THE EFFECTS OF NEW FIRM FORMATION ON REGIONAL
DEVELOPMENT OVER TIME: THE CASE OF GREAT BRITAIN
Max Planck Institute of Economics, Germany.
André van Stel,
Erasmus University Rotterdam, EIM Business & Policy Research, Netherlands, and Max Planck Institute of Economics, Germany.
David J. Storey,
CSME, Warwick Business School, University of Warwick, UK.
Paper submitted to Small Business Economics
1. INTRODUCTION This paper examines the link between new and small firms on the one hand, and
economic development on the other. In principle it is clear that if a new firm is created
and it employs, either as an owner or an employee, individuals who formerly were not
employed, then this adds to aggregate employment in the economy. This arithmetic often
provides a superficial attraction, particularly to politicians facing high levels of un-
employment or under-employment. It suggests that raising new firm formation will lead
subsequently to additional employment.
This paper, however, makes it clear that this outcome is achieved only when certain key
assumptions are made. The first is that it is possible for policy makers to take actions
which do raise new firm formation. A second assumption is that macro economic
circumstances remain unchanged. If this assumption is violated – such that macro
economic circumstances become either more or less benign – then this will influence
rates of new firm formation. On balance the evidence is that less benign macro-economic
conditions lead to a rise in new firm formation. The third assumption is that it is new
firm formation which causes increased employment and not vice versa. Often, increased
employment is not the consequence but rather the cause of higher start-up rates. This is
because local demand is likely to be higher in regions with employment increases, and
therefore such regions are attractive locations to found new firms. As these two
mechanisms differ greatly in their policy implications it is of vital importance to
disentangle the correct direction of causality. Fourthly, as Fritsch & Mueller (2004) show,
the creation of a new firm has both immediate and longer term consequences. As noted
earlier, new firms have an easily identifiable short term direct effect in creating
employment, but they also have two longer term consequences. The first is that new
firms displace inefficient incumbents, leading to job losses. However, a second medium
term consequence is that new firms enhance the competitiveness of firms that remain in
1business, by exerting a powerful threat upon incumbent firms inducing improved
1 Disney, Haskel and Hedon (2003) find that in the UK between 1980 and 1992 about half of productivity
gain was because of internal factors – such as introducing new technology and organisational changes. The
remaining half was because of external factors most notably that the entrants were more productive than
performance from them. As Fritsch and Mueller (2004) show, the effect of new entrants is therefore threefold: the first effect is to increase employment, the second is to lower employment and the third is to increase employment. The total effect upon employment can therefore be either positive or negative and depends upon the magnitude of the three elements.
This paper develops the Fritsch and Mueller (2004) findings for Germany by applying a broadly similar estimation procedure to data for Great Britain. It recognises that the net effect of new firm formation on subsequent employment can be either positive or negative. Furthermore, this study investigates whether there are circumstances in which the effects of new firm formation on subsequent employment change are more likely to be positive or more likely to be negative. In particular it assesses whether policies to raise new firm formation are more effective in geographical areas that already have high rates of new firm formation or whether they are more effective in low enterprise areas. To test this, the paper examines the effects of new firm formation on subsequent employment change between 1981 and 2003 for Great Britain, making a distinction between England on the one hand and Scotland and Wales on the other.
Our broad key finding is that, for Great Britain as a whole and England in particular, the effect of new firm formation on subsequent employment growth is positive. Hence raising new firm formation, given that this is possible, will lead to subsequent increased employment in this country. However, we do not get the same results for Scotland and
Wales. Here the impact of increased new firm formation is negative – causing lower
Unfortunately both Scotland and Wales have, historically, had low rates of new firm formation. Believing this to be undesirable, both countries implemented policies, independent of those of England, which have sought to raise new firm formation. The evidence from this paper is that increased rates of new firm formation in Scotland and Wales lead to less, rather than more, employment. The paper therefore raises the question of whether it is possible to have „the wrong type of entrepreneurship‟ and if
those exiting. However amongst single plant independent firms almost all the gains were attributable to external factors.
public policy should focus on quantity or quality of new firms. In Scotland and Wales
new firm formation appears to lead to fewer jobs whereas in England – or at least parts of
England – it seems to lead to more jobs.
2. POLICY CONTEXT
This section discusses the changing enterprise policy context in Great Britain over the last
three decades, which is vital for understanding the context of the paper. We use the
categorisation of policy provided by Greene et al. (2004). They identify four separate
policy periods. The first is the 1970s when enterprise policy – defined as providing assistance and support for new and small firms – hardly existed. The 1970s is therefore viewed as a “policy off” period.
The arrival of Margaret Thatcher as Prime Minister in 1979 changed matters radically. A
key tenet of the philosophy of her government was to move Britain away from what she
saw as a dependency culture – defined as one where individuals relied on the state. Her
objective was for Britain to have an enterprise culture – in which individuals relied on themselves rather than the state. Thatcher therefore saw enterprise creation as highly
desirable in itself, but also as a direct mechanism for lowering unemployment levels. The
1980s can therefore be considered as a decade characterised by support for new start-up
The 1990s saw a shift in this policy, away from new firms and towards providing support
2. For GB as a whole therefore, the mid to for existing small firms with growth potential
late 1990s sees considerably less emphasis placed by policy makers upon new firms. The
1990s can therefore be viewed as a decade in which the focus was on smaller firms with
growth potential. Finally, following the election of a Labour government in 1997, the
decade of 2000 sees a further change with policies being focussed more specifically on
seeking to enhance productivity and increase social inclusion. In this period, both some
3types of new firms and some types of growing firms are supported.
2 For the reasons underlying this shift see Greene et al (2004). 3 These policies were formalised in Small Business Service (2004). This identified the so-called seven
pillars: building an enterprise culture, encouraging a dynamic start up market, building the capability for
Whilst these policies are being implemented in Great Britain – and specifically in
England – different policies were being adopted in both Scotland and Wales, particularly
in the 1990s. At a time when GB policies focussed upon “growth businesses”, Scottish
policy makers became concerned that their new firm formation was much lower than that
of England, as shown in Figure 1 below. In 1993 Scotland launched its Business Birth
Rate Policy towards developing business start-ups and the development of
entrepreneurship seeking to eliminate the difference between the two countries. This
policy was recently reviewed and an overhaul took place. The new approach was
launched in 2002 focusing on three priorities. Firstly, innovative high-growth start-ups
shall be encouraged to increase their number and value. Secondly, quality volume 20022002
20012001business support services are offered to encourage more people to start businesses. 20002000
Finally, the third focus is on entrepreneurship education. Other targets are to generate at 19991999
19981998least 30 new starts that achieve the highest growth within three years, to achieve a 70 19971997
19961996percent survival rate for start-ups over a three-year period, and to increase the number of 19951995start-ups by women and young people. 19941994
4519911991Great BritainGreat BritainGreat BritainGreat Britain4519901990EnglandEnglandEnglandEngland1989198940ScotlandScotlandScotlandScotland1988198840WalesWalesWalesWales1987198735198619863519851985
20VAT registration per 10,000 inhabitantsVAT registration per 10,000 inhabitants
Figure 1: VAT Registration per 10,000 inhabitants, 1980-2003
small business growth, improving access to finance for small businesses, encouraging more enterprise in
disadvantaged communities and under-represented groups, improving small businesses‟ experience of
government services and developing better regulation and policy
Over the period 1980 to 2003, new firm formation rates – measured as VAT Registrations
per 10,000 inhabitants – are consistently lower in Wales and Scotland than in England (Figure 1). The overall rate of Great Britain is very close to the England rate because
England is the numerically dominant country. On average about 90 percent of the VAT
registration in Great Britain take place in England. Wales counts for only about four
percent of the registrations and Scotland consequently for six percent. Figure 1 also
shows that there is a similar pattern over time in England, Scotland and Wales, with new
firm formation rates rising in all three countries at least until 1989, then falling until
1994-95 and subsequently being broadly stable. Hence, although the new firm formation
rates greatly changed over time the distribution between the three regions is very constant
over the period 1980 to 2003. The reason for the steep decrease at the beginning of the
1990s and in 1994 is an increase of the VAT registration threshold of more than
30 percent and 20 percent, respectively.
Although the introduction of the Scottish Business Birth Rate Policy in 1993 seems to
have resulted in realising higher new firm formation rates than Wales (Figure 1) it had
little obvious impact upon new firm formation and it certainly did not appear to close the
gap with England. In spite of the Scottish experience, Wales also sought to raise new
firm formation. It introduced its Entrepreneurial Action Plan (EAP) in 2001 but any
effect is difficult to identify given its short duration. Wales‟ Entrepreneurial Action Plan
is built around three key objectives. The first target is to encourage more people to start
or grow a firm by creating a greater awareness of the opportunities and benefits of
entrepreneurship in order to develop an entrepreneurial culture. Secondly, the focus is on
sustainable start-ups and to increase their number. Thirdly, the program is ought to
increase the number of start-ups with growth potential and to create employment.
3. LITERATURE REVIEW The lagged effects of new firm formation on employment growth already became
apparent in the study of Audretsch and Fritsch (2002). They found that employment
growth in the 1990s could not be explained by start-up rates in the 1990s but rather by
start-up rates of the 1980s in West Germany. Fritsch and Mueller‟s work on Germany,
referred to above, formally showed the importance of time lags in examining the
relationship between new firm formation and employment change. They theorised that
new firm formation had three effects reflected in Figure 2 below.
The first effect was the short term direct effect of employment created by the new firm.
The second was the displacement effect of the entrant which causes some existing firms
to go out of business, hence incurring job losses. Finally, there is a third effect of the
entrant which is to stimulate better performance from surviving firms. This is called the
Applying this theory to German data Fritsch and Mueller (2004) find support for this
three phase model. They concluded that the peak of positive impact of new firms on
regional economic development was reached about eight years after entry, and
disappeared after about a decade. Using a broadly similar approach, van Stel and Storey
(2004) examined Great Britain for the period 1981 to 1998. Their conclusions were
similar to those of Fritsch and Mueller. They found the employment impact of new firm
4 formation was maximised in year five but had declined to zero by year nine.
4 A disadvantage of their exercises was that they restricted the Almon lag polynomial to be of second
degree while Fritsch and Mueller showed that a third degree polynomial may actually be more realistic.
direct effectdirect effect
induced effectinduced effectII00
displacement effectdisplacement effect
IIIIon employment changeon employment changeImpact of new business formationImpact of new business formation
Lag (year)Lag (year)
Figure 2: Schematic effects of new firm formation on employment change (according to
Fritsch and Mueller, 2004)
However, it is Van Stel and Storey‟s findings for „un-entrepreneurial areas‟ such as North-East England and Scotland, which is of greatest interest in the context of the
current paper. Here they found that although there was, in the 1990s evidence that for
England as a whole, new firm formation led to additional employment, this was not the
case for all geographical areas. In areas with low rates of new firm formation such as
Scotland and North East England, the overall impact of new firm formation on
employment change was negative. This implied that policies to raise new firm formation
in low enterprise areas could be counter-productive in terms of employment change.
Rather than creating new employment opportunities, the new firms displace existing
firms leading to employment losses or those new firms exit the market shortly after entry
thereby degrading the employment just created.
4. DATA SOURCES
The Van Stel and Storey (2004) analysis ended in 1998, and so there was insufficient time to examine the impact of the Scottish Business Birth Rate Policy. The Wales Entrepreneurial Action Plan (EAP) was not implemented until 2001 after the data set had ended. The current paper extends the time period examined up until 2003, which does enable the Scottish policy to be more effectively reviewed. The EAP is however too recent for any of its effects to be included.
This paper formally merges the methodology of Van Stel and Storey, with that of Fritsch and Mueller. It has the specific objective of examining the impact of policy in the two low enterprise countries of Scotland and Wales, in comparison with high enterprise, England. As in the Van Stel and Storey (2004) paper, data on start-ups is taken from the UK Small Business Service. Start-ups have to exceed a turnover threshold in order to be registered for VAT purposes, for instance, a threshold of ?56,000.00 had to be exceeded in 2003. Businesses below this threshold may also choose to register, but this is optional and unlikely. While about 174,000 start-ups (including agriculture) registered in England, about 7,000 new firms registered in Wales and 12,000 in Scotland during the year 2003. Data on employment is taken from the Census of Employment and, from 1995 onwards, the Annual Employment Survey. Data are supplied by Nomis with the self-employed and unpaid family workers being excluded from the data. Furthermore, following Van Stel and Storey, hourly wages and population density were included as control variables. Wage data are derived from the New Earnings Survey Panel (NESP) Dataset and were changed into constant prices 2003. Data on population density were taken from the Office for National Statistics.
The unit of analysis is 60 British regions – 46 English Counties, 4 Welsh Regions and 10
Scottish Local Authority Regions over the period 1981-2003. Because of missing (employment) data, the region Orkney/Shetland/Western Isles had to be excluded, therefore generating 59 observations. Different regional and sectoral classifications in the original data files meant some linking operations were performed to ensure uniformity for 1980–2003 (see Van Stel and Storey, 2004 for details). The agricultural sector is
excluded as it is fundamentally different from the rest of the economy having, during this
period, exceptionally low start-up and death rates.
5. DATA ANALYSIS
The regions in England and the regions in Scotland and Wales clearly show the same
pattern regarding the correlation of new firm formation over time (see figure 3). Firstly,
there is a strong correlation in the short term between start-up rates, but this correlation
weakens over time. Secondly, there is a strong variation over space, while some regions
count for only four start-ups per 100 employees, other regions had even more than 15
registrations per 100 employees. Especially the English regions experience higher start-
up rates. Most Scottish or Welsh regions cluster around lower values. Figure 3 plots the
correlation of new firm formation for two time periods. The first is the relationship
between start-ups in t and in t-1and the second is between t and t-5. A distinction is made New firm formation rate in t-5
between the England, and Scotland and Wales data.
New firm formation rate in t-166
New firm formation rate in tNew firm formation rate in t
Figure 3a: Correlation of new firm formation rates over time in England (per 1,000