Ross2015 A Panel Data

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 17

Feature

Local Economy
Entrepreneurship and the 2015, Vol. 30(6) 672–688
! The Author(s) 2015

spatial context: A panel data Reprints and permissions:


sagepub.co.uk/journalsPermissions.nav

study into regional DOI: 10.1177/0269094215600135


lec.sagepub.com

determinants of small growing


firms in Scotland

Andrew G Ross
Edinburgh Napier University, UK

John Adams
British University in Egypt, Egypt

Kenny Crossan
Edinburgh Napier University, UK

Abstract
Entrepreneurship is widely recognised as a key determinant of economic growth, regional pros-
perity and sustainable development. Using a panel model with data from the value added tax
register, this paper estimates spatial variation in small growing firms across 32 Scottish regions for
the period 1998–2012. Results show there is considerable variation in small growing firms across
Scottish regions and may be explained by demand, supply, policy, cultural factors and agglomer-
ation benefits. Scotland has historically suffered from low levels of entrepreneurial activity; how-
ever, this paper provides relevant and timely findings into a form of entrepreneurial activity that
until now has largely been overlooked.

Keywords
entrepreneurship, panel data, regions, Scotland, small firms

Introduction

In modern market economies, entrepreneur-


Corresponding author:
ship is widely recognised to be a key deter-
Andrew G Ross, School of Management, Edinburgh
minant of economic growth, regional Napier University, Craiglockhart Campus, Edinburgh
prosperity and sustainable development EH14 1DJ, UK.
(OECD, 2008; Scottish Government Email: a.ross2@napier.ac.uk

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


Ross et al. 673

Economic Strategy, 2011; US Department rates of entrepreneurship between regions


of Commerce, 2007). Yet given the apparent within Scotland (Ross et al., 2012) and
importance of entrepreneurship, it is given that the Scottish Government is
remarkable how little is actually known responsible for economic development and
about the concept of entrepreneurship regeneration, it is surprising that there has
beyond start-up activity, both in terms of been a distinct lack of empirical research
types and forms of entrepreneurial activity addressing spatial aspects of entrepreneurial
and secondly the place-specific drivers that activity in Scotland, given that small enter-
explain geographical differences in regional prises account for 99% of all enterprises and
rates of entrepreneurial activity. This paper 53% of employment (Scottish Corporate
takes significant steps to address both of the Sector Statistics, 2012). However, although
aforementioned aspects by modelling regio- few studies exist on the extent and location
nal variation in determinants of continuing of entrepreneurial activity in Scotland,1 as a
entrepreneurial activity using panel data result of the perceived importance of entre-
across Scottish regions for the period preneurship to economic growth and
1998–2012. While the paper adds to the lim- employment, the last two decades have led
ited supply of entrepreneurial literature on research to focus increasingly on the spatial
small European countries, it is hoped the aspects of entrepreneurship. A review of the
research findings will be able to provide aforementioned literature reveals that the
guidance for government, local authorities vast majority of studies addressing the loca-
and other associate organisations involved tion of entrepreneurial activity have focused
in the design and implementation of enter- on the determinants of new firms or the loca-
prise policy at the local level. It is argued tion of existing industry. This literature has
here that a better understanding of the failed to address locational determinants of
environment for entrepreneurial activity in continuing entrepreneurial activity and/or
Scotland has never been more important for what may also be referred to as small grow-
Scotland’s long-term economic viability ing firms despite the evidence that growing
given that devolved powers are now to be businesses are likely to have a positive effect
extended significantly following the referen- on economic development (Botham and
dum result in September 2014. Graves, 2011).
While research on entrepreneurship has Indeed, NESTA (2009) highlighted that
increased significantly in recent decades, pre- while young firms are more likely to be
vious studies assessing the importance of growth orientated, the majority of growth
entrepreneurship have tended to concentrate firms (70%) are at least five years old.
on employment, innovation and growth Furthermore, a detailed examination by
effects (Botham and Graves, 2011; Carree NESTA of almost a quarter of a million
and Thurik, 2008; Van Praag and Versloot, UK start-ups founded in 1998 shows that
2007). Yet while entrepreneurship is widely the majority do not survive 19 years (62%),
accepted to be an important driver of both and of those that do, most remain small.
economic growth and development, empir- Only 10% of surviving businesses had more
ical studies strongly suggest that rates of than 10 employees 10 years later and fewer
entrepreneurial activity differ significantly than 5% had more than 20 employees. The
between regions and within countries implication being that simply encouraging
(Cheng and Li, 2011; Kangasharju, 2000; start-ups or new firm formation is unlikely
Okamuro and Kobayashi, 2006; Reynolds to lead to dramatic growth if they fail to
et al., 1994). This is particularly the case in expand. This is a simple yet important obser-
Scotland where despite distinct differences in vation implying regions that are not

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


674 Local Economy 30(6)

conducive to small growing businesses run a factors, agglomeration effects and institu-
greater risk of lower economic growth, tional and cultural determinants. Demand-
employment and regional competitiveness. side opportunities focus on the availability
Therefore, the ‘dilemma’ that is faced by of entrepreneurial opportunities linked to
policy makers is to better understand why the structural features of a regional econ-
regions with similar characteristics and size omy, while supply-side characteristics
vary so significantly in terms of rates of small focus on a region’s entrepreneurial capacity
growing firms, which we refer to as continu- (Audretsch, 2002). Institutional and cul-
ing entrepreneurial activity. tural variables are important as they often,
In order to address these shortcomings, respectively, underpin the incentive struc-
we put forward a framework which seeks to ture of an economy and the extent to
explain locational determinants of continu- which an activity is legitimised by society.
ing entrepreneurial activity, thus explaining
regional variation in rates of small firms in Demand-side factors
Scotland transitioning from the entrepre-
neurial start-up stage towards a growth Long-standing theoretical frameworks and
orientation for the first time. In short, we more recent empirical studies have suggested
argue that differences in regional rates of demand conditions are likely to be important
continuing entrepreneurial activity are a determinants of where firms choose to locate
reflection of region-specific characteristics (Audretch and Fritsch, 1994; Krugman, 1991).
and therefore entrepreneurial activity is pre- It is also possible to put forward the argument
dominately a regional event, whereby the that high or increasing demand is likely to
propensity for firms to grow is not only a encourage not only individuals to start busi-
reflection of individual preferences, but that nesses, but also to take advantage of positive
those decisions are also shaped and influ- market conditions and create opportunities for
enced by region-specific factors. the expansion of existing businesses.
The paper is structured as follows. The Therefore, it may be expected that an increase
next section presents the literature review in a region’s population growth rate will have a
and hypotheses, the subsequent section the positive effect on entrepreneurial activity
data and method, ‘Spatial patterns of conti- (Armington and Acs, 2002; Reynolds et al.,
nuing entrepreneurial activity’ section 1995). Furthermore, income levels can also
describes regional variation in rates of conti- be expected to affect demand for local busi-
nuing entrepreneurship, the empirical results nesses (Reynolds et al., 1994). As wages
are presented in the following section and the increase demand may also rise as a result of
final section concludes with some comments increased purchasing power thereby having a
and recommendations for further research. positive impact on the rate of continuing entre-
preneurial activity. Therefore, spatial differ-
Literature review and ences in the demand for goods and services
hypotheses are likely to influence the demand for continu-
ing entrepreneurship and it is likely that spatial
The conceptual framework adopted in this variations occur as a result of differences or
study is similar to those adopted in other changes in local demand conditions. To sum-
spatial studies (Audretsch and Fritsch, marise this argument:
1994; Bosma et al., 2008). In a similar fash- H1: High demand has a positive effect on
ion to Bosma et al. (2008), the explanatory the rate of continuing entrepreneurial
variables are grouped into four broad cate- activity.
gories, including demand and supply

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


Ross et al. 675

Supply-side factors changing environment, then this should be


evidenced via variations in entrepreneurs’
Previous studies have highlighted that the level of education or human capital. Some
level of unemployment may impact the empirical studies (Armington and Acs,
supply of entrepreneurship. However, 2002; Lee et al., 2004) have found that
based on previous findings the relationship regions with a higher number of university
between unemployment and entrepreneurial graduates are more likely to have higher
activity is not clear. On the one hand, ‘push’ levels of entrepreneurial activity than those
theory suggests increasing levels of regions with less skilled workers. Thus, in
unemployment reduce the prospects for our theoretical model of continuing entre-
finding paid employment and as a result preneurial activity, if we follow Schultz’s
the expected returns from entrepreneurship argument, it is more likely that regions
become more attractive pushing people into that have a higher stock of educated
undertaking entrepreneurial activity (Evans people generate greater human capital
and Leighton, 1990; Storey, 1991). Thus, embodied by their general and specific
the recession-push hypothesis suggests that skills, for implementing new ideas; not
worsening economic conditions activate only for creating new businesses, but also
previously dormant entrepreneurial ambi- importantly growing existing businesses.
tions, pushing individuals towards self- Furthermore, regions with a rich stock of
employment as a result of an actual or human capital can generate environments
likely negative change in labour market rich in local spillovers which are another
conditions. support mechanism by which existing busi-
On the other hand, according to ‘pull’ nesses can grow and be sustained.
theory individuals are more likely to con- The indicator of human capital used in
duct entrepreneurial activity under condi- this study is the percentage of the popula-
tions of economic expansion, when tion with a National Vocational
incomes are growing and opportunities are Qualification 4 or above (NVQ4þ).2 This
strong for market specialisation. proxy indicator of human capital accounts
Accordingly, there should be far less confu- for both skill-based vocational qualifica-
sion around the impact of unemployment tions, gained by tradesmen, who are highly
on continuing entrepreneurial activity since skilled in their field of work yet do not hold
high unemployment will inhibit market a tertiary qualification such as a degree; it
demand for goods and services within a also accounts for traditional academic
region. It will expose existing entrepreneurs qualifications including undergraduate and
to greater risk of falling incomes and pos- postgraduate qualifications which are more
sible bankruptcy. This logically implies that: closely associated with knowledge acquisi-
tion. Therefore, we hypothesis that:
H2: Increasing unemployment has a nega-
tive effect on the rate of continuing entre- H3: Higher human capital has a positive
preneurial activity. effect on the rate of continuing entrepre-
neurial activity.
Human capital has sometimes been viewed
as the key variable that influences economic The likelihood and practical reality of both
development (Lucas, 1988) and if we adopt starting and growing any business is likely
Schultz’s theoretical argument that human to be constrained by the availability of or
capital ought to be viewed mostly as the access to finance. However, unlike start-up
capacity of the entrepreneur to adapt in capital the availability of capital for busi-
order to deal with disequilibrium and/or a ness expansion usually represents an

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


676 Local Economy 30(6)

amount larger than what is available to be of industry and new firm formation are
financed through borrowing from family likely to be influenced by a range of external
and friends or by using personal credit agglomeration factors. Our model similarly
(Mason, 1991; Morky, 1988) with debt cap- advocates that the benefits of agglomeration
ital as the most likely source of external or centrifugal forces, as Krugman (1991)
financing among small growing firms puts it, will have a positive effect on conti-
(Equinox Management Consultants Ltd, nuing entrepreneurial activity, as growing
2000); even though there are barriers asso- firms will attempt to take advantage of
ciated with debt financing for small firms, pecuniary and non-pecuniary externalities
such as collateral guarantees required in such as access to a pool of well-qualified
order to obtain financing from banks and labour, the existence of specialised suppliers
the higher interest rates typically charged and knowledge spillovers, thus allowing
to smaller firms. them to lower transaction and transporta-
Following resource-based theory which tion costs and ultimately benefit from
argues that an entrepreneur will start a busi- increasing returns in order to remain com-
ness when he has sufficient resources to do petitive. Agglomeration economies can be
so (Braunerhjelm, 2007; Cooper et al., 1994; sub-divided into urbanisation economies
Penrose, 1959), in our model we suggest and specialisation economies. However,
there is strong reason to believe access to results regarding whether specialisation or
finance will influence the supply of continu- urbanisation economies are strongest have
ing entrepreneurial activity, whereby the proved inconclusive (Beaudry and
wealthier a region is the easier it should be Schiffauerova, 2009) and while it can be
to obtain finance for business expansion. accepted that economic activity in general
Specifically, we adopt the dynamic indi- is not evenly distributed, it cannot be
cator of capital gains in house prices. This agreed as to the factors that cause this
allows that an individual may live in a resi- uneven distribution of economic activity.
dence, which on paper has a high net worth, Therefore, at this stage although our
yet the individual may bear the financial model predicts agglomeration economies
responsibility of a large mortgage making should have a positive effect on continuing
it less likely that a bank will lend to a entrepreneurial activity we cannot say
highly geared individual, whereas regions whether continuing entrepreneurial activity
with an increasing growth rate in house is driven by specialisation economies or
prices indicate net gains in an individual’s urbanisation economies. However, we
personal wealth and the greater likelihood hypothesise that:
of raising finance based on the value of their H5: Economic agglomeration has a posi-
property. Therefore, we hypothesise that: tive effect on rates of continuing entrepre-
H4: Rates of continuing entrepreneurial neurial activity.
activity are higher in regions with greater
levels of personal wealth.
Institutional and cultural variables

Agglomeration factors Mason (1991) highlighted that culture is


relevant to entrepreneurship from two per-
A large body of theoretical and empirical spectives. Firstly, it gives an indication of
studies tells us that economic activities are attitudes towards entrepreneurship among
clustered in space (Krugman, 1991; the local population and secondly shows
Marshall, 1920) and that both the location the extent to which local institutions are

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


Ross et al. 677

supportive in terms of political leadership. entrepreneurial activity. In particular,


Therefore, the institutional context is there is good reason to believe the presence
important as political and economic institu- of a large existing small business population
tions underlie and determine the incentive will imply a positive attitude and culture
structure of an economy (North, 1991) towards entrepreneurship, whereby existing
which in turn can either encourage or firms are able to act as seedbeds and incu-
hinder an entrepreneurial culture. bators for entrepreneurs looking to grow
One of the key aspects of product market their businesses. Our model argues a large
institutions is the level of competition. small business population allows growing
However, competition refers not only to businesses to build contacts and develop
competition between firms in the private relationships which help facilitate the
sector but also between private firms and exchange of information relating to various
the public sector. Indeed, previous studies aspects of their business with increased
have found that a large government sector access to a variety of relevant and reliable
can have a negative impact on entrepreneur- information, thereby helping to improve the
ship (Bjørnskov and Foss, 2008; Nystrom, speed and quality of decision-making pro-
2008). Our own model argues, given that cesses within a business, and thus increasing
28%3 of Scotland’s workforce is employed their prospects for growth. Therefore, we
in the public sector this makes it less likely predict that the number of existing small
that public sector workers will be exposed businesses measured as the proportion of
to the skills required to run and grow a small businesses with <50 employees in the
small business and that it may be expected overall business population within a region
that a large public sector is also likely to acts as a proxy for a positive entrepreneurial
crowd out private sector firms as a result culture and it is expected that a region with
of competition. a positive culture towards entrepreneurship
We capture the effect of the institutional is more likely to have higher rates of con-
environment on continuing entrepreneurial tinuing entrepreneurial activity. Hence,
activity by using the proxy measure of the H7: A large proportion of small firms have
proportion of the work force employed in the a positive effect on rates of continuing
public sector in each region to assess the entrepreneurial activity.
influence of the institutional environment
on continuing entrepreneurial activity in
Scotland. We hypothesis that: Methodology
H6: A large public sector will be negatively
associated with the rate of continuing Data and method
entrepreneurial activity.
Dependent variable. Concerned not with the
Social and cultural norms will also influence broad concept of entrepreneurship, but with
the number of people who have already a subset of the overall entrepreneurial popu-
undertaken entrepreneurial activity in a lation our framework assesses regional fac-
region (Aoyama, 2009; Etzioni, 1987; tors that determine the location of existing
Malecki, 2009). Therefore, it is likely that small businesses demonstrating signs of
levels of entrepreneurship will be influenced small firm growth. Therefore, the indicator
by the way the local population perceives of entrepreneurial activity adopted in this
the economic climate with higher rates of study is the annual rate of firms transition-
entrepreneurship acting as an indicator of ing to Value Added Tax (VAT) registered
positive attitudes and opportunities for status; that is firms passing a turnover

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


678 Local Economy 30(6)

threshold requiring them to pay VAT with Independent variables. In order to explain the
these firms acting as a proxy measure of causes of regional variation in continuing
continuing entrepreneurial activity and entrepreneurial activity, the theoretical
small business growth. While VAT registra- framework and research hypotheses out-
tion statistics are publically available via the lined in the previous section consisted of
ONS ‘Business Demography’ publication four major factors: (1) demand-side factors,
and have been widely used in entrepreneur- (2) supply-side factors, (3) agglomeration
ial studies before (Ashcroft et al., 1991; factors and (4) institutional and cultural
Gleave and Mitra, 2010; Keeble and factors. The 10 specific indictors used to
Walker, 1994) VAT registration data are assess the four major factors are sum-
not without its limitations as highlighted marised in Table 1. Their choice reflects
by Ross et al. (2012). Nevertheless, VAT both the availability of suitable data and
registration data remain the most compre- informed judgement regarding processes
hensive measure of continuing entrepre- likely to have a major impact on continuing
neurial activity available in the UK and entrepreneurial activity in Scotland.
are officially recognised as a measure of Following Reynolds et al. (1994) and
entrepreneurial activity by the Scottish Sutaria and Hicks (2004) a one-year lag is
Government and local authorities. incorporated into the independent variables
As a result of differences in population as continuing entrepreneurial activity in
and region size, it is necessary to standard- part is likely to be a reflection of what has
ise the dependent variable (Ashcroft et al., happened previously.
1991; Ross et al., 2012). The dependent
variable can be standardised by two Empirical approach. In order to identify and
approaches: the ‘labour’ market approach analyse regional determinants of continuing
and the ‘ecological’ approach. The labour entrepreneurial activity, a panel data set
market approach standardises the number was constructed for Scotland’s 32 council
of new entrants relative to the size of a regions for the period 1998–2012. Panel
region’s population and/or workforce. The data estimation differs from regular cross
benefit of this approach is that it can indi- section and time series estimation in that
cate a region’s entrepreneurial potential, panel data is a combination of both cross
based on the assumption that entrepreneur- section and time series. In order to mitigate
ial activity is most likely to arise from the against the weaknesses of the basic OLS
actions of an individual/individuals within a pooled model, a robust individual effects
given region (Cheng and Li, 2011; Sutaria estimator is used. The individual effects
and Hicks, 2004). Alternatively, the eco- model can be written as
logical approach standardises entrants rela-
tive to the stock of existing businesses. Yit ¼ i þ Xit  þ it
ð1Þ
Selecting an appropriate method can be cru- i ¼ 1, 2, . . . , N; t ¼ 1, 2, . . . , T
cial, as alternative methods often produce
differing results. Given we are concerned where i denotes the individual region and t
with a region’s socio-economic characteris- denotes time. i is the individual intercept, 
tics and the extent to which they influence is k  1 and Xit the itth observation on K
an individual’s likelihood to grow and explanatory variables. When modelling
expand their business, the dependent vari- regional heterogeneity with an individual
able is standardised by the working popula- effect model one must decide between a
tion, as this is the most likely source of fixed or random effects estimator.
entrepreneurial activity. However, major differences exist in the

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


Ross et al. 679

Table 1. Explanatory variables.

Expected
Explanatory variable Operational definition effect Data source

Demand and supply factors


Wage rate Annual wage rate þ Office of National Statistics
Population growth Annual population growth þ General Register Office for
change Scotland
Unemployment rate (log) Annual % of unemployed  Office of National Statistics
Education % of population with NVQ 4 or þ ONS: Annual population
higher Survey
Access to finance Annual growth in house prices þ Scottish Neighbourhood
Statistics
Agglomeration factors
Specialisation economies Number of firms in manufac- þ/ IDBR: VAT register
(manufacturing) turing sector relative to the
total business population
(location quotient)
Specialisation economies Number of firms in the busi- þ/ IDBR: VAT register
(business services) ness sector relative to the
total business population
(location quotient)
Urbanisation economies Population density, measured þ/ General Register Office for
as inhabitants divided by Scotland
regional area (km2)
Institutional and cultural
factors
Size of public sector % of work force employed in  ONS: Annual population
the public sector Survey
Small business population % of small businesses with <50 þ IDBR: VAT register
employees in the overall
business population

way i is assigned and in the assumptions Therefore, the fixed effect model can be
made between fixed and random effects expressed as
models. The key differentiating factor
between a fixed and random effects model Yit ¼ ðþ i Þ þ Xit  þ it
ð2Þ
is the way the individual-specific error com- i ¼ 1, 2, . . . , N; t ¼ 1, 2, . . . , T
ponent is modelled. In the fixed effect
model, it is assumed to be part of the inter- where in the fixed effect model i are
cept, while in the random model it forms assumed to be fixed parameters to be esti-
part of the error variance. mated, it is the remaining stochastic dis-
The fixed effect model assumes that the turbance, which is assumed to be
individual effect is captured by the intercept independent and identically distributed
term i, which means that every individual IID (0 2 ). In the fixed effect model, the
region has their own intercept and that this individual effect ( þ i) is allowed to be
individual effect varies across groups. correlated with the independent variables

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


680 Local Economy 30(6)

Xit, while assuming that Xit remains uncor- Gateway services which provide assistance
related and independent of idiosyncratic to all businesses and are one of the two
error (it) for all i and t. However, in the main mechanisms offering support for
random effects model i is assumed to be entrepreneurs in Scotland.4 Alternative
randomly distributed with a constant options included utilising NUTS 25 or
mean and variance, but crucially that i is NUTS 3 data; however, the major weakness
uncorrelated with the regressors Cov of NUTS 2 data in the case of Scotland is
(Xit,ai) ¼ 0, for all t. Therefore, in the the failure to account for the entrepreneurs’
random model the individual effect is trea- local environment because of the degree of
ted as a random component and part of the aggregation and the limited number of
error structure, not the intercept. The NUTS 2 regions. NUTS 3 regions would
random effects model can be written as be more appropriate, but in Scotland these
do not correspond fully with the data sets
Yit ¼  þ Xit  þ ði þ it Þ
ð3Þ used in this study.
i ¼ 1, 2, . . . , N; t ¼ 1, 2, . . . , T
Spatial patterns of continuing
where the only difference i is now part of entrepreneurial activity
the error term and not the intercept.
Therefore, the random effects model meets Table 2 shows per capita rates of continuing
all of the same assumptions as the fixed entrepreneurial activity across UK
effects model plus the additional require- Government Office Regions. Table 2 reveals
ment that the individual effect i is uncor- that four regions have above average per
related with the regressors in all time capita start-up rates in the UK. These
periods (in other words, the individual regions are located in the South of
effect is constant). Ultimately, the common England and particularly in the South East
way of selecting between the fixed and of England: London (112.6), the South East
random effects model is to conduct a (88.6), East (81.2) and the South West
Hausman test but this test is only robust (76.5). London has by far the highest level
under the assumption of homoscedasticity. of continuing entrepreneurial activity in the
Instead Hansen’s J statistic is appropriate UK especially when compared with
when deciding between a fixed and Scotland, Wales and Northern Ireland.
random effect estimator when using robust Scotland lies ninth from 12th with 57.3
standard errors. On the basis of Hansen’s J registrations per 1000 of working-age popu-
statistic reported in ‘Empirical results’ sec- lation ahead only of Wales, Northern
tion a random effects estimator is used to Ireland and the North East region of
model variation in regional rates of continu- England which coincidently are among
ing entrepreneurship. some of the most underdeveloped regions
in the UK. Table 2 also reveals a clear
Unit of analysis. The principal unit of analysis North–South divide and elements of a
is the local authority council region, as this core–periphery relationship with continuing
is the most spatially disaggregated level entrepreneurial activity concentrated in the
at which to investigate the local environ- most economically prosperous parts of the
ment for entrepreneurship in Scotland. United Kingdom.
Furthermore, each council region is respon- While mapping rates of continuing entre-
sible for economic development and regen- preneurship at the GOR level is useful and
eration in their respective regions and provides a broad overview of continuing
has responsibility for delivering Business entrepreneurship at a UK level, as a result

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


Ross et al. 681

Table 2. Rates of continuing entrepreneurial the lowest rates of continuing entrepreneur-


activity, 1998–2012 per 1000 of working-age ship include: West Dunbartonshire (35.1),
population. Inverclyde (40.6) and Dundee City (43.2)
all of which are regions formally dominated
Region Rate
by heavy industry in the past. Appendix 1
London 112.6 also shows that some of the highest levels of
South East 88.6 entrepreneurial activity are located in the
East 81.2 more northern regions of Scotland and in
South West 76.5 rural areas reflected by 10 of the above aver-
East Midlands 68.4 age 13 regions being considered rural.6
West Midlands 68.3
However, unlike at the UK GOR level
North West 67.5
there is no evidence of a core–periphery
Yorkshire and The Humber 63.1
Scotland 57.3 relationship, with many of the most entre-
Wales 55.4 preneurial regions being considered as rural
Northern Ireland 52.4 and located on the periphery, while some of
North East 48.1 the least entrepreneurial regions are located
UK 75.7 in the urban central belt and south western
parts of Scotland. However, it is worth
Source: Office of National Statistics.
highlighting that both of Scotland’s major
cities, which are located in the Eastern and
of aggregation it fails to account for the Western parts of the central belt have
local environment of the entrepreneur and above-average entrepreneurial activity:
is not, therefore, able to advise about con- Edinburgh (70.8) and Glasgow (63.2).
tinuing entrepreneurial activity within Aberdeen City (71.6) is Scotland’s most
regions of a country, which is the principal entrepreneurial city, while Dundee (43.2) is
unit of analysis under investigation in this the least entrepreneurial. With the exception
paper. Therefore, the remainder of this sec- of the Edinburgh and Glasgow city regions,
tion maps spatial variation in rates of con- it is also worth noting the relatively low
tinuing entrepreneurship at the local levels of entrepreneurial activity in the cen-
authority council level in Scotland. tral belt overall and the south west of
Appendix 1 shows that there is consider- Scotland.
able spatial variation in continuing entre-
preneurial activity across Scottish regions. Empirical results
Average rates of continuing entrepreneur-
ship in Scotland range from a low of 35.1 As previously mentioned selecting the ‘cor-
(West Dunbartonshire) to a high of 76.0 rect’ econometric model is not always
(Stirling) with a Scottish average of 57.3 straightforward. However, the choice of
registrations per 1000 of working-age popu- whether to estimate coefficients using a
lation. In total there are 13 regions with fixed or random effects model is made
above average rates of continuing entrepre- using Hansen’s J Test, which tests if the i
neurship. The highest level of continuing are uncorrelated with the independent vari-
entrepreneurial activity is in Stirling (76.0) ables. In this case Hansen’s J test produced a
followed by two prosperous regions in the x2 result of 12.404 indicating we cannot
North East: Aberdeenshire (75.5) and reject the null hypothesis and that it is appro-
Aberdeen City (71.6), both regions benefit- priate to use the random effects model.
ing from being located at the centre of the Table 3 displays the econometric results.
UK oil and gas industry. Those regions with Overall, the model is statistically significant

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


682 Local Economy 30(6)

Table 3. Determinants of regional variation in continuing entrepreneurship: Scottish Regions,


1998–2012.

Independent variable All regions Urban regions Rural regions

Demand and supply factors


Wage rate 0.331 (0.011)*** 0.065 (0.014)*** 0.018 (0.014)
Population growth 0.243 (0.292) 1.671 (0.648)** 0.017 (0.302)
Unemployment rate 4.473 (1.144)*** 0.338 (1.677) 4.958 (2.618)**
Education 0.123 (0.069)* 0.192 (0.074)** 0.100 (0.111)
Access to finance 6.343 (2.159)*** 0.850 (2.625)* 7.812 (3.311)***
Agglomeration factors
Specialisation economies (manufacturing) 0.523 (0.590) 0.570 (0.766) 0.527 (0.726)
Specialisation economies (business services) 0.827 (2.275) 8.872 (2.757)*** 2.777 (3.440)
Urbanisation economies 0.003 (0.001)** 0.003 (0.000)*** 0.080 (0.062)
Policy and cultural factors
Size of public sector 0.188 (0.066)*** 0.325 (0.089)*** 0.210 (0.103)**
Small business population 1.551 (0.418)*** 0.576 (0.182)*** 1.486 (0.943)
Constant 195.725 (44.38)*** 62.90 (26.33)** 197.52 (107.508)
Breusch–Pagan test of independence 608.16 (0.0000) 7.25 (0.0000) 106.22 (0.0000)
Hansen’s J statistic (X2) 12.404 (0.2589) 47.546 (0.000) 17.135 (0.0714)
R2 0.42 0.74 0.36
N 480 270 210

Significance at the 0.01, 0.05 and 0.10 level indicated by ***, ** and *. 1. Robust standard errors used to correct for
heteroscedasticity.

explaining 42% of the variation in regional demand generates not only opportunities
levels of continuing entrepreneurial activity. for new businesses, but also small existing
The urban and rural models are able to businesses who can take advantage of posi-
explain 74 and 36% of variation, respect- tive market conditions and opportunities
ively. Results show that demand and for expansion. Income levels proxied by
supply, agglomeration and institutional regional wage rates were significant at the
and cultural factors are all significant fac- Scottish level and within urban regions
tors explaining regional differences in rates highlighting that increasing disposable
of continuing entrepreneurial activity. incomes imply increased purchasing
Perhaps most challenging from a policy power, generating growth opportunities
design perspective is that our results, as pre- for existing businesses. Population growth
dicted, highlight that regional rates of entre- as predicted has a positive and statistically
preneurial activity are not universal across significant effect, which is consistent with
regions with differences existing between earlier empirical studies addressing start-
determinants of entrepreneurial activity up entrepreneurship (Audretsch and
dependent upon whether a region is urban Fritsch, 1994; Reynolds et al., 1994;
or rural. Tamasy and Le Heron, 2008). However,
Similar to other studies related to general this effect was only significant in urban
levels of entrepreneurship both demand regions and supports the theoretical work
variables are in line with the business cycle of the New Economic Geography indicating
hypothesis and found to have a positive sig- market size matters whereby firms are
nificant effect on small growth-orientated drawn to locate where markets are largest
firms demonstrating that increasing as a means of reducing transportation costs

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


Ross et al. 683

and achieving economies of scale giving Economic Geography (Krugman, 1991)


individuals the best opportunity to grow a and to offer an explanation for statistical
business. Co-incidentally given demand was significance at the urban regional level,
not a significant factor in rural regions may entrepreneurs may be more attracted
indicate that local demand conditions might towards regions where they can access a
be less important over time for some critical mass of well-qualified labour allow-
growth-orientated businesses, as companies ing businesses to benefit from and be sus-
in these regions may be more reliant upon tained by continuing knowledge spillovers
inter-regional and international export leading to additional rounds of new innova-
markets. tive activity and business expansion.
The unemployment rate was found to be Access to finance was confirmed as a sig-
strongly negative and statistically significant nificant determinant of continuing entrepre-
at the Scottish level and in rural regions neurial activity in Scotland and within both
indicating that continuing entrepreneurial urban and rural regions suggesting the
activity in Scotland is opportunity driven greater the access to finance the lower the
and the effect of a given business cycle is barriers are to business growth and expan-
greater than any unemployment push sion. This would appear to be a particularly
effect. In particular, the results offer greater relevant point because there is some evi-
clarity compared with previous contradict- dence to suggest that debt finance is the
ory studies assessing the effect of unemploy- most likely source of external finance for
ment on start-up entrepreneurship business expansion (Equinox Management
(Audretsch and Fritsch, 1994; Audretsch Consultants Ltd, 2000) and that this form
et al., 2010; Bosma et al., 2008; Okamuro of financing is most likely to be obtained
and Kobayashi, 2006; Reynolds et al., 1994) from banks (Binks et al., 1988). Yet banks
by clearly indicating that regions with nega- are often reluctant to provide unsecured
tive labour market conditions are likely to lending, in case of default and the uncer-
feel a dampening effect on purchasing tainty associated with lending to small busi-
power as a result of lower levels of dispos- nesses where success and profit are never
able income. This re-enforces the argument guaranteed or certainly not realised until
that strong and stable economic conditions an unspecified later date. Furthermore,
are likely to lead to higher levels of entre- given that personal wealth is often used as
preneurial activity rather than weak or collateral to raise debt finance in a country
weakening economic conditions. such as Scotland, which has traditionally
Consistent with other studies had lower levels of homeownership relative
(Armington and Acs, 2002; Lee et al., to some other parts of the UK, this may
2004; Okamuro and Kobayashi, 2006) our mean banks are less willing to lend in
results indicate educational attainment sig- regions that are financially constrained
nificantly influences rates of entrepreneurial through high unemployment, low house
activity at both Scottish and urban regional prices and widespread social housing
level. From the perspective of the small which may reflect the low rates of continu-
growing firms, this suggests first, not only ing entrepreneurship found in the west of
is a more educated and skilled person likely Scotland.
to establish a business, but there is a greater Agglomeration externalities were found
likelihood of that business experiencing to have a significant positive effect.
growth. Second, in line with the theory of However, the form of agglomeration is
industrial districts (Marshall, 1920), dependent upon both the industry and
Growth Pole Theory and the New whether a region is urban or rural; and in

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


684 Local Economy 30(6)

some cases both specialisation and urban- private sector firms in Scotland are being
isation externalities were found to coexist crowded out by a large public sector and
within certain types of region. At national this is clearly an area which requires further
and urban region level the importance of a investigation.
critical mass of inter-industry spillovers was Finally, the existing small business popu-
detected, supporting the idea that diversity lation within a region, which acts as a proxy
and variety of industry is likely to be great- for culture, is strongly positive and statistic-
est in more densely populated regions and ally significant at the Scottish level and in
cities where firms can benefit from improved urban regions. This is an important finding
and advanced infrastructure, large labour for two reasons. First, a large small business
markets and the support provided by population acts as a proxy for how well
access to a wide range of support firms. entrepreneurial activity is accepted and
Yet simultaneously, at intra-industry level legitimised by society, implying a positive
within urban regions, Marshallian localisa- attitude and an enterprising culture and
tion economies play an important role in second, regions with a large body of existing
the location of where small growing firms small businesses can also act as role models
in the business services industry choose to and incubators providing advice, but
locate, appearing to take advantage of importantly allowing business owners to
Porter’s cluster effects. The findings indicate build contacts and develop networks both
densely populated regions and cities in par- formal and informal, which as previously
ticular benefit from both specialisation and discussed increases the variety of relevant
urbanisation economies in a way rural and reliable information leading to
regions do not and may help explain why improved speed and quality of decision
city regions often grow faster than other making, increasing not only the chances
regions. Indeed this view is supported by for success, but also business growth and
the evidence that three of Scotland’s four expansion.
city regions have above-average rates if con-
tinuing entrepreneurship. Conclusion and
From an institutional perspective, the recommendations
size of the public sector workforce was
found to have a negative statistically signifi- The model of continuing entrepreneurial
cant effect on the level of continuing entre- activity developed in this paper relates the
preneurial activity at national, urban and emergence of small growing firms to their
rural levels, which is a significant finding local environment and the factors that influ-
at the policy level from a number of per- ence that environment. The results highlight
spectives. First, it clearly identifies that the that rates of continuing entrepreneurial
institutional context is critical to levels of activity can most significantly be explained
entrepreneurial activity, given that it under- by local demand conditions, the number of
lies and determines the incentive structure existing small businesses (an indicator of
of any economy. Second, our findings and attitudes and culture towards entrepreneur-
Scotland’s relatively large public sector may ship), human capital, access to finance and
offer an insight into why Scotland has his- the presence of agglomeration externalities.
torically lagged behind other UK regions in It is also clear that the unemployment rate
both entrepreneurial activity and economic and Scotland’s large public sector work-
growth over a 30-year period. Third, and force has a strong negative and statistically
somewhat worryingly the findings may significant effect on rates of continuing
demonstrate that small growth-orientated entrepreneurship across Scottish regions.

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


Ross et al. 685

From a policy perspective our results, as encourages and actively supports an entre-
predicted, highlight that rates of continuing preneurial culture.
entrepreneurial activity are dependent upon Furthermore, it is important to ensure
region-specific factors and are therefore not that business numbers alone do not simply
universal across regions. It was also increase, but that these businesses are of a
observed that rates of continuing entrepre- high quality, capable of driving both eco-
neurial activity in Scotland are some way nomic growth and the creation of high-
below the UK average which may account quality employment opportunities which
for Scotland’s prolonged and low level of will ensure both higher and sustainable
economic growth over a 30-year period. levels of economic growth at a national
The emphasis placed on entrepreneur- and regional level in Scotland. As a conse-
ship at all levels of government suggests quence it is clear that entrepreneurship has
that our results have strong policy implica- a much wider impact on the social fabric
tions. However, rather than focussing on that underpins society and therefore polices
making specific policy recommendations, to promote and encourage dynamic innova-
which is likely to be counterproductive tive-driven regions are crucial for future
and counterintuitive given the heterogeneity prosperity and sustainable development.
in rates of continuing entrepreneurship Therefore, continued spatial research into
identified across regions, it is hoped the determinants of entrepreneurship in general
research findings will provide broad guid- and continuing entrepreneurship in particu-
ance for those involved in the design and lar are and always will be required on an
implementation of enterprise policy. Yet ongoing basis in each country.
because of regional heterogeneity our
results demonstrate that it is crucial that
individual regions should continue to have Declaration of conflicting interests
their own unique set of tailored policies The author(s) declared no potential conflicts of
which reflect the needs, capabilities and interest with respect to the research, authorship,
institutional framework of that region, as and/or publication of this article.
attempting to implant policy approaches
from other regions is no guarantee of suc-
cess (Sternberg, 2009). Indeed, if all policy Funding
recommendations were generalisable it The author(s) received no financial support for
would be unlikely that major differences in the research, authorship, and/or publication of
rates of continuing entrepreneurial activity this article.
would exist between countries and certainly
between regions within countries.
Notes
Nevertheless, at a broad level and on the
basis of the results reported here, those 1. Refer to Ashcroft et al. (1991).
responsible for implementing enterprise 2. HND, Degree and Higher Degree level quali-
policy in Scotland must recognise and do fications or equivalent.
3. Figures taken from the Annual Population
more to address the entrepreneurial deficit
Survey and Annual Labour Force Survey
that exists between regions and generate a based on local Authority figures for those
more positive culture towards entrepreneur- aged 16–64.
ship in Scotland. This should involve (at a 4. The others being both Scottish Enterprise and
greater speed) the continued development Highland’s & Island’s Enterprise, but they a
of institutions and levers that are capable primarily focused on supporting high growth
of providing an environment which trajectory companies.

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


686 Local Economy 30(6)

5. The NUTS classification (nomenclature of versus urbanisation debate. Research Policy


territorial units for statistics) is a hierarchical 38(2): 318–377.
system for dividing up the economic territory Binks M, Ennew C and Reed G (1988) Small
of the EU for the purpose of (1) the collec- business and banks: A two nation perspective,
tion, development and harmonisation of EU prepared for the Forum of Private Business,
regional statistics; (2) socio-economic ana- UK and the National Federation of
lyses of the regions and (3) framing of EU Independent Businesses, USA.
regional policies. Bjørnskov C and Foss NJ (2008) Economic free-
6. The Randall definition is based upon popula- dom and entrepreneurial activity: Some cross
tion density within a unitary authority country evidence. Public Choice 134(3): 307–328.
(council region). Where a unitary authority Bosma N, Van Stel A and Suddle K (2008) The
has a population density of less than one geography of new firm formation: Evidence
person per hectare it is considered rural. On from independent start-ups and new subsidi-
this basis there are 14 rural unitary authori- aries in the Netherlands. International
ties. Unitary Authority data are readily avail- Entrepreneurship and Management Journal
able and it is therefore very easy to apply this 4(2): 129–146.
definition to a wide range of data sources. Botham R and Graves A (2011) Regional vari-
One disadvantage, however, is since it is ations in new firm job creation: The contribu-
Unitary Authority based, some urban areas, tion of high growth start-ups. Local Economy
including Stirling and Inverness, are classified 26(2): 95–107.
as rural. Using the Randall definition of rur- Braunerhjelm P (2007) Academic entrepreneur-
ality 89% of Scotland’s landmass and 29% of ship: Social norms, university culture and
its population are classified as rural. policies. Science and Public Policy 34(9):
619–631.
Carree MA and Thurik AR (2008) The lag struc-
References ture of the impact of business ownership on
Aoyama Y (2009) Entrepreneurship and regional economic performance in OECD countries.
culture: Hamamatsu and Kyoto, Japan. Small Business Economics 30(1): 101–110.
Regional Studies 43(3): 495–512. Cheng S and Li H (2011) Spatially varying rela-
Armington C and Acs Z (2002) The determin- tionships of new firm formation in the United
ants of regional variation in new firm forma- States. Regional Studies 45(6): 773–789.
tion. Regional Studies 36(1): 33–45. Cooper AC, Gimeno-Gascon FJ and Woo CY
Ashcroft B, Love JH and Malloy E (1991) New (1994) Initial human and financial capital as
firm formation in the British counties with predictors of new venture performance.
special reference to Scotland. Regional Journal of Business Venturing 9(5): 371–395.
Studies 25(5): 395–409. Equinox Management Consultants Ltd (2000)
Audretsch D (2002) Entrepreneurship: A Survey Informal Equity Capital for SMEs: A Review
of the Literature. Prepared for the European of Literature. Industry Canada: Ottawa.
Commission, Enterprise Directorate General. Etzioni A (1987) Entrepreneurship, adaptation
Institute for Development Strategies. and legitimation: A macro behavioural per-
London: Indiana University and Centre for spective. Journal of Economic Behaviour and
Economic Policy Research (CEPR). Organisation 8(2): 175–189.
Audretsch D, Dohse D and Niebuhr A (2010) Evans D and Leighton L (1990) Small business
Cultural diversity and entrepreneurship: A formation by unemployed and employed
regional analysis for Germany. Annals of workers. Small Business Economics 2(4):
Regional Science 45(1): 55–85. 319–330.
Audretsch D and Fritsch M (1994) The geog- Gleave W and Mitra J (2010) The spatial deter-
raphy of firm births in Germany. Regional minants of entrepreneurship in the regions of
Studies 28(4): 359–365. England, 2000–2004. International Journal of
Beaudry C and Schiffauerova A (2009) Who’s Entrepreneurship and Small Business 9(2):
right, Marshall or Jacobs? The localisation 143–161.

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


Ross et al. 687

Kangasharju A (2000) Regional variations in Reynolds PD, Miller B and Maki WR (1995)
firm formation: Panel and cross section data Explaining regional variation in business
evidence from Finland. Papers in Regional births and deaths. Small Business Economics
Science 79(4): 355–373. 7(5): 389–407.
Keeble D and Walker S (1994) New firms, small Reynolds P, Storey D and Westhead P (1994)
firms and dead firms: Spatial patterns and Cross-national comparisons of the variation
determinants in the United Kingdom. in new firm formation rates: An editorial
Regional Studies 28(4): 411–427. overview. Regional Studies 28(4): 343–346.
Krugman P (1991) Geography and Trade. Ross AG, Crossan K and Juleff L (2012) How
Cambridge, MA: MIT Press. accurate are VAT registrations as a measure
Lee S, Florida R and Acs Z (2004) Creativity and of entrepreneurship? A spatial analysis of
entrepreneurship: A regional analysis of new Scotland and its regions. Local Economy
firm formation. Regional Studies 38(8): 27(3): 279–296.
879–891. Scottish Corporate Sector Statistics (2012)
Lucas R (1988) On the mechanics of economic Scottish corporate sector statistics for the
development. Journal of Monetary Economics Scottish Government. Available at: www.sco-
22: 3–42. tland.gov.uk/Resource/0041/00411283.pdf.
Malecki E (2009) Geographical environments for Scottish Government Economic Strategy (2011)
entrepreneurship. International Journal of Scottish government economic strategy.
Entrepreneurship and Small Business 7(2): Available at: www.scotland.gov.uk/
175–190. Resource/Doc/357756/0120893.pdf.
Marshall A (1920) Principles of Economics, 8th Sternberg R (2009) Regional dimensions of
ed. London: Macmillan. entrepreneurship. Foundations and Trends in
Mason C (1991) Spatial variations in enterprise: Entrepreneurship 5(4): 211–340.
The geography of new firm formation. Storey D (1991) The birth of new firms-does
In: Burrows R (ed.) Deciphering the unemployment matter? A review of the evi-
Enterprise Culture: Entrepreneurship, Petty dence. Small Business Economics 3(3):
Capitalism and the Restructuring of Britain. 167–178.
New York, NY: Routledge. Sutaria V and Hicks DA (2004) New firm for-
Morky BW (1988) Entrepreneurship and Public mation: Dynamics and determinants. Annals
Policy. New York, NY: Quorum Books. of Regional Science 38(2): 241–262.
North DC (1991) Institutions. Journal of Tamasy C and Le Heron R (2008) The geog-
Economic Perspectives 5(1): 97–112. raphy of firm formation in New Zealand.
Nystrom K (2008) Regional Institutional Tijdschrift Voor Economische En Sociale
Environment and Swedish Regional New Firm Geografie 99(1): 37–52.
Formation. CESIS Electronic Working Paper US Department of Commerce (2007) Economic
Series No 142. development America. Available at: www.
OECD (2008) Fostering Entrepreneurship. Paris: eda.gov/ImageCache/EDAPublic/documents/
OECD. pdfdocs2007/edamericasummer2007_2epdf/
Okamuro H and Kobayashi N (2006) The v1/edamericasummer2007.pdf.
impact of regional factors on the start-up Van Praag CM and Versloot PH (2007) What is
ratio in Japan. Journal of Small Business the value of entrepreneurship? A review of
Management 44(2): 310–313. recent research. Small Business Economics
Penrose ET (1959) The Theory of the Growth of 29(4): 351–382.
the Firm. New York, NY: John Wiley.

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015


688 Local Economy 30(6)

Appendix 1. Spatial variation in continuing entrepreneurial activity


rates per 1000 of working population 1998–2012

Downloaded from lec.sagepub.com at Flinders University on October 12, 2015

You might also like