms with non-zero employment in the 2011 BSD,
some 335 have 5-49 employees and 178 have 50-249 employees. The remaining
545 private sector workplaces in WERS belong to large firms with at least 250
employees in the 2011 BSD.

Each of the 513 SME firms in the WERS-BSD matched dataset has only one
workplace observation in WERS. In trying to explain the role of workplace
practices on SME growth, we use this single workplace to characterise the whole
of the firm, even though the firm may have other workplaces which we do not
observe and which might differ from the workplace observed in WERS. In
practice, the degree of such measurement error is likely to be small. If we
compare the employment numbers recorded for the workplace in WERS with
those recorded for the whole firm in the BSD, we find that the WERS workplace
accounts for 95% of the firm’s total employment on average within our SME
sample. For workplaces belonging to firms with 5-49 employees, the average is
100% and for those with 50-249 employees it is 84%. By comparison, the
average for workplaces belonging to firms with at least 250 employees is just
38%.

Five percent of the SME firms that we observe in 2011 (28 out of 513) exit over
the period 2011-2014. Among the remaining 485, we can regress growth in
employment between 2011 and 2014 on a variety of firm characteristics,
controlling for growth over the period 2008-11. The main demographic
characteristics of the firm are taken from the BSD, whilst all other firm
characteristics are taken from WERS. Although WERS includes indicators of many
potential drivers of firm growth, it does not include some drivers that are argued
to be important in the literature discussed in Section Two, notably measures of
access to finance. Nor does it contain measures of growth intentions.
Nevertheless, our range of firm characteristics is wider than is typically available.

We experiment with a variety of model specifications to arrive at a parsimonious

specification which is necessary given our modest sample size. The variables we
concentrate on are apposite for the SME population. So, for example, there is less
focus on unions than in the prior literature on workplace employment growth
(Bryson, 2004). These variables, which are discussed when presenting the
results, are listed below:

 Size, age, industry, single/multi (measured in the BSD)

 Product market characteristics
 Innovation and technology
 Incentives
 ‘High-performance’ work practices
 Training / skills

7
The WERS data have been used previously to study workplace employment
growth (e.g. Bryson and Nurmi 2011; Bryson, 2004) and SME employment
practices (e.g. Forth et al., 2006; Hoque and Bacon, 2006; Bryson, 1999).
However, it has not previously been used to study employment growth in SMEs.

10
  Voice
 Strategy
 Target setting
 Recruitment practices
 HR innovation
 The largest non-managerial occupational group.

In reporting results we focus on variables relevant to the hypotheses set out in

Section Two. Descriptive statistics for those variables which do feature in our final
models are presented in Appendix A, which includes statistics for the full sample
(Table A.1) as well as statistics for workplaces belonging to small and medium-
sized firms presented separately (Tables A.2 and A.3).

Different approaches have been taken to the measurement of firm growth within
the research literature. The two most commonly-used approaches are:

 The base-year method, in which firm size at the end of the period of
observation is divided by firm size at the start of the period

 The average-period method, in which the increase in size over the period
of observation is divided by the average of firm size at the start and end of
the period.

A formal exposition of the two methods is provided in Appendix B. The base-year

method is intuitive, as it provides a fairly straightforward ‘compound’ measure of
growth per annum (G). However, the average-year method provides a more
equitable treatment of small and large firms and is conventional in the more
academic literature (e.g. Hijzen et al, 2010; Haltiwanger et al, 2013). We thus
follow the established academic literature in using the ‘average-period’ method.

As noted above, we measure the growth of the firm in the three years following
the WERS survey (2011-2014). We also measure growth in the three years prior
to the WERS survey (2008-2011) and use this ‘prior trajectory’ as a control
variable in our regression analyses.

The BSD allows us to measure firm growth in terms of both employment and total
sales. Whilst much of the recent literature for the UK tends to focus on
employment growth (e.g. Anyadike-Danes et al, 2009; Hijzen et al, 2010), we
follow Bishop et al (2009) in looking at growth in employment and total sales.
This recognises that, whilst job creation is an important social and economic
outcome, firms are likely to have growth objectives focused around revenue,
since sales ultimately indicate market success.8

For the most part, we analyse growth in employment and growth in turnover
separately. However, we also analyse a binary variable which identifies those
firms that are in the top quartile of the employment growth distribution and the
top quartile of the turnover growth distribution (equating roughly to growth of at
least ten per cent per annum on both measures).

8
We acknowledge that gross value-added would be preferable to total sales.
However the former is not recorded in the BSD, except for the subset of firms
that are surveyed as part of the ONS’ Annual Business Survey.

11
Finally, in recognition that our measures of firm growth are computed only among
those firms that survive to 2014, we also examine the probability of firm closure
(exit). This is measured as the disappearance of the firm’s unique Entref from the
IDBR between 2011 and 2014. In most cases, this will arise through the closure
of the firm and its subsequent removal from the register. However, a firm’s Entref
may also disappear if the firm merges with, or is taken over by, another firm. Exit
is thus potentially measured with some degree of error.

4. RESULTS

4.1 ENTRY, EXIT AND GROWTH AMONG SMEs: ANALYSIS OF THE BSD

In this section we present descriptive analyses regarding employment growth in

the economy between 2004 and 2014, and the role played by firm entry and exit.
We do this through analyses of the Business Structure Database (BSD).

The total number of employees (as measured by the BSD) rose from 25.7m in
2004 to 27.4m in 2009 (Figure 1, left-hand axis). It then fell back to 26.9m in
2011 following the onset of recession before recovering to reach 28.4m in 2014.
Thus 2011 - the year the WERS went into the field - was the lowest point since
the onset of the crisis.

The total number of employees in private sector SMEs with 5-249 employees
followed a similar pattern over most of this period, growing from 8.3m in 2004 to
8.7m in 2009, then falling to 8.4m in 2011 before rising to 9.3m in 2013.

Figure 1: Employees by Firm Size, Business Structure Database

Source: BSD

12
The share of employees in SMEs was stable at around 31.5% from 2004-2011 but
it has since grown to around 33%, primarily because of the fall in public sector
employment since 2011, but also due to a small increase in the share of private
sector employment in firms with 5-249 employees (rising from 38.8% in 2008 to
40.0% in 2014).9

Figure 2 shows the contribution of firm birth and firm death to employment
among SMEs pre- and post-recession. All public sector organisations are excluded
from this and subsequent analyses. Prior to recession around 7% of employees
were in firms that had exited by the following year, but this fell steadily after the
crisis. The contribution of new firms entering the population was a little below 5%
prior to recession and fell further to under 4% with the crisis. However, it had
recovered to almost 5% by 2014. The remaining 90% or so of SME employment
in each year was in firms appearing in consecutive years, that is, firms that had
neither exited nor entered.

Figure 2: Percentage of Employment in SMEs Accounted for by Firm Entry

and Exit, BSD

Source: BSD

For those firms that are observed in two consecutive years we are able to
estimate their rate of employment growth. As noted above, there are two ways
that growth measures are typically constructed in the literature. The first method
divides employment change between t and t+1 by the employment level in time
t. This is what we term the base-year measure. The second measure is similar
but this time the denominator is the average employment in years t and t+1.
Growth trends for SMEs using both measures are presented in Figure 3. We also

9
The literature is not consistent on how employment in small and large firms
responds to the business cycle. On the one hand, Moscarini and Postel-Vinay
(2012) show that employment in large firms is more closely linked to changes in
the unemployment rate than those of small ones. On the other hand, Hardwick
and Adams (2002) find countercyclical variation in the relationship between firm
size and growth, suggesting that larger firms grow faster during recessions and
recoveries.

13
present their winsorized equivalents whereby outlier values beyond the 1st and
99th percentile values are recoded to these limits. Trends in growth are similar
whichever measure is used: growth rates are fairly stable at 2-4% per annum
prior to recession, dip in 2008/9 with the onset of recession, only to recover
quickly in 2010/11, peaking at rates of 3-5% per annum in 2012/13. In the rest
of the report we focus on the second measure because it is the measure most
commonly used in the literature (eg. Hijzen et al. 2010).10

Figure 3: Employment Growth among SMEs, BSD

Source: BSD

Table 1 Employment Levels and Growth Within the SME Population, BSD
Panel A: Firm-weighted

Employment in 2014 (row %) Employment

Employment
growth rate
in 2011:
Exit 1-4 5-49 50-249 250+ pa, 2011-14

All 17.1 11.8 63.8 7.0 0.3 -1.19%

5-49 emps 17.5 12.6 68.0 1.9 0.0 -1.17%

50-249 emps 12.0 2.2 11.4 70.6 3.8 -1.41%

10
Using this measure tends to result in lower levels of employment growth due to
the fact that the measure typically has a larger denominator than the base-
measure indicator and has an upper bound of <1.9999. We use the winsorized
version to avoid results being skewed by outlying cases experiencing exceptional
growth or contraction.

14
Panel B: Employment-weighted

Employment in 2014 (row %)

Employment
Employment
growth rate
in 2011:
Exit 1-4 5-49 50-249 250+ pa, 2011-14

All
14.2 5.6 46.7 30.8 2.7 3.26%

5-49 emps
16.0 7.9 71.1 5.0 0.1 3.69%

50-249 emps
11.4 2.0 8.1 71.6 6.9 2.59%
Source: BSD

Panels A and B of Table 1 identify what happened to SME firms between 2011 and
2014, the period of employment growth we focus on in the remainder of this
study. The tables are based on 452,595 firm-level observations. The left-hand
column in the tables identifies the status of the firm in 2011 while subsequent
rows show what had happened to them by 2014.

We can see from the firm-weighted data in Panel A that 17% of SMEs in 2011 had
exited by 2014: some of these will have died, either due to bankruptcy or
liquidation, while others may have exited for other reasons (as noted in Section
Three, we are unable to distinguish between the reasons for exit).

Exit was more common among the smallest SMEs: 17.5% of SMEs with 5-49
employees had exited by 2014, compared with 12% of those with 50-249
employees in 2011. A little over two-thirds of those in the 5-49 employee
category in 2011 remained in the same size band in 2014, as was the case for
those in the 50-249 size band. Very few of the SME sample had grown beyond
the employment threshold of 250 employees used to designate firms as SMEs
(0.3%). The average growth rate of firms was -1.19% between 2011 and 2014,
and was particularly low (-1.41% per annum) in the larger SMEs. The growth rate
is negative in part because the figures exclude new firm entrants, most of whom
would have grown.

By employment weighting the same data in Table 1 Panel B we can see what
percentage of employees were in SMEs that exited or grew. Fourteen per cent of
employees were employed by a SME that exited between 2011 and 2014: this
figure is lower than the firm-weighted 17.1% indicating that it was the smaller
SMEs employing fewer employees that were more likely to exit. When employee-
weighted, the average growth rate of SMEs was positive at around 3.3% per
annum, rising to 3.7% for employees in SMEs with 5-49 employees. The fact that
growth rates are positive in Panel B but negative in Panel A indicates that
employment growth was more positive among larger workplaces since it is these
workplaces that account for a greater proportion of employees.

15
4.2 EXIT AND GROWTH AMONG SMES: ANALYSIS OF BSD-WERS
LINKED DATA

Having presented exit and employment growth rates for SMEs in the BSD
population we turn to analyses of the subset of private sector firms which appear
in both WERS and the BSD. Analyses are weighted by the WERS workplace
sampling weights so that they reflect figures for the population from which they
were sampled, rather than merely within-sample estimates. The WERS sample is
atypical of all firms since they were sampled in 2010 and had to survive until
2011/12 in order to be interviewed in the survey. They are therefore a little more
stable than SMEs as a whole, as indicated by the lower exit rate of around 5%
(Table 2, column 4) when compared to the much higher exit rates in Table 1
Panel A.

Table 2: Employment Growth and Exit Rates 2011-2014 for the Matched
WERS-BSD Sample

Employment growth rate pa,

2011-14
2011 employment: Exit rate (%)
Mean SD

5-249 emps 0.80% 15.7 5.3%

5-49 emps 1.05% 16.0 5.4%

50-249 emps -0.56% 13.6 5.2%

Source: WERS-BSD

The employment growth rates presented in Table 2 for the WERS-BSD sample are
defined in the same way as those presented in the final column of Table 1 Panel
A. The growth rates are more positive in the WERS-BSD sample than they are for
all firms in the BSD. Mean growth is 0.80% per annum over the three years
2011-2014, which is modest, but it is notable that growth is stronger among
small firms than it is among medium-sized firms. It is also notable that there is
substantial variance in growth rates, as indicated by the large standard
deviations. Growth rates are normally distributed overall, however.

A fuller understanding of growth patterns among SMEs relies on multivariate

analyses as discussed earlier. In the literature on SME growth, employment
growth is one important metric, but another is growth in sales turnover. We
therefore estimate multivariate models as described earlier for employment
growth, on the one hand, and sales growth on the other. The mean and
distribution of turnover growth is very similar to those for employment growth.
Descriptive information on all the control variables used in the models is provided
in Appendix A.

Table 3 presents very simple models which use covariates found in the BSD to
identify some basic characteristics that are correlated with employment growth
and turnover in the period 2011-2014. The models are run on firms that did not
exit the BSD over the period. The models contain identical covariates, namely the

16
year in which the firm was born according to the BSD, a dummy variable
identifying whether the firm was in the Manufacturing sector, employment (or
turnover) levels in 2011 (expressed in logs) and growth in the period immediately
following the financial crisis (2008-2011).

Table 3: Growth in employment or turnover - BSD specification

[1] [2]
Employment Turnover
growth growth
2011-14 2011-14

Employment growth
2008-11 0.096*
[1.94]
Ln(employment in 2011) -0.019***
[-3.02]
Turnover growth 2008-11 -0.034
[-0.77]
Ln(turnover in 2011) -0.010**
[-1.99]
Manufacturing 0.049* 0.052
[1.94] [1.57]
Year of birth:
2006-10 0.006 0.008
[0.20] [0.23]
2001-05 0.003 0.051*
[0.13] [1.96]
1996-2000 0.002 -0.007
[0.10] [-0.34]
Ref. Pre 1996 0 0
[.] [.]

Number of observations 485 485

Adjusted R-squared 0.046 0.034
Source: WERS-BSD
Notes:
1. T-statistics in parentheses
2. Key to significance: * p<0.10, ** p<0.05, *** p<0.001
3. Constant not shown to prevent disclosure

Column 1 presents the employment growth equation. Coefficients can be

interpreted as independent correlations between the variable and employment
growth measured in terms of percentage changes. Employment growth in the
period immediately following the recession in 2008-11 is positively correlated with
employment growth in 2011-14. However, bigger SMEs in 2011 grew at a slower
rate in 2011-14. Manufacturers grew 5% more quickly than firms in the Service
sector. Firm age was not significant having conditioned on firm size.

17
A comparison of columns 1 and 2 reveals that larger firms grew less than smaller
ones, irrespective of whether scale is measured via employment or turnover.
Otherwise, the correlates of the two measures of growth were somewhat
different. Whilst employment growth in 2008-11 was positively associated with
growth from 2011-14, turnover growth in 2008-11 was actually negatively, albeit
not significantly, correlated with turnover growth in 2011-14. Firm age also
mattered somewhat in the turnover regression (but not the employment
regression), with firms set up in the early 2000s increasing turnover at a faster
rate than older firms set up prior to 1996. Finally, the Manufacturing coefficient
was statistically significant only in the employment regression; however it was
positive and of a similar size in the turnover regression.

Table 4 presents similar employment and turnover growth models to those

presented in Table 3 but this time we introduce additional control variables taken
from the WERS HR Manager.11 We experimented with a number of variables that
do not appear in the table which proved statistically non-significant. These
included target setting, strategic planning, representative forms of worker voice
(such as trade unions) and a plethora of "high performance" work practices such
as quality circles, just-in-time production and job rotation.

Table 4: Growth in employment or turnover - WERS specification

[1] [2]
Employment growth Turnover growth
2011-14 2011-14

Employment growth 2008-11 0.077*

[1.65]
Ln(employment in 2011) -0.027***
[-3.83]
Turnover growth 2008-11 -0.049
[-1.14]
Ln(turnover in 2011) -0.024***
[-4.38]
Manufacturing 0.028 0.03
[0.99] [0.92]
Year of birth:
2006-10 0.013 0.001
[0.44] [0.04]
2001-05 0.005 0.036
[0.23] [1.63]
1996-2000 0.000 -0.003
[-0.02] [-0.13]
Ref. Pre 1996 0 0
[.] [.]
Family owned -0.022 -0.053**
[-1.17] [-2.47]
Growing market 0.026 0.029

11
7 observations are lost due to missing data.

18
 [1.30] [1.33]
Degree of product innovation:
Rarely innovating 0.027 0.024
[1.21] [1.09]
Middle of the road 0 0
[.] [.]
Often innovating 0.039* 0.021
[1.78] [0.92]
Merit pay scheme 0.036* 0.029
[1.66] [1.28]
Payment-by-results scheme -0.001 0.043**
[-0.03] [2.02]
Arrangements for team
working:
No team working 0 0
[.] [.]
Non-autonomous teams -0.015 -0.021
[-0.73] [-0.93]
Autonomous teams -0.047** -0.059**
[-2.24] [-2.54]
Percentage of largest
occupation receiving training
in past year:
80%+ trained 0.060*** 0.050*
[2.59] [1.90]
40-79% trained 0.081*** 0.058**
[2.79] [2.05]
1-39% trained 0.034 0.058**
[1.36] [2.25]
None trained 0 0
[.] [.]
Attitude survey in last two
years 0.037** 0.015
[2.05] [0.70]
New technology introduced in
past two years 0.041** 0.033*
[2.53] [1.74]

Number of observations 478 478

Adjusted R-squared 0.139 0.128
Source: WERS-BSD
Notes:
1. T-statistics in parentheses
2. Key to significance: * p<0.10, ** p<0.05, *** p<0.001
3. Constant not shown to prevent disclosure
4. Identity of largest occupational group included among regressors but not
shown for brevity

19
Beginning with the employment growth analysis in column 1, it is apparent that
the introduction of the additional controls substantially increases the model's
ability to capture variance in employment growth across firms relative to the BSD
model in Table 3. This shows the value of the matched WERS-BSD dataset when
compared with administrative datasets which merely contain information on a
small range of demographic characteristics.12 Furthermore, the additional WERS
covariates do little to change the signs of the BSD variables: employment growth
in 2008-11 remains positively correlated with growth in 2011-14, size in 2011 is
negatively correlated with growth, and firm age is not significant. However,
Manufacturing firms no longer have significantly faster growth than their Service
sector counterparts.

Firms who introduced new technology in the two years prior to the WERS survey
experienced 4% higher growth per annum than observationally-equivalent firms
who had not introduced new technology. Similarly, those firms who said they
"often led the way" in response to the question "to what extent would you say
this workplace leads the way in terms of developing new products, services or
techniques" grew 4% per annum faster than those who scored 3 on the 5-point
scale (identified in the table as ‘middle of the road’). There is therefore some
support for Hypothesis Four which linked product innovation to higher
employment growth, but the link between process innovation and higher
employment growth runs counter to Hypothesis Five.

As predicted in Hypothesis One, investment in human capital brings rewards in

terms of employment growth: firms where at least 40% of core non-managerial
employees had undertaken off-the-job training in the year prior to the WERS
survey interview grew at a significantly faster rate than those that were less
training intensive.

In keeping with Hypothesis Two there is a positive association between

performance pay for employees and higher employment growth: those firms
paying at least some of their employees according to subjectively assessed merit
in 2011 grew 3.5% per annum faster than observationally equivalent firms
without such a scheme.

The concept of autonomous team-working is perhaps at the very heart of what

many term "High Involvement Management", and is usually deemed the key
instrument by which employees are able to maximise their contributions to output
with minimal oversight by supervisors. In much of the literature it is identified as
contributing positively to labour productivity and firm performance (e.g. Baron
and Kreps, 1999). Yet, according to the results in column 1, firms using
autonomous team-working grew at a rate of nearly 5% less per annum than 'like'
firms that did not use teams, while there was no association with non-
autonomous teams. These findings are consistent with Hypothesis Three that
suggested more formal mechanisms for employee involvement may not deliver
growth for SMEs. But it is also possible that, as other studies have shown (
Cordery et al., 2010; Langfred, 2004) too much worker autonomy can have
negative consequences for a firm. For instance, managers sometimes devolve too
much responsibility down to teams of workers with insufficient support, training
and incentives to carry out their more onerous tasks. The result can be a decline

12
Coad et al (2014: 99) are among those who bemoan the limited explanatory
power of models based on such sources.

20
in labour productivity which, ultimately, will affect the rate of growth in the firm.
Another possibility is that that firms with well-functioning autonomous team-
working are more productive than firms without, such that teams produce
additional output without firms having to employ additional workers to generate
that output. In other words, the firm may be more efficient and, as a consequent,
less reliant on employment growth. This is one reason why some firms are able to
engage in what economists sometimes refer to as "jobless growth".

Less formal, costly methods of engaging employees do appear to foster growth,

however. Firms that had conducted employee attitudes surveys in the two years
prior to the WERS survey interview experienced growth rates of almost 4% more
per annum than 'like' firms who had not surveyed their employees. If, as is often
argued in the HRM literature, employee attitude surveys are an indicator for
employer interest in engaging and motivating staff, then this finding supports the
proposition that firms can benefit from an engaged and satisfied workforce, a
contention supported by other empirical research (Bryson et al, 2014).

Column 2 of Table 4 presents a very similar model for sales turnover growth in
the period 2011-14. As with the employment growth analysis, the addition of the
WERS covariates increases the explanatory power of the model quite markedly
but has little direct impact on the sign nor the significance of the BSD variables.
Size, as indicated by log turnover in 2011, remains negatively associated with
sales growth whereas firm age, broad sector and lagged sales growth play little
role. In a number of ways the sales growth model is strikingly similar to that for
employment growth. Process innovation (the introduction of new technology) is
associated with higher sales growth, as predicted in Hypothesis Five, as is
investing in human capital via training, confirming Hypothesis One. The sales
growth returns to training are for undertaking any training, with little evidence of
additional returns for training a higher percentage of core employees.

Incentives are positively linked to sales growth, supporting Hypothesis Two,

though on this occasion it is the use of payments-by-results, rather than merit
pay, that proves statistically significant. Once again, autonomous team-working
has adverse consequences for the firm: the fact that they adversely affect sales
growth as well as employment growth suggests the mechanism could be reduced
productivity in the presence of autonomous team working. The finding is not
consistent with autonomous teams lowering employment through improved
productivity.

There are two results that differ across our two measures of growth. The first
relates to the use of attitude surveys: whereas these are linked to higher
employment growth they are positive but not statistically significantly related to
sales growth. The second is the role of family-ownership: family-owned firms
experience significantly lower sales growth than 'like' firms that are not family-
owned, whereas family ownership is negative but not significant for employment
growth. This negative effect of family ownership, predicted in Hypothesis Six,
may be related to the poorer management which characterises many family-
owned and family-run firms, an issue emphasised by Bloom and Van Reenen
(2010).

In some parts of the literature, analysts have followed the OECD in adopting a
simple (0,1) measure of firm growth whereby firms score '1' if they achieve
growth of at least 20% per annum in either employment or turnover (e.g.

21
Anyadike-Danes et al, 2009). The problem with such a measure is that firms
identified as 'high growth' may expand along one dimension at the expense of the
other resulting, for example, in jobless growth. We think there is value in using
both employment and turnover to identify high performing firms but we depart
from the OECD definition by identifying high performers as those firms that are in
the top quartile of employment growth and turnover growth rates (roughly 10
percent growth per annum in either case). Thirteen per cent of our 485 surviving
workplaces (49 out of 485, unweighted) have employment growth rates in the top
quartile and turnover growth rates in the top quartile, whereas around 16% of
firms in our sample meet the OECD criteria.

We estimate the influences on this 'high growth' measure in Table 5. Column 1

presents linear estimates of the (0,1) outcome whereas column 2 presents results
from the equivalent probit estimator. The signs and significance of coefficients are
very similar regardless of the estimation technique used so we comment below on
column 1 since the coefficients are easier to interpret as they capture percentage
changes in the probability of being high growth.

Table 5: High growth in employment and turnover

[1] [2]
OLS Probit
Employment growth 2008-11 0.288** 1.320**
[2.51] [2.46]
Ln(employment in 2011) -0.034*** -0.208**
[-2.59] [-2.21]
Manufacturing 0.06 0.115
[0.93] [0.31]
Year of birth:
2006-10 0.133** 0.796**
[2.12] [2.08]
2001-05 0.096* 0.675**
[1.80] [2.24]
1996-2000 -0.011 0.083
[-0.28] [0.25]
Ref. Pre 1996 0 0
[.] [.]
Family owned -0.080* -0.427**
[-1.90] [-2.12]
Growing market 0.129*** 0.640***
[2.70] [2.77]
Degree of product innovation:
Rarely innovating 0.005 -0.233
[0.10] [-0.71]
Middle of the road 0 0
[.] [.]
Often innovating 0.013 0.026
[0.29] [0.10]
Merit pay scheme 0.073 0.339

22
 [1.31] [1.32]
Payment-by-results scheme 0.02 0.112
[0.44] [0.44]
Arrangements for team working:
No team working 0 0
[.] [.]
Non-autonomous teams 0.053 0.296
[0.94] [1.01]
Autonomous teams -0.012 -0.024
[-0.27] [-0.09]
Percentage of largest occupation
receiving training in past year:
80%+ trained 0.121** 0.859**
[2.06] [2.28]
40-79% trained 0.160** 1.016**
[2.47] [2.37]
1-39% trained 0.085 0.692*
[1.62] [1.85]
None trained 0 0
[.] [.]
Attitude survey in last 2 years -0.052 -0.338
[-1.30] [-1.32]
New tech introduced -0.026 -0.164
[-0.67] [-0.70]

Number of observations 478 478

Adjusted R-squared 0.198
Pseudo R-squared 0.321
Source: WERS-BSD
Notes:
1. T-statistics in parentheses
2. Key to significance: * p<0.10, ** p<0.05, *** p<0.001
3. Constant not shown to prevent disclosure
4. Identity of largest occupational group included among regressors but not
shown for brevity

Firms that grew strongly in the immediate aftermath of the crisis were "high
growth" firms in 2011-14. However, growth was slower among those larger SMEs,
as indicated by their log employment in 2011. In keeping with the literature
(Caves, 1998), and in a departure from the findings presented above, younger
firms were more likely to be high growth than those born some time ago. The
model also confirms the adverse impact of family ownership on growth as
predicted in Hypothesis Six. It also points to the importance of market conditions
facing firms: those operating in growing markets experience significantly higher
growth rates than those firms whose HR Manager characterised the market for
their main product or service as "declining", "mature" or "turbulent". (Although
there was a positive coefficient on growing markets in the models reported above
it was not statistically significant).

23
Perhaps what is most striking about the "high growth" model is that most of the
choices management makes regarding HR practices, investment and the
organization of production, prove statistically non-significant in this model, even
though many had a statistically significant effect on incremental growth as
indicated in Table 4. In some instances, the reason for this is the larger standard
errors in the "high growth" model such that, even if point estimates on particular
coefficients are large, as in the case of merit pay, the estimates are insufficiently
precise to produce a statistically significant result. In other instances, such as the
introduction of new technology, the sign on the coefficient has switched to
negative, although again the effect is not statistically significant. The clear
exception to the non-significance of HR practices is the prevalence of training for
core non-managerial employees: this is strongly positively associated with being
a 'high growth' firm, with the returns being strongest where at least 40% of core
employees have received off-the-job training in the 12 months prior to the WERS
survey. The association between training and growth is thus a particularly robust
one supporting Hypothesis One.

Each of the models described above run on a selective sample in the sense that
they do not take account of the 28 firms that exited the sample after 2011. It is
conceivable that the results presented for surviving firms might have looked
different on the whole population if one took account of which firms exited and
whether HR practices also influenced the probability of exit. We therefore finally
estimated OLS and probit models to establish which firm traits predicted
subsequent exit.

It is apparent that, conditional on size in 2011, younger firms are more likely to
exit (Table 6), a finding that is typical in the literature. Family-ownership is also
positively correlated with exit, as is the HR Manager's perception that
organizational performance is below the average for the industry. However, very
few HR and organizational choice variables are correlated with the probability of
exit. There are two exceptions. First, undertaking employee attitude surveys are
positively correlated with exit (although this result is only statistically significant
in the probit regression, which may suffer from over-fitting given the small
number of exits and large number of discrete covariates). Second, off-the-job
training is linked to a lower likelihood of firm exit. This latter point is consistent
with earlier research by Collier et al (2005, 2011) using the WERS panel and
suggests the training effects on growth noted earlier are an underestimate of the
overall benefits of training to SMEs.

24
Table 6: Firm exit [1] [2]
OLS Probit
Ln(employment in 2011) 0.01 0.205**
[1.02] [2.12]
Manufacturing -0.084** -1.259**
[-2.02] [-2.37]
Year of birth:
2006-10 0.128** 1.540***
[2.43] [3.64]
2001-05 0.036 0.398
[1.57] [1.15]
1996-2000 0.023 0.166
[0.87] [0.44]
Ref. Pre 1996 0 0
[.] [.]
Family owned 0.033 0.490**
[1.46] [2.01]
Growing market -0.018 -0.059
[-0.77] [-0.25]
Degree of product innovation:
Rarely innovating -0.011 0.077
[-0.27] [0.26]
Middle of the road 0 0
[.] [.]
Often innovating -0.047 -0.375
[-1.42] [-1.29]
Merit pay scheme -0.005 -0.074
[-0.27] [-0.23]
Payment-by-results scheme -0.019 -0.176
[-0.75] [-0.63]
Arrangements for team working:
No team working 0 0
[.] [.]
Non-autonomous teams 0.002 -0.204
[0.07] [-0.55]
Autonomous teams 0.057 0.336
[1.44] [1.08]
Percentage of largest occupation
receiving training in past year:
80%+ trained -0.026 -0.228
[-0.71] [-0.71]
40-79% trained -0.083* -0.941*
[-1.90] [-1.66]
1-39% trained -0.080* -1.025**
[-1.93] [-2.44]
None trained 0 0

25
 [.] [.]
Attitude survey in last 2 years 0.038 0.908***
[0.87] [2.73]
New tech introduced -0.024 -0.372
[-1.10] [-1.47]
Perceived financial performance
relative to industry average:
A lot above average 0.05 0.188
[0.85] [0.36]
Above average 0.018 0.058
[0.67] [0.18]
About average 0 0
[.] [.]
Below average 0.083 0.937**
[1.01] [2.06]
Not known/no comparison possible 0.141* 1.444***
[1.66] [3.60]

Number of observations 506 506

Adjusted R-squared 0.115
Pseudo R-squared 0.385
Source: WERS-BSD
Notes:
1. T-statistics in parentheses
2. Key to significance: * p<0.10, ** p<0.05, *** p<0.001
3. Constant not shown to prevent disclosure
4. Identity of largest occupational group included among regressors but not
shown for brevity

5. CONCLUSIONS

This paper examines SME growth in the aftermath of perhaps the most severe
recession Britain has experienced in over a century. Much of the literature on SME
growth is based on evidence collected in more benign times, so it is particularly
instructive to revisit this literature and reappraise what factors appear to be
associated with growth. In doing so we combine administrative data taken from
the Business Structure Database and evidence from HR Managers and employees
surveyed as part of the 2011 Workplace Employment Relations Survey. The
richness of these data mean we can go beyond a focus on 'structural' features of
firms such as industrial sector, age, and initial size, by examining the role of
choices made by firms in terms of their HR management practices, investment
decisions, and the organization of production.

We find that 2011 – the year that WERS went into the field – saw the lowest
aggregate level of employment since the onset of the crisis in 2008. In that sense
the period 2011-2014, which is the basis for our later WERS-BSD analysis, was a
time of recovery. However, it was also a time in which the share of employees in
private sector SMEs was growing (from 38.8% in 2008 to 40.0% in 2014),
primarily because of the decline in public sector employment that took place
between 2011 and 2014.

26
Focusing only on private sector SMEs, we find that the extent of churning in the
population did not change a great deal through recession. In each pair of years
from 2004-5 to 2013-14, the percentage of employment in private sector SMEs
that continued in existence from year-to-year was always around 90%. Rates of
employment growth among these firms did dip in recession, however, only to
recover quickly in 2010/11. Again, this reinforces the view that our WERS-BSD
analysis covers a period in which the economy – and its constituent firms – were
returning to a more normal trajectory after the depths of the crisis

We then presented multivariate analyses of our sample of 513 SME firms from the
WERS-BSD matched dataset. The analyses covered four outcomes: employment
growth; turnover growth; being a ‘high-growth’ firm; and the probability of firm
exit. The reference period for all analyses was 2011-2014. We found that the
additional variables provided in the WERS data significantly extended the
explanatory power of our models when compared with specifications relying upon
the limited number of characteristics provided in the BSD – thereby showing the
value of the matched WERS-BSD dataset when compared with purely
administrative data sources. However, the set of additional workplace
characteristics and HR practices that were found to be statistically significant was
relatively small.

Looking across the wide range of workplace characteristics and HR practices that
were examined in our analyses, we found that investments in new technology and
the use of performance-related pay were positively and significantly correlated
with growth in employment and growth in turnover. Nonetheless, the single most
striking result was that the proportion of core non-managerial employees
engaged in off-the-job training was robustly associated with higher growth and
lower exit probabilities. Firms which engaged larger proportions of their core
employees in off-the-job training had higher levels of employment growth and
higher levels of turnover growth than other, similar firms. More extensive off-the-
job training was also positively associated with ‘high-growth’, using the OECD
definition which is a common reference point in the broader literature (that is, the
attainment of at least 20% growth in either employment or turnover per annum).
In addition, firms that engaged their core employees in off-the-job training were
also less likely to exit the population over the three-year period covered by our
analyses. The implication is that the amount of off-the-job training undertaken in
Britain's private sector SMEs is sub-optimal and that an increase in training is
likely to benefit firms and workers alike.

Other HR practices, such as communication and team-working, were not

correlated with growth rates or were, in some cases, linked to poorer growth
performance. We speculate that these practices may be more important once the
firm matures into a larger entity and, indeed, some supporting arguments have
been made in this regard in the broader literature. However, there is a need for
further research to investigate the extent to which certain HR practices impact
differentially on the firm’s ability to grow at different stages.

Among the SMEs that were our focus, however, it is apparent that what firms
chose to do in terms of investing in human capital, reorganising production and
incentivising workers made a difference to how these SMEs emerged from
recession. Ultimately, of course, firms are interested in profit maximisation as
opposed to employment growth. Yet the fact that firms' choices also influenced

27
sales growth is a reminder that these practices may also be beneficial for profit-
maximising firms and that more might be done to encourage firms to invest
where there appears to be sub-optimal investment.

Perhaps the chief limitation to our study is that we are unable to make causal
inferences about the link between HR practices and growth. We are able to
account for the growth trajectories of firms prior to 2011, so that the comparisons
we make partial out the potentially confounding effect of dynamic growth effects
that might otherwise bias our estimates of the links between practices and
growth. But we are unable to account for the fact that firms are making choices
that are partly based on their performance trajectory and what they think will
work for them, such that the practices we observe are endogenous. Future
research may be able to deploy methods which go beyond simple partial
correlations to establish the causal effects of management practices on SME
growth.

28
1. BIBLIOGRAPHY

Agarwal R, Echambadi F and Sarkar M (2004) “Knowledge transfer through

inheritance: spin-out generation, development and performance”, Academy of
Management Journal, 47: 501-522.

Aghion, P. and P. Howitt (1992), ‘A model of growth through creative destruction’

Econometrica, 60, 2: 323-351.

Anyadike-Danes, M., Bonner, K., Hart, M. and Mason, C. (2009) Measuring

Business Growth: High-growth firms and their contribution to employment in
the UK, London: NESTA.

Anyadike Danes M and Hart M (2014) “Moving on from the vital 6%”, Enterprise
Research Centre (ERC) Insight Bulletin, February.

Audretsch D (2012) “Determinants of high-growth entrepreneurship”, Report for

the OECD/DBA International Workshop on High-Growth Firms, Copenhagen,
28th March.

Baird L and Meshoulam I (1988) “Managing Two Fits of Strategic Human Resource
Management”, Academy of Management Review, 13, 1: 116-128.

Baron J and Kreps D (1999) Strategic Human Resources: Frameworks for General
Managers. New York: John Wiley & Sons.

Barringer B (2005) “A quantitative content analysis of the characteristics of rapid-

growth firms and their founders”, Journal of Business Venturing, 20: 663-687.

Birch D (1981) “Who Creates Jobs”, The Public Interest, 65: 3-14.

Bravo-Biosca A (2011) “A look at business growth and contraction in Europe”,

NESTA Working Paper 11/02, London: NESTA.

Bryson, A. (1999) ‘The impact of employee involvement on small firms’ financial

performance’, National Institute Economic Review, 3/99, No. 169, July, pp. 78-
95

Bryson, A. (2004) "Unions and Employment Growth in British Workplaces During

the 1990s: A Panel Analysis", Scottish Journal of Political Economy, 51, 4: 477-
506

Bloom N and Van Reenen J (2010) ‘Why do management practices differ across
firms and countries?’, Journal of Economic Perspectives, 24, 1: 203-24.

Bottazzi G, Secchi A and Tamagni F (2011), “Financial Constraints and Firm

Dynamics”, mimeo, LEM-Scuola Superiore Sant’Anna, Pisa, Italy. [cited in
Bravo-Bisoca et al, 2014]

Bravo-Biosca A, Criscuolo C and Menon C (2014) “What drives the dynamics of

business growth?”, NESTA Working Paper 14/03, London: NESTA.

29
Bryson, A. (2004) ‘Unions and Employment Growth in British Workplaces During
the 1990s: A Panel Analysis’, Scottish Journal of Political Economy, Vol. 51, No.
4, 477-506.

Bryson, A., Forth, J. and Stokes, L. (2015) Does Worker Wellbeing Affect
Workplace Performance?, London: Department for Business Innovation and
Skills, BIS/14/1120.

Bryson, A. and Nurmi, S. (2011) ‘Private sector employment growth, 1998-2004:

a panel analysis of British workplaces’, Cambridge Journal of Economics, 35, 1:
85-104 .

Cardon M (2003) “Contingent labor, an enabler of entrepreneurial growth”,

Human Resource Management, 42, 4: 357-373.

Cardon, M. and Stevens, C. (2004) "Managing human resources in small

organizations: What do we know?" Human Resource Management Review, 14:
295-323

Caves, R. E. (1998) 'Industrial Organization and New Findings on the Turnover

and Mobility of Firms", Journal of Economic Literature, 36, 4: 1947-1982

Coad A (2007) “Firm growth: a survey”, Papers on Economics and Evolution No.
703, Max Planck Institute of Economics.

Coad A and Rao R (2008) “Innovation and firm growth in high-tech sectors: a
quantile regression approach”, Research Policy, 37: 633-648.

Coad A, Daunfeldt S, Holzl W, Johanssen D and Nightingale P (2014) “High

growth firms: introduction to the special issue”, Industrial and Corporate
Change, 23, 1: 91-112.

Collier, W., Green, F. and Peirson, J. (2005) "Training and Establishment

Survival", Scottish Journal of Political Economy, 52, 5: 710-735

Collier W, Green F, Kim Y-B and Peirson J (2011) “Education, Training and
Economic Performance: Evidence from Establishment Survival Data”, Journal of
Labor Research, 32, 4: 336-361.

Cordery, J. L., Morrison, D., Wright, B. M., and Wall, T. D. (2010) 'The impact of
autonomy and task uncertainty on team performance: A longitudinal field
study', Journal of Organizational Behavior, Vol. 33, Issue 2-3: 240-258

Criscuolo C, Gal P and Menon C (2014) “The dynamics of employment growth:

new evidence from 18 countries”, CEP Discussion Paper No. 1274, London:
Centre for Economic Performance.

Department for Business Enterprise and Regulatory Reform (2008) High Growth
Firms in the United Kingdom: Lessons from an Analysis of Comparative UK
Performance, BERR Economics Paper No. 3, London: Department for Business
Enterprise and Regulatory Reform.

30
Department for Business, Innovation and Skills, Advisory, Conciliation and
Arbitration Service and National Institute of Economic and Social Research
(2014) Workplace Employee Relations Survey, 2011 [computer file]. 4th
Edition. Colchester, Essex: UK Data Archive [distributor], April 2014. SN:
7226, [Link]

Department for Business Innovation and Skills (2015) Business Population

Estimates for the UK and Regions 2015, London: Department for Business
Innovation and Skills.

Doms M, Dunne T and Roberts M (1995) “The role of technology use in the
survival and growth of manufacturing plants”, International Journal of
Industrial Organization, 13: 523-542.

Du J and Temouri Y (2015) “High-growth firms and productivity: evidence from

the United Kingdom”, Small Business Economics, 44: 123-143.

Dunne P and Hughes A (1994) “Age, Size, Growth and Survival: UK companies in
the 1980s”, Journal of Industrial Economics, 42, 2: 115-140.

Ericson R and Pakes A (1995) ‘Markov-perfect industry dynamics: a framework

for empirical work’, Review of Economic Studies, 62: 53-82.

Eurostat-OECD (2007) Business Demography Manual, Paris: Eurostat-OECD.

Forth, J., Bewley, H. and Bryson, A. (2006) Small and Medium-Sized Enterprises:
Findings from the 2004 Workplace Employment Relations Survey, London:
Department of Trade and Industry

Geroski P (2000) “The growth of firms in theory and practice”, pp. 168-186 in N
Foss and V Mahnke (eds.) Competence, Governance and Entrepreneurship,
Oxford: Oxford University Press.

Geroski P and Toker S (1996) “The turnover of market leaders in UK

manufacturing industry 1979-86”, International Journal of Industrial
Organization 14: 141-158.

Gibrat R (1931) Les Inégalités Économiques, Paris: Librairie du Receuil Sirey

Gilbert B, McDougall P and Audretsch D (2008) “Clusters, knowledge spillovers

and new venture performance: an empirical examination”, Journal of Business
Venturing, 23, 4: 405-422.

Greenaway D and Kneller R (2008) “Firm heterogeneity, exporting and foreign

direct investment”, The Economic Journal, 117: F134-161.

Haltiwanger J, Jarmin R and Miranda J (2013) “Who creates jobs? Small versus
large versus young”, Review of Economics and Statistics, 95, 2: 347-61.

Hardwick P and Adams M (2002) “Firm size and growth in the United Kingdom life
insurance industry”, Journal of Risk and Insurance, 69, 4: 577-593.

Hart M (2011) BIS Mid-Sized Business Research: Note on Regressions, London:

Department for Business Innovation and Skills.

31
Hay M and Kamshad K (1994) “Small firm growth: intentions, implementation
and impediments”, Business Strategy Review, 5, 3: 49-68.

Hijzen A, Upward R and Wright P (2010) “Job creation, job destruction and the
role of small firms: firm-level evidence for the UK”, Oxford Bulletin of
Economics and Statistics, 72, 5: 621-647.

Hoque K and Bacon N (2006) “The antecedents of training activity in British small
and medium-sized enterprises”, Work, Employment and Society, 20, 3: 531-
552.

Jovanovic B (1982) ‘Selection and the evolution of industry’, Econometrica, 50:

649-670.

Langfred, C. W. (2004) 'Too Much of A Good Thing? Negative Effects of High Trust
and Individual Autonomy in Self-Managing Teams', Academy of Management
Journal, 47, 3: 385-399

Marris R (1963) ‘A model of the ‘managerial’ enterprise’, Quarterly Journal of

Economics, 77, 2: 185-209.

Mason G, Bishop K and Robinson C (2009) Business Growth and Innovation: The
Wider Impact of Rapidly-Growing Firms in UK City-Regions, London: NESTA.

Moscarini G and Postel-Vinay F (2012) “The contribution of large and small

employers to job creation in times of high and low unemployment”, American
Economic Review, 102, 6: 2509-39.

NESTA (2009) The Vital 6 Per Cent: How High-Growth Innovative Businesses
Generate Prosperity and Jobs, London: NESTA.

OECD (2002) High Growth SMEs and Employment, Paris: OECD.

OECD (2015) Entrepreneurship at a Glance 2015, Figure 4.14, Paris: OECD.

Penrose E (1959) The Theory of the Growth of the Firm, Oxford: Basil Blackwell.

Phelps R, Adams R and Bessant J (2007) “Life cycles of growing organisations: a

review with implications for knowledge and learning”, International Journal of
Management Reviews, 9, 1: 1-30.

Porter M (1998) On Competition, Boston: Harvard Business School.

Roper S (1997) “Product innovation and small business growth: a comparison of

the strategies of German, UK and Irish Companies”, Small Business Economics
9: 523-537.

Roper S and Hart M (2013) “Supporting sustained growth among SMEs – policy
models and guidelines”, ERC White Paper No. 7, Warwick: Enterprise Research
Centre.

Syverson C (2011) "What Determines Productivity?", Journal of Economic

Literature, 49, 2: 326-365

32
Teece D, Pisano G and Shuen A (1997) “Dynamic capabilities and strategic
management”, Strategic Management Journal, 18, 7: p509–33.

Tether B (1997) “Growth diversity amongst innovative and technology-based new

and small firms: an interpretation’, New Technology, Work and Employment,
12, 2: 91-107.

Van Wanrooy, B., Bewley, H., Bryson, A., Forth, J., Stokes, L. and Wood, S.
(2013) Employment Relations in the Shadow of Recession: Findings from the
2011 Workplace Employment Relations Study, Palgrave MacMillan.

Welpton R (2009) “Business Structure Database: The Inter-Departmental

Business Register (IDBR) for Research”, Economic and Labour Market Review,
3, 6: 71-5.

Williamson O (1967) “Hierarchical control and optimum firm size’, Journal of

Political Economy, 75, 2: 123-138.

33
APPENDIX A: ADDITIONAL TABLES

Table A.1 Descriptive Statistics for Control Variables, WERS-BSD, full

sample

Unweighted Weighted
Mean Std. dev. Mean Std. dev. N
Employment growth 2008-11 0.064 0.194 0.091 0.201 485
Turnover growth 2008-11 0.021 0.253 0.038 0.256 485
Ln(employment in 2011) 3.485 1.513 2.748 1.285 485
Ln(turnover in 2011) 7.527 1.766 6.972 1.605 485
Manufacturing 0.122 0.327 0.147 0.355 485
2006-10 0.120 0.325 0.145 0.352 485
2001-05 0.249 0.433 0.348 0.477 485
1996-2000 0.196 0.397 0.174 0.379 485
Ref. Pre 1996 0.435 0.496 0.334 0.472 485
Family owned 0.513 0.500 0.571 0.496 485
Growing market 0.307 0.462 0.293 0.455 485
Rarely innovating 0.237 0.426 0.276 0.447 485
Middle of the road 0.280 0.450 0.284 0.452 485
Often innovating 0.482 0.500 0.440 0.497 485
Merit pay scheme 0.221 0.415 0.214 0.411 485
Payment-by-results scheme 0.303 0.460 0.293 0.456 485
No team working 0.262 0.440 0.363 0.481 485
Non-autonomous teams 0.340 0.474 0.265 0.442 485
Autonomous teams 0.398 0.490 0.372 0.484 485
80%+ trained 0.409 0.492 0.387 0.488 484
40-79% trained 0.128 0.335 0.099 0.299 484
1-39% trained 0.281 0.450 0.264 0.441 484
None trained 0.182 0.386 0.250 0.433 484
Attitude survey in last 2 years 0.304 0.461 0.205 0.404 483
New tech introduced 0.570 0.496 0.534 0.499 481
Professional 0.120 0.325 0.115 0.319 485
Assoc. Professional & Technical 0.122 0.327 0.129 0.336 485
Admin and Secretarial 0.118 0.322 0.117 0.321 485
Skilled Trades 0.151 0.358 0.188 0.391 485
Caring, Leisure and Service 0.118 0.322 0.093 0.291 485
Sales and Customer Service 0.153 0.360 0.169 0.375 485
Process and Plant Operatives 0.074 0.262 0.076 0.266 485
Elementary Occupations 0.146 0.354 0.113 0.317 485

34
Table A.2 Descriptive Statistics for Control Variables, WERS-BSD, sample
with 5-49 employees

Unweighted Weighted
Mean Std. dev. Mean Std. dev. N
Employment growth 2008-11 0.078 0.193 0.095 0.200 315
Turnover growth 2008-11 0.024 0.245 0.038 0.252 315
Ln(employment in 2011) 2.790 1.097 2.403 0.881 315
Ln(turnover in 2011) 6.890 1.429 6.659 1.350 315
Manufacturing 0.121 0.326 0.157 0.364 315
2006-10 0.121 0.326 0.148 0.356 315
2001-05 0.324 0.469 0.378 0.486 315
1996-2000 0.197 0.398 0.165 0.372 315
Ref. Pre 1996 0.359 0.480 0.309 0.463 315
Family owned 0.549 0.498 0.588 0.493 315
Growing market 0.295 0.457 0.296 0.457 315
Rarely innovating 0.260 0.440 0.284 0.452 315
Middle of the road 0.276 0.448 0.282 0.451 315
Often innovating 0.463 0.499 0.434 0.496 315
Merit pay scheme 0.222 0.416 0.216 0.412 315
Payment-by-results scheme 0.248 0.432 0.250 0.434 315
No team working 0.330 0.471 0.385 0.487 315
Non-autonomous teams 0.276 0.448 0.241 0.428 315
Autonomous teams 0.394 0.489 0.375 0.485 315
80%+ trained 0.369 0.483 0.378 0.486 314
40-79% trained 0.102 0.303 0.099 0.299 314
1-39% trained 0.287 0.453 0.251 0.434 314
None trained 0.242 0.429 0.273 0.446 314
Attitude survey in last 2 years 0.204 0.403 0.156 0.363 314
New tech introduced 0.497 0.501 0.510 0.501 312
Professional 0.092 0.290 0.102 0.304 315
Assoc. Professional & Technical 0.133 0.340 0.137 0.345 315
Admin and Secretarial 0.133 0.340 0.121 0.326 315
Skilled Trades 0.168 0.375 0.206 0.405 315
Caring, Leisure and Service 0.124 0.330 0.098 0.298 315
Sales and Customer Service 0.130 0.337 0.140 0.348 315
Process and Plant Operatives 0.063 0.244 0.074 0.263 315
Elementary Occupations 0.156 0.363 0.121 0.326 315

35
Table A.3 Descriptive Statistics for Control Variables, WERS-BSD, sample
with 50-249 employees

Unweighted Weighted
Mean Std. dev. Mean Std. dev. N
Employment growth 2008-11 0.038 0.195 0.071 0.205 170
Turnover growth 2008-11 0.016 0.269 0.040 0.275 170
Ln(employment in 2011) 4.773 1.324 4.637 1.490 170
Ln(turnover in 2011) 8.707 1.727 8.685 1.799 170
Manufacturing 0.124 0.330 0.093 0.291 170
2006-10 0.118 0.323 0.127 0.334 170
2001-05 0.112 0.316 0.180 0.385 170
1996-2000 0.194 0.397 0.222 0.417 170
Ref. Pre 1996 0.576 0.496 0.471 0.501 170
Family owned 0.447 0.499 0.478 0.501 170
Growing market 0.329 0.471 0.272 0.446 170
Rarely innovating 0.194 0.397 0.231 0.423 170
Middle of the road 0.288 0.454 0.297 0.458 170
Often innovating 0.518 0.501 0.472 0.501 170
Merit pay scheme 0.218 0.414 0.205 0.405 170
Payment-by-results scheme 0.406 0.493 0.528 0.501 170
No team working 0.135 0.343 0.247 0.433 170
Non-autonomous teams 0.459 0.500 0.398 0.491 170
Autonomous teams 0.406 0.493 0.355 0.480 170
80%+ trained 0.482 0.501 0.439 0.498 170
40-79% trained 0.176 0.382 0.098 0.299 170
1-39% trained 0.271 0.446 0.338 0.474 170
None trained 0.071 0.257 0.125 0.332 170
Attitude survey in last 2 years 0.491 0.501 0.475 0.501 169
New tech introduced 0.704 0.458 0.667 0.473 169
Professional 0.171 0.377 0.184 0.389 170
Assoc. Professional & Technical 0.100 0.301 0.083 0.277 170
Admin and Secretarial 0.088 0.284 0.096 0.296 170
Skilled Trades 0.118 0.323 0.085 0.279 170
Caring, Leisure and Service 0.106 0.309 0.068 0.253 170
Sales and Customer Service 0.194 0.397 0.325 0.470 170
Process and Plant Operatives 0.094 0.293 0.086 0.281 170
Elementary Occupations 0.129 0.337 0.072 0.259 170

36
APPENDIX B: MEASURES OF FIRM GROWTH

The two most commonly-used approaches to the measurement of firm growth (G)
are:

 The base-year method, in which firm size (N) at the end of the period of
observation (time ti) is divided by firm size at the start of the period (time
t1 )

 The average-period method, in which the increase in size over the period
of observation is divided by the average of firm size at the start and end of
the period

The advantages of each method are outlined in the text.

37