Professional Documents
Culture Documents
SME Performance Separating Myth From Reality (PDFDrive)
SME Performance Separating Myth From Reality (PDFDrive)
SME Performance Separating Myth From Reality (PDFDrive)
SME Performance
Separating Myth from Reality
John Watson
Professor of Accounting and Finance, The University of
Western Australia
Edward Elgar
Cheltenham, UK • Northampton, MA, USA
© John Watson 2010
Published by
Edward Elgar Publishing Limited
The Lypiatts
15 Lansdown Road
Cheltenham
Glos GL50 2JA
UK
A catalogue record for this book is available from the British Library
1 Introduction 3
4 Failure rates 47
5 Relating outputs to inputs 53
6 Adjusting for risk 59
7 A qualitative analysis 69
8 A quantitative analysis 87
PART VI CONCLUSIONS
References 140
Index 153
v
PART I
Background
1. Introduction
1.0 BACKGROUND AND MOTIVATION
The statistics on the longevity and mortality of small business in Australia show
a very disturbing picture. Nearly half go into receivership within three years
of commencement, and about 80 percent are out of business within ten years.
(Cohen 1987, p.6)
I had two problems with this quote and its use to promote the services of
chartered accountants. First, the extremely high failure rate referred to
did not seem to reflect the experiences of SME owners within my local
community. For example, I was aware of a number of SME owners who
had run very successful businesses over many years and who had eventu-
ally closed their businesses only when they felt it was time to retire. Were
these owners failures because they eventually retired and closed their busi-
nesses? I think not! Second, I was particularly concerned that such a highly
regarded professional body, of which I was a member, might have either
knowingly or unknowingly been using misleading statements to promote
the services of its members. While I strongly believe that SME owners can
benefit from obtaining professional advice, I do not believe the services of
chartered accountants (or any other service provider) should be promoted
on the basis of misleading statements.
Coincidentally, I also happened to be searching for a PhD topic at
that time, and therefore decided that an examination of SME failure
rates could potentially make a significant contribution to knowledge. As
3
4 SME performance
the first step down the PhD trail I undertook a literature review and it
soon became clear that the high mortality rate for SMEs referred to in
the Institute of Chartered Accountant’s brochure was simply reflecting
the consensus opinion at that time. The following quotes illustrate those
views: ‘The odds are well stacked against success, since in the United States
about 50 per cent of new ventures fail in the first two years and only a tiny
minority last ten years’ (Bannock 1981, p.34); ‘There is a lot of statistical
information in the literature on small business failures. The consensus of
opinion seems to be that between 50 and 60 per cent of such ventures fail
within three years of starting’ (Leslie, Magdulski and Champion 1985,
p.27); ‘Four out of five new firms fail within the first five years’ (Phillips
and Kirchoff 1989, p.65); ‘the literature . . . suggests that between one-
third and a half of new firms cease trading in their early years’ (Cromie
1991, p.44); and ‘Every year more than 100,000 new businesses open their
doors. The hazards, however, are so great that 95 percent eventually fail’
(Thankappan and Hammer 1980, p.1).2
However, I found much of the available literature confusing because
of the variety of definitions (or proxies) used to describe SME failure or
success. As a result, and in the absence of any contrary evidence, I felt that
dubious statistics suggesting very high failure rates for SMEs might have
become part of the folklore and received wisdom on this subject. It was
also common at that time (and, unfortunately, still occurs occasionally
today) for SME conference speakers to begin their presentations with a
justification of the importance of their topic based on the very high failure
rates prevalent within the sector.
The following quote further helped to convince me that an examination
of SME failure rates could make a useful contribution to our knowledge
concerning SME performance and the risks involved for the would-be
entrepreneur:
Like the weather, small business failure is the subject of much discussion . . .
But unlike the weather . . . there is . . . a dearth of timely, reliable, and relevant
information on small business failure rates. (Cochran 1981, p.50)
Therefore, while much had been written about SMEs, and in particular
about SME failure rates, reliable statistics on small business failure were
scarce and, as will be seen later, had typically been produced or inferred
from databases designed for other purposes.3 As stated by Scott and Lewis
(1984, p.49) ‘the absence of good statistical evidence leads to the growth
of myths and half truths’ and, as noted by Stanworth (1995, p.59), these
myths get ‘reported by the media, perpetuated by spokespeople for the
industry and subsequently accepted by the wider public’. Without reliable
Introduction 5
owners to access external funding; and female SME owners are not dis-
advantaged, relative to male SME owners, in terms of their networking
activities.
It should be noted, at the outset, that the focus of this book is on the
individual SME owner and, therefore, implicit in this book is the notion
that ‘failure’ is ‘bad’ and that reducing ‘failure’ rates is ‘good’. However,
from a societal perspective it can be argued that some level of ‘failure’
is ‘good’ because it allows inefficient operators to be replaced by more
efficient operators (Schumpeter 1942). Indeed Knott and Posen (2005,
p.638) found that, within the banking sector, ‘Excess entry and subse-
quent failure increase aggregate industry efficiency.’ Similarly, using real
options reasoning, McGrath (1999, p.16) noted that ‘A high failure rate
can even be positive, provided that the cost of failing is bounded.’
Having provided the background and motivation for this book, Part II
outlines the various issues that need to be considered if we want to get
a better understanding of: what constitutes failure; the rate of failure
within the SME sector (Chapter 2); and how economic and other factors
(such as age of business) are likely to impact reported SME failure rates
under alternative definitions of failure (Chapter 3). Part III compares
male- and female-controlled SMEs on a number of dimensions, such as:
business closure rates (Chapter 4); return on assets (Chapter 5); and risk-
adjusted returns (Chapter 6). Part IV examines the relationship between
external funding and firm growth using both a qualitative (Chapter 7)
and a quantitative approach (Chapter 8). This is followed in Part V by
an examination of the association between networking and firm perform-
ance (Chapter 9) and the differences in networking activities for male and
female SME owners (Chapter 10). Finally, Part VI (Chapter 11) concludes
the book with a summary of the key findings from the earlier chapters and
with some suggestions for future research.
I trust that the material provided in this book will help clarify a
number of important misconceptions relating to SME performance to
ensure that policy decisions by governments, bankers, service provid-
ers and any other groups with an interest in the SME sector are based
on reliable statistical analysis and not on unsubstantiated myths that
have been permitted to flourish in the absence of such evidence. Indeed,
anyone with an interest in SMEs should find the material presented in
the remainder of this book essential to a proper understanding of SME
performance.
8 SME performance
NOTES
1. It should be noted that in the literature (particularly the early literature) it is generally
assumed that firms that have not failed are successful. More recently some researchers
have moved away from a dichotomous definition of success/failure (such as bankrupt/
not bankrupt) to more continuous measures (for example, percentage growth in sales or
return on assets).
2. Further examples of similar comments can be found in Massel (1978) and Scott (1982).
3. Excellent literature reviews are provided by Berryman (1983) and Cochran (1981).
PART II
SME performance
In this section I intend to dispel the myth that SMEs suffer from exces-
sively high failure rates. Chapter 2 looks at how we might consider defin-
ing and measuring SME failure and success and then Chapter 3 examines
the likely impact of age, size, industry and the state of the economy on
reported failure rates. It is important that researchers, policy makers and
others with an interest in the SME sector adopt appropriate definitions of
failure and success or, at the very least, make clear any potential limita-
tions with the definition being used. Otherwise, as noted in Chapter 1,
inappropriate policies or actions might be adopted to the potential detri-
ment of SME owners and the future health of the economy.
For example, if we define failure as bankruptcy proceedings being initi-
ated against a firm, then it is likely that failure rates will increase during
periods of recession (as interest rates increase) and will decrease during
good economic times. However, if the sale of a business is used as the
definition of failure, then good economic times are likely to be associated
with an increase in the level of SME failures (as owners take the opportu-
nity to sell their businesses and retire, or to move into paid employment,
or to look for other opportunities) and periods of recession are likely to
be associated with lower failure rates (because owners will have limited
opportunities to sell their businesses). Such starkly contrasting outcomes
(depending on which failure definition is adopted) could, potentially, be a
source of some confusion for policy makers interested in the health of the
SME sector. Similarly, if as the result of an inappropriate definition being
used to measure SME failure, groups such as bankers are led to believe
that SMEs have very high failure rates, they might be unwilling to lend
to this sector or, if they do so, they might charge an interest rate premium
(Phillips and Kirchoff 1989) or insist on other conditions which are likely
to inhibit new firm start-ups and growth.
10 SME performance
How to adequately assess SME failure and success has long been a con-
troversial issue because the type of data routinely available to assess the
performance of large businesses has simply not been available for the SME
sector. Cochran (1981) suggested that the lack of a reliable measure of
failure was a major obstacle to understanding and alleviating the causes
of small business failure and Scott and Lewis (1984, p.49) noted that
‘[o]ne practical implication of this is that ill-founded policy must necessar-
ily follow’. Prior to looking more closely at some of the commonly used
indicators of SME failure and success, this chapter will consider various
attributes that might be considered when selecting a measure of perform-
ance for research or other purposes. Later in the chapter, I will discuss the
likely failure rates that might be expected using the various performance
measures suggested in the literature. This will enable readers, based on
the performance indicator(s) they believe to be the most appropriate, to
draw their own conclusions concerning the potential risks SME owners
confront.
11
12 SME performance
Objectivity/Verifiability
Relevance/Representational Faithfulness
Simplicity/Parsimony
As a rule, simple measures are less prone to error and should, there-
fore, be preferred to more complex measures. Studies adopting simple/
parsimonious measures are also more easily replicated and, therefore,
potentially more generalizable.
be a single measure that has all the desirable attributes and which meets the
needs of all users (such as: credit providers; owners and potential owners;
advisers to small business; and policy makers). For example, it is possible
that some objectivity might have to be sacrificed to obtain a measure that
is more relevant in a particular situation or for a particular user group.
We can now turn to an examination of a number of SME failure
measures commonly referred to in the literature.
Similarly, Bates (2005, p.344) notes that business closure ‘is not necessar-
ily rooted in failure or even performance that lags behind expectations;
departure requires only that a superior alternative has become available’.
In summary, using discontinuance as a measure of failure has the
advantage that it can be a relatively objective (verifiable) and simple
measure. However, it might be a biased measure if sole traders and part-
nerships are treated differently from companies. Further, it is difficult to
see how recording as failures all businesses that are closed or sold, irre-
spective of the reason for the sale or closure, is likely to provide useful or
relevant information for a number of key interest groups such as: credit
providers; owners and potential owners; advisers to small business; and
policy makers.
Bankruptcy/Losses to Creditors
Dun and Bradstreet (1979) classify all businesses that are placed into
bankruptcy, or cease operations with resulting losses to creditors, as
failed. The implication is that continuing businesses and businesses that
cease without any losses to creditors (although there might have been
losses to the owners) are regarded as successful (non-failed). This appears
to be a very narrow definition of failure and excludes many businesses that
might commonly be regarded as having failed, for example, businesses
that are barely breaking even and, therefore, not providing a reasonable
income or return for their owners (Land 1975). Examples of studies using
bankruptcy (losses to creditors) to define failure include: Massel (1978);
Cahill (1980); Hall and Young (1991); Lowe, McKenna and Tibbits
(1991); Harada (2007); and Hudson (1997). While reasonably homoge-
neous in terms of the way failure is defined, these studies generally only
examine a cohort of failed firms and typically provide no information on
the overall population of small businesses (that is, failed and non-failed
firms). These studies, therefore, do not usually provide annual failure rate
estimates. For the few studies where annual failure rates are estimated they
range from 0.43% to 1.3%.
As with discontinuance, bankruptcy has the advantage of being an
16 SME performance
Ulmer and Nielsen (1947, p.11) defined as failed ‘firms that were disposed
of (sold or liquidated) with losses to prevent losses’. Losses in this context
include the owner’s capital, and a business could therefore be regarded as
having failed even though there might not have been any losses to creditors.
Defining failure to include businesses that were sold or ceased to prevent
further losses appears more relevant for owners and potential owners,
advisers to small business and policy makers than using a measure based
on either discontinuance or bankruptcy. However, this measure is neither
as simple nor as objective as either bankruptcy or discontinuance as it
requires information from someone associated with the business. Because
such information would not generally be available from external (third-
party) sources, failure statistics reported using such a measure could be
difficult to verify and this might explain its limited use.
attending to customers while her husband spent most of his time in the
workshop. This business operated for some 30 years and provided the
couple with a good standard of living and a comfortable retirement. When
the couple were ready to retire they closed the business and the lease for
the premises was taken over by new owners, who established a coffee and
cake shop. Was the watch repair and jewellery shop a failure?
Example 2: Bookshop
A gentleman in his early forties started this business. It was very success-
ful and on the strength of its success the owner was able to borrow a large
sum of money from a bank to invest in a property deal. Unfortunately,
the property deal was not successful and the bank placed the owner (and
the bookshop) into bankruptcy and then sold the bookshop to recover
the monies owing. The proceeds from selling the bookshop were suffi-
cient to cover all the owner’s debts and the business continued under new
management/ownership. Should this business be classified as a failure,
either at the time it was placed into bankruptcy or when it was sold?
The manager of the shopping centre in which this business was located
felt that the business could be significantly improved. However, the owner
of the business rejected all of the manager’s suggestions. Given that the
rent paid for the premises occupied by this business included a component
based on gross sales, the manager ultimately decided not to renew the lease
and, as a result, the business closed. However, shortly afterwards, the busi-
ness reopened in another location. Should this business be classified as a
failure when it ceased to operate in the managed shopping centre?
Notes:
a. Note that business closure is a subset of discontinuance.
b. Assuming the researcher was aware that this business had relocated.
c. This business would most likely be classified as having failed unless the researcher was
able to access detailed information about its performance and the reason why it was the
subject of a forced (bankruptcy) sale.
d. Based on the opinion of the shopping centre manager, which was probably not shared
by the business owner.
there is little doubt that the business itself was very successful. A similar
conclusion regarding the bookshop would almost certainly be reached if
bankruptcy is the performance measure being used, unless the researcher
spent considerable time looking into all the circumstances surrounding the
bankruptcy proceedings and the reason for the forced sale of this business.
Interestingly, the fruit and vegetable shop would not have been considered
a failure under any of the definitions except failing to ‘make a go of it’, and
then only if the opinion of the shopping centre manager (rather than that
of the owner) was sought. In summary, it would seem that if we were to get
an expert opinion on the success or failure of each of these three businesses,
it is most likely that the fruit and vegetable shop would be considered the
least successful. However, this business is the least likely to be recorded
as a failure by researchers using secondary data where they are unable to
access any first-hand knowledge concerning individual businesses and the
reasons for their discontinuance.
The second point to note from Table 2.1 is the wide variation in
reported failure rates (from 0% to 66%) depending on the definition
selected and how that definition is implemented in specific cases. Of
20 SME performance
particular concern is the fact that the first and last definitions in Table
2.1 give exactly opposite results for each of the three examples provided.
I believe these examples are not uncommon, and we should therefore be
very careful in interpreting failure rate statistics derived from secondary
data sources where it has not been possible to obtain any input from a
person (or persons) knowledgeable about the businesses being examined.
This is an issue that should be of concern to policy makers and others
with an interest in the SME sector and suggests that researchers need
to be very careful in selecting and implementing performance measures
within this sector. Further, given that there is unlikely to be one single
performance measure that is the most appropriate in all cases, research-
ers need to carefully articulate any shortcomings in the measure(s) they
adopt.
I will now explore the likely SME failure rates that could be expected
based on the various definitions discussed above.
Note: a. Other included, for example, marriage breakdowns, and in such cases the
shopping centre manager was asked to give his/her opinion as to whether the business had
been successful prior to its sale/closure or whether the owner(s) had failed to ‘make a go
of it’. Based on the manager’s assessment, the business was then classified as failed/not
failed.
22
2. To avoid further losses 415 415 415 270 65
3. Did not make ‘a go of it’ 267 267 162 61
4. Retirement or ill health 126 37 29
5. To realize a profit 916 152 17
6a. Other – not failed 277 78 28
6b. Other – failed 34 34 23 68
179 594 895 2214 836 38
7. Unknown 166b 329 166 50
Totals 179 594 1061 2543 1002
% of Discontinuances 7 23 42 100 39
% of all businesses (n=5196) 3.4 11.4 20.4 49 19
Average annual failure rate (%) 0.7 2.3 4.1 9.4 3.9
Businesses < 5years old (%) 72 76 75 75 66
Notes:
23
a. Note that all businesses that are sold or closed due to bankruptcy, or to prevent further losses or because the owner(s) failed to ‘make a go of it’
are recorded as discontinuances. That is, the first three definitions of failure are assumed to be subsets of discontinuance. Similarly, bankruptcy
is a subset of both ‘to prevent further losses and ‘failed to “make a go of it”’ and ‘to prevent further losses’ is a subset of ‘failed to “make a go
of it”’. The last two columns of this table provide information with respect to the subset of businesses that were closed (rather than sold).
b. For the 329 businesses where the reason for their discontinuance was unknown I have assumed they failed to ‘make a go of it’ if the business
was closed (rather than sold to new owners). The direction and size of any bias caused by this assumption is unknown.
Source: Adapted from Watson and Everitt (1996, Tables 3 and 7).
24 SME performance
Table 2.4 SME failure rates within first five years of start-up, 1961–90a
(%)
Note: a. Note that firms that commenced during the period 1985–90 are excluded from
the analysis as five years of data post start-up are required for this analysis.
than 75% for all the failure definitions being considered. For example, if
we use business closure as our definition of failure, we can see from Table
2.4 that 20.7% of Australian businesses located within a managed shop-
ping centre failed within five years, or, alternatively, almost 80% of busi-
nesses survived beyond five years. These results clearly indicate that the
overwhelming majority of Australian SMEs are likely to survive beyond
five years.
Comparing the average annual rate of business closure in Table 2.3 for
all businesses (3.9%) with the rate in Table 2.4 for businesses less than
five years old (4.5%) we can see, as expected, that the closure rates are
higher for firms in their first five years of existence. This finding confirms
the expectation that firms are most vulnerable in their early years as they
learn about their industry. ‘The efficient grow and survive while the ineffi-
cient decline and fail’ (Jovanovic 1982, p.649). There have been numerous
studies confirming this proposition (referred to as the ‘liability of newness’,
Stinchcombe 1965), for example: Freeman, Carroll and Hannan (1983);
Stewart and Gallagher (1986); Evans (1987); Bates and Nucci (1989); and
Dunne, Roberts and Samuelson (1989).
In this section I will examine a number of studies that can provide com-
parative statistics of the rate of SME bankruptcy and closure for different
periods and geographical settings. I am not aware of any comprehensive
studies that can provide failure rate comparisons using either ‘failed to
“make a go of it”’ or ‘ceased to prevent further losses’ as a definition of
failure. Also, discontinuance (sale or closure), although often used in early
studies, is now seldom considered an appropriate measure of SME failure
and, therefore, will not be examined any further.
Bankruptcy Rates
Closure Rates
targeted for follow-up surveys in each of the three subsequent years.6 The
closure rates for these businesses in the subsequent three survey periods
are presented in Table 2.5, which reports an average annual closure rate
for Australian SMEs of 8%. While this rate is almost double the 3.9%
reported in Table 2.3 for Australian businesses located within managed
shopping centres, the higher rate for the broader population of Australian
SMEs is not unexpected given that SMEs operating within a managed
shopping centre would normally have been screened prior to start-up
and, in many cases, would receive ongoing advice and support.7 Again, it
should be emphasized that the 8% closure rate reported above should not
be interpreted as representing the failure rate for Australian SMEs because
many of these businesses are likely to have been successful at the time of
their closure. For example, many of them will have closed because the
owner(s) decided it was time to retire.
The second study of business closures I would like to discuss is a com-
prehensive US study by Headd (2003, p.51) that sought ‘to challenge
the widely held but often unsubstantiated belief that new firm closure rates
are high and that a closure is a negative outcome’. Headd estimates that
49.6% of US employing businesses that commenced operations during
the period 1989–98 closed within four years. This represents an average
annual closure rate of approximately 16% for firms less than four years
old. This closure rate for US employing businesses is double the rate
shown in Table 2.5 for Australian employing businesses. However, it is
important to note that the rate for the US relates only to newly formed
businesses and, therefore, it is reasonable to expect that this rate would be
28 SME performance
significantly higher than that applying to all businesses (old and young).
This issue will be explored further in the following chapter.
The third study of business closure I would like to discuss is a longitu-
dinal study undertaken by Box (2008) that included 2154 Swedish joint-
stock companies in seven birth cohorts that commenced in various years
between 1899 and 1950. Box’s results indicate that, on average, 75% of
firms survived beyond four years.8 Box also reports that the survival rates
varied significantly across the cohort groups, most likely as the result of
different environmental (economic) forces. While Box’s (2008) average
four-year survival rate is considerably higher than the 50% reported by
Headd (2003) for US firms, it should be noted that Box’s sample was of
incorporated businesses and it is reasonable to expect that they will have
higher survival rates (lower closure rates) because they are likely to have
undergone greater scrutiny prior to start-up compared to firms that begin
as sole traders or partnerships. Interestingly, the four-year survival rate of
75% reported by Box (2008) is similar to the 80% five-year survival rate for
Australian SMEs located within managed shopping centres, as shown in
Table 2.4.
The final study of business closures I would like to discuss was carried
out by Forsyth (2005), who examined a cohort of 4103 small firms that
started operations in 1992 (the beginning of a period of economic expan-
sion) in one of Washington’s 27 rural counties. Forsyth followed this
cohort of firms from their inception to 2000 (the peak of the economic
expansion), distinguishing between employing and non-employing busi-
nesses. Forsyth (2005) reported a 60% four-year survival rate on average
for all rural firms. This survival rate is reasonably similar to the 50% four-
year survival rate reported by Headd (2003), particularly given that the
period referred to by Headd included the 1990–91 recession. However,
Forsyth’s (2005) survival rate varied considerably by employment status:
from 56% for non-employers to 77% for firms with at least one employee.
2.6 SUMMARY
Table 2.6 Summary bankruptcy and closure rates from selected studies
NOTES
The previous chapter reviewed some of the more commonly used measures
of SME success and failure and showed how reported failure rates can
vary substantially depending on the definition of failure used. The broader
the definition, the higher the likely failure rate; the narrower the definition,
the lower the likely failure rate. While each of the definitions has certain
strengths, none is clearly superior. Further, the results of prior studies
suggest that reported failure rates (even where the same definition is used)
can vary substantially. For example, the annual bankruptcy rates reported
in Chapter 2 range from 0.7% to 1.8%, with business closure rates ranging
from 3.9% to 16%. This large variance in reported failure rates must surely
confuse policy makers and others interested in the SME sector. This
chapter will consider some of the potential reasons for these differences, in
particular, age, size, industry and the state of the economy.
Most studies of SME failure have found that a business is at greatest risk
in its first few years. Jovanovic (1982) argued that younger firms are more
likely to fail because they face greater variability in their cost functions
while they learn about their industry and management capabilities. The
efficient grow and survive; the inefficient decline and fail. For this reason,
younger firms are less likely to survive than older ones. There have been
numerous studies confirming this proposition: Stewart and Gallagher
(1986); Evans (1987); Bates and Nucci (1989); and Dunne, Roberts and
Samuelson (1989). I will now look at how failure rates can be expected to
vary with firm age.
For each failure definition, Figure 3.1 plots the average annual failure
rate by age of business for my sample of 5196 business start-ups operating
within 51 Australian managed shopping centres over the period 1961–90,
as described in Chapter 2. As can be seen from this figure, the failure rates
31
32 SME performance
18
16
Discontinuance of ownership
14
Probability of failure (%)
12
10
8
Failed to ‘make a go of it’
6
Discontinuance of business
4
To prevent further losses
2
Bankruptcy
0
0.5 1.5 2.5 3.5 4.5 5.5 6.5 7.5 8.5 9.5
Age (years)
generally reach a maximum at around three years of age and then gradu-
ally decline. Brüderl, Preisendörfer and Ziegler (1992, p.234) note that
this inverted U-shaped curve can be explained by the fact that ‘[m]ortality
rates are low immediately after starting a business because organizations
can survive on initial resources, increase to a maximum, and decline after-
wards’. This mortality process has been labelled the ‘liability of adoles-
cence’, in contrast to ‘liability of newness’, which depicts monotonically
declining failure rates (Brüderl and Schussler 1990). It should be noted
that Box (2008), Cressy (2006) and Ganguly (1985) also report failure rates
conforming to a bell-shaped distribution.
Also note that the failure rate distributions reported in Figure 3.1 for
‘discontinuance of business (closure)’ and ‘failed to “make a go of it”’ are
very similar. This finding further supports the observation in Chapter 2
that (at an aggregate level) where an objective measure of SME failure is
required, business closure might be an appropriate proxy. However, as
explained in Chapter 2, this measure is likely to overstate the ‘true’ SME
failure rate because many businesses close while successful (Headd 2003).
Given that failure rates appear to peak during the first five years of
a firm’s existence, Table 3.1 reports average annual failure rates for the
Effects of age, size, industry and the economy 33
Notes:
* Significantly different from average failure rate at 5% using a one-tailed test.
** Significantly different from average failure rate at 1% using a one-tailed test.
following three age classifications: less than five years, five to ten years
and over ten years. As expected, the results generally indicate significantly
higher than average failure rates for the cohort of firms less than 5 years
of age, with lower failure rates being reported for the two older age groups
and particularly for those firms over 10 years of age.
Turning again to the representative sample of 5030 Australian SMEs
surveyed by the Australian Bureau of Statistics (ABS) in the four consecu-
tive years from 1995 to 1998 (as described in Chapter 2) Table 3.2 also
presents the annual closure rates for three age groups: less than five years;
five to ten years; and over ten years. From Table 3.2 it appears that age has
an even more pronounced effect on the closure rates for the ABS sample,
compared to those reported in Table 3.1 for businesses located within a
managed shopping centre. The results presented in Table 3.2 indicate that
the average annual closure rate for Australian SMEs less than five years
old (16%) is over three times the closure rate for SMEs over ten years old
(5%). For businesses located within a managed shopping centre, however,
the average annual closure rate for businesses less than five years old
(4.3%) is less than double the closure rate for businesses over ten years
old (2.6%). It would seem reasonable to suggest, therefore, that the fact
that businesses located within a managed shopping centre are typically
screened and receive ongoing advice and support might have a significant
impact both on their overall survival rate and, more particularly, on the
survival rate of younger businesses. This finding has implications for gov-
ernment policy and highlights the important role service providers, such as
accountants, might play in potentially minimizing early stage failure rates
within the SME sector.
34 SME performance
Notes:
a. Note that in Australia the financial/tax year runs from 1 July to 30 June.
** Significantly different from overall (total) closure rate at 1% using a two-tailed test.
The findings presented in Table 3.2 emphasize the need (when reporting,
interpreting and comparing failure rates) to provide information about the
age distribution of the SMEs under investigation and to control for age in
any analysis being conducted.
Table 3.3 Australian SME closure rates 1995–96 to 1997–98 by size (%)
Notes:
* Significantly less than small business closure rate at 5% using a one-tailed test.
** Significantly less than small business closure rate at 1% using a one-tailed test.
et al. 1989), they have generally found that a size effect persists even after
controlling for age. Using the ABS data to further explore the relation-
ship between business closure and both size and age, Table 3.3 reports
separately the average annual closure rates for Australian small (20 or
fewer employees) and medium-sized businesses (more than 20 employees).
From the results presented in this table it is apparent that age of business
has a far greater impact on average annual closure rates than does size of
business (see the significance results for age reported in Table 3.2).
The results presented in Table 3.3 provide some support for Jovanovic’s
(1982) argument because for the two younger age groups there is no differ-
ence in the average annual closure rates of the small versus the medium-
sized firms. It is only in the oldest age group that there is a significant
difference in the closure rate by size of business. The fact that the overall
closure rate is significantly higher for the smaller firms is largely a function
of there being disproportionately more small firms in the youngest age
group, where the closure rate is much higher than average.
The fact that for SMEs over ten years old the closure rate appears higher
for smaller firms could be the result of closure rates being biased against
smaller firms and, therefore, the adoption of an alternative (more appropri-
ate) definition of failure could result in a finding that supports Jovanovic’s
(1982) proposition for all age groups. Similarly, the fact that many prior
studies have reported significant differences in failure rates by size of busi-
ness could also be a function of the failure definitions used by those studies.
For example, using either discontinuance of ownership or business
closure as a definition of failure is biased against smaller firms. The reason
for this is that, compared to larger businesses, smaller businesses typi-
cally have lower start-up and closure costs, and a greater dependency on
the life cycle of their owners.1 As noted by Hutchinson, Hutchinson and
36 SME performance
Table 3.4 Comparing failure rates for SMEs located in larger and smaller
shopping centres (%)
Notes: * Significantly different from average failure rate in larger shopping centres at 5%
using a two-tailed test.
Many studies examining the incidence of SME failure have reported sig-
nificant variations in failure rates between industry sectors. Furthermore,
the results from some studies are in direct conflict. For example, Lowe,
McKenna and Tibbits (1991) and Fredland and Morris (1976) examined
SME bankruptcies and reported that the manufacturing sector had the
highest failure rate, while Brüderl, Preisendörfer and Ziegler (1992) and
Phillips and Kirchoff (1989) looked at closure rates and found that the
manufacturing sector had the lowest failure rate. The significant variations
in reported failure rates and the apparent conflict between the findings of
some studies must surely be a source of confusion for policy makers and
others with an interest in the SME sector. To try to better understand the
conflicting results referred to above, I will now explore the likely associa-
tion between industry and reported failure rates for various definitions of
SME failure.
Stewart and Gallagher (1986) suggested that firms in sectors which
involve high capital costs are likely to have higher levels of liquidations
and bankruptcies and will find it harder to cease trading and simply
switch to something else. These comments suggest that using bankruptcy
as a measure of failure is likely to be biased against industries with high
start-up costs (capital requirements) and in favour of businesses with
lower start-up costs. Conversely, using closure as a definition of failure
is likely to be biased in the opposite direction, namely, in favour of busi-
nesses with high start-up costs and against businesses with low start-up
38 SME performance
High
Bankruptcy
Failure rate
Failed to ‘make
a go of it’
Business closure
Low
Figure 3.2 Expected relationship between failure rates and start-up costs
There are three types of risk that can ultimately impact a venture’s chances
of success: those that are unique to the firm (for example, the experience of
the owner); those that relate to the industry in which the firm operates (for
Effects of age, size, industry and the economy 39
example, whether the industry is in decline); and those that relate to the
state of the economy (for example, unemployment rates). There is little an
SME owner can do to protect a business from economic downturns and,
as noted by Fredland and Morris (1976, p.9), during ‘cyclical downturns
the marginal firm is more likely to fail’.
Following the call by Shailer (1989) to pay more attention to external
(exogenous) variables, such as interest rates and various other economic
indicators, Everett and Watson (1998) modelled SME failure (using
various failure definitions) to determine just how significant various
macro-economic factors are to SME mortality. Evidence from the US
reported by Sharpe (1981) suggests that market-based risk (as opposed to
firm- or industry-based risk) represents approximately 25% of the total risk
associated with listed companies. Similarly, Foster (1986, p.199) reported
that on average, in the US, external factors (industry and economy)
explain about 43% of the variation in business net income.
The results of modelling failure rates against various macro-economic
variables for my sample of 5196 business start-ups operating within 51
Australian managed shopping centres over the period 1961–90 are sum-
marized in Table 3.5. The table reports only those macro-economic vari-
ables that were found to be significant and, for each significant variable,
whether the relationship with SME failure was positive or negative. As can
be seen from the table, retail sales for the current six-month period and
retail sales lagged six months are the only macro-economic indicators to
feature in more than one model (failed to ‘make a go of it’; discontinuance
40 SME performance
There are at least two possible causes for this relationship. First, a high
unemployment rate might indicate problems in the economy which, in
turn, could lead to an increase in SME failure. Second, a high unemploy-
ment rate could result in an increase in the demand for self-employment
and, therefore, greater opportunities to sell both marginal and successful
businesses.
In summary, the findings presented by Everett and Watson (1998) indi-
cate that, on average, macro-economic factors appear to be associated
with about 30% to 50% of SME failures, depending on the definition of
failure used. As expected, failure was positively associated with interest
rates (where failure was defined as bankruptcy) and the rate of unem-
ployment (where failure was defined as discontinuance of ownership).
However, failure was also positively associated with lagged employment
rates (where failure was defined as to prevent further losses) and with retail
sales (where failure was defined as: failed to ‘make a go of it’; discontinu-
ance of ownership; or discontinuance of business). These findings are con-
sistent with Bhattacharjee et al. (2009), who reported, for a large sample
of UK companies, that bankruptcies were associated with economic
downturns and acquisitions with economic upturns.
These results suggest that many businesses are sold, or cease, volun-
tarily, with their proprietors able to time their exit to take advantage of
prevailing economic conditions. That is, some of the macro-economic
variables found to be significant in this study cannot be viewed as causing
failure, but rather they seem to provide the trigger for SME owners to take
the opportunity to sell or close their businesses. Thus, depending on the
definition of failure adopted, a positive economic outlook might be asso-
ciated with an increase in the rate of SME failure. Policy decisions made
in the absence of a sound understanding of how various macro-economic
variables are likely to impact SME failure rates (under various definitions
of failure) could, therefore, be suspect. Further, without a clear under-
standing of the relationship between key economic indicators and the
various definitions of failure, accurate evaluations of government policies
and programs designed to help SMEs will be problematic.
3.5 SUMMARY
NOTES
1. Ang (1992, p.187) noted that ‘[s]mall businesses can terminate due to the departure or
demise of a single individual or the dissolution of a partnership’.
2. Storey et al. (1987, p.42) suggested that ‘manufacturers are more likely to be placed into
liquidation’ because they are more likely to have purchased fixed assets in order to operate
their business. Also, Garrod and Miklius (1990) reported that bankruptcies represented
11% of discontinuances for manufacturers but only 6% for retailers. Similarly, Stewart and
Gallagher (1986, p.46) noted that: ‘[s]ectors which have high capital costs are likely to have
higher levels of liquidations and bankruptcies. The firms will be more likely to be in debt
and will find it harder to cease trading and simply switch to something else.’
PART III
NOTE
1. Du Rietz and Henrekson (2000) found that after controlling for industry the underper-
formance of female entrepreneurs disappeared for three variables (increased profitability;
46 SME performance
increased number of employees; and increased number of orders) but not for a fourth
(increased sales). Chell and Baines (1998), in a UK study of 104 micro-businesses, found
that after controlling for industry there was no significant difference in sales between
male- and female-owned businesses.
4. Failure rates
4.0 INTRODUCTION
In this chapter I aim to dispel the myth that female-owned SMEs are less
likely to survive (more likely to close) than male-owned SMEs. Carter,
Williams and Reynolds (1997) surveyed a sample of 203 retail firms from
two midwestern states of America in 1986 and then again in 1992. They
found that 34% of women-owned businesses but only 22% of men-owned
businesses ceased operations over the six-year period of their study. Boden
and Nucci (2000), in a large study of US sole proprietorships in the retail
and service industries that commenced operations in two different time
periods, found that the mean survival rate for male-owned businesses was
4–6% higher than for female-owned businesses. The findings of these two
studies suggest that female-owned retail and service businesses have higher
odds of closure than those owned by males.
However, Cooper et al. (1994) analysed a longitudinal study of 1053
new ventures (representative of all industry sectors and geographical
regions) in an attempt to predict the performance of new ventures based
on factors that could be observed at the time of start-up. Indicators of
initial human and financial capital were examined to determine how they
affected the probability of three possible performance outcomes: failure,
marginal survival or high growth. Cooper et al. (1994) argued that general
human capital (represented by the entrepreneur’s education, gender and
race) might reflect the extent to which the entrepreneur has had the oppor-
tunity to develop relevant skills and contacts. The results presented by
Cooper et al. (1994) suggest that while women-owned ventures are less
likely to grow, they are no more likely to close.
Kalleberg and Leicht (1991) tested several hypotheses concerning the way
the survival (and success) of small businesses headed by men and women was
related to industry differences, organizational structures and the attributes
of their owner-operators. Their analysis was based on businesses in south
central Indiana across three industries (food and drink, computer sales
and software, and health) for the period 1985–1987. As with Cooper et al.
(1994), Kalleberg and Leicht’s (1991) results also suggest that businesses
headed by women are no more likely to close than those headed by men.
47
48 SME performance
It would seem, therefore, that there are some conflicting results from
prior research into the relative performances of male- and female-owned
businesses. However, these past studies have generally been based on
limited samples (which might explain the conflicting results), making it
difficult to generalize from their findings. The closure rates presented and
discussed in this chapter are based on a large and representative data set
compiled by the Australian Bureau of Statistics (ABS) covering the four-
year period 1995–1998, as discussed in Chapter 2.
Of the 5030 SMEs selected by the ABS for an annual survey every year
for four consecutive years, there were 3046 SMEs for which the sex of
the major decision maker could be determined (this person was deemed
to control the business even though they might not have been the major
owner). Of these businesses, 2868 were male-controlled and 178 were
female-controlled. Table 4.1 presents the annual closure rates for these
SMEs.
Table 4.1 Annual closure rates for male- and female-controlled SMEs
Note: * Female closure rate significantly higher than male closure rate at 5% using a one-
tailed test.
From Table 4.1 it can be seen that there is some evidence to suggest that
the closure rate for female-controlled SMEs is higher than that for male-
controlled SMEs. At face value, this finding is consistent with Carter et
al. (1997) and Boden and Nucci (2000). However, the results presented in
Table 4.1 do not control for other potentially important variables such as
age, size and industry.
It is possible that the higher closure rate for female-controlled SMEs is the
result of key demographic differences: in particular age, size and industry.
For the female- and male-controlled SMEs in the ABS sample, Table 4.2
presents their age and size details and Table 4.3 reports the closure rates by
industry sector, together with the percentage of female-controlled SMEs
in each of those sectors.
Table 4.2 indicates, as expected, significant differences between the
male- and female-controlled SMEs in terms of both age and size of
business. In particular, the female-controlled SMEs are significantly
overrepresented in the two youngest age categories and significantly
underrepresented in the oldest age category. As discussed in Chapter 3,
SME failure rates appear to peak during the first five years and, therefore,
Table 4.2 Age and size demographics for the female- and male-controlled
SMEs
Male Female
Controlled Controlled
n = 2868 n = 178
Firm age (%)
Less than 2 years 13 22**
2 years to less than 5 years 15 22**
5 years to less than 10 years 23 21
10 years to less than 20 years 27 24
20 years and older 21 11**
Total 100 100
Firm size in 1994–95
Number of employees 27 13**
Source: ABS.
50 SME performance
exhausted – does this mean that all mining ventures are failures? Clearly
not! Further, many SMEs that were initiated to provide services to a
particular mine site will also almost certainly close when the mine closes
– are all of these ventures also failures? Again, the answer is clearly not!
As noted in Chapter 2, there is no generally accepted definition of failure
that meets the needs of all users (and for which data is readily available)
and it is important, therefore, that researchers understand (and explain)
the potential limitations with respect to the particular definition they have
adopted.
Given the industry differences noted above, it is important that indus-
try, together with age and size of business, is controlled in any comparison
of failure rates for male- and female-controlled SMEs. In the following
section I will control for age, size and industry in further examining the
closure rates reported in Table 4.1 for the 2868 male-controlled and 178
female-controlled SMEs from the ABS sample.
To examine the association between SME closure and the sex of the major
decision maker, a logistic regression analysis was undertaken incorporat-
ing age, size and industry as control variables. The results of the analysis
are reported in Table 4.4. As expected, they indicate that the effect of both
age and industry on the probability of SME closure (failure) is highly sig-
nificant. However, consistent with the arguments advanced by Jovanovic
(1982), size of the business is unrelated to its survival prospects. After
controlling for age, size and industry, the sex of the person responsible for
the major decision making within the business is also not significant, that
is, there is no association between SME closure and the sex of the major
decision maker.
4.4 SUMMARY
Note: In running the logistic regression the last category was used as the reference point
for each categorical variable and, therefore, the last category is not shown in the table. For
firm age the last category is ‘20 years or older’ and for industry the last category is ‘Personal
and other services’.
5. Relating outputs to inputs
5.0 INTRODUCTION
53
54 SME performance
Borrowing allows a firm to have an asset base that is bigger than could be
provided using just the owner’s investment (equity). This, in turn, allows
the firm to potentially earn higher profits, but it also increases the risk to
the firm. Therefore, when analysing a firm’s ROE, it is important to sepa-
rate out the effects of leverage to eliminate the potentially confounding
effects of differing debt policies and attitudes to risk. In the next chapter,
the issue of risk, when comparing the performance of male- and female-
controlled SMEs, will be explored further.
Table 5.1 Comparing the mean annual outputs and inputs for male- and
female-controlled SMEs
Table 5.2 Relating mean annual outputs to inputs in comparing male- and
female-controlled SMEs
the log of ROA and the log of ROE. For unprofitable businesses there was
no significant difference between the male- and female-controlled SMEs.
Therefore, there is some evidence to suggest, contrary to popular belief,
that female-controlled SMEs might actually outperform male-controlled
SMEs in terms of ROE and ROA.
However, while the results presented in Table 5.2 suggest that female-
controlled SMEs might outperform male-controlled SMEs, the results do
not control for a number of potentially important confounding variables
that would be expected to impact SME performance. For example, Chapter
3 demonstrated how SME performance could vary significantly across
industries and by age and size of business. Also, Fasci and Valdez (1998)
found that hours dedicated to a business on a weekly basis, a measure of
the owner’s labour input to the business (as opposed to financial input),
contributed significantly to earnings, and this is therefore another variable
that should be controlled in any evaluation of firm performance. In the
Relating outputs to inputs 57
Table 5.3 ANOVA of log ROA and log ROE and various business
demographics
following analysis of firm performance, the SMEs in the sample were clas-
sified into two groups – those operating less than five days per week and
those operating five or more days per week – and this was included as an
additional control variable (together with age and industry). Note that size
is not included as a control variable because the performance measures
themselves control for size by relating outputs to inputs.
Table 5.3 presents the results of analysing the variance (ANOVA) in Log
ROA and Log ROE, for profitable businesses only, to see if the significant
differences between the male- and female-controlled SMEs reported in
Table 5.2 remain after controlling for industry, age of business and the
number of days the business operated. As expected, industry and age
of business are highly significant in explaining the variation in perform-
ance across SMEs. Similarly, the number of days the business operated
was significant in explaining variations in Log ROA, but not Log ROE.
Presumably the reason why the number of days the business operated was
58 SME performance
5.4 SUMMARY
Again, the findings presented in this chapter clearly indicate that female-
controlled SMEs do not underperform male-controlled SMEs. Indeed,
there is some evidence to suggest that, with respect to ROA (asset man-
agement), female-controlled SMEs might outperform male-controlled
SMEs.
6. Adjusting for risk
6.0 INTRODUCTION
59
60 SME performance
profit) as the single most important number, without any explicit control
for risk. However, we know ‘there is risk in the world, and that inves-
tors generally dislike it’ (Sharpe 1975, p.29). While there is no doubting
the importance of sales and profit to a business, it is equally important
to explicitly relate these performance indicators to the underlying risks
involved in the business.
It should be noted that, although the ‘economics literature typically
defines risk as variability’ (Forlani and Mullins 2000, p.309), it can mean
different things to different people. Forlani and Mullins, in an experi-
mental study, asked a sample of entrepreneurs leading America’s fastest
growing firms to make choices among a series of hypothetical new ven-
tures. They found that while their subjects ‘tended to shun high levels
of variability . . . they appeared willing to accept a considerable degree
of hazard, or possible downside . . . presumably in pursuit of potentially
significant gains’ (2000, p.305). This finding indicates that variability is
central to the entrepreneur’s notion of risk.
Sharpe (1975) suggests that the reward-to-variability ratio (Sharpe
ratio) is an appropriate unambiguous measure of performance that
controls for risk. ‘The reward-to-variability ratio is simply the ratio of
reward (which is good) to variability (which is bad)’ (Sharpe 1975, p.29).
Other things being equal, the higher the ratio, the better the performance.
Although Sharpe discussed the reward-to-variability ratio in the context
of comparing the performances of individual securities and portfolios, it
would appear to be an appropriate measure for assessing the perform-
ance of SMEs. Taggart (1996, p.276) notes that because the Sharpe ratio
‘adjusts for total risk, it can be useful for assessing the performance of
a portfolio that is less than fully diversified’. Given that the majority of
SME owners have a less than fully diversified investment portfolio (with
the majority of their wealth being tied up in their business) the Sharpe
ratio would seem to be particularly well suited to assessing and comparing
SME performance.
In applying the Sharpe ratio, it is normal to measure ‘reward’ in terms
of stock price returns and ‘variability’ as the standard deviation in those
returns.1 However, many SMEs are not listed, and stock price information
is therefore not available. In the absence of stock price returns it seems rea-
sonable to suggest that profit might be an appropriate alternative measure
of reward and the standard deviation in profit an appropriate alternative
measure of variability (risk). In support of using profit (rather than stock
market returns) as the reward measure, the following two points should be
noted: first, for SME owners, the profit earned by their ventures is clearly
a significant reward; and second, stock prices are largely driven by profits
(particularly future expected profits).2
Adjusting for risk 61
are related to the standard deviation in those profits, using the Sharpe
ratio (Sharpe 1975), no difference exists between the male- and female-
controlled SMEs. This suggests that, after adjusting for risk, there is no
difference in the performance of male- and female-controlled SMEs. Put
another way, males take more risks and, as a result, on average their
firms earn higher profits.
The results presented in Table 6.1 do not control for industry effects,
the age of the business or the size of the business. Chapter 5 indicated
that each of these variables was potentially highly significant in explain-
ing variations in SME performance by gender. Therefore, to check
the robustness of the results presented in Table 6.1, Table 6.2 presents
the results of examining the log of average annual profit the log of the
standard deviation of annual profits and the log of the Sharpe ratio
using analysis of variance (ANOVA) to see if the sex of the person in
control of the business has any significant impact after controlling for
industry, age and size of business. The industry categories used are as
described in Chapter 3. The age categories used are: zero to less than
two years old; two to less than five years old; five to less than ten years
old; ten to less than 20 years old; and 20 or more years old. The number
of employees in the business at the time the business was first surveyed
is used to measure firm size. The SMEs were classified into the following
size groups: those with one to four employees; those with five to nine
employees; those with ten to 19 employees; and those with 20 or more
employees.3
As can be seen from Table 6.2, after controlling for industry, age and
size of business, the sex of the person in control of the business still has
an impact on the profits earned, that is, male-controlled firms typically
have higher profits than female-controlled firms. However, the table also
indicates that, after controlling for industry, age and size of business,
the sex of the person in control of the business also has an impact on
the standard deviation in annual profit, that is, male-controlled firms are
typically more risky than female-controlled firms. However, if the Sharpe
ratio is used as the basis of comparison, then the sex of the person in
control of the business has no impact. In this case industry is the only
variable associated with firm performance, but only at a 5% level of
significance.
The results presented in Table 6.2 confirm the robustness of the find-
ings reported in Table 6.1 and again support the proposition that there
is no significant difference in the performances of male- and female-
controlled SMEs, provided performance is measured appropriately. In
other words, although females may adopt different strategies in found-
ing, running and growing their businesses, they are likely to be no less
effective than males in terms of the risk-adjusted rewards earned by their
firms.
6.5 SUMMARY
Once again, the findings presented in this chapter clearly indicate that
female-controlled SMEs do not underperform male-controlled SMEs. I
trust that the evidence provided in Chapters 4, 5 and 6 will have convinced
the reader, once and for all, that provided performance is measured appro-
priately, and adequate controls are included in the analysis, male- and
female-controlled SMEs perform equally well.
Adjusting for risk 65
NOTES
1. Stock price returns are usually calculated as: (the gain in stock price over a period +
dividends during the period)/beginning of period stock price.
2. It should also be noted that the Sharpe ratio (1975) is often adjusted by the risk-free
rate to allow comparison across periods where the underlying risk-free rate differs. In
this chapter, however, no such adjustment is made because the analysis covers the same
period for all firms.
3. Although these categories are somewhat arbitrary, they are similar to those used in previ-
ous studies. Also, the definition of a medium-sized business in Australia is one with 20 or
more employees and this is captured in the last category. This variable was also entered
into the ANOVA as a covariate (continuous) variable to see if it improved the analysis,
but it did not.
PART IV
Part IV is aimed at dispelling the myth that SME growth, particularly for
female-controlled SMEs, is constrained by an inability to access appropri-
ate levels of external (bank) funding. The available literature suggests a
strong link between the availability of finance and SME growth, and this
has led to the notion of a ‘finance gap’, implying that ‘there may be major
“barriers” preventing an owner-manager’s access to equity’ (Hutchinson
1995, p.231). This notion of a ‘finance gap’ within the SME sector has been
supported by a number of researchers (see, for example, Berger and Udell
1998; Pissarides 1999; Becchetti and Trovato 2002; Carter et al. 2003). It
has also been suggested that the ‘barriers’ to finance might be even more
acute for female-owned SMEs, as there is a perception that financial insti-
tutions (banks) discriminate against female business owners (Riding and
Swift 1990; Breen, Calvert and Oliver 1995; Brush et al. 2001).
Given that SMEs are responsible for significant levels of employ-
ment, innovation and productivity, it is important that policy makers
and advisers are well-informed about the determinants of SME growth,
including the various demand-side issues surrounding the provision of
growth funding for this sector (Becchetti and Trovato 2002). Carpenter
and Petersen (2002) examined more than 1600 US small manufacturing
firms and found that the growth of these firms appeared to be constrained
by a lack of (internal) finance. Similarly, Bruno and Tyebjee (1985) found
that ventures that had received external capital achieved statistically sig-
nificantly higher sales and employment growth (compared to ventures
without external capital). With respect to women-owned businesses,
Carter and Allen (1997) noted that the availability of financial resources is
a major influence on their growth.
While there is no doubting that firms need finance to grow, it is also
the case that not all firms have the capacity, or desire, to grow. Simply
68 SME performance
69
70 SME performance
Table 7.2 Reasons SME owners might choose not to seek external
funding
Notes:
Figures do not always add up to 100% because of rounding.
** ‘Discouraged’ group significantly different from ‘successful’ group at 1% using a two-
tailed test.
owners’ desire for control is strong enough their businesses will be entirely
self-funded. Again, this indicates the importance of internal factors when
SME owners are considering growth options that require significant levels
of external funding.
Table 7.4 sets out the participants’ responses when asked to list the
reasons why a bank might refuse a loan application. The majority of
reasons provided by the participants can be grouped under two broad
headings: perceived inadequacies in the owner’s business acumen (busi-
ness plan and track record); and the perceived risks (including lack of
74 SME performance
Notes:
Figures do not always add up to 100% because of rounding.
** ‘Discouraged’ group significantly different from ‘successful’ group at 1% using a two-
tailed test.
Table 7.5 confirms just how important maintaining control was for the
majority of the focus group participants. In the ‘discouraged’ group, all
participants rated the importance of maintaining control as a ‘1’ (very
important). In the ‘successful’ group there was a little more dispersion in
the ratings, but the majority of the group still attached a very high level of
importance to maintaining control. The one exception was a female partic-
ipant who indicated that, for her, maintaining control was unimportant.
Table 7.6 reports the views of the group with respect to the risk-taking
propensities of men and women. The responses for the male and female
focus group participants are shown separately. Interestingly, over half
the participants (mainly the males) did not express a view on this issue.
The consensus of those who spoke was that women were likely to be more
conservative (risk averse) than men, although a significant number of par-
ticipants believed either that there was no difference, or that it depended
on the personality of the individual owner, rather than the sex.1
Finally, Table 7.7 sets out the key thoughts/ideas expressed by the
participants concerning the factors (issues) that might influence an SME
owner’s decision to seek external funding. Two main thoughts emerged
from this discussion. First, that growth for growth’s sake, without a growth
in profit, was not worthwhile. Second, owners who were planning to exit
76 SME performance
Notes:
Figures do not always add up to 100% because of rounding.
(1 = very important, 7 = unimportant)
Table 7.6 Comparing male and female SME owners’ attitudes to risk
Note: Where a participant spoke on more than one occasion, his/her view was only
recorded once.
the firm were unlikely to want to raise additional funding for fear of over-
capitalizing their business. In terms of funding sources, it is interesting
to note that the few negative comments made about banks (for example:
banks will only lend you money when you don’t need/want it; and banks
won’t lend to businesses they don’t understand) came from the ‘discour-
aged’ group. It is difficult to know on what basis they formed their views
(perhaps it was from their previous experiences or, alternatively, from
A qualitative analysis 77
hearing about the experiences of others), but clearly the participants in the
‘successful’ group did not share their views. However, the comment that
banks don’t lend to businesses they don’t understand raises two important
issues. First, it suggests that banks should not discount the importance of
relationship (compared to transactions-based) lending for the SME sector.
‘Relationship lending is generally associated with the collection of “soft”
information over time through relationships with the firm, the owner, and
the local community’ (Berger and Udell 2002, p.F38). By way of contrast,
transactions-based lending is ‘generally associated with the use of “hard”
information’ (Berger and Udell 2002, p.F38), such as financial ratios, and
may not be appropriate for many SMEs, particularly for non-traditional
businesses. Second, it is important that SME owners seeking external
funding ensure that they have a clearly articulated business plan that
makes it as easy as possible for a loan officer to understand the nature of
their business and the risks involved. SME owners with limited expertise in
this area should consider obtaining professional help. This is an area that
government policy makers might consider investigating further.
In summary, a number of interesting findings emerged from the focus
group sessions. First, it seems that the majority of SME owners are acutely
aware of the various risks associated with business ownership, and this is
therefore foremost in their minds when they consider the merits of seeking
external funding. Related to this notion of risk is the issue of control, the
second major theme that seemed to be at the ‘top of the mind’ of virtu-
ally all the focus group participants. It seems that many SME owners are
unlikely to consider external funding if there is a reasonable likelihood
that they could lose control of their business. The focus group results
also suggest that (compared to male owners) female owners are more
risk averse, and therefore less inclined to access external sources of funds.
Based on the results of these focus groups, it appears that the so-called
‘finance gap’ (believed by many to be a major barrier inhibiting SME
growth) is more ‘myth’ than ‘reality’. This issue is pursued further in the
next section, which looks at the findings of a survey designed to further
explore the issues raised by the focus group participants.
This second study was designed to further investigate the various demand-
side issues that arose from the focus group discussions. Based on the focus
group results and a review of the literature, a mail survey was sent to 534
SME owners. Excluding the 69 ‘dead letters’ that were returned ‘address
unknown’, the 123 usable responses represented a response rate in excess
A qualitative analysis 79
Sex of Owner
Female Male
Applied for funding in last 3 years? Yes 13 54% 55 56%
No 11 46% 43 44%
Total 24 98
Pearson c2 = 0.030, p value = 0.863
Application for funding successful? Yes 11 85% 46 88%
No 2 15% 6 12%
Total 13 52
Pearson c2 = 0.143, p value = 0.706
Note: One respondent failed to specify sex and three male respondents failed to indicate
whether their applications had been successful.
of 25%. This was a good response rate for a target group of this type (given
the length of the questionnaire – 16 pages) and, no doubt, was aided by the
offer of a A$30 payment in return for a completed questionnaire (Newby,
Watson and Woodliff 2003).
The results presented in Table 7.8 indicate that a little over 50% of
the SME owners had applied for funding in the last three years, with
the application rates about the same for both females and males. This
finding suggests that female SME owners are not being ‘discouraged’ from
applying for bank funding by perceptions of bank discrimination against
women. Further, the results indicate that about 12% of SME owners have
their applications for funding denied, again with no significant difference
between the female and male SME owners. The fact that there was no
significant difference in the rejection rates for the female and male SME
owners suggests that the banking sector does not routinely discriminate
against women. It should also be noted that the 12% rejection rate for this
Australian sample is similar to the 10% reported by Levenson and Willard
(2000) for the US and the 11% reported by Fraser (2006) for the UK.
Given that Levenson and Willard (2000, p.83) conclude that the ‘extent
of true credit rationing appears quite small’ and Fraser (2006, p.123) con-
cludes ‘that most SMEs are getting the finance they want’, this suggests
that it is also unlikely (given the similar rejection rates) that any substan-
tial supply-side ‘finance gap’ exists in Australia.
To further explore the ‘finance gap’ issue, Table 7.9 presents the
main reasons why external (debt) funding had not been sought by the
80 SME performance
Table 7.9 Reasons for not applying for external funding in the last three
years
respondents during the last three years. The relevant respondents were
given a list of items and asked to rate each using a seven-point scale
ranging from ‘strongly disagree’ (1) to ‘strongly agree’ (7).
The results presented in Table 7.9 indicate two primary reasons why
SME owners do not apply for funding from financial institutions: first,
because they have access to sufficient funding under their current arrange-
ments; and second, because they do not require additional funding. These
results are consistent with Fraser’s (2006) finding that 95% of businesses
that had not sought new finance reported that they did not need additional
funding. Interestingly, the results presented in Table 7.9 suggest that, com-
pared to the male SME owners, the female SME owners were more likely
to believe they already had sufficient funds under existing arrangements
and, therefore, had no need to apply for additional external funding. Also
of interest were the responses to the last two items in Table 7.9, which
indicate that the females strongly disagreed with the premise that the
reason they had not applied for external funding in the last three years was
because they had been rejected previously or because they were unlikely to
get the full amount requested. Taken together, and consistent with Fraser’s
(2006) finding in the UK, these results strongly suggest that female SME
owners are not discriminated against by Australian financial institutions,
A qualitative analysis 81
and they are not therefore discouraged from applying for external funding
because of any perceptions of bias against them. They might, however, be
less inclined to access external funding for other reasons, such as a desire
to maintain control and/or to minimize the risks they are exposed to.
Tables 7.8 and 7.9 suggest that financial institutions in Australia do
not discriminate against female SME owners and female SME owners
are no more likely to be discouraged from applying for external funding
than their male counterparts. Tables 7.10, 7.11 and 7.12 present further
evidence to support this conclusion.
For the female and male SME owners in the sample, Table 7.10 pro-
vides a comparison of the time taken to approve a loan, loan duration
82 SME performance
and the interest rate charged. As can be seen from the results, there are
no significant differences between the female and male SME owners with
respect to these factors. This provides further evidence that Australian
financial institutions do not routinely discriminate against female SME
owners.
Table 7.11 provides a comparison of the levels of satisfaction experi-
enced by both the females and males with respect to their lending institu-
tion in terms of issues such as time taken to process an application and
the interest rate charged. As can be seen from the results, there are no
significant differences in the average satisfaction levels for the female and
male SME owners and for all but the last item (the service fees charged)
the female and male SME owners rated their level of satisfaction above the
mid-point on the scale. This again suggests that financial institutions do
not routinely discriminate against female SME owners and that the major-
ity of SME owners (both female and male) are reasonably satisfied with
their financial institution.
Finally, Table 7.12 provides a comparison of the responses given by
both the female and male SME owners concerning their views on potential
differences in the treatment men and women might receive when apply-
ing for funding from a financial institution. These responses suggest that
neither the female nor the male respondents believed women suffer any
A qualitative analysis 83
Notes:
a Risk-taking propensity was assessed using statements taken from the risk sub-scale
of the Jackson Personality Inventory (1976).
b Internal locus of control was assessed using statements taken from Levenson
(1974).
Table 7.14 Factors discouraging SME owners from applying for funding
in the future
funding in the future. From the results presented in Table 7.14, it seems
that for all groups of SME owners (that is, females, males, non-applicants
and applicants) the prospect of unreasonable terms and conditions is likely
to be the most relevant factor deterring them from applying for external
funding in the future. The second most relevant factor for all groups is the
risk of not being able to repay the loan, with the non-applicants ranking
the relevance of this factor significantly higher than the applicants. The
potential to lose control of their business was the third most important
factor for the females and non-applicants and the fourth most important
factor for the males and applicants. Interestingly, the females ranked this
concern significantly higher than their male counterparts, suggesting,
contrary to the internal locus of control results reported in Table 7.13,
that (compared to males) female SME owners might indeed have a greater
desire to maintain control over their businesses.
It should also be noted in Table 7.14 that the two least relevant factors
likely to discourage SME owners from applying for external funding in the
future are the probability of not getting the loan and apprehension over
having their loan application rejected. These two factors were rated as the
A qualitative analysis 85
least relevant for all four groups. In summary, the evidence provided in
Tables 7.13 and 7.14 supports the proposition that it is more likely that
individual owner characteristics (such as desire to maintain control and
risk aversion) are responsible for variations in the level of external funding
across SMEs rather than these variations being the result of systematic
discrimination by financial institutions.
Finally, Table 7.15 provides the mean scores given by the SME owners
(female/male and non-applicants/applicants) when asked how they would
most likely use any surplus funds generated by their business. As can be
seen from the table, the repayment of debt was the priority for all owners,
and particularly for the female owners and the non-applicants. This indi-
cates that SME owners have an aversion to borrowing funds from finan-
cial institutions and would, therefore, prefer to repay debt rather than use
surplus funds for other purposes.
7.3 SUMMARY
NOTE
1. Note that this view is consistent with the argument advanced by Watson and Newby
(2005) that biological sex might not be an appropriate discriminator when examining
differences in the psychological attributes of SME owners. Instead, they suggest that the
use of masculine and feminine traits might prove more useful in future research.
8. A quantitative analysis
8.0 INTRODUCTION
This chapter builds on the key findings of the previous chapter suggesting
that banks do not routinely discriminate against female SME owners and
that the notion of a ‘finance gap’ might be more ‘myth’ than ‘reality’. In
particular, this chapter will contrast the implications that arise from Myers’
Pecking Order Theory (1984) with those that might be expected if, indeed,
there is a ‘finance gap’ negatively impacting growth in the SME sector.
87
88 SME performance
High
Finance gap
Relative
level of
external
funding
Low
Figure 8.1 Pecking Order Theory versus the notion of a ‘finance gap’
male-controlled SMEs, with this gap narrowing as firms get older (and
become less ‘informationally opaque’). Alternatively, if Pecking Order
Theory applies more strongly to females than males (and banks don’t
routinely discriminate against females) we would expect both male- and
female-controlled SMEs to have relatively similar debt levels in their early
years, but with a more noticeable reduction in external debt over time for
the female-controlled SMEs. These competing propositions will be tested
in section 8.5, but before that the relationship between external funding
and firm growth is discussed in the next section.
Table 8.1 Descriptive statistics for firms with high/low dtar and high/low
growth
Table 8.2 Logistic regression models for SMEs with high/low growth
Note: In running the logistic regression, the last category was used as the reference point
for each categorical variable and, therefore, is not shown in the table. For industry the last
category is ‘Personal and other services’ and for education the last category is ‘Tertiary
(business)’.
variables, except for a firm’s dtar, which is added in model 2. The results
in model 2 (consistent with the descriptive statistics provided in Table
8.1) suggest that a firm’s dtar is not a significant factor in explaining firm
growth. Running a separate analysis for the female- and male-controlled
SMEs confirmed the results for the whole group.
The logistic regression results presented in Table 8.2 indicate that, rather
than being associated with a firm’s dtar, firm growth appears to be highly
associated with a firm’s average profits and, to a lesser extent, with firm
age, industry and type of legal organization. Firm growth also appears
94 SME performance
to be associated with the owner’s sex (but only weakly), but not with the
owner’s human capital (education or experience). The positive associa-
tion between growth and average profits is expected because (other things
being equal) profitable firms will have more funds available to fund growth
opportunities. This result is also consistent with Carpenter and Petersen’s
(2002) finding that firm growth, for a sample of 1600 US small manufac-
turing firms, appeared to be constrained by the lack of internal funds. The
association between firm growth and firm age is also as expected, with
younger SMEs being more likely to be in the high-growth group.
Somewhat surprisingly, given that most of the literature suggests that
the growth of female-controlled SMEs is constrained by a lack of external
finance, the findings reported in Table 8.2 indicate that female-controlled
SMEs are more likely (than male-controlled SMEs) to be in the high-
growth group (although this finding is only significant at the 10% level).
Also somewhat unexpectedly, incorporated (limited liability) firms
appeared less likely to have a high dtar compared to unincorporated
(unlimited liability) firms. This result suggests that for unincorporated
firms the banks are probably requiring additional security over the
owner’s personal assets.
Table 8.3 presents the results of examining the mean and median dtar for
younger and older female- and male-controlled SMEs to determine the
possible impact of age on the relative debt levels of these two groups. The
results in this table indicate that while male-controlled SMEs have a sig-
nificantly higher dtar, overall, there is no significant difference in the dtar
Table 8.4 Four-year growth rates for female- and male-controlled SMEs
by age
for the younger male- and female-controlled SMEs. This finding suggests
that banks do not routinely discriminate against females but, rather, it is
likely that female SME owners prefer to have lower debt levels and will,
therefore, take the opportunity to repay debt over time from the cash flows
generated by their firms. This conclusion is consistent with the qualitative
findings presented in Chapter 7.
Table 8.4 presents the results of examining mean and median four-year
growth rates for younger and older female- and male-controlled SMEs to
determine the possible impact of firm age on firm growth for these two
groups. The results indicate that, for all firms, there is no significant differ-
ence in the growth rates for female- and male-controlled SMEs. However,
for older firms, it would appear that female-controlled SMEs experience
higher growth rates than male-controlled SMEs. Further, this result
appears to be driven by a decline in growth rates for older male-controlled
SMEs (which was significant at p = 0.05, not reported) rather than an
increase in growth rates for older female-controlled SMEs (which was not
significant).
This finding of a decline in growth rates over time for male-controlled
SMEs is consistent with most of the previous literature suggesting a
negative relationship between firm growth and age (Becchetti and Trovato
2002). However, the finding of no significant change in growth rates over
time for the female-controlled SMEs is also consistent with Cliff’s (1998)
conclusion that female SME owners prefer steady rather than fast-paced
growth. It is also possible that many female SME owners are unable to
devote the necessary time and energy required to achieve rapid growth in
the early years of their business because of family responsibilities. As these
diminish, female SME owners are more likely to have the time, energy and
financial resources needed to develop and expand their ventures.
96 SME performance
8.6 SUMMARY
The findings reported in this chapter, and Chapter 7, indicate that Pecking
Order Theory, rather than bank discrimination, might provide a better
explanation for any observed differences in the level of external funding
between male- and female-controlled SMEs. Because female SME owners
are, on average, more risk averse and have a greater need to feel in control
of their businesses (than male SME owners) they will be less inclined to
access external funding. The fact that the relatively lower levels of external
funding in female-controlled SMEs is most noticeable in older firms (with
established track records) is consistent with the proposition that lower
levels of external funding in female-controlled SMEs is the result of per-
sonal choice (Pecking Order Theory) rather than bank discrimination.
Contrary to some prior research, the findings reported in this chapter
indicate that growth is not significantly associated with a firm’s relative
level of external (bank) funding, but is associated with a number of other
firm-level variables, in particular, firm profitability. Further, the findings
suggest that there is no significant difference in the overall growth rates for
female- and male-controlled SMEs in Australia. However, for older firms,
female-controlled SMEs appear to have significantly higher growth rates
than male-controlled SMEs. For younger firms, there appears to be no
difference in the growth rates for the female- and male-controlled SMEs.
This result is somewhat puzzling, given that the older female-controlled
SMEs also had relatively lower levels of external (bank) debt available to
fund firm growth. The implication is that, compared to male-controlled
SMEs, female-controlled SMEs have more internally generated funds.
Given that female- and male-controlled firms have similar levels of relative
profitability (see Chapter 5), the higher level of internally generated funds
available to female-controlled SMEs can only be the result of female SME
owners withdrawing less funds from their businesses (for example, in the
form of wages, fees, dividends or drawings). As the ABS data does not
include any information on amounts paid to (withdrawn by) the owner,
this is an issue that future research might usefully investigate.
Finally, while the findings reported in this and the previous chapter are
limited to external (bank) debt, it is conceivable that they could equally
apply to external equity funding. That is, female SME owners might be
less inclined to access external equity funding because of the risks involved
and the potential loss of control that might follow.
Given that there has only been a limited amount of research focusing
‘on small, growing entrepreneurial companies and the factors affecting the
capital structure of these firms’ (Michaelas, Chittenden and Poutziouris
1999, p.114), particularly female-controlled SMEs (Brush et al. 2001), the
A quantitative analysis 97
results provided in these two chapters should help to better inform finance
providers, business advisers and policy makers about the determinants of
SME growth and the various demand-side issues surrounding the provi-
sion of growth funding for this sector.
I trust the material presented in this and the previous chapter will have
convinced the reader that banks do not routinely discriminate against
female SME owners and that the notion of a ‘finance gap’ might be more
‘myth’ than ‘reality’.
NOTE
1. Note that Cressy (1995) argues that for owners to experience a loss of control does not
require the bank to have an equity stake in the firm; it might, for example, take the form
of monitoring activities imposed on the firm.
PART V
While there are many factors that can influence the success of a venture
and there are various risk reduction strategies that can be employed to
increase a firm’s chances of survival (see, for example, Duchesneau and
Gartner 1990; Cooper 1993; Cooper et al. 1994; Robson and Bennett
2000; Shepherd, Ettenson and Crouch 2000; Larsson, Hedelin and Garling
2003), only recently have researchers begun to highlight the potential
significance to SME performance of an owner-manager’s networking
involvement. Coleman (1988) notes that information is important to deci-
sion making but is costly to obtain and that networks provide a means
by which important information can potentially be acquired in a cost-
effective manner. Therefore, networking can enhance an SME owner’s
social capital (Coleman 1988) because it provides access to information
embedded within the networks accessed. Further, Granovetter (1983)
argues that individuals whose networks (and, therefore, main source of
information) comprise primarily family and friends (strong ties) are likely
to have access to less information than individuals whose networks include
many acquaintances (weak ties). Presumably for this reason, Fischer and
Reuber (2003) suggest that owners of high-growth firms need to develop
ties beyond their personal circle of contacts and local communities.
Similarly, innovation theory suggests that networks (particularly those
comprised of many weak ties) are important in diffusing innovations and,
therefore, SMEs whose owners are heavily involved in networking should
outperform SMEs whose owners make limited (or no) use of networks
(Havnes and Senneseth 2001).
In support of the foregoing propositions (and despite Aldrich and Reese
(1993) and Cooper et al. (1994) being unable to find a significant relation-
ship between networking and firm performance), there have been a limited
number of studies that have documented a positive association between
networking and various aspects of firm performance. For example,
Duchesneau and Gartner (1990) found that successful firms were more
likely to have used professional advice. Potts (1977) noted that successful
101
102 SME performance
Seibert, Kraimer and Liden (2001) provide a useful summary and discus-
sion of the three conceptualizations of social capital found in the litera-
ture. First, there is weak tie theory as proposed by Granovetter (1973).
Here the focus is on the strength of social ties and it is argued that net-
works comprising strong ties (such as family and friends) are more likely
to be the source of redundant information than would be the case where
networks comprise weak ties (such as acquaintances). Second, is Burt’s
(1992) notion of structural holes. A structural hole is deemed to exist
where two individuals are not connected in any way. Here the focus is not
on the direct ties between SME owners and individual members of their
network but, rather, on the relationships between the various members in
an SME owner’s network. An SME owner whose network contains many
structural holes (that is, few of the other members of the network are con-
nected) is likely to have ‘more unique and timely access to information’
(Seibert et al. 2001, p.221). Third, is social resource theory (Lin, Ensel
and Vaughn 1981), which focuses on the nature of the resources embed-
ded within a network, rather than on the strength of ties or the existence
of structural holes. So, while weak tie theory and structural hole theory
examine the links between members of a network, social resource theory
is concerned with the nature of information (social resources) held by
individual members of the network.
A variety of terms can be found in the literature to describe the
Networking and performance 103
however, an SME owner has only recently started in business, and survival
is of utmost importance, the owner might require more intense help from
a reduced set of networks, particularly from formal network sources such
as external accountants and lawyers.
Again, the analysis presented in this (and the following) chapter is based
on the data set compiled by the Australian Bureau of Statistics (ABS), as
described in Chapter 2. The ABS data from the 1995–96 (second) survey
contained information relating to the frequency (never – coded 0; between
one and three times – coded 1; or more than three times – coded 2) with
which owners had sought advice during the year from seven formal sources
(banks; business consultants; external accountants; industry associations;
the Small Business Development Corporation (SBDC); solicitors/lawyers;
and the tax office) and three informal sources (family and friends; local
businesses; and others in the industry). There were 5027 responses to the
1995–96 (second) survey. However, on examining the data, it was found
that 13 businesses had no income (sales or other income). Therefore, they
were excluded from the analysis on the assumption that they were not
active businesses. This left 5014 firms (representing approximately 1.25%
Table 9.1 Range and intensity of formal and informal network access (%)
1.1
1.0
Probability of Survival
0.9
0.8
0.7
Linear
Quadratic
0.6
–10 0 10 20 30
Networking Score
Figure 9.1 depicts the results of estimating the relationship between firm
survival and networking using both a linear and quadratic model. The
networking score variable depicted in Figure 9.1 can vary from 0 to 20 and
is the product of network range (which can vary from 0 to 10, based on
the total number of formal and informal networks) and network intensity
(which can vary from 0 to 2, based on the frequency of network contact –
with no contact coded 0, contact between one and three times coded 1 and
contact more than three times coded 2).
The figure indicates that an inverted U-shaped function might indeed
best represent the relationship between firm survival and networking.
Watson (2007) notes that the probability of firm survival peaks when
the SME owner is involved in about six networks beyond this level the
probability of survival declines.
Similarly, Figures 9.2 and 9.3 depict the results of estimating the rela-
tionship between networking and both income growth and return on
equity (ROE), respectively. To assess the relationship between network-
ing and both firm growth and ROE, the analysis focuses on firms in the
upper and lower quartiles for these two performance measures. Firms in
Networking and performance 107
0.58
0.56
0.54
Probability of High Growth
0.52
0.50
0.48
0.46
0.44
0.42 Linear
Quadratic
0.40
–10 0 10 20 30
Networking score
Source: ABS.
the upper quartile are coded 1 and those in the lower quartile are coded 0.
It should be noted that, unlike the analysis of firm survival, the analysis
for growth and ROE is restricted to only those firms that survived to the
last year of the ABS’s four-year longitudinal study. The results depicted
in Figure 9.2 indicate that the same inverted U-shaped relationship that
applies to survival and networking also applies to growth and networking.
However, as can be seen from Figure 9.3, the same cannot be said for ROE
and networking.
The findings presented in Figures 9.1 and 9.2 suggest that both the sur-
vival and growth of SMEs can be enhanced by owners being involved, up
to a limit, in a range of networks. However, the same relationship does
not appear to exist between ROE and networking. It seems that the costs
involved with networking (particularly in terms of the SME owner’s time)
might have a negative impact on overall firm profitability. This issue will
be examined further in the following analysis.
Given the findings above, the SME owners’ networking scores were
entered into logistic regression models as both first and second order vari-
ables. Table 9.2 provides the results of examining the relationship between
108 SME performance
0.54
0.52
Probability of High ROE
0.50
0.48
0.46
0.44
Linear
Quadratic
0.42
–10 0 10 20 30
Networking score
Source: ABS.
firm performance (survival, growth and ROE) and the level of networking
activity undertaken by SME owners. In the first model, only the demo-
graphic variables (age, industry and size) are included. In the second
model, networking is added. As can be seen from the table, the first order
networking variable is significantly positively related to the probability of
firm survival and, to a lesser extent, growth, but not ROE. The results also
indicate that the second order networking variable is significantly nega-
tively associated with survival and growth, but not ROE. These results
add further support to the proposition that the relationship between SME
performance and networking resembles an inverted U-shaped function
for both survival and growth, but not ROE (where there was simply no
significant relationship between networking and performance). Perhaps
the additional revenues gained through networking help the firm survive
and grow but any additional profit earned is offset by the additional costs
involved with networking (both time and financial).
In Table 9.3, the overall networking score variable is replaced by its
various constituent parts. First, the overall networking score is broken
down into a formal and informal network variable (model 3) and, second,
Table 9.2 Logistic regression models of survival, growth and ROE against networking
109
10
10 years to less 0.05 1.05 0.10 1.10 0.22 1.24 0.23 1.26 0.37 1.45** 0.36 1.43*
than 20
Industry * * ** **
Mining −0.24 0.79 −0.20 0.82 −0.47 0.63 −0.43 0.65 −1.02 0.36 −1.06 0.35*
Manufacturing 0.02 1.02 −0.04 0.96 −0.25 0.78 −0.28 0.76 −0.70 0.50* −0.70 0.50*
Construction 0.28 1.33 0.14 1.15 0.45 1.57 0.44 1.55 −0.15 0.86 −0.14 0.87
Wholesale trade 0.45 1.56 0.38 1.46 −0.18 0.83 −0.21 0.81 −0.56 0.57 −0.55 0.58
Retail trade 0.11 1.12 0.06 1.06 −0.22 0.81 −0.23 0.80 −0.63 0.53* −0.64 0.53*
Accom., cafes, −0.10 0.90 −0.11 0.90 −0.66 0.52 −0.66 0.52 −0.69 0.50 −0.71 0.49*
rest’s
Table 9.2 (continued)
110
serv.
Firm size 0.00 1.00 0.00 1.00 0.00 1.00 0.00 1.00 0.00 1.00 0.00 1.00
Networking 0.37 1.45** 0.10 1.10** −0.02 0.98
Networking2 −0.02 0.98** −0.01 1.00* 0.00 1.00
Nagelkerke R Square 0.327 0.392 0.032 0.039 0.024 0.027
Notes:
* Significant at 5%; ** Significant at 1%.
In running the logistic regression, the last category was used as the reference point for each categorical variable and, therefore, the last category is
not shown in the table. For firm age the last category is ‘20 years or older’ and for industry the last category is ‘Personal and other services’.
111
than 5
5 years to less −0.27 0.76 −0.26 0.77 0.30 1.35* 0.27 1.31* 0.33 1.39* 0.33 1.39*
than 10
10 years to less 0.09 1.09 0.07 1.08 0.23 1.25 0.22 1.25 0.36 1.43* 0.36 1.43*
than 20
Industry ** ** ** **
Mining −0.25 0.78 −0.22 0.81 −0.46 0.63 −0.42 0.66 −1.06 0.35* −1.06 0.35*
Manufacturing −0.08 0.92 −0.03 0.97 −0.29 0.75 −0.29 0.75 −0.70 0.50* −0.69 0.50*
Construction 0.11 1.12 0.06 1.07 0.45 1.56 0.43 1.54 −0.15 0.86 −0.13 0.88
Wholesale 0.33 1.39 0.40 1.50 −0.23 0.80 −0.23 0.80 −0.55 0.58 −0.56 0.57
trade
Table 9.3 (continued)
112
storage
Finance & −0.24 0.79 −0.25 0.78 −0.11 0.90 −0.11 0.90 −0.51 0.60 −0.52 0.60
insurance
Property & 0.11 1.11 0.10 1.10 0.02 1.02 0.02 1.02 −0.32 0.73 −0.33 0.72
bus. serv.
Cultural & rec. 0.19 1.21 0.23 1.26 −0.15 0.86 −0.17 0.84 −1.22 0.29** −1.23 0.29**
serv.
Firm size 0.00 1.00 0.00 1.00 0.00 1.00 0.00 1.00 0.00 1.00 0.00 1.00
Formal networks 0.46 1.59** 0.13 1.14** −0.03 0.98
Formal networks2 −0.03 0.97** −0.01 0.99* 0.00 1.00
Informal 0.19 1.20* −0.07 0.94 0.01 1.01
networks
Informal −0.02 0.98 0.01 1.01 −0.01 1.00
networks2
Network range 0.04 1.04 0.23 1.26** −0.06 0.95
Network range2 0.00 1.00 −0.02 0.98* 0.00 1.00
Network intensity 2.19 8.92** −0.25 0.78 −0.07 0.93
Network intensity2 −0.65 0.52** 0.09 1.09 0.06 1.06
113
Nagelkerke R 0.392 0.406 0.041 0.041 0.027 0.027
Square
Notes:
* Significant at 5%; ** Significant at 1%.
In running the logistic regression, the last category was used as the reference point for each categorical variable and, therefore, the last category is
not shown in the table. For firm age the last category is ‘20 years or older’ and for industry the last category is ‘Personal and other services’.
into a network range and intensity variable (model 4). Consistent with
expectations, the results in Table 9.3 indicate that a firm’s survival and
growth (but not ROE) are more strongly associated with an owner’s
involvement in formal rather than informal networks. This finding sup-
ports the argument that weak ties are likely to be more important than
strong ties in the dissemination of information and, therefore, firm
performance (Granovetter 1983). Also, as expected, firm survival was
significantly associated with network intensity (but not network range),
while firm growth was significantly associated with network range (but not
network intensity).
The results presented in Table 9.3 indicate that SME owners need to be
strategic in terms of the nature of their networking involvement. If growth
is of paramount concern, SME owners would be well advised to consider
developing a broad range of networks, although the significance of the
second order network range variable suggests that there is a limit beyond
which further networking involvement is likely to be counterproductive.
Alternatively, if survival is of paramount concern (as might be the case
early in the life of a new venture), SME owners would be well advised
to develop closer ties with a smaller range of networks. Again, however,
SME owners should carefully monitor the time and cost associated with
networking because the results indicate that very high levels of network
intensity can be counterproductive and are unlikely to benefit overall
profitability (ROE).
9.3 SUMMARY
The findings reported in this chapter indicate that (after allowing for
age, industry and size of business) networking appears to be significantly
positively associated with firm survival and, to a lesser extent, growth
(consistent with the results of Brüderl and Preisendörfer 1998 for newly
established firms). This finding confirms the importance of social capital
in providing SME owners with information critical to the success of their
ventures. However, there appears to be no significant association between
networking and ROE (profitability). Further, the findings with respect
to both survival and growth suggest that there might be some optimum
level of resources (both time and financial) that an owner should allocate
to networking. For example, accessing more than six networks during a
year is likely to be counterproductive. Similarly, accessing any individual
network on more than three occasions during a year is also likely to be
counterproductive. Therefore, given that business failure generally results
in heavy personal loss (Bannock 1981), owners need to seriously consider
Networking and performance 115
the range and intensity with which they access various potential networks
(formal and informal).
The results also indicate that both formal and informal networks are
associated with firm survival, but that only formal networks are associated
with growth (and neither formal nor informal networks are associated
with ROE). The finding with respect to formal networks highlights the
particular importance of weak ties (Granovetter 1983) in building an SME
owner’s social capital.
Further, the results show that network intensity is more critical to firm
survival than network range. Conversely, network range is more critical to
firm growth than network intensity, again confirming the importance of
weak ties (Granovetter 1983) in disseminating information, and provid-
ing support for the assertion by Fischer and Reuber (2003) that owners of
rapid-growth firms should be interested in (and should support) govern-
ment policy aimed at developing a network-based approach to facilitating
firm growth.
Having explored the relationship between networking and SME per-
formance for all firms, the following chapter will specifically look at
possible networking differences in male- and female-controlled SMEs.
10. Networking: comparing female-
and male-controlled SMEs
10.0 INTRODUCTION
Cromie and Birley (1992) argue that because the majority of women enter
self-employment from a domestic and/or non-managerial background it is
likely that their personal network contacts will not be as extensive, or well
developed, as their male counterparts. Similarly, Munch, McPherson and
Smith-Lovin (1997) note that housework and childrearing are extremely
lonely forms of work and this isolation results in many women having
limited network contacts compared to men. Even where women move
directly from paid employment into self-employment, it is likely they will
have fewer network contacts because females typically occupy lower level
positions within the organizations they leave, compared to the typical
male (Cromie and Birley 1992).
Aldrich (1989) also argues that past research, and much of the literature,
indicates that female entrepreneurs might not only have fewer networks
than their male counterparts, but they are likely to be embedded in differ-
ent types of networks. For example, Munch et al. (1997) suggest that as a
116
Networking: female- and male-controlled SMEs 117
Note: Chi-Square test comparing males and females not significant at 5%.
Source: ABS.
of females) accessing five or more networks during the year. This finding
is consistent with Cooper, Woo and Dunkelberg (1989) and Robson and
Bennett (2000), who reported that entrepreneurs sought information
from a variety of different sources. However, the results also indicate no
significant differences between the male and female owners in terms of the
number of networks they accessed during the year. This result appears at
odds with most of the literature on gender and networking but supports
Cromie and Birley’s (1992) finding that the personal contact networks of
women are just as diverse as those of men. Note that a separate analysis of
the sub-set of SMEs that accessed three or fewer networks during the year
also failed to find any gender difference and the same applied to the sub-set
of SMEs that accessed seven or more networks during the year.
Table 10.2 provides a summary of the frequency with which both the
male and female SME owners made contact with a variety of formal and
informal networks during the year. As expected, the male SME owners,
on average, made significantly more frequent contact with formal network
sources, particularly with banks, business consultants, industry associa-
tions and solicitors. Unexpectedly, the female SME owners (on average)
did not make significantly more frequent contact with informal network
sources in total, although they did make significantly more frequent contact
Networking: female- and male-controlled SMEs 119
Table 10.2 Frequency of formal and informal network contact for male
and female SME owners (%)
Note: *, ** Chi-Square test significantly different for males and females at 5% and 1%,
respectively.
Source: ABS.
with family and friends. These findings are consistent with Robson, Jack
and Freel (2008), who reported that male Scottish business owners were
significantly more likely to seek advice from consultants and chambers of
commerce, while female Scottish business owners were significantly more
likely to turn to friends and relatives. Shaw, Lam and Carter (2008) also
reported that female owners were significantly more likely (than male
owners) to identify a family member as their prime network contact.
When the overall frequency of contact with all network sources is
examined, the result is again contrary to expectations as there is no sig-
nificant difference between the male and female SME owners. This result,
although inconsistent with the majority of the literature, again confirms
Cromie and Birley’s (1992) finding that females are just as active in their
networking relationships as men. As noted by Cromie and Birley, once in
120 SME performance
Tables 10.3 and 10.4 present the results of modelling the relationship
between a firm’s networking score and its chances of surviving and achiev-
ing high growth, respectively. Note from Chapter 9 that a firm’s network-
ing score can range from zero (if no networks had been accessed during
the year) to 20 (if all ten networks had been accessed on more than three
occasions during the year). Growth is measured as the percentage increase
in total income over the three-year period being examined. Surviving firms
are coded 1, while discontinued firms are coded 0. In terms of growth, the
analysis presented in this chapter focuses on those firms in the top 25%
(upper quartile – coded 1) compared to those in the bottom 25% (lower
quartile – coded 0).
The results in Table 10.3 indicate a significant positive relationship
between networking and firm survival (and a negative relationship between
age of business and firm survival). Similarly, the results in Table 10.4 indi-
cate a significant positive relationship between networking and firm growth
(with younger businesses also more likely to achieve high growth). Note
that, consistent with previous studies that have incorporated appropriate
Networking: female- and male-controlled SMEs 121
Note: In running the logistic regression, the last category was used as the reference point
for each categorical variable and, therefore, the last category is not shown in the table. For
education the last category is ‘Tertiary (business)’ for industry the last category is ‘Personal
and other services’; and for firm age the last category is ‘20 years or older’.
Source: ABS.
122 SME performance
Note: In running the logistic regression, the last category was used as the reference point
for each categorical variable and, therefore, the last category is not shown in the table. For
education the last category is ‘Tertiary (business)’ for industry the last category is ‘Personal
and other services’; and for firm age the last category is ‘20 years or older’.
Source: ABS.
Table 10.5 Modelling firm survival and individual network contact for
male- and female-controlled SMEs
Note: In running the logistic regression, the last category was used as the reference point
for each categorical variable and, therefore, the last category is not shown in the table. For
firm age the last category is ‘20 years or older’ and for each network the last category is
‘More than 3 times’.
Source: ABS.
Networking: female- and male-controlled SMEs 125
Table 10.6 Modeling firm growth and individual network contact for
male- and female-controlled SMEs
Note: In running the logistic regression, the last category was used as the reference point
for each categorical variable and, therefore, the last category is not shown in the table. For
firm age the last category is ‘20 years or older’; for industry the last category is ‘Personal
and other services, and for each network the last category is ‘More than 3 times’.
Source: ABS.
126 SME performance
year compared to only accessing this source on between one and three
occasions. This finding suggests that there might be some optimal level of
networking with external accountants beyond which there is no additional
benefit to be gained (but nor is there any evidence that more frequent
contact does any harm). The only other formal network that showed up
in the models was industry association, but only for the male-controlled
SMEs. As was the case with external accountants, it would seem that pro-
vided male SME owners access industry associations on between one and
three occasions during the year there is no additional benefit to accessing
this network source more frequently.
The results with respect to the use of informal networks were also quite
interesting, with the males benefiting from networking with others in the
industry and the females from family and friends. In this case, however,
the results strongly suggest that that excessive networking might be coun-
terproductive. For both the male- and female-controlled SMEs, it would
appear that accessing informal networks (others in the industry for males
and family and friends for females) on between one and three occasions
during the year is significantly more likely to be associated with firm
survival than accessing such sources more frequently (or not at all). This
finding suggests that the association between accessing informal networks
and firm survival resembles an inverted U-shaped function for both male
and female SME owners.
In summary, the final model for predicting the survival of male-
controlled SMEs incorporates (along with the age of the business)
both formal networks (external accountants and industry associations)
and informal (others in the industry). Accessing other networks (the
Australian tax office, banks, business consultants, family and friends,
local businesses, the SBDC and solicitors) does not add significantly to the
explanatory power of the model. Similarly, the final model for predicting
the survival of female-controlled SMEs incorporates (along with the age
of the business) both formal networks (external accountants) and informal
(family and friends).
Consistent with Granovetter’s (1973) weak tie theory and Burt’s (1992)
notion of structural holes, for both male- and female-controlled SMEs,
the results show a stronger relationship between survival and formal
network sources than between survival and informal network sources,
although both networking sources are clearly important. This result is
contrary to Brüderl and Preisendörfer’s (1998) finding that strong ties are
more important than weak ties in explaining firm survival. However, the
results support the suggestion by Uzzi (1996) that networks consisting of
a balance of both weak and strong ties might ultimately be more valu-
able than networks that are focused on only weak or only strong ties. It
Networking: female- and male-controlled SMEs 127
should also be noted that the model for predicting the survival of female-
controlled SMEs appears to be superior to that for male-controlled SMEs
(in terms of the Nagelkerke R Square value).
Table 10.6 provides the results of undertaking a similar analysis using
sales growth as the dependent variable. In terms of formal networks,
male-controlled high-growth SMEs appear to gain some advantage from
accessing both external accountants and industry associations. However,
the results again indicate, with respect to external accountants, that there
might be some optimum level of networking beyond which there is no
further benefit to be gained and, in the case of industry associations,
excessive networking (more than three times during a year) might be
counterproductive. That is, there is no difference (in terms of firm growth)
between accessing external accountants one to three times during the year
and more often. However, accessing industry associations between one
and three times during the year appears to be significantly more beneficial
than accessing this source more often (or not at all). This suggests that, for
growth-oriented male-controlled SMEs, accessing both external account-
ants and industry associations up to three times during a year might be
an optimal strategy; any further interaction with these formal network
sources is likely to be counterproductive (particularly with respect to
networking with industry associations).
For the female-controlled SMEs, the results indicate that accessing
an external accountant for advice on more than three occasions during
the year is significantly associated with high sales growth compared to
never accessing this source, or only doing so one to three times. Beyond
noting that accessing an external accountant on more than three occasions
during the year appears beneficial, it is not possible to indicate what the
optimum level of networking with external accountants might be for high-
growth female-controlled SMEs. These results suggest that while male
SME owners make effective use of both external accountants and industry
associations, female SME owners tend to rely more heavily on exter-
nal accountants (possibly because of problems associated with access-
ing industry associations which typically meet after hours, or perhaps
because they see little value in accessing this network). Interestingly, no
informal networks (which typically consist of stronger ties and fewer
structural holes) appear to be related to firm growth for either the male- or
female-controlled SMEs.
In summary, the final model for predicting high-growth male-controlled
SMEs incorporates (along with age and industry) two formal networks
(external accountants and industry associations) but no informal net-
works. The final model for predicting high-growth female-controlled
SMEs incorporates only one formal network (external accountants)
128 SME performance
10.3 SUMMARY
Conclusions
11. Conclusions, implications and
areas for future research
11.0 INTRODUCTION
When I agreed to write this book I had two primary motivations. First,
was to summarize the key findings from the research I have been involved
with over the past 20 years. Second, drawing on those findings and the
work of other scholars, to try and dispel a number of myths that have been
allowed to perpetuate in ‘the absence of good statistical evidence’ (Scott
and Lewis 1984, p.49). Without such evidence, there is the risk that these
myths will get ‘reported by the media, perpetuated by spokespeople for
the industry and subsequently accepted by the wider public’ (Stanworth
1995, p.59). Further, policy decisions by governments and others with an
interest in SMEs are likely to be suspect if they are based on such misper-
ceptions. For instance, the assumed high risk of failure within the SME
sector has been cited as justification for the high rates of return demanded
from this sector by bankers and venture capitalists (Phillips and Kirchoff
1989).
133
134 SME performance
Further, the SME literature suggests that the goals and expectations of
owner-operators impact on how they evaluate their firm’s performance.
For example, Buttner and Moore (1997, p.34) discovered that female
small-business owners measure success in terms of ‘self-fulfilment and
goal achievement. Profits and business growth, while important, were less
substantial measures of their success.’
We need to also recognize that each SME owner is likely to have a unique
set of goals related to his/her individual situation (Naffziger, Hornsby and
Kuratko 1994) and, consequently, it could be argued that the perform-
ance of an SME can only be appropriately assessed based on the extent to
which those specific goals have been (are being) met (Murphy et al. 1996).
This clearly is an area that future research could usefully explore.
Having examined the various definitions of SME failure found in the
literature and noted that reported failure rates vary considerably across
Conclusions 135
studies, even where the same failure definition is used, Chapter 3 then con-
sidered some of the more likely reasons for such variations. In particular,
the relationship between the various failure definitions and firm age, size,
industry and the state of the economy were examined.
In terms of age, the available evidence presents a compelling case to
suggest that failure rates peak at around three years of age, irrespective of
the failure definition being used. It is important, therefore, when compar-
ing and analysing SME failure rates, that researchers control for the age
of the business.
In terms of the size of the business, the results indicate that some failure
definitions can be biased either for or against larger or smaller businesses.
In particular, bankruptcy appears to be biased against larger businesses,
while discontinuance is biased against smaller firms. This suggests that,
when comparing and analysing SME failure rates, researchers should
control for both age and size of business and should make clear any limi-
tations associated with the failure definition being used. The fact that the
majority of past studies have used discontinuance (which is biased against
SMEs) as their definition of failure is the most likely reason why the myth
that SMEs have unacceptably high failure rates (compared to large busi-
nesses) has become established as part of the folklore on this subject.
Similarly, it would appear that some failure definitions are likely to be
biased either for or against certain industry sectors. In particular, indus-
tries with significant start-up costs are likely to report higher bankruptcy
rates but lower discontinuance rates. Conversely, industries with relatively
small start-up costs are likely to report lower bankruptcy rates but higher
discontinuance rates.
Finally, in terms of the effect of macro-economic factors on the rate of
SME failure, the evidence again suggests some potentially confounding
signals depending on the definition of failure being used. For example, and
as would be expected, the rate of SME bankruptcies appears to be posi-
tively related to interest rates. However, for all other failure definitions,
the evidence suggests that improvements in the economy can provide the
trigger for SME owners to move out of self-employment (resulting in an
apparent increase in failure rates). This indicates that policy makers need
to exercise considerable care, for example, in assessing the impact (on
SME failure rates) of any measures introduced to stimulate the economy
as a means of promoting business, particularly in the SME sector.
Part III was devoted to a comparison of male- and female-controlled
SMEs in terms of: failure rates (Chapter 4); various return measures
(Chapter 5); and a risk-adjusted measure (Chapter 6). In each case (and
contrary to popular belief), after controlling for key demographic differ-
ences (such as age and industry), there was no evidence to suggest that
136 SME performance
was particularly true of the association between firm survival and the use
of certain informal networks (others in the industry for male-controlled
SMEs and family and friends for female-controlled SMEs).
In summary, although SME owners appear to access a number of dif-
ferent networks, few of these networks appear to be associated with firm
performance (survival or growth). The only networks to show up as being
significantly associated with firm performance are: external accountants
(for firm survival and growth, for both male- and female-controlled SMEs);
industry associations (for the survival and growth of male-controlled
SMEs); others in the industry (for the survival of male-controlled SMEs);
and family and friends (for the survival of female-controlled SMEs).
11.3 CONCLUSION
I trust that the material I have presented in this book will help dispel a
number of myths relating to SME performance that have been allowed
to perpetuate in the absence of well-constructed research. In particular,
I hope I have been able to convince the reader that: SMEs do not suffer
from excessive failure rates; female-owned SMEs do not underperform
male-owned SMEs (when appropriate performance measures are adopted
and key control variables are incorporated into the analysis); female SME
owners do not find it more difficult than male SME owners to access exter-
nal funding; SME growth is not limited by a lack of external funding; and
female SME owners are not disadvantaged, relative to male SME owners,
in terms of their networking activities.
References
Aldrich, H. (1989). Networking Among Women Entrepreneurs, in
Women-Owned Businesses. Ed. O. Hagan, C. Rivchun and D. Sexton,
New York: Praeger.
Aldrich, H. E. and Reese, P. R. (1993). Does Networking Pay Off? A
Panel Study of Entrepreneurs in the Research Triangle, in Frontiers
of Entrepreneurship Research 1993. Ed. N. C. Churchill, S. Birley, J.
Doutriaux, E. J. Gatewood, F. S. Hoy and W. E. Wetzal, Babson Park,
MA: Wellesley. 325–339.
Ang, J. S. (1992). On the Theory of Finance for Privately Held Firms. The
Journal of Small Business Finance, 1(3), 185–203.
Anna, A. L., Chandler, G. N., Jansen, E. and Mero, N. P. (2000). Women
Business Owners in Traditional and Non-Traditional Industries. Journal
of Business Venturing, 15(3), 279–303.
Baldwin, J. R. and Gorecki, P. K. (1991). Firm Entry and Exit in
the Canadian Manufacturing Sector, 1970–1982. Canadian Journal of
Economics, 24(2), 300–323.
Ballantine, J. W., Cleveland, F. W. and Koeller, T. C. (1993). Profitability,
Uncertainty, and Firm Size. Small Business Economics, 5(2), 87–100.
Bannock, G. (1981). The Economics of Small Firms: Return from the
Wilderness. Oxford, UK: Basil Blackwell.
Barber, B. M. and Odean, T. (2001). Boys Will Be Boys: Gender,
Overconfidence, and Common Stock Investment. The Quarterly Journal
of Economics, 116(1), 261–292.
Barton, S. L. and Matthews, C. H. (1989). Small Firm Financing:
Implications From a Strategic Management Perspective. Journal of
Small Business Management, 27(1), 1–7.
Bates, T. (1994). Social Resources Generated by Group Support Networks
May Not Be Beneficial to Asian Immigrant-Owned Small Businesses.
Social Forces, 72(3), 671–689.
Bates, T. (1995). Analysis of Survival Rates Among Franchise and
Independent Small Business Startups. Journal of Small Business
Management, 33(2), 26–36.
Bates, T. (2005). Analysis of Young, Small Firms That Have Closed:
Delineating Successful From Unsuccessful Closures. Journal of Business
Venturing, 20(3), 343–358.
140
References 141
153
154 SME performance