DO MORE HEAVILY REGULATED ECONOMIES HAVE POORER SPAIN

advertisement
DO MORE HEAVILY REGULATED ECONOMIES HAVE POORER
PERFORMING NEW VENTURES?: EVIDENCE FROM BRITAIN AND
SPAIN
Joan-Lluis Capelleras, Universitat Autonoma, Barcelona, Spain
Kevin F. Mole, Francis J. Greene and David J. Storey*, Warwick
Business School, University of Warwick, UK.
*Address for Correspondence: CSME, Warwick Business School,
University of Warwick, Coventry,
United Kingdom CV4 7AL
Tel: +44 (0)24 765 220 74
Fax: +44 (0)24 765 237 47
Email: david.storey@wbs.ac.uk
1
DO MORE HEAVILY REGULATED ECONOMIES HAVE POORER
PERFORMING NEW VENTURES?: EVIDENCE FROM BRITAIN AND
SPAIN1
ABSTRACT
This paper examines new venture performance in two countries,
Britain and Spain.
Each has radically different regulatory
environments: Britain is the fifth least regulated out of 85 countries,
whereas Spain is in fifty-fifth place. Hypotheses are derived that
suggest that in Spain there would be fewer start-ups, they would be
larger at start and would grow more slowly than those in Britain. The
key finding is that new ventures appear to be almost identical in both
economies, both in terms of numbers and post start up performance.
This questions attempts to ease bureaucratic burdens on the process
of new venture start-up and growth.
1. INTRODUCTION
Countries vary markedly in the way in which they regulate and
provide an environment for enterprise. Djankov et al (2002) document
stark differences between, for example, Italy and Canada. In the
former, an individual wishing to start a five employee business has to
The authors would like to express their thanks to the Leverhulme Foundation for
their generous funding of the English element of this project. The Spanish element
has been financed by the “SME Observatory” of the Institut d’Estudis Regionals i
Metropolitans de Barcelona and the Departament d’Economia de l’Empresa,
Universitat Autonoma de Barcelona. The paper has benefited considerably from
comments received from Andre van Stel, Francis Chittenden and participants at the
4th International Entrepreneurship Forum in Paris, the FSF group in Halmstad,
Sweden, the 28th National Conference of Political and Industrial Economists in
Ancona, Italy, and seminars at Technical University of Lisbon and the UK
Department of Trade and Industry, London. We are, of course, responsible for the
paper.
1
2
follow 16 different procedures, pay the equivalent of nearly US$4,000
and wait 62 days for the necessary permits.
In Canada, the
procedures can be completed in two days by paying US$280.
Djankov et al’s conclusion was that these types of regulation were not
in the public interest. They found that countries where regulations
were most burdensome were less likely to be democratic, more
characterized
by
official
corruption,
and
had
larger
unofficial
economies and lower levels of wealth. Similarly, in the case of a single
country, Bertrand and Kramarz (2002) find the entry restrictions
imposed by regional zoning boards in France lead to slower retail
employment growth. The case for lowering the barriers to establishing
new businesses seems clear.
But regulation may affect not only venture creation.
It may also
influence subsequent survival and growth (Gimeno et al, 1997). For
example, business owners may choose not to grow their business
beyond a given size threshold for fear of triggering eligibility for
employee rights or other forms of legislation (Pryor, 2000). Support
for the impact of such legislation on the growth of new firms is
provided by the highly influential work of Scarpetta et al (2002). They
showed that ‘burdensome regulations’ on entrepreneurial activity
explained why, in the US, entrant firms started smaller than those in
European Union countries, but those which survived grew faster and
quickly surpassed them in size. The EU now uses this research to
3
drive forward its efforts to reduce ‘bureaucratic burdens’ on firms as
one element in its efforts to close the productivity gap between Europe
and the United States (European Union, 2002).
However, neither the academic nor the policy literature has yet sought
to model the mechanism by which the regulatory environment
influences the formation and growth of new and small firms (NSFs), in
order to address the question: what differences would we expect to see
in NSFs in a heavily regulated (HR) economy, compared with a lightly
regulated (LR) economy?
This paper does address this question.
has two effects.
The first is to add to the cost of starting and
operating a business.
entrepreneurs.
It theorizes that regulation
The second is to influence the “skill set” of
It then uses these assumptions to formulate six
hypotheses, which it tests using data from Britain – or more
accurately England - and Spain. These countries are chosen to reflect
clear differences in their regulatory regimes.
As shown in the first
column of Table 1, according to the Djankov et al index, Great Britain
– can be considered to be the fifth least regulated economy in the
world for NSFs. In contrast, Spain is a much more heavily regulated
4
economy, occupying position 55 out of the 85 countries.2
Table 1: Measures of the Scale of the Regulatory Environment
Country
Djankov et al
Nicoletti and
Kauffman et al
(2002)
Scarpetta
(1999)
Cost + Time of
(2003)
Economy-wide
meeting
Overall
product market
government
regulatory
regulation.
requirements.
approach.
Position in
Position in
Position in
OECD
world
OECD
Britain
5
1
1
Spain
55
13
15
Number of
85
21
21
countries
Given these sharp differences in the regulatory approaches of the two
countries it might be expected that there would be equally stark
differences between the characteristics of their NSFs.
The paper
reports the results of a parallel survey of new firms in the two
countries using, in so far as possible, identical research approaches.
Perhaps surprisingly, new firms in the two countries show much
greater similarities than differences. In both countries, the surveyed
firms were, on average, between four and five years old. In contrast to
the theory, the initial start up size of the firms was not significantly
different in the two countries and neither was their current size.
2
Other reviews of overall regulatory approaches adopted by countries, by Nicoletti
and Scarpetta [2003] and by Kaufman et al [1999], also point to wide differences
between GB and Spain. However, it is important to note that, since 2002, a number
of large high income countries, most notably France, have sought to reduce the time
and cost of business start up. Nevertheless, the most recent World Bank data
(Doing Business in 2005) indicates that, whilst it takes 18 days to start a business in
Britain, it takes 108 in Spain. Equally, the cost of starting a business in Britain is
0.9% of GNP per capita, compared with 16.5% for Spain.
5
Furthermore, the pattern of employment growth was also strikingly
similar, as were the factors that ‘explained’ this growth.
The paper concludes by speculating on the reasons why these results
differ so markedly from the theory and prior empirical findings.
It
concludes that the focus of prior work has been exclusively on new
‘registered’ or official businesses, whereas the current work includes a
high proportion of new firms that are below the size threshold at
which internationally comparable statistics are collected.
The fact
that such firms begin, and often remain, below the radar screens of
officialdom may mean they are more likely to be influenced by local
conditions and attitudes, rather than by the official regulatory regime.
Our findings are compatible with an explanation that sees the
regulatory framework as a secondary influence upon the performance
of new and small firms. Of greater significance are the characteristics
of new and small firm owners. Their skills and determination appear
to transcend national boundaries and, by implication, regulatory
regimes.
The paper begins by theorizing about the implications for new and
small firms (NSFs) of differences in regulatory regimes. Specifically, it
makes a distinction between heavily regulated (HR) and lightly
regulated (LR) economies and formulates five hypotheses to be tested.
Section 3 sets out the procedures used for data collection. Section 4
6
tests the hypotheses and the implications of the findings are
considered in Section 5.
2. THEORY
Using an Evans and Jovanovic (1989) framework, we assume venture
creation takes place when the individuals expected utility as a
business owner exceeds that from employment as an employee3.
E (U SE ) > E (U E )
[1]
Whilst utility as a business owner comprises both pecuniary and nonpecuniary benefits4, for simplicity we assume it corresponds to income
(Y).
Income from the business/self-employment YSE, is assumed to
depend upon entrepreneurial talent θ, and access to capital, K.
YSE = ƒ(θ, K)
[2]
A second Evans and Jovanovic assumption is that access to capital, K
depends upon the personal wealth of an individual (W) and that all
individuals, irrespective of their entrepreneurial talent, are able to
borrow a fixed proportion λ against that wealth.
We exclude unemployment for ease of exposition. For the same reason, we exclude
a time dimension. In other words, current income or utilities are compared rather
than expected future utility.
4 See the work of Morton and Podolny (2002) on wine producers in California.
3
7
K= W[1+λ]
[3]
So, in what respects would we expect the venture creation decision
and the venture continuation decision5, to differ between heavily
regulated [HR] and lightly regulated [LR] economies?
From Equations [1] and [2] we can see that business ownership rates
in country i will depend upon alternative employment opportunities,
the distribution of entrepreneurial talent and access to capital.
Although this will be modified later, we begin by assuming that, in the
short term, governments seeking to raise business ownership are
unable to influence θ. We also assume governments do not wish to
reduce the attractiveness of paid employment [UE], since this is the
main provider of jobs in most economies.
We define an HR economy as one where it is more ‘expensive’ in terms
of time and cost to start a business than in an LR economy. Since
this is a fixed cost, the effect is clearly to lower owners’ wealth. The
effect from Equation [3] is to lower K and hence YSE in an HR
economy. This implies:
5
See Gimeno et al (1997)
8
H1: Fewer businesses will start in a highly regulated (HR)
economy than in a lightly regulated (LR) economy.
A second effect of the higher fixed costs of starting a business in an
HR compared with an LR economy is that it is less worthwhile starting
very small businesses because the income generated is insufficient to
cover the fixed start-up costs. This implies:
H2: The size distribution of (legal) start-ups will differ: Highly
regulated (HR) economies will have a higher proportion of larger
start-ups than lightly regulated (LR) economies.
From Equations [2] and [3] it is clear that both λ and θ influence new
venture creation choices and both are likely to differ between HR and
LR economies.
λ is likely to be higher in an LR economy because
banks, which are the main source of external borrowings for NSFs
(Petersen and Rajan, 1994), will be both better informed and have a
higher expectation of loan repayments than banks in an HR economy.
Governments can also be a source of capital for new and small firms.
However it seems unlikely that HR governments, which impose
regulations upon the starting of a business, are likely to be a more
important source of new firm funding than LR governments that seek
to promote new and small enterprises.
The effect, therefore, is to
reinforce the difficulties of new firms in HR economies, leading to
lower YSE. Hence, individuals with the same levels of θ will become
business owners in an LR economy, yet choose to remain as
9
employees in an HR economy.
Conversely, those who do choose to
become business owners in an HR economy are more likely to be only
those capable of earning low incomes as employees. All else equal,
they are likely to be individuals with low levels of human capital.
Equation [2] assumes entry into self-employment is determined by λ,
but also by entrepreneurial talent, θ. Hence, differences between HR
and LR countries could reflect a different distribution of θ, which is
independent of the regulatory regime. For example, the ‘skill set’ of
the HR economy entrepreneur could differ from that of the LR
economy entrepreneur. The former has to be skilled in ‘dealing with
the bureaucracy’ by being able to complete forms and being willing to
devote time to these procedures. In contrast, the skill set of the LR
entrepreneur would typically focus more upon the customer for the
business.
A second dimension is that HR economies are frequently characterized
by entry into a self employed trade or profession requiring a very
lengthy ‘apprenticeship’, but without which it is impossible to trade.
This selection process is likely to attract more patient individuals and
to focus upon technical and administrative skills6.
Since lengthy
An example may be the lengthy apprenticeship system used in some countries for
electricians, plumbers and other building workers. In HR economies, these trades
are more likely to attract entry from individuals prepared to wait for some years to
obtain relatively certain and secure income. They are likely to be less attractive to an
individual who identifies a business opportunity and seeks to exploit it by gaining
first mover advantage. The differences in apprenticeship patterns between HR
Germany and LR United States are examined by Acemoglu and Pischke (1998).
6
10
apprenticeships are much more characteristic of an HR than an LR
economy, this is also expected to be reflected in the personal
characteristics of the self-employed.
These differences in θ, the entrepreneurial skill set and in λ, the
owners borrowing power, imply that NSFs, would differ between HR
and LR economies.
H3: The characteristics of new firms and their owners in HR
economies will differ from those in LR economies.
As well as being a fixed cost on start-ups, some forms of regulation
may lead to higher variable costs for new and small businesses.
Others may act as disincentives on firms to grow beyond a certain
scale. An example is Value Added Tax (VAT) which is a sales tax paid
only by [virtually all] registered businesses over a minimum size7. An
example of scale-based disincentives is where employment ‘thresholds’
exist in which firms beyond the threshold become responsible for
payments of workers’ pensions or have to comply with other forms of
employment legislation, whereas the firms below the threshold do
not8. The effect of such thresholds is hypothesized to constrain the
growth of new and small firms.
In 2004 this threshold in the UK was sales in excess of £58,000.
These size thresholds also vary according to the nature of the legislation. For
example, in the UK, firms with less than 5 workers are exempt from parts of the
Maternity and Parental Leave Regulations, 1999; those with less than 21 employees
are exempt from parts of the Employment Relations Act 1999; and, until 2004, firms
with less than 15 employees were exempt from the Disability Discrimination Act
1995. A full review of these exemptions is provided in Keter (2004).
7
8
11
A second factor which may lead to slower growth rates in new and
small firms in HR compared with LR economies is that, as a result of
these higher costs, expected profits would be lower in the former
economies. Since retained profits are a key influence on wealth (W)
this, from Equation (3), has a direct effect on (K) and an indirect effect
on (λ) since it makes banks less willing to lend to small firms. This
implies:
H4: The growth rates of new and small firms will be slower in
HR economies than in LR economies.
There are three possible explanations for these different growth rates.
The first is that the human capital of (legal) new business owners will
be lower in an HR than an LR economy, because those with high
human capital will obtain alternative employment. Second, the costs
of start up and the continuing costs of business operation are higher
in an HR economy. Thirdly, regulation may be particularly onerous
on new and small firms in certain sectors in an HR economy, so the
sectoral composition of firms is also likely to differ. This implies:
H5: The factors influencing the growth of new and small firms
in HR and LR economies will differ.
12
3. DATA AND METHODOLOGY
Data for this study were obtained from two surveys of wholly new
independent firms conducted separately in selected areas of England
and Spain. Given the earlier data presented in Table 1, Spain will be
taken as the HR economy and Great Britain, and specifically England,
is taken as an LR economy9. A key purpose is to select not only new
firms that appear in official statistics, but also to include those which
are below the minimum size threshold and, hence, are omitted from
official statistics.
It is also to select geographical areas within the
countries that can be regarded as broadly representative of those
countries.
The British survey was conducted in 2001 in three English counties,
Buckinghamshire, Shropshire and Tees Valley (the former county of
Cleveland plus Darlington).
The three counties were specifically
chosen to reflect high, medium and low firm entry rates, by the
standards of official statistics10 (For more details see Greene et al,
2004).
This survey used the identical approach of two previous
studies on new firms founded in the county of Cleveland, one
undertaken in 1980 and one in 1991 (Storey, 1982; Storey and
9 Although Table 1 refers to Britain the regulatory data are derived for London, the
capital city of England. Britain comprises England, Wales and Scotland, but not
Northern Ireland
10 Using the measure of new VAT Registrations per 10,000 population the rates for
Tees Valley, Shropshire and Buckinghamshire in 2001 were 18, 32 and 52
respectively. The English average was 33. Equally importantly, as shown by Greene
et al (2004), whilst formation rates do vary with the economic cycle, Tees Valley is
13
Strange, 1993). One key benefit of the approach was its success in
surveying new firms that were not registered for VAT and, hence,
never appeared in any UK official statistics
11.
The UK study was replicated in 2003 in Spain. A geographical and
administrative area (Western Valley, adjacent to Barcelona) with
comparable economic characteristics to the English counties by
national standards was selected.
In particular, the Western Valley
exhibits a firm density which is similar to the Spanish national level
(63.1 firms per 1,000 inhabitants)12.
The population of the selected areas in both countries is also quite
similar. The population of Western Valley was approximately 740,000
in 2001. The number of inhabitants in Tees Valley was 640,000 in the
same
year.
Buckinghamshire
had
480,000
inhabitants.
In
Shropshire, this figure was 440,000. Western Valley has a lower GDP
per
head
than
the
three
English
areas13,
reflecting
national
differences14. The unemployment rate is higher in the Spanish area
than in the English counties (7.1 in Western Valley, 5.4 in Tees Valley,
always low, Shropshire is always average and Buckinghamshire is always high, over
the period 1980-2001.
11 Storey and Strange (1993) found that only 50% of the new firms were registered
for VAT. Of the unregistered 50% of firms, 16% had sales below the VAT threshold
and so were not required to register, 4% had sales above the limit, and the
remaining 30% refused to disclose their sales.
12 Unfortunately, start-up rates are available in Spain only at national level. Firm
density is, therefore, used as the best measure of entrepreneurial activity at regional
and county levels.
13 The GDP per head is higher in Buckinghamshire than in Shropshire and Tees
Valley.
14 In 2001, Spanish GDP per head was $20,155 compared with $25,479 in the UK.
14
1.9 in Shropshire and 1.4 in Buckinghamshire, all in year 2001). This
broadly reflects unemployment rates in each country (5.1 in the UK
and 10.5 in Spain for the year 2001).
Neither Britain nor Spain has a single, comprehensive and publicly
available list of new firms. Existing lists of limited companies, which
are publicly available, exclude numerous small start-ups and so are of
no value in this context. The VAT data is not publicly available but
even this excludes firms with sales of less than £58,000.
Our
reservations about private databases for Britain is their greater
likelihood of including firms seeking credit, so generating sample bias.
For Spain, these databases are recognized to be even more imperfect.
Databases, appropriate for the purpose, were, therefore, constructed
by the researchers in each country. The English study compiles a list
of new firms in the same way as in the two previous Cleveland studies.
A list of new firms was derived through comparisons of BT telephone
directories for 2000 with those from 1995.
Those firms in the
directories for 2000, but not present in 1995, were considered to be
potential new firms to the area.
In the Spanish area, an initial list of new firms was derived using
three sources of information in order to include both limited and non
limited companies. A list of new firms based on local tax payments,
the Chambers of Commerce and Industry directory and a commercial
15
database based on the Official Register of Enterprises was compiled.
A careful analysis of these lists was made and overlaps between the
three databases were detected. From this cross-checking process, a
list of potential new firms, which were defined as those founded
between 1998 and 2000, was obtained.
Having then derived a list of potential new firms, identical procedures
were used in England and Spain to produce lists of new firms.
Researchers contacted businesses by phone in order to determine
whether they were wholly new independent firms. The study excluded
firms that were ‘in-moves’ to the area, subsidiaries, affiliates and firms
created for reducing tax burdens. Face-to-face interviews were then
conducted with new firm founders on the grounds that the researcher
was certain these were ‘real’ businesses.
The questionnaire was initially designed in English and was translated
for the Spanish study. The Spanish questionnaire was tested through
a series of extended interviews. During this process it became clear
that some questions were not applicable in the Spanish context, and
these were eliminated. In addition, some questions were asked only of
the Spanish firms, but the vast majority of questions were common to
both countries. The questionnaire took around an hour to complete
and was administered at the normal place of work of the respondent.
16
The final sample consists of 624 new firms in England15 and 182 in
Spain. However, only 490 English firms provided data on employment
at start up, whereas this was provided by all Spanish firms, and it is
these responses that are used in this paper.16
4. TESTING THE HYPOTHESES
H1: Fewer businesses will start in a highly regulated (HR)
economy than in a lightly regulated (LR) economy.
Ideally, it would be desirable to test whether there are differences in
the rate of new business start ups in HR and LR economies, but this
is not possible using our data.
This is because documenting the
population of new businesses, even in our chosen areas, is difficult for
15Given
the multi-stage processes necessary to identify this group of new firms,
there is no single statistic which is an ideal measure of response rate. The Tees
Valley results illustrate this. A total of 2,490 ‘new firms’ were identified from the
Telephone Directory and contact was attempted with 2,112. Of these, 791 were
ineligible on grounds of age, ownership or sector, 278 had ceased trading or moved
outside the area, implying an eligible population of 1,043. In total, 320 firms were
interviewed. One estimate of response rate is therefore 320/1043, implying a
response rate of 32%. However, a total of 336 telephone lines were either
disconnected or re-allocated which is likely to be a powerful sign that the business
has ceased. If these are also assumed to be ‘dead’ businesses, and the denominator
only includes businesses known to be eligible at the time of the survey, then the
response is 320/707 or a rate of 45%. Our, highly subjective, judgment is that the
real response rate is likely to be around 40% for Tees Valley. For Buckinghamshire
the response rate, comparable to the Tees Valley figure of 45%, is 69% and for
Shropshire the comparable figure is 75%. For the Western Valley, the comparable
estimation of the response is 182/404, implying a rate of 43%.
16 The reason for this discrepancy is the English survey asked about employment
one, three and five years prior to the interview, but did not ask a specific question
about start up. Where the available data corresponded with the year of start up this
data was used, but clearly it was not appropriate to use it for all cases. There were
no significant differences between the 624 and 490 cases in terms of sector, age or
geography. For Spain, a specific question on number of employees at start-up was
used.
17
the reasons given in Section 3 above.
As an alternative, we examine the relationship between the Djankov et
al measures of regulation and the Global Entrepreneurship Monitor
(GEM) measure of actual new firm formation – referred to as ‘baby
businesses’ that have been running for between 4 and 42 months and
have not been paying salaries longer than that17.
The merit of the
baby business measure in this context is that it also captures
businesses that are not included in official statistics, either because
they are too small or because they choose not to register.
Data for both measures are available for thirty six countries and are
presented in Figure 1. Simple correlation coefficients are reported in
Table 2.
Figure 1 identifies the United Kingdom and Spain, but for comparative
purposes it also includes the United States. It shows the three
countries occupying ‘corner positions’. Thus the UK and Spain have
broadly similar formation rates but radically different entry barriers.
Yet the UK and the US have very similar entry barriers but very
different formation rates.
This is not the same as the more widely quoted Total Entrepreneurship Activity
(TEA) measure which combines both nascent and baby businesses. We use the baby
measure because our focus is on actual businesses started rather than on the prestart phase. We also use the Djankov et al (2002) measure of burdens, rather than
the more recent World Bank (2005) measure because the new businesses will have
started during an earlier regulatory regime, captured in Djankov et al (2002).
17
18
Figure 1: Scatter plot of UK, US and Spain
8
USA
7
GEMnascent firms %2002
6
5
4
3
UK
Spa
2
0,0
,1
,2
,3
,4
,5
,6
Djankov et al (2002) Cost + time 1999
Table 2: Correlation coefficients: Four measures of Bureaucratic
burdens and New Firm formation
Number of
Procedures
Time
Cost
Cost + Time
.005
.099
.005
.062
GEM New
firms
N=36 countries.
Table 2 then shows that for the 36 countries that appear in both GEM
and Djankov et al there is no association between new firm formation
rate and the various measures chosen to reflect the bureaucratic
difficulties of starting a new enterprise.
H1, therefore, cannot be
supported from this evidence.
19
H2: The size distribution of (legal) start-ups will differ: Highly
regulated (HR) economies will have a higher proportion of larger
start-ups than lightly regulated (LR) economies.
To test this, and all subsequent hypotheses, the data used will be for
England and Spain only, the derivation of which was described in
Section 3 above. Table 3 provides means, standard deviations and the
range (min, max) for employment variables; Figure 2 shows the
distribution of start-up employment sizes18. Table 3 shows there is no
significant difference between the mean initial start up size of firm in
England and in Spain. However, Figure 2 does suggest that a slightly
lower percentage (50%) of Spanish new firms began with less than
three employees, than was the case for England (65%).
Whilst H2
cannot, therefore, be supported, the pattern of the distribution
suggests Spain may have fewer very tiny start-ups.
Table 3: Employment: current and at start up
Variables
England (a)
Spain (b)
Mean
Std.
Dev.
Min
Max
Mean
44
Std.
Dev.
T-test
Min
Max
3.26 2.65
1
16
-1.31
Start-up size
2.89 3.46
1
‘Current’ size
5.77 9.96
1
109 5.87 5.16
1
34
-0.13
Absolute
change
2.88 8.84
-8
104 2.61 4.16
-6
26
0.39
a
N=490,
b
N=182
A wide variety of different new venture performance measures have been used in
the literature, such as employment, profitability, sales or more subjective measures
of performance. We favor the employment measure even though it is not an objective
the owner seeks to maximize. This is because in an international study it minimizes
currency and accounting problems, and yet is broadly correlated with sales.
18
20
Figure 2: Number of jobs at foundation
35
30
25
England
15
Spain
%
20
10
5
0
1
2
3
4
5
6
7
8
9
10+
No. Jobs
H3: The characteristics of new firms and their owners in HR
economies will differ from those in LR economies.
H3 is tested in Table 4.
Appendix A shows the definitions of all
variables used in this and all subsequent tables.
Following Storey
(1994) a distinction is made between ‘pre-start’ and ‘at –start’
variables.
Previous papers that have used the ‘pre-start’ variables
used in this paper are tabulated in Appendix B. Appendix C shows
previous papers using the ‘at start’ variables.
Table 4 shows there are several significant differences between
England and Spain both in terms of the business type and the
business owners.
Compatible with H3, there are more business
owners with formal qualifications, more with higher managerial
experience, and founders are also older in the English sample.
As
expected, the use of external advice, both before and after start-up, is
also higher amongst the English firms.
21
Table 4: Characteristics of the samples
Variables
England
Spain
0.76
43.60
1999.70
0.90
0.62
0.22
0.89
0.55
0.78
0.28
0.17
0.14
0.69
38.24
1555.18
0.72
0.40
0.24
0.70
0.43
0.78
0.36
0.09
0.02
0.35
0.17
0.08
0.15
0.07
0.03
0.03
0.27
0.03
0.03
0.13
49.79
23.88
26.33
0.77
0.21
0.10
0.26
0.03
0.04
0.02
0.18
0.03
0.03
0.10
4.30
4.88
Pre-start variables
Male*
Foundage***
Foundage2d***
Formal qualification*
Manager***
Unemployed
Support***
Formal written business plan**
Start-up personal savings
Start-up clearing bank**
Start-up friends or relatives**
Start-up public organizations***
At-start variables
Limited co***
Manufacturing
Construction
Wholesale and retail trade, motor repair***
Hotels and restaurants**
Transport, storage and communication
Financial intermediation
Real estate, renting & business activities**
Education
Health and social work
Other social and personal services
Tees Valley
Shropshire
Buckinghamshire
Firm age***
Note: table shows mean values for each variable.
Significance levels: *p<0.10, **p<0.05, ***p<0.01.
However, in contrast with H3, the proportion of limited companies is
significantly higher amongst the Spanish than the English firms.
Furthermore, new Spanish firms were more likely than English firms
to use bank loans or overdrafts to finance the business prior to startup.
22
A second test of H3 is presented in Table 5. Pre and at-start variables
are used to explain start-up size19.
Initial size is measured as the
logarithmic transformation of the initial number of jobs. A general-tospecific approach is used, beginning with the inclusion of all variables
and then moving to a ‘best’ model. The models have good explanatory
power, especially for the Spanish data.
Table 5 shows the
unemployment variable is significant in explaining start-up size in
both samples, implying that the unemployed in both countries start
smaller businesses but the variables themselves differ between the
two countries. It is difficult to explain why either of these variables
might be an outcome from different regulatory regimes.
Even less compatible with H3 is the finding that, in both England and
Spain, firms founded with a limited liability legal form start bigger
than other firms. Industry sector also appears to play a significant
role in determining start-up size in the two samples, but it is only in
hotels and restaurants that new firms start on a larger scale than
other firms in both England and Spain. In the English counties the
dummy variable for health is positively related to initial size, while
financial Intermediation and education are negatively related to startup size.
In Spain, it is construction and manufacturing firms that
begin large.
19 Appendices B, C and D present research findings from fourteen prior multivariate
studies conducted in a number of countries on start-up size and new firm growth.
These studies have been used to derive the variables used in the current study.
23
Other specific variables are also significant in explaining start-up size.
Location matters for the scale of entry in the English case: firms in
Tees Valley are more likely to start bigger than in Shropshire and
Buckinghamshire.
Having a business plan, prior to starting the
business, is characteristic of larger Spanish new firms.
In sum, the initial size of new firms in both countries is influenced by
sectoral composition. A second key similarity between the countries is
that start-up size is strongly influenced by legal status. Given that
this is the clearest indicator of ‘regulation’ available to us, and limited
liability is more expensive to obtain in Spain than in England, it
suggests that making it more expensive does not seem to either
reduce its importance or its use. Conversely, it is difficult to see the
benefits to Spain of lowering the cost of obtaining limited liability.
Thirdly, the unemployed in both countries start smaller businesses
than those moving directly from employment.
Overall, therefore, there is little support for H3, implying that, when
they start, there are few differences between new firms and their
owners in England and in Spain. Furthermore, even where there are
differences, it is not obvious that these are linked to differences in the
regulatory regime, since limited liability status exercises a positive
influence
on
start
up
size
in
both
England
and
Spain.
24
Table 5: Multiple regression analysis for start-up size
England a
Variable
General model
Coef.
t
(Constant)
.221
0.36
Male
.078
Foundation age
.012
Spain
Specific model
Coef.
t
Coef.
t
.75
.114
1.50
0.98
-.118
-2.33**
-.109
-2.54**
0.47
-.003
-.17
-.001
-0.21
.000
.04
Qualification
.053
0.42
.029
.26
Manager
.072
1.15
.010
.20
-. 099
-1.84*
-.124
-2.64***
.133
3.35***
Unemployed
-.223
-2.92***
Support
.084
0.80
-.031
-.57
Formal written business plan
.032
0.49
.120
2.48**
-.069
-0.83
-.051
-.99
Start-up clearing bank
.183
2.37**
.043
.85
Start-up friends or relatives
-.021
-0.25
-.023
-.30
Start-up public organizations
-.123
-1.25
.105
.67
.066
0.76
-.109
-1.35
.263
3.56***
Start-up personal savings
Shropshire
Buckinghamshire
Limited Company
Manufacturing
Construction
.037
-.079
0.30
-0.57
-.212
8.62***
Coef.
Specific model
.248
Foundation age2d
.701
t
General model
.170
-2.98***
2.36**
.311
4.42***
.267
4.40***
.263
5.44***
.099
0.88
.275
3.30***
.279
3.71***
-.037
-0.28
.222
2.29**
.220
2.58**
25
Trade
-.338
-3.08***
-.272
-2.73***
.081
1.00
.093
1.28
.297
1.99*
.335
2.48**
.354
2.33**
.282
2.16**
Transport
-.276
-1.83*
-.233
-1.58
.098
.80
.116
1.02
Financial intermediation
-.455
-2.48**
-.332
-1.92*
.022
.14
.033
.23
Business activities
-.275
-2.37**
-.189
-1.77*
.014
.17
.017
.22
Education
-.501
-3.21***
-.428
-3.13***
.082
.58
.147
1.21
.472
1.72*
.680
2.39**
.185
1.34
.165
1.27
Hotels & restaurants
Health
R2
0.187
0.177
0.376
0.369
n/a
n/a
0.280
0.321
F
4.96***
7.75***
3.93***
7.57***
Ramsey RESET test: F
2.74**
2.49*
1.26
0.51
Cook-Weisberg test: chi2(1)
8.17***
18.49***
1.65
1.07
174
182
Adjusted R2
Number of cases
452
484
aNote: Regression with robust standard errors hence no adjusted R2.
Significance levels: *p<0.10, **p<0.05, ***p<0.01.
26
H4: The growth rates of new and small firms will be slower in
HR economies than in LR economies.
To conduct our first test of this hypothesis we return to the data
presented previously in Table 3, supplemented by Figure 3 below.
Table 3 shows that there is no significant difference either in the
current size or in the absolute change in employment since start-up
in the English and Spanish firms. Similarly Figure 3 shows that the
current employment size distribution of the English and Spanish
firms is strikingly similar. This result provides no support for H420.
25
20
15
%
England
Spain
10
5
0
1
2
3
4
5
6
7
8
9
10+
No. Jobs
Figure 3: Number of jobs ‘currently’
As the first step in a second test of this hypothesis we repeat the
analysis of Bruderl and Preisendorfer (2000) who examined new firms
in Munich in 1990. They classified firms into four groups based on
It is important to note that the differences between Britain and Spain identified in
Table 1 relate only to the cost of starting a business, and so may not influence the
growth of a business. World Bank (2005), however, suggests that new firms in
Spain face considerably more ‘red tape’ after start up than those in Britain.
20
27
employment change since start-up:
1. Decliners: Firms with negative employment change i.e.
firms with fewer jobs than when they began.
2. Statics: Firms with no change in the number of jobs.
3. Slow growers: Firms with some, but only small employment
growth.
4. Fast growers: Firms which have created at least five
additional jobs since starting21.
Table 6 presents descriptive statistics for these four categories of
employment change for the English and Spanish data. It shows the
proportion of firms losing jobs (i.e. decliners) is 7% in England and 8%
in Spain. Fast growing firms are 15% of the English and 19% of the
Spanish sample. Statics are the largest group in the English dataset,
whereas slow growers are the largest group in the Spanish case. No
significant differences in employment change are found between the
four groups of new firms, as shown in Table 6, implying no support for
H422.
An alternative presentation is made in Figure 4. It shows the patterns
of employment change in the four groups of new firms in the English
and Spanish data set. Again, it is the similarity of the growth patterns
that is striking, rather than the differences.
Figure 4 shows the
largest firms at start-up in both countries are the decliners; the
An alternative approach adopted by Delmar et al (2003) is to use cluster analysis
to derive the groups. Our view is that this creates somewhat artificial clusters and
that our simpler and easily interpretable classification is, on balance, preferable.
22 Bruderl and Preisendorfer (2000) find that 8% are decliners which is virtually the
same as our results. However, they find 60% of firms are statics, 7% are fast
growers and 25% are slow growers, which does differ from the findings of this paper.
21
28
second largest group are the fast-growers, followed by the slow
growers. The firms that are smallest at start are the statics. This is
the case for both Spain and England.
Table 6: Employment change in four groups of firms
England
Group
Spain
a
T-test
b
N
%
Mean
Std.
Dev.
Min.
Max.
N
%
Mean
Std.
Dev.
Min.
Max.
33
6.7
-1.82
1.67
-8
-1
15
8.2
- 2.13
1.77
-6
-1
0.60
Statics
204
41.6
0.00
0.00
0
0
50
27.5
0.00
0.00
0
0
-
Slow
growers
178
36.3
2.09
1.06
1
4
83
45.6
2.24
1.16
1
4
-1.02
Fast
growers
75
15.3
14.68
18.3
6
5
104
34
18.7
9.47
4.57
5
26
1.63
490
100
2.88
8.84
-8
104
182
100
2.61
4.16
-6
26
0.40
Decliners
Total
F-test = 83.251***, b F-test = 160.702***.
Significance levels: *p<0.10, **p<0.05, ***p<0.01.
a
The largest firms at start-up are most likely to become decliners in
both countries. Their employment declines from seven to five jobs in
the Spanish case and from five to three in the English one. Thus,
these firms have lost an average of two jobs each. The group
constitutes 7% and 8% of new firms in England and Spain,
respectively.
In both countries, decliners converge over time with the group of slow
growers. These firms have grown from three to five jobs over time.
They are the biggest group of businesses in Western Valley (46%).
29
Zero employment change is observed in a group of very small
businesses.
These firms have on average about three jobs in the
English sample and about two in the Spanish one. This static group
constitutes 42% of the English businesses and 28% of those in Spain.
Figure 4: Patterns of employment change
England
Spain
Number of jobs
Number of jobs
20
20
15
15
y=3
y=3
10
10
y=0
y=2
y=0
5
5
y=1
y=1
t
t=0
y=0: Decliners
y=2
t+4
t=0
y=1: Statics
y=2: Slow growers
growers
t
t+5
y=3: Fast
Finally, fast-growing firms constitute 15% and 19% of all new firms
respectively in the two economies. These firms started with less than
five jobs, but have expanded to nineteen jobs in the English sample
and fourteen in Spain.
30
To sum up, Figure 4 clearly shows that the four patterns of
employment change are both very different from one another, but very
similar amongst English and Spanish new firms. This similarity both
of growth patterns and employment numbers provides little support
for H4 that the growth rate of new firms would be expected to be lower
in HR economies than in LR economies.
H5: The factors influencing the growth of new and small firms
in HR and LR economies will differ.
To test H5, an ordered probit model is used to estimate the probability
of a firm having negative, neutral, slow or fast employment growth.
Tables 7a and 7b show marginal effects for England and Spain. In
addition to the pre-start and at-start variables, the post-start
variables are also included. Prior work which has utilized these poststart variables is identified in Appendix D.
It is clear that almost all the signs of the significant variables are the
same for decliners and statics, while opposite signs are found in the
estimations of the two remaining groups (slow and fast growers). This
suggests a basic split between the determinants of new firms with job
gains and those of new firms with neutral or negative employment
change.
In terms of H5, two variables are significant in both the LR and HR
sample. In both the Spanish and English new firms, workforce
31
training is positive and significant23.
Second, the choice of limited
liability has a positive and significant marginal effect on employment
growth in both English and Spanish firms.
Strong significant sectoral effects are also found in both countries,
but the sectors where growth is faster varies. In the English sample,
construction and manufacturing firms are more likely to be slow
growers, whereas transport and business service firms are less likely
to be decliners. In Spain, health sector firms have a positive sign and
education a negative sign.
There are, of course, some significant variables in the ordered probit
estimations that are country-specific, but again these are not clearly
related to the characteristics of LR and HR economies. For example,
employment growth is positively correlated with prior managerial
experience of the founder only in the English new firms. No founderspecific characteristics appear in the Spanish equation.
In the English case there is no significant relationship between
employment change and the use of external advice, as previous
findings in the UK indicate (Westhead and Birley, 1995). In contrast,
those Spanish founders establishing either a fast or slow growing
business (i.e. positive employment change) are more likely to use
23 This may indicate that new firms who invest in training their employees are those
that are able to finance this from their profits or that training is a determinant of
growth. Our data are not able to distinguish between these two explanations.
32
external advice once the firm begins trading (1st year Support) while
decliners and statics are more likely to use advice prior to start-up
(Support).
Table 7a shows the English firms focusing on low prices were less
likely to be in the growing groups, whereas the growers were more
likely to currently have a bank loan24.
Table 7b shows that, for Spain, firms with limited liability grow faster
than those choosing other legal forms.
It is, of course, interesting
that this result is found for HR Spain, but not for LR England, since it
is our assumption that company status is more difficult/expensive to
obtain in a HR economy.
those
new
firms
which
Finally, in contrast to our expectations,
‘currently’
use
finance
from
public
organizations are more likely to be statics and less likely to become
fast growers.
Overall, the evidence on the factors influencing the growth of new and
small firms in England and Spain provides little clear support for H5.
In both countries, sector, human capital, choice of legal form and
strategy variables all influence new firm performance.
Of course,
differences between the countries also exist. Prior management
24 Again, as in the relationship with workforce training, it must be emphasized that
this does not imply that access to bank finance causes the growth.
33
Table 7a: Ordered probit model, marginal effects. England
Variable
Managera
Incorporated new firma
Manufacturinga
Constructiona
Tradea
Hotels & restaurantsa
Transporta
Financial intermediationa
Business activitiesa
Educationa
Healtha
Buckinghamshirea
Shropshirea
Traininga
Clearing bank loana
Lower prices
Firm age
Start-up size
Decliners
dy/dx
-.021
-.026
-.027
-.027
-.019
-.030
-.026
-.037
-.026
.026
-.033
.015
-.001
-.066
-.020
.009
-.007
.008
(y = 0)
z
-1.82*
-2.45**
-2.31**
-2.33**
-1.46
-2.82***
-1.88*
-4.09***
-1.98*
0.57
-3.17***
1.11
-0.03
-4.07***
-2.12**
1.93*
-2.97***
1.09
Statics (y = 1)
Slow growers (y = 2)
dy/dx
z
dy/dx
z
-.070
-2.02**
.041
1.87*
-.101
-2.57**
.051
2.64***
-.123
-2.01**
.049
2.87***
-.134
-1.85*
.045
3.53***
-.080
-1.28
.036
1.63
-.155
-2.24**
.046
3.87***
-.140
-1.33
.041
3.80***
-.261
-3.35***
-.003
-0.05
-.108
-1.86*
.050
2.21**
.074
0.74
-.052
-0.61
-.204
-2.17**
.031
0.86
.049
1.21
-.029
-1.13
-.001
-0.03
.001
0.03
-.208
-5.98***
.122
4.82***
-.076
-2.11**
.039
2.22**
.033
2.00**
-.018
-1.94*
-.024
-3.27***
.013
3.02***
.029
1.09
-.016
-1.08
Fast growers (y = 3)
dy/dx
z
.049
2.04**
.076
2.41**
.101
1.74*
.115
1.54
.063
1.16
.138
1.81*
.125
1.07
.302
2.05**
.084
1.70*
-.049
-0.81
.206
1.53
-.034
-1.22
.001
0.03
.153
5.89***
.057
2.01**
-.024
-2.01**
.017
3.30***
-.021
-1.09
Number of cases = 416. LR chi2(18) = 94.05***. Pseudo R2 = 0.094. Log-likelihood = -453.436.
a dy/dx is for discrete change of dummy variable from 0 to 1
Significance levels: *p<0.10, **p<0.05, ***p<0.01.
34
Table 7b: Ordered probit model, marginal effects. Spain
Variable
Supporta
Incorporated new firma
Manufacturinga
Constructiona
Tradea
Hotels & restaurantsa
Transporta
Financial intermediationa
Business activitiesa
Educationa
Healtha
1st year supporta
Traininga
Plan nowa
Public organizations
Owner managerial skills
Firm age
Start-up size
Decliners (y = 0)
dy/dx
z
.036
2.41**
-.151
-2.70***
-.028
-1.50
-.018
-0.93
.022
0.69
-.032
-2.20**
-.010
-0.32
.093
0.73
.064
1.22
.175
1.14
-.040
-2.86***
-.027
-2.01**
-.058
-2.66***
-.036
-2.20**
.352
1.25
-.020
-2.49**
-.002
-0.47
.102
2.73***
Statics
dy/dx
.142
-.261
-.116
-.075
.067
-.174
-.037
.169
.151
.217
-.273
-.113
-.196
-.131
.196
-.070
-.005
.049
(y = 1)
z
2.68***
-4.78***
-1.39
-0.81
0.79
-1.72*
-0.28
1.41
1.81*
3.82***
-6.98***
-1.97**
-3.55***
-2.48**
1.81*
-2.92***
-0.47
3.75***
Slow growers (y = 2)
dy/dx
z
-.054
-2.19**
.255
3.35***
.041
1.90*
.028
1.34
-.043
-0.69
-.013
-0.12
.017
0.36
-.173
-0.86
-.123
-1.33
-.281
-1.61
-.365
-1.69*
.039
2.01**
.092
2.56**
.059
2.21**
-.427
-2.40**
.039
2.35**
.003
0.47
-.193
-2.47**
Fast growers (y = 3)
dy/dx
z
-.124
-2.32**
.157
4.65***
.103
1.16
.065
0.69
-.046
-0.84
.219
1.01
.030
0.26
-.090
-1.83*
-.091
-2.09**
-.111
-3.54***
.678
2.97***
.102
1.65*
.161
3.23***
.108
2.22**
-.121
-4.48***
.052
2.92***
.004
0.47
-.259
-3.28***
Number of cases = 178. LR chi2(18) = 81.08***. Pseudo R2 = 0.185. Log-likelihood = -178.645.
a dy/dx is for discrete change of dummy variable from 0 to 1
Significance levels: *p<0.10, **p<0.05, ***p<0.01.
35
experience exerts a positive influence on firm performance in England,
but not in Spain.
Equally, a focus on lowering prices is a
characteristic of slow growing firms in England, but not in Spain.
Nevertheless, whilst these differences are identified, it is difficult to
clearly link them to the regulatory environment, as implied by H5.
5. CONCLUDING REMARKS
This paper theorized that in heavily regulated (HR) economies the cost
of both starting and operating a (legal) small business would be higher
than in a low regulation [LR] economy. It also argued that the effect of
these differences in regulations would influence the characteristics of
those
founding
new
subsequent growth.
businesses,
and
so
also
influence
their
In short, HR economies would have fewer new
firms, which would start larger but grow more slowly than those in an
LR economy. The factors influencing these growth rates would also
differ.
Testing for these differences was undertaken on a parallel sample of
new small firms [NSFs] in England and Spain. These countries were
selected on the grounds that the former could be viewed as a clear LR
economy, whereas Spain was a clear HR economy. In a world ‘league
table’, of regulatory burdens, England (Britain) was the fifth most
lightly regulated, but Spain was in fifty-fifth place, out of eighty-five
countries.
The key findings of the paper, however, point to the similarities rather
than to the differences in the NSFs established in LR and HR
economies. Not only are the firms of similar size at start-up, and after
four to five years of life, but the factors that explain initial size and
subsequent growth are also broadly similar.
Our results show that initial size is strongly influenced by human
capital variables.
In both economies, the background and starting
resources of the entrepreneur are very important in determining startup size. A particularly interesting result is that limited liability legal
form has a significant and positive effect on start-up size in both
countries yet, because of its greater cost, it is this legal form which
might be expected to deter entrepreneurs in an HR economy.
Our results on growth in new firms also show strong similarities
between the LR and HR economies of England and Spain.
The
proportion of new firms that become growers, statics and decliners is
very similar.
It is also the case that patterns of growth are very
similar, as are the factors explaining that growth.
There are, of course, some differences between the LR and HR
economies. For example, some founder-specific characteristics, some
37
industrial sectors and the use of external advice and sources of
finance differ between the countries, but not in a way that links them
clearly to regulatory burdens.
We also find instances of where our results directly conflict with our
theory. For example, we find that finance from public organizations is
a significant variable explaining employment change in NSFs in Spain.
The use by Spanish NSFs of publicly provided external advice in their
first year of trading is positively and significantly related to
subsequent employment growth.
This contrasts with the English
firms where it is not significant. This is the reverse of our theorizing
about LR and HR economies.
Our overall judgment, therefore, is that the simplistic view that an HR
economy produces smaller, slower growing firms in comparison with
an LR economy is not supported by the evidence in this paper. It does
not necessarily imply that regulation is irrelevant to the formation and
growth of NSFs, but perhaps suggests that for developed democratic
economies such as Spain and England the differences are less than
might have been expected.
One possible reason for this may be
because this paper, unlike official statistics on new firms, includes a
much wider coverage of very small firms. It may be that enterprise is
equally present in all regulatory regimes, but simply less likely to
appear in official data in an HR than in a LR economy.
38
This not-proven verdict points to the need in future research, as a
minimum, to extend the range and number of countries included.
But, given the importance of the topic to those currently formulating
public policy, our view is that the not-proven verdict should be a
cause for policy makers to pause and look again at the real impact
that regulation has on small and new firms.
39
6. REFERENCES
Acemoglu, D. and Pischke, J-S. (1998) “Why do firms train? Theory
and Evidence”, Quarterly Journal of Economics, Vol. 113 (1), pp.79119.
Almus, M. (2002) “What Characterizes a Fast Growing Firm”, Applied
Economics, Vol. 34, pp.1497-1508.
Almus, M. and Nerlinger, E.A. (1999) “The Growth of New Technologybased Firms: Which Factors Matter?”, Small Business Economics, Vol.
13, pp.141-154.
Basu, A. and Goswami, A. (1999) “Determinants of South Asian
Entrepreneurial Growth in Britain: A Multivariate Analysis”, Small
Business Economics, Vol. 13, pp. 57-70.
Baum, J.R., Locke, E.A. and Smith, K.G. (2001) “A Multidimensional
Model of Venture Growth”, Academy of Management Journal, Vol. 44,
No. 2, pp. 292-303.
Bertrand, M. and Kramarz, F. (2002) “Does Entry regulation hinder
job creation? Evidence from the French Retail Industry”, Quarterly
Journal of Economics, Vol. 117, No.4, November, pp. 1369-1413.
Brixy, U and Kohaut, S. (1999) “Employment Growth Determinants in
New Firms in Eastern Germany” Small Business Economics, Vol.13,
No.2, September, pp.155-170.
Bruderl, J and Preisendorfer, P. (2000) “Fast Growing Businesses:
Empirical Evidence from a German Study”, International Journal of
Sociology, Vol. 30, No 3, pp. 45-70.
Cooper, A.C., Gimeno-Gascon, F.J. and Woo, C.Y. (1994) “Initial
Human and Financial Capital As Predictors Of New Venture
Performance”, Journal of Business Venturing, Vol. 9, No 5, pp. 371395.
Djankov, S., La Porta, R., Lopez-de-Silanes, F. and Schleifer, A. (2002)
‘The Regulation of Entry’, Quarterly Journal of Economics, Vol. CXVII,
February, No.1, pp. 1-37.
Davidsson, P., Kirchhoff, B., Abulnassor, H.J. and Gustausson, H.
(2002) “Empirical Analysis of Business Growth Factors Using Swedish
Data”, Journal of Small Business Management, Vol. 40, No2, October,
pp. 332-349.
Delmar, F, Davidsson, P., and Gartner, W.B. (2003) “Arriving at the
high growth firm”, Journal of Business Venturing, Vol. 18, pp 189-216.
40
European Union (2002) Green Paper: Entrepreneurship in Europe,
European Commission: Brussels.
Evans, D.S. and Jovanovic, B. (1989) “An Estimated Model of
Entrepreneurial Choice under Liquidity Constraints”, Journal of
Political Economy, Vol. 97, No. 4, pp. 808-827.
Gimeno, J, Folta, T.B., Cooper, A.C. and Woo, C.Y. (1997) “Survival of
the fittest? Entrepreneurial human Capital and the Persistence on
Underperforming firms”, Administrative Science Quarterly, Vol. 42, pp.
750-783
Gorg, H., Strobl, E. and Ruane, F. (2000) “Determinants of Firm StartUp Size: An Application of Quantile Regression for Ireland”, Small
Business Economics, Vol. 14, pp. 211-222.
Greene, F.J., Mole, K. and Storey, D.J. (2004) “Does More Mean
Worse? Three Decades of Enterprise Policy In The Tees Valley”, Urban
Studies, Vol. 41, no. 7, pp. 1207-1228
Harhoff, D. Stahl, K. and Woywode, M. (1998) “ Legal form, Growth
and Exit of West German firms- Empirical results for manufacturing,
trade and service industries, Journal of Industrial Economics, Vol. 46,
no. 4, pp 453-488.
Kaufmann, D., Kray, A. and Zoldo-Lobaton, P. (1999) ‘Aggregating
Governance Indicators’, www.worldbank.org/wbi/governance/
Keter, V. (2004) Small Firms: Red Tape, House of Commons Library,
Research Paper, 04/52. House of Commons, London.
Mata, J. (1996) “Markets, Entrepreneurs and the Size of New Firms”,
Economic Letters, Vol. 52, pp. 89-94.
Mata, J. and Machado, J.A.F. (1996) “Firm Start-up Size: A
Conditional Quantile Approach”, European Economic Review, Vol. 40,
pp. 1305-1323.
Morton, F.M.S. and Podolny, J.M. (2002) “Love or Money? The effects
of owner motivation in the California wine industry”, Journal of
Industrial Economics, Vol. 50, No. 4, December, pp. 431-456
Nicoletti, G. and Scarpetta, S. (2003) ‘Regulation, Productivity and
Growth: OECD Evidence’, World Bank. New York.
Petersenm, M.A. and Rajan, R.G. (1994) “The benefits of lending
relationships – evidence from small business data”, Journal of
Finance, Vol. 49, No. 1, pp. 3-37.
41
Pfeiffer, F and Reize, F (2000) “Business start ups by the unemployedan econometric analysis based on firm data”, Labour Economics, Vol.
7, No. 5, pp. 629-663.
Pryor, F.L. (2000) “Quantitative notes on the extent of governmental
regulations in various OECD nations”, International Journal of
Industrial Organization, Vol. 20, pp. 693-714
Reid, G.C. and Smith, J.A. (2000) “What Makes a New Business StartUp Successful?” Small Business Economics, Vol.14, No.3, pp.165-182.
Scarpetta, S, Hemmings, P, Tressel, T and Woo, J. (2002) “The role of
policy and Institutions for Productivity and Firm Dynamics: evidence
from Micro and Industry data”, Economics Dept Working Paper, No.
329, OECD, Paris
Storey, D.J. (1982) Entrepreneurship and The New Firm, London:
Croom Helm.
Storey, D.J. (1994a) “The Role of Legal Status in Influencing Bank
Financing and New Firm Growth”, Applied Economics, Vol. 26, pp.129136.
Storey, D.J. (1994b) Understanding the Small Business Sector,
London: Routledge.
Storey D.J. and Strange, A. (1993) Entrepreneurship in Cleveland
1979-1989: A Study of the Effect of the Enterprise Culture, Employment
Department Research Series No.3, January.
Westhead, P., Birley, S. (1995) “Employment Growth In New
Independent Owner-Managed Firms in Great Britain”, International
Small Business Journal, Vol. 13, No. 3, pp. 11-34.
Wiklund, J., Shepherd, D. (2003) “Aspiring For, and Achieving
Growth: The Moderating Role of Resources and Opportunities”,
Journal of Management Studies, Vol. 40, No. 8, pp. 1919-1941.
World Bank (2005) “Doing Business in 2005: OECD High Income:
Regional Profile”, Monitoring, Analysis and Policy Unit, Investment
Climate Department.
42
Appendix A: Variable definitions and predicted signs
Predicted
sign
Pre-start variables
Definition
Male
= 1 if male; 0 if female
+
Age of founder
= age of founder in years when started the business
+
Age2 of founder
= (age of founder in years when started the business)2
-
Formal qualification
= 1 if founder has formal qualifications; 0 otherwise
+
Manager
= 1 if founder was a manager prior to the business starting; 0
otherwise
+
Unemployed
= 1 if founder was unemployed prior to the business starting; 0
otherwise
-
Support
= 1 if founder used external advice before start-up; 0 otherwise
+
Business plan
= 1 if founder had a formal written business plan prior to the business
starting; 0 otherwise
-
Start-up personal
savings
= 1 if firm used personal savings to establish the firm at start-up; 0
otherwise
+
Start-up clearing
bank loan
= 1 if firm used loans or overdrafts to establish the firm at start-up; 0
otherwise
+
Start-up friends or
relations
= 1 if firm used loans from friends/relations to establish the firm at
start-up; 0 otherwise
+
Start-up public
organizations
= 1 if firm had finance from public organizations to establish the firm
at start-up; 0 otherwise
+
43
Appendix A (continued): Variable definition and predicted sign
Predicted
sign
At-the-start variables
Definition
Start-up size
= number of persons employed at start-up, including founder
-
Limited co
= 1 if firm is a limited company; 0 otherwise
+
Manufacturing
= 1 if firm is in manufacturing; 0 otherwise
Construction
= 1 if firm is in construction; 0 otherwise
Trade
= 1 if firm is in wholesale and retail trade; 0 otherwise
Hotels and
restaurants
= 1 if firm is in hotels and restaurants; 0 otherwise
Transport
= 1 if firm is in transport and communication; 0 otherwise
Financial
intermediation
= 1 if firm is in financial intermediation; 0 otherwise
Business activities
= 1 if firm is in renting, real state or business activities; 0 otherwise
Education
= 1 if firm is in education; 0 otherwise
Health
= 1 if firm is in health and social work; 0 otherwise
Shropshire
= 1 if firm is in the county of Shropshire; 0 otherwise
Buckinghamshire
= 1 if firm is in the county of Buckinghamshire; 0 otherwise
44
Appendix A (continued): Variable definition and predicted sign
Predicted
sign
Post-start variables
Definition
Firm age
= number of years the firm has been trading
-
New products
= 1 if firm has introduced new products since founding; 0 otherwise
+
Formal plan now
= 1 if firm has a formal written business plan; 0 otherwise
+
Formal training for
workers
= 1 if firm conducts formal training for employees; 0 otherwise
+
1st year support
= 1 if firm has used external advice during first year of operation; 0
otherwise
+
Lower prices
= from 1 (much worse than competition) to 5 (much better than
competition)
-
Owner managerial
skills
= from 1 (much worse than competition) to 5 (much better than
competition)
+
Personal savings
= 1 if firm uses personal savings as a ‘current’ source of finance; 0
otherwise
-
Clearing bank loan
= 1 if firm uses bank loans or overdrafts as a ‘current’ source of
finance; 0 otherwise
+
Friends or relations
= 1 if firm uses loans from friends/relations as a ‘current’ source of
finance; 0 otherwise
-
Public organizations
= 1 if firm uses finance from public organizations as a ‘current’ source
of finance; 0 otherwise
+
45
Appendix B: ‘Pre-start’ variables affecting start-up size and employment growth: prior research findings
Study
Dependent variable
Firm age
Male
Cooper et al.
(1994), USA
Storey
(1994a), UK
Westhead
and Birley
(1995), UK
Relative employment
growth
Employment size per
year of life
(Present employmentemployment size when
firm received first
order)/age
Log of start-up size
Not given
+
Age of
founder
(Age of
founder)
Education
Unemployment
+
Prior
managerial
experience
n.s.
Use of
external
advice
n.s.
+
n.s.
-
+
2
Mata (1996),
Portugal
Almus and
Nerlinger
(1999),
Germany
Basu and
Goswami
(1999), UK
Brixy and
Kohault
(1999), East
Germany
Bruderl and
Preisendorfer
(2000),
Germany (B)
Mean of 4
years
Not given
+
-
n.s.
Not given
+
Gi=lnEti2-lnEti1/(ti2-ti1)
where E is employment
and ti is time
Firms up to 7
years old
+
Log yi=(Yt/Ys)1/(t-s)-1
where Yt is the ith
firm’s sales turnover in
period t and Ys is the
sales turnover in the
first year after start-up
Log of 1996
employment less
employment when first
registered
Fastest growing 4% of
start-ups. Growth is
>100% and at least 5
jobs
Mean of 19
years
+
+
+
n.s.
n.s.
+
n.s.
Up to 6 years
old
4 years
n.s
n.s.
n.s.
46
Study
Dependent variable
Firm age
Male
Age of
founder
(Age of
founder)
Pfeiffer and
Reize(2000)
Germany
Reid and
Smith (2000),
UK
Employment growth
Between 2
and 3 years
old
Mean of 2
years
n.s.
n.s.
n.s.
Education
2
Baum et al.
(2001), USA
Almus
(2002),
Germany
Davidsson et
al. (2002),
Sweden
Wiklund and
Sheperd
(2003),
Sweden
3 clusters of firms
reflecting performance
in profitability,
productivity and
employment change
Factor obtained using
sales, employment and
profit growth
Firms in the upper
10% of employment
growth distribution,
firms in the upper 10%
of the Birch Index
distribution
1996 employment
minus initial
employment divided by
the average of 1996
and initial employment
An index composed of
four variables (relative
change in sales and
employment, selfreported rate of sales
and employment
growth compared to
competitors)
Prior
managerial
experience
Unemployment
Use of
external
advice
n.s.
n.s.
-
-
Not given
Between 6
and 9 years
+
Some are
more than 25
years of age
Not given
n.s.
+
47
Appendix C: ‘At-the-start’ variables affecting start-up size and employment growth: prior research findings
Dependent variable
Firm age
Sector
Not given
Retail and personal services
Manufacturing
+
Log of industry size
(+); Log of MES (+) and
Suboptimal scale (-)
Study
Cooper et al.
(1994), USA
Storey (1994a),
UK
Mata (1996),
Portugal
Westhead and
Birley (1995),
UK
Almus and
Nerlinger (1999),
Germany
Basu and
Goswami (1999),
UK
Brixy and
Kohault (1999),
East Germany
Bruderl and
Preisendorfer
(2000), Germany
Pfeiffer and
Reize (2000),
Germany
Employment size per year of
life
Log of start-up size
Mean of 4
years
Not given
(Present employmentemployment size when firm
received first order)/age
Gi=lnEti2-lnEti1/(ti2-ti1) where E
is employment and ti is time
Not given
Firms up to
7 years old
High tech
+
Log yi=(Yt/Ys)1/(t-s)-1 where Yt is
the ith firm’s sales turnover in
period t and Ys is the sales
turnover in the first year after
start-up
Log of 1996 employment less
employment when first
registered
Mean of 19
years
Wholesale
-
Fastest growing 4% of startups. Growth is >100% and at
least 5 jobs
Employment growth
4 years
Up to 6
years old
Between 2
and 3 years
old
Energy/mining,
manufacturing,
construction, transportation,
business service
+
Data processing
+
Limited
company
Firm age
+
n.s.
+
-
Start-up
size
-
+
-
-
+
+
+
-
48
Appendix C (continued): ‘At-the-start’ variables affecting start-up size and employment growth: prior research
findings
Study
Dependent variable
Firm age
Reid and Smith
(2000), UK
3 clusters of firms reflecting
performance in profitability,
productivity and employment
change
Factor obtained using sales,
employment and profit
growth
Firms in the upper 10% of
employment growth
distribution, firms in the
upper 10% of the Birch
Index distribution
1996 employment minus
initial employment divided
by the average of 1996 and
initial employment
An index composed of four
variables (relative change in
sales and employment, selfreported rate of sales and
employment growth
compared to competitors)
Mean of 2
years
Baum et al.
(2001), USA
Almus (2002),
Germany
Davidsson et al.
(2002), Sweden
Wiklund and
Sheperd (2003),
Sweden
Sector
Limited
company
Construction, transport and
communication, bus. rel.
services (not-knowledge
based)
+
Healthcare, technical
consultants, knowledge
intensive services
+
n.s.
+
Firm age
Start-up
size
Not given
Between 6
and 9 years
Some are
more than 25
years of age
Not given
+
-
-
-
n.s.
n.s.
49
Appendix D: ‘Post-start’ variables affecting employment growth: prior research findings
Dependent variable
Firm age
Employment size per year of
life
Log of start-up size
Mean of 4 years
Study
Cooper et al.
(1994), USA
Storey
(1994a), UK
Mata (1996),
Portugal
Westhead and
Birley (1995),
UK
Almus and
Nerlinger
(1999),
Germany
Basu and
Goswami
(1999), UK
Brixy and
Kohault
(1999), East
Germany
Bruderl and
Preisendorfer
(2000),
Germany (A)
Pfeiffer and
Reize (2000),
Germany
(Present employmentemployment size when firm
received first order)/age
Gi=lnEti2-lnEti1/(ti2-ti1) where
E is employment and ti is
time
New
products
Formal
planning
Workforce
training
‘Current’
sources of
finance
Owner
managerial
skills
Not given
Not given
Personal
savings
-
Firms up to 7
years old
+
Log yi=(Yt/Ys)1/(t-s)-1 where Yt
is the ith firm’s sales
turnover in period t and Ys is
the sales turnover in the first
year after start-up
Log of 1996 employment less
employment when first
registered
Mean of 19
years
Up to 6 years
old
+
Fastest growing 4% of startups. Growth is >100% and at
least 5 jobs
4 years
+
Employment growth
Between 2 and
3 years old
+
50
Appendix D (continued): ‘Post-start’ variables affecting employment growth: prior research findings
Study
Dependent variable
Firm age
Reid and
Smith (2000),
UK
3 clusters of firms reflecting
performance in profitability,
productivity and employment
change
Factor obtained using sales,
employment and profit
growth
Firms in the upper 10% of
employment growth
distribution, firms in the
upper 10% of the Birch Index
distribution
1996 employment minus
initial employment divided by
the average of 1996 and
initial employment
An index composed of four
variables (relative change in
sales and employment, selfreported rate of sales and
employment growth
compared to competitors)
Mean of 2 years
Baum et al.
(2001), USA
Almus (2002),
Germany
Davidsson et
al. (2002),
Sweden
Wiklund and
Sheperd
(2003),
Sweden
Not given
New
products
Formal
planning
+
+
Workforce
training
‘Current’
sources of
finance
Bank loan
n.s.
Owner
managerial
skills
+
Between 6 and
9 years
Some are more
than 25 years
of age
Not given
51
Download