entrepreneurs versus incumbents: who creates the better jobs? · 2016-04-14 · the words of...
TRANSCRIPT
The Rockwool Foundation Research Unit
Study Paper No. 101
Entrepreneurs versus Incumbents:
Who Creates the Better Jobs?
Johan M. Kuhn, Nikolaj Malchow-Møller and
Anders Sørensen
University Press of Southern Denmark
Odense 2015
Entrepreneurs versus Incumbents: Who Creates the Better Jobs?
Study Paper No. 101
Published by:
The Rockwool Foundation research Unit and
University press of Southern Denmark
Address:
The Rockwool Foundation Research Unit
Soelvgade 10, 2.tv.
DK-1307 Copenhagen K
Telephone +45 33 34 48 00
E-mail [email protected]
web site: www.en.rff.dk
ISBN 978-87-93119-28-4
ISSN 0908-3979
November 2015
1
1
EntrepreneursversusIncumbents:WhoCreatetheBetterJobs?
Johan M. Kuhn, Department of International Economics and Management, Copenhagen
Business School, Porcelænshaven 16A, DK-2000 Frederiksberg, Denmark (tel: +45 3815 3467,
email: [email protected])
Nikolaj Malchow-Møller, Department of Business and Economics, University of Southern
Denmark, Campusvej 55, 5230 Odense M, Danmark (tel: +45 6550 2109, email:
Anders Sørensen (corresponding author), Department of Economics, Copenhagen Business School, Porcelænshaven 16 Porcelænshaven 16A, DK-2000 Frederiksberg, Denmark (tel: +45 3815 3493, email: [email protected]).
Abstract: What are the characteristics of jobs in entrepreneurial firms as compared to jobs in incumbent firms?
Even though this question has been addressed by many researchers before us, we provide new evidence
to the field since we measure the entrepreneur as the organic new firm. In the literature, the majority of
studies have focused on entrepreneurs as measured by small or new firms. By organic new firm, we
mean new firms that are not the result of restructurings or organising existing or additional activities in a
formally new firm. Moreover, we distinguish entrepreneurial firms by different types and distinguish between growing and declining industry‐region clusters. Our results differ from the findings in the
existing literature. Specifically, we find that compared to incumbents, entrepreneurial firms have higher
total factor productivity, are more skill intensive, and pay higher wages. The differences are more
pronounced in growing clusters. Moreover, the results show important differences between different
types of entrepreneurial firms. Specifically, spin‐offs are found to enjoy the largest productivity
advantage. The wage and skill premiums at the firm level disappear at the job level, as larger
incumbents are both more skill intensive and pay higher wages than smaller incumbents.
Keywords: Entrepreneurship, job quality, productivity
JEL: L1
Acknowledgements: We are greatly indebted to the Rockwool Foundation for funding of this project, and Statistics for providing the data. Thanks to Jan Rose Skaksen, Søren Leth‐Petersen, Mette Ejrnæs, Pernille Bang and Zuzanna Tilewska for helpful comments.
3
2
1. IntroductionEntrepreneurs are often considered to play an important role as an engine for growth and prosperity. In
the words of Schumpeter (1934, 1943), entrepreneurs create combinations of inputs and outputs to
pioneer new activities, exploit new market opportunities and allocate labour to its most productive use.
However, entrepreneurship is not always found to be good business for those involved in
entrepreneurial projects. The majority of studies in the economic literature have found that
entrepreneurs pay lower wages than other firms and hire employees with lower levels of human capital.
Moreover, the productivity levels of entrepreneurs are found to be of similar magnitude or lower than in
established firms. See Van Praag and Versloot (2007) for an overview of the literature.
In this paper, we reconsider the characteristics of jobs in entrepreneurial firms as compared to jobs in
established firms. Even though this question has been addressed by many researchers before us, we
provide new evidence to the field since we measure the entrepreneur as the organic new firm. By
organic new firm, we mean that the new firm is not a result of restructurings or a result of organising
existing or additional activities in a formally new firm. Thereby, the organic new firm must not have
existed previously under a different name, with a different owner, or in another legal form (personally
owned, incorporated, etc.). Furthermore, the new firm must not have been started by persons who are
already registered as business owners. Nor may the new firm be a re‐start of a business after closure, or
a result of changes in the firm‐registration information. In sum, the set of new firms used in this paper is
much more likely to reflect organic start‐ups than if we had simply used the set of all new or small firms,
which have been common practice in the literature.
In the literature, the majority of studies have focused on entrepreneurs as measured by small or new
firms. Small firms are not necessarily start‐ups; it can also be old firms that have not grown large. In this
way, old firms without employment growth are mixed up with entrepreneurial firms. This is
unfortunately and may influence the established evidence for jobs created by entrepreneurs. Using new
firms is in principle better, but still problematic because many of the formally new firms may be the
result of restructurings or the result of organising existing or additional activities in formally new firms.
New firms that are not organic new firms should not be included in empirical studies of the
characteristics of jobs generated by entrepreneurs as they may influence the empirical findings.
As our main purpose is to characterize jobs and to compare them across firm types, we compare the
following variables across entrepreneurs and established firms: wages, skills, labour productivity as well
as total factor productivity (TFP). The latter measure is not a job specific measure, however, it is partly
4
3
based on firms’ labour inputs and therefore indirectly includes important information on job
characteristics.
The following results are established in the literature that is based on the entrepreneur as measured by
establishment/firm size or age: Brixy et al. (2007) and Koch and Späth (2009) both find that start‐ups or
younger firms pay lower wages, whereas Nyström and Elvung (2014) find that there is a wage penalty
associated with employment in a new firm for labor market entrants. In an earlier study, Brown and
Medoff (2003) find that firms that have been in business for more years pay higher wages, but once they
control for worker characteristics, the relationship becomes insignificant or negative. Kölling et al.
(2002) find a similar result. A number of studies have also examined the relationship between firm size
and wages, and here it is well established that smaller firms pay lower wages, see, e.g., Troske (1999).
Just as small firms are found to hire less skilled workers, see Troske (1999), several studies also find that
new firms hire less skilled workers. Koch and Späth (2009) thus find that the use of high‐qualified labor is
lower for young firms than for incumbents, and Brown and Medoff (2003) also find that young firms
employ workers with lower levels of education.
Most studies show that new establishments or firms have lower levels of labor productivity, see, e.g.,
Brouwer et al. (2005), Jensen et al. (2001) and Foster et al. (2006), and also lower levels of TFP, see, e.g.,
Castany et al. (2005) and Brouwer et al. (2005). An exception is Disney et al. (2003) who find that labor
productivity of new plants is relatively high, and also that TFP levels are higher for new plants than for
old plants. Moreover, Huergo and Jaumandreu (2004) find that young firms have higher growth of
productivity than old firms and that this difference can last for many years but also that productivity
growth rates of surviving firms converge.
As mentioned, we shed new light on the characteristics of jobs created by entrepreneurs as a
consequence of the applied measure of entrepreneur; the organic new firm. In doing this, we further
extend the existing literature in two important ways. First, we distinguish between firms located in
growing industry‐region clusters and firms located in declining industry‐region clusters. The motivation
behind distinguishing between growing and declining clusters is to take the business environment of the
firms into account because a large share of entrepreneurial activity is not pulled by new market
opportunities. Push factors like dissatisfaction with or separation from an earlier employment
relationship or lack of alternative employment opportunities also play an important role, see, e.g., Amit
and Muller (1995). As a consequence, a large share of the jobs created by entrepreneurs are not created
5
4
in business environments reflecting new market opportunities, but rather in stable or contracting
markets, and this is likely to affect both the performance of the entrepreneurial venture and the quality
and characteristics of the jobs created.
Second, we distinguish between three types of entrepreneurs: spin‐offs, entrepreneurial start‐ups and
other new firms. This is motivated by previous findings in the literature where spin‐offs have been found
to be relatively successful in terms of survival and growth; see, e.g., Dahl and Reichstein (2007) and
Eriksson and Kuhn (2006). Spinoffs are throughout defined as organic new firms where the founders
have recent industry experience from the same industry as that of the entrepreneurial firm. From a
policy perspective, the distinction between spin‐offs and other types of entrepreneurial start‐ups is also
highly relevant given that most (OECD) countries are trying to figure out how to support
entrepreneurship most effectively. Finally, other new firms are a residual category where it has not been
impossible to identify the entrepreneur(s).
We base our analysis on Danish worker‐firm register data for the time period 2001‐2010. The data are
drawn from different Danish registers administered by Statistics Denmark and cover almost the entire
private sector of the Danish economy for the period 2001‐10. It contains detailed data on revenue,
inputs, worker information, industry and regional location for all firms, including whether firms are
organic new firms or not.
The empirical methodology is quite simple. We regress the characteristics of the jobs and the
performance measures that we wish to compare on a set of dummies for firm and cluster types. We also
include a number of additional controls for other firm or worker characteristics, and we look at within‐
cluster differences between firm types.
Finally, we extend the definition of clusters to contain a third dimension – the education of the worker.
That is, a cluster is defined as jobs with certain skill contents in a given industry and region. We then
compare jobs in entrepreneurial and incumbent firms within these education‐specific clusters.
Throughout the paper, we consider job characteristics in the final year of our sample, i.e., 2010, and
entrepreneurial firms are in most of the analysis defined as organic new firms established after 2000.
That is, they can be up to nine years old. This choice involves a selection on relatively successful
entrepreneurial firms as it includes surviving firms and excludes firms that have not survived up until
2010. E.g., entrepreneurial firms that are established in 2002 and have survived until 2010 are 8 years of
age. These firms constitute a relatively selected group as compared to firms that are established in 2002
6
5
and have closed down again before 2010. To check the robustness of our results to this assumption, we
also look at alternative scenarios where organic new firms are firms established after 2005, i.e., they can
be up to five years old.
We find several interesting results. First, when comparing entrepreneurs and incumbents, without
distinguishing between different types of entrepreneurs, we find that entrepreneurs have higher TFP
(but lower labour productivity), are more skill intensive and tend to pay higher wages than incumbents.
Furthermore, we find that these differences are more pronounced in growing industry‐region clusters.
Second, distinguishing between different types of entrepreneurial firms, we find that spin‐offs enjoy the
largest productivity advantage compared to old firms, whereas higher wages are mainly found within
other new firms, although spin‐offs fare better than old firms when located in growing clusters.
Third, the wage and skill premiums at the firm level disappear in the worker‐level regressions, which
generally show lower wages in spin‐offs and entrepreneurial start‐ups than in old firms. The difference
in results between the two approaches reflects that the larger old firms are given more weight in the
worker‐level regressions, and these old firms are different from the smaller old firms.
Fourth, when we look separately at manufacturing firms, we find more robust TFP advantages and
higher average wages of entrepreneurial firms compared to old firms.
Fifth, introducing the third dimension in our cluster does not change results qualitatively, and neither
does defining new firms as those established after 2005 (instead of 2000). In the latter case, the only
difference is that we now find that entrepreneurial firms are less productive than old firms. This lower
productivity is driven by entrepreneurial firms in declining clusters, whereas entrepreneurial firms have
productivity levels of similar magnitude to old firms in growing clusters. Moreover, spin‐offs still enjoy
productivity advantages compared to old firms, especially in growing clusters, whereas entrepreneurial
start‐ups and other new firms have similar or lower productivity levels. These results suggest that it
takes time for organic new firms to catch‐up and exceed incumbents in terms of productivity as also
found by Huergo and Jaumandreu (2004).
The main contribution of this paper is to investigate the characteristics of jobs in entrepreneurial firms
compared to established firms using a measure of the entrepreneur as the organic new firm. Our results
differ from the findings in the existing literature. Specifically, we find that compared to incumbents,
entrepreneurial firms have higher TFP (and lower labour productivity), are more skill intensive, and pay
7
6
higher wages. However, the wage and skill premiums at the firm level disappear at the worker level, as
the larger old firms are both more skill intensive and pay higher wages than the smaller old firms. This
latter result is more consistent with the existing literature. It should be emphasised that the purpose of
this paper is not to present a causal relationship for job quality and firm types. The study is rather a
study of facts for the raw and conditional differences between jobs in entrepreneurial and incumbent
firms. It is therefore not clear whether the results are biased by for example selection of different types
of individuals into different types of firms. We leave this important issue for future research.
The rest of the paper is organised as follows. In Section 2, we describe the data and the definitions
applied. Section 3 presents the empirical approach. In Section 4, we present our baseline results, where
we compare new and old firms. In Section 5, we introduce our distinction between different types of
new firms, and in Section 6, we consider the manufacturing sector separately. Sections 7 and 8 contain
our robustness analyses. Section 9 concludes.
2. DataIn the analyses below, we rely on matched worker‐firm data covering almost the entire private sector of
the Danish economy for the period 2001‐10. The data are drawn from different Danish registers
administered by Statistics Denmark.
First, we use the General Enterprise Statistics, which builds on the Central Business Register and contains
annual information about all active firms in the Danish economy. From this database, we get
information about, e.g., industry, region, sales, capital input, intermediate inputs and number of
workers of all firms in the private sector.
Second, we use the Statistics on New Enterprises, which identifies, for each year between 2001 and
2010, all the new start‐ups. This includes both personally‐owned and incorporated firms that fulfil a
number of conditions that allow us to consider them as being organic new firms in a given year. We use
this database to identify all the start‐ups among all the active firms in the General Enterprise Statistics.
Note that Statistics Denmark has undertaken extensive efforts to identify these organic new firms. Many
of the formally new firms may thus be the results of restructurings or the results of organising existing or
additional activities in formally new enterprises. As a consequence, for firms to appear in the Statistics
on New Enterprises, they must not only be newly registered at the business authorities for VAT‐taxation,
it is also required that the firms must not have existed previously under a different name, with a
8
7
different owner, or in another legal form (personally owned, incorporated, etc.). Furthermore, they
must not have been started by persons who are already registered as business owners at the VAT
authorities. The data are also cleaned for re‐starts of businesses after closure, and changes in the firm‐
registration information; see Statistics Denmark (2002) for more details. Finally, we have removed an
additional 0.2 per cent of observations in the Statistics on New Enterprises, where supplementary
information in the General Enterprise Statistics suggests that the firm was established before 2001. In
sum, the set of new firms used in this paper is much more likely to reflect organic start‐ups than if we
had simply used the set of all new establishments or all new firms, which has been common practice in
the literature.
The Statistics on New Enterprises is restricted to industries that Statistics Denmark categorises as
“business‐related industries”. This excludes the public sector and (most of) the primary sector, as well as
industries with activities that are not liable to VAT, such as dentists, transportation of persons, banking
etc. Furthermore, the Statistics on New Enterprises is restricted to firms with standard ownership types.
To ensure valid comparisons, we impose the same sampling conditions on the General Enterprise
Statistics, i.e., we exclude firms in non‐business‐related industries and firms with non‐standard
ownership types.
For the majority of the new firms in the Statistics on New Enterprises, Statistics Denmark has been able
to identify the entrepreneurs behind the firms. For the personally‐owned firms identification is
straightforward: The entrepreneur is simply identified as the owner of the firm. For incorporated firms,
Statistics Denmark uses a prioritized list of criteria to identify the principal entrepreneur(s).1 For 17 per
cent of the cases, Statistics Denmark has not been able to identify the entrepreneur behind the firm.
Third, we use the Firm Integrated Database (FIDA), which identifies all the individuals working in a given
firm in the last week of November each year. From this database, we can get information about wages,
education, age, and gender of all the workers in a given firm in a given year. Furthermore, as workers
1 If information about a founder is available from the registration information, this person is identified as the principal entrepreneur. In case of more than one founder, Statistics Denmark selects the one who has the highest salary – or if none of the founders are employed in the firm, they pick the founder who appears first in the registration database. If information about founders is not available, they look for a member of the board (or the executive board) who is also employed in the firm. Again, they pick the one with the highest salary in the case where more than one board (or executive board) member is employed in the firm. If no board (or executive board) member is employed in the firm, they pick the board (or executive board) member who appears first in the registration database.
9
8
are linked to firms each year, we can identify all the individuals that leave (or join) a firm between two
years.
From the FIDA we construct firm‐level measures of the average age and wage of the workers, the
gender composition and the share of workers who have completed tertiary education, i.e., individuals
who have obtained at least a bachelor degree (we refer to these as “highly educated”). We also use the
FIDA to construct a worker‐level dataset where the information about the firm is linked to each worker.2
Note that individuals can be associated with more than one firm in a given year as Statistics Denmark
records both a primary and (potentially) a secondary job of each individual, where the distinction
between the two is based on the wage income generated by the jobs. In the present paper, we discard
information about secondary jobs. We deviate from this “one‐job‐only”‐rule in one case: for owners of
personally‐owned firms with employees. These individuals can have two jobs: As an employer in their
own firm and as a wage worker in another firm. However, any given individual can only have one job per
firm.
In the analyses of this paper, we use cross sections for 2010 at either the firm or the job level, but we
use the historical information in the databases to identify the new firms and to construct our industry‐
region clusters. More about this below.
In the main part of the analysis, we define the new firms as those established after 2000, i.e., new firms
can be up to nine years old in 2010, whereas established/old firms are those already present in 2000.
This leaves us with 108,322 old firms and 79,712 new firms. To check the robustness of our results to
these assumptions, we also look at an alternative scenario where new firms are defined as firms
established after 2005, which leaves us with 130,056 old firms and 57,988 new firms.
We further divide the new firms into three types of new firms that are mutually exclusive: spin‐offs,
entrepreneurial start‐ups, and other new firms. Spin‐offs are defined as new firms where at least one of
the entrepreneurs has been employed in the same industry prior to establishing the firm. More
precisely, we identify all (primary and secondary) jobs of the founders in the year prior to start‐up of the
2 Note that when we talk about “workers”, this may include both the employees and the owners of personally‐owned firms. Technically speaking, the latter are not employed in their own firms. While wage information only exists for employees (and is only reliable for approximately 1.2 million employees – corresponding to 75 per cent of the observations), information about age, gender and education is available for both types of workers. It should be mentioned, however, that the educational level is unknown for approximately 4 per cent of the observations.
10
9
new firm, and if any of these jobs were in the same 2‐digit‐NACE industry as the new firm, the start‐up is
categorised as a spin‐off. This leaves us with 24,017 spin‐offs among the 79,712 new firms.
Entrepreneurial start‐ups, on the other hand, are new firms where none of the entrepreneurs have been
employed in the same 2‐digit‐NACE industry as the new firm in the year prior to starting up the new
venture, i.e., these are firms that cannot be categorized as spin‐offs as defined above. This results in
46,637 entrepreneurial start‐ups.
Finally, we are left with a residual category of 9,058 firms, where Statistics Denmark has been unable to
identify the entrepreneur(s) behind the firms. We label this residual group “other new firms”. These
new firms are all incorporated firms, and are thus likely to be larger than the average entrepreneurial
firm.
In the analysis, we divide the economy into a number of clusters to try to capture local market
conditions. Specifically, we are interested in determining whether a given firm or job is in a declining or
a growing cluster. We measure the local market conditions by the employment growth at the cluster
level between 2005 and 2010. The clusters allow comparison of job characteristics not just by whether
or not the given job is in a new or established firm, but also by whether or not the job is in a cluster with
growing or declining economic activity as measured by changes in employment. We believe that this
provides a more appropriate assessment of the role of new firms relative to old firms.
The most obvious way of defining a cluster is according to industries (Goos and Manning, 2007), but also
geography may be important for local producers and service industries. In the following, we therefore
define clusters by the industry and region of the firm. For our baseline cluster definition, we use
industries at the 3‐digit NACE level, which gives us 233 different industries. Together with five different
geographical regions, this results in 1,165 potential clusters of which 74 do not contain any firms in
2010. In Appendix D, we also consider the robustness of our results to changes in the number of clusters
by using 77 industries at the 2‐digit NACE level with the 5 regions.
In Section 8, we also introduce the education level of the employee as a third dimension in the cluster
definition. In this case, jobs are uniquely assigned to clusters, but a firm can then be “located” in more
than one cluster – with some of its jobs in one cluster and other jobs in another cluster.
Before turning to the details of the estimation approach and our empirical findings, we present some
descriptive statistics in Tables 2.1 to 2.4 below. In Table 2.1, we present the number of clusters, the
11
10
number of jobs and the number of firms applied in the analysis, broken down by the employment
growth of the clusters. In total, the 1,081 clusters contain approximately 1.56 million jobs and 188,000
firms in 2010. Of the 1,081 clusters, more than 50 per cent (including new clusters) had negative or
positive employment growth of more than 25 per cent during the period 2005‐2010. However, “only”
around 24 per cent of the firms and 28 per cent of the jobs were located in these clusters. Furthermore,
the large number of clusters with negative growth rates may in part reflect the outbreak of the financial
crisis during this period. Less than 2.5 per cent of all clusters in 2010 were new, i.e., without any activity
back in 2005.3 The shares of jobs and firms in these clusters in 2010 were only 0.05 and 0.02 per cent,
respectively.
[Table 2.1 around here]
Table 2.2 splits up the firms on the four categories of firms used in this paper: Established firms, spin‐
offs, entrepreneurial start‐ups and other new firms. Among the 188,000 firms, 80,000 (42 per cent) are
identified as new firms, i.e., established after 2000 and hence can be up to nine years old.
Approximately 24,000 (13 per cent) are categorised as spin‐offs according to our definition, whereas
47,000 (25 per cent) are categorised as entrepreneurial start‐ups, leaving approximately 9,000 (5 per
cent) in the residual category (other new firms).4
From the lower part of the Table, we can see that while new firms constitute 42 per cent of the firms.
They only represent around 16 per cent of the jobs. Most of the jobs (1.31 million) are found in
established firms, whereas 114,000 jobs (7 per cent) are found in entrepreneurial start‐ups followed by
spin‐offs with 82,000 jobs and other new firms with 56,000 jobs.
Another observation from Table 2.2 is that there are relatively many entrepreneurial start‐ups and other
new firms in clusters with more than 10 per cent employment growth. Somewhat surprisingly, spin‐off
firms are overrepresented in clusters with more than 10 per cent negative employment growth.
[Table 2.2 around here]
3 In the following tables new clusters are included in the 0.5+ interval as a consequence of confidentiality requirements when the number of observations in a cell is too low. 4 With the alternative definition of new firms where these are defined as firms established after 2005, 58,000 (31 per cent) firms are categorized as new firms. Approximately 17,000 (9 per cent) are categorised as spin‐offs according to our definition, whereas 35,000 (18 per cent) are categorised as entrepreneurial start‐ups, leaving approximately 6,000 (3 per cent) in the residual category (other new firms).
12
11
Table 2.3 presents the average firm size (as measured by jobs or workers per firm) for the four different
firm types and by cluster‐growth intervals. Established firms have on average 12.1 jobs per firm. Among
the new firms, the average “other new firm” is considerably larger than the two other types of new
firms, where spin‐offs tend to be slightly larger than entrepreneurial start‐ups. However, there is
considerable variation across clusters with different levels of employment growth, where there seems to
be a tendency for most firm types to be larger in the more extreme growth intervals.
[Table 2.3 around here]
Table 2.4 summarizes some of the basic characteristics of the workers in the different firm types and by
growing and declining clusters, respectively. It is seen that jobs in growing clusters have higher wages,
have higher shares of women, and have higher shares of highly‐educated workers. Established firms and
other new firms pay the highest salaries to their employees, while spinoffs in declining clusters pay the
lowest. Spin‐offs in declining clusters are also characterized by the lowest shares of highly‐educated
workers.
[Table 2.4 around here]
3. EstimationApproachIn the analysis, we apply simple linear regression models to compare the productivity levels and
characteristics of jobs in new and old firms, and to determine how this depends on cluster growth.
Specifically, we estimate a number of regressions of the following type at either the firm level or the job
level:
(1)
where Y is the dependent variable, e.g., the average wage (or productivity level) in the firm, or the wage
in the individual job if the estimation is at the job level. X is a vector of control variables, such as age and
gender – either at the firm or the job level.
We operate with dummy variables for established firms and (different types of) new firms and interact
them with dummy variables for whether the firm (or the job) is located in a growing or declining cluster.
For ease of exposition, equation (1) only contains one dummy variable for the firm type, DNF, which
equals one for new firms and zero for old firms. Later, DNF is split into three dummy variables for the
13
12
three different types of new firms. Similarly, the dummy variable, DGC, equals 1 for growing clusters and
0 for declining clusters. Note that we only distinguish between growing and declining clusters (not
between more detailed growth intervals of the clusters) in the regression analysis. This reduces the
number of coefficients to be estimated considerably and makes it easier to interpret the results.
In equation (1), the parameter �� (the constant term) measures the average value of Y in the reference
category, which is an old firm in a declining cluster. An old firm in a growing cluster on the other hand is
measured by �� � ��, i.e., �� measures the difference between an old firm in a growing cluster and an
old firm in a declining cluster. The value of a new firm in a declining cluster is given by �� � ��, i.e., �� measures the difference between new firms and old firms in a declining cluster. A new firm in a growing
cluster is measured by �� � �� � �� � ��, i.e., �� � �� measures the difference between new firms and
old firms in a growing cluster, whereas �� measures the difference between two differences, namely the
difference between new and old firms in growing clusters and the difference between new and old firms
in declining clusters.
In the following Sections, we estimate (1) for a number of different dependent variables. Our main
purpose is to characterize jobs and to compare them across firm types. To this end, we describe job
characteristics using a number of measures; wages, skills, and productivity. In firm‐level regressions, we
use the average hourly wage of the firm, the skill intensity of the firm measured as the share of
employees with a tertiary education among all employees, as well as labour productivity as measured by
firm sales per employee. For the latter measure, we also control for capital and intermediate inputs to
get a measure of TFP. In job‐level regressions, we use the hourly wage and the education level of the
individual that possesses the job.
In each case, we present a number of different regressions. First, we present a regression that only
distinguishes between old and new firms, i.e., we do not take differences between growing and
declining clusters into account. This implies that we implicitly impose the restriction �� � �� � � in equation (1). This is the approach that resembles the existing literature most closely. Second, we add
the distinction between growing and declining clusters and hence estimate all four parameters (��to ��). In neither of these regressions do we include additional control variables.
In the third regression, we also include controls for the age, gender and education level of the worker(s)
– except in the case where the dependent variable is the education level of the worker(s) herself. By
including this information, we take worker heterogeneity into account when comparing job
14
13
characteristics of different firms. If this is an important aspect, the estimated firm‐type dummies are
expected to be of less importance after controlling for these worker characteristics.
In the fourth regression, we also include industry and region dummies thereby taking (potentially
unobserved) differences across industries and regions into account. If, e.g., certain industries or regions
tend to have high employment growth, the dummy variable for growing clusters will become less
important as part of the variation will now be picked up by the industry and region dummies.
In the fifth regression, we take this one step further and include cluster dummies, implying that we
control for observed and unobserved differences at the cluster level. This regression thereby yields
within‐cluster differences between established and new firms. As a consequence, cannot be
estimated in this regression. If the differences in job characteristics between new and old firms depend
on different (unobserved) cluster characteristics, the estimated difference between new and old firms
will become smaller in this case.
Finally, in the job‐level regressions we re‐estimate the third‐fifth regressions taking firms size into
account. This is motivated by the finding that firm size may play an important role, see, e.g., Troske
(1999) who finds that smaller firms pay lower wages. I.e., in the worker‐level regressions we present 8
regressions per dependent variable.
In Section 4, we only distinguish between new and old firms, whereas we apply the three different types
of new firms (spin‐offs, entrepreneurial start‐ups, and other new firms) in Section 5. In Section 6 we
consider the manufacturing and service sector separately, whereas we consider the case where new
firms are only those established after 2005 in Section 7. Section 8 introduces education as a third
dimension in our cluster definition.
4. NewFirmsvs.OldFirmsIn this Section, we investigate differences in the characteristics of jobs in new and old firms, taking the
growth in employment in the industry‐region cluster into account. In Section 4.1, we use the firm‐level
dataset to consider differences in average hourly wages, the proportion of workers with a tertiary
education degree, sales per worker and TFP. In Section 4.2, we turn to our job‐level dataset, where we
look at wages and education.
15
14
4.1Firm‐levelanalysisAs our first measure of job quality, we use the wage rate. Higher wages can reflect a number of effects
but are thought of as being equal to the value of the marginal product of labour. In Table 4.1, we
therefore present the results of estimating (1) using the log of the average hourly wage as the
dependent variable. As some firms do not have employees, and since wage information is not reliable
for all firms with employees, the sample is reduced to around 95,000 observations (firms) in this case
(out of a total of 188,000 firms).5 For the same reason, the number of clusters is reduced from 1,081 to
1,054.
[Table 4.1 around here]
In column 1, only the dummy for a new firm, DNF, is included. In this case, we find that new firms pay
slightly higher average wages (around 2 per cent) than old firms. This result is somewhat surprising as
entrepreneurs are typically found to pay lower wages than established firms in the literature.
In column 2, we condition on the cluster growth by including the dummy for positive employment
growth in the cluster, DGC, and its interaction with the new‐firm dummy, DNF. In this case, we can
estimate all the four beta coefficients from equation (1). We find that average wages of old firms are
higher in growing clusters than in declining clusters, as is significantly positive. The difference,
however, is less than 2 per cent. New firms, on the other hand, now exhibit lower wages (1 per cent)
than established firms in declining clusters ( is negative), but higher wages than established firms
(around 7%) in growing clusters ( is positive). In other words, the local business environment
does seem to make a difference for the relative performance of new vs. old firms.
In the following columns, we check the robustness of these findings when including additional controls.
In column 3, we control for the share of women among the workers, the share of workers with a tertiary
education degree, and the average age among workers. The inclusion of these controls removes the
difference in wages between old firms in declining and growing clusters, as is no longer statistically
different from zero. The differences between new and old firms also change a bit. When we control for
5 Statistics Denmark’s hourly wage variable is based on the available information from the wage income registers and different sources of information to determine working hours. The measurement of the latter can easily become highly uncertain, for which reason Statistics Denmark attaches a quality estimate to its hourly wage variable. If this estimate makes the wage information of a given individual too doubtful, the given individual’s wage record is not considered for the analysis.
16
15
gender, education and age, new firms actually pay higher wages in both growing and declining clusters,
but the difference is still largest in growing clusters, where it remains around 7 per cent ( ).
Finally, Columns 4 and 5 include industry and region dummies and industry x region dummies,
respectively, where the latter corresponds to fixed effects at the cluster level. In these cases, both
and become insignificant. That is, we do not observe any differences between new and old firms
when we control for (possibly unobserved) industry‐ and region‐specific characteristics. In other words,
within a cluster, we do not observe any (significant) differences between new and old firms when it
comes to average wages.
In Table 4.2, we consider the skill intensity in new vs. old firms, where we use the share of employees in
the firm with a tertiary education degree as the dependent variable. Note that in this case, the sample
size almost doubles compared to Table 4.1.
[Table 4.2 around here]
According to the first column of Table 4.2, the unconditional skill‐intensity is 2 percentage points higher
in new firms than in old firms (the estimate of is both positive and significantly different from zero).
When distinguishing between growing and declining clusters, it is found that the education‐intensity is
around 10 percentage points higher in old firms in growing clusters compared to old firms in declining
clusters, as the estimate of equals 0.098. Moreover, new firms are found to be more education‐
intensive than old firms – both in declining clusters (0.4 percentage points) and especially in growing
clusters (0.4 + 2.4 = 2.8 percentage points).
Controlling for worker heterogeneity in column 3 does not change this picture qualitatively, and neither
does the inclusion of region and industry dummies in column 4, although the difference between new
and old firms in growing clusters become quantitatively less important. When looking at within cluster
differences in column 5, we also still find that new firms have a higher skill intensity – around 1
percentage point in declining clusters and around 1.5 percentage points in growing clusters – but the
difference between growing and declining clusters (i.e., ) is no longer significant.
In Table 4.3, we present the results for the case where the dependent variable is the (log of) sales (or
revenue) per worker. From column 1, we can see that sales per worker are, in general, lower in new
firms (16 per cent). When we distinguish between growing and declining clusters (column 2), we still
find that new firms have lower sales per worker – and there is no statistically significant difference
17
16
between declining and growing clusters (the estimate of is insignificant). Controlling for worker
heterogeneity and fixed effects (columns 3‐5) does not alter this conclusion qualitatively.
[Table 4.3 around here]
Table 4.4 also has (the log of) sales per worker as the dependent variable, but includes (the log of)
capital per worker, (the log of) intermediate inputs per worker and (the log of) number of workers as
additional controls. In this case, the estimates of to can be interpreted as (differences in) total
factor productivity. In column 1, we can see that this results in new firms having slightly higher TFP (1.5
percent) on average. When distinguishing between growing and declining clusters (column 2), it turns
out that the “advantage” of new firms is found only within growing clusters as the estimated sum of
and amounts to 5.5 per cent), whereas the estimate of is negative, implying that new firms in
declining clusters have slightly lower TFP (1 per cent) than old firms in declining clusters. When we
control for worker differences (column 3), however, new firms are found to have higher productivity in
both growing and declining clusters, and the same holds when controlling for industry and region fixed
effects (column 4). Even with cluster fixed effects (column 5), new firms are found to have a productivity
advantage in growing clusters.6
[Table 4.4 around here]
Thereby, we find that new firms have lower sales per worker, but tend to have higher TFP (at least in
growing clusters). This indicates that old firms are either more capital intensive or use more
intermediate inputs per worker. In regressions not reported, we find that new firms are indeed less
capital intensive and use less intermediate input per worker.7
Before turning to our worker‐level regressions, we summarize the findings at the firm level: We find that
new firms on average have higher wages, are more skill intensive, have lower sales per worker but have
higher TFP. While the difference in wages disappears when controlling for industry and region (or
cluster) characteristics, the findings that new firms are more skill intensive, have lower sales and higher
TFP are more robust. Furthermore, we find evidence that the higher wages, skill intensity and
6 Note that throughout Table 4.4, the coefficients of capital per worker and input per worker are as expected, whereas the coefficient of (the log of) the number of workers is found to be significantly positive, which indicates increasing returns to scale. 7 Results are available upon request.
18
17
productivity of new firms are more pronounced in growing clusters. Even when we use only within‐
cluster variation, we find differences between new and old firms. Specifically, we find that new firms are
more skill intensive, have lower sales but higher TFP (in growing clusters).
4.2Job‐levelanalysisTable 4.5 presents results where we use the (log of the individual) hourly wage as the dependent
variable. The first five columns are comparable to the five columns in Table 4.1.
Note that the results for individual wages are quite different from those presented for average firm
wages in Table 4.1. Throughout the first five columns, we find that individual wages are lower in new
firms than in old firms, but less so when we control for worker heterogeneity in column 3. When
controlling for industry and regional fixed effects in column 4 (or cluster fixed effects in column 5), we
even find that the difference is larger in growing clusters. Thus, in column 5, we find that wages in new
firms in declining clusters are 3.7 per cent lower than in old firms in the same declining clusters, and that
the difference is 4.7 per cent in growing clusters (the sum of and ).
[Table 4.5 around here]
The difference between the results in Table 4.1 and Table 4.5 can be explained by the fact that the
regressions at the job level (Table 4.5) are weighted versions of the regressions at the firm level (Table
4.1), where the weights used are the numbers of jobs at the firm level. In other words, when large firms
are weighted more than small firms (as it is the case in Table 4.5), we find that wages are higher in old
firms.
In light of this, it might be interesting to see how results are affected if we control directly for firm size.
In the literature, it is thus a relatively robust finding that wages are increasing with firm size, see, e.g.,
Troske (1999) and Brown and Medoff (2003). The last three columns in Table 4.5 are therefore similar to
columns 3‐5 but include firm size (as measured by total employment) as an additional control.8
When controlling for firm size, new firms are again found to pay higher wages, at least in declining
clusters, where the difference is between 1.6 and 1.9 per cent. In growing clusters, on the other hand,
there is only a minor wage difference between new and old firms when we control for firm size. Thus,
8 We apply a non‐parametric approach and include log(employment), log(employment)2,.., and log(employment)10 in regressions 6‐8.
19
18
individual wages are not necessarily higher in old firms because they are old, but perhaps because the
old firms tend to be larger.
In Table 4.6, we use a linear probability model where the dependent variable is a dummy for the
individual having a tertiary education. That is, we estimate the probability that a given job is filled with
an individual that has a tertiary education and how this probability is affected by whether the firm is
new or old, and whether the firm is located in a growing or declining cluster. The results can thus be
seen as job‐level versions of the estimations of skill intensities in Table 4.2.9
When we do not distinguish between growing and declining clusters (column 1), we do not find any
statistically significant difference between new and old firms. In column 2, however, we find that the
probability of having completed tertiary education is marginally lower in new firms than in old firms in
declining clusters, whereas the probability is slightly higher (but not significant) in growing clusters.
These results do not change when we control for worker heterogeneity (column 3). When we add
industry and region dummies (column 4) or cluster dummies (column 5), the positive effect of new firms
in growing clusters disappears.
[Table 4.6 around here]
In sum, using our job‐level dataset, we find a slightly smaller tendency of the workers in new firms to
have completed tertiary education. This result is in some contrast to the results at the firm level in Table
4.2, where it was found that new firms are more skill intensive than old firms. As in the case of Tables
4.1 and 4.5, this difference again reflects the different weighting of the observations in the worker‐level
and firm‐level regressions. Thus, among the old firms, especially the larger firms tend to be more skill
intensive, and these firms are given more weight in the worker‐level regressions. This is also evident
from columns 6‐8 where we control directly for firm size. In this case, the estimate of is no longer
significant (although remains significantly negative but numerically small when controlling for cluster
fixed effects in column 8). Thus, parallel to the case of Table 4.5, the higher probability of having a
tertiary education in old firms may not be due to these firms being old but because old firms tend to be
larger.
9 The reason for choosing a linear probability model is that it is difficult to estimate probit or logit models with several hundred industry, region and cluster dummies. In column 4 there are 233 industry dummies and five regional dummies, whereas there are more than 1,000 cluster dummies in column 5.
20
19
To summarize the findings at the job level, we find that individual wages are lower in new firms
compared to old firms (as long as we do not control for firm size). In addition, we find hardly any
differences across new and old firms in the probability that a job is occupied by an individual with
tertiary education.
Taken together, the findings in this section have shown that new firms are more productive and skill
intensive and pay higher wages than old firms – and these differences are typically more pronounced in
growing clusters. However, the wage and skill premiums at the firm level disappear at the job level, as
the larger old firms are both more skill intensive and pay higher wages than the smaller old firms.
5. DifferentTypesofNewFirmsIn this Section, we extend the analysis of Section 4, and distinguish between different types of new
firms: spin‐offs, entrepreneurial start‐ups and other new firms. The results are organised in a parallel set
of Tables to those in Section 4. Thus, Table 5.1 is comparable to Table 4.1 etc., with the only difference
being that and are each split into three different dummies ( and )
representing the three different types of new firms. We will not go through all results in detail but will
focus on the findings that differ from or qualify those of Section 4.
5.1Firm‐levelanalysisIn Table 5.1 we use the log of the average hourly firm wage as the dependent variable. Compared to
Table 4.1, we find that it is the group of “other new firms” that drives the result of higher average wages
in new firms (column 1). Thereby, the results in Column 1 partly confirm existing research findings of
entrepreneurial jobs being relatively low paid. At least this is the case when we look at the average
wages in entrepreneurial start‐ups, which are slightly lower (1.5 per cent) than in old firms. Average
wages in other new firms are, on the other hand, 18 per cent higher than in old firms.
[Table 5.1 around here]
When we distinguish between growing and declining clusters (columns 2 and 3), entrepreneurial start‐
ups and spin‐offs seem to fare worse than or as well as old firms in declining clusters (as seen from the
estimates of and ), but better than old firms in growing clusters (as seen from the estimates of
and ). In this sense, entrepreneurial firms in growing clusters are generating high‐
wage jobs.
21
20
Most of these differences, however, disappear when we add industry and region (or cluster) fixed
effects (columns 4 and 5). Other new firms are still found to pay higher wages (11 per cent) in both
growing and declining clusters, while entrepreneurial firms pay on average 3 per cent lower wages than
established firms. There are no significant differences between spin‐offs and old firms.
Table 5.2 considers the skill intensity of firms as measured by the share of employees with a tertiary
education. The main impression from Table 5.2 is that new firms are relatively more skill intensive than
old firms, and this holds for all three types of new firms. However, only in the case of spin‐offs do we
find an additional positive effect in growing clusters when we add cluster fixed effects (column 5).
[Table 5.2 around here]
Table 5.3 presents results for sales per worker. It is evident that the lower sales per worker found for
new firms in Table 4.3 is due to spin‐offs and entrepreneurial firms, whereas other new firms in general
have higher sales per worker than old firms (although the difference is not statistically significant when
including cluster fixed effects in column 5). Note, however, that in the case of spin‐offs, they only have
lower sales compared to old firms when located in declining clusters. In growing clusters, the sales of
spin‐offs are at par with those of established firms (as can be seen from the estimate of ).
[Table 5.3 around here]
Table 5.4 provides an interesting qualification of the findings in Table 4.4. The overall higher productivity
of new firms found in Table 4.4 is apparently due to spin‐offs and other new firms (column 1 of Table
5.4). When we distinguish between growing and declining clusters, we can see that in particular spin‐
offs enjoy even higher productivity in growing clusters compared to old firms – and this result is robust
to the inclusion of region and industry (or cluster) fixed effects. In column 5, the productivity advantage
of spin‐offs in growing clusters compared to old firms in growing clusters thus adds up to around 10 per
cent. For comparison, the productivity advantage of other new firms disappears as we include fixed
effects for region and industry (column 4) or clusters (column 5). Finally, entrepreneurial start‐ups are
less productive than established firms in declining clusters, but in growing clusters, they seem to be
equally productive as can be seen from the estimate of .
[Table 5.4 around here]
22
21
The overall conclusion of this sub‐section is that the higher wages in new firms are driven by other new
firms, although there is some evidence that spin‐offs and entrepreneurial start‐ups fare better than old
firms when located in growing clusters. Higher skill intensity is found in all three types of new firms, but
spin‐offs also exhibit additional skill intensity when located in growing clusters. Finally, the higher
productivity of new firms is driven in particular by spin‐offs.
5.2Job‐levelanalysisEstimation results for individual wages are presented in Table 5.5. Compared to Table 4.5, we can see
that the lower wages in new firms are due to spin‐offs and entrepreneurial start‐ups, whereas other
new firms actually exhibit higher wages in declining (but not growing) clusters when we control for
industry and region (or cluster) fixed effects (columns 4 and 5).
As in Table 4.5, the lower wages can be explained by spin‐offs and entrepreneurial start‐ups being
smaller than old firms. Thus, when we include controls for firm size (columns 6 to 8), spin‐offs are not
found to pay lower wages, and entrepreneurial start‐ups are found to pay lower wages in growing
clusters only. Other new firms are now found to pay higher wages (in particular in declining clusters).
[Table 5.5 around here]
Table 5.6 estimates the probability that a given individuals has completed tertiary education. Some
differences across the three types of new firms are observed. Specifically, the probability seems to be
higher in other new firms in declining clusters – and also in growing clusters as long as we do not control
for industry and region (or cluster) fixed effects. For spin‐offs and entrepreneurial start‐ups, on the
other hand, we find a lower probability – at least when these are located in declining clusters. In
growing clusters, the probability is not lower for spin‐offs compared to old firms.
Columns 6‐8 include controls for firm size. This does not change the results qualitatively.
[Table 5.6 around here]
In sum, the job‐level analysis shows that the lower individual wages in new firms that we found in
Section 4.2 are driven mainly by spin‐offs and entrepreneurial start‐ups. These two types of firms also
tend to imply a lower probability of having tertiary education
23
22
Overall, the analyses in this section have shown important differences between the three types of new
firms. Spin‐offs enjoy the largest productivity advantage compared to old firms, whereas higher wages
are mainly found within other new firms, although spin‐offs and entrepreneurial start‐ups fare better
than old firms when located in growing clusters. However, these wage premiums disappear in the job‐
level regressions, which generally show lower wages in spin‐offs and entrepreneurial start‐ups than in
old firms, whereas other new firms are at par with old firms. The lower wage rates however can mainly
be explained by firm size. Skill intensities are higher in all three types of new firms, but the skill
premiums in the new firms again disappear in the job‐level analyses.
6. ManufacturingSectorvs.ServiceSectorIt might be hypothesized that the difference in performance between new and old firms looks different
in the manufacturing sector and in the service sector. Thus, many of the new firms in the service sector
consist of retailers and other small service firms such as hairdressers. In the manufacturing sector, on
the other hand, it takes both a larger investment and perhaps also a more novel idea to set up a new
firm. If that is the case, then the performance and jobs of new firms in the manufacturing sector may
look quite differently.
To analyse any differences across the two sectors, we have re‐estimated the tables from Sections 4 and
5 separately for each sector, where the manufacturing sector is defined as NACE code section C (C:
Manufacturing), and the service sector as NACE code sections G‐N (G: Wholesale, Retail Trade; Repair of
Motor Vehicles and Motorcycles, H: Transportation and storage services, I: Accommodation and food
service activities, J: Information and communication services, K: Financial and insurance activities, L:
Real estate activities, M: Professional, scientific and technical activities, N: Administrative and support
service activities). To save space, the tables are not included in the paper but are presented in Appendix
A and B.
The results for the service sector turn out to be very similar to the overall results from Sections 4 and 5.
This is a consequence of the fact that most of the firms and workers in the data belong to the service
sector. Thus, out of the approximately 188,000 firms and 1,560,000 jobs in Table 2.1, approximately
120,000 firms and 1 million jobs belong to the service sector, whereas 13,500 firms and 120,000 jobs
belong to the manufacturing sector. As a consequence, in what follows we focus on the differences in
the manufacturing sector compared to the general results in Sections 4 and 5.
24
23
First, the finding of higher average wages in new firms is now more robust to the inclusion of fixed
effects (Table A4.1). On the other hand, average wages do not seem to depend on cluster growth.
However, in the individual wage regressions (Table A4.5), we now also find that new firms pay as high
(or even higher) wages than old firms when located in growing clusters (but lower when located in
declining clusters). When we distinguish between the three types of new firms as in Section 5,
entrepreneurial start‐ups and spin‐offs now pay average wages of the same magnitude as old firms in
both growing and declining clusters in the firm‐level regressions (Table A5.1), while the advantage are
for other new firms. In the job‐level regressions (Table A5.5), the negative wage premium in
entrepreneurial start‐ups is now also less pronounced, while spin‐offs in growing clusters are now found
to pay higher wages than old firms.
Second, sales per worker (Table A4.3) are still found to be lower in new firms – and now even more in
declining clusters (although results are not very significant). The productivity advantage of new firms is
more pronounced and significant within manufacturing, but again does not depend on cluster growth
(Table A4.4). Interestingly, when we distinguish between different types of new firms (Table A5.4), we
find that the productivity advantage is now found across all three types of new firms – not just spin‐offs
(and other new firms).
Third, the higher skill intensity of new firms in firm‐level regressions is also found in the manufacturing
sector (Table A4.2), but not in spin‐offs (Table A5.2). In the job‐level regressions, we again find a general
disadvantage of new firms (Table A4.6), and the skill advantage of other new firms (Table A5.6)
disappears.
Overall, the results for the manufacturing sector look much like the general results. However, the
productivity advantage and higher average wages of new firms appear more robust in the
manufacturing sector.
7.Robustness:AgeofNewFirmsAs a robustness check of our results, we have re‐estimated the tables of Sections 4 and 5, where we
define new firms as those established after 2005, instead of all firms established after 2000. The results
are presented in Appendix C. This reduces the number of new firms in the analysis from around 80,000
to 60,000, and it increases the number of old firms correspondingly. Using this definition of new firms,
we find very similar results to those of Sections 4 and 5.
25
24
The only difference between the baseline results in Sections 4 and 5 and the results based on the
alternative definition of entrepreneurs is with respect to productivity. For this measure, it is found that
new firms in general are now less productive than old firms. The lower productivity is driven by new
firms in declining clusters, whereas new firms in growing clusters have productivity levels of similar
magnitude to old firms.
The analyses also show important differences between the three types of new firms. Spin‐offs still enjoy
productivity advantage compared to old firms, especially in growing clusters, whereas entrepreneurial
start‐ups and other new firms have similar or lower productivity compared to old firms.
As the new firms considered in this section are younger than the new firms considered in the previous
sections, the difference in results suggests that it takes some time for organic new firms to catch‐up and
exceed incumbents in terms of productivity. This finding is also consistent with the finding that
entrepreneurs have higher growth of productivity, see Huergo and Jaumandreu (2004).
8.Robustness:3‐dimensionalclustersAs an additional robustness check of our results, we extend the definition of clusters to contain a third
dimension – the education of the individual. In this case, a cluster is defined as jobs (individuals) with a
certain educational background in a given industry and region. We can then compare jobs in new and
old firms within these education‐specific clusters.
Specifically, we use 233 industries, five regions and five education categories. This results in 6,900
clusters of which 312 clusters are new in 2010. This also allows us distinguish between declining,
growing and new clusters in the analysis.
The introduction of an educational dimension in the cluster makes comparisons at the firm level
impossible, as a firm can now be located in up to five different clusters. Similarly, skill‐intensity
estimations do not make sense when clusters are defined by education. Hence, in this section, we focus
on job‐level results for wages.
Table 8.1 below is comparable to Table 4.5 from Section 4, whereas Table 8.2 is comparable to Table 5.5
from Section 5. Instead of having one dummy for growing clusters, we now have two dummies: One for
growing clusters and one for new clusters. We still have one dummy for new firms (Table 8.1) and three
26
25
dummies for three types of new firms (Table 8.2), but these are now each interacted with both the
dummy for growing clusters and the dummy for new clusters, which results in more interaction terms.
[Tables 8.1 and 8.2 around here]
Still, we find that wages are lower in new firms however the difference disappears and becoming
positive when controlling for firm size (Table 8.1). We also find that wages are lower in spin‐offs and
entrepreneurial start‐ups are lower in declining and growing clusters (Table 8.2). The results for new
clusters are less significant for entrepreneurial start‐ups but might reflect that the number of these
clusters is relatively limited. The results for other new firms are more mixed. When we control for region
and industry fixed effects (or cluster fixed effects), they are found to pay higher wages than old firms in
declining clusters. The results for growing clusters are less clear. When it comes to new clusters, other
new firms seem to pay higher wages as long as we do not control for region and industry fixed effects
(or cluster fixed effects), in which case the results are reversed. All in all, the results in Tables 8.1 and 8.2
seem to confirm the picture from Table 4.5 and Table 5.5.
The results in this section, thereby, do not change the overall conclusions from the previous sections.
9.ConclusionDespite a general belief that entrepreneurs are important for job creation and growth, existing research
has indicated that jobs created by entrepreneurs are often unsecure and relatively low paid. Employees
in start‐ups often earn lower wages than employees in other firms, and productivity levels in
entrepreneurial start‐ups are often lower than in established firms.
This paper has shed new light on the quality of jobs created by start‐up firms compared with similar
measures in established firms. Compared to previous studies, we measure the entrepreneur as the
organic new firm instead of using firm age or firm size as criteria to distinguish entrepreneurs from
incumbent firms. Moreover, we have taken the business environment of the firms into account by
distinguishing between firms located in growing industry‐region clusters and firms in declining industry‐
region clusters. Thus, instead of just comparing new firms to old firms, we compare the jobs created by
entrepreneurial firms in growing (declining) clusters to jobs (or productivity levels) in old firms in
growing (declining) clusters. We believe that this provides a more appropriate assessment of the
contribution of entrepreneurial firms. Furthermore, we also distinguish between different types of
27
26
organic new firms (spin‐offs, entrepreneurial start‐ups and other new firms) to investigate whether
some types are more similar to old firms than others.
The analyses are based on Danish worker‐firm register data for the time period 2001‐2010, which cover
almost the entire private sector of the Danish economy. A particularly interesting feature of these data
is that Statistics Denmark has undertaken extensive efforts to identify the organic new firm among the
formally new firms.
Overall, we find that entrepreneurial firms are more productive, more skill intensive, and pay higher
wages than old firms. Moreover, these differences are more pronounced in growing industry‐region
clusters. We also find that spin‐offs enjoy the largest productivity advantage compared to old firms,
whereas higher wages are mainly found within other new firms, although spin‐offs and entrepreneurial
start‐ups fare better than old firms when located in growing clusters. However, the wage and skill
premiums at the firm level disappear in the job‐level regressions, which generally show lower wages in
spin‐offs and entrepreneurial start‐ups than in old firms. The difference in results between the two
approaches reflects that the larger old firms are given more weight in the individual‐level regressions,
and these are different from the smaller old firms. When we look specifically at the manufacturing
sector, we find more robust productivity advantages and higher average wages of new firms compared
to old firms.
The broader perspective of the results presented in this paper is that jobs in organic new firms are not
low‐quality jobs in the sense that they pay lower wages and have lower skill content. Moreover, organic
new firms do not seem to have lower productivity. This is especially the case for firms in growing
clusters. These results indicate that entrepreneurial firms do play an important role as an engine for
growth and prosperity. The overall result of the paper is thus consistent with Schumpeter (1934, 1943)
who argues that entrepreneurs create combinations of inputs and outputs to pioneer new activities,
exploit new market opportunities and allocate labour to its most productive use.
The viewpoint that entrepreneurs are of key importance for generating new jobs and economic growth
in an increasingly competitive international environment have given rise to initiatives such as the Small
Business Act for Europe and the SBIR program in the USA; programmes that reallocate substantial
resources from large established firms to small firms and start‐ups. While the majority of studies in the
literature have found that jobs created in such firms are less attractive with lower levels of human
capital and lower wages (see Van Praag and Versloot, 2007), the present study provides some evidence
28
27
for the opposite. However, the results depend to some extent on the definition of entrepreneurs both
with respect to type of entrepreneurial firm and with respect to for how long firms can be considered to
still be entrepreneurial firms.
29
28
REFERENCESBrixy, U., Kohaut, S. and Schnabel, C., (2007), “Do Newly Founded Firms Pay Lower Wages? First
Evidence from Germany” Small Business Economics, Vol. 29, No. ½, 161‐171
Brouwer, P., de Kok, J. and Fris, P., (2005), “Can firm age account for productivity differences?”, SCALES‐
paper N200421, Zoetermeer, Netherlands
Brown, C., & J.L. Medoff (2003). Firm age and wages. Journal of Labor Economics, 21(3), 677‐697
Castany, L., López‐Bazo, E. and Moreno, R. (2005). “Differences in Total Factor Productivity Across Firm
Size. A Distributional Analysis,” University of Barcelona Working Paper
Dahl, M. S., & Reichstein, T. (2007). Are You Experienced? Prior Experience and the Survival of New
Organizations. Industry & Innovation, 497‐511
Disney, R., Haskel, J. and Heden, Y., (2003), “Restructuring and Productivity Growth in UK
Manufacturing”, The Economic Journal, Vol. 113, No. 489, 666‐694
Eriksson, T., & Kuhn, J. M. (2006). Firm spin‐offs in Denmark 1981‐2000 ‐ patterns of entry and exit.
International Journal of Industrial Organization, 1021‐1040
Foster, L., Haltiwanger, J. and Krizan, C.J., (2006), “Market selection, reallocation, and restructuring in
the U.S. Retail sector in the 1990s”, The Review of Economics and Statistics, 88(4), 748‐758
Goos, M., & Manning, A. (2007), "Lousy and Lovely Jobs: The Rising Polarization of Work in Britain", The
Review of Economics and Statistics, 89(1), 118‐133
Huergo, E., and J. Jaumandreu (2004), "Firms’ age, process innovation and productivity growth",
International Journal of Industrial Organization, 22, 541– 559
Jensen, J.B., McGuckin, R.H. and Stiroh, K.J. (2001), “The impact of vintage and survival on Productivity:
Evidence from cohorts of U.S. Manufacturing plants”, The Review of Economics and Statistics,
83(2), pp. 323‐332
Koch, A. and Späth, J., (2009), “New Firms – Different Jobs? An Inquiry into the Quality of Employment in
Start‐ups and Incumbents”, IAW‐Discussion Paper, No. 50
30
29
Kölling, A., Schnabel, C. and Wagner, J., (2002), “Establishment Age and Wages: Evidence from German
Linked Employer‐Employee Data”, IZA Discussion Paper, No. 679
Nyström, K. and Elvung, G.Z., (2014), “New firms and labor market entrants: Is there a wage penalty for
employment in new firms?”, Small Business Economics, Vol. 43, pp. 399‐410
Schumpeter, J. A. (1934). The Theory of Economic Development. Cambridge, Mass.: Harvard University
Press.
Schumpeter, J. A. (1943). Capitalism, Socialism and Democracy. New York: Harper.
Troske, K.R., (1999), “Evidence on the Employer Size‐Wage Premiun from Worker‐Establishment
Matched Data”, The Review of Economics and Statistics, Vol. 81, No. 1, pp.15‐26
van Praag, M., & Versloot, P. H. (2007). What is the value of entrepreneurship? A review of recent
research. Small Business Economics, s. 351‐382.
31
TABLE 2.1.: N
umbe
r of clusters a
nd jo
bs. B
y cluster g
rowth
Num
ber o
f clusters
Num
ber o
f job
sNum
ber o
f firm
s
Cluster g
rowth
#%
#%
#%
‐0.50‐
124
11.47
38,334
2.46
4,31
42.29
‐0.50;‐0.25
202
18.69
222,93
314
.29
19,820
10.54
‐0.25;‐0.1
190
17.58
366,29
823
.47
47,918
25.48
‐0.1‐0.1
253
23.40
634,57
940
.67
76,068
40.45
0.1‐0.25
857.86
121,98
47.82
19,838
10.55
0.25‐0.5
656.01
113,09
77.25
14,610
7.77
0.5+
137
12.67
62,396
4.00
5,43
72.89
New
252.31
831
0.05
390.02
Totals
1,08
110
0.00
1,56
0,45
210
0.00
188,04
410
0.00
32
TABLE 2.2: Num
ber o
f firm
type
s and
jobs. B
y cluster g
rowth and
firm
type
sNum
ber o
f firm
s
Establish
ed firm
sSpin‐offs
Other new
firm
sCluster g
rowth
#%
#%
#%
#%
‐0.50‐
2439
2.3
436
1.8
1,14
52.5
294
3.2
‐0.50;‐0.25
1192
511
.02,
337
9.7
4,68
310
.087
59.7
‐0.25;‐0.1
2836
026
.28,
072
33.6
9,81
221
.01,67
418
.5 ‐0
.1‐0.1
4495
441
.58,
884
37.0
18,2
7039
.23,96
043
.7 0.1‐0.25
1099
710
.22,
167
9.0
5,44
411
.71,23
013
.6 0.25‐0.5
7092
6.5
1,75
37.3
5,09
010
.967
57.5
0.5+
2565
2.4
368
1.5
2,19
34.7
350
3.9
Totals
108,33
2
100.0
24,017
10
0.0
46,637
10
0.0
9,05
8
100.0
Num
ber o
f job
s
Establish
ed firm
sSpin‐offs
Other new
firm
sCluster g
rowth
#%
#%
#%
#%
‐0.50‐
31,250
2.4
2,04
22.5
3,29
72.9
1,74
53.1
‐0.50;‐0.25
195,99
415
.08,
013
9.8
12,9
1011
.46,01
610
.8 ‐0
.25;‐0.1
308,05
723
.524
,958
30.4
23,3
1020
.59,97
317
.9 ‐0
.1‐0.1
535,10
440
.933
,322
40.6
43,5
9038
.422
,563
40.6
0.1‐0.25
94,782
7.2
7,50
49.1
12,1
4410
.77,55
413
.6 0.25‐0.5
92,587
7.1
5,24
26.4
10,3
179.1
4,95
18.9
0.5+
51,228
3.9
1,08
51.3
8,08
87.1
2,82
65.1
Totals
1,30
9,00
2
100.0
82,166
10
0.0
113,65
6
100.0
55,628
100.0
Entrep
rene
urial
Entrep
rene
urial
33
TABLE 2.3: Average
num
ber o
f job
s per firm
. By cluster g
rowth after firm
type
sJobs per firm
Establish
ed firm
sSpin‐offs
Entrep
rene
urial
start‐up
sOther
new
Cluster g
rowth
##
##
‐0.50‐
12.8
4.7
2.9
5.9
‐0.50;‐0.25
16.4
3.4
2.8
6.9
‐0.25;‐0.1
10.9
3.1
2.4
6.0
‐0.1‐0.1
11.9
3.8
2.4
5.7
0.1‐0.25
8.6
3.5
2.2
6.1
0.25‐0.5
13.1
3.0
2.0
7.3
0.5+
20.0
2.9
3.7
8.1
Totals
12.1
3.4
2.4
6.1
34
Table 2.4: Average
individu
al cha
racteristics. By firm ty
pe and
cluster growth
Cluster g
rowth
Firm
type
Num
ber o
f workers
Num
ber o
f firms
Mean age
Share of
females
Share of
highly
educated
Mean ho
urly
wage (DKK
)
Negative
Establish
ed firm
s81
0,10
371
,316
40.7
0.32
0.04
211.69
Negative
Spin‐offs
55,271
16,272
36.0
0.30
0.02
170.57
Negative
Enterprene
urial start‐ups
64,376
25,656
37.6
0.30
0.04
175.45
Negative
Other new
firm
s30
,740
5,33
337
.90.34
0.06
216.75
Positive
Establish
ed firm
s49
8,89
937
,016
38.1
0.42
0.11
224.28
Positive
Spin‐offs
26,895
7,74
536
.10.39
0.11
193.38
Positive
Enterprene
urial start‐ups
49,280
20,981
37.5
0.39
0.11
188.83
Positive
Other new
firm
s24
,888
3,72
537
.00.40
0.14
224.55
35
TABLE 4.1: Average
Wages in
New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)5.10
6***
5.10
2***
4.73
1***
6.01
9***
4.94
1***
(0.002
)(0.002
)(0.009
)(0.089
)(0.021
)
Positive cluster g
rowth (β
2)0.01
6***
0.00
4‐0.017
***
(0.005
)(0.004
)(0.005
)
New
firm
(β3)
0.02
2***
‐0.008
**0.01
8***
0.00
40.00
3(0.003
)(0.004
)(0.004
)(0.003
)(0.004
)
Positive cluster g
rowth x New
firm
(β4)
0.08
1***
0.04
9***
0.00
10.00
4(0.008
)(0.007
)(0.007
)(0.008
)Co
ntrols for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns95
,228
95,228
95,228
95,228
95,228
R‐squared
0.00
00.00
40.13
30.27
40.03
2Notes: The
dep
ende
nt variable is the log of th
e average ho
urly wage in th
e firm. The
sample is redu
ced to
arou
nd 95,00
0 firms a
nd th
e nu
mbe
r of clusters to 10
54, because so
me firms d
o no
t have em
ployees a
nd
because reliable wage inform
ation cann
ot be ob
tained
for a
ll firms w
ith employees. See
text fo
r more
details. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
36
TABLE 4.2: Skill‐Intensity
in New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)0.07
1***
0.03
7***
‐0.071
***
0.15
80.03
0**
(0.001
)(0.001
)(0.002
)(0.118
)(0.012
)
Positive cluster g
rowth (β
2)0.09
8***
0.09
5***
0.01
0***
(0.002
)(0.002
)(0.002
)
New
firm
(β3)
0.02
0***
0.00
4***
0.01
8***
0.00
7***
0.00
9***
(0.001
)(0.001
)(0.001
)(0.001
)(0.002
)
Positive cluster g
rowth x New
firm
(β4)
0.02
4***
0.02
2***
0.00
8***
0.00
6(0.003
)(0.003
)(0.002
)(0.003
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns18
8,04
418
8,04
418
7,63
318
7,63
318
7,63
3R‐squared
0.00
20.04
90.06
00.23
60.00
4Notes: The
dep
ende
nt variable is the the share of employees w
ith a te
rtiary edu
catio
n de
gree
in th
e firm.
The nu
mbe
r of clusters is 1
081, since some clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
37
TABLE 4.3: Sales per W
orker in New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)13
.404
***
13.469
***
13.747
***
18.117
***
13.518
***
(0.003
)(0.003
)(0.009
)(0.845
)(0.027
)
Positive cluster g
rowth (β
2)‐0.193
***
‐0.145
***
0.00
1(0.006
)(0.006
)(0.007
)
New
firm
(β3)
‐0.158
***
‐0.143
***
‐0.169
***
‐0.116
***
‐0.097
***
(0.004
)(0.005
)(0.005
)(0.005
)(0.008
)
Positive cluster g
rowth x New
firm
(β4)
‐0.004
0.00
1‐0.002
0.01
6(0.009
)(0.009
)(0.008
)(0.013
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns17
8,15
217
8,15
217
7,79
918
1,42
617
7,79
9R‐squared
0.00
80.02
00.04
40.31
30.00
6Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker. The nu
mbe
r of clusters is 1
021, since
some clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
38
TABLE 4.4: Produ
ctivity
in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)7.05
1***
6.97
7***
6.79
0***
6.92
6***
7.08
6***
(0.022
)(0.022
)(0.024
)(0.059
)(0.093
)
Positive cluster g
rowth (β
2)0.05
2***
0.03
8***
‐0.019
***
(0.005
)(0.005
)(0.005
)
New
firm
(β3)
0.01
5***
‐0.010
**0.01
7***
0.00
8**
0.00
7(0.003
)(0.004
)(0.004
)(0.004
)(0.007
)
Positive cluster g
rowth x New
firm
(β4)
0.06
5***
0.05
4***
0.02
4***
0.02
7*(0.007
)(0.007
)(0.007
)(0.015
)
log(Interm
ediate inpu
ts per worker)
0.39
7***
0.40
2***
0.40
8***
0.37
6***
0.37
6***
(0.002
)(0.002
)(0.002
)(0.002
)(0.012
)log(capital per worker)
0.11
7***
0.11
6***
0.11
1***
0.13
4***
0.13
3***
(0.001
)(0.001
)(0.001
)(0.001
)(0.007
)log(nu
mbe
r of w
orkers)
0.05
9***
0.05
8***
0.07
5***
0.06
0***
0.05
9***
(0.001
)(0.001
)(0.001
)(0.001
)(0.005
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns13
3,08
713
3,08
713
2,84
213
2,84
213
2,84
2R‐squared
0.52
70.53
00.53
70.62
10.46
9Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker in the firm. See
text fo
r more de
tails.
Robu
st standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
nu
mbe
r of clusters is 9
75. ***
p<0.01, ** p<
0.05
, * p<0.1.
39
TABLE 4.4A
: Intermed
iate In
puts per W
orker in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)12
.617
***12
.765
***13
.415
***13
.655
***12
.941
***
(0.005
)(0.005
)(0.016
)(0.117
)(0.044
)
Positive cluster g
rowth (β
2)‐0.480
***‐0.378
***
‐0.007
(0.010
)(0.010
)(0.010
)
New
firm
(β3)
‐0.293
***‐0.231
***‐0.298
***‐0.127
***‐0.130
***
(0.007
)(0.008
)(0.008
)(0.007
)(0.013
)
Positive cluster g
rowth x New
firm
(β4
‐0.101
***‐0.072
***
‐0.029
**‐0.025
(0.016
)(0.016
)(0.013
)(0.023
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
nono
nono
yes
Observatio
ns14
3,86
614
3,86
614
3,60
014
3,60
014
3,60
0R‐squared
0.01
10.04
40.07
40.35
60.01
0Notes: The
dep
ende
nt variable is the log of interm
ediate inpu
ts per worker in the firm. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is
compu
ted on
the de
‐meane
d data. The
num
ber o
f clusters is 9
79. ***
p<0.01, ** p<
0.05
, *
p<0.1.
40
TABLE 4.4B
: Ca
pital per W
orker in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)12
.168
***12
.186
***10
.628
***13
.761
***11
.695
***
(0.006
)(0.007
)(0.020
)(0.116
)(0.039
)
Positive cluster g
rowth (β
2)‐0.058
***
‐0.033
**0.00
2(0.013
)(0.013
)(0.013
)
New
firm
(β3)
‐0.438
***‐0.464
***‐0.248
***‐0.177
***‐0.181
***
(0.009
)(0.011
)(0.011
)(0.009
)(0.020
)
Positive cluster g
rowth x New
firm
(β4)
0.07
9***
0.03
2*0.04
5***
0.06
0*(0.019
)(0.018
)(0.015
)(0.032
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cno
nono
noyes
Observatio
ns14
7,21
014
7,21
014
6,91
514
6,91
514
6,91
5R‐squared
0.01
60.01
60.07
30.39
90.01
2Notes: The
dep
ende
nt variable is the log of capita
lper worker in the firm. See
text fo
r more
details. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
num
ber o
f clusters is 9
82. ***
p<0.01, ** p<
0.05
, * p<0.1.
41
TABLE 4.5: W
ages in
New
and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)5.33
4***
5.31
1***
4.58
3***
5.21
2***
4.68
1***
4.41
0***
5.02
8***
4.48
7***
(0.000
)(0.007
)(0.010
)(0.022
)(0.003
)(0.011
)(0.022
)(0.005
)
Positive cluster g
rowth (β
2)0.06
6***
0.02
5***
0.01
2**
0.01
2*0.00
3(0.020
)(0.010
)(0.005
)(0.007
)(0.005
)
New
firm
(β3)
‐0.098
***
‐0.117
***
‐0.070
***
‐0.037
***
‐0.037
***
0.01
9***
0.01
8***
0.01
6***
(0.001
)(0.009
)(0.006
)(0.003
)(0.001
)(0.004
)(0.003
)(0.001
)
Positive cluster g
rowth x New
firm
(β4)
0.04
8**
0.01
0‐0.016
**‐0.010
***
0.00
9‐0.016
**‐0.012
***
(0.022
)(0.012
)(0.008
)(0.002
)(0.009
)(0.007
)(0.002
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,17
0,48
91,17
0,48
91,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
8R‐squared
0.00
50.01
10.35
90.41
30.28
00.38
10.42
30.29
2Notes: The
dep
ende
nt variable is the log of th
e ho
urly wage. The
sample is redu
ced to aroun
d 1,17
0,00
0 individu
als a
nd th
e nu
mbe
r of clusters to
1053
, because re
liable wage inform
ation cann
ot be ob
tained
for a
ll workers. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐
squared in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
42
TABLE 4.6: Skills in
New
and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)0.06
9***
0.04
4***
‐0.031
***
0.15
50.03
5***
‐0.006
0.16
1*0.04
0***
(0.003
)(0.002
)(0.003
)(0.097
)(0.001
)(0.004
)(0.097
)(0.001
)
Positive cluster g
rowth (β
2)0.06
6***
0.07
4***
0.00
5**
0.07
3***
0.00
3(0.009
)(0.007
)(0.002
)(0.006
)(0.002
)
New
firm
(β3)
0.00
1‐0.007
***
‐0.005
**‐0.004
***
‐0.003
***
0.00
1‐0.001
‐0.001
(0.004
)(0.002
)(0.002
)(0.001
)(0.001
)(0.002
)(0.001
)(0.001
)
Positive cluster g
rowth x New
firm
(β4)
0.01
6*0.01
3*‐0.003
‐0.003
**0.01
1‐0.002
‐0.002
**(0.010
)(0.008
)(0.003
)(0.001
)(0.007
)(0.003
)(0.001
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,56
0,45
21,56
0,45
21,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
0R‐squared
0.00
00.01
80.04
30.14
40.01
30.04
70.14
50.01
4Notes: The
mod
el estim
ated
is a line
ar probability mod
el and
the de
pend
ent v
ariable is a du
mmy for w
hether th
e pe
rson
has com
pleted
tertiary
education. The
num
ber o
f clusters is 1
081, since some clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard
errors in
paren
theses. The
R‐squ
ared
in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
43
TABLE 5.1: Average
Wages in
Differen
t Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)5.10
6***
5.10
2***
4.73
4***
6.02
0***
4.93
9***
(0.002
)(0.002
)(0.009
)(0.089
)(0.021
)
Positive cluster g
rowth (β
2)0.01
6***
0.00
5‐0.016
***
(0.005
)(0.004
)(0.005
)
Firm
types:
Spin‐off (β
31)
‐0.006
‐0.039
***
‐0.002
‐0.000
‐0.001
(0.005
)(0.005
)(0.005
)(0.005
)(0.006
)Entrep
rene
urial start‐up (β
32)
‐0.015
***
‐0.042
***
‐0.019
***
‐0.031
***
‐0.031
***
(0.005
)(0.005
)(0.005
)(0.005
)(0.006
)Other new
firm
(β33)
0.17
9***
0.16
4***
0.17
1***
0.10
9***
0.10
6***
(0.007
)(0.009
)(0.008
)(0.007
)(0.009
)
Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.10
3***
0.05
3***
0.01
30.01
5(0.011
)(0.010
)(0.009
)(0.013
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up
0.07
0***
0.04
8***
‐0.004
‐0.001
(0.010
)(0.010
)(0.009
)(0.009
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.03
2**
0.01
2‐0.025
**‐0.023
(0.015
)(0.014
)(0.012
)(0.017
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns95
,228
95,228
95,228
95,228
95,228
R‐squared
0.00
80.01
10.13
90.27
70.03
6Notes: The
dep
ende
nt variable is the log of th
e average ho
urly wage in th
e firm. The
sample is redu
ced to aroun
d 95
,000
firm
s and
the nu
mbe
r of clusters to 10
54, because so
me firms d
o no
t have em
ployees a
nd because re
liable
wage inform
ation cann
ot be ob
tained
for a
ll firms w
ith employees. See
text fo
r more de
tails. R
obust stand
ard errors
in paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
44
TABLE 5.2: Skill‐Intensity
in Differen
t Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)0.07
1***
0.03
7***
‐0.071
***
0.15
80.02
9**
(0.001
)(0.001
)(0.002
)(0.118
)(0.012
)
Positive cluster g
rowth (β
2)0.09
8***
0.09
5***
0.00
9***
(0.002
)(0.002
)(0.002
)Firm
types:
Spin‐off (β
31)
0.00
6***
‐0.010
***
0.00
8***
0.00
7***
0.00
7***
(0.002
)(0.001
)(0.001
)(0.001
)(0.002
)Entrep
rene
urial start‐up (β
32)
0.02
2***
0.00
5***
0.01
7***
0.00
5***
0.00
6***
(0.001
)(0.001
)(0.001
)(0.001
)(0.002
)Other new
firm
(β33)
0.04
1***
0.03
8***
0.05
1***
0.02
3***
0.02
3***
(0.003
)(0.003
)(0.003
)(0.003
)(0.004
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.05
6***
0.05
4***
0.02
9***
0.02
6***
(0.005
)(0.005
)(0.004
)(0.008
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐
0.01
6***
0.01
3***
0.00
2‐0.000
(0.003
)(0.003
)(0.003
)(0.004
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.011
*‐0.007
‐0.005
‐0.007
(0.006
)(0.006
)(0.005
)(0.007
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns18
8,04
418
8,04
418
7,63
318
7,63
318
7,63
3R‐squared
0.00
20.05
00.06
10.23
60.00
4Notes: The
dep
ende
nt variable is the the share of employees w
ith a te
rtiary edu
catio
n de
gree
in th
e firm. The
nu
mbe
r of clusters is 1
081, since some clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more
details. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data.
*** p<
0.01
, ** p<
0.05
, * p<0.1.
45
TABLE 5.3: Sales per W
orker in Diffe
rent Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)13
.404
***
13.469
***
13.731
***
14.059
***
13.508
***
(0.003
)(0.003
)(0.009
)(0.079
)(0.027
)
Positive cluster g
rowth (β
2)‐0.193
***
‐0.144
***
‐0.005
(0.006
)(0.006
)(0.006
)Firm
types:
Spin‐off (β
31)
‐0.046
***
‐0.085
***
‐0.120
***
‐0.035
***
‐0.042
***
(0.006
)(0.007
)(0.007
)(0.006
)(0.010
)Entrep
rene
urial start‐up (β
32)
‐0.270
***
‐0.232
***
‐0.252
***
‐0.157
***
‐0.158
***
(0.005
)(0.006
)(0.006
)(0.006
)(0.010
)Other new
firm
(β33)
0.12
7***
0.11
0***
0.10
8***
0.03
3**
0.02
8(0.012
)(0.016
)(0.016
)(0.013
)(0.022
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.11
0***
0.10
8***
0.04
9***
0.06
4***
(0.013
)(0.013
)(0.011
)(0.019
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42‐0.036
***
‐0.026
***
0.00
20.01
0(0.010
)(0.010
)(0.009
)(0.014
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.07
7***
0.04
3*0.04
1**
0.03
9(0.024
)(0.024
)(0.020
)(0.046
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns17
8,15
217
8,15
217
7,79
917
7,79
917
7,79
9R‐squared
0.01
90.03
00.05
40.32
20.01
1Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker. The nu
mbe
r of clusters is 1
021, since some
clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squared in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
46
TABLE 5.4: Produ
ctivity
in Differen
t Types of N
ew Firm
s and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)7.06
8***
6.99
1***
6.80
1***
6.93
4***
7.09
0***
(0.022
)(0.022
)(0.024
)(0.059
)(0.093
)
Positive cluster g
rowth (β
2)0.05
1***
0.03
7***
‐0.021
***
(0.005
)(0.005
)(0.005
)Firm
types:
Spin‐off (β
31)
0.06
2***
0.02
4***
0.05
4***
0.04
7***
0.04
5***
(0.005
)(0.005
)(0.005
)(0.005
)(0.008
)Entrep
rene
urial start‐up (β
32)
‐0.021
***
‐0.047
***
‐0.019
***
‐0.020
***
‐0.020
***
(0.004
)(0.005
)(0.005
)(0.005
)(0.008
)Other new
firm
(β33)
0.05
3***
0.04
7***
0.06
5***
0.01
7*0.01
5(0.009
)(0.012
)(0.012
)(0.010
)(0.016
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.13
3***
0.11
0***
0.06
3***
0.06
6***
(0.011
)(0.011
)(0.010
)(0.025
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐
0.05
8***
0.05
2***
0.02
0**
0.02
2*(0.009
)(0.009
)(0.008
)(0.013
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.00
1‐0.006
0.00
50.00
8(0.019
)(0.019
)(0.017
)(0.027
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns13
3,08
713
3,08
713
2,84
213
2,84
213
2,84
2R‐squared
0.52
80.53
10.53
80.62
10.47
0Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker in the firm. See
text fo
r more de
tails.
Robu
st standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
nu
mbe
r of clusters is 9
75. ***
p<0.01, ** p<
0.05
, * p<0.1.
47
TABLE 5.5: W
ages in
Differen
t Types of N
ew Firm
s and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)5.33
4***
5.31
1***
4.58
4***
5.21
4***
4.68
2***
4.41
6***
5.03
3***
4.49
1***
(0.000
)(0.007
)(0.010
)(0.022
)(0.003
)(0.011
)(0.022
)(0.005
)
Positive cluster g
rowth (β
2)0.06
6***
0.02
5***
0.01
2***
0.01
2*0.00
3(0.020
)(0.010
)(0.005
)(0.007
)(0.005
)Firm
types:
Spin‐off (β
31)
‐0.155
***
‐0.176
***
‐0.097
***
‐0.051
***
‐0.051
***
‐0.003
0.00
50.00
2(0.002
)(0.009
)(0.006
)(0.005
)(0.002
)(0.005
)(0.004
)(0.002
)Entrep
rene
urial start‐up (β
32)
‐0.118
***
‐0.136
***
‐0.094
***
‐0.064
***
‐0.062
***
0.00
40.00
1‐0.000
(0.002
)(0.010
)(0.006
)(0.005
)(0.002
)(0.005
)(0.004
)(0.002
)Other new
firm
(β33)
0.00
6**
0.00
50.00
90.02
4***
0.02
3***
0.07
3***
0.06
0***
0.05
8***
(0.002
)(0.015
)(0.010
)(0.007
)(0.002
)(0.010
)(0.007
)(0.002
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.07
7***
0.01
6‐0.009
‐0.005
0.02
3**
‐0.002
0.00
1(0.024
)(0.013
)(0.009
)(0.004
)(0.010
)(0.008
)(0.004
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐
0.03
60.01
4‐0.017
‐0.012
***
0.01
0‐0.018
**‐0.015
***
(0.024
)(0.016
)(0.011
)(0.003
)(0.011
)(0.009
)(0.003
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.009
‐0.024
‐0.036
***
‐0.026
***
‐0.021
‐0.036
***
‐0.029
***
(0.032
)(0.016
)(0.013
)(0.004
)(0.016
)(0.012
)(0.004
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,17
0,48
91,17
0,48
91,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
8R‐squared
0.00
80.01
30.36
00.41
30.28
10.38
10.42
30.29
2Notes: The
dep
ende
nt variable is the log of th
e ho
urly wage. The
sample is redu
ced to aroun
d 1,17
0,00
0 individu
als a
nd th
e nu
mbe
r of clusters to 10
53,
because reliable wage inform
ation cann
ot be ob
tained
for a
ll workers. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in
columns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
48
TABLE 5.6: Skills in
Differen
t Types of N
ew Firm
s and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)0.06
9***
0.04
4***
‐0.031
***
0.15
60.03
5***
‐0.003
0.16
3*0.04
1***
(0.003
)(0.002
)(0.003
)(0.097
)(0.001
)(0.004
)(0.097
)(0.001
)
Positive cluster g
rowth (β
2)0.06
6***
0.07
4***
0.00
5**
0.07
3***
0.00
3(0.009
)(0.007
)(0.002
)(0.006
)(0.002
)Firm
types:
Spin‐off (β
31)
‐0.016
***
‐0.020
***
‐0.017
***
‐0.006
***
‐0.006
***
‐0.010
***
‐0.002
*‐0.003
***
(0.004
)(0.002
)(0.002
)(0.001
)(0.001
)(0.002
)(0.001
)(0.001
)Entrep
rene
urial start‐up (β
32)
0.00
0‐0.008
***
‐0.007
***
‐0.007
***
‐0.006
***
‐0.004
*‐0.005
***
‐0.004
***
(0.004
)(0.002
)(0.002
)(0.001
)(0.001
)(0.002
)(0.001
)(0.001
)Other new
firm
(β33)
0.02
7***
0.01
8***
0.01
8***
0.00
60.00
6***
0.02
7***
0.00
9**
0.00
9***
(0.005
)(0.005
)(0.005
)(0.004
)(0.001
)(0.005
)(0.004
)(0.001
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.02
4**
0.02
1**
0.00
20.00
5**
0.02
1***
0.00
40.00
6***
(0.010
)(0.008
)(0.004
)(0.002
)(0.008
)(0.004
)(0.002
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐
0.01
10.00
7‐0.005
‐0.005
***
0.00
4‐0.005
‐0.005
***
(0.010
)(0.008
)(0.004
)(0.002
)(0.007
)(0.003
)(0.002
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.01
00.00
7‐0.007
‐0.009
***
0.00
6‐0.006
‐0.008
***
(0.013
)(0.011
)(0.006
)(0.002
)(0.011
)(0.006
)(0.002
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,56
0,45
21,56
0,45
21,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
0R‐squared
0.00
10.01
80.04
30.14
40.01
40.04
80.14
50.01
4Notes: The
mod
el estim
ated
is a line
ar probability mod
el and
the de
pend
ent v
ariable is a du
mmy for w
hether th
e pe
rson
has com
pleted
tertiary edu
catio
n.
The nu
mbe
r of clusters is 1
081, since some clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
parenthe
ses. The
R‐squ
ared
in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
49
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)5.29
0***
4.86
0***
4.58
6***
5.21
7***
4.89
4***
4.40
9***
5.03
1***
4.69
6***
(0.005
)(0.005
)(0.010
)(0.023
)(0.003
)(0.012
)(0.022
)(0.004
)
Positive cluster g
rowth, existing clusters (β
21)
0.12
4***
0.13
2***
0.01
8**
0.01
1***
0.00
5(0.013
)(0.009
)(0.007
)(0.004
)(0.006
)Po
sitive cluster g
rowth, new
clusters (β 2
2)0.09
20.06
5‐0.007
‐0.027
0.00
1(0.063
)(0.058
)(0.051
)(0.051
)(0.050
)
New
firm
(β3)
‐0.099
***
‐0.070
***
‐0.068
***
‐0.038
***
‐0.039
***
0.02
1***
0.01
6***
0.01
3***
(0.006
)(0.006
)(0.005
)(0.003
)(0.001
)(0.003
)(0.003
)(0.001
)
Firm
type
x Positive cluster g
rowth, existing clu
sters:
Positive cluster g
rowth, existing clusters x New
firm
(β4A)
‐0.013
‐0.016
0.00
2‐0.014
**‐0.004
*0.00
6‐0.010
**‐0.005
**(0.015
)(0.011
)(0.008
)(0.007
)(0.002
)(0.007
)(0.005
)(0.002
)
Firm
type
x Positive cluster g
rowth, new
clusters:
Positive cluster g
rowth, new
clusters x
New
firm
(β4B)
0.30
0***
0.23
5***
0.21
5**
‐0.034
‐0.030
0.11
1*‐0.098
***
‐0.041
(0.096
)(0.086
)(0.089
)(0.055
)(0.042
)(0.064
)(0.018
)(0.042
)Co
ntrol for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,17
0,48
91,17
0,48
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
8R‐squared
0.02
20.23
80.35
90.41
30.16
30.38
10.42
30.17
6
TABLE 8.1: W
ages in
Differen
t Types of N
ew Firm
s and
Old Firm
s, W
orker Level, 201
0
Notes: The
dep
ende
nt variable is the log of th
e ho
urly wage. The
sample is redu
ced to aroun
d 1,17
0,00
0 individu
als a
nd th
e nu
mbe
r of clusters to 57
60, because re
liable wage
inform
ation cann
ot be ob
tained
for a
ll workers. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
50
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)5.29
0***
4.86
1***
4.58
7***
5.21
8***
4.89
4***
4.41
4***
5.03
1***
4.70
0***
(0.005
)(0.005
)(0.010
)(0.023
)(0.003
)(0.012
)(0.022
)(0.004
)
Positive cluster g
rowth, existing clusters (β
21)
0.12
4***
0.13
2***
0.01
8**
0.01
1***
0.00
50.00
7*(0.013
)(0.009
)(0.007
)(0.004
)(0.006
)(0.004
)Po
sitive cluster g
rowth, new
clusters (β 2
2)0.09
20.06
5‐0.007
‐0.028
0.00
2‐0.027
(0.063
)(0.058
)(0.051
)(0.051
)(0.050
)(0.051
)
Spin‐off (β
31)
‐0.143
***‐0.102
***‐0.091
***‐0.048
***‐0.050
***
0.00
10.00
7*0.00
2(0.008
)(0.007
)(0.006
)(0.004
)(0.002
)(0.004
)(0.004
)(0.002
)Entrep
rene
urial start‐up (β
32)
‐0.117
***‐0.093
***‐0.090
***‐0.064
***‐0.062
***
0.00
8*0.00
1‐0.001
(0.008
)(0.007
)(0.006
)(0.005
)(0.002
)(0.005
)(0.004
)(0.002
)Other new
firm
(β33)
0.00
30.01
9**
0.00
40.01
8***
0.01
5***
0.06
9***
0.05
5***
0.04
9***
(0.009
)(0.009
)(0.007
)(0.005
)(0.002
)(0.008
)(0.006
)(0.002
)
Firm
type
x Positive cluster g
rowth, existing clu
sters:
Positive cluster g
rowth, existing clusters x Spin‐off (β 4
1A)
‐0.021
‐0.024
*‐0.003
‐0.018
**‐0.009
**0.01
3‐0.009
‐0.001
(0.017
)(0.013
)(0.011
)(0.009
)(0.004
)(0.009
)(0.008
)(0.004
)Po
sitive cluster g
rowth, existing clusters x Entrepren
euria
l start‐up (β
42A
‐0.022
‐0.024
*0.00
4‐0.017
*‐0.010
***
0.00
2‐0.020
**‐0.013
***
(0.018
)(0.013
)(0.012
)(0.010
)(0.003
)(0.009
)(0.008
)(0.003
)Po
sitive cluster g
rowth, existing clusters x Other new
firm
(β43
A)‐0.032
‐0.029
*‐0.020
*‐0.020
*‐0.006
‐0.014
‐0.022
**‐0.009
**(0.025
)(0.017
)(0.012
)(0.011
)(0.004
)(0.014
)(0.010
)(0.004
)
Firm
type
x Positive cluster g
rowth, new
clusters:
Positive cluster g
rowth, new
clusters x
Spin‐off (β 4
1B)
‐0.160
‐0.111
‐0.070
‐0.128
‐0.046
‐0.068
‐0.123
‐0.073
(0.102
)(0.090
)(0.099
)(0.093
)(0.092
)(0.100
)(0.093
)(0.091
)Po
sitive cluster g
rowth, new
clusters x
Entrepren
euria
l start‐up (β
42B)
0.07
90.03
7‐0.056
‐0.076
0.05
3‐0.051
‐0.076
0.03
9(0.100
)(0.093
)(0.075
)(0.077
)(0.059
)(0.077
)(0.078
)(0.059
)Po
sitive cluster g
rowth, new
clusters x
Other new
firm
(β43
B)0.28
1***
0.21
8***
0.22
4***
‐0.056
‐0.142
**0.11
2**
‐0.096
*‐0.138
**(0.067
)(0.062
)(0.056
)(0.054
)(0.064
)(0.052
)(0.054
)(0.064
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,17
0,48
91,17
0,48
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
8R‐squared
0.02
40.24
00.36
00.41
30.16
40.38
10.42
30.17
6Notes: The
dep
ende
nt variable is the log of th
e ho
urly wage. The
sample is redu
ced to aroun
d 1,17
0,00
0 individu
als a
nd th
e nu
mbe
r of clusters to 57
60,
because reliable wage inform
ation cann
ot be ob
tained
for a
ll workers. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in
columns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
TABLE 8.2: W
ages in
Differen
t Types of N
ew Firm
s and
Old Firm
s, W
orker Level, 2010
30
AppendixA.EstimationResultsfortheManufacturingSector
52
TABLE A2
.1.: Num
ber o
f clusters a
nd jo
bs. B
y cluster g
rowth
Num
ber o
f clusters
Num
ber o
f job
sNum
ber o
f firm
s
Cluster g
rowth
#%
#%
#%
‐0.50‐
7517
.61
14,221
4.87
1,32
49.79
‐0.50;‐0.25
116
27.23
96,439
33.03
5,21
638
.55
‐0.25;‐0.1
8018
.78
73,424
25.14
3,39
125
.06
‐0.1‐0.1
7317
.14
66,014
22.61
2,39
717
.72
0.1‐0.25
184.23
8,06
52.76
452
3.34
0.25‐0.5
92.11
20,186
6.91
230
1.70
0.5+
4310
.09
12,885
4.41
498
3.68
New
122.82
780
0.27
220.16
Totals
426
100.00
292,01
410
0.00
13,530
100.00
53
TABLE A2
.2: N
umbe
r of firm
type
s and
jobs. B
y cluster g
rowth and
firm
type
sNum
ber o
f firm
s
Establish
ed firm
sSpin‐offs
Other new
firm
sCluster g
rowth
#%
#%
#%
#%
‐0.50‐
876
9.5
667.9
312
11.2
7011
.0 ‐0
.50;‐0.25
3611
39.0
309
36.9
1068
38.2
228
35.8
‐0.25;‐0.1
2450
26.5
210
25.1
585
20.9
146
23.0
‐0.1‐0.1
1607
17.4
166
19.8
506
18.1
118
18.6
0.1‐0.25
280
3.0
273.2
127
4.5
182.8
0.25‐0.5
146
1.6
202.4
421.5
223.5
0.5+
290
3.1
394.7
157
5.6
345.3
Totals
9260
100.0
837
100.0
2797
100.0
63610
0.0
Num
ber o
f job
s
Establish
ed firm
sSpin‐offs
Other new
firm
sCluster g
rowth
#%
#%
#%
#%
‐0.50‐
12,784
4.7
297
6.3
724
7.5
416
7.3
‐0.50;‐0.25
90,159
33.1
1238
26.4
3142
32.5
1,90
033
.5 ‐0
.25;‐0.1
68,084
25.0
1537
32.8
2248
23.2
1,55
527
.4 ‐0
.1‐0.1
62,411
22.9
894
19.1
1736
17.9
973
17.2
0.1‐0.25
7,24
82.7
330
7.0
388
4.0
991.7
0.25‐0.5
18,897
6.9
234
5.0
914
9.4
141
2.5
0.5+
12,407
4.6
151
3.2
526
5.4
581
10.3
Totals
271,99
010
0.0
4681
100.0
9678
100.0
5,66
510
0.0
Entrep
rene
urial
Entrep
rene
urial
54
TABLE A2
.3: A
verage
num
ber o
f job
s per firm
. By cluster g
rowth after firm
type
sJobs per firm
Establish
ed firm
sSpin‐offs
Entrep
rene
urial
start‐up
sOther
new firm
sCluster g
rowth
##
##
‐0.50‐
14.6
4.5
2.3
5.9
‐0.50;‐0.25
25.0
4.0
2.9
8.3
‐0.25;‐0.1
27.8
7.3
3.8
10.7
‐0.1‐0.1
38.8
5.4
3.4
8.2
0.1‐0.25
25.9
12.2
3.1
5.5
0.25‐0.5
129.4
11.7
21.8
6.4
0.5+
42.8
3.9
3.4
17.1
Totals
29.4
5.6
3.5
8.9
55
TABLE A2
.4: A
verage
individu
al cha
racteristics. By firm ty
pe and
cluster growth
Cluster g
rowth
Firm
type
Num
ber o
f workers
Num
ber o
f firms
Mean age
Share of
females
Share of
highly
educated
Mean ho
urly
wage (DKK
)
Negative
Establish
ed firm
s19
7,58
87,90
242
.50.28
0.04
215.82
Negative
Spin‐offs
3,49
667
738
.60.28
0.01
172.66
Negative
Enterprene
urial sta
7,08
52,26
339
.30.23
0.03
193.79
Negative
Other new
firm
s4,24
151
340
.10.27
0.03
209.93
Positive
Establish
ed firm
s74
,402
1,35
841
.80.35
0.12
243.42
Positive
Spin‐offs
1,18
516
039
.30.19
0.06
237.42
Positive
Enterprene
urial sta
2,59
353
438
.60.22
0.06
224.29
Positive
Other new
firm
s1,42
412
340
.80.25
0.09
268.40
56
TABLE A4
.1: A
verage
Wages in
New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)5.15
0***
5.14
8***
4.81
0***
5.06
7***
5.09
4***
(0.005
)(0.005
)(0.034
)(0.055
)(0.053
)
Positive cluster g
rowth (β
2)0.01
60.00
6‐0.020
(0.015
)(0.014
)(0.015
)
New
firm
(β3)
0.04
8***
0.04
5***
0.04
3***
0.02
6***
0.02
6**
(0.010
)(0.011
)(0.010
)(0.010
)(0.011
)
Positive cluster g
rowth x New
firm
(β4)
0.01
40.01
10.00
80.00
8(0.026
)(0.024
)(0.023
)(0.024
)Co
ntrols for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns9,33
79,33
79,33
79,33
79,33
7R‐squared
0.00
30.00
30.13
20.28
60.02
3Notes: The
dep
ende
nt variable is the log of th
e average ho
urly wage in th
e firm. The
sample is redu
ced to
arou
nd 9,000
firm
s and
the nu
mbe
r of clusters to 42
0, because so
me firms d
o no
t have em
ployees a
nd because
reliable wage inform
ation cann
ot be ob
tained
for a
ll firms w
ith employees. See
text fo
r more de
tails. R
obust
standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, **
p<0.05
, * p<0.1.
57
TABLE A4
.2: Skill‐Intensity
in New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)0.02
1***
0.02
0***
‐0.004
‐0.004
0.01
8***
(0.001
)(0.001
)(0.005
)(0.006
)(0.006
)
Positive cluster g
rowth (β
2)0.01
0***
0.01
1***
0.00
3(0.003
)(0.003
)(0.003
)
New
firm
(β3)
0.00
8***
0.00
7***
0.01
0***
0.00
8***
0.00
8***
(0.002
)(0.002
)(0.003
)(0.003
)(0.003
)
Positive cluster g
rowth x New
firm
(β4)
0.00
20.00
3‐0.000
0.00
1(0.007
)(0.007
)(0.007
)(0.006
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns13
,530
13,530
13,516
13,516
13,516
R‐squared
0.00
10.00
20.00
60.08
70.00
1Notes: The
dep
ende
nt variable is the the share of employees w
ith a te
rtiary edu
catio
n de
gree
in th
e firm.
The nu
mbe
r of clusters is 4
20, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
58
TABLE A4
.3: Sales per W
orker in New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)13
.624
***
13.613
***
13.690
***
14.382
***
13.878
***
(0.008
)(0.008
)(0.036
)(0.097
)(0.057
)
Positive cluster g
rowth (β
2)0.07
4***
0.06
2***
0.01
4(0.023
)(0.023
)(0.024
)
New
firm
(β3)
‐0.103
***
‐0.092
***
‐0.113
***
‐0.128
***
‐0.098
***
(0.014
)(0.015
)(0.016
)(0.017
)(0.019
)
Positive cluster g
rowth x New
firm
(β4)
‐0.078
*‐0.080
**‐0.066
‐0.064
(0.040
)(0.040
)(0.042
)(0.048
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns13
,364
13,364
13,352
13,516
13,352
R‐squared
0.00
40.00
50.01
90.18
20.00
9Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker. The nu
mbe
r of clusters is 4
20, since
some clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
59
TABLE A4
.4: P
rodu
ctivity
in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)6.08
6***
6.08
6***
6.05
7***
6.54
1***
6.54
5***
(0.086
)(0.086
)(0.089
)(0.113
)(0.149
)
Positive cluster g
rowth (β
2)‐0.001
‐0.003
‐0.014
(0.014
)(0.014
)(0.015
)
New
firm
(β3)
0.05
8***
0.05
9***
0.07
3***
0.04
6***
0.04
5***
(0.011
)(0.012
)(0.012
)(0.012
)(0.015
)
Positive cluster g
rowth x New
firm
(β4)
‐0.003
‐0.002
0.00
90.00
2(0.028
)(0.028
)(0.027
)(0.030
)
log(Interm
ediate inpu
ts per worker)
0.48
3***
0.48
3***
0.47
7***
0.45
1***
0.45
0***
(0.008
)(0.008
)(0.008
)(0.008
)(0.011
)log(capital per worker)
0.10
2***
0.10
2***
0.09
9***
0.10
1***
0.10
1***
(0.004
)(0.004
)(0.004
)(0.004
)(0.006
)log(nu
mbe
r of w
orkers)
0.03
6***
0.03
6***
0.04
7***
0.04
5***
0.04
4***
(0.003
)(0.003
)(0.003
)(0.003
)(0.005
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns12
,158
12,158
12,150
12,150
12,150
R‐squared
0.60
90.60
90.61
30.64
50.56
2
Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker in the firm. See
text fo
r more de
tails.
Robu
st standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
nu
mbe
r of clusters is 4
20. ***
p<0.01, ** p<
0.05
, * p<0.1.
60
TABLE A4
.4A: In
term
ediate In
puts per W
orker in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)12
.920
***12
.906
***13
.108
***14
.086
***13
.245
***
(0.011
)(0.012
)(0.050
)(0.111
)(0.077
)
Positive cluster g
rowth (β
2)0.10
0***
0.08
6***
0.03
0(0.030
)(0.030
)(0.030
)
New
firm
(β3)
‐0.203
***‐0.182
***‐0.214
***‐0.165
***‐0.158
***
(0.021
)(0.023
)(0.024
)(0.023
)(0.032
)
Positive cluster g
rowth x New
firm
(β4)
‐0.138
**‐0.148
**‐0.118
**‐0.114
(0.058
)(0.058
)(0.056
)(0.069
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns12
,823
12,823
12,812
12,812
12,812
R‐squared
0.00
80.00
90.01
70.15
50.00
9Notes: The
dep
ende
nt variable is the log of interm
ediate inpu
ts per worker in the firm. See
text
for m
ore de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted
on th
e de
‐meane
d data. The
num
ber o
f clusters is 4
23. ***
p<0.01, ** p<
0.05
, * p<0.1.
61
TABLE A4
.4B: Cap
ital per W
orker in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)12
.209
***12
.172
***12
.125
***12
.724
***12
.182
***
(0.015
)(0.016
)(0.063
)(0.128
)(0.085
)
Positive cluster g
rowth (β
2)0.24
9***
0.23
6***
0.07
5*(0.042
)(0.042
)(0.043
)
New
firm
(β3)
‐0.359
***‐0.354
***‐0.369
***‐0.323
***‐0.313
***
(0.026
)(0.029
)(0.029
)(0.029
)(0.042
)
Positive cluster g
rowth x New
firm
(β4)
‐0.088
‐0.093
‐0.114
*‐0.154
**(0.070
)(0.070
)(0.068
)(0.072
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns12
,664
12,664
12,655
12,655
12,655
R‐squared
0.01
40.01
80.02
70.11
60.01
4Notes: The
dep
ende
nt variable is the log of capita
lper worker in the firm. See
text fo
r more
details. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the
de‐m
eane
d data. The
num
ber o
f clusters is 4
21. ***
p<0.01, ** p<
0.05
, * p<0.1.
62
TABLE A4
.5: W
ages in
New
and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)5.36
2***
5.33
5***
4.74
5***
4.81
9***
4.79
0***
4.54
8***
4.57
6***
4.57
7***
(0.001
)(0.007
)(0.020
)(0.017
)(0.008
)(0.026
)(0.028
)(0.013
)
Positive cluster g
rowth (β
2)0.09
9***
0.04
3**
0.00
40.00
4‐0.001
(0.031
)(0.018
)(0.014
)(0.013
)(0.013
)
New
firm
(β3)
‐0.051
***
‐0.070
***
‐0.047
***
‐0.026
***
‐0.022
***
0.03
9***
0.03
8***
0.03
7***
(0.003
)(0.013
)(0.010
)(0.009
)(0.003
)(0.008
)(0.007
)(0.003
)
Positive cluster g
rowth x New
firm
(β4)
0.06
60.06
6*0.04
50.04
5***
0.06
0**
0.02
40.02
5***
(0.051
)(0.036
)(0.029
)(0.006
)(0.025
)(0.023
)(0.006
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns26
4,28
226
4,28
225
7,54
725
7,54
725
7,54
725
7,54
725
7,54
725
7,54
7R‐squared
0.00
10.01
60.34
80.37
70.31
50.36
90.38
60.32
5Notes: The
dep
ende
nt variable is the log of th
e ho
urly wage. The
sample is redu
ced to aroun
d 26
4,28
2 individu
als a
nd th
e nu
mbe
r of clusters to 41
2,
because reliable wage inform
ation cann
ot be ob
tained
for a
ll workers. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in
columns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
63
TABLE A4
.6: Skills in
New
and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)0.06
1***
0.03
7***
‐0.018
**‐0.031
***
0.02
4***
‐0.024
***
‐0.052
***
0.01
0**
(0.010
)(0.003
)(0.007
)(0.006
)(0.001
)(0.005
)(0.008
)(0.004
)
Positive cluster g
rowth (β
2)0.08
7***
0.08
2***
0.00
30.03
5***
‐0.002
(0.028
)(0.026
)(0.005
)(0.009
)(0.004
)
New
firm
(β3)
‐0.023
**‐0.009
**‐0.009
**‐0.000
0.00
20.01
0***
0.00
7**
0.00
7***
(0.011
)(0.004
)(0.004
)(0.003
)(0.002
)(0.004
)(0.003
)(0.002
)
Positive cluster g
rowth x New
firm
(β4)
‐0.046
‐0.043
‐0.007
‐0.011
***
‐0.006
‐0.001
‐0.003
(0.031
)(0.030
)(0.010
)(0.004
)(0.018
)(0.010
)(0.004
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns29
2,01
429
2,01
428
6,18
128
6,18
128
6,18
128
6,18
128
6,18
128
6,18
1R‐squared
0.00
10.02
60.04
30.11
10.01
30.07
70.11
40.01
6
Notes: The
mod
el estim
ated
is a line
ar probability mod
el and
the de
pend
ent v
ariable is a du
mmy for w
hether th
e pe
rson
has com
pleted
tertiary
education. The
num
ber o
f clusters is 4
26, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard
errors in
paren
theses. The
R‐squ
ared
in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
64
TABLE A5
.1: A
verage
Wages in
Differen
t Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)5.15
0***
5.14
8***
4.81
6***
5.06
9***
5.09
7***
(0.005
)(0.005
)(0.034
)(0.055
)(0.053
)
Positive cluster g
rowth (β
2)0.01
60.00
6‐0.020
(0.015
)(0.014
)(0.015
)
Firm
types:
Spin‐off (β
31)
‐0.054
***
‐0.062
***
‐0.040
**‐0.025
‐0.025
(0.018
)(0.020
)(0.018
)(0.017
)(0.018
)Entrep
rene
urial start‐up (β
32)
0.03
4***
0.03
5***
0.03
2**
0.00
60.00
8(0.012
)(0.013
)(0.013
)(0.012
)(0.012
)Other new
firm
(β33)
0.19
4***
0.18
3***
0.15
9***
0.12
8***
0.12
9***
(0.019
)(0.022
)(0.021
)(0.019
)(0.021
)
Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.04
10.02
70.01
80.01
1(0.046
)(0.047
)(0.041
)(0.043
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up
‐0.010
‐0.009
‐0.007
‐0.005
(0.032
)(0.029
)(0.028
)(0.027
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.05
50.05
20.04
50.04
4(0.047
)(0.044
)(0.040
)(0.036
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns9,33
79,33
79,33
79,33
79,33
7R‐squared
0.01
40.01
40.14
00.29
10.03
0Notes: The
dep
ende
nt variable is the log of th
e average ho
urly wage in th
e firm. The
sample is redu
ced to aroun
d 9,00
0 firms a
nd th
e nu
mbe
r of clusters to 41
3, because so
me firms d
o no
t have em
ployees a
nd because re
liable wage
inform
ation cann
ot be ob
tained
for a
ll firms w
ith employees. See
text fo
r more de
tails. R
obust stand
ard errors in
parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
65
TABLE A5
.2: Skill‐Intensity
in Differen
t Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)0.02
1***
0.02
0***
‐0.003
‐0.004
0.01
8***
(0.001
)(0.001
)(0.005
)(0.006
)(0.006
)
Positive cluster g
rowth (β
2)0.01
0***
0.01
1***
0.00
3(0.003
)(0.003
)(0.003
)Firm
types:
Spin‐off (β
31)
‐0.006
*‐0.005
‐0.002
‐0.000
0.00
0(0.003
)(0.004
)(0.004
)(0.004
)(0.004
)Entrep
rene
urial start‐up (β
32)
0.00
6**
0.00
5*0.00
9***
0.00
7**
0.00
8**
(0.003
)(0.003
)(0.003
)(0.003
)(0.004
)Other new
firm
(β33)
0.03
4***
0.02
8***
0.03
0***
0.02
0***
0.02
2***
(0.007
)(0.007
)(0.007
)(0.007
)(0.008
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)‐0.008
‐0.007
‐0.005
‐0.007
(0.009
)(0.009
)(0.009
)(0.009
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐
‐0.001
0.00
0‐0.001
‐0.001
(0.008
)(0.008
)(0.008
)(0.008
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.02
50.02
50.01
10.01
4(0.022
)(0.023
)(0.021
)(0.020
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns13
,530
13,530
13,516
13,516
13,516
R‐squared
0.00
40.00
60.00
90.08
80.00
3Notes: The
dep
ende
nt variable is the the share of employees w
ith a te
rtiary edu
catio
n de
gree
in th
e firm. The
nu
mbe
r of clusters is 4
26, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails.
Robu
st standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<
0.01
, ** p<
0.05
, * p<0.1.
66
TABLE A5
.3: Sales per W
orker in Diffe
rent Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)13
.624
***
13.613
***
13.697
***
14.344
***
13.882
***
(0.008
)(0.008
)(0.036
)(0.090
)(0.056
)
Positive cluster g
rowth (β
2)0.07
4***
0.06
2***
0.00
6(0.023
)(0.023
)(0.022
)Firm
types:
Spin‐off (β
31)
‐0.120
***
‐0.119
***
‐0.131
***
‐0.088
***
‐0.084
***
(0.026
)(0.028
)(0.028
)(0.026
)(0.029
)Entrep
rene
urial start‐up (β
32)
‐0.166
***
‐0.152
***
‐0.178
***
‐0.164
***
‐0.156
***
(0.017
)(0.018
)(0.019
)(0.018
)(0.022
)Other new
firm
(β33)
0.19
5***
0.20
8***
0.18
7***
0.11
9***
0.12
4***
(0.035
)(0.037
)(0.037
)(0.035
)(0.039
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)‐0.023
‐0.035
‐0.010
‐0.033
(0.073
)(0.073
)(0.071
)(0.092
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐
‐0.092
**‐0.093
**‐0.055
‐0.058
(0.045
)(0.045
)(0.044
)(0.055
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.083
‐0.080
‐0.102
‐0.136
*(0.099
)(0.100
)(0.090
)(0.082
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns13
,364
13,364
13,352
13,352
13,352
R‐squared
0.01
30.01
40.02
80.19
30.01
5Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker. The nu
mbe
r of clusters is 4
24, since so
me
clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
67
TABLE A5
.4: P
rodu
ctivity
in Differen
t Types of N
ew Firm
s and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)6.09
3***
6.09
2***
6.06
4***
6.55
0***
6.55
1***
(0.087
)(0.087
)(0.089
)(0.113
)(0.150
)
Positive cluster g
rowth (β
2)‐0.001
‐0.003
‐0.014
(0.014
)(0.014
)(0.015
)Firm
types:
Spin‐off (β
31)
0.02
60.02
00.03
8**
0.04
3**
0.04
0**
(0.018
)(0.020
)(0.020
)(0.019
)(0.020
)Entrep
rene
urial start‐up (β
32)
0.05
8***
0.05
8***
0.07
2***
0.03
4**
0.03
4**
(0.013
)(0.014
)(0.015
)(0.014
)(0.017
)Other new
firm
(β33)
0.09
9***
0.11
5***
0.12
3***
0.09
5***
0.09
6***
(0.029
)(0.032
)(0.032
)(0.031
)(0.030
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.03
00.02
70.02
20.01
6(0.048
)(0.048
)(0.048
)(0.073
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐
0.00
30.00
30.01
90.02
0(0.033
)(0.033
)(0.033
)(0.035
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.078
‐0.070
‐0.055
‐0.093
(0.074
)(0.074
)(0.069
)(0.065
)
log(Interm
ediate inpu
ts per worker)
0.48
2***
0.48
2***
0.47
6***
0.45
1***
0.45
0***
(0.008
)(0.008
)(0.008
)(0.008
)(0.011
)log(capital per worker)
0.10
2***
0.10
2***
0.09
9***
0.10
1***
0.10
0***
(0.004
)(0.004
)(0.004
)(0.004
)(0.006
)log(nu
mbe
r of w
orkers)
0.03
6***
0.03
6***
0.04
7***
0.04
5***
0.04
4***
(0.003
)(0.003
)(0.003
)(0.003
)(0.005
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns12
,158
12,158
12,150
12,150
12,150
R‐squared
0.60
90.60
90.61
30.64
50.56
2Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker in the firm. See
text fo
r more de
tails.
Robu
st standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
nu
mbe
r of clusters is 4
20. ***
p<0.01, ** p<
0.05
, * p<0.1.
68
TABLE A5
.5: W
ages in
Differen
t Types of N
ew Firm
s and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)5.36
2***
5.33
5***
4.74
6***
4.81
9***
4.79
1***
4.54
9***
4.57
6***
4.57
7***
(0.001
)(0.007
)(0.020
)(0.017
)(0.008
)(0.026
)(0.028
)(0.013
)
Positive cluster g
rowth (β
2)0.09
9***
0.04
3**
0.00
40.00
4‐0.001
(0.031
)(0.018
)(0.014
)(0.013
)(0.013
)Firm
types:
Spin‐off (β
31)
‐0.094
***
‐0.131
***
‐0.080
***
‐0.058
***
‐0.058
***
0.01
20.00
90.00
5(0.006
)(0.015
)(0.012
)(0.011
)(0.006
)(0.011
)(0.011
)(0.006
)Entrep
rene
urial start‐up (β
32)
‐0.072
***
‐0.075
***
‐0.058
***
‐0.037
***
‐0.031
***
0.04
3***
0.03
9***
0.04
0***
(0.005
)(0.018
)(0.012
)(0.011
)(0.005
)(0.010
)(0.010
)(0.005
)Other new
firm
(β33)
0.00
9‐0.020
‐0.009
0.01
20.01
5***
0.05
1***
0.05
5***
0.05
5***
(0.005
)(0.023
)(0.019
)(0.017
)(0.005
)(0.016
)(0.013
)(0.005
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.13
3**
0.12
7**
0.12
8**
0.11
5***
0.11
8**
0.09
00.08
1***
(0.066
)(0.059
)(0.060
)(0.012
)(0.049
)(0.055
)(0.012
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐
0.00
40.02
50.03
60.02
6***
0.00
70.00
8‐0.005
(0.067
)(0.054
)(0.044
)(0.008
)(0.030
)(0.029
)(0.008
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.11
80.09
1‐0.000
0.03
5***
0.10
0**
‐0.001
0.04
5***
(0.078
)(0.057
)(0.038
)(0.012
)(0.039
)(0.037
)(0.012
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns26
4,28
226
4,28
225
7,54
725
7,54
725
7,54
725
7,54
725
7,54
725
7,54
7R‐squared
0.00
20.01
70.34
90.37
70.31
60.36
90.38
70.32
5Notes: The
dep
ende
nt variable is the log of th
e ho
urly wage. The
sample is redu
ced to aroun
d 26
0,00
0 individu
als a
nd th
e nu
mbe
r of clusters to 41
2,
because reliable wage inform
ation cann
ot be ob
tained
for a
ll workers. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in
columns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
69
TABLE A5
.6: Skills in
Differen
t Types of N
ew Firm
s and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)0.06
1***
0.03
7***
‐0.018
**‐0.030
***
0.02
4***
‐0.024
***
‐0.052
***
0.00
9**
(0.010
)(0.003
)(0.007
)(0.006
)(0.001
)(0.006
)(0.008
)(0.004
)
Positive cluster g
rowth (β
2)0.08
7***
0.08
2***
0.00
30.03
5***
‐0.002
(0.028
)(0.026
)(0.005
)(0.009
)(0.004
)Firm
types:
Spin‐off (β
31)
‐0.037
**‐0.025
***
‐0.022
***
‐0.008
***
‐0.007
*‐0.001
‐0.001
‐0.001
(0.015
)(0.004
)(0.004
)(0.003
)(0.004
)(0.003
)(0.003
)(0.004
)Entrep
rene
urial start‐up (β
32)
‐0.021
*‐0.005
‐0.006
0.00
20.00
5*0.01
4**
0.00
9**
0.01
1***
(0.012
)(0.006
)(0.006
)(0.004
)(0.003
)(0.006
)(0.004
)(0.003
)Other new
firm
(β33)
‐0.013
‐0.003
‐0.005
0.00
20.00
30.01
2*0.00
9*0.00
9**
(0.013
)(0.006
)(0.006
)(0.005
)(0.004
)(0.006
)(0.005
)(0.004
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)‐0.041
‐0.039
‐0.014
‐0.018
**0.00
2‐0.008
‐0.012
(0.048
)(0.046
)(0.010
)(0.008
)(0.038
)(0.010
)(0.008
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐
‐0.056
*‐0.053
*‐0.004
‐0.008
‐0.019
0.00
4‐0.000
(0.033
)(0.032
)(0.016
)(0.006
)(0.025
)(0.016
)(0.006
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.032
‐0.030
‐0.007
‐0.008
0.01
1‐0.002
0.00
0(0.036
)(0.035
)(0.018
)(0.008
)(0.027
)(0.019
)(0.008
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns29
2,01
429
2,01
428
6,18
128
6,18
128
6,18
128
6,18
128
6,18
128
6,18
1R‐squared
0.00
10.02
60.04
30.11
10.01
30.07
70.11
40.01
6Notes: The
mod
el estim
ated
is a line
ar probability mod
el and
the de
pend
ent v
ariable is a du
mmy for w
hether th
e pe
rson
has com
pleted
tertiary edu
catio
n.
The nu
mbe
r of clusters is 4
26, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
parenthe
ses. The
R‐squ
ared
in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
31
AppendixB.EstimationResultsfortheServiceSector
72
TABLE B2
.1.: Num
ber o
f clusters a
nd jo
bs. B
y cluster g
rowth
Num
ber o
f clusters
Num
ber o
f job
sNum
ber o
f firm
s
Cluster g
rowth
#%
#%
#%
‐0.50‐
336.92
21,106
2.10
2,44
42.04
‐0.50;‐0.25
6112
.79
88,797
8.84
7,97
86.67
‐0.25;‐0.1
7515
.72
195,25
919
.44
24,473
20.47
‐0.1‐0.1
142
29.77
493,40
149
.11
58,251
48.73
0.1‐0.25
5210
.90
95,815
9.54
14,714
12.31
0.25‐0.5
357.34
75,646
7.53
7,62
36.38
0.5+
7114
.88
34,613
3.45
4,03
93.38
New
81.68
400.00
120.01
Totals
477
100.00
1,00
4,67
710
0.00
119,53
410
0.00
73
TABLE B2
.2: N
umbe
r of firm
type
s and
jobs. B
y cluster g
rowth and
firm
type
sNum
ber o
f firm
s
Establish
ed firm
sSpin‐offs
Other new
firm
sCluster g
rowth
#%
#%
#%
#%
‐0.50‐
1,22
71.8
346
2.4
688
2.3
183
2.7
‐0.50;‐0.25
4,81
87.0
982
6.8
1,71
7
5.8
461
6.9
‐0.25;‐0.1
15,039
21.8
3,38
3
23.3
4,96
4
16.9
1,08
7
16.3
‐0.1‐0.1
34,243
49.7
6,55
8
45.1
14,268
48.5
3,18
2
47.7
0.1‐0.25
8,11
911
.81,85
9
12.8
3,71
8
12.6
1,01
8
15.3
0.25‐0.5
3,69
35.4
1,14
9
7.9
2,28
3
7.8
498
7.5
0.5+
1,74
12.5
270
1.9
1,80
2
6.1
238
3.6
Totals
68,880
100
14,547
100
29,440
100
6,66
7
100
Num
ber o
f job
s
Establish
ed firm
sSpin‐offs
Other new
firm
sCluster g
rowth
#%
#%
#%
#%
‐0.50‐
16,102
1.9
1,69
13.1
2,23
02.9
1,08
32.7
‐0.50;‐0.25
76,432
9.2
4,16
47.6
5,31
87.0
2,88
37.1
‐0.25;‐0.1
165,70
319
.911
,375
20.7
12,279
16.1
5,90
214
.6 ‐0
.1‐0.1
412,58
049
.526
,809
48.7
35,895
46.9
18,117
44.9
0.1‐0.25
73,397
8.8
6,52
311
.99,34
312
.26,55
216
.2 0.25‐0.5
62,375
7.5
3,71
56.8
5,53
17.2
4,02
510
.0 0.5+
26,244
3.2
721
1.3
5,87
47.7
1,81
44.5
Totals
832,83
310
0.0
54,998
100.0
76,470
100.0
40,376
100.0
Entrep
rene
urial start‐
Entrep
rene
urial start‐
74
TABLE B2
.3: A
verage
num
ber o
f job
s per firm
. By cluster g
rowth after firm
type
sJobs per firm
Establish
ed firm
sSpin‐offs
Entrep
rene
urial
start‐up
sOther new
firms
Cluster g
rowth
##
##
‐0.50‐
13.1
4.9
3.2
5.9
‐0.50;‐0.25
15.9
4.2
3.1
6.3
‐0.25;‐0.1
11.0
3.4
2.5
5.4
‐0.1‐0.1
12.0
4.1
2.5
5.7
0.1‐0.25
9.0
3.5
2.5
6.4
0.25‐0.5
16.9
3.2
2.4
8.1
0.5+
15.1
2.7
3.3
7.6
Totals
12.1
3.8
2.6
6.1
75
TABLE B2
.4: A
verage
individu
al cha
racteristics. By firm ty
pe and
cluster growth
Cluster g
rowth
Firm
type
Num
ber o
f workers
Num
ber o
f firms
Mean age
Share of
females
Share of
highly
educated
Mean ho
urly
wage (DKK
)
Negative
Establish
ed firm
s46
8,37
242
,087
39.8
0.38
0.05
214.13
Negative
Spin‐offs
32,528
8,35
736
.00.35
0.03
170.00
Negative
Enterprene
urial st a
39,281
14,642
37.2
0.37
0.05
170.19
Negative
Other new
firm
s20
,330
3,72
137
.30.36
0.07
224.17
Positive
Establish
ed firm
s36
4,46
126
,793
36.6
0.43
0.10
221.79
Positive
Spin‐offs
22,470
6,19
036
.00.37
0.12
193.49
Positive
Enterprene
urial st a
37,189
14,798
37.3
0.36
0.12
192.71
Positive
Other new
firm
s20
,046
2,94
636
.60.39
0.14
228.61
76
TABLE B4
.1: A
verage
Wages in
New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)5.10
3***
5.08
9***
4.69
5***
4.92
5***
4.97
5***
(0.003
)(0.003
)(0.012
)(0.015
)(0.028
)
Positive cluster g
rowth (β
2)0.03
8***
0.01
4***
‐0.024
***
(0.006
)(0.005
)(0.006
)
New
firm
(β3)
0.03
6***
‐0.008
0.01
8***
‐0.001
‐0.001
(0.005
)(0.006
)(0.005
)(0.005
)(0.006
)
Positive cluster g
rowth x New
firm
(β4)
0.09
3***
0.05
9***
0.00
60.00
7(0.009
)(0.009
)(0.008
)(0.010
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns61
,943
61,943
61,943
61,943
61,943
R‐squared
0.00
10.00
70.13
50.29
30.02
6Notes: The
dep
ende
nt variable is the log of th
e average ho
urly wage in th
e firm. The
sample is redu
ced to
arou
nd 60,00
0 firms a
nd th
e nu
mbe
r of clusters to 46
7, because so
me firms d
o no
t have em
ployees a
nd
because reliable wage inform
ation cann
ot be ob
tained
for a
ll firms w
ith employees. See
text fo
r more
details. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
77
TABLE B4
.2: Skill‐Intensity
in New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)0.07
8***
0.04
2***
‐0.051
***
‐0.023
***
0.06
2***
(0.001
)(0.001
)(0.003
)(0.003
)(0.006
)
Positive cluster g
rowth (β
2)0.09
2***
0.09
3***
0.01
1***
(0.002
)(0.002
)(0.002
)
New
firm
(β3)
0.03
0***
0.01
4***
0.02
6***
0.00
5***
0.00
7***
(0.002
)(0.002
)(0.002
)(0.001
)(0.002
)
Positive cluster g
rowth x New
firm
(β4)
0.01
7***
0.01
4***
0.00
9***
0.00
6(0.003
)(0.003
)(0.003
)(0.004
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns11
9,53
411
9,53
411
9,22
611
9,22
611
9,22
6R‐squared
0.00
30.03
90.04
80.23
60.00
1Notes: The
dep
ende
nt variable is the the share of employees w
ith a te
rtiary edu
catio
n de
gree
in th
e firm.
The nu
mbe
r of clusters is 4
77, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
78
TABLE B4
.3: Sales per W
orker in New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)13
.498
***
13.553
***
13.816
***
14.848
***
13.620
***
(0.004
)(0.005
)(0.012
)(0.027
)(0.037
)
Positive cluster g
rowth (β
2)‐0.139
***
‐0.127
***
0.00
4(0.007
)(0.007
)(0.008
)
New
firm
(β3)
‐0.137
***
‐0.137
***
‐0.158
***
‐0.102
***
‐0.064
***
(0.005
)(0.007
)(0.008
)(0.007
)(0.010
)
Positive cluster g
rowth x New
firm
(β4)
0.02
4**
0.02
2**
‐0.012
0.01
0(0.011
)(0.011
)(0.011
)(0.016
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns11
3,91
511
3,91
511
3,65
911
7,25
111
3,65
9R‐squared
0.00
60.01
10.01
80.30
30.00
4Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker. The nu
mbe
r of clusters is 4
36, since so
me
clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses.
The R‐squared in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
79
TABLE B4
.4: P
rodu
ctivity
in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)7.18
1***
7.12
9***
6.91
0***
7.89
4***
7.32
8***
(0.025
)(0.026
)(0.028
)(0.037
)(0.104
)
Positive cluster g
rowth (β
2)0.02
2***
0.00
9*‐0.023
***
(0.005
)(0.005
)(0.006
)
New
firm
(β3)
0.02
1***
‐0.016
***
0.01
9***
0.01
6***
0.01
5(0.004
)(0.006
)(0.006
)(0.005
)(0.010
)
Positive cluster g
rowth x New
firm
(β4)
0.07
6***
0.06
5***
0.02
4***
0.02
7*(0.008
)(0.008
)(0.008
)(0.016
)
log(Interm
ediate inpu
ts per worker)
0.38
7***
0.39
0***
0.39
7***
0.35
1***
0.35
1***
(0.002
)(0.002
)(0.002
)(0.002
)(0.013
)log(capital per worker)
0.12
0***
0.12
0***
0.11
3***
0.14
3***
0.14
2***
(0.001
)(0.001
)(0.001
)(0.002
)(0.007
)log(nu
mbe
r of w
orkers)
0.05
9***
0.06
0***
0.08
4***
0.05
7***
0.05
6***
(0.002
)(0.002
)(0.002
)(0.002
)(0.007
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns93
,464
93,464
93,266
93,266
93,266
R‐squared
0.51
90.52
0.53
0.63
10.44
4Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker in the firm. See
text fo
r more de
tails.
Robu
st standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
nu
mbe
r of clusters is 4
31. ***
p<0.01, ** p<
0.05
, * p<0.1.
80
TABLE B4
.4A: In
term
ediate In
puts per W
orker in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)12
.568
***12
.776
***13
.534
***14
.022
***12
.897
***
(0.006
)(0.007
)(0.020
)(0.028
)(0.059
)
Positive cluster g
rowth (β
2)‐0.519
***‐0.431
***
‐0.004
(0.012
)(0.012
)(0.011
)
New
firm
(β3)
‐0.330
***‐0.271
***‐0.333
***‐0.095
***‐0.099
***
(0.009
)(0.012
)(0.012
)(0.010
)(0.016
)
Positive cluster g
rowth x New
firm
(β4)
‐0.046
**‐0.030
‐0.047
***
‐0.042
(0.019
)(0.019
)(0.015
)(0.026
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns10
1,02
910
1,02
910
0,81
710
0,81
710
0,81
7R‐squared
0.01
20.04
50.07
80.39
90.01
0Notes: The
dep
ende
nt variable is the log of interm
ediate inpu
ts per worker in the firm. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the
de‐m
eane
d data. The
num
ber o
f clusters is 4
31. ***
p<0.01, ** p<
0.05
, * p<0.1.
81
TABLE B4
.4B: Cap
ital per W
orker in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)12
.221
***12
.342
***10
.665
***12
.015
***11
.736
***
(0.007
)(0.010
)(0.023
)(0.042
)(0.051
)
Positive cluster g
rowth (β
2)‐0.312
***‐0.266
***
0.00
3(0.015
)(0.015
)(0.014
)
New
firm
(β3)
‐0.425
***‐0.479
***‐0.236
***‐0.127
***‐0.134
***
(0.011
)(0.015
)(0.015
)(0.012
)(0.025
)
Positive cluster g
rowth x New
firm
(β4)
0.17
0***
0.10
1***
0.02
20.04
1(0.022
)(0.021
)(0.017
)(0.037
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns10
6,16
310
6,16
310
5,92
210
5,92
210
5,92
2R‐squared
0.01
40.01
90.09
00.43
80.01
1Notes: The
dep
ende
nt variable is the log of capita
lper worker in the firm. See
text fo
r more de
tails.
Robu
st standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
num
ber o
f clusters is 4
36. ***
p<0.01, ** p<
0.05
, * p<0.1.
82
TABLE B4
.5: W
ages in
New
and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)5.34
1***
5.32
3***
4.58
3***
4.56
0***
4.70
0***
4.40
5***
4.35
7***
4.49
7***
(0.001
)(0.012
)(0.012
)(0.013
)(0.004
)(0.016
)(0.015
)(0.006
)
Positive cluster g
rowth (β
2)0.04
40.01
30.01
3**
0.00
70.00
1(0.027
)(0.013
)(0.005
)(0.009
)(0.005
)
New
firm
(β3)
‐0.086
***
‐0.116
***
‐0.078
***
‐0.038
***
‐0.038
***
0.01
2*0.01
5***
0.01
3***
(0.002
)(0.014
)(0.009
)(0.005
)(0.002
)(0.006
)(0.005
)(0.002
)
Positive cluster g
rowth x New
firm
(β4)
0.06
2**
0.01
5‐0.023
**‐0.016
***
0.01
4‐0.018
**‐0.014
***
(0.029
)(0.014
)(0.009
)(0.003
)(0.011
)(0.008
)(0.003
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns70
9,67
670
9,67
669
3,20
969
3,20
969
3,20
969
3,20
969
3,20
969
3,20
9R‐squared
0.00
40.00
70.35
90.42
10.26
10.38
10.43
10.27
3Notes: The
dep
ende
nt variable is the log of th
e ho
urly wage. The
sample is redu
ced to aroun
d 1,17
0,00
0 individu
als a
nd th
e nu
mbe
r of clusters to 46
7,
because reliable wage inform
ation cann
ot be ob
tained
for a
ll workers. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in
columns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
83
TABLE B4
.6: Skills in
New
and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)0.07
3***
0.04
9***
‐0.025
***
‐0.042
***
0.04
2***
0.00
5‐0.042
***
0.04
6***
(0.004
)(0.003
)(0.003
)(0.004
)(0.001
)(0.005
)(0.005
)(0.001
)
Positive cluster g
rowth (β
2)0.05
5***
0.06
6***
0.00
30.06
7***
0.00
3(0.011
)(0.008
)(0.002
)(0.006
)(0.002
)
New
firm
(β3)
0.00
8*‐0.004
‐0.002
‐0.005
**‐0.004
***
0.00
5‐0.001
‐0.001
(0.004
)(0.003
)(0.003
)(0.002
)(0.001
)(0.003
)(0.002
)(0.001
)
Positive cluster g
rowth x New
firm
(β4)
0.02
2*0.01
4‐0.002
‐0.003
*0.01
0‐0.002
‐0.002
*(0.012
)(0.009
)(0.004
)(0.001
)(0.007
)(0.004
)(0.001
)
female
‐0.017
***
‐0.024
***
‐0.024
***
‐0.017
***
‐0.024
***
‐0.024
***
(0.002
)(0.002
)(0.001
)(0.002
)(0.002
)(0.001
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,00
4,67
71,00
4,67
797
3,61
097
3,61
097
3,61
097
3,61
097
3,61
097
3,61
0R‐squared
0.00
00.01
30.04
80.15
30.01
70.05
20.15
30.01
7Notes: The
mod
el estim
ated
is a line
ar probability mod
el and
the de
pend
ent v
ariable is a du
mmy for w
hether th
e pe
rson
has com
pleted
tertiary
education. The
num
ber o
f clusters is 4
77, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard
errors in
paren
theses. The
R‐squ
ared
in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
84
TABLE B5
.1: A
verage
Wages in
Differen
t Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)5.10
3***
5.08
9***
4.69
4***
4.92
1***
4.97
2***
(0.003
)(0.003
)(0.012
)(0.015
)(0.028
)
Positive cluster g
rowth (β
2)0.03
8***
0.01
5***
‐0.024
***
(0.006
)(0.005
)(0.006
)
Firm
types:
Spin‐off (β
31)
0.01
2*‐0.032
***
‐0.002
0.00
0‐0.000
(0.006
)(0.008
)(0.007
)(0.007
)(0.009
)Entrep
rene
urial start‐up (β
32)
‐0.013
**‐0.064
***
‐0.034
***
‐0.047
***
‐0.045
***
(0.006
)(0.008
)(0.007
)(0.007
)(0.008
)Other new
firm
(β33)
0.20
6***
0.18
5***
0.19
3***
0.10
7***
0.10
4***
(0.009
)(0.012
)(0.011
)(0.010
)(0.012
)
Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.10
3***
0.05
7***
0.01
00.01
0(0.014
)(0.012
)(0.011
)(0.016
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.10
4***
0.07
5***
0.01
20.01
2(0.013
)(0.012
)(0.011
)(0.011
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.03
9**
0.01
5‐0.019
‐0.015
(0.018
)(0.016
)(0.015
)(0.021
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns61
,943
61,943
61,943
61,943
61,943
R‐squared
0.01
00.01
60.14
20.29
60.03
0Notes: The
dep
ende
nt variable is the log of th
e average ho
urly wage in th
e firm. The
sample is redu
ced to aroun
d 60
,000
firm
s and the nu
mbe
r of clusters to 46
7, because so
me firms d
o no
t have em
ployees a
nd because re
liable wage inform
ation cann
ot
be obtaine
d for a
ll firms w
ith employees. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in
column 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
85
TABLE B5
.2: Skill‐Intensity
in Differen
t Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)0.07
8***
0.04
2***
‐0.051
***
‐0.024
***
0.06
1***
(0.001
)(0.001
)(0.003
)(0.003
)(0.006
)
Positive cluster g
rowth (β
2)0.09
2***
0.09
3***
0.01
1***
(0.002
)(0.002
)(0.002
)Firm
types:
Spin‐off (β
31)
0.05
8***
0.05
6***
0.03
6***
0.03
3***
0.00
3(0.005
)(0.005
)(0.005
)(0.008
)(0.003
)Entrep
rene
urial start‐up (β
32)
0.02
9***
0.01
8***
0.02
9***
0.00
20.00
4(0.002
)(0.002
)(0.002
)(0.002
)(0.003
)Other new
firm
(β33)
0.04
0***
0.03
6***
0.04
9***
0.02
4***
0.02
4***
(0.003
)(0.004
)(0.004
)(0.003
)(0.004
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.05
6***
0.05
3***
0.03
5***
0.03
2***
(0.005
)(0.005
)(0.004
)(0.008
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.00
2‐0.001
0.00
2‐0.002
(0.004
)(0.004
)(0.004
)(0.005
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.003
‐0.002
‐0.008
‐0.010
(0.007
)(0.007
)(0.006
)(0.008
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns11
9,53
411
9,53
411
9,22
611
9,22
611
9,22
6R‐squared
0.00
30.04
00.04
90.23
70.00
2Notes: The
dep
ende
nt variable is the the share of employees w
ith a te
rtiary edu
catio
n de
gree
in th
e firm. The
num
ber o
f clusters is 477
, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard
errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
86
TABLE B5
.3: Sales per W
orker in Diffe
rent Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)13
.498
***
13.553
***
13.797
***
14.719
***
13.608
***
(0.004
)(0.005
)(0.012
)(0.022
)(0.037
)
Positive cluster g
rowth (β
2)‐0.139
***
‐0.127
***
‐0.003
(0.007
)(0.007
)(0.007
)Firm
types:
Spin‐off (β
31)
‐0.019
**‐0.044
***
‐0.073
***
0.00
6‐0.006
(0.008
)(0.011
)(0.011
)(0.009
)(0.010
)Entrep
rene
urial start‐up (β
32)
‐0.265
***
‐0.267
***
‐0.282
***
‐0.131
***
‐0.135
***
(0.006
)(0.009
)(0.009
)(0.008
)(0.015
)Other new
firm
(β33)
0.18
2***
0.18
1***
0.16
6***
0.08
8***
0.08
2***
(0.014
)(0.019
)(0.019
)(0.015
)(0.021
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.07
2***
0.07
0***
0.02
4*0.04
3*(0.016
)(0.016
)(0.013
)(0.022
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.03
5***
0.03
4***
‐0.011
0.00
2(0.013
)(0.013
)(0.011
)(0.019
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.01
80.00
80.04
8**
0.05
1(0.027
)(0.027
)(0.023
)(0.048
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns11
3,91
511
3,91
511
3,65
911
3,65
911
3,65
9R‐squared
0.01
90.02
40.03
00.32
10.01
0Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker. The nu
mbe
r of clusters is 4
36, since so
me
clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐
squared in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
87
TABLE B5
.4: P
rodu
ctivity
in Differen
t Types of N
ew Firm
s and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)7.20
7***
7.15
5***
6.93
1***
7.90
2***
7.33
6***
(0.025
)(0.026
)(0.028
)(0.037
)(0.104
)
Positive cluster g
rowth (β
2)0.02
0***
0.00
8‐0.025
***
(0.005
)(0.005
)(0.006
)Firm
types:
Spin‐off (β
31)
0.08
8***
0.03
6***
0.07
1***
0.06
5***
0.06
2***
(0.006
)(0.008
)(0.008
)(0.007
)(0.011
)Entrep
rene
urial start‐up (β
32)
‐0.027
***
‐0.070
***
‐0.032
***
‐0.020
***
‐0.019
**(0.005
)(0.007
)(0.007
)(0.006
)(0.010
)Other new
firm
(β33)
0.06
7***
0.06
2***
0.08
8***
0.03
0**
0.02
6(0.011
)(0.015
)(0.015
)(0.012
)(0.021
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.12
2***
0.10
1***
0.05
2***
0.05
5**
(0.013
)(0.013
)(0.011
)(0.026
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.08
4***
0.07
6***
0.02
3**
0.02
4*(0.010
)(0.010
)(0.009
)(0.013
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.00
7‐0.004
0.01
60.02
0(0.021
)(0.021
)(0.018
)(0.031
)
log(Interm
ediate inpu
ts per worker)
0.38
5***
0.38
9***
0.39
6***
0.35
1***
0.35
0***
(0.002
)(0.002
)(0.002
)(0.002
)(0.013
)log(capital per worker)
0.12
0***
0.12
0***
0.11
3***
0.14
3***
0.14
3***
(0.001
)(0.001
)(0.001
)(0.002
)(0.007
)log(nu
mbe
r of w
orkers)
0.05
6***
0.05
6***
0.08
1***
0.05
5***
0.05
4***
(0.002
)(0.002
)(0.002
)(0.002
)(0.007
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns93
,464
93,464
93,266
93,266
93,266
R‐squared
0.52
00.52
20.53
10.63
20.44
5Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker in the firm. See
text fo
r more de
tails. R
obust
standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
num
ber o
f clusters is
431. ***
p<0.01, ** p<
0.05
, * p<0.1.
88
TABLE B5
.5: W
ages in
Differen
t Types of N
ew Firm
s and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)5.34
1***
5.32
3***
4.58
4***
4.56
0***
4.70
0***
4.41
1***
4.36
2***
4.50
1***
(0.001
)(0.012
)(0.011
)(0.013
)(0.004
)(0.016
)(0.015
)(0.006
)
Positive cluster g
rowth (β
2)0.04
40.01
30.01
3**
0.00
70.00
1(0.027
)(0.013
)(0.005
)(0.009
)(0.005
)Firm
types:
Spin‐off (β
31)
‐0.139
***
‐0.174
***
‐0.108
***
‐0.053
***
‐0.055
***
‐0.016
**‐0.000
‐0.002
(0.003
)(0.015
)(0.010
)(0.006
)(0.003
)(0.007
)(0.006
)(0.003
)Entrep
rene
urial start‐up (β
32)
‐0.116
***
‐0.154
***
‐0.114
***
‐0.073
***
‐0.072
***
‐0.012
*‐0.011
*‐0.012
***
(0.003
)(0.015
)(0.010
)(0.006
)(0.003
)(0.007
)(0.006
)(0.003
)Other new
firm
(β33)
0.01
9***
0.01
80.01
50.03
2***
0.02
9***
0.08
2***
0.06
7***
0.06
4***
(0.003
)(0.023
)(0.015
)(0.010
)(0.003
)(0.015
)(0.011
)(0.003
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.08
6***
0.01
9‐0.020
**‐0.014
***
0.03
0**
‐0.007
‐0.002
(0.030
)(0.016
)(0.010
)(0.004
)(0.012
)(0.009
)(0.004
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.07
2**
0.03
7*‐0.014
‐0.009
**0.03
1**
‐0.010
‐0.007
(0.030
)(0.020
)(0.012
)(0.004
)(0.014
)(0.010
)(0.004
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.007
‐0.032
*‐0.048
***
‐0.036
***
‐0.035
*‐0.047
***
‐0.041
***
(0.041
)(0.019
)(0.017
)(0.005
)(0.020
)(0.015
)(0.005
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns70
9,67
670
9,67
669
3,20
969
3,20
969
3,20
969
3,20
969
3,20
969
3,20
9R‐squared
0.00
60.01
00.36
00.42
20.26
20.38
10.43
10.27
3Notes: The
dep
ende
nt variable is the log of th
e ho
urly wage. The
sample is redu
ced to aroun
d 71
0,00
0 individu
als a
nd th
e nu
mbe
r of clusters to 46
7, because re
liable
wage inform
ation cann
ot be ob
tained
for a
ll workers. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
ns 5 and
8 are
compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
89
TABLE B5
.6: Skills in
Differen
t Types of N
ew Firm
s and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)0.07
3***
0.04
9***
‐0.024
***
‐0.042
***
0.04
2***
0.00
8‐0.040
***
0.04
8***
(0.004
)(0.003
)(0.003
)(0.004
)(0.001
)(0.005
)(0.005
)(0.001
)
Positive cluster g
rowth (β
2)0.05
5***
0.06
6***
0.00
30.06
7***
0.00
2(0.011
)(0.008
)(0.002
)(0.006
)(0.002
)Firm
types:
Spin‐off (β
31)
‐0.007
‐0.018
***
‐0.014
***
‐0.008
***
‐0.007
***
‐0.007
**‐0.004
**‐0.004
***
(0.005
)(0.003
)(0.003
)(0.002
)(0.001
)(0.003
)(0.002
)(0.001
)Entrep
rene
urial start‐up (β
32)
0.00
6‐0.003
‐0.002
‐0.010
***
‐0.008
***
0.00
2‐0.007
***
‐0.006
***
(0.004
)(0.003
)(0.003
)(0.002
)(0.001
)(0.003
)(0.002
)(0.001
)Other new
firm
(β33)
0.03
3***
0.01
9***
0.01
8***
0.01
0*0.01
0***
0.02
7***
0.01
3**
0.01
3***
(0.007
)(0.007
)(0.007
)(0.005
)(0.002
)(0.007
)(0.005
)(0.002
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.03
1***
0.02
4***
0.00
70.00
9***
0.02
2***
0.00
8*0.00
9***
(0.012
)(0.009
)(0.004
)(0.002
)(0.008
)(0.004
)(0.002
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.01
40.00
4‐0.005
‐0.006
***
‐0.001
‐0.005
‐0.006
***
(0.012
)(0.009
)(0.004
)(0.002
)(0.007
)(0.004
)(0.002
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.02
00.01
3‐0.010
‐0.012
***
0.01
0‐0.010
‐0.012
***
(0.016
)(0.013
)(0.007
)(0.003
)(0.013
)(0.007
)(0.003
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,00
4,67
71,00
4,67
797
3,61
097
3,61
097
3,61
097
3,61
097
3,61
097
3,61
0R‐squared
0.00
10.01
30.04
80.15
30.01
70.05
30.15
30.01
8Notes: The
mod
el estim
ated
is a line
ar probability mod
el and
the de
pend
ent v
ariable is a du
mmy for w
hether th
e pe
rson
has com
pleted
tertiary edu
catio
n. The
nu
mbe
r of clusters is 4
77, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐
squared in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
32
Appendix C. Estimation Results for Alternative Definition ofEntrepreneurs(Maxageof5years).
92
TABLE C2
.1.: Num
ber o
f clusters a
nd jo
bs. B
y cluster g
rowth
Num
ber o
f clusters
Num
ber o
f job
sNum
ber o
f firm
s
Cluster g
rowth
#%
#%
#%
‐0.50‐
124
11.47
38,334
2.46
4,31
42.29
‐0.50;‐0.25
202
18.69
222,93
314
.29
19,820
10.54
‐0.25;‐0.1
190
17.58
366,29
823
.47
47,918
25.48
‐0.1‐0.1
253
23.40
634,57
940
.67
76,068
40.45
0.1‐0.25
857.86
121,98
47.82
19,838
10.55
0.25‐0.5
656.01
113,09
77.25
14,610
7.77
0.5+
137
12.67
62,396
4.00
5,43
72.89
New
252.31
831
0.05
390.02
Totals
1,08
110
0.00
1,56
0,45
210
0.00
188,04
410
0.00
93
TABLE C2
.2: N
umbe
r of firm
type
s and
jobs. B
y cluster g
rowth and
firm
type
sNum
ber o
f firm
s
Establish
ed firm
sSpin‐offs
Other new
firm
sCluster g
rowth
#%
#%
#%
#%
<‐0.5
2,95
2
2.3
290
1.7
855
2.5
217
3.5
‐0.5;‐0
.25
14,271
11.0
1,57
9
9.3
3,35
9
9.7
611
9.8
‐0.25;‐0.1
33,790
26.0
5,75
8
33.8
7,17
8
20.7
1,19
2
19.2
‐0.1‐0.1
53,597
41.2
6,23
0
36.6
13,593
39.1
2,64
8
42.6
0.1‐0.25
13,489
10.4
1,53
9
9.0
4,02
8
11.6
782
12.6
0.25‐0.5
8,86
6
6.8
1,33
2
7.8
3,91
4
11.3
498
8.0
0.5+
3,09
1
2.4
314
1.8
1,79
8
5.2
273
4.4
Totals
130,05
6
100.0
17,042
100.0
34,725
100.0
6,22
1
100.0
Num
ber o
f job
s
Establish
ed firm
sSpin‐offs
Other new
firm
sCluster g
rowth
#%
#%
#%
#%
<‐0.5
33,989
2.4
924
1.9
2,33
1
3.3
1,09
0
3.6
‐0.5;‐0
.25
208,30
3
14.8
4,33
4
9.0
7,08
5
10.0
3,21
1
10.7
‐0.25;‐0.1
331,48
8
23.5
14,778
30
.714
,478
20.4
5,55
4
18.6
‐0.1‐0.1
573,27
5
40.6
20,107
41
.828
,741
40.4
12,456
41.7
0.1‐0.25
107,29
7
7.6
3,91
3
8.1
7,33
0
10.3
3,44
4
11.5
0.25‐0.5
101,02
1
7.2
3,22
7
6.7
6,63
0
9.3
2,21
9
7.4
0.5+
56,022
4.0
792
1.6
4,48
4
6.3
1,92
9
6.5
Totals
1,41
1,39
5
100.0
48,075
100.0
71,079
100.0
29,903
100.0
Entrep
rene
urial start‐ups
Entrep
rene
urial start‐ups
94
TABLE C2
.3: A
verage
num
ber o
f job
s per firm
. By cluster g
rowth after firm
type
sJobs per firm
Establish
ed
firms
Spin‐offs
Entrep
rene
urial start‐ups
Other new
firms
Cluster g
rowth
##
##
<‐0.5
11.5
3.1
2.8
5.0
‐0.5;‐0
.25
14.6
2.7
2.1
5.3
‐0.25;‐0.1
9.8
2.5
2.0
4.7
‐0.1‐0.1
10.7
3.1
2.1
4.7
0.1‐0.25
8.0
2.5
1.8
4.4
0.25‐0.5
11.4
2.3
1.7
4.5
0.5+
18.1
2.4
2.5
7.1
Totals
10.9
2.7
2.1
4.8
95
TABLE C2
.4: A
verage
individu
al cha
racteristics. By firm ty
pe and
cluster growth
Cluster g
roFirm
type
Num
ber o
f workers
Num
ber o
f firms
Mean age
Share of
females
Share of
highly
educated
Mean
hourly
wage
(DKK
)
Negative
Establish
ed firm
s87
0,99
9
84,661
40.48
0.32
0.04
210.39
Negative
Spin‐offs
32,222
11,387
35.34
0.29
0.02
164.63
Negative
Enterprene
urial start‐ups
40,077
18,859
37.09
0.30
0.12
165.69
Negative
Other new
firm
s17
,192
3,67
0
37.39
0.32
0.06
218.89
Positive
Establish
ed firm
s54
0,39
6
45,395
38.06
0.41
0.11
223.36
Positive
Spin‐offs
15,853
5,65
5
35.74
0.39
0.04
180.50
Positive
Enterprene
urial start‐ups
31,002
15,866
37.41
0.40
0.12
182.21
Positive
Other new
firm
s12
,711
2,55
1
36.28
0.38
0.13
218.04
96
TABLE C4
.1: A
verage
Wages in
New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)5.11
2***
5.10
3***
4.74
1***
6.02
2***
4.94
5***
(0.002
)(0.002
)(0.009
)(0.089
)(0.021
)
Positive cluster g
rowth (β
2)0.03
0***
0.01
4***
‐0.017
***
(0.004
)(0.004
)(0.005
)
New
firm
(β3)
0.00
9**
‐0.019
***
0.00
9**
‐0.003
‐0.003
(0.004
)(0.004
)(0.004
)(0.004
)(0.005
)
Positive cluster g
rowth x New
firm
(β4)
0.07
1***
0.04
0***
0.00
10.00
4(0.009
)(0.008
)(0.008
)(0.010
)Co
ntrols for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns95
,228
95,228
95,228
95,228
95,228
R‐squared
0.00
00.00
30.13
20.27
40.03
2Notes: The
dep
ende
nt variable is the log of th
e average ho
urly wage in th
e firm. The
sample is redu
ced to
arou
nd 95,00
0 firms a
nd th
e nu
mbe
r of clusters to 10
54, because so
me firms d
o no
t have em
ployees a
nd
because reliable wage inform
ation cann
ot be ob
tained
for a
ll firms w
ith employees. See
text fo
r more de
tails.
Robu
st standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<
0.01
, ** p<
0.05
, * p<0.1.
97
TABLE C4
.2: Skill‐Intensity
in New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)0.07
3***
0.03
8***
‐0.068
***
0.15
80.03
0**
(0.001
)(0.001
)(0.002
)(0.118
)(0.012
)
Positive cluster g
rowth (β
2)0.10
1***
0.09
7***
0.00
9***
(0.002
)(0.002
)(0.002
)
New
firm
(β3)
0.02
0***
0.00
3***
0.01
8***
0.00
7***
0.00
8***
(0.001
)(0.001
)(0.001
)(0.001
)(0.002
)
Positive cluster g
rowth x New
firm
(β4)
0.02
5***
0.02
3***
0.01
4***
0.01
2***
(0.003
)(0.003
)(0.003
)(0.004
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns18
8,04
418
8,04
418
7,63
318
7,63
318
7,63
3R‐squared
0.00
10.04
90.06
00.23
60.00
4Notes: The
dep
ende
nt variable is the the share of employees w
ith a te
rtiary edu
catio
n de
gree
in th
e firm. The
nu
mbe
r of clusters is 1
081, since some clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more
details. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
98
TABLE C4
.3: Sales per W
orker in New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)13
.410
***
13.474
***
13.779
***
18.148
***
13.558
***
(0.002
)(0.003
)(0.009
)(0.846
)(0.026
)
Positive cluster g
rowth (β
2)‐0.189
***
‐0.139
***
0.00
5(0.005
)(0.005
)(0.006
)
New
firm
(β3)
‐0.237
***
‐0.220
***
‐0.251
***
‐0.202
***
‐0.167
***
(0.004
)(0.006
)(0.006
)(0.006
)(0.008
)
Positive cluster g
rowth x New
firm
(β4)
‐0.008
‐0.006
‐0.017
*0.00
7(0.009
)(0.009
)(0.009
)(0.011
)Co
ntrols for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns17
8,15
217
8,15
217
7,79
918
1,42
617
7,79
9R‐squared
0.01
60.02
70.05
20.31
90.01
3Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker. The nu
mbe
r of clusters is 1
021, since
some clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, *
p<0.1.
99
TABLE C4
.4: P
rodu
ctivity
in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)7.11
0***
7.02
9***
6.87
6***
7.00
5***
7.16
1***
(0.022
)(0.022
)(0.024
)(0.060
)(0.093
)
Positive cluster g
rowth (β
2)0.06
2***
0.04
7***
‐0.017
***
(0.004
)(0.004
)(0.005
)
New
firm
(β3)
‐0.041
***
‐0.066
***
‐0.042
***
‐0.046
***
‐0.047
***
(0.004
)(0.004
)(0.005
)(0.004
)(0.007
)
Positive cluster g
rowth x New
firm
(β4)
0.06
2***
0.05
1***
0.02
5***
0.02
8*(0.008
)(0.008
)(0.007
)(0.016
)
log(Interm
ediate inpu
ts per worker)
0.39
6***
0.40
1***
0.40
7***
0.37
6***
0.37
5***
(0.002
)(0.002
)(0.002
)(0.002
)(0.012
)log(capital per worker)
0.11
5***
0.11
4***
0.11
0***
0.13
3***
0.13
2***
(0.001
)(0.001
)(0.001
)(0.001
)(0.007
)log(nu
mbe
r of w
orkers)
0.05
4***
0.05
3***
0.06
8***
0.05
3***
0.05
3***
(0.001
)(0.001
)(0.001
)(0.001
)(0.005
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns13
3,08
713
3,08
713
2,84
213
2,84
213
2,84
2R‐squared
0.52
80.53
00.53
70.62
10.46
9Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker in the firm. See
text fo
r more de
tails.
Robu
st standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
nu
mbe
r of clusters is 9
75. ***
p<0.01, ** p<
0.05
, * p<0.1.
100
TABLE C4
.4A: In
term
ediate In
puts per W
orker in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)12
.605
***
12.759
***
13.425
***
13.688
***
12.976
***
(0.004
)(0.005
)(0.015
)(0.117
)(0.046
)
Positive cluster g
rowth (β
2)‐0.488
***
‐0.381
***
0.00
1(0.009
)(0.010
)(0.010
)
New
firm
(β3)
‐0.368
***
‐0.301
***
‐0.375
***
‐0.192
***
‐0.194
***
(0.008
)(0.009
)(0.009
)(0.008
)(0.016
)
Positive cluster g
rowth x New
firm
(β4)
‐0.112
***
‐0.087
***
‐0.072
***
‐0.068
**(0.017
)(0.017
)(0.015
)(0.033
)Co
ntrol for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns14
3,86
614
3,86
614
3,60
014
3,60
014
3,60
0R‐squared
0.01
50.04
80.07
90.35
90.01
5Notes: The
dep
ende
nt variable is the log of interm
ediate inpu
ts per worker in the firm. See
text fo
r more de
tails.
Robu
st standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
nu
mbe
r of clusters is 9
79. ***
p<0.01, ** p<
0.05
, * p<0.1.
101
TABLE C4
.4B: Cap
ital per W
orker in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)12
.113
***
12.127
***
10.569
***
13.715
***
11.647
***
(0.006
)(0.007
)(0.019
)(0.115
)(0.040
)
Positive cluster g
rowth (β
2)‐0.044
***
‐0.022
*0.01
0(0.012
)(0.012
)(0.012
)
New
firm
(β3)
‐0.424
***
‐0.440
***
‐0.214
***
‐0.146
***
‐0.149
***
(0.009
)(0.012
)(0.012
)(0.010
)(0.022
)
Positive cluster g
rowth x New
firm
(β4)
0.04
9**
‐0.002
0.03
4**
0.05
0(0.019
)(0.019
)(0.015
)(0.034
)Co
ntrol for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns14
7,21
014
7,21
014
6,91
514
6,91
514
6,91
5R‐squared
0.01
30.01
30.07
20.39
90.01
1
Notes: The
dep
ende
nt variable is the log of capita
lper worker in the firm. See
text fo
r more de
tails. R
obust stand
ard
errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
num
ber o
f clusters is 9
82.
*** p<
0.01
, ** p<
0.05
, * p<0.1.
102
TABLE C4
.5: W
ages in
New
and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)5.33
0***
5.30
4***
4.57
9***
5.21
1***
4.68
1***
4.41
6***
5.03
4***
4.49
1***
(0.000
)(0.007
)(0.010
)(0.022
)(0.003
)(0.011
)(0.022
)(0.005
)
Positive cluster g
rowth (β
2)0.07
2***
0.02
8***
0.01
3***
0.01
10.00
1(0.020
)(0.009
)(0.005
)(0.007
)(0.005
)
New
firm
(β3)
‐0.121
***
‐0.119
***
‐0.080
***
‐0.052
***
‐0.051
***
0.00
60.00
60.00
5***
(0.002
)(0.009
)(0.006
)(0.004
)(0.001
)(0.005
)(0.004
)(0.002
)
Positive cluster g
rowth x New
firm
(β4)
‐0.008
‐0.019
*‐0.027
***
‐0.021
***
0.02
4***
‐0.003
0.00
1(0.021
)(0.010
)(0.007
)(0.002
)(0.008
)(0.006
)(0.002
)Co
ntrols for g
ende
r, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,17
0,48
91,17
0,48
91,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
8R‐squared
0.00
40.01
00.35
90.41
30.28
10.38
10.42
30.29
2Notes: The
dep
ende
nt variable is the log of th
e ho
urly wage. The
sample is redu
ced to aroun
d 1,17
0,00
0 individu
als a
nd th
e nu
mbe
r of clusters to 10
53,
because reliable wage inform
ation cann
ot be ob
tained
for a
ll workers. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in
columns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
103
TABLE C4
.6: Skills in
New
and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)0.06
9***
0.04
3***
‐0.031
***
0.15
50.03
5***
‐0.006
0.16
1*0.03
9***
(0.003
)(0.002
)(0.003
)(0.097
)(0.001
)(0.004
)(0.097
)(0.001
)
Positive cluster g
rowth (β
2)0.06
7***
0.07
4***
0.00
5**
0.07
3***
0.00
3(0.009
)(0.007
)(0.002
)(0.006
)(0.002
)
New
firm
(β3)
0.00
1‐0.005
**‐0.004
*‐0.004
**‐0.002
***
‐0.001
‐0.001
‐0.001
(0.003
)(0.002
)(0.002
)(0.001
)(0.001
)(0.002
)(0.002
)(0.001
)
Positive cluster g
rowth x New
firm
(β4)
0.01
20.01
0‐0.005
*‐0.005
***
0.01
2*‐0.002
‐0.003
***
(0.010
)(0.008
)(0.003
)(0.001
)(0.006
)(0.003
)(0.001
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,56
0,45
21,56
0,45
21,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
0R‐squared
0.00
00.01
80.04
30.14
40.01
30.04
70.14
50.01
4Notes: The
mod
el estim
ated
is a line
ar probability mod
el and
the de
pend
ent v
ariable is a du
mmy for w
hether th
e pe
rson
has com
pleted
tertiary
education. The
num
ber o
f clusters is 1
081, since some clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard
errors in
paren
theses. The
R‐squ
ared
in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
104
TABLE C5
.1: A
verage
Wages in
Differen
t Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)5.11
2***
5.10
3***
4.74
4***
6.02
2***
4.94
5***
(0.002
)(0.002
)(0.009
)(0.089
)(0.021
)
Positive cluster g
rowth (β
2)0.03
0***
0.01
4***
‐0.017
***
(0.004
)(0.004
)(0.005
)Firm
types:
Spin‐off (β
31)
‐0.018
***
‐0.049
***
‐0.009
‐0.003
‐0.004
(0.006
)(0.006
)(0.006
)(0.006
)(0.007
)Entrep
rene
urial start‐up (β
32)
‐0.028
***
‐0.054
***
‐0.030
***
‐0.041
***
‐0.040
***
(0.006
)(0.006
)(0.006
)(0.006
)(0.007
)Other new
firm
(β33)
0.16
8***
0.15
5***
0.16
2***
0.10
2***
0.09
9***
(0.009
)(0.011
)(0.010
)(0.009
)(0.010
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.09
3***
0.04
0***
0.00
90.01
2(0.014
)(0.012
)(0.011
)(0.015
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.06
2***
0.04
3***
0.00
20.00
5(0.012
)(0.011
)(0.011
)(0.011
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.02
50.00
7‐0.026
*‐0.021
(0.019
)(0.017
)(0.016
)(0.019
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns95
,228
95,228
95,228
95,228
95,228
R‐squared
0.00
50.00
80.13
70.27
60.03
5Notes: The
dep
ende
nt variable is the log of th
e average ho
urly wage in th
e firm. The
sample is redu
ced to aroun
d 95
,000
firms a
nd th
e nu
mbe
r of clusters to 10
54, because so
me firms d
o no
t have em
ployees a
nd because re
liable wage
inform
ation cann
ot be ob
tained
for a
ll firms w
ith employees. See
text fo
r more de
tails. R
obust stand
ard errors in
parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
105
TABLE C5
.2: Skill‐Intensity
in Differen
t Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)0.07
3***
0.03
8***
‐0.069
***
0.15
80.02
9**
(0.001
)(0.001
)(0.002
)(0.118
)(0.012
)
Positive cluster g
rowth (β
2)0.10
1***
0.09
7***
0.00
9***
(0.002
)(0.002
)(0.002
)Firm
types:
Spin‐off (β
31)
0.00
9***
‐0.010
***
0.00
9***
0.00
7***
0.00
7***
(0.002
)(0.001
)(0.002
)(0.001
)(0.003
)Entrep
rene
urial start‐up (β
32)
0.02
3***
0.00
6***
0.01
9***
0.00
5***
0.00
7***
(0.002
)(0.002
)(0.002
)(0.001
)(0.002
)Other new
firm
(β33)
0.03
4***
0.02
9***
0.04
2***
0.01
8***
0.01
8***
(0.003
)(0.003
)(0.003
)(0.003
)(0.005
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.06
1***
0.05
8***
0.03
5***
0.03
2***
(0.005
)(0.005
)(0.005
)(0.009
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.01
4***
0.01
3***
0.00
7**
0.00
5(0.004
)(0.003
)(0.003
)(0.005
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.002
0.00
20.00
80.00
7(0.007
)(0.007
)(0.007
)(0.008
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns18
8,04
418
8,04
418
7,63
318
7,63
318
7,63
3R‐squared
0.00
20.04
90.06
10.23
60.00
4Notes: The
dep
ende
nt variable is the the share of employees w
ith a te
rtiary edu
catio
n de
gree
in th
e firm. The
num
ber o
f clusters is 108
1, since some clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard
errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
106
TABLE C5
.3: Sales per W
orker in Diffe
rent Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)13
.410
***
13.474
***
13.767
***
14.101
***
13.550
***
(0.002
)(0.003
)(0.009
)(0.080
)(0.026
)
Positive cluster g
rowth (β
2)‐0.189
***
‐0.139
***
‐0.001
(0.005
)(0.005
)(0.006
)Firm
types:
Spin‐off (β
31)
‐0.116
***
‐0.155
***
‐0.200
***
‐0.103
***
‐0.110
***
(0.007
)(0.008
)(0.008
)(0.007
)(0.011
)Entrep
rene
urial start‐up (β
32)
‐0.347
***
‐0.310
***
‐0.333
***
‐0.226
***
‐0.227
***
(0.005
)(0.007
)(0.007
)(0.006
)(0.010
)Other new
firm
(β33)
0.05
2***
0.04
2**
0.03
0‐0.037
**‐0.042
*(0.014
)(0.019
)(0.019
)(0.016
)(0.022
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.10
9***
0.10
7***
0.04
6***
0.06
2***
(0.015
)(0.015
)(0.013
)(0.018
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
‐0.034
***
‐0.027
**‐0.007
0.00
2(0.011
)(0.011
)(0.010
)(0.015
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.05
3*0.01
70.01
70.01
9(0.029
)(0.029
)(0.025
)(0.043
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns17
8,15
217
8,15
217
7,79
917
7,79
917
7,79
9R‐squared
0.02
40.03
50.05
90.32
50.01
6Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker. The nu
mbe
r of clusters is 1
021, since some
clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐
squared in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
107
TABLE C5
.4: P
rodu
ctivity
in Differen
t Types of N
ew Firm
s and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)7.12
0***
7.03
9***
6.88
2***
7.01
0***
7.16
3***
(0.022
)(0.022
)(0.024
)(0.060
)(0.093
)
Positive cluster g
rowth (β
2)0.06
2***
0.04
7***
‐0.018
***
(0.004
)(0.004
)(0.005
)Firm
types:
Spin‐off (β
31)
0.00
9*‐0.030
***
‐0.004
‐0.006
‐0.008
(0.006
)(0.006
)(0.006
)(0.006
)(0.008
)Entrep
rene
urial start‐up (β
32)
‐0.074
***
‐0.100
***
‐0.075
***
‐0.073
***
‐0.073
***
(0.005
)(0.006
)(0.006
)(0.005
)(0.008
)Other new
firm
(β33)
‐0.020
*‐0.017
‐0.000
‐0.040
***
‐0.042
**(0.011
)(0.015
)(0.014
)(0.013
)(0.016
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.13
8***
0.11
5***
0.06
9***
0.07
2***
(0.014
)(0.013
)(0.012
)(0.027
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.05
6***
0.05
0***
0.02
3***
0.02
6*(0.010
)(0.010
)(0.009
)(0.014
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.021
‐0.028
‐0.013
‐0.010
(0.023
)(0.023
)(0.021
)(0.029
)
log(Interm
ediate inpu
ts per worker)
0.39
5***
0.40
0***
0.40
6***
0.37
5***
0.37
5***
(0.002
)(0.002
)(0.002
)(0.002
)(0.012
)log(capital per worker)
0.11
5***
0.11
4***
0.11
0***
0.13
3***
0.13
2***
(0.001
)(0.001
)(0.001
)(0.001
)(0.007
)log(nu
mbe
r of w
orkers)
0.05
3***
0.05
2***
0.06
7***
0.05
3***
0.05
2***
(0.001
)(0.001
)(0.001
)(0.001
)(0.005
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns13
3,08
713
3,08
713
2,84
213
2,84
213
2,84
2R‐squared
0.52
80.53
10.53
80.62
10.47
0Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker in the firm. See
text fo
r more de
tails. R
obust
standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
num
ber o
f clusters is
975. ***
p<0.01, ** p<
0.05
, * p<0.1.
108
TABLE C5
.5: W
ages in
Differen
t Types of N
ew Firm
s and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)5.33
0***
5.30
4***
4.58
0***
5.21
2***
4.68
1***
4.42
1***
5.03
8***
4.49
5***
(0.000
)(0.007
)(0.010
)(0.022
)(0.003
)(0.011
)(0.022
)(0.005
)
Positive cluster g
rowth (β
2)0.07
2***
0.02
8***
0.01
3***
0.01
10.00
1(0.020
)(0.009
)(0.005
)(0.007
)(0.005
)Firm
types:
Spin‐off (β
31)
‐0.180
***
‐0.176
***
‐0.104
***
‐0.061
***
‐0.059
***
‐0.015
***
‐0.003
‐0.004
*(0.003
)(0.009
)(0.006
)(0.005
)(0.002
)(0.005
)(0.005
)(0.002
)Entrep
rene
urial start‐up (β
32)
‐0.147
***
‐0.141
***
‐0.108
***
‐0.080
***
‐0.077
***
‐0.015
***
‐0.015
***
‐0.014
***
(0.003
)(0.009
)(0.008
)(0.006
)(0.002
)(0.006
)(0.005
)(0.002
)Other new
firm
(β33)
0.00
50.00
40.00
30.00
70.00
6**
0.06
5***
0.05
0***
0.04
8***
(0.003
)(0.020
)(0.013
)(0.010
)(0.003
)(0.012
)(0.009
)(0.003
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)‐0.003
‐0.024
**‐0.028
***
‐0.024
***
0.02
6***
0.00
40.00
5(0.022
)(0.012
)(0.009
)(0.003
)(0.010
)(0.008
)(0.003
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
‐0.023
‐0.021
‐0.037
***
‐0.032
***
0.02
0*‐0.011
‐0.008
***
(0.022
)(0.015
)(0.011
)(0.003
)(0.011
)(0.009
)(0.003
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.006
‐0.017
‐0.015
‐0.006
**0.02
10.00
00.00
6*(0.031
)(0.013
)(0.013
)(0.003
)(0.013
)(0.012
)(0.003
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,17
0,48
91,17
0,48
91,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
8R‐squared
0.00
60.01
20.36
00.41
30.28
10.38
10.42
30.29
2Notes: The
dep
ende
nt variable is the log of th
e ho
urly wage. The
sample is redu
ced to aroun
d 1,17
0,00
0 individu
als a
nd th
e nu
mbe
r of clusters to 10
53, because
reliable wage inform
ation cann
ot be ob
tained
for a
ll workers. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
ns 5 and
8
are compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
109
TABLE C5
.6: Skills in
Differen
t Types of N
ew Firm
s and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)0.06
9***
0.04
3***
‐0.031
***
0.15
50.03
5***
‐0.003
0.16
3*0.04
1***
(0.003
)(0.002
)(0.003
)(0.097
)(0.001
)(0.004
)(0.097
)(0.001
)
Positive cluster g
rowth (β
2)0.06
7***
0.07
4***
0.00
5**
0.07
3***
0.00
3(0.009
)(0.007
)(0.002
)(0.006
)(0.002
)Firm
types:
Spin‐off (β
31)
‐0.014
***
‐0.014
***
‐0.011
***
‐0.004
**‐0.003
***
‐0.008
***
‐0.001
‐0.002
(0.003
)(0.002
)(0.002
)(0.001
)(0.001
)(0.002
)(0.002
)(0.001
)Entrep
rene
urial start‐up (β
32)
0.00
3‐0.002
‐0.003
‐0.006
***
‐0.004
***
‐0.003
‐0.004
**‐0.004
***
(0.003
)(0.003
)(0.002
)(0.002
)(0.001
)(0.003
)(0.002
)(0.001
)Other new
firm
(β33)
0.01
9***
0.00
60.00
60.00
30.00
4**
0.01
3*0.00
60.00
7***
(0.005
)(0.007
)(0.006
)(0.005
)(0.002
)(0.007
)(0.005
)(0.002
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.01
20.01
1‐0.001
0.00
10.01
6**
0.00
30.00
4**
(0.010
)(0.009
)(0.004
)(0.002
)(0.007
)(0.004
)(0.002
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.00
40.00
2‐0.009
**‐0.008
***
0.00
1‐0.007
*‐0.007
***
(0.010
)(0.008
)(0.004
)(0.001
)(0.007
)(0.003
)(0.001
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.02
5*0.02
2**
‐0.003
‐0.005
***
0.02
7**
‐0.000
‐0.003
(0.013
)(0.011
)(0.006
)(0.002
)(0.011
)(0.005
)(0.002
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,56
0,45
21,56
0,45
21,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
0R‐squared
0.00
00.01
80.04
30.14
40.01
40.04
70.14
50.01
4Notes: The
mod
el estim
ated
is a line
ar probability mod
el and
the de
pend
ent v
ariable is a du
mmy for w
hether th
e pe
rson
has com
pleted
tertiary edu
catio
n. The
nu
mbe
r of clusters is 1
081, since some clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐
squared in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
33
AppendixD.EstimationResults forAlternativeDefinitionofClusters(77NACE2industries).
112
TABLE D2
.1.: Num
ber o
f clusters a
nd jo
bs. B
y cluster g
rowth
Num
ber o
f clusters
Num
ber o
f job
sNum
ber o
f firm
s
Cluster g
rowth
#%
#%
#%
‐0.50‐
338.80
3351
92.15
3091
1.64
‐0.50;‐0.25
7219
.20
1192
587.64
1059
05.63
‐0.25;‐0.1
8622
.93
5047
0732
.34
5884
831
.29
‐0.1‐0.1
105
28.00
7412
8047
.50
8864
747
.14
0.1‐0.25
349.07
7672
94.92
1731
69.21
0.25‐0.5
174.53
4824
83.09
5609
2.98
0.5+
287.47
3671
12.35
3943
2.10
Totals
375
100.00
1,56
0,45
210
0.00
188,04
410
0.00
113
TABLE D2
.2: N
umbe
r of firm
type
s and
jobs. B
y cluster g
rowth and
firm
type
sNum
ber o
f firm
s
Establish
ed firm
sSpin‐offs
Other new
firm
sCluster g
rowth
#%
#%
#%
#%
‐0.50‐
1705
1.6
349
1.5
801
1.7
236
2.6
‐0.50;‐0.25
6352
5.9
758
3.2
2,97
16.4
509
5.6
‐0.25;‐0.1
3476
332
.110
,020
41.7
11,933
25.6
2132
23.5
‐0.1‐0.1
5267
348
.610
,019
41.7
21,324
45.7
4631
51.1
0.1‐0.25
8428
7.8
1,73
07.2
6,21
413
.394
410
.4 0.25‐0.5
2520
2.3
921
3.8
1,76
83.8
400
4.4
0.5+
1891
1.7
220
0.9
1,62
63.5
206
2.3
Totals
108,33
2
100.0
24,017
10
0.0
46,637
10
0.0
9,05
8
100.0
Num
ber o
f job
s
Establish
ed firm
sSpin‐offs
Other new
firm
sCluster g
rowth
#%
#%
#%
#%
‐0.50‐
2752
22.1
1,65
32.0
2,62
42.3
1720
3.1
‐0.50;‐0.25
1038
487.9
3,91
84.8
8,36
57.4
3127
5.6
‐0.25;‐0.1
4292
5532
.830
,502
37.1
30,893
27.2
1405
725
.3 ‐0
.1‐0.1
6249
9947
.738
,141
46.4
50,987
44.9
2715
348
.8 0.1‐0.25
5590
54.3
4,75
75.8
10,559
9.3
5508
9.9
0.25‐0.5
3818
72.9
2,57
13.1
4,90
04.3
2590
4.7
0.5+
2928
62.2
624
0.8
5,32
84.7
1473
2.6
Totals
1,30
9,00
2
100.0
82,166
10
0.0
113,65
6
100.0
55,628
100.0
Entrep
rene
urial
Entrep
rene
urial
114
TABLE D2
.3: A
verage
num
ber o
f job
s per firm
. By cluster g
rowth after firm
type
sJobs per firm
Establish
ed firm
sSpin‐offs
Entrep
rene
urial
start‐up
sOther new
firms
Cluster g
rowth
##
##
‐0.50‐
12.8
4.5
2.9
5.9
‐0.50;‐0.25
16.4
3.3
2.8
6.9
‐0.25;‐0.1
10.9
3.0
2.4
6.0
‐0.1‐0.1
11.9
3.6
2.4
5.7
0.1‐0.25
8.6
3.3
2.2
6.1
0.25‐0.5
13.1
2.8
2.1
7.3
0.5+
20.0
2.9
3.7
8.1
Totals
12.1
3.3
2.5
6.1
115
TABLE D2
.4: A
verage
individu
al cha
racteristics. By firm ty
pe and
cluster growth
Cluster g
rowth
Firm
type
Num
ber o
f workers
Num
ber o
f firms
Mean age
Share of
females
Share of
highly
educated
Mean ho
urly
wage (DKK
)
Negative
Establish
ed firm
s84
5,38
6
70,119
40.96
0.31
0.04
214.48
Negative
Spin‐offs
54,3
80
16,1
96
36.5
40.
270.
0217
3.67
Negative
Enterprene
urial start‐ups
65,5
72
25,1
46
37.6
60.
280.
0317
9.58
Negative
Other new
firm
s31
,179
5,24
7
38.27
0.31
0.06
219.36
Positive
Establish
ed firm
s46
3,61
6
38,213
37.37
0.45
0.12
220.05
Positive
Spin‐offs
27,7
86
7,82
1
35.1
40.
440.
1218
6.63
Positive
Enterprene
urial start‐ups
48,0
84
21,4
91
37.4
50.
430.
1218
3.28
Positive
Other new
firm
s24
,449
3,81
1
36.49
0.43
0.15
221.40
116
TABLE D4
.1: A
verage
Wages in
New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)5.10
6***
5.12
2***
4.74
1***
5.97
9***
4.89
4***
(0.002
)(0.002
)(0.009
)(0.089
)(0.032
)
Positive cluster g
rowth (β
2)‐0.050
***
‐0.043
***
‐0.027
***
(0.005
)(0.005
)(0.006
)
New
firm
(β3)
0.02
2***
‐0.010
***
0.01
5***
0.00
7**
0.00
7(0.003
)(0.004
)(0.004
)(0.003
)(0.006
)
Positive cluster g
rowth x New
firm
(β4)
0.09
6***
0.06
2***
0.01
3*0.01
5(0.008
)(0.007
)(0.007
)(0.009
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns95
,228
95,228
95,228
95,228
95,228
R‐squared
0.00
00.00
30.13
30.25
20.03
9Notes: The
dep
ende
nt variable is the log of th
e average ho
urly wage in th
e firm. The
sample is redu
ced to
arou
nd 95,00
0 firms a
nd th
e nu
mbe
r of clusters to 37
3, because so
me firms d
o no
t have em
ployees a
nd
because reliable wage inform
ation cann
ot be ob
tained
for a
ll firms w
ith employees. See
text fo
r more de
tails.
Robu
st standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<
0.01
, ** p<
0.05
, * p<0.1.
117
TABLE D4
.2: Skill‐Intensity
in New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)0.07
1***
0.02
9***
‐0.078
***
0.14
80.02
9**
(0.001
)(0.001
)(0.002
)(0.118
)(0.012
)
Positive cluster g
rowth (β
2)0.12
0***
0.12
0***
0.01
8***
(0.002
)(0.002
)(0.002
)
New
firm
(β3)
0.02
0***
0.00
4***
0.01
8***
0.00
9***
0.01
0***
(0.001
)(0.001
)(0.001
)(0.001
)(0.002
)
Positive cluster g
rowth x New
firm
(β4)
0.02
1***
0.01
8***
0.00
0‐0.002
(0.003
)(0.003
)(0.003
)(0.004
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns18
8,04
418
8,04
418
7,63
318
7,63
318
7,63
3R‐squared
0.00
20.06
80.07
90.19
40.00
3Notes: The
dep
ende
nt variable is the the share of employees w
ith a te
rtiary edu
catio
n de
gree
in th
e firm.
The nu
mbe
r of clusters is 3
75, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
118
TABLE D4
.3: Sales per W
orker in New
and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)13
.404
***
13.495
***
13.764
***
18.156
***
13.554
***
(0.003
)(0.003
)(0.009
)(0.848
)(0.036
)
Positive cluster g
rowth (β
2)‐0.264
***
‐0.207
***
‐0.000
(0.006
)(0.006
)(0.008
)
New
firm
(β3)
‐0.158
***
‐0.146
***
‐0.173
***
‐0.128
***
‐0.108
***
(0.004
)(0.005
)(0.005
)(0.005
)(0.009
)
Positive cluster g
rowth x New
firm
(β4)
0.01
10.01
5*‐0.016
*‐0.000
(0.008
)(0.008
)(0.008
)(0.015
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns17
8,15
217
8,15
217
7,79
918
1,42
617
7,79
9R‐squared
0.00
80.02
90.04
90.27
70.00
9Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker. The nu
mbe
r of clusters is 3
52, since
some clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
119
TABLE D4
.4: P
rodu
ctivity
in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)7.05
1***
7.00
9***
6.81
6***
6.95
6***
7.09
1***
(0.022
)(0.022
)(0.024
)(0.056
)(0.123
)
Positive cluster g
rowth (β
2)0.01
2***
0.00
1‐0.022
***
(0.004
)(0.005
)(0.006
)
New
firm
(β3)
0.01
5***
‐0.011
***
0.01
7***
0.01
7***
0.01
6*(0.003
)(0.004
)(0.004
)(0.004
)(0.009
)
Positive cluster g
rowth x New
firm
(β4)
0.07
2***
0.05
8***
0.01
3*0.01
5(0.007
)(0.007
)(0.007
)(0.021
)
log(Interm
ediate inpu
ts per worker)
0.39
7***
0.40
0***
0.40
7***
0.38
1***
0.38
1***
(0.002
)(0.002
)(0.002
)(0.002
)(0.015
)log(capital per worker)
0.11
7***
0.11
7***
0.11
1***
0.12
6***
0.12
6***
(0.001
)(0.001
)(0.001
)(0.001
)(0.007
)log(nu
mbe
r of w
orkers)
0.05
9***
0.05
8***
0.07
5***
0.06
9***
0.06
8***
(0.001
)(0.001
)(0.001
)(0.001
)(0.007
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns13
3,08
713
3,08
713
2,84
213
2,84
213
2,84
2R‐squared
0.52
70.52
80.53
60.60
50.47
6Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker in the firm. See
text fo
r more de
tails.
Robu
st standard errors in parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
nu
mbe
r of clusters is 3
26. ***
p<0.01, ** p<
0.05
, * p<0.1.
120
TABLE D4
.4A: In
term
ediate In
puts per W
orker in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)12
.617
***12
.777
***13
.423
***13
.760
***13
.013
***
(0.005
)(0.005
)(0.016
)(0.111
)(0.056
)
Positive cluster g
rowth (β
2)‐0.511
***‐0.415
***‐0.042
***
(0.010
)(0.010
)(0.012
)
New
firm
(β3)
‐0.293
***‐0.214
***‐0.288
***‐0.155
***‐0.156
***
(0.007
)(0.008
)(0.009
)(0.007
)(0.017
)
Positive cluster g
rowth x New
firm
(β4)
‐0.137
***‐0.096
***
‐0.019
‐0.015
(0.016
)(0.016
)(0.013
)(0.029
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns14
3,86
614
3,86
614
3,60
014
3,60
014
3,60
0R‐squared
0.01
10.05
10.07
80.31
80.01
5Notes: The
dep
ende
nt variable is the log of interm
ediate inpu
ts per worker in the firm. See
text
for m
ore de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is
compu
ted on
the de
‐meane
d data. The
num
ber o
f clusters is 3
26. ***
p<0.01, ** p<
0.05
, *
p<0.1.
121
TABLE D4
.4B: Cap
ital per W
orker in New
and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)12
.168
***12
.229
***10
.658
***13
.627
***11
.600
***
(0.006
)(0.007
)(0.020
)(0.110
)(0.053
)
Positive cluster g
rowth (β
2)‐0.202
***‐0.147
***‐0.064
***
(0.013
)(0.013
)(0.015
)
New
firm
(β3)
‐0.438
***‐0.491
***‐0.273
***‐0.197
***‐0.200
***
(0.009
)(0.011
)(0.011
)(0.009
)(0.023
)
Positive cluster g
rowth x New
firm
(β4)
0.17
8***
0.11
7***
0.05
6***
0.06
7(0.019
)(0.018
)(0.016
)(0.045
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns14
7,21
014
7,21
014
6,91
514
6,91
514
6,91
5R‐squared
0.01
60.01
70.07
40.35
40.01
9Notes: The
dep
ende
nt variable is the log of capita
lper worker in the firm. See
text fo
r more
details. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the
de‐m
eane
d data. The
num
ber o
f clusters is 3
27. ***
p<0.01, ** p<
0.05
, * p<0.1.
122
TABLE D4
.5: W
ages in
New
and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)5.33
4***
5.32
1***
4.58
6***
5.19
2***
4.67
5***
4.41
1***
5.01
7***
4.47
8***
(0.000
)(0.007
)(0.010
)(0.023
)(0.003
)(0.011
)(0.022
)(0.005
)
Positive cluster g
rowth (β
2)0.04
0*0.01
30.02
2***
0.00
90.01
0(0.022
)(0.010
)(0.007
)(0.007
)(0.006
)
New
firm
(β3)
‐0.098
***
‐0.112
***
‐0.067
***
‐0.040
***
‐0.040
***
0.02
3***
0.01
7***
0.01
7***
(0.001
)(0.008
)(0.006
)(0.004
)(0.001
)(0.004
)(0.004
)(0.001
)
Positive cluster g
rowth x New
firm
(β4)
0.03
40.00
2‐0.013
‐0.010
***
0.00
1‐0.011
‐0.009
***
(0.024
)(0.012
)(0.008
)(0.002
)(0.009
)(0.007
)(0.002
)
Controls for g
ende
r, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,17
0,48
91,17
0,48
91,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
8R‐squared
0.00
50.00
70.35
90.40
70.28
50.38
10.41
70.29
7Notes: The
dep
ende
nt variable is the log of th
e ho
urly wage. The
sample is redu
ced to aroun
d 1,17
0,00
0 individu
als a
nd th
e nu
mbe
r of clusters to 37
3,
because reliable wage inform
ation cann
ot be ob
tained
for a
ll workers. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
123
TABLE D4
.6: Skills in
New
and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)0.06
9***
0.04
4***
‐0.031
***
0.15
50.03
5***
‐0.006
0.16
1*0.04
0***
(0.003
)(0.002
)(0.003
)(0.097
)(0.001
)(0.004
)(0.097
)(0.001
)
Positive cluster g
rowth (β
2)0.07
2***
0.08
5***
0.00
7**
0.08
6***
0.00
6**
(0.010
)(0.008
)(0.003
)(0.007
)(0.003
)
New
firm
(β3)
0.00
1‐0.010
***
‐0.009
***
‐0.002
‐0.001
*0.00
00.00
10.00
2**
(0.004
)(0.002
)(0.002
)(0.001
)(0.001
)(0.002
)(0.002
)(0.001
)
Positive cluster g
rowth x New
firm
(β4)
0.02
0*0.01
6*‐0.003
‐0.004
***
0.01
2*‐0.002
‐0.003
***
(0.011
)(0.008
)(0.003
)(0.001
)(0.007
)(0.003
)(0.001
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,56
0,45
21,56
0,45
21,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
0R‐squared
0.00
00.02
00.04
90.13
20.01
40.05
40.13
20.01
4Notes: The
mod
el estim
ated
is a line
ar probability mod
el and
the de
pend
ent v
ariable is a du
mmy for w
hether th
e pe
rson
has com
pleted
tertiary
education. The
num
ber o
f clusters is 3
75, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard
errors in
paren
theses. The
R‐squ
ared
in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
124
TABLE D5
.1: A
verage
Wages in
Differen
t Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)5.10
6***
5.12
2***
4.74
5***
5.97
9***
4.89
3***
(0.002
)(0.002
)(0.009
)(0.089
)(0.033
)
Positive cluster g
rowth (β
2)‐0.050
***
‐0.042
***
‐0.026
***
(0.005
)(0.005
)(0.006
)
Firm
types:
Spin‐off (β
31)
‐0.006
‐0.043
***
‐0.009
*‐0.001
‐0.002
(0.005
)(0.005
)(0.005
)(0.005
)(0.009
)Entrep
rene
urial start‐up (β
32)
‐0.015
***
‐0.046
***
‐0.022
***
‐0.029
***
‐0.029
***
(0.005
)(0.005
)(0.005
)(0.005
)(0.008
)Other new
firm
(β33)
0.17
9***
0.17
0***
0.17
8***
0.12
6***
0.12
5***
(0.007
)(0.009
)(0.008
)(0.008
)(0.011
)
Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.11
5***
0.07
6***
0.02
7***
0.03
0**
(0.011
)(0.010
)(0.009
)(0.014
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.09
1***
0.06
3***
0.01
20.01
5(0.011
)(0.010
)(0.009
)(0.012
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.03
3**
0.00
3‐0.031
**‐0.031
*(0.015
)(0.014
)(0.012
)(0.018
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns95
,228
95,228
95,228
95,228
95,228
R‐squared
0.00
80.01
00.14
00.25
60.04
3Notes: The
dep
ende
nt variable is the log of th
e average ho
urly wage in th
e firm. The
sample is redu
ced to aroun
d 95
,000
firms a
nd th
e nu
mbe
r of clusters to 37
3, because so
me firms d
o no
t have em
ployees a
nd because re
liable wage
inform
ation cann
ot be ob
tained
for a
ll firms w
ith employees. See
text fo
r more de
tails. R
obust stand
ard errors in
parenthe
ses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
125
TABLE D5
.2: Skill‐Intensity
in Differen
t Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)0.07
1***
0.02
9***
‐0.078
***
0.14
70.02
8**
(0.001
)(0.001
)(0.002
)(0.118
)(0.013
)
Positive cluster g
rowth (β
2)0.12
0***
0.12
0***
0.01
7***
(0.002
)(0.002
)(0.002
)Firm
types:
Spin‐off (β
31)
0.00
6***
‐0.007
***
0.01
1***
0.00
8***
0.00
9***
(0.002
)(0.001
)(0.001
)(0.001
)(0.003
)Entrep
rene
urial start‐up (β
32)
0.02
2***
0.00
4***
0.01
7***
0.00
6***
0.00
8***
(0.001
)(0.001
)(0.001
)(0.001
)(0.002
)Other new
firm
(β33)
0.04
1***
0.03
4***
0.04
8***
0.02
7***
0.02
7***
(0.003
)(0.003
)(0.003
)(0.003
)(0.004
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.05
0***
0.04
8***
0.03
1***
0.02
7**
(0.005
)(0.005
)(0.004
)(0.011
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.01
2***
0.00
9***
‐0.011
***
‐0.013
**(0.003
)(0.003
)(0.003
)(0.005
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.004
‐0.002
0.00
1‐0.001
(0.006
)(0.006
)(0.005
)(0.008
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns18
8,04
418
8,04
418
7,63
318
7,63
318
7,63
3R‐squared
0.00
20.06
90.08
00.19
50.00
5Notes: The
dep
ende
nt variable is the the share of employees w
ith a te
rtiary edu
catio
n de
gree
in th
e firm. The
num
ber o
f clusters is 375
, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard
errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
126
TABLE D5
.3: Sales per W
orker in Diffe
rent Types of N
ew Firm
s and
Old Firm
s, Firm
Level, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)13
.404
***
13.495
***
13.749
***
14.119
***
13.542
***
(0.003
)(0.003
)(0.009
)(0.076
)(0.037
)
Positive cluster g
rowth (β
2)‐0.264
***
‐0.206
***
‐0.008
(0.006
)(0.006
)(0.007
)Firm
types:
Spin‐off (β
31)
‐0.046
***
‐0.091
***
‐0.128
***
‐0.053
***
‐0.059
***
(0.006
)(0.007
)(0.007
)(0.006
)(0.012
)Entrep
rene
urial start‐up (β
32)
‐0.270
***
‐0.242
***
‐0.264
***
‐0.169
***
‐0.172
***
(0.005
)(0.006
)(0.007
)(0.006
)(0.011
)Other new
firm
(β33)
0.12
7***
0.15
9***
0.15
1***
0.05
7***
0.05
6**
(0.012
)(0.016
)(0.016
)(0.013
)(0.025
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.12
1***
0.12
1***
0.04
9***
0.06
2***
(0.013
)(0.012
)(0.011
)(0.023
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.00
20.00
9‐0.014
‐0.003
(0.010
)(0.010
)(0.009
)(0.016
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.033
‐0.060
**‐0.013
‐0.020
(0.023
)(0.023
)(0.021
)(0.052
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns17
8,15
217
8,15
217
7,79
917
7,79
917
7,79
9R‐squared
0.01
90.03
90.05
90.28
80.01
4Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker. The nu
mbe
r of clusters is 3
52, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in
column 5 is compu
ted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
127
TABLE D5
.4: P
rodu
ctivity
in Differen
t Types of N
ew Firm
s and
Old Firm
s, 201
0(1)
(2)
(3)
(4)
(5)
Constant (β
1)7.06
8***
7.02
4***
6.82
7***
6.96
3***
7.09
5***
(0.022
)(0.023
)(0.024
)(0.056
)(0.123
)
Positive cluster g
rowth (β
2)0.01
1**
0.00
0‐0.024
***
(0.004
)(0.005
)(0.006
)Firm
types:
Spin‐off (β
31)
0.06
2***
0.02
5***
0.05
4***
0.05
6***
0.05
4***
(0.005
)(0.005
)(0.005
)(0.005
)(0.011
)Entrep
rene
urial start‐up (β
32)
‐0.021
***
‐0.055
***
‐0.025
***
‐0.015
***
‐0.016
*(0.004
)(0.005
)(0.005
)(0.005
)(0.009
)Other new
firm
(β33)
0.05
3***
0.07
3***
0.09
0***
0.04
5***
0.04
4**
(0.009
)(0.012
)(0.012
)(0.011
)(0.018
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.12
5***
0.10
4***
0.05
4***
0.05
5(0.011
)(0.011
)(0.010
)(0.034
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.08
4***
0.07
2***
0.01
6**
0.01
9(0.009
)(0.008
)(0.008
)(0.018
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.054
***
‐0.062
***
‐0.050
***
‐0.049
(0.019
)(0.019
)(0.017
)(0.032
)
log(Interm
ediate inpu
ts per worker)
0.39
5***
0.39
9***
0.40
6***
0.38
0***
0.38
0***
(0.002
)(0.002
)(0.002
)(0.002
)(0.015
)log(capital per worker)
0.11
7***
0.11
7***
0.11
1***
0.12
6***
0.12
6***
(0.001
)(0.001
)(0.001
)(0.001
)(0.007
)log(nu
mbe
r of w
orkers)
0.05
6***
0.05
6***
0.07
3***
0.06
8***
0.06
6***
(0.001
)(0.001
)(0.001
)(0.001
)(0.007
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
Observatio
ns13
3,08
713
3,08
713
2,84
213
2,84
213
2,84
2R‐squared
0.52
80.53
00.53
70.60
60.47
7Notes: The
dep
ende
nt variable is the log of sa
les (revenu
e) per worker in the firm. See
text fo
r more de
tails. R
obust stand
ard
errors in
paren
theses. The
R‐squ
ared
in colum
n 5 is compu
ted on
the de
‐meane
d data. The
num
ber o
f clusters is 3
26. ***
p<0.01,
** p<0.05, * p<0.1.
128
TABLE D5
.5: W
ages in
Differen
t Types of N
ew Firm
s and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)5.33
4***
5.32
1***
4.58
7***
5.19
4***
4.67
6***
4.41
6***
5.02
2***
4.48
2***
(0.000
)(0.007
)(0.010
)(0.023
)(0.003
)(0.011
)(0.022
)(0.005
)
Positive cluster g
rowth (β
2)0.04
0*0.01
40.02
2***
0.01
00.01
0(0.022
)(0.010
)(0.007
)(0.008
)(0.006
)Firm
types:
Spin‐off (β
31)
‐0.155
***
‐0.172
***
‐0.096
***
‐0.056
***
‐0.057
***
0.00
10.00
30.00
1(0.002
)(0.009
)(0.006
)(0.005
)(0.002
)(0.005
)(0.004
)(0.002
)Entrep
rene
urial start‐up (β
32)
‐0.118
***
‐0.129
***
‐0.087
***
‐0.063
***
‐0.062
***
0.00
90.00
20.00
2(0.002
)(0.010
)(0.009
)(0.006
)(0.002
)(0.006
)(0.005
)(0.002
)Other new
firm
(β33)
0.00
6**
0.00
90.00
80.01
9**
0.01
9***
0.07
3***
0.05
8***
0.05
9***
(0.002
)(0.014
)(0.010
)(0.008
)(0.002
)(0.010
)(0.008
)(0.002
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.05
6**
0.01
1‐0.005
‐0.002
0.01
20.00
10.00
3(0.026
)(0.014
)(0.010
)(0.004
)(0.011
)(0.009
)(0.004
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.02
6‐0.004
‐0.019
‐0.017
***
‐0.001
‐0.013
‐0.012
***
(0.026
)(0.016
)(0.012
)(0.003
)(0.011
)(0.010
)(0.003
)Po
sitive cluster g
rowth x Other new
firm
(β43)
‐0.017
‐0.022
‐0.028
**‐0.024
***
‐0.022
‐0.030
**‐0.027
***
(0.033
)(0.016
)(0.014
)(0.004
)(0.016
)(0.012
)(0.004
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,17
0,48
91,17
0,48
91,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
81,14
4,58
8R‐squared
0.00
80.01
00.36
00.40
80.28
60.38
10.41
70.29
8Notes: The
dep
ende
nt variable is the log of th
e ho
urly wage. The
sample is redu
ced to aroun
d 1,17
0,00
0 individu
als a
nd th
e nu
mbe
r of clusters to 37
3, because re
liable
wage inform
ation cann
ot be ob
tained
for a
ll workers. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐squ
ared
in colum
ns 5 and
8 are com
puted
on th
e de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.
129
TABLE D5
.6: Skills in
Differen
t Types of N
ew Firm
s and
Old Firm
s, W
orker Level, 201
0(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constant (β
1)0.06
9***
0.04
4***
‐0.035
***
0.14
80.03
3***
‐0.013
***
0.15
40.03
7***
(0.003
)(0.002
)(0.003
)(0.096
)(0.001
)(0.005
)(0.097
)(0.001
)
Positive cluster g
rowth (β
2)0.07
2***
0.08
5***
0.00
7**
0.08
6***
0.00
6**
(0.010
)(0.008
)(0.003
)(0.007
)(0.003
)Firm
types:
Spin‐off (β
31)
‐0.016
***
‐0.024
***
‐0.020
***
‐0.007
***
‐0.006
***
‐0.011
***
‐0.004
***
‐0.004
***
(0.004
)(0.002
)(0.002
)(0.001
)(0.001
)(0.002
)(0.001
)(0.001
)Entrep
rene
urial start‐up (β
32)
0.00
0‐0.010
***
‐0.009
***
‐0.004
**‐0.003
***
‐0.003
‐0.002
‐0.001
(0.004
)(0.003
)(0.003
)(0.002
)(0.001
)(0.003
)(0.002
)(0.001
)Other new
firm
(β33)
0.02
7***
0.01
2**
0.01
2**
0.01
0**
0.01
1***
0.02
3***
0.01
4***
0.01
4***
(0.005
)(0.005
)(0.005
)(0.004
)(0.001
)(0.005
)(0.004
)(0.001
)Firm
type
x Positive cluster g
rowth:
Positive cluster g
rowth x Spin‐off (β 4
1)0.02
6**
0.02
4***
0.00
8**
0.00
8***
0.02
2***
0.01
0**
0.00
9***
(0.011
)(0.009
)(0.004
)(0.002
)(0.008
)(0.004
)(0.002
)Po
sitive cluster g
rowth x Entrepren
euria
l start‐up (β
42)
0.01
30.00
7‐0.010
***
‐0.011
***
0.00
2‐0.010
**‐0.012
***
(0.011
)(0.009
)(0.004
)(0.002
)(0.008
)(0.004
)(0.002
)Po
sitive cluster g
rowth x Other new
firm
(β43)
0.02
00.01
6‐0.003
‐0.005
**0.01
3‐0.002
‐0.005
**(0.014
)(0.012
)(0.006
)(0.002
)(0.012
)(0.006
)(0.002
)
Control for gen
der, age and ed
ucation
nono
yes
yes
yes
yes
yes
yes
Indu
stry and
Region fixed
effe
cts
nono
noyes
nono
yes
noIndu
stry x Region (= cluster) fixed
effe
cts
nono
nono
yes
nono
yes
Control for firm
size
nono
nono
noyes
yes
yes
Observatio
ns1,56
0,45
21,56
0,45
21,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
01,52
0,01
0R‐squared
0.00
10.02
10.04
90.13
20.01
40.05
40.13
20.01
4Notes: The
mod
el estim
ated
is a line
ar probability mod
el and
the de
pend
ent v
ariable is a du
mmy for w
hether th
e pe
rson
has com
pleted
tertiary edu
catio
n. The
nu
mbe
r of clusters is 3
75, since so
me clusteres d
o no
t con
tain any
observatio
ns in
201
0. See
text fo
r more de
tails. R
obust stand
ard errors in
paren
theses. The
R‐
squared in colum
ns 5 and
8 are com
puted on
the de
‐meane
d data. ***
p<0.01, ** p<
0.05
, * p<0.1.