Professional Documents
Culture Documents
The modern corporation has far reaching influence on our lives in an increasingly
globalised economy. This series will provide an invaluable forum for the publica-
tion of high quality works of scholarship covering the areas of:
Alex Coad
Evolutionary Economics Group, Max Planck Institute of
Economics, Germany
Edward Elgar
Cheltenham, UK • Northampton, MA, USA
© Alex Coad 2009
Published by
Edward Elgar Publishing Limited
The Lypiatts
15 Lansdown Road
Cheltenham
Glos GL50 2JA
UK
1 Introduction 1
2 Firm size distributions 14
3 Growth rate distributions 25
4 Gibrat’s law 39
5 Profits, productivity and firm growth 49
6 Innovation and firm growth 76
7 Other determinants of firm growth 84
8 Theoretical perspectives 100
9 Growth strategies 111
10 Growth of small and large firms 129
11 Conclusion 143
Notes 152
Bibliography 161
Index 191
v
Figures and tables
FIGURES
TABLES
vi
Acknowledgements
A large number of people helped in the preparation of this book, at various
stages of its development. I am very grateful for their help. In particular, I
am very much indebted to (listed alphabetically): David Audretsch, Erkko
Autio, Christian Cordes, Giovanni Dosi, Elizabeth Garnsey, Vojislav
Maksimovic, Rie Nemoto, Bernard Paulré, Christos Pitelis, Rekha Rao,
Angelo Secchi, Agusti Segarra, Erik Stam, Fede Tamagni, Mercedes
Teruel, and Ulrich Witt. I am also very grateful to Matt Pitman, commis-
sioning editor at Edward Elgar, for his remarkable patience and support.
Any remaining shortcomings in this book are my own responsibility,
however.
vii
1. Introduction
Research into firm growth has been accumulating at a terrific pace, and is
being published in a growing range of outlets, such as journals relating to
the disciplines of economics, management, sociology, entrepreneurship, as
well as disciplines as diverse as statistical physics and psychology.
The past few decades have witnessed much progress in empirical research
into firm growth, in particular, for a number of reasons. First, datasets
documenting economic phenomena are growing in terms of their level of
detail, sample size and availability. The rise of information technology has
played a major role in this trend. Many countries have statistical offices that
undertake censuses of business firms and establishments, creating longitudi-
nal databases that track individual firms over time, and make these records
available to researchers (under restrictions of confidentiality, of course).
Firms are required to provide information on themselves and their opera-
tions at a level of detail that is quite remarkable. For example, many firms
are required to file financial reports that describe their operations not just
at the aggregated level, but disaggregated by production plant or by line of
business. Even ‘soft’ variables, such as entrepreneurial growth aspirations,
are becoming commonplace in quantitative statistical analyses – these
variables can be measured using subjective responses of individuals to large-
scale questionnaires. A second development favouring empirical research
into economics is that econometric techniques have kept pace with the
availability of increasingly informative datasets. Modern econometric work
is able to deal with such complicated issues of endogeneity, unobserved het-
erogeneity within individuals, and sample selection bias. The progress that
has been made in this domain has been reflected by the number of Nobel
memorial prizes awarded to econometricians in recent years. Breakthroughs
in econometric theory have granted more legitimacy to empirical findings,
and have also allowed researchers to investigate more elaborate hypotheses.
Old results have been turned on their heads when new, improved statistical
methods have been applied. A third major development is that continual
increases in computational power have been able to match developments
in databases and econometric techniques. Bootstrapping methods, for
example, are particularly computationally intensive and their use has only
become feasible thanks to developments in the performance of computers.
1
2 The growth of firms
Early theoretical work into the size and growth of firms was placed in a
comparative statics framework, and by reason of its static nature did not
really deal with the dynamic phenomenon of growth. Firms were supposed
to be at their ‘optimal size’; and if they weren’t there already, they were
assumed to grow instantaneously to reach it. In this way, firm growth
received a cursory treatment as an appendage to the optimal size theory.
Firms were considered to grow only inasmuch as this enabled them to
Introduction 5
reach their optimal size. However, dissatisfaction with this theory of firm
behaviour has grown in recent decades. Notions of an ‘optimal size’ have
been rejected in almost any interpretation of the phrase that one might
subscribe to (see section 8.1).
Growing dissatisfaction with the conventional static approach of eco-
nomic theory has led to the ascendancy of new themes in theoretical work.
Emphasis has been placed on the prevalence of uncertainty, change and
bounded rationality in the context of a turbulent and restless economy. It
has been suggested that firm growth has replaced firm size as the central
variable in industrial economics (Marris, 1999).
Uncertainty, and also bounded rationality on the part of firms, are
important foundations for an analysis of firm growth, because growth
inevitably involves expansion into new areas. Uncertainty is magnified,
of course, in dynamic markets that are continually being transformed by
technological innovation and competition. In addition, we note that the
firm itself is changing, through growth, in ways it cannot foresee.
Path-dependency is also an important theme. Firms can be seen as
bundles of specific capabilities, or as the repositories of organizational
routines (Nelson and Winter, 1982; Dosi et al., 2000). Firms tend to be
specialized in what they do and they cannot easily change from one day to
the next. What a firm did in the past defines what it can do in future, and
so a firm’s growth opportunities are very much constrained by its current
production activities. Competitive advantage rests to a large extent on
accumulated firm-specific resources as well as production capabilities that
have been carefully developed over time, and the gradual nature of this
process places a limit on the ability of firms to adapt rapidly to a changing
environment.
In addition, we feel it is necessary to recognize the great heterogeneity
that exists between firms, whether we consider productivity levels, prof-
itability, or a large number of other key dimensions. As Griliches and
Mairesse (1995) explain (p. 23):
We also thought that one could reduce aggregation biases by reducing the heter-
ogeneity as one goes down from such general mixtures as ‘total manufacturing’
to something more coherent, such as ‘petroleum refining’ or the ‘manufacture
of cement’. But something like Mandelbrot’s fractals phenomenon seems to be
at work here also: the observed variability-heterogeneity does not really decline
as we cut our data finer and finer. There is a sense in which different bakeries
are just as much different from each other, as the steel industry is from the
machinery industry.
(See also Dosi and Grazzi (2006) for further evidence of pervasive
heterogeneity of firms, even at finely disaggregated levels.) We should
6 The growth of firms
that will take this into account. This is especially true given that Chapter
6 focuses specifically on firm-level innovation. Fourth, the low rationality
assumptions that form the basis of the evolutionary framework strike us as
simply being far more judicious than the ‘Olympian’ rationality frequently
assumed in the neoclassical paradigm. Uncertainty is unquestionably one
of the most fundamental features of the modern economy, and it seems
to us to be also one of the defining characteristics of firm growth. Firm
growth is essentially a venture into unfamiliar territory. Indeed, in Chapter
5 (section 5.1) we criticize the mainstream literature that takes the assump-
tion of infinitely rational profit-maximizing firms as a foundation for its
empirical work into firm-level investment patterns. Instead, we delve into
evolutionary theory to obtain a guiding theory. In section 8.4 we investi-
gate the evolutionary principle of ‘growth of the fitter’ and it is astonish-
ing to observe that even this general principle, when taken literally, does
not appear to hold. It seems that even evolutionary economics, which has
genuinely mild rationality assumptions, may be overstating the capacity of
the forces of economic development.
A final motivation for basing our analysis in the evolutionary perspec-
tive is that it appears to be more or less in accordance with the empirical
facts. One of the few regularities that has emerged from research into the
growth of firms is that Gibrat’s ‘law of proportionate effect’ appears to
provide a better description of industrial development than any other
alternative theory. Although Gibrat’s law is frequently criticized as having
no theoretical content (due to the emphasis on purely stochastic shocks),
on the contrary it is our (controversial) view that Gibrat’s law does have
a theoretical basis, and that it is not too far-fetched to consider that this
basis is of an ‘evolutionary’ flavour. We have three reasons for making this
association. First, Gibrat’s law emphasizes heterogeneity between firms
that stems from the variance of the growth shocks. Second, the stochastic
nature of Gibrat’s law can be seen to emphasize the inherent uncertainty
that permeates modern capitalism. Third, Gibrat’s law accommodates the
evolutionary principle of path dependency by the fact that a firm’s current
size is viewed as the mere amalgamation of all previous growth shocks.
The evolutionary setting of this book does not go far in predicting how
much a particular firm will grow, however – instead it provides a theoreti-
cal setting in which empirical work can be grounded. The prevalence of
uncertainty and also pervasive heterogeneity of firms in the context of a
turbulent and restless economy suggests to us that the state of the economy
cannot be worked out from the armchair. Instead, our understanding of
the growth of firms must progress through solid empirical analysis. This
necessarily involves ‘getting one’s hands dirty’ and working with data.
We feel obliged to reiterate an exhortation that is dated but nonetheless
8 The growth of firms
is not just a ‘statistic’ but actually has a much deeper significance. This is
indeed one of the dangers of empirical work – one can get so accustomed
to dealing with numbers that one may forget what the numbers actu-
ally represent. (Indeed, this has led some individuals to be unnecessarily
apprehensive about empirical work in general.)
But without further ado, we now discuss how firm size and growth can
be measured.
Measuring Size
The number of different indicators of firm size is rather vast, and is limited
only by the imagination of the researcher. Employment and total sales,
however, are the most commonly used indicators (Delmar, 1997). This is in
part because data on employment and sales is among the easiest to obtain.
In the majority of cases, it will make little difference which firm size indica-
tor is taken, as they tend to give similar results. (This should not be taken
for granted, however.)
Among the candidate indicators of firm size and growth, a major advan-
tage of employment is that, unlike financial quantities, it does not need to
be deflated. This is useful for multi-sector analyses, where sector-specific
deflators need not be sought out. It is also useful for the cross-country
analyses, or investigations involving multinational corporations, because
exchange rate complications are avoided. A drawback of employment,
however, is that indivisibilities are substantial for small firms that have
only a few employees. Sales is also frequently taken as a measure of firm
size. One disadvantage of sales, though, is that it need not necessarily
correspond to the actual value-added of a company. Consider the case
of a firm that buys an almost finished product (for example computers)
that is modified or repackaged in some minor way before being sold on to
others. Such a firm will have a high sales figure, because of the high cost
of the final product, even though its contribution to the overall economy
in terms of value-added will be low. If this firm then goes on to acquire its
upstream components suppliers, its total sales will not change but its share
of value-added will of course have increased. Value-added may thus be a
better indicator of firm size than sales, because it takes into account the cost
of materials used in the production process. In practice, however, data on
value-added is not always available, and the peculiar scenario described
above does not occur very often.
There are also many other measures of firm size in use. Another popular
measure is total assets – although this indicator encounters difficulties if
the firms in the sample have different capital intensities. Some authors (for
example Little, 1962; Baumol et al., 1970) speak of firm growth as referring
10 The growth of firms
Measuring Growth
St St21
5 2
St21 St21
St
5 21
St21
where Sit is the size of firm i at time t. Taking logs, and remembering that
log(1) 5 0, we obtain:
The observation that the firm size distribution is positively skewed proved
to be a useful point of entry for research into the structure of industries. (See
Figures 2.1 and 2.2 for some examples of aggregate firm size distributions.)
Gibrat (1931) considered the size of French firms in terms of employees
and concluded that the lognormal distribution was a valid heuristic. Hart
and Prais (1956) presented further evidence on the size distribution, using
data on quoted UK firms, and also concluded in favour of a lognormal
model. The lognormal distribution, however, can be viewed as just one of
several candidate skew distributions. Although Simon and Bonini (1958)
maintained that the ‘lognormal generally fits quite well’ (p. 611), they
preferred to consider the lognormal distribution as a special case in the
wider family of ‘Yule’ distributions. The advantage of the Yule family of
distributions was that the phenomenon of arrival of new firms could be
incorporated into the model. Steindl (1965) applied Austrian data to his
14
Firm size distributions 15
1
1998
2000
2002
0.1
0.01
Pr
0.001
le-04
–4 –2 0 2 4 6
S
Source: Bottazzi et al. (2008a).
Figure 2.1 Kernel estimates of the density of firm size (total sales) in
1998, 2000 and 2002, for French manufacturing firms with
more than 20 employees
10–1
10–4
Frequency
10–7
10–10
10–13
1 10 102 103 104 105 106
Firm size (employees)
analysis of the firm size distribution, and preferred the Pareto distribution
to the lognormal on account of its superior performance in describing the
upper tail of the distribution. Similarly, Ijiri and Simon (1964, 1971, 1974)
apply the Pareto distribution to analyse the size distribution of large US
firms.
Efforts have been made to discriminate between the various candidate
skew distributions. One problem with the Pareto distribution is that the
empirical density has many more middle-sized firms and fewer very large
firms than would be theoretically predicted (Vining, 1976). This leads to
a shape for the firm size distribution that is concave to the origin when
plotted on log–log axes. Other research on the lognormal distribution
has shown that the upper tail of the empirical size distribution of firms is
too thin relative to the lognormal (Stanley et al., 1995), which implies, a
fortiori, that the tails are too thin relative to the Pareto (evidence on this
can be found in Rossi-Hansberg and Wright (2007); see also Ramsden
and Kiss-Haypal (2000) for international comparisons of the upper tail
of the firm size distribution). Marsili (2005) reports that the Pareto
is a good fit for the upper tail of the aggregate firm size distribution,
whereas the lognormal seems to be a better fit for the smaller firms.
Quandt (1966) compares the performance of the lognormal and three
versions of the Pareto distribution, using data disaggregated according
to industry. He reports the superiority of the lognormal over the three
types of Pareto distribution, although each of the distributions produces
a best-fit for at least one sample. Furthermore, it may be that some
industries (for example the footwear industry) are not fitted well by any
distribution.
More generally, Quandt’s results on disaggregated data lead us to
suspect that the regularities of the firm size distribution observed at the
aggregate level do not hold with sectoral disaggregation. Silberman (1967)
also finds significant departures from lognormality in his analysis of 90
four-digit SIC sectors in his analysis of US firms. Similarly, Marsili (2005)
observes significant differences across sectors in the firm size distribution.
It has been suggested that, while the firm size distribution has a smooth
regular shape at the aggregate level, this may merely be due to a statis-
tical aggregation effect rather than a phenomenon bearing any deeper
economic meaning (Dosi et al., 1995; Dosi, 2007). Empirical results lend
support to these conjectures by showing that the regular unimodal firm
size distributions observed at the aggregate level can be decomposed
into much ‘messier’ distributions at the industry level, some of which are
visibly multimodal (Bottazzi and Secchi, 2003a; Bottazzi et al., 2008a).
For example, Bottazzi and Secchi (2005) present evidence of significant
bimodality in the firm size distribution of the worldwide pharmaceutical
Firm size distributions 17
industry, and relate this to a cleavage between the industry leaders and
fringe competitors.
Other work on the firm size distribution has focused on the evolution
of the shape of the distribution over time. It would appear that the initial
size distribution for new firms is particularly right-skewed, although the
log-size distribution tends to become more symmetric as time goes by.
This is consistent with observations that small young firms grow faster
than their larger counterparts. As a result, it has been suggested that the
lognormal can be seen as a kind of ‘limit distribution’ to which a given
cohort of firms will eventually converge. Lotti and Santarelli (2001)
present support for this hypothesis by tracking cohorts of new firms
in several sectors of Italian manufacturing. Cabral and Mata (2003)
find similar results in their analysis of cohorts of new Portuguese firms.
However, Cabral and Mata interpret their results by referring to finan-
cial constraints that restrict the scale of operations for new firms, but
become less binding over time, thus allowing these small firms to grow
relatively rapidly and reach their preferred size. The empirical analysis
presented in Angelini and Generale (2008) is similar to that of Cabral and
Mata (2003), in the sense that they observe that the firm size distribu-
tion for an entering cohort is extremely skewed, but that the distribution
becomes less skewed over time and approaches the lognormal. Angelini
and Generale (2008) then examine the financial constraints hypothesis
put forward by Cabral and Mata (2003), but they observe that financial
constraints play only a minor role in explaining this evolution, especially
in developed countries.
Although the skewed nature of the aggregate firm size distribution is a
robust finding, there may be some other features of this distribution that
are specific to countries. Table 2.1, taken from Bartelsman et al. (2005),
highlights some differences in the structure of industries across countries.
For instance, one observes that large firms account for a considerable
share of French industry, whereas in Italy firms tend to be much smaller
on average. (These international differences cannot simply be attributed to
differences in sectoral specialization across countries.)
Absolute number (%) Share of employment (%) Ave. no. employees per firm
Total Manufac- Business Total Manufac- Business Total Manufac- Business
economy turing services economy turing services economy turing services
US 86.7 69.9 87.9 16.6 5.8 20.6 26.4 80.3 21.4
Western Germany 87.9 77.9 90.2 23.6 11.3 33.8 17.0 39.1 11.5
France 78.6 73.6 78.8 13.9 17.0 12.1 33.5 32.1 35.7
18
Italy 93.1 87.5 96.5 34.4 30.3 46.3 10.5 15.3 6.8
UK – 74.9 – – 8.3 – – 40.7 –
Canada – – – – – – 12.7 40.5 12.0
Denmark 90.0 74.0 90.8 30.2 16.1 33.4 13.3 30.4 12.7
Finland 92.6 84.8 94.5 25.8 13.0 33.0 13.0 27.8 9.9
Netherlands 95.8 86.7 96.8 31.2 16.9 41.9 6.5 18.3 5.3
Portugal 86.3 70.5 92.8 27.7 15.7 39.8 16.8 31.0 11.4
Note: The columns labelled ‘share of employment’ refer to the employment share of firms with fewer than 20 employees.
In the limit, as t becomes large, the log(x0) term will become insignificant,
and we obtain
t
log (xt) < a es (2.4)
s51
In this way, a firm’s size at time t can be explained purely in terms of its
idiosyncratic history of multiplicative shocks. If we further assume that all
firms in an industry are independent realizations of independent and iden-
tically distributed (i.i.d.) growth shocks, then this stochastic process leads
to the emergence of a lognormal firm size distribution, that is:
1 (lnxt 2e# t) 2
b
P (xt) 5 e2a 2s2t
(2.5)
xt"2ps2
There are of course several serious limitations to such a simple vision of
industrial dynamics. We have already seen that the distribution of growth
rates is not normally distributed, but instead resembles the Laplace or
‘symmetric exponential’. Furthermore, contrary to results implied by
Gibrat’s model, it is not reasonable to suppose that the variance of firm
20 The growth of firms
0.1
0.01
Pr
0.001
le-04
le-05
–20 0 20 40 60 80 100 120
Age
0.1
0.01
0.001
Pr
le-04
le-05
le-06
0 20 40 60 80 100 120
Age
Source: Author’s elaboration of the data in Segarra et al. (2008), page 104.
The dearth of studies into the age distribution of firms could be because
data on age is difficult to obtain, or alternatively because data on age is not
entirely reliable due to the subjective nature of the reporting of age (Phillips
and Kirchhoff, 1989). In any case, more work on the age distribution is
clearly warranted. We anticipate that future work will show more of an
interest in the age distribution, and will also be able to clarify the nature of
the firm age distribution.
P (xt) 5 C # x2b
t (2.10)
2.5 CONCLUSION
We began this chapter by looking at empirical work into the firm size dis-
tribution (section 2.1). Two distributions, in particular, stand out as can-
didates for approximations to the empirical distribution – the lognormal
or the Pareto distribution. We then turned to theoretical models of firm
24 The growth of firms
It has long been suspected that the distribution of firm growth rates is
fat-tailed. In an early contribution, Ashton (1926) considers the growth
patterns of British textile firms and observes that ‘In their growth they
obey no one law. A few apparently undergo a steady expansion . . . With
others, increase in size takes place by a sudden leap’ (Ashton, 1926, pp.
572–3). Little (1962) investigates the distribution of growth rates, and
also finds that the distribution is fat-tailed. Similarly, Geroski and Gugler
(2004) compare the distribution of growth rates to the normal case and
comment on the fat-tailed nature of the empirical density. Recent empiri-
cal research, from an ‘econophysics’ background, has discovered that the
distribution of firm growth rates closely follows the parametric form of
the Laplace density. Using the Compustat database of US manufacturing
firms, Stanley et al. (1996) observe a ‘tent-shaped’ distribution on log-log
plots that corresponds to the symmetric exponential, or Laplace distribu-
tion (see also Amaral et al., 1997 and Lee et al., 1998). The quality of the
fit of the empirical distribution to the Laplace density is quite remarkable.
The Laplace distribution is also found to be a rather useful representation
when considering growth rates of firms in the worldwide pharmaceutical
industry (Bottazzi et al., 2001).
25
26 The growth of firms
1998
2000
2002
1
Prob
0.1
0.01
0.001
–3 –2 –1 0 1 2
Conditional growth rate
1998
2000
2002
1
Prob
0.1
0.01
0.001
–2 –1.5 –1 –0.5 0 0.5 1 1.5 2
Conditional growth rate
during recessions. Higson et al. (2002) and Higson et al. (2004) consider
the evolution of the first four moments of distributions of the growth of
sales, for large US and UK firms over periods of 30 years or more. They
observe that higher moments of the distribution of sales growth rates have
significant cyclical patterns. In particular, evidence from both US and UK
firms suggests that the variance and skewness are countercyclical, whereas
the kurtosis is procyclical. Higson et al. (2002) explain the countercyclical
movements in skewness in these words (p. 1551):
The central mass of the growth rate distribution responds more strongly to
the aggregate shock than the tails. So a negative shock moves the central mass
closer to the left of the distribution leaving the right tail behind and generates
positive skewness. A positive shock shifts the central mass to the right, closer to
the group of rapidly growing firms and away from the group of declining firms.
So negative skewness results.
Giulio Bottazzi and Angelo Secchi have played a major role in bringing
attention to the empirical growth rate distribution, and their theoretical
model is also an important contribution to the literature.
The Bottazzi and Secchi (2006a) model3 is in line with previous ‘island
models’ of industry evolution in the sense that it conceives of firm growth
as a random process – ‘in our model luck is the principal factor that
finally distinguishes winners from losers among the contenders’ (Bottazzi
and Secchi, 2006a, p. 236). The main point of departure from the ‘island’
models, however, is that their model emphasizes the role of competition
between organizations in shaping the growth rate distribution. While
previous models tended to treat firms as independent entities, competi-
tive effects between firms are prominent in the Bottazzi and Secchi model.
According to this model, firms compete in a given industry for a finite
number of pre-existing business opportunities, that, once obtained, can
be translated into growth. Of central importance is the positive feedback
mechanism which assigns the growth opportunities to the different firms.
This assignment mechanism postulates the existence of dynamic increas-
ing returns to growth (because of economies of scale, economies of scope,
network externalities, knowledge accumulation, and so on) and posits that
market success is cumulative or self-reinforcing. Firms that have already
received a growth opportunity are more likely to obtain another one. As a
result, many opportunities tend to concentrate in a few firms. Bottazzi and
Secchi show that, under certain conditions, the growth rate distribution
approaches the empirically-observed Laplace distribution.
It should be noted, however, that the positive feedback mechanism is
assumed to operate in the short term only (that is for periods of up to one
year). This short duration of the increasing returns mechanism is, effec-
tively, required in order to reconcile the model with empirical work into the
Growth rate distributions 31
A simplified model
This model considers the special case of the propagation of employment
growth throughout the various levels of a firm’s hierarchy. The organiza-
tion of production in a hierarchy is indeed a general feature of all firms – in
fact, in the Transaction–Cost–Economics literature, the words ‘firm’ and
‘hierarchy’ are used almost interchangeably. The firm is characterized as
being composed of a relatively large number of hierarchies.7 The bottom
layer of the firm (that is, the very lowest hierarchical level) is composed
exclusively of productive workers, whilst all of the other levels are com-
posed of managers whose task is to supervise either productive workers or
subordinate managers (see Figure 3.3 for an illustration). A firm grows by
adding a productive worker. The number of managers is determined by the
number of productive workers and also by limits on the efficient span of
control, a, which correspond to the maximum number of subordinates that
Growth rate distributions 33
Supervisors Supervisors
Productive Productive
workers workers
a manager can effectively supervise. ‘At executive levels [the span of control]
is seldom less than three, and seldom more than ten, and usually lies within
narrower bounds – particularly if we take averages over all executives in an
organization at a given level.’ (Simon, 1957, p. 32). In this model, though,
we do not need to attribute any specific numerical value to a and so we leave
it in algebraic form. It is computationally helpful, and also theoretically
meaningful, however, to assume that a is a whole number that is strictly
greater than unity (that is, a [ N1, a . 1). For analytical simplicity, we
assume that a is a constant and does not vary either within a hierarchical
level or across levels (for a discussion of the plausibility of this assumption,
see Williamson, 1967, p. 128). For the purposes of this model, we also must
assume that adjustment of the firm’s hierarchical organization to additional
productive workers occurs within one time period. Finally, we assume that
the firm is initially at a stable state, such that it is already efficiently organ-
ized in the sense that it is not possible for it to employ fewer managers given
the number of productive workers and its given value of a (that is the limit
on the efficient span of control). The reader may notice major similarities
between the model developed here and the executive compensation model
of Simon (1957) and the information flows model of Williamson (1967).
The fact that the same hierarchical model has been applied in quite different
contexts lends credibility to its use here – indeed, we cannot be accused of
having conclusions that emerge from ad hoc modelling assumptions.
A summary understanding can be obtained by looking at Figure 3.3.
34 The growth of firms
Formal model Let us begin with the simplest possible case, considering
one firm that grows by adding just one productive worker (Δn 51). If new
productive workers can be integrated without having to add a supervisor,
we have Δn 5 Δx; that is the number of productive workers added is equal to
change in total employment. It is possible, however, that all of the managers
in the second hierarchical level (that is, those that supervise the productive
workers) are already fully occupied. This will occur when the number of
productive workers (before adding the new one) is exactly a multiple of a.
If this is the case, the arrival of the supplementary worker will require that
one supplementary manager be hired at the next hierarchical level. This
scenario will occur with probability 1/a. However, the arrival of this new
manager at the second level may add to the workload of managers on the
third hierarchical level, and so on. The probability that the addition of a
productive worker leads to at least two managers being hired at two succes-
sive levels is 1/a 3 1/a 5 1/a2. We can continue with this reasoning to end
up with the following distribution of employment growth:
(Log) Prob.
1-(1/)
-1[1-(1/)]
-2[1-(1/)]
-3[1-(1/)]
1 2 3 4
Growth of total employment (x)
Discussion The model is admittedly far too simple to be realistic, yet its
simplicity makes for greater visibility of the source of the emergence of
the heavy-tailed distribution. The model can be seen as the simplest model
in a family of possible models that view firms as coherent collections of
resources that are complementary and discrete. These latter are subject to
localized interactions and embedded in an organization that tends to a crit-
ical state of full utilization of its resources. In this context, a small growth
stimulus working through local interaction channels can be transmitted
throughout a firm to produce potentially large-scale effects. We argue that
it is these properties that explain the emergence of the observed fat-tailed
growth rate distributions.
The model describes the dynamics of a single, ‘autistic’ organization and
makes no attempt to account for competitive interactions between firms.
In our view, this is not a serious flaw. Other explanations of the fat-tailed
growth rate distribution have emphasized the complex nature of inter-firm
competition as the source of the emergence of the observed distribution
(for example Bottazzi and Secchi, 2006a; McKelvey and Andriani, 2005).
Recent empirical work has nonetheless cast doubt on the importance of
Growth rate distributions 37
3.3 CONCLUSION
In the limit, as t becomes large, the log(x0) term will become insignificant,
and we obtain
39
40 The growth of firms
t
log (xt) < a es (4.4)
s51
In this way, a firm’s size at time t can be explained purely in terms of its
idiosyncratic history of multiplicative shocks. If we further assume that all
firms in an industry are independent realizations of i.i.d. normally distrib-
uted growth shocks, then this stochastic process leads to the emergence of
a lognormal firm size distribution.
Although Gibrat’s (1931) seminal book did not provoke much of an imme-
diate reaction, in recent decades it has spawned a flood of empirical work.
Nowadays, Gibrat’s ‘Law of Proportionate Effect’ constitutes a bench-
mark model for a broad range of investigations into industrial dynamics.
Another possible reason for the popularity of research into Gibrat’s law,
one could suggest quite cynically, is that it is a relatively easy paper to
write. First of all, it has been argued that there is a minimalistic theoretical
background behind the process (because growth is assumed to be purely
random). Then, all that needs to be done is to take the IO economist’s
‘favourite’ variable (that is, firm size, a variable which is easily observable
and readily available) and regress the difference on the lagged level. In
addition, few control variables are required beyond industry dummies and
year dummies, because growth rates are characteristically random.
Empirical investigations of Gibrat’s law rely on estimation of equations
of the type:
firms have a higher exit hazard which may obfuscate the relationship
between size and growth. The second version, on the other hand, consid-
ers only those firms that survive. Research along these lines has typically
shown that smaller firms have higher expected growth rates than larger
firms. The third version considers only those large surviving firms that
are already larger than the industry Minimum Efficient Scale of produc-
tion (with exiting firms often being excluded from the analysis). Generally
speaking, empirical analysis corresponding to this third approach suggests
that growth rates are more or less independent from firm size, which lends
support to Gibrat’s law.
The early studies focused on large firms only, presumably partly due to
reasons of data availability. A series of papers analysing UK manufactur-
ing firms found a value of b greater than unity, which would indicate a
tendency for larger firms to have higher percentage growth rates (Hart,
1962; Samuels, 1965; Prais, 1974; Singh and Whittington, 1975).
However, the majority of subsequent studies using more recent datasets
have found values of b slightly lower than unity, which implies that, on
average, small firms seem to grow faster than larger firms. This result is fre-
quently labelled ‘reversion to the mean size’ or ‘mean-reversion’.2 Among
a large and growing body of research that reports a negative relationship
between size and growth, we can mention here the work by Kumar (1985)
and Dunne and Hughes (1994) for quoted UK manufacturing firms, Hall
(1987), Amirkhalkhali and Mukhopadhyay (1993) and Bottazzi and Secchi
(2003a) for quoted US manufacturing firms (see also Evans, 1987a and
1987b, for US manufacturing firms of a somewhat smaller size), Gabe
and Kraybill (2002) for establishments in Ohio, Goddard et al. (2002) for
quoted Japanese manufacturing firms, and Sleuwaegen and Goedhuys
(2002) for manufacturing firms from Côte d’Ivoire.
Studies focusing on small businesses have also found a negative rela-
tionship between firm size and expected growth – see for example Yasuda
(2005) for Japanese manufacturing firms, Segarra and Callejon (2002) and
Calvo (2006) for Spanish manufacturing, McPherson (1996) for Southern
African micro businesses, and Wagner (1992) and Almus and Nerlinger
(2000) for German manufacturing. Dunne et al. (1989) analyse plant-level
data (as opposed to firm-level data) and also observe that growth rates
decline along size classes. Research into Gibrat’s law using data for specific
sectors also finds that small firms grow relatively faster (see for example
Barron et al. (1994) for New York credit unions, Weiss (1998) for Austrian
farms, Liu et al. (1999) for Taiwanese electronics plants, and Bottazzi and
Secchi (2005) for an analysis of the worldwide pharmaceutical sector).
Indeed, there is a lot of evidence that a slight negative dependence of
growth rate on size is present at various levels of industrial aggregation.
42 The growth of firms
smaller firms initially grow faster, it becomes more difficult to reject the
independence of size and growth as time passes. Similarly, results reported
by Becchetti and Trovato (2002) for Italian manufacturing firms, Geroski
and Gugler (2004) for large European firms and Cefis et al. (2007) for
the worldwide pharmaceutical industry also find that the growth of large
firms is independent of their size, although including smaller firms in the
analysis introduces a dependence of growth on size. After digging around
in the available evidence, Caves (1998) concludes his survey of industrial
dynamics with the ‘substantive conclusion’ that Gibrat’s law holds for
firms above a certain size threshold, whilst for smaller firms growth rates
decrease with size. You surveyed the literature and arrived at a similar
conclusion (You, 1995).
Concern about econometric issues has often been raised. Sample selec-
tion bias, or ‘sample attrition’, is one of the main problems, because
smaller firms have a higher probability of exit. Failure to account for the
fact that exit hazards decrease with size may lead to underestimation of
the regression coefficient (that is, b). Hall (1987) was among the first to
tackle the problem of sample selection, using a generalized Tobit model.
She concludes that selection bias does not seem to account for the nega-
tive relationship between size and growth in her data. Similar conclusions
were reached by McPherson (1996). An alternative way of correcting for
sample selection is by applying Heckman’s two-step procedure. This is the
methodology used by Harhoff et al. (1998), who also observe that selec-
tion bias has only a small influence on the Gibrat coefficient. In short, the
‘problem of sample selection does not seem to significantly affect the rela-
tionship between growth rate and size of firm’ (Marsili, 2001, p. 15). The
possibility of heteroskedasticity is also frequently mentioned, although it
can be corrected for quite easily, for example by applying White’s (1980)
procedure. In any case, heteroskedasticity does not introduce any asymp-
totic bias in the coefficient estimates. Serial correlation in growth rates
can lead to biased estimates, although Chesher (1979) proposes a simple
framework for dealing with this. Finally, Hall (1987) investigates whether
‘errors-in-variables’ may be influencing the regression results, but con-
cludes that measurement error does not appear to be an important factor
in her dataset.
Hymer and Pashigian (1962) were among the first to draw attention to the
negative relationship between growth rate variance and firm size. If firms
can be seen as a collection of ‘components’ or ‘departments’, then the
44 The growth of firms
overall variance of the growth rate of the firm is a function of the growth
rate variance of these individual departments. In many cases, the variance
of the firm’s growth rate will decrease with firm size. For example, in the
case where these departments (i) are of approximately equal size, such that
the size of the firm is roughly proportional to the number of components;
and (ii) have growth rates that are perfectly independent from each other,
then Central Limit Theorem leads us to expect a decrease in growth rate
variance that is proportional to the inverse square root of the firm’s size.
However, Hymer and Pashigian (1962) were puzzled by the fact that the
rate of decrease of growth rate variance with size was lower than the rate
that would be observed if large firms were just aggregations of independent
departments. At the same time, they found no evidence of economies of
scale. They saw this as an anomaly in a world of risk-averse agents. Why
should firms grow to a large size, if there are no economies of scale, and if
the growth rate variance of a large firm is higher than the corresponding
variance of an equivalent group of smaller firms? Subsequent studies did
not attempt to answer this question, but they did bear in mind the exist-
ence of a negative relationship between growth rate variance and firm size.
As a consequence, empirical analyses of Gibrat’s law began to correct for
heteroskedasticity in firm growth rates (for example Hall, 1987; Evans,
1987a,b; Dunne and Hughes, 1994; Hart and Oulton, 1996; and Harhoff
et al., 1998).
In recent years efforts have been made to quantify the scaling of the
variance of growth rates with firm size. This scaling relationship can be
summarized in terms of the following power law: s (gi) ~ebsi; where s(gi)
is the standard deviation of the growth rate of firm i, b is a coefficient
to be estimated, and si is the size (total sales) of firm i. Values of b have
consistently been estimated as being around 20.2 for US manufacturing
firms (Amaral et al., 1997, 1998; Bottazzi and Secchi, 2003a) and also for
firms in the worldwide pharmaceutical industry (Bottazzi et al., 2001; De
Fabritiis et al., 2003; Matia et al., 2004; Bottazzi and Secchi, 2006b). Lee
et al. (1998) find that a scaling exponent of 20.15 is able to describe the
scaling of growth rate variance for both quoted US manufacturing firms
and the GDP of countries. For French manufacturing firms, the analysis
in Bottazzi et al. (2008a) yields estimates of b of around 20.07, although
in the case of Italian manufacturing firms Bottazzi et al. (2007) fail to find
any relation between growth rate variability and size.
The discussion in Lee et al. (1998, p. 3277) gives us a better understand-
ing of the values taken by b, the scaling exponent. If the growth rates of
divisions of a large diversified firm are perfectly correlated, we should
expect a value of b 5 0. On the other hand, if a firm can be viewed as an
amalgamation of perfectly independent sub-units, we expect a value of b
Gibrat’s law 45
5 20.5. The fact that the estimated exponents are between these extreme
values of 0 and −0.5 suggests that the constituent departments of a firm
have growth patterns that are somewhat correlated.
Matia et al. (2004) and Bottazzi and Secchi (2006b) return to the scaling-
of-variance puzzle by considering firms as being composed of a certain
number of products that correspond to independent sub-markets. The
average size of the sub-markets increases with firm size, but the growth
rates are independent across sub-markets. These authors provide support
for their model by examining evidence from the worldwide pharmaceutical
industry, where a firm’s portfolio of activities can be decomposed to a fine
level of aggregation. As a result,
the explanation of the relationship between the variance of the growth rates
distribution and the size of the firm based on the Central Limit Theorem is valid,
as long as one considers the actual number of sub-markets a firm operates in,
instead of assuming that this number is somehow proportional to the size of the
firm. (Bottazzi and Secchi, 2006b, p. 860)
The model described in Matia et al. (2004) and Bottazzi and Secchi
(2006b) bears a certain similarity with the model in Amaral et al. (1998,
2001), who explain scaling of variance in terms of firms being composed of
independent ‘divisions’ in a diversified firm, rather than independent ‘sub-
markets’. Relatedly, Sutton (2002) explains the scaling of variance with size
by considering that firms are composed of independent business lines that
display great heterogeneity in size. Coad (2008b) conducts an empirical
investigation that can be situated along these lines, relating the variance of
growth rates to a firm’s multiplant structure. The variance of growth rates
is observed to decline with number of production plants, although interest-
ingly enough this decline is not monotonic. It may be that the tendency for
large size to be associated with lower variance is partially offset by a greater
propensity to take risks in larger firms.
Early empirical studies into the growth of firms considered serial correla-
tion when growth was measured over a period of four to six years. Positive
autocorrelation of 33 per cent was observed by Ijiri and Simon (1967) for
large US firms, and a similar magnitude of 30 per cent was reported by
Singh and Whittington (1975) for UK firms. However, much weaker auto-
correlation was later reported in comparable studies by Kumar (1985) and
Dunne and Hughes (1994).
46 The growth of firms
and observe that firms do in fact have idiosyncratic growth patterns that
are not visible simply by looking at averages across firms. They create a
purely random ‘benchmark’ case in which the growth rates of all firms
are pooled together and then growth rates are extracted randomly to con-
struct growth patterns for ‘artificial firms’. Bootstrap resampling methods
allow them to generate a distribution of autocorrelation coefficients for
this random scenario. They then compare this stochastic benchmark with
the empirical distribution of autocorrelation coefficients (see for example
Figure 5 in Bottazzi et al. (2002) for the case of autocorrelation of employ-
ment growth). The differences between the distributions are supported by
formal statistical tests (that is, Kolmogorov–Smirnov tests). The authors
conclude that firm growth patterns are indeed idiosyncratic, that they do
have a memory process, and that there are indeed persistent asymmetries
in growth dynamics across firms.
Coad (2007a) also explores the issue of heterogeneous growth profiles
across firms and goes some way towards finding regularities in growth rate
autocorrelation patterns. A firm’s growth dynamics are seen to depend
on two dimensions – a firm’s size and its lagged growth rate. First of all,
it is demonstated that smaller firms are more prone to experience nega-
tive autocorrelation, whilst larger firms have a tendency towards positive
autocorrelation. This is consistent with propositions that small and large
firms operate on a different ‘frequency’ or timescale, with the actions of
large firms unfolding over a longer time horizon. This dependence of
autocorrelation on firm size helps to explain why the studies reviewed
above yielded different autocorrelation coefficients for databases with
different firm size compositions. Second, Coad (2007a) demonstrates that
the autocorrelation coefficient depends on the growth rate. Firms whose
growth rate is close to the average in one year are likely not to experience
any autocorrelation in the following year. However, those firms that expe-
rience extreme growth rates (either extreme positive or negative growth
rates) are likely to experience considerable negative autocorrelation. This
is especially true for fast-growth small firms, whose growth patterns are
particularly erratic. These findings are consonant with work by Garnsey
and Heffernan (2005), who highlight the ‘prevalence of interruptions to
growth’ (p. 675) for small firms, after observing that only a small fraction
of small firms grew continuously over their period of analysis. Large firms,
however, undergo a smoother growth experience – they are likely to experi-
ence positive autocorrelation irrespective of their growth rate in the previ-
ous period. Mrs Penrose offered this explanation: ‘Large firms often plan
further ahead than do smaller firms, partly because their greater financial
strength enables them to afford it, and partly because the nature of their
operations more or less forces them to do so.’ (Penrose, 1959, p. 244).
48 The growth of firms
4.5 CONCLUSION
Like it or not, Gibrat’s law still reigns supreme as the leading model of
firm growth. Gibrat’s Law is essentially a random statistical process, and
as such it is often criticized by economists as having no theoretical founda-
tion. In this chapter we began by introducing Gibrat’s model of the ‘law
of proportionate effect’ (section 4.1) before reviewing empirical work on
the law and its implications. In section 4.2 we considered the relationship
between firm growth and size. The empirical literature is huge – much work
has been done on many different datasets. If it were indeed possible to gen-
eralize, then we would suggest the following recapitulation of this work. It
appears that smaller firms tend to grow faster than larger firms, although
above a certain size threshold these differences fade out, such that expected
growth rates are more or less independent of firm size. Gibrat’s law also
yields some other testable implications. In section 4.3 we looked at the
relationship between firm size and growth rate variance. In nearly all data-
sets considered, growth rate variance decreases with firm size. In section
4.4 we searched for an autocorrelation in firm growth rates. In many cases
there appears to be an autocorrelation structure in growth rates, although
the results are far from harmonious and suggest that autocorrelation can
either be negative or positive, or even in between (that is, no significant
autocorrelation). These conflicting results are reconciled somewhat by the
observation that small firms tend to display negative autocorrelation while
that of larger firms is more positive. Furthermore, it appears that small
firms that experienced rapid growth in one year are unlikely to be able to
repeat this in the following year.
5. Profits, productivity and firm
growth
Theoretical work has long been interested in the relationship between a
firm’s relative performance (measured either in terms of profits or pro-
ductivity) and its growth rate. In fact, a number of theoretical authors
tend to take it for granted that firms with higher performance will rein-
vest their profits into growth, such that the more efficient firms will have
higher growth rates. Empirical work into the matter, however, suggests
that the expected positive relationship between performance and growth
is generally lower than expected, or even non-existent.
We begin this chapter by looking at the relationship between profits
and growth (section 5.1). There remains a controversy, however, over
how the relationship between financial performance and growth should be
interpreted. Neoclassical economists, for example, think that in a perfect
world, current financial performance should not be related to investment.
If investment is related to financial performance, then financial constraints
are preventing firms from undertaking their optimal expansion plans.
Evolutionary economists take an opposite view – ideally, firm expansion
should be related to current profits. This is because progressive industrial
development requires the reallocation of scarce resources towards the more
efficient producers.
In addition to profits, many theorists have focused on the relationship
between productivity and firm growth. In fact, profits and productivity are
quite closely correlated and they can be considered as alternative indicators
of relative performance. The relationship between productivity and growth
is addressed in section 5.2.
In section 5.3 we look at models that look at the coevolution of both
profits, productivity, and other aspects of firm growth. These models
consider not only the effect of profits and productivity on subsequent firm
growth, but also the association between firm growth and subsequent
growth in profits or productivity. Section 5.4 concludes.
49
50 The growth of firms
0 0
–1 –1
–0.5 0 0.5 1 –0.5 0 0.5 1
profit rate (t-1) profit rate (t-1)
ISIC24: chemicals and chemical products ISIC25: rubber and plastic products
1 1
growth rate (t-1:t)
0.5 0.5
0 0
–1 –1
–0.5 0 0.5 1 –0.5 0 0.5 1
profit rate (t-1) profit rate (t-1)
0.5 0.5
0 0
–1 –1
–0.5 0 0.5 1 –0.5 0 0.5 1
profit rate (t-1) profit rate (t-1)
5.1.1 q Theory
Equation Description
(1) Intertemporal capital market arbitrage condition
(2) Solving (1) on an infinite horizon
(3) Defining the discount factor b over an infinite horizon
(4) Substituting for dividend payments in the firm’s stock market value
(5) Defining the firm’s after-tax net revenue
(6) First-order condition for investment
(7) The evolution of the shadow price of capital
(8) Rearranging (6) to obtain marginal q
(9) Rearranging (8) assuming a quadratic functional form for
adjustment costs
(10) Rewriting marginal q assuming linear homogeneity of production
and adjustment costs
(11) Expressing the expected depreciation allowances on an infinite
horizon
(12) Expressing the expected present value of all cash flows associated
with debt
(13) Regression equation
2000; Gomes, 2001). This may occur if the stock market is not perfectly effi-
cient at foreseeing a firm’s fortunes or allocating resources. Furthermore,
the denominator of q includes only fixed capital, and regrettably it does
not include those elements that are truly valued by shareholders and that
cannot be easily bought or sold on asset markets, such as management
skill, human capital or R&D capital. Furthermore, q may not be a good
predictor of investment if managers are boundedly rational, or if they just
don’t choose to grow on the basis of maximizing shareholder value. q may
also fail to predict investment if the other assumptions mentioned above
do not hold.
only the largest and most mature firms are likely to face a smoothly increasing
loan interest rate . . . Small and medium-sized firms are less likely to have access
to impersonal centralized debt markets. . . . during periods of tight credit, small
and medium-sized borrowers are often denied loans in favor of better-quality
borrowers. (Fazzari et al., 1988a, p. 153)6
flow, such as R&D investment (see for example Himmelberg and Petersen,
1994; Bougheas et al., 2003; Bond et al., 2003a; and Brown et al., 2007).
A common theme in many of these studies is that, whenever investment
(or firm growth) is associated with changes in cash flow, it tends to be pre-
sented as ‘bad news’. It is implicitly assumed that any investment–cash flow
sensitivities are signs of financial constraints, that investment opportunities
have been forgone, and also that these investment opportunities would
have been ‘optimal’.7 An interpretation based on market imperfection is
evoked, and policy makers have frequently been urged to intervene to help
constrained firms to grow.
However, the Fazzari et al. (1988a) approach to investment research has
recently met an extensive criticism by Kaplan and Zingales (1997, 2000).8
To begin with, Kaplan and Zingales present a theoretical model to show
that any sensitivity of investment to cash flow should not be interpreted
as evidence of financial constraints (see also the theoretical model by Alti,
2003). They also re-examine the original Fazzari et al. (1988a) database in
conjunction with a scrutiny of annual company reports of these companies,
and observe that the highest investment–cash flow sensitivities actually
belong to those firms that seem to be the least financially constrained.
Indeed, ‘wrong-way’ differential investment–cash flow sensitivities (that is
where firms classified as more constrained a priori display lower sensitiv-
ity of investment to cash flow) have also been found by a number of other
researchers, such as Gilchrist and Himmelberg (1995), Kadapakkam et al.
(1998), Cleary (1999) and Erickson and Whited (2000). One notable example
mentioned by Kaplan and Zingales (2000) is that, in 1997, Microsoft would
have been labelled as ‘financially constrained’ according to the classifica-
tion schemes of Fazzari et al. (1988a, 2000) even though it had almost $9
billion in cash, corresponding to 18 times its capital expenditures!
where d stands for the variation in the infinitesimal interval (t, t 1 dt), and
xi represents the market share of firm i in a population of competing firms.
Fi is the level of ‘fitness’ of the considered firm, where fitness corresponds
to relative financial performance or perhaps some measure of relative pro-
– –
S
ductivity. F is the average fitness in the population, i.e. F 5 ixiFi, and a
is a parameter. It is straightforward to see that this equation favours the
above-average firms with increasing market share, whilst reducing that of
‘weaker’, less profitable firms.
Although empirical investigations of the evolutionary principle of
‘growth of the fitter’ are rather scarce, some empirical studies of evolu-
tionary inspiration can be found in Dosi (2007), Bottazzi et al. (2008b)
and Coad et al. (2008) for Italian manufacturing firms; Coad (2007b,d)
for French manufacturing firms; and Coad and Rao (2009) for US manu-
facturing. Dosi (2007) and Bottazzi et al. (2008b) present scatterplots of
the relationships between profits, productivity and sales growth. While
there appears to be a relationship between profits and productivity, these
variables seem to be unrelated to firm growth. Coad (2007d) applies panel-
data instrumental variable techniques to account for a lag structure, firm-
specific time invariant effects, and endogeneity in the relationship between
profits and growth. He finds a statistically significant relationship between
financial performance and sales growth for French manufacturing firms.
Nevertheless, the magnitude of the coefficient is so small that he concludes
‘it may be more useful to consider a firm’s profit rate and its subsequent
growth rate as entirely independent’ (p. 385). Coad (2007d) also observes
that the effect of profits on firm growth appears to be overshadowed by
the effect of firm growth on subsequent profitability. Coad (2007b), Coad
and Rao (2009) and Coad et al. (2008) focus on the co-evolution over time
58 The growth of firms
60
explanation could be that firms are financially constrained.
Imperfect markets (I/K)it 5 b1qit 1 b2 (CF/K)it 1 eit Any explanatory power of cash flow over and above that of q
(e.g. Fazzari et al., indicates financial constraints.
1988a)
Reduced form models (DS/S) 5 b2 (CF/S) it 1 eit Cash flow taken as a proxy for financial constraints (without
(e.g. Fagiolo and controlling for q). Any sensitivity of sales growth to cash flow is
Luzzi, 2006) interpreted as financial constraints.
Evolutionary approach (DS/S) 5 b2 (OM/S) it 1 eit Sales growth should be associated with operating margin according
(e.g. Coad, 2007c) to the principle of ‘growth of the fitter’.
Notes: I is investment for firm i at time t, K is fixed assets, q is Tobin’s q, Y is output, CF is cash flow, (DS/S) is the growth rate of sales, OM is
operating margin, e is the residual error term. ∏it 5 pitF(Kit, Lit) − pitG(Iit, Kit) − witLit (see Bond et al., 2003b, p. 156).
Profits, productivity and firm growth 61
they want. To sum up, one caricature of the neoclassical approach could
be: ‘Assume firms are efficient. Financial-market imperfections prevent
them from getting enough funding. Policy should intervene, perhaps by
subsidizing firm investment.’
A major caveat of the mainstream neoclassical literature is that it takes
as a starting point the assumption that firms are perfectly rational and
will invest only if this increases their long-term profits. In this stream of
literature, questions regarding the existence of imperfections in the finan-
cial system take centre stage, while questions regarding suboptimality of
firms are largely ignored. Evolutionary economics, in constrast, discards
assumptions of hyper-rationality and starts from the hypothesis that firms
are heterogeneous, boundedly rational, and that not all firms deserve to
grow.
But how important are financial constraints for economic development,
really? In an attempt to answer this question, we focus our investigations
where financial constraints are alleged to matter the most: the case of
small young entrepreneurial firms. We begin by referring to a growing
literature on the theme of excessive start-up (that is, entry beyond the
socially optimum level) which is surveyed extensively in Santarelli and
Vivarelli (2007). These authors observe that ‘at the end of the 90s, in
the UK the incidence of people starting a firm not because of a market
opportunity but just because they had no better choice was about 22%’ (p.
461). This can be compared to statistics (admittedly from other sources)
reporting that there is only ‘a minority of firms (about 15–20%) indicating
the desire to introduce product and/or process innovation as a funda-
mental reason to start a new independent economic activity’ (Santarelli
and Vivarelli, 2007, p. 463). Santarelli and Vivarelli (2007) emphasize
the important role that market selection has to play, and consider that
‘(early) failure should be seen as socially optimal rather than the result of
either financial market imperfections or other market failures.’ (p. 473).
As a result, they judge that financial constraints are not a major problem
affecting entrepreneurship and the growth of (small) firms, and conclude
that ‘modern developed economies are affected by too many start-ups
and that policy subsidies have contributed to an overall “excess of entry”
which – far from fostering economic growth – may just fuel turbulence
and market churning.’ (p. 475).
Other contributions, of a theoretical nature, have related over-entry
to over-optimistic forecasts of entrepreneurs (Dosi and Lovallo, 1998;
Camerer and Lovallo, 1999 and Arabsheibani et al., 2000). These articles
go on to suggest that entry of over-confident low-quality entrepreneurs
may even crowd out higher-quality entrepreneurs. It has been argued
that marginal entrepreneurs can free-ride on the credentials of more able
62 The growth of firms
entrepreneurs, thus bringing down the average quality of the credit pool
(de Meza and Webb, 1987, 1999; de Meza, 2002). Marginal entrants should
thus be discouraged from entering. This body of theoretical work also sug-
gests that the use of internal finance to fund start-ups has beneficial effects
on start-up survival rates (through ‘incentive effects’) and also plays a role
in reducing moral hazard. As a result, it has been suggested that start-
ups should not be subsidized.12 Empirical evidence on excessive start-up
should also be taken into consideration (for example Dunne et al., 1988;
Bartelsman et al., 2005). These studies highlight the waste associated with
entry of new firms, by showing that a large proportion of entrants can be
expected to fail only a few years.
It is nonetheless rather unsettling to observe that the recommendations
emerging from the neoclassical literature have, to a certain extent, been
able to guide policy.
A belief that capital-market failure has held back enterprise has been a factor
behind policies designed to encourage start-ups and business expansion.
Examples include the grants and subsidies provided by the Federal Small
Business Administration in the United States and the Loan Guarantee Scheme
in the United Kingdom. (de Meza and Webb, 1999, p. 153)
In the United States, for example, there have been public initiatives
to provide finance to small firms that are suspected of being ‘financially
constrained’. According to Lerner (2002), these ‘public venture capital
programmes are often characterised by a considerable number of under-
achieving firms. . . . The end result can be a stream of government funding
being awarded to companies that consistently underachieve.’ (pp. 81–82).
Levenson and Willard (2000) are also critical of schemes such as the
provision of guaranteed loans to small firms.13 They remark that ‘there is
no direct evidence that small firms are, in fact, credit rationed in formal
capital markets’ (p. 84). Using data from a national survey in 1988–89, they
calculate an upper bound for the share of small businesses that were credit-
rationed as 6.36 per cent, and conclude that ‘the extent of true credit ration-
ing appears quite limited’ (p. 83). Likewise, Hughes (1997) focuses on UK
SMEs and concludes that ‘the evidence for general equity or debt gaps in
the UK is weak. If anything, SME funding was too easy in the boom of
the late 1980s.’ (p. 151). Cressy (1996) reaches a similar conclusion: ‘an
appropriate government policy should be to make business startups more
difficult, rather than less’ (Cressy, 1996, p. 1266).
Finally, it should also be noted that apart from being a potential waste of
funds, government initiatives to alleviate financial constraints also have the
drawback of encouraging socially wasteful rent-seeking behaviour (Little,
1987; Lerner, 2002).
Profits, productivity and firm growth 63
Summary
It is perhaps quite natural to assume that the most productive firms will
grow while the least productive will decrease in size. A number of theo-
retical contributions have made strong statements along these lines (see
in particular the evolutionary literature in section 5.1.3). This assumption
does not seem to be borne out by empirical work, however. A number of
studies have cast doubt on the validity of the evolutionary principle of
‘growth of the fitter’, when relative productivity is taken as a proxy for
fitness. One explanation for this is that while some firms become more
productive through expansion, others become more productive through
downsizing. An illustration of this is provided by Baily et al. (1996) who
observe that, among plants with increasing labour productivity between
1977 and 1987, firms that grew in terms of employees were balanced out
64 The growth of firms
by firms that decreased employment. They find that about a third of labour
productivity growth is attributable to growing firms, about a third to
downsizing firms, and the remaining third is attributable to the processes
of entry and exit. Similarly, Foster et al. (1998) also fail to find a robust
significant relationship between establishment-level labour productivity or
multifactor productivity and growth (see also the review in Bartelsman and
Doms, 2000, pp. 583–4). In addition, using a database of Italian manufac-
turing firms, Bottazzi et al. (2002, 2008b) fail to find a robust relationship
between productivity and growth.14
Other researchers working on other datasets have been able to detect a
positive association between relative productivity and firm growth. Pavcnik
(2002) investigates productivity growth among Chilean manufacturing plants
as the Chilean economy was undergoing significant liberalization and deregu-
lation. She observed that aggregate productivity increased by 19 per cent
over a seven-year period, and that most of this productivity growth was due
to the real-location of resources from the less to the more efficient producers.
Sleuwaegen and Goedhuys (2002) also observe a positive relationship between
productive efficiency and sales growth in their sample of Ivorian manufactur-
ing firms (although the effect is not always statistically significant). Likewise,
Liu et al. (1999) observe that labour productivity has a positive effect on
growth in their sample of Taiwanese electronics plants. Maksimovic and
Phillips (2002) find evidence that selection effects also operate within con-
glomerate firms. They observe that the growth rates of plants in both single-
and multiple-segment firms are significantly and positively related to their
productivity. The sensitivity of plant growth to productivity is greater for
single-segment firms than it is for multiple-segment firms, though.
There is ample evidence suggesting that low productivity helps to predict
exit.15 Bellone et al. (2008) observe that both productivity and profitabil-
ity are positively related to the probability of survival, although the chief
factor that influences survival is profitability. Schlingemann et al. (2002)
report that conglomerates are more likely to divest peripheral segments
that are performing poorly (see their Table 6). Maksimovic and Phillips
(2008) perform a plant-level analysis; they find that more productive plants
are less likely to be shut down (see in particular their Table 11).
Relatedly, many researchers have provided indirect evidence on the
matter, by reporting that industries experience positive productivity
growth that is due, in part, to the exit of relatively inefficient establishments
(see for example Griliches and Regev, 1995; Foster et al., 2006; and Foster
et al., 2008). Although Foster et al. (2008) find that low productivity is
associated with exit, they nonetheless stress that it is profitability, rather
than productivity, that is the true determinant of survival.
Nonetheless, it seems that productivity levels are not very helpful in
Profits, productivity and firm growth 65
predicting growth rates. Put differently, it appears that selection only oper-
ates via elimination of the least productive firms or establishments, while
the mechanism of selection via differential growth does not appear to be as
strong. As a result, the mechanism of selection appears to be rather ‘sub-
optimal’ in the sense that its effectiveness is lower than it could conceivably
be. For Baily and Farrell (2006), the lack of a positive relationship between
relative productivity and growth corresponds to a lack of competition.
In an ideal scenario, firms would compete for growth opportunities, and
selective pressures would attribute these growth opportunities discrimi-
nating in favour of the most productive firms. In this way, there would be
some sort of dynamic efficient reallocation at work, whereby an economy’s
scarce resources are redistributed to those firms that are able to employ
them most efficiently. In reality, however, this mechanism does not seem
to be operating. Instead, the evidence is consistent with the hypothesis that
many of the more productive firms may not actually seek to grow, or may
be unable to grow. As a consequence, the absence of selection via differen-
tial growth is evidence of missed productivity growth opportunities for the
economy as a whole. Whilst we can put forward here that stimulating the
growth of high-productivity firms might constitute an objective for policy,
it is evident that there are large question marks surrounding how such a
policy intervention might be engineered.
1 a DLPitDqit (5.2)
i[C
1 a qit (LPit 2 LPt21) 2 a qi,t21 (LPi,t21 2 LPt21)
i[N i[X
1 a DTFPitDqit (5.3)
i[C
1a qit (TFPit2TFPt21) 2 a qi,t21 (TFPi,t212TFPt21)
i[N i[X
66 The growth of firms
67
LP 1977–1987 0.45 20.13 0.37 0.31 1.00
Disney et al. (2003) UK manufacturing TFP 1980–92 0.05 0.15 0.26 0.54 1.00
LP 1980–92 0.48 0.04 20.01 0.49 1.00
Foster et al. (2006) US retail LP 1987–1997 0.16 0.24 20.39 0.98 1.00
Foster et al. (2008) US manufacturing TFP 1977–1997 0.36 20.17 0.50 0.30 1.00
Notes: columns (1) to (4) may not add to 1.00 because numbers in columns (1) to (4) are rounded off. In some cases, the establishment shares (qit)
are weighted by employment (Disney et al., 2003; Foster et al., 2006), whereas in others they are weighted by output (Foster et al., 1998, 2008).
Foster et al. (2008) estimates are pooled for the four five-year periods over the interval 1977–1997.
68 The growth of firms
t t+1
time
5.4 CONCLUSION
t t+1
time
Figure 5.3 A stylized depiction of the process of firm growth and R&D
investment, based on 1-lag VAR regression results. The
thickest line corresponds to the most significant association
(i.e. a coefficient of $0.3), while the other thinner lines
correspond to relatively ‘minor’ associations (coefficients of
magnitude $0.10). Coefficients lower than 0.05 in magnitude,
as well as autocorrelation coefficients, are not represented
76
Innovation and firm growth 77
“tried and succeeded”, “tried and failed”, and “not tried”’ (Freel, 2000, p.
208; see also Freel and Robson, 2004).
products if new goods are developed and are associated with the creation
of new jobs. As a result, the direct labour-saving effects of innovation need
to be weighed up against the indirect effects, and so the overall effect of
innovation on employment needs to be investigated empirically. Although
Van Reenen recently lamented the ‘dearth of microeconometric studies on
the effect of innovation on employment’ (Van Reenen, 1997, p. 256), the
situation has improved over the last decade.
Research into technological unemployment has been undertaken in dif-
ferent ways. As a consequence, the results emerging from different studies
are far from harmonious – ‘[e]mpirical work on the effect of innovations on
employment growth yields very mixed results’ (Niefert, 2005, p. 9). Doms
et al. (1995) analyse survey data on US manufacturing establishments, and
observe that the use of advanced manufacturing technology (which would
correspond to process innovation) has a positive effect on employment. At
the firm level of analysis, Hall (1987) observes that employment growth is
related positively and significantly to R&D intensity in the case of large
US manufacturing firms. Coad and Rao (2007) also observe a positive and
significant influence of innovation on employment growth in their analysis
of four high-tech US manufacturing industries. In addition, they observe
that when weights are given to firms according to their size, the coefficient
increases in magnitude – which suggests that larger firms, who have the
power to create or destroy a larger absolute number of jobs, are more likely
to convert innovative activity into employment gain. Greenhalgh et al.
(2001) observe that R&D intensity and also the number of patent publica-
tions have a positive effect on employment for British firms. Nevertheless,
Evangelista and Savona (2002, 2003) observe a negative overall effect of
innovation on employment in the Italian services sector. When the distinc-
tion is made between product and process innovation, the former is usually
linked to employment creation whereas the consequences of the latter are
not as clear-cut. Evidence presented in Brouwer et al. (1993) reveals a small
positive employment effect of product-related R&D although the com-
bined effect of innovation is imprecisely defined. Relatedly, work by Van
Reenen (1997) on listed UK manufacturing firms and Smolny (1998) for
West German manufacturing firms show a positive effect on employment
for product innovations. Smolny also finds a positive employment effect
of process innovations, whereas Van Reenen’s analysis yields insignificant
results. Harrison et al. (2005) consider the relationship between innovation
and employment growth in four European countries (France, Italy, the
UK and Germany) using data for 1998 and 2000 on firms in the manufac-
turing and services industries. Whilst product innovations are consistently
associated with employment growth, process innovation appears to have a
negative effect on employment, although the authors acknowledge that this
Innovation and firm growth 83
6.3 CONCLUSION
7.1 AGE
The relationship between size and growth has received a great deal of atten-
tion in empirical work, as we discussed above in section 4.2. Relatedly,
the relationship between a firm’s age and its growth rate has also been
frequently investigated. Age and size are certainly closely related, and in
some cases they are both taken to represent what is essentially the same
phenomenon (see for example the model in Greiner, 1972).
Older firms are often assumed to be more inert and less capable of
adapting to a changing environment. Older firms may lack the drive and
entrepreneurial spark that is required to observe new business opportu-
nities and then build upon them. The routinized nature of production
in firms may lead a firm to stick to what it knows best and become
increasingly distanced from external developments. Inertia can also
accumulate as people in the organization become entrenched in their
positions and resist change. Nonetheless, it is also conceivable that age
can be an advantage for small, young firms: since older firms are more
experienced, they have an established reputation and a proven track
record that confer more credibility in factor markets. This latter point
84
Other determinants of firm growth 85
Human Capital
The human capital embodied in the proprietor has long been suspected
of having an effect on firm growth. Education may be instrumental for
entrepreneurs to carry out their growth aspirations (Wiklund, 2007), or
indeed it may help entrepreneurs nurture such aspirations in the first place.
It may also make the founders more aware of their business environment
and the opportunities available to them, or it may put entrepreneurs in
a better standing vis-à-vis lenders (Robson and Obeng, 2008) or other
stakeholders.
Almus (2002) identifies a positive effect of human capital (that is, univer-
sity degree or above) on growth for fast-growing German firms. Robson
and Bennett (2000), however, fail to find a significant effect of skill level
in explaining employment or profitability growth in their sample of UK
small businesses.
Research on firm growth undertaken in developing countries has also
shown an interest in the human capital of the founding entrepreneur.
McPherson (1996) investigates small firms in five southern African nations
and observes that the level of human capital embodied in the proprietor
has a positive and significant influence on the growth of micro and small
businesses in five Southern African nations. Robson and Obeng (2008)
focus on small businesses in Ghana and report that better educated found-
ers faced fewer obstacles to expansion. They observe that education had a
stronger effect than age or sex on the importance of barriers facing growing
small businesses. Mead and Liedholm’s (1998) survey of the evidence
Other determinants of firm growth 89
Sex
The sex of the founding entrepreneur has also been linked to firm growth,
and the main finding seems to be that businesses headed by female entrepre-
neurs experience slower growth. This broad relationship has received much
attention in samples from developed countries (see Catley and Hamilton,
1998 for a survey) and has also been found among less developed countries
such as Indonesia (Singh et al., 2001), India (Coad and Tamvada, 2008),
and South Africa, Swaziland and Botswana (McPherson, 1996). Although
women entrepreneurs tend to be found in lower growth industries, the
lower growth of these entrepreneurs still remains after controlling for
industry (Mead and Liedholm, 1998).
Why is it that businesses led by female entrepreneurs tend to grow at
slower rates? One reason for this could be that females are less ambitious
than their male counterparts. Questionnaire evidence from Ghana suggests
that while women entrepreneurs encounter barriers preventing them from
starting their business, these ventures do not face higher barriers to growth
once they are already in business (Robson and Obeng, 2008). If established
female businesses do not encounter higher barriers to growth, then their
lower growth rates may well be due to less ambitious attitudes towards
growth. Another possible explanation is that females are more sensitive to
both domestic and professional responsibilities, and that since they rely on
their business for their household income, they will not seek to take risks
with their business and may be more willing to pass up business opportuni-
ties if these are perceived as too risky (Mead and Liedholm, 1998). There
may be still other differences between male and female entrepreneurs.
For example, while female entrepreneurs have strong skills in dealing
with people, their financial skills would appear to be weak in comparison
90 The growth of firms
(Catley and Hamilton, 1998; Hisrich and Ozturk, 1999). Female entrepre-
neurs seem to be less interested in making money (Cromie, 1987).
for small firms whose size is close to this threshold. This usually affects
firms whose size is somewhere in the range of 8–15 employees, depending
upon the country (see Schivardi and Torrini, 2008). In developing coun-
tries, firms can avoid or evade taxes by remaining small and informal.
Larger firms, on the other hand, can effectively lobby governments to
reduce their tax burden. As a result, the size distribution has a lot of weight
corresponding to small firms and large firms, and with a ‘missing middle’
which testifies to the disadvantages associated with a medium-sized scale of
operations (Tybout, 2000; Sleuwaegen and Goedhuys, 2002). In this case,
small firms will tend to allay their growth aspirations, while medium-sized
firms will have incentives to grow.
Still other determinants of firm growth can be mentioned here. Almus
(2004) observes that small German firms have lower growth rates when
there is ‘the shadow of death sneaking around the corner’ (Almus, 2004, p.
199). Employment growth rates are observed to be significantly lower up to
three years before a firm’s exit. There is also some evidence that uncertainty
may dampen a firm’s investment. Guiso and Parigi (1999) present convinc-
ing evidence that uncertainty of demand plays a significant role in reduc-
ing firm-level investment in the case of Italian manufacturing firms. Their
measure of demand uncertainty is constructed by referring to the subjective
probability distribution of future demand for a firm’s products according
to the firm’s leading managers. Relatedly, Lensink et al. (2005) use survey
data on Dutch SMEs to show that uncertainty has a mixed effect on invest-
ment. They observe that uncertainty increases the probability of investing
(in the context of a binary ‘invest or not’ model); it is seen to reduce the
overall amount of investment. Finally, Robson and Bennett (2000) show
that the use of external business advice is also associated with superior
growth, although the direction of causality is not easily ascertained. They
also present evidence that firms with an ‘established reputation’ experience
lower employment growth and higher turnover growth.
fairly constant across industries, there is more variation when one com-
pares the levels of growth via acquisition across industries.
Some investigators have expressed reservations about the standard
industrial classification scheme, because considerable heterogeneity can
be observed between firms in the same sector. If firms can be grouped
together with comparable firms, however, then it may well be easier to find
regularities and to generalize across firms within these groups. As a con-
sequence, other ways of grouping together similar firms have been sought.
Techniques such as Principal Components Analysis and Cluster Analysis
have been applied in attempts to group like firms together, often resulting
in new taxonomies of firms. In this spirit, Delmar et al. (2003) and Birley
and Westhead (1994) consider the case of small firms, while de Jong and
Marsili (2006), Leiponen and Drejer (2007) and Srholec and Verspagen
(2008) examine the case of innovating firms.
traded firms have experienced a rise in volatility over this period, this is
overwhelmed by declining volatility among privately held firms.
Regional effects have also been observed to have an influence on firm
growth. McPherson (1996) reports that Southern African small businesses
grow faster in urban areas than in rural areas. Similarly, Sleuwaegen
and Goedhuys (2002) observe that firms in the relatively industrialized
Abidjan region experience a faster growth than firms from rural areas of
Côte d’Ivoire, presumably because of the better availability of resources.
Reichstein and Dahl (2004) analyse the growth of Danish limited liability
firms and find that firms have higher expected growth rates when they are
located in a region that is increasing its specialization in the firm’s specific
industry. The aforementioned studies notwithstanding, however, Gabe
and Kraybill (2002) observe that both the county growth rate and a metro-
politan area dummy do not appear to have a statistically significant effect
on growth rates, for their sample of plants in Ohio.
Hart and Pearce (1986) analyse the growth patterns of very large firms
in several countries and observe that firms in Japan and Germany differed
from firms hailing from the USA and the UK, in that their Japanese and
German firms showed no tendency for Galtonian regression towards the
mean. In other words, while small firms grow faster than larger ones in the
US and UK, this was not observed in Japan or Germany. Subsequent work
found that small firms tended to grow faster than larger ones, for both US
data (Hall, 1987; Evans, 1987a,b; Amirkhalkhali and Mukhopadhyay, 1993;
Bottazzi and Secchi, 2003a) and also UK data (Kumar, 1985; Dunne and
Hughes, 1994). In other countries, Galtonian regression to the mean may
be less of an issue. For instance, regression to the mean is not observed in
the cases of Italian (Bottazzi et al., 2007) or French firms (Bottazzi et al.,
2008a). Bartelsman et al. (2005) contribute to this literature by exploring
differences in firm growth rates in a number of developed countries. They
observe that the post-entry growth of successful entrants is much higher in
the USA than in Europe. In particular, they observe that ‘[a]fter 7 years of
life, the average cohort of firms in manufacturing experience more than 60%
growth in employment, while in European countries the increase is in the
5–35% range’ (Bartelsman et al., 2005, p. 386). This is partly because new
firms tend to be relatively smaller upon entry in the US, thus having a larger
gap between their entry size and the industry minimum efficient scale (MES).
The authors suggest that this difference in post-entry growth rates is due to
institutional barriers to growth that are in place in Europe, such as the lack
of market-based financial systems, relatively high administrative costs that
may deter smaller firms at entry, and tighter hiring-and-firing restrictions.
Several other interesting results relating to cross-country differences in
firm growth rates can be found in the study by Beck et al. (2005b), which
96 The growth of firms
In the last few chapters we sought to look for the determinants of firm
growth. Chapter 4 presented the literature relating firm size to growth rate,
while Chapters 5 and Chapter 6 sought to clarify the relationship between
growth, on the one hand, and relative performance (profits and productiv-
ity) and innovation, on the other. In this chapter, we sought to find other
factors that might influence firm growth, such as firm age, inter-firm com-
petition, characteristics of the entrepreneur, and other firm-specific factors
(such as multi-plant structure or legal status) and also industry-specific and
macroeconomic factors. Although many variables are associated with firm
growth, to an extent that is statistically significant, it is not easy to predict
a firm’s future growth rate with much precision.
Without doubt, the main result that emerges from our survey of empiri-
cal work into firm growth is that the stochastic element is predominant.
Marsili (2001) summarizes in this way: ‘In short, the empirical evidence
suggests that although there are systematic factors at the firm and indus-
try levels that affect the process of firm growth, growth is mainly affected
by purely stochastic shocks’ (Marsili, 2001, p. 18). Geroski (2000) makes
an even bolder statement: ‘The most elementary “fact” about corporate
growth thrown up by econometric work on both large and small firms is
that firm size follows a random walk.’ (p. 169). The R2 statistic in growth
rate regressions is characteristically low, especially for databases containing
many small firms whose growth is particularly erratic. Including a long list
of explanatory variables and lags does little to help raise the R2 value, as is
evident from the survey provided in Table 7.1. Firm growth thus appears to
be remarkably idiosyncratic, even if the assumption of a purely stochastic
process of firm growth is often rejected on purely statistical grounds. Even
Table 7.1 A survey of R2 values obtained from regressions where the dependent variable is the growth rate of a firm or
plant
97
Geroski et al. (1997) 271 large quoted UK firms Market value, lagged firm growth, innovations, patents, 17–19%
1976–82 industry growth
Harhoff et al. (1998) About 10 000 West German Size, age, subsidiary, diversification, legal status, industry 8%
firms
Liu et al. (1999) Over 900 Taiwanese Age, size, industry dummies, capital–labour ratio, sales per 19–22%
manufacturing plants worker, dummies for R&D and exporting activity
Robson & Bennett Over 1000 SMEs in Britain in Size, age, exports, profits, industry, innovation and 4–8%
(2000) 1997 technology, use of external advice, strategy variables
Geroski & Gugler Large firms in 14 European Size, age, subsidiaries, diversification, growth of rivals 5–6%
(2004) countries, over 100 000 obs.
1994–98
Reichstein and Dahl 8739 Danish firms, 1994–96 Size, age, regional specialization, concentration, industry 1–2%
(2004)
Table 7.1 (continued)
98
technology, sample selection
Fagiolo and Luzzi 14 277 Italian firms 1995–2000 Size, age, cash flow, dummies for multi-plant firms, year and 2–3%
(2006) industry dummies
Coad (2007d) 8405 French manufacturing Gross operating margin, lagged growth, lagged size, industry 4–8%
firms, 1996–2004 and year dummies
Notes: Control variables include the constant term (though this is not mentioned above). Where fixed-effect regressions have been employed, we
refer to the overall R2 and not the within R2 or between R2. Where we have the choice, we prefer the adjusted R2 to the basic R2. Although growth
rates are mostly obtained by measuring size at annual intervals, this is not always the case. For example, McPherson (1996) takes the average
annual growth rate for the whole of the period since start-up, whereas Liu et al. (1999) take a yearly average of the growth rate over four years.
Other determinants of firm growth 99
if we were to compare firms of a similar size and age in the same industry,
with similar financial resources and similar patterns of innovation activity,
the growth rates of these firms would be largely idiosyncratic.
Research into firm growth has made progress on several fronts, however.
Gibrat’s model of proportionate growth shocks (described in section
4.1) serves as a useful benchmark model for describing firm growth. The
empirical literature has suggested a list of determinants of growth, and
they seem to go some way (although admittedly not very far) in explaining
differences in firm growth rates. Statistical analyses that focus on specific
hypotheses have provided nuances to our understanding of both the char-
acteristics of growth rates, and also the determinants and consequences
of growth. Unexpected results have also highlighted puzzles that deserve
more investigation in future.
With these insights in mind, in the following chapter we now turn to
compare our insights with theoretical work which, we hope, will be able to
piece these findings together in a coherent conceptual framework, as well
as providing explanations of the associations. Theoretical approaches to
firm growth can also play a role in suggesting which causal relationships
might be at work, as well as providing further testable hypotheses that can
be investigated in future work.
8. Theoretical perspectives
In the following we briefly present five distinct theoretical perspectives,
discussing their predictions for firm growth and judging them according
to the available empirical evidence. These five theories are the neoclassi-
cal theory (in particular, propositions based on the notion of an ‘optimal
size’), Penrose’s (1959) ‘theory of the growth of the firm’, the managerial
approach, evolutionary economics and its principle of ‘growth of the
fitter’, and also the population ecology approach.
100
Theoretical perspectives 101
compensation, bonuses, and other perquisites very often increase with firm
size.6 Furthermore, non-pecuniary incentives such as prestige, likelihood
of promotion, social status and power are also associated with firm size.
As a result, firm size (and firm growth) are seen to be important factors in
the ‘managerial utility function’, alongside the financial performance of
the firm. For some firms, such as small young firms, the pursuit of growth
maximization may coincide with that of profit maximization, so that a
manager has no conflict of interest between his duties to shareholders and
his own objectives (Mueller, 1969). In other cases, however, managers
have to choose between fulfilling their mandate of profit-maximization
(in service of shareholders) or pursuing their own interests of growth-
maximization. According to the managerial theory, utility-maximizing
managers are assumed to maximize the growth rate of the firm subject to
the constraint of earning a satisfactory profit rate, which should be large
enough to avoid being dismissed by shareholders or being taken over by
stock market ‘raiders’.
In the influential managerial model developed by Marris (1963, 1964),
firms are assumed to grow by diversification only. Marris posits a quad-
ratic relationship between profits and firm growth.7 Above a certain level
of growth, additional diversification has a lower expected profitability
because managers have less time and attention to devote to the operating
efficiency of existing activities and the development of new activities.
The managerial theory has also been extended to the case of growth
by conglomerate merger (Mueller, 1969). Mergers are a faster (and more
expensive) way of growth than internal growth – so managerial arguments
are a fortiori relevant for this type of growth.
Testing the ‘managerial hypothesis’ is a difficult task because the
theoretical models (for example Marris, 1964) propose a non-linear hump-
shaped relationship between growth rate and profit rate, with additional
growth having a negative effect on profits only beyond a certain ‘profit-
maximizing’ growth rate. Nonetheless, one basic prediction that emerges is
that the growth rates of manager-controlled firms will be higher than those
of owner-controlled firms, whilst profit rates are likely to be lower. Some
early studies thus tried to find performance differences between owner-
controlled and manager-controlled firms. The results, however, did not
offer unequivocal support in favour of the theoretical predictions. Radice
(1971) tests the hypothesis that owner-controlled firms have lower growth
rates and higher profit rates than management-controlled firms, using a
sample of 89 large UK firms over the period 1957–67. Perhaps surpris-
ingly, he observes that owner-controlled firms have both higher growth
rates and profit rates. Holl’s (1975) analysis also focuses on large UK firms,
but he fails to detect any significant difference in performance between
Theoretical perspectives 105
grow whilst less viable firms lose market share and exit.8 The notion of
selection via differential growth is also a central theme in the books by
Downie (1958) and Nelson and Winter (1982). Downie (1958) models
industrial development by assuming that firms grow by reinvesting their
earnings. Growth rates thus rise with profitability. Nelson and Winter’s
(1982) influential book contains a formal micro-founded simulation model
in which firms compete against each other in a turbulent market environ-
ment. In this model, firms can gain competitive advantage through either
the discovery of cost-reducing innovations or by imitating the industry
best practice. Firms that are more profitable are assumed to grow, whilst
firms that are less successful are assumed to lose market share. Agent-based
simulation modelling has since remained a dominant tool in the evolution-
ary literature (see, among others, Chiaromonte and Dosi, 1993; Dosi et
al., 1995; Marsili, 2001; and Dosi et al., 2006; see also Kwasnicki, 2003;
and Dawid, 2006 for surveys). In addition to computer simulation models,
the principle of ‘growth of the fitter’ has also formed the foundations of
analytical evolutionary models (see, for example, Winter, 1964, 1971; and
Metcalfe, 1993, 1994, 1998;9 see also Jovanovic, 1982; Hopenhayn, 1992;
and Melitz, 2003 for more elaborate mathematical models).
The evolution of firms and industries, as depicted in this family of
models, is guided by the mechanism of ‘replicator dynamics’, by which
growth is imputed according to some broad measure of ‘fitness’ (or ‘viabil-
ity’). This mechanism can be presented formally by Fisher’s ‘fundamental
equation’, which states that:
–
xi 5 xi (Fi 2 F ) (8.1)
where d stands for the variation in the infinitesimal interval (t, t 1 dt), and
xi represents the market share of firm i in a population of competing firms.
Fi is the level of ‘fitness’ of the considered firm, where fitness corresponds
to relative financial performance or perhaps some measure of relative pro-
– –
S
ductivity.10 F is the average fitness in the population, i.e. F 5 ixiFi, and
a is a parameter. It is straightforward to see that this equation favours the
above-average firms with increasing market share, whilst reducing that of
‘weaker’, less profitable firms.
This ‘replicator dynamics’ does sound intuitively appealing, because
implicit in it is the idea that selective pressures act with accuracy, that
financial constraints prevent inefficient firms from growing, and that the
economic system adapts so as to allocate resources efficiently amongst
firms, such that firms ‘get what they deserve’. However, these assumptions
may not find empirical validation for a number of reasons. First of all, it
cannot be assumed that all firms have the same propensity to grow. Some
Theoretical perspectives 107
This should not be taken to mean that the scholars deny the existence of
differences between organizations.11 Instead, this is due to the fact that
the fundamental unit of analysis here is the population of organizations
within a niche, rather than the individual organizations that make up
the population. As a consequence, population ecologists tend to explain
the performance of organizations by referring to features common to all
organizations within the same niche, rather than firm-specific factors.12 Of
course, there are clear limits to a theory of firm growth rates based solely
on industry-wide characteristics, because large differences in growth rates
can be observed between firms in the same industries. Notwithstanding
the analytical starting point, however, some work in this stream of litera-
ture relates the performance of organizations to idiosyncratic rather than
environmental factors.
Broadly speaking, the empirical strategy in the ‘population ecology’
literature takes place by gathering life-history data on populations of
organizations that are arguably in the same ‘niche’. This niche may refer
to specific industries (for example automobile producers, Hannan et al.,
1995), niches within industries (such as biotechnology drug discovery
companies, Sorensen and Stuart, 2000), or even non-commercial ideologi-
cal organizations (Minkoff, 1999; Simons and Ingram, 2004). Most studies
focus on the effects of characteristics of organizations,13 populations, and
the environment on organizational performance by examining birth and
death rates of organizations. However, efforts have been made to explain
differences in growth rates between firms in the same industry. Barron et al.
(1994) analyse data on New York Credit Unions over the period 1914–90
and observe that larger firms have lower expected growth rates than their
smaller counterparts. The interpretation they offer is that larger organiza-
tions have become less efficient and less well adapted to the current busi-
ness environment, thus being more vulnerable to young competitors. This
builds upon a key population ecology tenet that firms are fundamentally
inert (Hannan and Freeman, 1984), being both averse to and relatively
incapable of strategic or organizational change.
8.6 DISCUSSION
resources, and managers divert time and effort into drawing up their
growth projects, taking into account the growth opportunities that are
available. The managerial perspective on firm growth was presented in
section 8.3, according to which managers take pleasure in being in charge
of a large organization and seek to grow as much as possible, beyond the
profit maximizing level. In section 8.4 we presented the evolutionary princi-
ple of ‘growth of the fitter’, which seems quite plausible at first sight but, so
far, it has not performed well in empirical investigations. Finally, in section
8.5 we presented the population ecology perspective on firm growth.
Although some theories are more useful than others, none of the theories
can provide a comprehensive explanation of firm growth. Firm growth is
indeed a multifaceted phenomenon, it has a strong idiosyncratic character,
and as a result it is difficult to generalize across firms and circumstances.
9. Growth strategies
After starting the book by reviewing the empirical evidence on firm growth,
the previous chapter contrasted the evidence with some broad-based
theoretical predictions on firm growth. These theories did not fare excep-
tionally well in accounting for the variation in growth rates across firms,
presumably because it is not easy to generalize across heterogeneous firms
facing dissimilar circumstances.
In this chapter (and also the next) we take an approach that can perhaps
be described as ‘appreciative theorizing’ (Nelson and Winter, 1982, pp.
45–8); we will seek to provide tailored descriptions of certain aspects of the
firm growth process, without trying to unify these aspects together under
one centralized theoretical monolith.
We begin by discussing the attitudes of firms towards growth (section
9.1), and then the available means of achieving growth (such as diversifica-
tion, acquisition and internationalization – see sections 9.2–9.4). It appears
useful to relate these two topics to the distinction between ‘demand’ for
growth and ‘supply’ of growth opportunities, respectively. Firm growth
requires both a willing attitude to take up growth opportunities, and also
the availability of suitable opportunities. However, in the long run, the dis-
tinction between supply and demand determinants of growth may become
blurred (Penrose, 1960). Entrepreneurs and managers with a strong desire
to grow will surely find suitable growth opportunities if they search for
them. Correspondingly, one could suppose that even firms with a marked
aversion to growth will eventually take up additional growth opportunities
if these are attractive enough.
111
112 The growth of firms
expound why firms may or may not want to grow, as well as discussing the
intentionality of growth.
Advantages of growth
Growth of an organization can be seen as a means of alleviating tensions in
its internal management. Employees appreciate the opportunities for pro-
motion as well as the higher salaries and prestige that accompany growth.
Aoki (1990) writes that employees may even be willing to forgo current
earnings in exchange for future benefits made possible by promotion in an
expanding hierarchy. In addition, work is likely to become more challeng-
ing as the firm ‘breaks from its routines’ and expands into new business
areas. ‘Work is more fun in a growing company’ as Roberts (2004, p. 243)
bluntly puts it. Bronars and Deere (1993) point out that growing firms
are better able to maintain worker morale, and as a result they are less
susceptible to unionization activity on the part of the employees.
Conversely, a lack of growth can create an uninspiring and stultifying
business environment which depresses managerial efficiency (Hay and
Morris, 1979). As a result, in growing firms it is ‘easier to obtain com-
mitment to organizational goals and priorities from various factions and
to resolve conflicts between those factions’ (Whetten, 1987, p. 340). An
organization may thus seek a positive growth rate in order to keep its
members satisfied. Indeed, it has been conjectured that firms that take
their employees’ interests seriously are likely to have higher growth rates
(Aoki, 1990).
The managerial vision of the firm can be considered as an extension of
this line of reasoning. Managers attach positive utility to the growth rate
of the firm, because an increase in firm size is associated with increases in
compensation, power, prestige, bonuses and perquisites. One difference,
however, is that managers have the power to determine a firm’s growth
strategy themselves, and so they can pursue a growth rate above that which
would be optimal for the shareholders. For more on the managerialist
theory of the firm, see section 8.3.
Firms may also seek growth as a means of attaining other objectives
related to its production of goods and services. Lower production costs
may be achieved if expansion leads to economies of scale (due to a larger
scale of production), or economies of scope (because of a wider range of
products or services). Growth may also take place if firms wish to expand
their productive capacity or boost their output so as to deter entry from
potential competitors (Dixit, 1980).1 Furthermore, a larger, more diversi-
fied firm is better able to spread its risk among its various activities. (This
Growth strategies 113
will be an advantage for managers whose fortunes are tied to those of the
firm (Amihud and Lev, 1981), although it is not necessarily an advantage
for shareholders because they can reduce their risk by investing in a diversi-
fied portfolio including other firms.) In this way, growth can be considered
to be a basis for security (Whetten, 1987).
Other reasons have also been advanced to suggest why firms might
want to grow. One reason might be because growth is sometimes a more
suitable metric of performance than profits – this is particularly true for
high-volatility markets. A firm’s management may thus set its performance
goals in terms of percentage increases in sales rather than profit margins or
share prices. Other firms may grow for want of a better alternative. This
might be the case for firms who grow by reinvesting profits in the company,
as a means of avoiding heavy taxes (on dividends, for example).
There is some empirical evidence that demonstrates the positive effect
of growth on firm performance. Coad (2007d) analyses a large sample of
French manufacturing firms and observes that growth is associated with
short-lived increases in profit rates, whether growth is measured in terms
of employment, sales, or value-added. Perhaps surprisingly, there seems to
be a larger effect of growth on profits than that of profits on growth. This
finding of a beneficial and temporary influence of firm growth on profit
rates is consistent with the Kaldor–Verdoorn ‘dynamic increasing returns’,
Penrose’s (1959) theory of ‘economies of growth’, and Starbuck’s (1971)
‘will-o’-the-wisp’ models of firm growth.
Disadvantages of growth
Despite the aforementioned advantages linked to growth, some managers
or owner-managers may be wary of increasing the size of their firm. One
major reason for this is what we could call the ‘control-loss’ argument.
Loss of control may originate from the increased size or the rate of growth.
As a firm increases in size, as employees are added and the number of hier-
archical levels increases, the manager has less control of the firm and is less
well informed of its current state (Williamson, 1967). Problems of control
and coordination are also increasing functions of the growth rate. Whilst it
has been advanced that problems of coordination vanish under truly static
conditions (Kaldor, 1934), fast-growth firms may experience difficulties in
coordinating operations in a complex and changing environment.2
Family-owned and traditional firms may have an especially cautious
approach to growth if they are keen to keep the firm under tight control or
if they are reluctant to integrate a large number of employees and managers
from outside the family. Furthermore, they may be particularly risk-averse
because failure of the enterprise may take on connotations of ruining
the family tradition. Managers whose training and experience have been
114 The growth of firms
scientists that founded the firms are often not well prepared for the man-
agement roles that larger firms require. In the printing sector, many firms
choose not to grow simply because the owners use their business as a means
to support a relaxed and independent lifestyle. More generally, Greiner
(1998) provides the following description of the ‘lifestyler’ manager’s
attitude to growth:
When C complains in turn of being overworked (as he certainly will) A will, with
the concurrence of C, advise the appointment of two assistants to help C. But he
can then avert internal friction only by advising the appointment of two more
assistants to help D, whose position is much the same. With this recruitment of
E, F, G and H the promotion of A is now practically certain.
Seven officials are now doing what one did before. . . . For these seven make so
much work for each other that all are fully occupied and A is actually working
harder than ever. (Parkinson, 1957, p. 5)
markets. There are thus a number of issues and complications that accom-
pany a firm’s decision to grow. These issues are discussed in the following
sections.
Theoretical perspectives
An early view of diversification considered that managerial competences
were the key to superior firm performance, irrespective of the sector of
activity. In other words, this perspective holds that ‘management is an
amorphous substance which can be applied with equal success to totally
unrelated lines of business’ (Mueller, 1969, p. 651). In order to take full
advantage of these scarce assets, successful firms sought to spread their
superior management capabilities across several different industries. In
this way, diversification was guided by a logic of synergies of managerial
competence as opposed to synergies of a technological nature. As a result,
the large diversified conglomerate became a popular organizational form,
especially in the 1950s and 1960s.
Penrose’s (1959) vision of firm growth by diversification can be placed
within this context. Managerial attention is seen to be the main factor limit-
ing firm growth. As a firm continues its operations, incumbent managers
gradually gain experience, and new managers can be trained and integrated
120 The growth of firms
into the firm, thus expanding the firm’s resource base. In this way, mana-
gerial resources are continually being freed up over time. Growth thus
constitutes a responsible use for excess managerial attention – it challenges
managers to focus their attention on generating profits in new activities.
However, Penrose also gives clear recommendations as to the direction of
diversification. A key element of Penrose’s theory of firm growth is that
firms are composed of indivisible resources, which are specialized and spe-
cific to the firm. A firm’s diversification strategy should therefore focus on
how best to exploit the idiosyncracies of the firm’s current resource base.
In other words, growth by diversification is most effective when the new
activities are related to the existing resource base.
The notion of related or ‘synergistic’ diversification is central to Igor
Ansoff’s (1987) celebrated book. Ansoff advocated a prudent approach
for diversification at a time when, in retrospect, it appears that general
management synergies were overestimated. According to him, firms should
only consider diversification when they have no other option in order to
realize their growth objectives – ‘if a firm can meet all of its objectives by
measures short of diversification or internationalization, it should do so’
(Ansoff, 1987, p. 131). Indeed, in many cases a firm can discover growth
opportunities by re-evaluating and re-formulating its strategies within its
present portfolio of activities, instead of expanding the portfolio by com-
mencing new activities. Firms that choose to diversify, however, can do this
in one of three ways: by exporting the firm’s traditional products or services
into new markets, which constitues the ‘highest synergy move’ (Ansoff,
1987, p. 125), or by diversifying according to synergies of demand or syner-
gies of technology. In each case, attention must be paid to the coherence of
the diversified firm’s portfolio of activities. Candidate new businesses must
display synergies with the existing portfolio of activities along dimensions
such as operations, R&D, or marketing and distribution. These synergies
may be due to lower expected fixed costs of starting up, or alternatively
due to anticipated operating economies. Furthermore, efforts should be
made to convert the ex ante ‘potential synergy’ into ‘realized synergy’, by
actively seeking to integrate the new activity alongside the firm’s existing
activities. If these guidelines are successfully applied, synergistic diversifi-
cation allows firms to earn superior profits by leveraging their capabilities,
know-how and general experience in new markets. It should be pointed out
that synergistic diversification is not incompatible with corporate refocus-
ing, but is instead closely related (Batsch, 2003). Both of these view the
firm as a coherent portfolio of related activities based on a small number of
core competences. Refocusing can be seen as a corrective strategic measure
undertaken after excessive unrelated diversification – it is a modification
(but not necessarily a reduction) in a firm’s activities as the firm seeks to
Growth strategies 121
prefer to spend it on pet projects even if these have a low expected return
(Jensen, 1986).
Empirical evidence
A large body of research in the financial economics literature has focused
on the relative performance of diversified firms vis-à-vis stand-alone firms
or less-diversified firms, generally using data on large US firms. The general
message that emerges from this literature is that diversification is associ-
ated with inferior performance, although more recent work has challenged
this view, and some scholars have even suggested that diversification may
even be associated with superior performance (see the survey in Martin
and Sayrak, 2003).
A historical perspective provides a helpful context in which research
into diversification can be framed. In the 1950s and especially the 1960s,
diversification was actually a popular strategy, for several reasons. First,
and perhaps most important, antitrust law imposed limits on the market
shares of firms in specific industries, which meant that firms who were
willing to grow had to do so in new industries. Second, capital markets were
relatively undeveloped and firms had incentives to organize several busi-
nesses around an ‘internal capital market’ – this is also known as the ‘deep
pockets’ argument. Third, the multidivisional or ‘M-form’ organization
was growing in popularity. Fourth, there is evidence that early diversifica-
tion announcements actually received a positive stock market reaction,
which encouraged further diversification. As a result, the 1960s have been
described as a ‘wave of unrelated acquisitions’ (Montgomery, 1994, p. 170).
Unrelated diversification by acquisition became increasingly common
during the 1960s (Shleifer and Vishny, 1990), whereas in previous times
firms had been more wary of entering markets where their learned capabili-
ties did not give them a distinct competitive advantage (Chandler, 1992).
The 1970s were also characterized by unrelated acquisitions and overdi-
versification. The 1980s, however, have been associated with a ‘return to
corporate specialization’ (Bhagat et al., 1990). During this time, changes in
the business environment made diversification less appealing (in particular,
financial markets became more developed, and antitrust law changed its
stance on measures of absolute market share). Furthermore, increasing
attention was being drawn to the relatively poor financial performance
of large diversified conglomerates. As a result, the takeover wave of the
1980s can be characterized by firms refocusing on their core capabilities by
acquiring businesses in related activities. During this time, it was common
for ‘corporate raiders’ to acquire a large diversified conglomerate and to
sell on the constituent segments individually, often at a considerable profit
(Shleifer and Vishny, 1990).
Growth strategies 123
in Oviatt and McDougall, 1994; and also McDougall et al., 1994; Autio et
al., 2000 and Sapienza et al., 2006; see also Zahra and George, 2002 and
Jones and Coviello, 2005 for surveys of international entrepreneurship).
Although it is recognized that the accumulation of knowledge and experi-
ence is an incremental process, it has been argued that small firms are more
flexible and are able to learn rapidly, so they can expand in foreign markets
at a faster pace than older firms. SMEs may have established home-based
advantages (Kuemmerle, 2002) that they leverage in new markets. In this
view, SMEs can expand abroad by applying existing capabilities to profit
opportunities in new foreign markets. However, they may also treat their
expansion projects as learning opportunities, providing them with valu-
able knowledge and allowing them to augment their existing capabilities
(Kuemmerle, 2002; Zahra et al., 2000). Indeed, internationalization is a sig-
nificant event that has the power to change the firm itself. While older firms
tend to be more inert and are prone to being ‘locked in’ to the exploitation
of established capabilities, international SMEs may be more flexible and
also more receptive to knowledge gained from internationalization, using
their knowledge to develop their core capabilities further.
9.5 CONCLUSION
129
130 The growth of firms
Phase
1 2 3 4 5
large
collaboration
coordination
“?”
Size of Organization
delegation
red tape
direction
control
creativity
autonomy
structure. They benefit from efficient information flow, relatively quick deci-
sion-making and proximity to their customer base (You, 1995). As a result,
small firms may have an advantage in serving niche markets for specialized
products, whereas larger firms may be better adapted to catering for larger,
standardized markets. Small firms also tend to be less capital-intensive
(Caves, 1998; Bellone et al., 2008) or, similarly, more labour-intensive (You,
1995) than their larger counterparts. These factors notwithstanding, smaller
firms seem to be associated with lower productivity than larger firms (Idson
and Oi, 1999), especially when one considers small firms in developing coun-
tries (Little, 1987). One reason for this could be because small, young firms
tend to charge lower prices for the same goods (Foster et al., 2008).
In developed countries, entrepreneurial small firms have been ascribed
an important role in introducing new products and new techniques into
the market, through technological innovations (Pavitt et al., 1987; Acs
and Audretsch, 1990). Pavitt et al. (1987) observe that small firms in the
UK account for a share of major innovations that is disproportionately
large when compared to their R&D expenditure levels. Acs and Audretsch
(1990) analyse data on US firms and emphasize the innovative prowess of
small businesses. More specifically, small firms seem to play an especially
important role in highly innovative and skill-intensive industries which are
in early stages of their life cycles (You, 1995).
Even in developed countries, however, many entrepreneurs are not true
entrepreneurs, in the sense that they do not bring innovations or bring
about reform in stagnant markets (Santarelli and Vivarelli, 2007). Many
enter for less noble reasons, such as over-optimism on the part of the
founder, the pursuit of a relaxed lifestyle, or the flight from unemploy-
ment. In fact, many entrants are far less productive than incumbents (even
taking into account their liability of small scale). The concept of new firm
entry gathers together a particularly heterogeneous group of enterprises. In
many cases, and especially in developing countries, micro and small firms
exist because they offer individuals a livelihood and a source of independ-
ent revenue. In many cases, new small businesses are founded as a last
resort rather than as a first choice (Beck et al., 2005a). It has even been
argued that, in the case of India, better educated individuals are quick to
leave self-employment and pursue alternative career paths, whereas less
educated individuals have fewer opportunities to leave self-employment
(Nafziger and Terrell, 1996). Workers in such enterprises are relatively
badly paid, or may even be unpaid family members.
According to Penrose (1959), small firms can thrive in the ‘interstices’ of
major markets, in niche sub-markets that are not large enough to support
large firms. As a result, they are often sheltered from direct competition
with large firms. This is not to say that they are entirely protected from the
132 The growth of firms
It is fitting to begin this section on the growth of small and large firms with
a well-known passage written by Alfred Marshall:
[W]e may read a lesson from the young trees of the forest as they struggle
upwards through the benumbing shade of their older rivals. Many succumb on
Growth of small and large firms 133
the way, and a few only survive; those few become stronger with every year, they
get a larger share of light and air with every increase of their height, and at last
in their turn they tower above their neighbours, and seem as though they would
grow on for ever, and for ever become stronger as they grow. But they do not.
One tree will last longer in full vigour and attain a greater size than another;
but sooner or later age tells on them all. Though the taller ones have a better
access to light and air than their rivals, they gradually lose vitality; and one after
another they give place to others, which, though of less material strength, have
on their side the vigour of youth. (Marshall, 1961, p. 263; first edition published
1890)
Marshall is quite right in pointing out the perils that afflict small young
firms. The growth of small firms indeed involves a struggle to obtain
vital resources (the light and air in Marshall’s analogy), and many firms
will not survive. One direction in which Marshall’s analogy might need
further clarification, however, concerns the fact that not all firms strug-
gle for growth. In fact, many small firms don’t seem to want to grow. In
the remainder of this section, we discuss three key aspects of the growth
of small firms: the struggle to survive, desire to grow, and also structural
change in growing organizations.
rates hovered at about 50 per cent after around four years, in their sample
of small US firms. Bartelsman et al. (2005) examine the post-entry perform-
ance of new firms in seven OECD countries and observe that about 20–40
per cent of entering firms fail within the first two years, while only about
40–50 per cent survive beyond the seventh year.
Although survival rates are lower than one might have hoped for, not all
cases of firm exit should be considered as failures. Headd (2003) presents
evidence that up to a third of business exits are considered by owner-
managers to be successful events, corresponding to cases such as planned
closures, sale of the business, or retirement from the work force. As such,
raw figures on survival rates may actually be understating the success rate
of new businesses, because they do not take into account these cases of
successful closures. Headd (2003) analyses data on new employer firms
in the US and shows that while half of these new firms can be expected to
survive after four years, as many as one-third of cases correspond to suc-
cessful closures, with the remaining two-thirds being failures. The failure
rate remains quite high, however, even when successful closures are taken
into consideration.
New firms enter on a small scale relative to that of incumbents – around
40–60 per cent of the average size of incumbents (Bartelsman et al., 2005).
Their small size puts them at a disadvantage vis-à-vis their larger coun-
terparts, and so they must expand rapidly, as if their life depended on it.
Wiklund (2007) explains that, for new small firms, ‘growth and survival
go hand in hand’ (p. 145). Garnsey et al. (2006) acknowledge that growth
creates problems, but they add that the problems accompanying growth
are less dangerous to a firm’s survival than the absence of growth. The
larger they grow, the smaller their cost disadvantage relative to firms above
the MES, and thus the higher their chances of survival. For such firms, the
growth objective coincides with survival and the pursuit of profits. These
firms tend to have a higher average growth rate than larger firms, despite
the difficulties they may face in financing their expansion.
Some influential theoretical models have attempted to describe the
chaotic process of small firms growing larger. Jovanovic (1982) presents
what is known as the ‘passive learning’ model, in which small firms have a
fixed, firm-specific productivity level. Their growth and survival prospects
are bound to this productivity variable. Although firms do not know how
productive they are upon entry, they learn about their relative productivities
once they have entered. It is shown that this model is able to account for the
faster growth and also the higher exit hazards associated with small firms.
Hopenhayn (1992) presents a similar model in which a firm’s productivity
level evolves in random fashion, according to a Markov process. Finally,
the ‘active learning’ model (Ericson and Pakes, 1995; see also Pakes and
Growth of small and large firms 135
Although research into small firms has placed great emphasis on fast-
growth firms,5 the majority of small firms do not experience fast growth.
136 The growth of firms
Small firms have many excuses for staying at a small size. Lack of finance
could be one reason, problems finding customers could be another, or
problems finding honest and hardworking employees could be yet another.
In many cases, however, small firms don’t grow because they just don’t
have growth ambitions.
For many owner-managers of small businesses, the main objective is
independence. These individuals simply view their enterprise as instrumen-
tal in guaranteeing an independent and relaxed lifestyle where they can ‘be
their own boss’. Owner-managers may be more interested in developing
a personal friendship with existing customers than looking for new profit
opportunities. They may also be wary of relinquishing control of the firm
and delegating tasks to new employees. If they feel that growth would com-
promise their independence or the enjoyment of their work, then they will
not be motivated to grow (Wiklund et al., 2003). Once these firms survive
infancy, therefore, they may experience no further growth; and with the
passing of time, these firms may become increasingly inert and locked in
to their present scale of operations, letting growth opportunities fly away
from under their noses. These firms have been dubbed as ‘lifestyler’ firms
by some (for example Hay and Kamshad, 1994), while others have referred
to these small firms as the ‘living dead’ (O’Farrell and Hitchens, 1988, p.
1372).
While some firms aim for high growth, others lack the ambition. Growth
aspirations, one might suppose, play a major role in separating the busi-
nesses that grow from those that don’t. A number of researchers have
used questionnaire evidence to investigate the association between growth
ambitions and realized growth, finding the expected positive association
(Miner et al., 1994; Wiklund and Shepherd, 2003; Delmar and Wiklund,
2008). Wiklund and Shepherd (2003) show that while growth ambitions are
positively associated with growth, these positive effects are magnified when
interacted with factors such as the entrepreneur’s education and business
experience. Delmar and Wiklund (2008) observe that growth motivation
is positively associated with subsequent growth, and also, interestingly
enough, that realized growth feeds back to have a positive impact on subse-
quent growth motivation. As such, growth may be an acquired taste, such
that past growth tends to increase the desire for further growth.
Chapin (1957) stands out in the literature as one of the most bizarre and
esoteric attempts at finding optimality in the growth of organizations.
Chapin analyses data on the two largest sub-groups (regular membership
and Sunday-school enrolment) in 80 Minneapolis churches. Rudimentary
Growth of small and large firms 137
statistical analysis cannot reject the hypothesis that the ratio of these two
sub-groups tends to the Fibonacci proportion of 0.6180. Chapin refers
to principles of harmony and symmetry of structure, which have been
observed in the growth of snail shells and sunflower seeds, to postulate
that organizations grow by a proportional scaling up of existing depart-
ments. More specifically, he asserts that it is the Fibonacci proportion that
governs the optimal logarithmic growth spiral of the relative share of the
two largest sub-groups within organizations.
Whatever the reason, Chapin’s model did not have a major impact upon
scholars of firm growth. One major shortcoming of his model is that firms
do not stay in the same proportions as they grow, but instead they undergo
tremendous restructuring. Small firms and large firms are very different in
structure, and they should not be considered as merely scaled down ver-
sions of larger firms. Small firms are not just scaled down versions of larger
firms. As small firms grow and become larger, their growth is accompanied
by considerable organizational stresses, which leads them to undergo
substantial transformations. Hannan and Freeman (1977) provide an ani-
mated analogy of the structural change that accompanies the growth of
organizations. They write that ‘a mouse could not possibly maintain the
same proportion of body weight to skeletal structure while growing as big
as a house. It would look neither like a mouse nor operate physiologically
like a mouse’ (Hannan and Freeman, 1977, p. 938).
An alternative model of organizational growth, briefly sketched out in
Andriani and McKelvey (2007), considers firm size in terms of the concepts
of surface and volume.
Employees dealing with people outside the firm are surface employees – they
bring in the resources from the environment. Volume employees are those inside
who produce and coordinate: they are resource users. As firms grow, then, they
have to maintain the square-cube ratio by adding more surface units, or by
making them more efficient. (Andriani and McKelvey, 2007, p. 1219)
As we have seen from the previous section, small and large firms grow for
different reasons and in different ways. Indeed, it has been observed that
138 The growth of firms
expands further. If the firm does not grow, it remains at what they call
the ‘success-disengagement’ stage. If the firm decides to grow, however,
it experiences a ‘take-off’ and must deal with issues of decentralization
and delegation before reaching the ultimate stage, ‘resource maturity’.
Churchill and Lewis (1983) also emphasize a fundamental transforma-
tion that takes place in growing firms – the fact that although the owner’s
abilities are important at the start of the enterprise, they become less so as
the firm becomes mature. Conversely, delegation is not important in small
firms but it becomes increasingly important as the firm grows. It follows
that the ‘inability of many founders to let go of doing and begin managing
and delegating’ (p. 42) is a major obstacle to the development and growth
of small firms.
The model developed by Garnsey (1998) bears some similarities to
that of Churchill and Lewis (1983), although it focuses more on the early
growth and development of new firms. Garnsey places emphasis on the
high hazard rates that confront new firms, and their effort and struggle
to quickly access, mobilize and deploy resources before they can generate
resources for growth. Once a firm’s operations are set up, however, the
initial burst of energy required to get things going is no longer required,
and resources are released for growth. Garnsey (1998) also discusses the
phenomenon of routinization of operations in small growing firms. To
begin with, ‘[n]ew firms are hampered by their need to make search proc-
esses a prelude to every new problem they encounter’ (p. 541). As time
goes by, however, firms learn about their business and develop problem-
solving repertoires that make demanding situations appear more routine.
Problems can be identified as recurrent and require less time and energy,
and ‘early challenges are replaced by repetitive grind’ (p. 542). As a con-
sequence, this routinization found in growing small firms can engender
disillusionment, and growth can be hindered by morale problems (which
may even lead to spin-outs of new ventures).
Garnsey and co-authors argue in favour of a process theory of growth
rather than a theory in discrete stages, however, because although certain
developmental patterns are common in new growing firms, firms face dif-
ferent challenges and deal with them in different ways (Hugo and Garnsey,
2005; Stam and Garnsey, 2005). Firms may struggle with the same recur-
ring problems, or they may face several issues at the same time. As such,
the growth of small firms is better described in terms of processes than
successive stages (Garnsey, 1998; Druilhe and Garnsey, 2006).
Although the ‘stages of growth’ models have largely escaped empirical
attention, it is worthwhile to mention here the work by Kazanjian and
Drazin (1989).8 The essence of their test is to observe how small new firms
evolve through four discrete growth stages – conception and development;
Growth of small and large firms 141
10.4 CONCLUSION
One often gets the impression from the popular press that nothing is as
exciting as the rapid growth of successful small firms. Small growing firms
do not stay small for long, however. This chapter focused on the differences
between small and large firms, as well as the organizational transforma-
tions that accompany the growth as small firms become larger. We began
in section 10.1 by highlighting some of the differences between small and
large firms, before moving on to a discussion of how small and large firms
differ both in their attitudes to growth and also the characteristics of their
growth (section 10.2). Small firms must struggle to survive, and growth can
reduce their chances of exit. Many small firms are daunted by the prospects
of growth, and prefer to stay at a small size. Once firms start to grow,
however, their growth may awaken in them the desire for further growth.
In section 10.3 we presented the ‘stages of growth’ models. According to
these models, small entrepreneurial firms must deal with a number of dif-
ficulties (such as issues of delegation, monitoring and coordination) as they
move from one stage to the next, eventually becoming large-scale bureau-
cratic businesses. An advantage of these models is that they highlight the
fundamental transformations that accompany growth. Indeed, small firms
cannot be seen as simply ‘scaled-down’ versions of larger firms, as the
growth stage models vividly illustrate. The growth stage models also have a
number of drawbacks, however. Firms do not automatically progress from
one stage to another, but may face recurring problems, or several obstacles
at the same time. Empirical investigations into growth stage models have
drawn attention to the limited practical relevance of these models, if they
are taken too literally. The stage of growth models should therefore be
taken with a grain of salt.
11. Conclusion
What have we learned about firm growth? To conclude this book, we begin
by reviewing the main themes encountered (section 11.1) and make some
final comments concerning theoretical (section 11.2) and empirical (section
11.3) work. Section 11.4 concludes with a synthetic discussion about the
nature of firm growth, guided by the main findings of the book.
143
144 The growth of firms
It may be that the majority of the total variation in firm growth rates is
within firms over time (Geroski and Gugler, 2004; see however Davis et al.,
2006). As a consequence, it would make sense for future empirical work
to attempt to explain growth by referring to variables that vary more over
time within particular firms than they vary between firms (in the cross-
section) at any given time. Unfortunately, however, firm-specific variables
that display such properties are not easy to think of, and data on such
variables is even harder to obtain.
Our survey has also emphasized a number of surprising and perhaps
counterintuitive findings. For example, we saw in Chapter 5 that finan-
cial performance and productivity are poor predictors of growth. The
evolutionary mechanism of selection by differential growth does not
seem to work very effectively at all. Instead, selection appears to operate
mainly via the channel of exit – that is ‘survival of the fitter’ rather than
‘growth of the fitter’ – and this considerably reduces the power of selective
forces. Although there are strong implications hinging on the relationship
between relative ‘fitness’ (usually profits or productivity) and growth, there
is nonetheless a shortage of empirical research that has been conducted in
this domain. As a result, I feel obliged to reiterate Caves’ (1998) recom-
mendation: ‘Because reallocations of activity from the less efficient to
the more efficient are so important for the optimal use of resources, more
evidence is needed on how competitive conditions within an industry affect
the speed with which the more efficient displace the less efficient.’ (Caves,
1998, p. 1977). The nature of the relationship between relative performance
(whether it be financial performance or productivity) and firm growth is
indeed bewildering in terms of the conspicuous gap between theoretical
work and empirical findings. This is a puzzle that remains to be tackled.
We suggest that cohort studies, where cohorts of firms having the same age
are tracked over time, might be a useful approach because a firm’s attitude
to growth is assumed to vary over the life cycle.
Firms are surprisingly heterogeneous with regards to their propensities
for growth, and there is some evidence to suggest that, in several cases,
poorly performing firms are often more ambitious in their growth plans
than better firms. Business firms are certainly not homogeneous, rational
profit maximizers. What are the implications of this? Should inefficient
firms be discouraged, and efficient firms be encouraged to grow? How could
such a policy be operationalized? More work on this would be welcome.
Another puzzle in the literature concerns the link between innovation
and firm growth. While much theoretical work, as well as questionnaire
evidence from managers, stresses the crucial role of innovation in explain-
ing growth, empirical studies have not really picked up on this in a satisfac-
tory manner. One explanation for this discrepancy between theoretical and
Conclusion 145
meaningful to follow Penrose and suppose that growth is not just a means
to obtain a certain size, but rather it is an end in itself, a constructive appli-
cation of spare resources. Indeed, in the presence of learning-by-doing and
dynamic increasing returns, a lack of growth would be akin to stagnation.
As a result, we consider the notion of an ‘equilibrium growth rate’ to be
closer to the truth than that of an ‘equilibrium size’.
We also argue in favour of descriptive theorizing, ‘appreciative’ theoriz-
ing in the spirit of Nelson and Winter (1982), where the objective is not on
mathematical formalism or obtaining testable hypotheses, but an earnest
desire to provide as realistic a description of the phenomenon as possible.
Theoretical models can also play a valuable role in explaining specific
aspects of firm growth. Theoretical modellers, we suggest, should try to
compare the predictions of their model with as many stylized facts as pos-
sible. A large number of regularities can be found in the growth of firms,
as this survey has testified. For example, a model of firm growth can be
evaluated by considering the shape of the growth rate distribution, for
instance, or by looking at growth rate autocorrelation. A theoretical model
can be better understood if as many implications as possible are tried and
tested, and even if in some dimensions the model does not agree with the
empirical facts, these shortcomings should be acknowledged. In this way,
the gains and limits of the model can be better appreciated, and room is
left for further progress.
We are less interested in theoretical models that merely attempt to
explain the firm size distribution, however – there are many other styl-
ized facts available that can be useful in evaluating the contribution of a
theoretical model. There are indeed many statistical mechanisms that can
explain aggregate distributions (Brock, 1999), but the richness of a model
rests on other criteria, such as its ability to reproduce a number of stylized
facts, or its ability to describe the dynamics of the economic system. In
addition, we are sceptical of elaborate models trying to reconcile empirical
behaviour with rational maximizing behaviour, because the plain truth is
that firms are not perfectly rational profit maximizers.
Perhaps the main message that seemed to emerge from this monograph,
and especially the survey of empirical work, was that growth rates
appeared to be remarkably random by nature, reflecting the existence of
strong idiosyncratic components in the statistical series of firm growth. The
challenge researchers face, of course, is to further our knowledge of firm
growth by making original contributions to this literature that correspond
Conclusion 147
the motion of balls rolling down inclined planes (albeit without taking into
account such factors as friction, air resistance and magnetic fields). This
seems to us to be a sensible position to take.
Gibrat’s law has indeed proven to be a very useful model of firm growth,
and has spawned a truly vast empirical literature that seeks to determine
whether or not Gibrat’s law holds in any one particular database. It seems
to us that further work that aims to test Gibrat’s law is not the most fruitful
avenue of further progress, however. Instead, we recommend that research-
ers look for ways to gain new insights into firm growth, equipped with a
solid grasp of econometric techniques, and – more importantly – being
driven by a curiosity and an imagination that comes from a genuine desire
to glimpse further into the obscure cloud of confusion that surrounds the
subject of firm growth. In order to make progress in this field, therefore,
we feel obliged to reiterate an exhortation that is dated but nonetheless
still very relevant: ‘The subject of organizational growth has progressed
beyond abysmal darkness. It is ready for – and badly needs – solid, system-
atic empirical research directed toward explicit hypotheses and utilizing
sophisticated statistical methods’ (Starbuck, 1971, p. 126).
We wrap up by, once again, arguing in favour of Herbert Simon’s (1968)
research strategy, which emphasizes the need for solid empirical work to
first produce the ‘stylized facts’ that theory can then attempt to explain. At
this stage, we consider that research into the growth of firms could benefit
greatly from gathering of statistical regularities and ‘stylized facts’. We
consider that theory without any solid empirical basis – what we might call
‘armchair axiomatics’ (Dosi, 2004) – will be of little use in furthering our
knowledge of the growth of firms and the evolution of industries.
One of the more useful theories of firm growth was the descriptive theory
formulated by Edith Penrose in her celebrated book in 1959. The essence
of Penrose’s vision was that firms will always have internal resources for
growth because of learning-by-doing effects and, more specifically, the
freeing up of managerial attention as managers become increasingly accus-
tomed to their tasks. Unless the firm decides to grow, however, and unless
it chooses to make use of these spare resources, it appears to us that these
newly-liberated managerial resources will be absorbed as organizational
slack. Firms need to decide on the direction into which they can channel
these excess resources.
Growth can be seen as an entrepreneurial venture, no matter how large a
firm is. It is the quest for new opportunities. Growth requires imagination
Conclusion 149
1. Note also that, in a very small number of cases, value-added can take
on negative values. This would be the case, for instance, of a firm that
sells goods for less than the cost of labour.
CHAPTER 2
CHAPTER 3
152
Notes 153
CHAPTER 4
CHAPTER 5
firms can readily afford to pay the exit costs and promptly close down
the plant, and also because the higher abilities of the managers have a
higher opportunity cost and can quickly be put to good use elsewhere
in the economy.
16. Alternative schemes of decomposing productivity growth can be
found in Griliches and Regev (1995), Olley and Pakes (1996) and Aw
et al. (2001).
17. Traditional productivity indicators measure output in terms of total
sales rather than production, which makes establishments that charge
higher prices appear more productive.
18. Conventional econometric techniques that are applied to investi-
gate the causality between variables rely on instrumental variables.
Instrumental variable techniques may not be very effective in this
particular case, however, because firm growth is notoriously random
and it is unlikely that a suitable instrumental variable can be found.
For instance, in many applications of panel data instrumental variable
estimators, lagged variables are taken as instruments. It would hardly
be appropriate to take lags as instruments in the case of growth rates,
however, because of the low autocorrelation in growth rate series.
19. Due to the data construction procedure (growth rates calculated by
taking log-differences), firms with negative profits cannot be included
in the analysis.
CHAPTER 6
CHAPTER 7
1. For a survey of this work, see for example Carl Shapiro’s confidently-
titled article, ‘The theory of business strategy’ (Shapiro, 1989).
2. Other less common candidate variables for inter-firm competition
include import penetration (e.g. Haskel et al., 2007) and ‘profit elastic-
ity’ (Boone et al., 2007).
3. That is, when the combined market share of these two firms is greater
than 80 per cent.
4. It is interesting to also read onwards in Schumpeter (1942). The longer,
and more complete quotation, is as follows: ‘The businessman feels
himself to be in a competitive situation even if he is alone in his field
or if, though not alone, he holds a position such that investigating gov-
ernment experts fail to see any effective competition between him and
any other firms in the same or a neighboring field and in consequence
conclude that his talk, under examination, about his competitive
sorrows is all make-believe.’
5. When firms are pooled together in a cross-section, unionized firms
have lower expected growth rates, although this could simply reflect
the fact that unionized firms are more likely to be found in low-growth
sectors. When controlling for other influences (such as industry effects
and business cycle activity) the negative association between unioniza-
tion and firm growth disappears, such that, for a given firm, unioniza-
tion may not have any impact on its subsequent growth rates (Bronars
and Deere, 1993).
6. Pavitt (1984), Malerba (2002), Malerba and Orsenigo (1997) and a
number of other scholars have emphasized that innovation regimes
are sector-specific and that innovative activity undertaken by firms
varies considerably across sectors. Empirical evidence in Leiponen and
Drejer (2007) and Srholec and Verspagen (2008), however, finds that
there is a considerable amount of heterogeneity of firm-level innova-
tion strategies even within specific sub-sectors. The importance of the
sector in explaining innovative activity is therefore a matter of debate.
CHAPTER 8
1. One might see a resemblance here with some theories to be found in the
Vatican, which consider that people only have sex because they intend
to reach an ‘optimal’ family size . . .
2. Jacques Lesourne puts it this way – ‘L’entreprise cherchera à employer
ces ressources inutilisées, mais en le faisant en créera d’autres, en
158 The growth of firms
CHAPTER 9
1. The ‘entry deterrence’ argument is of limited relevance, because
entrants are usually too small to pose a serious threat. However,
the argument may hold as long as large firms in other industries are
deterred from diversifying into the sector under consideration.
2. Using survey evidence for Dutch SMEs, Lensink et al. (2005) observe
that higher growth firms perceive that they have more idiosyncratic
uncertainty than other firms.
3. In fact, it is precisely because of the intentionality attributed to the
growth of firms that Penrose (1955) rejects biological analogies as valid
descriptions of firm growth.
4. An unpublished comparison of sectoral growth rate distribution
parameters (at the 3-digit level) for Italy and France reveals that there
is very little in common in the growth rate distributions for same
sectors across countries. This hints that the underlying sector-specific
production technology does not go far in explaining growth rates –
instead it may well be that human factors play a major role.
5. For empirical evidence on the heterogeneity of firm productivity levels,
even within narrowly-defined industrial sectors, see Dosi and Grazzi
(2006). See also Dosi (2007) for evidence on the dispersion of profit
margins within industries.
6. A similar conclusion is reached in Chandler (1992, p. 94).
7. There is evidence that joint ventures undertaken with firms from the
host country seem to do better than alliances undertaken with firms
from the same country, or with firms from a third country (Lu and
Beamish, 2001).
CHAPTER 10
CHAPTER 11
161
162 The growth of firms
and SME Research: On its Way to the Next Millennium, Aldershot, VA:
Avebury, pp. 190–216.
Delmar, F. and J. Wiklund (2008), ‘The effect of small business man-
agers’ growth motivation on firm growth: a longitudinal study’,
Entrepreneurship Theory and Practice, 32(3), 437–57.
Delmar, F., P. Davidsson and W. Gartner (2003), ‘Arriving at the high-
growth firm’, Journal of Business Venturing, 18(2), 189–216.
Dickerson, A.P., H.D. Gibson and E. Tsakalotos (2000), ‘Internal vs exter-
nal financing of acquisitions: do managers squander retained profits?’,
Oxford Bulletin of Economics and Statistics, 62(3), 417–31.
Dierickx, I. and K. Cool (1989), ‘Asset stock accumulation and sustainabil-
ity of competitive advantage’, Management Science, 35(12), 1504–11.
Disney, R., J. Haskel and Y. Heden (2003), ‘Restructuring and productiv-
ity growth in UK manufacturing’, Economic Journal, 113, 666–94.
Dixit, A. (1980), ‘The role of investment in entry-deterrence’, Economic
Journal, 90(357), 95–106.
Dixon, R.L. (1953), ‘Creep’, Journal of Accountancy, July, 48–55.
Dobson, S. and B. Gerrard (1989), ‘Growth and profitability in the
Leeds engineering sector’, Scottish Journal of Political Economy, 36(4),
334–52.
Doms, M. and T. Dunne (1998), ‘Capital adjustment patterns in manufac-
turing plants’, Review of Economic Dynamics, 1, 409–29.
Doms, M., T. Dunne and M.J. Roberts (1995), ‘The role of technology
use in the survival and growth of manufacturing plants’, International
Journal of Industrial Organization, 13, 523–42.
Dosi, G. (2004), ‘A very reasonable objective still beyond our reach: eco-
nomics as an empirically disciplined social science’, in M. Augier and J.
March (eds), Models of a Man: Essays in Memory of Herbert A. Simon,
Cambridge, MA: MIT Press.
Dosi, G. (2007), ‘Statistical regularities in the evolution of industries:
a guide through some evidence and challenges for the theory’, in F.
Malerba and S. Brusoni (eds), Perspectives on Innovation, Cambridge,
UK: Cambridge University Press.
Dosi, G. and M. Grazzi (2006), ‘Technologies as problem-solving pro-
cedures and technologies as input-output relations: some perspectives
on the theory of production’, Industrial and Corporate Change, 15(1),
173–202.
Dosi, G. and D. Lovallo (1998), ‘Rational entrepreneurs or optimistic
martyrs? Some considerations on technological regimes, corporate
entries, and the evolutionary role of decision biases’, in R. Garud, P.
Nayyar and Z. Shapiro (eds), Foresights and Oversights in Technological
Change, Cambridge, UK: Cambridge University Press, pp. 41–68.
Bibliography 171
Evans, D.S. (1987a), ‘The relationship between firm growth, size and
age: estimates for 100 manufacturing industries’, Journal of Industrial
Economics, 35, 567–81.
Evans, D.S. (1987b), ‘Tests of alternative theories of firm growth’, Journal
of Political Economy, 95(4), 657–74.
Fagiolo, G. and A. Luzzi (2006), ‘Do liquidity constraints matter in
explaining firm size and growth? Some evidence from the Italian manu-
facturing industry’, Industrial and Corporate Change, 15(1), 1–39.
Fazzari, S., R. Hubbard and B. Petersen (1988a), ‘Financing constraints
and corporate investment’, Brookings Papers on Economic Activity,
1(1988), 141–95.
Fazzari, S., R. Hubbard and B. Petersen (1988b), ‘Investment, financ-
ing decisions, and tax policy’, American Economic Review Papers and
Proceedings, 78(2), 200–205.
Fazzari, S., R. Hubbard and B. Petersen (2000), ‘Investment-cash flow
sensitivities are useful: a comment on Kaplan and Zingales’, Quarterly
Journal of Economics, 115(2), 695–705.
Feldman, M. and B. Pentland (2003), ‘Reconceptualizing organizational
routines as a source of flexibility and change’, Administrative Science
Quarterly, 48(1), 94–121.
Feldman, M.S. (2004), ‘Resources in emerging structures and processes of
change’, Organization Science, 15(3), 295–309.
Fishman, A. and R. Rob (2003), ‘Consumer inertia, firm growth and indus-
try dynamics’, Journal of Economic Theory, 109(1), 24–38.
Fizaine, F. (1968), ‘Analyse statistique de la croissance des entreprises
selon l’âge et la taille’, Revue d’économie Politique, 78, 606–20.
Fleck, J. (1984), ‘The adoption of robots in industry’, Physics in Technology,
15, 4–11.
Fluck, Z. and A. Lynch (1999), ‘Why do firms merge and then divest? A
theory of financial synergy’, Journal of Business, 72(3), 319–46.
Foster, L., J. Haltiwanger and C. Krizan (2006), ‘Market selection, real-
location, and re-structuring in the US retail trade sector in the 1990s’,
Review of Economics and Statistics, 88(4), 748–58.
Foster, L., J. Haltiwanger and C.J. Krizan (1998), ‘Aggregate productivity
growth: lessons from microeconomic evidence’, NBER Working Paper
6803.
Foster, L., J. Haltiwanger and C. Syverson (2008), ‘Reallocation, firm
turnover, and efficiency: selection on productivity or profitability?’,
American Economic Review, 98(1), 394–425.
Freel, M. and P. Robson (2004), ‘Small firm innovation, growth and per-
formance: evidence from Scotland and Northern England’, International
Small Business Journal, 22(6), 561–75.
Bibliography 173
Hopenhayn, H. (1992), ‘Entry, exit, and firm dynamics in long run equilib-
rium’, Econometrica, 60(5), 1127–50.
Horngren, C.T. (1984), Introduction to Financial Accounting (2nd edn),
Englewood Cliffs, NJ: Prentice Hall.
Hoshi, T., A. Kashyap and D. Scharfstein (1991), ‘Corporate structure,
liquidity, and investment: evidence from Japanese industrial groups’,
Quarterly Journal of Economics, 106(1), 33–60.
Hu, X. and F. Schiantarelli (1998), ‘Investment and capital market imper-
fections: a switching regression approach using US firm panel data’,
Review of Economics and Statistics, 80(3), 466–79.
Hubbard, R. (1998), ‘Capital-market imperfections and investment’,
Journal of Economic Literature, 36(1), 193–225.
Huberman, B.A. and L.A. Adamic (1999), ‘Growth dynamics of the world-
wide web’, Nature, 401, 131.
Hughes, A. (1997), ‘Finance for SMEs: a UK perspective’, Small Business
Economics, 9(2), 151–68.
Hugo, O. and E. Garnsey (2005), ‘Problem-solving and competence crea-
tion in the early development of new firms’, Managerial and Decision
Economics, 26(2), 139–48.
Hyland, D.C. and J.D. Diltz (2002), ‘Why firms diversify: an empirical
examination’, Financial Management, 31, 51–81.
Hymer, S. and P. Pashigian (1962), ‘Firm size and rate of growth’, Journal
of Political Economy, 70(6), 556–69.
Idson, T. and W. Oi (1999), ‘Workers are more productive in large firms’,
American Economic Review Papers and Proceedings, 89(2), 104–8.
Ijiri, Y. and H.A. Simon (1964), ‘Business firm growth and size’, American
Economic Review, 54(2), 77–89.
Ijiri, Y. and H.A. Simon (1971), ‘Effects of mergers and acquisitions on busi-
ness firm concentration’, Journal of Political Economy, 79(2), 314–22.
Ijiri, Y. and H.A. Simon (1967), ‘A model of business firm growth’,
Econometrica, 35(2), 348–55.
Ijiri, Y. and H.A. Simon (1974), ‘Interpretations of departures from the
Pareto curve firm-size distributions’, Journal of Political Economy, 82(2),
315–31.
Ijiri, Y. and H.A. Simon (1977), Skew Distributions and the Sizes of
Business Firms, Amsterdam: North Holland.
Jensen, M. and W. Meckling (1976), ‘Theory of the firm: managerial
behavior, agency costs, and capital structure’, Journal of Financial
Economics, 3(4), 305–60.
Jensen, M.C. (1986), ‘Agency costs of free cash flow, corporate finance,
and takeovers’, American Economic Review Papers and Proceedings,
76(2), 323–9.
178 The growth of firms
Lotti, F., E. Santarelli and M. Vivarelli (2003), ‘Does Gibrat’s law hold
among young, small firms?’, Journal of Evolutionary Economics, 13(3),
213–35.
Lu, J. and P. Beamish (2001), ‘The internationalization and performance
of SMEs’, Strategic Management Journal, 22(6–7), 565–86.
Lucas, R. (1978), ‘On the size distribution of business firms’, Bell Journal
of Economics, 9(2), 508–23.
Luttmer, E. (2007), ‘Selection, growth, and the size distribution of firms’,
Quarterly Journal of Economics, 122(3), 1103–44.
MacDonald, C.J. (1998), ‘The evolution of Intel’s copy EXACTLY!
Technology transfer method’, Intel Technology Journal, quarter 4, 1–6.
Mairesse, J. and P. Mohnen (2002), ‘Accounting for innovation and meas-
uring innovativeness: an illustrative framework and an application’,
American Economic Review Papers and Proceedings, 92(2), 226–30.
Maksimovic, V. and G. Phillips (2002), ‘Do conglomerate firms allocate
resources inefficiently across industries? Theory and evidence’, Journal
of Finance, 57(2), 721–67.
Maksimovic, V. and G. Phillips (2008), ‘The industry life cycle, acqui-
sitions and investment: does firm organization matter?’, Journal of
Finance, 63(2), 673–708.
Malerba, F. (2002), ‘Sectoral systems of innovation and production’,
Research Policy, 31(2), 247–64.
Malerba, F. and L. Orsenigo (1997), ‘Technological regimes and sectoral
patterns of innovative activities’, Industrial and Corporate Change, 6(1),
83–118.
Mansfield, E. (1962), ‘Entry, Gibrat’s law, innovation, and the growth of
firms’, American Economic Review, 52(5), 1023–51.
Mansfield, E., J. Rapoport, A. Romeo, E. Villani, S. Wagner and F. Husic
(1977), The Production and Application of New Industrial Technology,
New York: Norton.
Markides, C. (1992), ‘Consequences of corporate refocusing: ex ante evi-
dence’, Academy of Management Journal, 35(2), 398–412.
Markides, C. (1995), ‘Diversification, restructuring and economic per-
formance’, Strategic Management Journal, 16(2), 101–18.
Marris, R. (1963), ‘A model of the managerial enterprise’, Quarterly
Journal of Economics, 77(2), 185–209.
Marris, R. (1964), The Economic Theory of Managerial Capitalism,
London: Macmillan.
Marris, R. (1999), ‘Edith Penrose and economics’, Contributions to
Political Economy, 18, 47–65.
Marshall, A. (1961), Principles of Economics, 8th edn, London, UK:
Macmillan and Co., First published 1890.
Bibliography 181
Minkoff, D. (1999), ‘Bending with the wind: strategic change and adapta-
tion by women’s and racial minority organizations’, American Journal
of Sociology, 104(6), 1666–703.
Mitra, R. and V. Pingali (1999), ‘Analysis of growth stages in small firms: a
case study of automobile ancillaries in India’, Journal of Small Business
Management, 37(3), 62–3.
Mitzenmacher, M. (2003), ‘A brief history of generative models for power
law and lognormal distributions’, Internet Mathematics, 1(2), 226–51.
Montgomery, C.A. (1994), ‘Corporate diversification’, Journal of Economic
Perspectives, 8(3), 163–78.
Mowery, D. (1983), ‘Industrial research and firm size, survival, and growth
in American manufacturing, 1921–1946: an assessment’, Journal of
Economic History, 43(4), 953–80.
Mueller, D.C. (1969), ‘A theory of conglomerate mergers’, Quarterly
Journal of Economics, 83(4), 643–59.
Myers, S. (1984), ‘The capital structure puzzle’, Journal of Finance, 39(3),
575–92.
Nafziger, E. and D. Terrell (1996), ‘Entrepreneurial human capital and the
long-run survival of firms in India’, World Development, 24(4), 689–96.
Nelson, R.R. (1987), Understanding Technical Change as an Evolutionary
Process, Amsterdam: North Holland.
Nelson, R.R. and S.G. Winter (1982), An Evolutionary Theory of Economic
Change, Cambridge, MA: Belknap Press of Harvard University Press.
Niefert, M. (2005), ‘Patenting behavior and employment growth in
German start-up firms: a panel data analysis’, Discussion Paper 05-03,
ZEW Centre for European Economic Research, Mannheim.
O’Farrell, P.N. and D.M.W.N. Hitchens (1988), ‘Alternative theories of
small-firm growth: a critical review’, Environment and Planning A, 20(3),
1365–83.
Oliner, S. and G. Rudebusch (1992), ‘Sources of the financing hierarchy for
business investment’, Review of Economics and Statistics, 74(2), 643–54.
Olley, G. and A. Pakes (1996), ‘The dynamics of productivity in the tele-
communications equipment industry’, Econometrica, 64(6), 1263–97.
Oviatt, B. and P. McDougall (1994), ‘Toward a theory of international
new ventures’, Journal of International Business Studies, 25(1), 45–64.
Reprinted in Journal of International Business Studies (2005), 36(1),
29–41.
Packard, D. (1995), The HP way: How Bill Hewlett and I built our company,
New York: HarperCollins.
Pakes, A. and R. Ericson (1998), ‘Empirical implications of alternative
models of firm dynamics’, Journal of Economic Theory, 79(1), 1–45.
Parkinson, C.N. (1957), Parkinson’s Law, Boston, MA: Houghton Mifflin.
Bibliography 183
Reichstein, T. and M. Jensen (2005), ‘Firm size and firm growth rate
distributions – the case of Denmark’, Industrial and Corporate Change,
14(6), 1145–66.
Rivkin, J. (2001), ‘Reproducing knowledge: replication without imitation
at moderate complexity’, Organization Science, 12(3), 274–93.
Roberts, J. (2004), The Modern Firm: Organizational Design for
Performance and Growth, Oxford, UK: Oxford University Press.
Robson, P. and R. Bennett (2000), ‘SME growth: the relationship with
business advice and external collaboration’, Small Business Economics,
15(3), 193–208.
Robson, P. and B. Obeng (2008), ‘The barriers to growth in Ghana’, Small
Business Economics, 30(4), 385–403.
Roll, R. (1986), ‘The hubris hypothesis of corporate takeovers’, Journal of
Business, 59(2), 197–216.
Roper, S. (1997), ‘Product innovation and small business growth: a com-
parison of the strategies of German, UK and Irish companies’, Small
Business Economics, 9(6), 523–37.
Rossi-Hansberg, E. and M. Wright (2007), ‘Establishment size dynam-
ics in the aggregate economy’, American Economic Review, 97(5),
1639–66.
Rumelt, R. (1982), ‘Diversification strategy and profitability’, Strategic
Management Journal, 3(4), 359–69.
Salop, S.C. (1979), ‘Strategic entry deterrence’, American Economic Review
Papers and Proceedings, 69(2), 335–38.
Salter, W.E.G. (1960), Productivity and Technical Change, Cambridge,
UK: Cambridge University Press.
Samuels, J. (1965), ‘Size and the growth of firms’, Review of Economic
Studies, 32(1), 105–12.
Santarelli, E. and M. Vivarelli (2007), ‘Entrepreneurship and the process
of firms’ entry, survival and growth’, Industrial and Corporate Change,
16(3), 455–88.
Sapienza, H., E. Autio, G. George and S. Zahra (2006), ‘A capabilities
perspective on the effects of early internationalization on firm survival
and growth’, Academy of Management Review, 31(4), 914–33.
Sarno, D. (2008), ‘Capital structure and growth of the firms in the back-
ward regions of the [sic] south Italy’, Applied Financial Economics, 18,
821–33.
Sastry, M. (1997), ‘Problems and paradoxes in a model of punctu-
ated organizational change’, Administrative Science Quarterly, 42(2),
237–75.
Schaller, H. (1990), ‘A re-examination of the q theory of investment using
US firm data’, Journal of Applied Econometrics, 5(4), 309–25.
Bibliography 185
191
192 The growth of firms
Liu, J. et al. 41, 64, 85, 90, 97, 98 Miner, J. et al. 136
Lockett, A. et al. 124, 147 Minimum Efficient Scale (MES) 41, 93
lognormal distribution 14, 16, 20, 101 Minkoff, D. 109
Lotti, F. and E. Santarelli 17 Mitzenmacher, M. 22
Lotti, F. et al. 39, 42–3 Montgomery, C. 102–3, 121, 122, 123
Lucas, R. 101 motivation, lack of, in larger firms 114,
Luttmer, E. 22 117
Mowery, D. 42, 78
McCombie, J. 70 Mueller, D. 104, 119, 125
MacDonald, C. 118, 119 multiplant firms 68, 90, 96, 97, 98
McDougall, P. et al. 127, 128 Myers, S. 58
McKelvey, B. and P. Andriani 29, 36,
86, 137 Nafziger, E. and D. Terrell 131
McPherson, M. 41, 43, 88, 89, 94, 95, Nelson, R. 114
97, 98 Nelson, R. and S. Winter 5, 77, 106,
Maksimovic, V. and G. Phillips 64, 111, 146
93–4, 123, 124, 125, 147 neoclassical economics 7, 8, 61, 62, 63,
management 69, 145
cognitive leadership 139 ‘optimal size’ 56, 60, 100–101, 102,
control and small firms 105 103, 107, 112, 145
and diversification 105 q-theory 51–4, 55, 57, 59, 60, 117
entrenchment 121 Netherlands 18, 42, 92, 114
incentives 104, 121 networks 91, 126
pursuit of growth 58, 69, 70, 104 Niefert, M. 82
resources 3, 54, 91, 102, 112–14, 115,
116–17, 119–20, 121, 125, 138–9 O’Farrell, P. and D. Hitchens 136
shareholder-wealth-maximizing Oliner, S. and G. Rudebusch 55
managers 59–61 ‘optimal size’ theory 56, 60, 100–101,
talent distribution and small firms 102, 103, 107, 112, 145
101 organizational
theory of the firm 103–5 change 3, 5, 11, 112, 114, 141
Mansfield, E. 19, 40, 42, 77–8, 147 slack 31, 32, 70
manufacturing industries see under Oviatt, B. and P. McDougall 128
individual countries ownership structure 90
market
concentration 93 Pareto distribution 16, 22–3, 28
power 107 Parkinson, C. 115–16
selection 61 partnership, inter-firm 91
value 4, 51, 52, 53, 57, 76, 97 patents 77, 78, 79–80, 82, 97
Markides, C. 123 Paulré, B. 121
Marris, R. 4, 5, 58, 103–5, 119, 125 Pavcnik, N. 64
Marshall, A. 132–3 Pavitt, K. 131
Marsili, O. 16, 43, 94, 96 Penrose, E. 4, 31, 32, 47, 69, 91, 111,
Martin, J. and A. Sayrak 122 119–20, 131, 135, 145–6, 148
Matia, K. et al. 44, 45 theory of growth of firm 102–3, 113,
Matsusaka, J. 123 116–17
Mead, D. and C. Liedholm 88–9 pharmaceutical industry 16–17, 41, 43,
mergers and acquisitions see 44, 45, 46, 79
acquisitions Phillips, B. and B. Kirchhoff 22, 42, 133
Metcalfe, J. 76, 106 population ecology 108–9
Index 197