UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Uddannelses- og Forskningsudvalget 2015-16
UFU Alm.del
Offentligt

Econometric Analysis of the

Danish Innovation Incubator

Programme

June 2015

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Published

Danish Agency for Science, Technology and

Innovation

Bredgade 40

1260 København K

Telefon: +45 3544 6200

E-mail: [email protected]

www.ufm.dk

Layout

Tobias Gebetsberger

Publication can be downloaded at ufm.dk/publikationer

ISBN (electronic publication):

978-87-93151-60-4

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Contents

Executive summary

Background

Analytical focus of the analysis

Earlier highly related analyses

The survival and growth of IM-firms

Mobility of individuals

Scope and limitations of the analysis

Data limitations

Limitations of the treatment-control analysis design

Main findings

Sammenfatning

Analysens fokus

Tidligere resultater

Analysedesign

1. Introduction

1.1 A short description of the Danish Innovation Incubator Programme

(Innovationsmiljøordningen)

1.2 Theoretical considerations, earlier research and related studies

1.2.1 General challenges of venture capital finance

1.2.2 Earlier results on Danish venture capital finance

1.2.3 Measurement issues of venture capital’s potential growth effects

1.2.4 Earlier studies on aggregated data

1.2.5 Firm-level comparison studies - general issues

1.2.6 Earlier results of firm-level comparison studies

1.2.7 Potential knowledge spill-overs of high-tech firms

1.2.8 There is no convergence in the literature on the effects of venture

capital

2. Data

2.1 The Statistics Denmark data

2.2 The Experian data

2.3 The presence of IM-firms in the Statistics Denmark register data

2.4 The treatment of the raw data prior to the analysis

2.5 Data limitations

2.6 Basic descriptive statistics

3. Survival and growth of IM-firms in comparison with a group of reference

firms

3.1 Selection of reference firms for the subsequent comparisons

3.2 Survival

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

3.3 Employment growth

3.4 Turnover developments

3.5 Value added developments

3.6 Annual Earnings

4. Individual mobility and generation of start-ups

4.1 Data on individuals in the Statistics Denmark Databases

4.2 Job mobility and firm creation

4.3 New firm creation by the staff of IM-firms

4.4 The number of IM-firm ‘spin-offs’

4.5 Individual mobility in association with firm exit

4.6 Emigration decisions of IM-staff

4.7 Short summary of the analysis of individual mobility

5. References

6. Appendix

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Executive summary

Background

Externalities to innovation often leave innovators underpaid relative to the value of their inven-

tions, and innovations are often not realized in established organisations but in new firms. This

creates a rationale for public investments in small innovative firms, which is reflected in almost

every OECD country having one or several support programmes for small innovative firms. Ex-

amples include the U.S.

Small Business Innovation Research

(SBIR)-programme or the Europe-

an-level

EUROSTARS-programme.

In Denmark, public investments in small innovative firms take different forms. One of the

programmes to support small new innovative firms is the Danish Innovation Incubator Pro-

gramme (IM-programme,

Innovationsmiljøordningen

in Danish). This programme, which is

the subject of the present report, was established in 1998, and is under the Danish Agency for

Science, Technology and Innovation (DASTI). The programme has a decentralized structure, as

it is implemented with a number of different incubators financing newly started innovative

firms.

The innovation incubators,

Innovationsmiljøer

in Danish, in the following IMs, provide

loans and equity finance to newly started firms. The IMs operate in the very early stages of the

commercialization process where the risk is high and private investors are reluctant to engage.

Until now (2013/2014), approximately 1,000 firms have been found eligible for a total support

volume of approximately DKK1.75 billion (€250 million).

Analytical focus of the analysis

The main objective of the IM-programme is to bridge the funding gap in the early and most risky

stages of the venture market, by supporting the creation of new innovative firms. As part of its

ongoing monitoring of the goal attainment of the IMs, DASTI performs a yearly benchmark

analysis on a number of selected key performance indicators.

In addition to the benchmark

analyses, DASTI has initiated this present report.

The purpose of the report is to investigate to

what extent the IMs increase the economic performance of the firms in the IM-programme. This

question is addressed by analyzing the survival and growth of firms that participate in the IM-

programme. Also, the report establishes evidence of whether there is indirect firm creation activ-

ity related to the IM-programme. This potential indirect effect of the IM-programme would re-

main unnoticed in a standard evaluation analysis.

From 2014 to 2017 the programme is operated by four incubators, with main offices at the major Danish cities.

BOREAN Innovation (Aalborg), Pre-Seed Innovation (Kgs. Lyngby), Syddansk Teknologisk Innovation (Odense) ,

and CAPNOVA (Aarhus and Roskilde). For further

information see http://ufm.dk/forskning-og-

innovation/samspil-mellem-viden-og-innovation/fa-hjaelp-til-kommercialisering/innovationsmiljoer.

For further information

see http://ufm.dk/forskning-og-innovation/samspil-mellem-viden-og-innovation/fa-

hjaelp-til-kommercialisering/innovationsmiljoer/tal-om-innovationsmiljoer-1.

The present analysis and report is by Johan Kuhn, PhD, who has consulted different Danish and European-level

organizations on data-based evaluation practice, and has done a set of similar analyses for the DASTI and the

Danish Ministry for Business Affairs.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

In sum, the present study considers the survival and growth of start-ups participating in

the IM-programme, and the propensity of participating firms and individuals to generate new

firms in the wake of IM-firms.

Earlier highly related analyses

The additional effects of the IMs on the economic performance of firms have earlier been evalu-

ated by CEBR (2009). The CEBR-analysis failed to detect any significant differences in the eco-

nomic performance of IM-firms and a reference group of comparable firms.

This is in line with

international studies on venture capital. It is a fact that the IM-programme is costly in the sense

of negative financial performance of the IMs, as most loans and investments in the IM-firms

need to be written off (Oxford Research, 2013). This again is partly due to low survival rates of

the IM-firms.

The present study is similar to the earlier CEBR-analysis. However, the present report

gains from more data: for example, the statistical robustness of results benefits from a substan-

tial increase in the number of observations for the present analysis relative to the CEBR-

analysis. Also, the richness of the data has increased substantially, including additional perfor-

mance measures, specific information on, e.g., firm employees’ background characteristics,

business transitions in association with firm exit, and the option to track individuals’ mobility

across different firms.

The survival and growth of IM-firms

The present study helps evaluating potential growth effects of the IM-programme by describing

the growth in employment, value added, turnover and annual earnings (profits) in participating

firms on basis of register data information. In order to estimate any potential additional effects

of the IM-programme, the IM-firms’ survival and growth is compared with a group of other firm

start-ups that do not participate in the programme. This group of other firms is selected such

that it resembles the group of participating firms in their observable characteristics.

Mobility of individuals

While the comparison study of IM-programme participants and non-participants is a standard

evaluation exercise, the present study also casts light on additional aspects: e.g. to what extent

individuals of IM-firms are involved in the startup of other firms. This is possible because Dan-

ish register data allows tracking individuals who move from one firm to another firm. This in-

formation is used to generate evidence on the mobility of individuals of the IM-firms. Of special

interest are the following questions:

(i)

(ii)

How many new firms are created by the individuals of the IM-firms?

What happens to the individuals in IM-firms that close operation? Do they start

new firms, or do they return to established organisations?

The first set of questions is motivated by the argument that public investments in innovation

might not be good business for society when measured

by financial returns and at the level of the

participating firm, but good business when one takes into account positive externalities. The

present study will by no means be able to establish monetary estimates on these externalities,

but it might give a first rough indication of business activity in the wake of IM-firms.

The CEBR report bases its conclusions on a comparison of firms that participate in the programme with other,

similar firms that do not participate.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

The second set of questions is partly motivated by the same externality argument. But it is

also interesting because of the incentive scheme of the IM-programme encouraging firm closure

as a means of getting rid of the IM as an outside owner or the obligation to repay loans. The

presence of this incentive is of particular importance to high-tech firms whose assets are often

uncodified ideas and knowledge which are easily transferred from one organization to another

one. And it might be one of the reasons for the high closure rates of IM- firms.

Also, given that the ambition of many portfolio firms is to compete on global markets, and

given that Denmark is a high-tax country, there is a strong incentive for entrepreneurs to move

the headquarters abroad by closing operation in Denmark and restarting the firm in countries

with more favourable business environments. International mobility of entrepreneurs and staff

of high-tech start-ups that close down gives an indication of any significance of this issue.

Scope and limitations of the analysis

This report is to be read as a contribution to our general knowledge of the functioning of the IM-

programme. Although the richness and scope of Danish business and other register data is out-

standing in an international comparison, there is still a lot we do not know about the IM-firms

and other firms in our data.

Data limitations

First, there might be success stories of IM-firms that escape our view by simply not being regis-

tered in the data.

However, we are not aware of any anecdotal evidence on salient successes that are not

reflected in our data, and there are indeed a number of records in the data with outstanding

annual earnings numbers in association with successful transactions of IM-firms. Yet, there is a

risk of extreme success stories of the IM-programme not being in the data. So this report might

be supplemented with the anecdotal evidence of extreme success stories – if they exist - and

related to other success parameters, like IMs’ return on investments that reflect these financial

successes.

The observation period for the analysis starts in 1999 and ends in 2011 for some parts of

the analysis and 2012 for others. So the report cannot describe very early projects of the IM-

programme and the developments of the most recent years. And it is important to note that it is

not just the case that the analysis might not do justice to some single success stories in case they

exist, but, generally, is unable to describe the growth and success of all IM-firms that are not in

the register data: A first result of the investigation is that there is a substantial share of IM-firms

that never reaches critical size and activity levels to become sampled in the official register da-

tasets, and never hands in any annual financial report to the business authorities that collect the

firm-level financial data for our analysis.

So the present report misses the - presumably not very significant - economic activity of

these IM-firms when investigating IM-firms’ total economic activity, and it might be assumed

that it paints a more positive picture of the economic performance of the IM-firms in the sample

than otherwise would be the case, simply because it cannot consider the least successful IM-

firms.

For the analysis of mobility of the individuals of the IM-firms, a share of missing IM-firms

in the registers also implies a limited number of worker-firm relationships in the Statistics Den-

mark employer-employee databases. This comes on top of a significant share of IM-firms not

having any individuals being associated with them in the data and a low representation of IM-

firms in the Statistics Denmark registers on firm creation and entrepreneurship.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

There is no information in the present analysis’ data on whether or not IM-firms would

have come to life without the IM-programme and how many of them would have been started if

the IM-programme did not exist.

Also, the present analysis on register data can not detect the

effects of the IM-programme on firms before they occur in the data for the first time. In sum, the

data used for the present analysis is not able to evaluate the potential effects of the IM-

programme on firm creation or any potential effects in the time period between creation and

being in the data for the first time.

Limitations of the treatment-control analysis design

Finally, part of the following analysis does not just provide an overview of IM-firms’ survival and

growth, but also compares their survival and growth to other, similar, start-ups. These similar

start-ups might help the reader to better evaluate to what extent additional effects of the IM-

programme on the performance of IM-firms might exist. But, obviously, there are elements of

randomness in the selection of this reference group for comparisons: there is freedom in terms

of the formulation of the models behind the selection procedure, the selection is based on a lim-

ited set of firm characteristics for firms with missing information in the data, and there is a ran-

dom selection of firms into the reference group in case of more than one firm being equally qual-

ified for selection into the reference group.

Thus, the reference group should not be taken for more than it is: a point of reference that

offers the opportunity to relate the survival and growth of participant firms with other firms that

can be shown to be highly similar in a number of observable characteristics like the first year in

which they occur in the data, the industry, and financial and other characteristics in the first year

in which they occur in the data.

Main findings

Keeping the above-described limitations with respect to data and methodology in mind, the

general characteristics of the IM-firms can be summarized as follows:

ii)

Approximately one third of the IM-firms do not reach critical mass to become sam-

pled in the Statistics Denmark register databases.

Approximately 70 percent of the IM-firms under observation leave the data over a

twelve-year horizon, where twelve years is the maximum time span that we can fol-

low these firms in the data. More than half of the IM-firms’ exits are registered as

being bankruptcies.

In 2011, which is the last year for which there is consistent employment information

in the data, IM-firms employ 1,600 individuals. This employment is for its largest

part created in association with firm creation: 1,200 of the 1,600 employment rela-

tionships are already present at the first time the firms figure in the data.

After firm creation, there is no sustainable employment growth in IM-firms. Surviv-

ing firms grow from on average from 2.5 to six employees over a 10-year time peri-

od, but this growth is offset by the exit of other IM- firms. In sum, the aggregate

iii)

iv)

Based on a survey, the Oxford Research (2013) report concludes that roughly 50 percent of IM-firms would not

have come to life in the (hypothetical) absence of the programme. The relationship between the respondents’

answers to this survey question and their performance is an obvious issue for further investigation, but is not part

this report.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

vi)

employment growth in the group of IM-firms cannot be shown to be positive in the

long run.

At the end of the analysis' observation period, IM-firms have an aggregate turnover

of approximately DKK2.5 billion and a value added of DKK500 million.

Although there is a couple of highly successful transactions in our data with annual

earnings (profits) well above DKK500 million, in total IM-firms generate an accu-

mulated earnings loss of approximately DKK5 billion over this variables’ observa-

tion period 1999-2012.

As mentioned, the analysis compares IM-firms with other similar firms of the same industry,

started in the same year as the IM-firms and similar in a number of observable

characteristics

like firm size and share of

highly educated individuals in their first year of existence. The com-

parison with these ‘reference firms’ gives us the following results:

vii)

IM-firms have higher exit rates from the data, higher closure rates, and higher

bankruptcy rates. In the beginning of the observation period, i.e., up to approxi-

mately 2006, IM-firms have lower growth in employment, turnover, and value add-

ed. Afterwards, they catch up and end up at similar levels at the end of the observa-

tion period in 2011 for employment and turnover and 2012 for value added and an-

nual earnings.

While survival is statistically significantly lower for IM-firms in comparison with

the reference group, the growth patterns in the other variables are not statistically

distinguishable from each other.

However, there are more successful firms in the group of IM-firms for firm ages ap-

proximately 5 years and above. There is tentative evidence of surviving IM-firms

being characterised by lower growth performance in terms of value added and earn-

ings than the reference group in the short run and higher performance in the long

run. Yet, it is important to note that these findings are associated with a considera-

ble uncertainty.

viii)

ix)

The Danish register data link individuals to firms. This allows generating the following find-

ings on the mobility of IM-firms’ staff:

IM-firms are characterised by high employee turnover: those firms for which work-

er-firm relationships (jobs) can be identified in the Danish register data dissolve 40

percent of their jobs per year, which is partly due to firm closures. Not all jobs in

IM-firms are the individuals’ only occupation: almost 25 percent of the individuals

in IM-firms are registered to have other jobs in other firms.

The analysis finds that, on average, roughly one new firm is started by or with IM-

firm employees for each IM-firm in the Statistics Denmark’s employer-employee-

database (FIDA).

There are only few new firms in the data that inherit groups of individuals from IM-

firms and therefore may be considered ‘spin-offs’. So the report finds no evidence of

any significant ‘spin-off’ activity in the wake of IM-firms.

xi)

xii)

These numbers hide a large degree of heterogeneity at the level of the individual firm. Individual surviving firms

increase their turnover by on average DKK8 million over a 11-year time period, which is the longest time span

available for the analysis of turnover. They increase their value added by on average DKK3.5 million over a 11-

year period and DKK8.8 million over a 12 year period, where the difference is due to a few highly successful

transactions of IM-firms at the very end of the analysis’ observation period.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

xiii)

Also, it rarely happens that groups of individuals of closed-down IM-firms move to

the same other firms. So there are no indications of artificial firm closures, i.e. clo-

sures as means of re-organization rather than termination of business. Individuals

of closed-down IM-firms rarely move abroad in association with firm closure.

In sum, for the time period under investigation and the data limitations in mind, the present

report cannot establish any significant evidence of the IM-programme leading to additional

effects on the performance of the IM-firms. In the first years of the IM-programme, IM-firms

were characterised by lower performance than the other start-ups in the reference group. More

recently, they were catching up; however, the analysis cannot establish evidence on whether this

development is a trend or just fortunate coincidence. In order to fully understand this seemingly

positive development in the IM-firms’ performance, it is recommended to follow the develop-

ments in the IM’s portfolios in the years to come.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Sammenfatning

Eksternaliteter til innovation medfører, at opfindere og entreprenører ofte er underbetalte i

forhold til værdien af deres opfindelser, og at innovationer ofte ikke bliver realiseret i etablerede

organisationer, men i nye virksomheder. Disse forhold motiverer offentlig intervention i form af

investeringer i små innovative virksomheder, og næsten alle OECD-lande har et eller flere støt-

teprogrammer til små innovative virksomheder. Som eksempler kan nævnes det amerikanske

Small Business Innovation Research (SBIR) program, eller, på europæisk plan, EUROSTARS-

programmet.

I Danmark antager offentlige investeringer i små innovative virksomheder forskellige for-

mer. Et af programmerne til fremme af små nye innovative virksomheder er Innovationsmiljø-

ordningen (IM-ordningen). IM-ordningen, som foreliggende rapport handler om, blev etableret i

1998 og hører i dag under Styrelsen for Forskning og Innovation (FI). IM-ordningen har en de-

central struktur, og bliver implementeret igennem danske Innovationsmiljøer, som er private

aktieselskaber, hvor universiteterne direkte eller indirekte indgår i ejerkredsen.

Innovationsmiljøerne finansierer nystartede innovative virksomheder (IM-virksomheder

eller porteføljevirksomheder i følgende), typisk i form af indskud af egenkapital (ansvarig låne-

kapital/risikovillig kapital). Innovationsmiljøerne opererer i de meget tidlige faser af virksom-

hedernes liv, hvor risikoen er høj og private investorer ofte holder sig tilbage. Indtil videre

(2013/2014) har cirka 1.000 virksomheder modtaget finansiering i størrelsesorden ca. 1,75 mil-

liarder kr.

Analysens fokus

Innovationsmiljøernes primære opgave er at afhjælpe dette

funding gap

i de tidligste og mest

risikofyldte faser af det private marked for risikovillig kapital ved at understøtte dannelsen af

nye innovative virksomheder. Som en del af dens løbende tilsyn og monitorering af innovati-

onsmiljøerne, udgiver FI en årlig opgørelse (Performanceregnskab) over innovationsmiljøernes

aktiviteter opgjort på en række nøgletal.

Som et supplement til de årlige performanceregnskaber har FI bestilt nærværende rapport.

Formålet med rapporten er at undersøge, om det kan sandsynliggøres, at innovationsmiljøerne

skaber yderligere effekter ift. IM-virksomhedernes økonomiske præstation. Dette spørgsmål

søges besvaret ved at sammenligne overlevelsesrate og en række økonomiske nøgletal for IM-

virksomhederne og en referencegruppe af sammenlignelige virksomheder. Derudover undersø-

ger rapporten omfanget af iværksætteraktiviteten blandt de personer, der har været tilknyttet

IM-virksomhederne.

Tidligere resultater

Hvorvidt innovationsmiljøerne kan siges at skabe yderlige effekter ift. IM-virksomhedernes

økonomiske præstation er tidligere blevet undersøgt af CEBR (2009). CEBR-analysen kunne

ikke finde stærke indikationer for tilstedeværelsen af yderligere effekter af IM-ordningen ift.

virksomhedernes økonomiske performance.

For yderligere information se http://ufm.dk/forskning-og-innovation/samspil-mellem-viden-og-innovation/fa-

hjaelp-til-kommercialisering/innovationsmiljoer/tal-om-innovationsmiljoer-1.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Nærværende rapport har samme metodiske udgangspunkt som CEBR-analysen, men har

adgang til mere omfattende data, fx en væsentlig stigning i antallet af observationer. Også in-

formationsmængden i data er øget i forhold til CEBR-analysen, såsom supplerende resultatmål,

oplysninger om medarbejderes baggrundsvariabler som f.eks. uddannelse, virksomhedsover-

gange i forbindelse med virksomhedslukninger, og muligheden for at spore den enkeltes mobili-

tet på tværs af forskellige virksomheder.

Analysedesign

Nærværende rapport undersøger, om det ved hjælp af registerdata kan sandsynliggøres, at inno-

vationsmiljøerne skaber yderligere effekter ift. IM-virksomhedernes overlevelse og økonomiske

præstation, ved at sammenligne væksten i beskæftigelse, værditilvækst, omsætning og indtjening

(profit) for IM-virksomheder og en referencegruppe af sammenlignelige virksomheder, der ikke

har deltaget i IM-ordningen. Referencegruppen er udvalgt så de ligner IM-virksomhederne på en

række observerbare karakteristika.

De tidligere resultater fra CEBR-analysen vedrørende tilstedeværelsen af yderligere effekter

af IM-ordningen er i overensstemmelse med de foreliggende internationale undersøgelser af

risikovillig kapital. Dog er det et faktum, at IM-ordningen er underskudsgivende i den forstand,

at de fleste investeringer i portefølje- virksomhederne skal afskrives med lave finansielle tilbage-

løb til følge (Oxford Research, 2013). Dette skyldes til dels lave overlevelsesrater blandt porteføl-

jeselskaberne. På den anden side, så vides lidt om de potentielt positive effekter af IM-

ordningen, f.eks. ift. vækst, innovation, eller iværksætteraktiviteten blandt individerne tilknyttet

IM-virksomhederne. Nærværende analyse belyser følgende valg af aspekter:

(i) Overlevelse og vækst af nystartede virksomheder, der deltager i IM-

ordningen. Dette bygger videre på tidligere CEBR-analyse.

(ii) Tilbøjelighed af IM-virksomheders enkeltindivider til at skabe nye virksom-

heder i kølvandet af IM–virksomhederne.

Mens selve sammenligningen af deltagende og ikke-deltagende virksomheder er en standard

evalueringsøvelse, så kaster undersøgelsen også lys på, i hvilket omfang individerne bag IM-

virksomheder skaber nye virksomheder. Dette belyser iværksætterdynamikken blandt projekt-

deltagerne og er mulig, da de danske registerdata tillader at følge individer, der flytter fra et

firma til et andet. Disse oplysninger bruges til at dokumentere mobiliteten blandt IM-

virksomhedernes individer.

Af særlig interesse er, hvor mange nye virksomheder er skabt af individerne i IM-

virksomheder. Dvs. om IM-virksomheder genererer "spin-off" virksomheder, og hvis ja, hvor

mange? Derudover ser rapporten nærmere på, hvad der sker med de individer, som forlader IM-

virksomheder, f.eks. når disse lukker. Starter de nye virksomheder, eller vender de tilbage til

etablerede organisationer?

Svaret på disse spørgsmål skal give et fingerpeg om omfanget af eventuelle positive ekster-

naliteter af IM-ordningen ift. iværksætteraktivitet. Argumentet bag er, at de offentlige investe-

ringer i innovation måske ikke er en overskudsforretning for samfundet målt ved det økonomi-

ske afkast for deltagende virksomheder, men en god forretning, når der tages hensyn til de posi-

tive afledte effekter, som f.eks. dannelse af nye virksomheder. Nærværende undersøgelse er på

ingen måde i stand til at generere monetære estimater på disse eksternaliteter, men giver en

første grov indikation af niveauet af iværksætteraktiviteten i kølvandet af IM-virksomhederne.

At belyse medarbejdermobilitet i IM-virksomhederne er også interessant, da IM-

ordningens investeringer i form af risikovillig kapital måske kan tilskynde lukninger af virksom-

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

heder, hvor iværksætteren ”sætter eksterne ejere uden for døren”. Dette incitament kunne være

af særlig betydning for højteknologiske virksomheder, hvis aktiver ofte er ideer og viden, som let

kan overføres fra en virksomhed til en anden. Derudover kan kombinationen af at mange IM-

virksomheder sigter efter globale markeder og et forholdsvis højt skattetryk i Danmark medføre

at iværksættere flytter udenlands ved at lukke operation i Danmark og genstarte selskaberne i

lande med et gunstigere forretningsmiljø. Så et kig på international mobilitet af iværksættere og

ansatte i de højteknologiske nystartede IM-virksomheder giver en indikation af betydningen af

denne problemstilling.

Afgrænsning

Denne analyse skal læses som et bidrag til den generelle viden om IM-ordningen. Selvom om-

fanget af datamaterialet bag analysen er enestående i en international sammenligning, så skal

begrænsninger i informationsmængden holdes in mente ved fortolkning af analysens resultater.

For det første kan der være succeshistorier blandt IM-virksomheder, som undslipper analy-

sen ved ikke at være registreret i data; umiddelbart vides dog ikke noget om anekdotiske succe-

ser, som ikke er med i analysens data. Og der er faktisk en række observationer i analysens data

med udestående årlige indtjeninger, typisk i forbindelse med vellykkede transaktioner i mest

biotek branchen som f.eks. salg af licenser. Men der er altid en risiko for at succeshistorier fra

IM-ordningen ikke er i data grundet bogføringsmæssige forhold. Denne rapport kan altså med

fordel suppleres med cases af ekstreme succeshistorier - hvis de findes - og kan med fordel rela-

teres til andre parametre, som fx innovationsmiljøernes finansielle afkast (beskrevet i Oxford

Research, 2012), der afspejler disse finansielle succeser.

Analysens observationsperiode er 1999 og frem til 2011 for nogle variable og 2012 for an-

dre. Så data kan ikke beskrive de allerførste projekter, som er iværksat før 1999 derudover

mangler data efter 2012. Det er også vigtigt at bemærke, at analysen ikke er i stand til at beskrive

vækst og succes i en forholdsvis stor delmængde af små IM-virksomheder: et første resultat af

undersøgelsen er nemlig, at en væsentlig andel af IM-virksomheder aldrig når den kritiske stør-

relse og aktivitetsniveau til at blive samplet i Danmarks Statistiks officielle registerdatasæt, og

aldrig afleverer en årsrapport til Erhvervsstyrelsen, der står bag de finansielle oplysninger på

virksomhedsniveau for analysen.

Dermed misser analysen den økonomisk aktivitet i disse IM-virksomheder, som formentlig

ikke er særlig stor, når den ser på den samlede økonomiske aktivitet i IM-virksomheder. Yderli-

gere må det forventes, at analysen tegner et mere positivt billede af IM-ordningens potentielle

bidrag til deltagende virksomheder, end det ellers ville være tilfældet - simpelthen fordi analysen

ikke betragter de mindst succesrige IM-virksomheder, som aldrig når ind i datamaterialet.

For så vidt angår analysen af mobiliteten af IM-virksomheders medarbejdere, er det væ-

sentligt at have in mente, at en betydelig andel af IM-virksomhederne ikke har individer tilknyt-

tet i Danmarks Statistiks iværksætter- eller medarbejder-databaserne.

Data siger heller ikke noget om, hvorvidt IM-virksomhederne ville være etableret uden til-

stedeværelse af IM-ordningen, dvs. hvor mange af dem ville være startet i en hypotetisk situati-

on, hvor IM-ordningen ikke fandtes.

Og en analyse på registerdata kan naturligvis ikke finde

eventuelle effekter af IM-ordningen for virksomhederne, inden de optræder i data første gang.

Dermed er analysen ikke i stand til at beskrive programmets eventuelle betydning for virksom-

hedsopstart og potentielle effekter i perioden mellem opstart og tidspunktet hvor virksomheden

optræder i data for første gang.

Dog viser en spørgeskemaundersøgelse foretaget blandt IM-virksomhederne, at omkring 50% ikke ville være

startet uden hjælp fra innovationsmiljøerne, Oxford Research (2012).

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Endelig giver en væsentlig del af følgende analyse ikke bare et overblik over IM-

virksomhedernes overlevelse og vækst, men sammenligner deres overlevelse og vækst med an-

dre lignende nystartede virksomheder (referencegruppen). Denne sammenligning kan hjælpe

ved vurderingen af, i hvilket omfang IM-virksomhedernes præstation skal betragtes som til-

fredsstillende.

Men referencegruppen bør ikke tages for mere end den faktisk er: et referencepunkt, der

giver mulighed for at forholde sig til spørgsmålet, om overlevelse og vækst i deltagende virksom-

heder er på højde med andre virksomheder, som kan vises at være meget lig i en række obser-

verbare karakteristika, f.eks. deres første år, hvor de optræder i de data, deres industri, uddan-

nelsesniveau og en række finansielle variable i det første år, hvor de optræder i data.

Resultater

Med de ovenfor beskrevne begrænsninger ift. data og metode in mente kan

resultaterne fra den

deskriptive analyse opsummeres som følgende:

ii)

Ca. en tredjedel af alle IM-virksomheder opnår ikke den kritiske mini-

mum-aktivitet til at optræde i Danmarks Statistiks registrerdatabaser.

Ca. 70 procent af IM-virksomhederne forlader stikprøven over en 12-

årig tidshorisont, hvor 12 år er den maksimale tidsperiode, som vi kan

følge IM-virksomhederne i data. Mere end halvdelen af disse virksom-

hedsophør er registreret som at være konkurser.

I 2011, som er det sidste år, for hvilket der er konsistente informationer

vedrørende beskæftigelsen i data, beskæftiger IM-virksomheder ca.

1.650 individer. Denne beskæftigelse er størstedels skabt i forbindelse

med virksomhedsopstart: Ca. 1.200 ansættelsesforhold er allerede til

stede første gang virksomhederne optræder i data.

Når virksomheder følges over tid, finder analysen, at IM-virksomheder

som helhed ikke generer vedvarende vækst i beskæftigelsen. Overleven-

de virksomheder vokser fra i gennemsnit fra 2,5 til seks medarbejdere i

løbet af en 10-års periode, men denne vækst opvejes af jobtabet i andre

IM-virksomheder, som lukker. Sammenfattende kan den samlede vækst

i beskæftigelsen for IM-virksomheder ikke vises at være positiv i det

lange løb.

I slutningen af observationsperiode måles IM-virksomhedernes samlede

årlige omsætning til at være på ca. 2,5 milliarder kr. og deres værditil-

vækst (dækningsbidrag) til 500 millioner kr.

Selvom der er en række meget vellykkede transaktioner i analysens data

med årlige indtjening (profit) over 500 millioner kr., så måles IM-

virksomhederne som helhed at generere et akkumuleret finansielt tab på

ca. 5 milliarder kr. over perioden 1999-2012.

iii)

iv)

vi)

Disse tal skjuler en stor grad af heterogenitet blandt de enkelte IM-virksomheder. Overlevende IM-virksomheder

øger deres omsætning med 8 millioner kr. i gennemsnit over en 11-årig periode. Overlevende IM-virksomheder

øger i gennemsnit deres værditilvækst med 3,5 millioner kr. over en 11.-årig periode og 8,8 millioner kr. over en

12-årig periode, hvor forskellen skyldes nogle få meget vellykket transaktioner blandt IM-virksomhederne i slut-

ningen af observationsperioden.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Som nævnt, så sammenligner analysen IM-virksomheder med en referencegruppe af tilsvarende

virksomheder, som er i samme branche, startet i samme år som IM-virksomheder og er sam-

menlignelig i en række andre observerbare karakteristika.

vii)

I sammenligning med referencegruppen forlader en højere andel af IM-

virksomheder datamaterialet over observationsperioden, og IM-

virksomheder er kendetegnet ved højere andele af virksomhedsluknin-

ger og konkurser. IM-virksomheder har i begyndelsen af observations-

perioden, dvs. op til ca. 2006, en lavere samlet vækst i beskæftigelsen,

omsætning og værditilvækst. Bagefter haler de ind og ender på omtrent

samme niveau ved afslutningen af observationsperioden i 2011 for be-

skæftigelse og omsætning, og i 2012 for værditilvækst og den årlige ind-

tjening.

Mens overlevelse er signifikant lavere for IM-virksomheder i sammen-

ligning med referencegruppen, så er de andre vækstvariablene for IM-

virksomhederne og referencegruppen ikke signifikant forskellige fra

hinanden.

Blandt virksomheder, som overlever deres første fem år er der flere suc-

cesfulde virksomheder blandt IM-virksomheder end i referencegruppen.

Overlevende IM-virksomheder er kendetegnet ved lavere vækst i værdi-

tilvækst og indtjening på kort sigt og højere vækst i det lange løb sam-

menlignet med referencegruppen. Det er dog vigtigt at bemærke, at dis-

se positive resultater er forbundet med en betydelig statistisk usikker-

hed.

viii)

ix)

De danske registerdata tillader at se på mobilitet blandt IM-virksomheders medarbejdere. Her

finder analysen følgende:

IM-virksomheder er kendetegnet ved en høj medarbejderomsætning: I

de virksomheder, for hvilke der kan identificeres

worker-firm

relationer

(jobs) i de danske registerdata, forlader 40 procent af medarbejderne

deres job om året, hvilket til dels skyldes virksomhedslukninger. Ikke al-

le jobs i IM-virksomheder er medarbejderens eneste beskæftigelse: næ-

sten 25 procent af medarbejderne i IM-virksomheder er registreret til

også at have jobs i andre virksomheder.

Analysen finder, at der i gennemsnit etableres ca. én ny virksomhed af

eller med medarbejdere fra IM-virksomheder for hver IM-virksomhed i

FIDA-databasen.

Der findes forholdsvis få nye virksomheder i registerdata, som er startet

af IM-virksomheders medarbejdere, og som blev startet ved at mere end

én medarbejder flyttede fra IM-virksomheden til den nye virksomhed

(spin-offs). Så der findes ikke tegn på nævneværdig dannelse af spin-off-

virksomheder i kølvandet af IM-virksomhederne.

Det sker sjældent, at grupper af medarbejder fra lukkede IM-

virksomheder flytter til den samme anden virksomhed. Så der findes

heller ikke tegn på, at en nævneværdig del af IM-virksomhedernes luk-

ninger i virkeligheden er skjulte omorganiseringer. Det sker også sjæl-

dent, at individer fra lukkede IM-virksomheder flytter til udlandet i for-

bindelse med virksomhedslukning.

xi)

xii)

xiii)

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Så for den givne tidsperiode og de datamæssige begrænsninger in mente, finder analysen ikke

tegn på at innovationsmiljøerne skaber yderligere væksteffekter for IM-virksomhederne. Efter

programmet blev startet, har IM-virksomheder været kendetegnet ved svagere performance end

andre opstartsvirksomheder som ligner IM-virksomheden i en række observerbare karakteristi-

ka. I de seneste år af analysens observationsperiode er IM-virksomheder ved at indhente disse

andre virksomheder, men det er for tidligt at afgøre om det er tale om egentlig trend. Derfor

anbefales det at følge udviklingen i IM-virksomhederne i de kommende år for at finde svar på,

om denne tilsyneladende positive udvikling er del af en længerevarende trend.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

1. Introduction

The following report presents the data, methodology, and results of the analysis of the Danish

Innovation Incubator Programme (Innovationsmiljøordningen, IM-programme in the follow-

ing). The analysis was completed by Johan Kuhn/EPAC for the Danish Agency for Science,

Technology and Innovation (DASTI) in 2015. It contributes to

DASTI’s

strategy to continuously

monitor and evaluate its innovation support programmes, to develop and improve the designs of

its programmes, and to improve programme evaluation techniques.

The IM-programme was launched in

1998

and provides finance and counselling to newly

started innovative firms. It is under the Danish Agency for Science, Technology and Innovation,

DASTI, and administered by locally embedded incubators called Innovationsmiljøer in Danish

(IM, “Innovation

environments”).

The IM-programme offers publicly funded risk finance and counselling to researchers,

entrepreneurs and others who have an innovative business idea with a commercial potential.

The IMs operate in the very early stages of the commercialisation process where the risk is high

and private investors are reluctant to engage. Hence, the IM-programme compensates for a

potential market failure in the private market for risk finance, by supporting the creation of new

innovative start-ups.

For any analysis, it is of course imperative to define the dimensions by which the pro-

gramme in question is to be analysed. This present report considers the survival and growth of

firms that participate in the IM-programme. Growth is measured by the number of employees,

turnover, value added and annual earnings (profit) at firm level. Obviously, if IM-firms create a

lot of jobs or economic activity, then society gains directly from tax income and indirectly from

externalities like demand for services and intermediate products from other firms.

Also, small innovative firms are often short-lived, located in dynamic environments, and

ideas can be easily transferred from one firm to the other. So one obvious way by which IM-firms

might be assumed to generate positive externalities is by encouraging the creation of other new

firms that benefit from assets like knowledge or ideas that are transferred from the IM-firm to

the new firm. The present report follows this presumption, and investigates to what extent IM-

firm individuals are creating new firms in the wake of existing and closed-down IM-firms.

1.1 A short description of the Danish Innovation Incubator Programme (Innova-

tionsmiljøordningen)

The purpose of the IM-programme is to support commercialisation of business ideas and re-

search and development output. IMs offer consulting and capital, where capital is provided as

any combination of two different kinds of loans, distinguished by their priority in case of default,

and equity investments (venture capital).

Obviously, in case of equity investments, the IM assumes ownership in the IM-firm. How-

ever, for this analysis, it was not possible to take into consideration to what extent IMs are repre-

sented in the boards of their portfolio firms and to what extent they exercise influence in these

firms. Still, in comparison to other Danish public venture capital initiatives, there is high a

transparency of the IMs’ investments and identification of IM-firms is univocal. This is because

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

IMs typically invest directly in their portfolio firms rather than indirectly through private ven-

ture capital firms.

In terms of their focus on providing capital for start-ups the IMs are related to another

state-owned investor called Vækstfonden (the Danish Growth Fund) and a recently established

fund called Dansk Vækstkapital which is primarily financed by state-guaranteed loans from

Danish pension funds. In addition to these there are a number of private venture capital firms,

some of which can draw on the resources of the industrial foundations of some of the largest

Danish firms.

The most significant difference between the IMs and most other investors is that the IMs

are focusing exclusively on financing the very early stages of commercialisation processes by

limiting their portfolios to newly started firms of a maximum age of one year. So the IMs offer

services only to newly started firms and act in the most risky segment of the venture capital

market.

With respect to the additional economic effects on the IM-firms, the IM-programme has

been evaluated once before, in 2009, by the Centre of Economic and Business Research (CEBR).

The CEBR report compared the performance of IM-firms with other firms, and remains largely

inconclusive regarding any performance differences.

1.2 Theoretical considerations, earlier research and related studies

1.2.1 General challenges of venture capital finance

The IMs invest venture capital, for which reason it is relevant to place the current study in the

literature on the potential effects of venture capital on recipient firms. Also, they provide coun-

selling in association with finance, which in itself should have positive effect on the performance

of participating firms (Nielsen and Keuschnigg, 2007). They also provide access to financing

through loans. The IM’s investments are often levered by private venture capital investors (Ox-

ford Research, 2012).

Investments in start-ups are risky, and venture capital is sometimes considered the only

means of finance for firms based on new technologies and is characterised by long development

horizons. Yet, venture capital investments in small high-tech start-ups are characterised by typi-

cal principal-agents problems that occur when there is uncertainty and private information. The

more risky the project, and the lower its growth potential, the more attractive is venture capital

compared to loan finance from the point of view of the entrepreneur. Venture capital implies

that a share of the financial gains of a project is going to the venture capital investors. So venture

capital is an expensive form of finance for highly successful projects. This implies that entrepre-

neurs who have a choice of how to finance their projects will typically prefer other means of

financing than venture capital (Berger and Udell, 1998) – especially those entrepreneurs with a

strong belief in their business idea.

On the other hand, venture capital is not leaving any financial obligation in case of business

failure, making it a cheap form of finance for unsuccessful start-ups. This implies that, from the

entrepreneur’s point of view, it is the preferred choice for risky projects or projects with low

growth potential. So notwithstanding the effort of investors to screen projects – manifested in

‘proof-of-business’ and ‘due diligence’ procedures – it needs to be kept in mind that there is a

simple microeconomic adverse selection problem associated with venture finance that can be

assumed to lead to poor quality firms being over-represented in the pool of firms financed by

venture capital.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Venture capital finance of high-tech start-ups requires detailed contracts (see, e.g., Kaplan and

Strömberg, 2002), as there is an incentive for the entrepreneur to close the business to get rid of

the venture capital investor, and transfer the firm’s intangible assets like knowledge and ideas to

another firm afterwards. The critical point in these contracts is the extent to which investors can

monitor the entrepreneur and exercise control in the firm.

There is a theoretical argument that any division of ownership and control in venture capi-

tal financed firms might lead to opportunistic behaviour like rent extraction (Lerner (2012),

CEBR (2009)).

On the other hand, venture capital is often considered the only source of capital for highly

risky projects with sometimes long development horizons. For IMs, this view is supported by

Oxford Research’s (2012) report where they present results of a survey of the firms that have

cooperated with an IM: approximately 50 percent of the respondents claim that they would not

have been started without the IM.

In sum, venture capital investments put high demands on the investors in terms of select-

ing projects, contracting and management.

1.2.2 Earlier results on Danish venture capital finance

These challenges of venture capital finance notwithstanding, there are of course a number of

highly successful firms being funded and fostered with venture capital. The historical success of

(public) venture capital finance is an empirical question, and this and the following subsections

will make an attempt to summarise some of the experiences with venture capital.

Venture capital is a financial instrument, so an obvious parameter to gauge its contribution

to society is whether or not its financial returns are positive or negative. If positive, this would be

a sufficient condition for recommending an increase in the volume of public venture capital.

For Denmark, the question of whether or not financial returns are sufficiently high to make

an argument for venture capital has a clear answer: for the Danish venture capital market as

such, the rate of return is, according to the Ministry of Business Affairs

, minus 6.7 percent per

year, leaving investors with a forty percent loss of their initial investments. In this context the

IMs perform no better than the private venture market. Oxford Research (2012) documents that

the different IMs, the portfolio firms of which are the subject of this study, generate losses of

between 73 and 91 percent of their initial investments over the (economic downturn) period

2007-2010. One reason of private venture capital outperforming the IMs with respect to return

of investment, might be that the IMs operate in the earliest and most risky stages of the venture

market (pre-seed and seed capital), where private investors are reluctant to invest, and thus

invest in firms with higher risk of failure compared to private venture capital investors.

Given these findings, there is a need for other arguments than positive financial payoffs if

one is to invest in venture capital. These arguments can be divided into two parts: first, Danish

investors have, until now, been unexperienced or unlucky, or both, and it is only a question of

time until Danish business successes financed by venture capital turn financial returns positive.

Second, one should adopt a wider focus than just concentrating on financial returns, as there

might be other positive effects of venture capital investments that are large enough to over-

compensate the financial losses.

Økonomi og Erhvervsministeriet (2010).

Strictly speaking, the success threshold is not just whether or not venture capital reaches above zero returns, but

higher risk-adjusted returns than alternative investments.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

To get an indication of the relevance of the first argument, one might turn one’s view to the in-

ternational evidence on financial returns to venture capital investments. There is no clear pic-

ture on this issue, yet, in one of the most recent and comprehensive surveys of the international

literature by Da Rin et al. (2013), the authors conclude that “there is an emerging consensus that

average returns of venture capital funds do not exceed market returns”. And the second argu-

ment can be investigated on the basis of alternative growth effects of venture capital such as

innovation or employment creation. Studies on these issues are discussed further below.

1.2.3 Measurement issues of venture capital’s potential growth effects

Unfortunately, studies on venture capital are to their greatest part subject to the problems of

unrepresentative samples and statistical omitted variable biases.

The first problem arises when data is based on surveys with insufficient response rates or

selected such that only successful venture-firms or firms financed by venture capital are ana-

lysed. For example, survey evidence is typically biased by survivorship bias (only surviving firms

are asked to participate) and response bias

, as it must be presumed that the decision to re-

spond to a survey is positively correlated to the success of the venture.

But to find representative data for an analysis of venture capital is a general problem that is

not just limited to survey data. Finding and isolating venture investors and their portfolio firms

is not a straightforward task in many studies that, as a consequence of this, might be feared to

miss out the smallest or least successful venture investors or firms backed by venture capital.

This applies to studies like Peneder, 2009, while other studies like, e.g., Botazzi and Rin, 2002,

focus on a small and non-representative share of venture-capital-backed firms that are publicly

listed.

Omitted variable biases are the result of violations of statistical all-else equal conditions.

They are equivalent to alternative explanations of the analysis’ findings that the modeller has not

taken care of. Alternative explanations of a study’s findings cannot be ruled out in any empirical

study the data of which is not generated by a lottery design. Statistical analyses are innocent

comparisons of subpopulations, however, interpretations of their findings as causal effects rest

on all-else equal assumptions.

1.2.4 Earlier studies on aggregated data

Studies, the interpretations of which are prone to omitted-variable biases are studies on country

comparisons, like Meyer, 2010. Cross-country studies can document positive correlations be-

tween venture capital investments and GDP growth in given years, but these correlations have

no simple interpretation and can typically not be used for policy recommendations.

Markets for

financial instruments fluctuate with business cycles, and the issue of whether venture capital

generates innovation activity or innovation activity creates markets for venture capital is far

from resolved.

See, e.g., Phalippou, L. and O. Gottschalg, 2009.

Two examples of these potential biases are Vækstfonden, 2006 and Vækstfonden, 2009.

The Meyer study documents contemporary correlations or correlations with short time lags. This is at odds with

venture capital being ‘patient’ capital typically invested in projects with relatively long development time hori-

zons. One reason for mentioning this study at this place is that it previously has been cited in, e.g., Oxford Re-

search, 2012.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Yet, the question of whether the size of venture capital markets is positively related to innovation

activity and the number of highly successful start-ups is of course legitimate. E.g., it is an obvi-

ous question of whether the exponential growth in the size of venture capital market over the last

decades in the U.S. has been associated with any proportionate increase in innovation activity.

Studies addressing this question remain inconclusive. A heavily cited study that uses a

change in legislation as an exogenous shock to identify causality by Kortum and Lerner, 2000,

suggests the presence of positive effects of venture capital on patenting activity.

But this is not equivalent to the international evidence univocally suggesting positive ef-

fects of venture capital on recipient firms. The commercialisation focus of venture capital inves-

tors is not necessarily making recipient firms more innovative, and, e.g., Stuck and Weingarten,

2005 have argued that innovation activity in the U.S. has not followed track with the increase in

the volume of the venture capital market.

A more recent study by Ueda and Hirukawa, 2008, that uses the same analysis set-up,

suggests the presence of labour productivity effects, but cannot detect any signs of positive total

factor productivity effects. This supports notions of more capital intensive production technolo-

gies in venture capital recipient firms compared to other firms. And Zucker, Darby, and Brewer,

1998, even document a negative relationship between venture capital market size and the crea-

tion of new biotechnology firms.

1.2.5 Firm-level comparison studies - general issues

The most straightforward way to reduce omitted variable biases is to analyse potential effects of

venture capital at the level of the individual firm, and to compare firms that receive venture capi-

tal with other firms that do not receive venture capital.

This is the ‘industry-standard’ in the academic evaluation literature. However, although

academic papers on the subject (and most reports) typically claim to establish evidence of causal

relationships in the data, it needs to be kept in mind that any non-experimental study is – to

some extent – comparing apples with bananas: no matter how carefully the modeller has select-

ed firms for comparisons, there will always be differences in the two groups of firms, some of

which might be suspected to drive the later results of the study.

In other words, the results of comparison studies are always a combination of the true

(value adding) effects of venture capital and selection effects, i.e., firms with specific characteris-

tics having a higher propensity of receiving venture capital. So part of the findings of these stud-

ies is due to venture capital changing receiving firms’ growth paths, and part of the findings is

due to venture capital financed firms having specific characteristics to start with.

1.2.6 Earlier results of firm-level comparison studies

For Denmark, there are currently two firm comparison studies of the potential effects of venture

capital on firm growth.

First, CEBR, 2009, which cannot document venture capital recipient firms showing higher

growth than other comparable firms. And, second, a study by Ernst and Young, 2010, that con-

For Denmark, CEBR, 2012, calculates job creation estimates of public investments in firms on basis of a Keynes-

ian macroeconomic model. This treats these investments as standard fiscal policies and calculates job creation by

the model’s Keynesian multipliers. The numerical results of the study are for the hypothetical scenario of these

public investments being just as strongly related to job creation as other investments, and rely on the properties

of the macroeconomic model.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

centrates on firms receiving capital from the public investor Growth Fund/Vækstfonden that

also finds similar or lower increases in economic activity in firms receiving capital from the in-

vestor.

Examples of international studies that compare start-ups financed with venture capital

with other start-ups are Penneder, 2009, on Austrian data, Engel and Keilbach, 2007, on Ger-

man data, and Puri and Zarutskie, 2009, on U.S. data.

When comparing venture capital recipient firms with other firms, Engel and Keilbach,

2007, find high innovation activity in venture capital financed firms prior to the venture capital

investment, but not afterwards. They document higher growth rates in venture capital financed

firms than other firms that, however, the firms are not equivalent in their patenting activity prior

to receiving venture capital.

Peneder, 2009, finds that venture capital financed firms grow faster than other similar

firms. His study also addresses an important aspect of venture capital finance: the question of

how large a share of venture capital recipient firms would exist in the first place in the absence of

venture capital finance. This question has of course no general answer, as capital supply is highly

volatile over time, but it is also a question typically evading stringent analysis.

Puri and Zarutskie,2012, give evidence of the relatively few firms that have received ven-

ture capital in the U.S. having created substantial numbers of jobs.

This evidence finds coun-

terparts in the statements of various venture capital associations (including the Danish one),

which, however, have not been scrutinized by independent research.

Puri and Zarutskie also

find that venture-capital financed firms grow faster in terms of employees and sales than other

highly similar firms, and have lower exit risk.

The Puri and Zarutskie study is interesting as it gives evidence of venture capital investors

concentrating on growth rather than profitability. This agrees with the tentative findings for IMs

of this present report and the comparison of the studies of Kortum and Lerner, 2000, on the one

hand, and Hirukawa, M, and M. Ueda, 2011, on the other: while the former of the two gives evi-

dence of venture capital investments leading to higher levels of innovation (measured by patent-

ing activity) the second finds venture capital investments being negatively related to productivity

growth.

The results of higher growth in venture capital backed firms are in line with other often

cited studies like Brav and Gompers, 1997, Hellman and Puri, 2000, and Kortum and Lerner’s,

Note the job creation argument of venture capital rests on the assumption that substantial shares of employees

in venture capital recipient firms were unemployed in the counterfactual case of an absence of venture capital.

This assumption is not necessarily realistic, especially in the light of venture capital firms employing high shares

of high-skilled individuals who are characterized by low unemployment. Unfortunately, there is little general

evidence on the extent to which new firms create new economic activity and to what extent they crowd out activi-

ty in established firms. Technological innovations, like, for example, the power loom, can have negative effects on

employment as well. And new firms as such are generally not being more productive than established ones (van

Praag and Versloot, 2007), indicating that there is no simple relationship between entrepreneurship activity and

economic growth.

There is no convergence towards any harmonized Danish database on venture capital recipient firms. This data-

base would be a first step towards commonly agreed evidence on the most basic features of the Danish venture

capital market.

Their study gives indications of venture-financed firms being characterized by relatively low exit risk of in the

first years, and relatively high exit risk after the first years of existence. So venture capital investors are suggested

to be patient at first, but get determined when closing unsuccessful projects.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

2000, study on innovation increases in association with an exogenous increase in the volume of

the venture capital market.

1.2.7 Potential knowledge spill-overs of high-tech firms

As mentioned in the outset of this section, an important aspect of public innovation support

programmes like public venture capital, is that innovations create externalities. Thus, mobility of

knowledge implies that even firms that go bankrupt can make positive contributions to society,

for example if the ideas developed in the unsuccessful firm lead to the foundation of new (and

successful) firms afterwards. A single and ground-breaking analysis that follows this argument is

Møen, 2004, who, however, cannot establish evidence of positive growth effects on firms that

were created in the wake of a government innovation support programme that was generally

deemed unsuccessful.

1.2.8 There is no convergence in the literature on the effects of venture capital

So, in sum, making general statements on the economic effects of venture capital is difficult.

Studies typically reporting low financial returns on venture capital investments meet a number

of studies that find positive growth developments in venture-capital financed firms.

So at present, we know too little about how important venture capital is to create growth

for being able to making policy recommendations. Insiders statements like “taken together, the

evidence supporting the positive impact of venture capital on innovation is weak at best (Ueda

and Hirukawa, 2009)” or “We believe that the role of government in venture capital remains

under-researched (Da Rin et al, 2012)” indicate that the question of whether public intervention

in the venture capital market is good business for society remains unanswered in the academic

literature.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

2. Data

The data

for the analysis comes from DASTI, Experian A/S, and Statistics Denmark:

DASTI supplied information on 1,034 individual IM-projects from the period 1998 to

2013. Information includes firm-identifiers, and the start date of the project. This will

henceforth be called

the DASTI IM data.

Firm background information comes from the Danish Business Authority, including in-

dustry and date of establishment. These data, typically referred to as ‘Stamdata’, are

made available by the business intelligence company Experian A/S. Of the 1,034 firms

in the DASTI IM data, 1,017 can be found in the firm background information.

Data from annual financial reports that incorporated firms are obliged to submit to the

Business authority. Just like it is the case for the firm background information, these

data were made available by Experian A/S. Of the 1,034 individual IM-firms, 877 hand

in at least one financial report that is in the financial data. So there are 140 firms that

are in the firm background information but are not represented in the financial report

data. Of these, a number were started too recent to be in the accounting data; 91 were

started after 2011 and not so important for the analysis that is supposed to follow firms

over a couple of years after start-up.

Statistics Denmark

from 1999 onward. This is matched em-

ployer-employee data including information on individuals (demographic information,

information on education, wage and occupation) and firms (e.g. size, turnover). These

data will be referred to as the

Statistics Denmark data.

2.1 The Statistics Denmark data

Characteristics for individuals and firms are drawn from Statistics Denmark’s register. Data is

available up to 2011, which implies that there is no information on the most recent projects.

Statistics Denmark data is available on an annual basis, with census date in mid-November. The

data comes with individual and firm identifiers; these allow associating individuals with their

firms. Over the last decades, the data resources of Statistics Denmark have been continuously

extended. For example, the present analysis benefits from Statistics Denmark’s individual-level

information on education (degrees, focus of electives, grades) and firm-level information on

turnover.

The most relevant Statistics Denmark databases for the analysis are as follows:

1. The Statistics Denmark education register data has information for all Danish citizens

and most immigrants, and distinguishes between 2,800 different educations. For the

analysis, these will be categorised according to the Danish Education Classification sys-

tem (DUN) similar to the International Standard Classification of Educations (ISCED).

2. The Statistics Denmark

FIRM-database

has information on turnover and a few other

variables for all firms above some minimum activity levels for all private sector firms

with the exception of a few industries, that Statistics Denmark considers non-business-

related (e.g. social institutions).

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

The Statistics Denmark

Entrepreneurship-database

is a sub-sample of the FIRM-

database and available from 2001 onwards. It associates new firms in the FIRM-

database with individuals, that, according to Statistics Denmark’s algorithm (which,

again, is based on individual-related background information), are the founders of these

firms.

The

FIDA-database

that links firms with individuals.

The Statistics Denmark Migration Database having individual-id’s and emigration and

immigration dates for all (registered) individuals moving in and out of Denmark. For

the project, there is information on approximately 580,000 emigration events of

440,000 different individuals over the period 1973 to 2012.

2.2 The Experian data

The Experian data consists of approximately 1.7 million financial records in the period from

2000 to 2010.

The timing of the records is based on the closing dates of the financial report pe-

riods. In case of firms filing multiple reports in a calendar year, only one of these is selected for

the analysis. The merge of the information from Statistics Denmark and the Experian data is

based on firm registration numbers and the calendar year of Statistics Denmark’s census dates

in November on the one hand and the closing dates of the financial report periods on the other.

2.3 The presence of IM-firms in the Statistics Denmark register data

There are 1,034 registered projects in the DASTI sample, but not all of these are in the Experian

data, and a substantial share is not in the register databases maintained by Statistics Denmark.

Of the 1,034 firms, 44 firms never hand in any financial report, and 55 firms cannot be found in

any of the different Statistics Denmark register datasets. It is most likely that the records in the

DASTI data belong to firms that never reach critical mass for registration at Statistics Denmark.

Of the 1,034 firms, 834 are at some point in time in Statistics Denmark’s

FIRM-database,

683 are in the

FIDA-database,

and

395

are in the Entrepreneurship database. It is remarkable

that such large shares of IM-firms are not in the Statistics Denmark data which is supposed to

collect largely all economic activity.

With regards to the Statistics Denmark Entrepreneurship database, and to a lesser extent

the FIRM-database, the low representation of IM-firms might be due to firms being required to

be above minimum activity thresholds to get sampled in these databases. With regards to the

FIDA-database, part of the attrition might be assumed to individuals in IM-firms having other

jobs (individual-firm-matches) in other firms, and the FIDA-database sampling a maximum two

jobs per individual.

This implies that it should be kept in mind that associating firms with individuals is not

possible for all programme participant firms. As a consequence of the fact that the Statistics

Denmark registers are not able to identify the entrepreneurs behind a share of IM-firms, we

consider the union of entrepreneurs identified by the Entrepreneurship database and the staff

identified by the FIDA database when analysing individual mobility in the second part of the

analysis.

Most firms have their closing date at the end of December, which implies a short time difference between the

Statistics Denmark information (of end-November) and the financial report information. However, there are also

firms that have chosen other dates, e.g. end of March, to close their books. For these firms, the information from

the Statistics Denmark registers comes with a time lag of up to 11 months.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

The extent to which IM-firms are represented in the different databases has implications for the

analysis: for example, both the selection of reference firms for the statistical comparisons is

based on information in the data, and we can thus only analyse success variables for the sub-

samples of firms that have the relevant pieces of information in the data.

In total, 926 IM-firms have information in either the Statistics Denmark or the Experian

data at some point in time, with the number of participating firms in the data increasing from 81

in 1999 to 453 in 2011. Not all IM-participant firms are considered for the analysis: first, firms

started after 2011 are not considered, as 2011 is the last year with full representation in the Sta-

tistics Denmark registers.

Moreover, there are firms started in 1998 for which there is no in-

formation in the Statistics Denmark registers either.

2.4 The treatment of the raw data prior to the analysis

Before being able to analyse the data, there is a need for making decisions regarding the defini-

tion of variables and the treatment of extreme observations.

For example, extracting estimates of firm starting dates from the different data sources is

an exercise in itself, as there is a degree of uncertainty in the data with regards to the year in

which a given firm is founded: establishment year information is collected in the DASTI data,

the business authority’s firm background data (STAMDATA), and Statistics Denmark’s General

firm Statistics and Entrepreneurship database. The information on starting years from these

sources is not necessarily consistent, and may further deviate from the first years the given firm

occurs in the Statistics Denmark FIDA-employer-employee data or hands in its first annual re-

port to the business authority.

In the following, we follow firms after the first year in which they occur in either the Statis-

tics Denmark or the Experian data. This year will be referred to as “year zero” or “base year”. In

the following, firm age is defined by the number of years after the base year, that is, firm age is

zero in the base year.

Firm age is not the only variable that can be defined in different ways. The same is true for

the number of employees. The realisations of this variable will be primarily defined on the basis

of the Statistics Denmark registers, and supplemented with information of the Experian data-

base. Also, there are variables which describe the staff characteristics of the firms, e.g., the aver-

age age of the firm’s individuals and the share of women or employees with a long tertiary-level

education degree. The firms’ individuals are identified on basis of the Statistics Denmark FIDA-

database which links individuals to firms. For firms that are not in the FIDA database, but in the

Entrepreneurship database, individuals are identified on basis of this database instead. E.g., for

the latter group of firms, the firm’s individuals’ average age is defined by the age of the entrepre-

neur who is registered in the Entrepreneurship database.

The analysis requires being able to follow firms over time, and this possibility is obviously not given for very

recent projects.

For one of the IM-firms, DASTI has no date information. For the remaining firms, the DASTI date information is

close to the first year in which the firm occurs in the registers for the first time. For only three percent of observa-

tions, start date information in the two data sources is deviating by more than one year. In absolute numbers,

there are 28 firms that occur in the Statistics Denmark data two or more years after they are registered as started

by either DASTI or the business authority’s firm background data. In three cases, the firm’s first year in the Sta-

tistics Denmark data precedes the information of the DASTI or Experian data. And in 308 cases, the first year of

the firm is in the Statistics Denmark data is just one year later than the registration year of either the DASTI or

Experian data.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Of specific importance for the analysis is the treatment of extreme observations (outliers).

Investments in small innovative firms are risky, and any public programme might be considered

a success, if only it makes the difference between success and failure for one single highly suc-

cessful firm. On the other hand, the statistical tools of the analysis are designed to describe gen-

eral features of the data, and it is impossible to tell whether one single highly successful firm is a

coincidence or a general feature of the data generating process of the programme.

For the evaluation exercise of the analysis, the sample of IM-firms is reduced by a very

small number of firms having less than DKK-50 million in gross profit or earnings before taxes

in year zero, and a very small number of firms on the top of the earning distribution in year zero

with earnings greater than DKK750,000.

Also, industrial foundations (not showing any salient

features in the data) are deleted. This leaves us with 888 IM-firms for the comparison analysis.

2.5 Data limitations

The sampling condition that only active firms can be followed in the data naturally implies that

any potential effects of the IM-programme can only be analysed after year zero. In other words,

the IM-programme might have an effect on firms before they are registered for the first time,

and an effect on firms that never get registered. These potential effects are not covered by the

analysis. The following analysis cannot estimate the importance of the IM-programme on the

formation of the firms.

Also, we ignore the fact that firm closure and registration of a new firms is an easy adminis-

trative exercise. As a consequence, managers might decide to close and re-open start-ups to

qualify for participation in the IM-programme to get access to finance. In this analysis, a firm is

considered as new when it becomes registered as a new firm and occurs in the data as stretched

out above, independently of any potential strategic closures and re-openings.

2.6 Basic descriptive statistics

The following describes the sample of IM-firms in a few dimensions. These descriptions can

stand alone, and qualify our general knowledge on these firms. The variables that are described

is somehow arbitrary, however, we are of course interested to developing some ‘feel’ on these

firms, i.e., their industries, their size, and the characteristics of their employees. Later, we are of

course interested in their performance, for example their survival and growth, but these varia-

bles will be described in comparison to other firms that have not participated in the IM-

programme and considered in section 3 of the report.

We set out with a look at the industry distribution of IM-firms in FIGURE 2.1. Note all

descriptive statistics with the exception of firm growth figures are from year zero, which is the

first year the firm is in the data. All monetary figures are inflated or deflated by the Danish Con-

sumer price index to prices as of 2009.

Not considering firms on basis of year zero characteristics is motivated by the wish to extract the most similar

reference group for comparisons. The firms that were deleted are not characterised by any salient developments

in their success parameters. The DKK750,000 threshold for year zero annual earnings might be considered strict,

however, it only affects a very small number of IM-firms. The deleted firms are characterised by low performance

growth after year zero, with, e.g., annual earnings dropping to below DKK-50 million.

Based on a survey, Oxford Research (2012) conclude that approximately 50 percent of IM-firm would not have

been started in a hypothetical situation of an absence of the IM-programme.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 2.1: Industry distribution of IM-firms

Scien)ﬁc research and development

Computer programming, consultancy and related ac)vi)es

Other

Wholesale trade, except of motor vehicles and motorcycles

Publishing ac)vi)es

Manufacture of computer, electronic and op)cal products

Other manufacturing

Architectural and engineering ac)vi)es; technical tes)ng and…

Manufacture of machineryand equipment n.e.c.

Ac)vi)es of extraterritorial organisa)ons and bodies

Informa)on service ac)vi)es

Manufacture of electrical equipment

Manufacture of basic pharmaceu)cal products and…

Ac)vi)es of head oﬃces; management consultancy ac)vi)es

Retail trade, except of motor vehicles and motorcycles

Financial service ac)vi)es, except insurance and pension…

Manufacture of rubber and plas)c products

Manufacture of chemicals and chemical products

100

150

200

250

We find that large shares of IM-firms are in industries commonly described as high-skilled, with

23 percent of firms in scientific research and development and 20 percent in IT.

We also find, cf. FIGURE 2.2, that the largest numbers of IM-firms were financed at the

turn of the century, with a peak at year 2000 with 150 firms having received finance. Recall

starting years are defined as the first year in which the firm is either in the DASTI, Experian, or

Statistics Denmark data, and may for this reason deviate from earlier investigations. After 2003,

numbers of new firms in the data average at approximately 50 new entries per year. The drop in

the number of observations at the end of the observation period is a result of incomplete data

coverage and not decreasing start-up activity; and firms started after 2011 are not part of the

subsequent analyses – which are supposed to follow firms in the years after start-up.

FIGURE 2.2: IM-firms by the first year they occur in the data

160

140

120

100

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Earlier, it was found that many of the IM-firms are in industries that often are considered as

‘high-tech’ industries.

This is confirmed by an investigation of the individuals behind the IM-firms, the character-

istics of which are described in TABLE 2.1: IM-firms are far more skill-intensive than most other

firms, with an almost ten percent share of individuals associated with these firms having holding

a PhD-degree and almost thirty percent having completed a tertiary (university-level) education.

TABLE 2.1: IM-firms firm-level average characteristics of first-year individuals

Number of

observations

Number of employees

Average age

Share of women

Share of immigrants

Share of highly (tertiary-level) educated

Share of PhDs

798

597

616

640

Standard

deviation

2.59

9.87

0.26

0.21

0.40

0.27

Mean

1.58

37.50

0.10

0.07

0.29

0.09

Minimum

>16

Maximum

<40

<75

Also, we find that IM-firms are registered to be started with on average less than two individuals,

and for a large number of these start-ups, no employees are registered at all.

There are single

firms that are started with more employees than would be expected for start-ups, which might be

the result of these firms being funded as part of organisational changes of existing firms rather

than being ‘greenfield-‘start-ups.

The average age of the individuals involved and the share of immigrants might be consid-

ered as being in the expected range by being roughly equal to the Danish working population

averages. However, the share of women in newly started IM-firms is only ten percent and, thus,

is significantly lower than might be expected on basis of women’s general labour market partici-

pation rates in Denmark being similar to men’s.

There is another important aspect to be kept in mind that is illustrated by TABLE 2.1: the

lack of information for a substantial share of IM-firms in the Statistics Denmark registers. For

example, for almost a third of all firms there is no individual-level information. There are two

reasons for this: first, the firm never reaches sufficient activity levels to become part of the FIRM

statistics database (which has the Entrepreneurship database as one of its subsamples). Second,

no employee has her highest or second highest paid job in the IM-firm. In other words: many of

the IM-firms are closed before reaching critical thresholds for firm registration and employment

in the registers.

We close the description of IM-firms in their first year of registration by taking a look at

their financial figures in TABLE 2.2.

Average turnover (this is variable is from VAT registers and part of Statistics Denmark’s

FIRM database) is at DKK3.6 million, which is quite high but in line with the average number of

employees. There are few firms with very high turnover from the very beginning, but we do not

have access to any additional information which might make us consider these records as firms

that are misreporting or simply registration errors.

Employment information is both from the Experian data and the Statistics Denmark data (variable GF_inkl).

The Statistics Denmark registers allow for maximum two jobs (firm-individual-relationships) per individual, or-

dered by pay, at census date in mid-November.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

We also find that IM-firms are often relatively capital intensive, with an average equity per

employee of approximately DKK1 million and more than DKK2 million total assets per employee

– this might be presumed to be related to the IMs’ capital investments in these firms.

TABLE 2.2: IM-firms financial statistics

Number of

observations

Turnover (DKK 1,000)

Value added (DKK 1,000)

Annual earnings (DKK 1,000)

Total assets; balance (DKK 1,000)

Equity (DKK 1,000)

777

334

484

Standard

deviation

2211.78

524.79

2067.29

2134.87

1648.2

Mean

1220.89

-214.590

-722.58

1304.86

526.59

Minimum

>-4,000

>-40,000

>20,000

>-3,500

Maximum

<26,000

<3,000

<600

<30,000

<20,000

After having described IM-firms in their first year of existence in the data, a straightforward

question is of course how IM-firms are developing over time. In the following, we take a few

looks at some of the most obvious performance criteria: Survival and growth in employment,

turnover, value added and annual earnings (profit).

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

3. Survival and growth of IM-

firms in comparison with a

group of reference firms

In the following, IM-firms are analysed in terms of their survival and growth in employment,

turnover, value added and annual earnings (before-tax profits). Their performance is bench-

marked to a group of similar firms that serve as a point of reference. These other ‘reference

firms’ help evaluating the extent to which the survival and growth of IM-firms reflect any poten-

tial effects of the IM-programme on these variables. The reference firms are selected such that

they resemble the IM-firms in a number of observable characteristics in the first year of exist-

ence.

Before turning to the analysis of the single variables, the following section 3.1 shortly de-

scribes the procedure by which non-participating firms are selected into the reference group.

Readers not interested in the somewhat technical details can jump directly to the analysis of

survival in section 3.2.

3.1 Selection of reference firms for the subsequent comparisons

Prior to the analysis, a reference group of firms is selected from the data for latter comparisons.

We attempt to make the comparisons with the reference group as meaningful as possible, by

selecting firms that can be shown to be highly similar in a number of variables in the first year in

which they occur in the data.

The point of departure for the selection of reference firms is the universe of firms in the

combined Experian-Statistics Denmark firm level data. As a first step, these databases are ad-

justed by only considering firms in the first year in which they occur in the register data, by only

considering firms in industries in which there is at least one IM-firm, and by not considering

firms much larger than the largest IM-firm in its first year of existence.

These conditions are motivated by the objective to make the reference group as similar as

possible to the group of participants in most possible dimensions. The resulting sample is de-

scribed in some dimensions in TABLE 3.1.

It is no surprise that the above-mentioned adjustments to the data are not sufficient to

isolate a reference group of firms that can be claimed to be highly similar to the IM-firms, as

TABLE 3.1 gives evidence of IM-firms and potential reference firms being very different from

each other in their observable characteristics.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

TABLE 3.1: IM firms. Comparison of IM-firms and all other new firms in the Experian-Statistics Den-

mark database that share basic characteristics. New firms in the first year they occur in the combined

Experian/Statistics-Denmark data

firms

Number of employ-

ees

Average age

Share of women

Share of immigrants

Share of highly

(tertiary-level)

educated

Share of PhDs

Turnover (DKK

1,000)

Value added (DKK

1,000)

Annual earnings

(DKK 1,000)

Total assets; balance

(DKK 1,000)

Equity (DKK 1,000)

798

597

616

640

777

334

484

Mean

1.58

37.50

0.10

0.07

0.29

0.09

1220.89

-214.59

-722.58

1304.86

526.58

Std

2.59

9.87

0.26

0.21

0.40

0.27

2211.78

524.79

2067.29

2134.87

1648.20

Minimum

-3913

-38168

-3406

Maximum

25742

2749

574

28002

17184

Firms in the adjusted sample

388,633

332,108

348,544

351,716

222,605

90,033

152,629

152,990

152,976

Mean

1.67

37.99

0.29

0.08

0.09

0,00

4,109

864

-151

21,599

9,216

Std

3.28

11.28

0.42

0.25

0.28

0.07

28,600

2,665

2,598

1,912,593

442,419

Minimum

-24,985

-49,570

-18,010

-311,507

Maximum

713,000

24,964

5,000

724,000,000

127,000,000

***

T-

test

Notes: ***; **: significant at 1,5 percent significance level;

*: significant at 10 percent significance level

So there is a need to find more similar reference firms, and this is achieved by means of a ‘pro-

pensity score matching’-procedure. This is standard in the evaluation literature, and e.g. de-

scribed in greater in detail in Kopeinig and Caliendo, 2008.

In short, the procedure first estimates a statistical binary choice model for each firm in the

sample. This allows for calculating the estimated IM-programme participation probability,

called propensity score, on the basis of observed firm characteristics included as explanatory

variables in the binary choice model. Then the procedure selects, for each IM-programme partic-

ipant firm, one non-participant firm into the reference group. This non-participant firm is char-

acterised by (a) being in the same (NACE 2-digit) industry, (b) occurring in the data in the same

year for the first time and (c) having as similar as possible a propensity score as the participant

firm.

The statistical binary choice model used to calculate the propensity scores is displayed in

TABLE A.1 in the appendix of this report. Its main variables are again industry, but also individ-

ual characteristics like the share of female or highly educated employees, or average age and

financial characteristics.

It is found that there are strong associations of firms’ of IM-programme participation

probabilities with their worker characteristics. For example, a low share of women and a high

share of highly educated employees significantly increase the probability of IM-programme par-

ticipation. Also, financial indicators are important, with relatively weak initial performance in

terms of value added and net income increasing the probability of any randomly chosen firm in

the sample of TABLE 3.1 being a IM-programme participant firm.

In practice, the selection of reference firms iterates on the specification of the binary choice

model. A list of observable variables for which the highest possible similarity is wanted is speci-

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

fied, and explanatory variables are added to the binary choice model until firms in the reference

group share most of their observed characteristics with the group of IM-firms.

Again, it should be acknowledged that individual-level information can only be included in

the selection procedure for firms the individuals of which are in the data. We have seen earlier,

that this is not the case for a considerable share of firms in the data. In these cases, matching is

on other, available variables like industry or financial information.

The success of the matching procedure is just the similarity of IM-firms and firms in the

reference group in terms of observable characteristics – in the first year the given firm is in the

data. It might be noted at this place that, as a result of balancing on industry and starting year in

the procedure, IM-firms and reference firms are exactly equal in their industry distribution (at a

2-digit NACE level) and the distribution of starting years.

Other variables are compared in TABLE 3.2. It is found that a number of firms in both

groups of firms (IM-firms and reference firms) have missing information for a couple of varia-

bles. This should be kept in mind when interpreting the results of the analysis. Besides that,

average values of IM-firms and reference firms are highly similar, with only one of the t-test

statistics indicating a significant difference for the share of highly educated.

The statistical distributions of the variables can be tested by Kolmogorov-Smirnov (KS)

tests. These indicate differences for the distributions of financial variables. So although averages

are highly similar, there are for example relatively more firms in the participant group with rela-

tively poor net income figures in the first year they occur in the data. So this should also be kept

in mind for the subsequent analysis.

It might also be noted that a number of participant firms are matched to the same refer-

ence firms; these reference firms’ records are weighted accordingly in following analysis – simp-

ly by expanding the records (i.e. creating duplicates) of the relevant reference firms by the num-

ber of times they have been chosen as reference firms.

In terms of interpretation of later results, it can be noted that, if the two groups were in-

deed equal on average in all their characteristics, any differences in outcomes between the IM-

firms and firms in the reference group would be the IM programme’s causal effect. In the pre-

sent case, there is equality in industry, and a high similarity in other variables like firm size. But

there are of course differences between the two groups of firms in unobserved variables. And

some of these might be correlated both to IM-programme participation and later performance,

and it is thus suggested that the reference group is used as a point of reference for benchmarking

and not a means to draw causal inference.

Similarity is evaluated by two-sided t-tests. Testing for statistically significantly differences between IM-firms

(“treatments”) and reference firms (“controls”) by two-sided t-tests is a standard procedure in the literature.

However, note that this test of non-rejection of the null of equality is not equivalent to ‘proving’ equality: instead,

it just indicates that average differences in the two groups are small relative to the statistical variation of the vari-

able in question.

Growth statistics of surviving firms will be relative to the given firms’ realisations of the success variables in year

zero, so any initial level differences net out. However, firm survival statistics will be affected by different distribu-

tions of year 0 characteristics.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

TABLE 3.2: Comparison of IM-firms and reference firms

firms

Number of employ-

ees

Average age

Share of women

Share of immigrants

Share of highly

(tertiary-level)

educated

Share of PhDs

798

597

616

640

Mean

1.58

37.50

0.10

0.07

0.29

0.09

Std

2.59

9.87

0.26

0.21

0.40

0.27

Minimum

Maximum

Reference firms

797

591

620

643

Mean

1.54

38.13

0.09

0.32

0.05

Std

2.27

10.11

0.23

0.43

0.20

Minimum

Maximum

T-test

(p-

values)

0.793

0.276

0.697

0.532

0.253

0.001***

0.475

0.539

0.923

0.551

0.873

KS-test

0.095*

0.971

1.00

0.23

0.141

0.024**

0.001***

0.000***

Turnover (DKK 1,000)

777 1,221 2,212

25,742

783 1,324 3,369

42,226

Value added (DKK

334

-215

525

-3,913

2,749

345

-180

912

-7,183

1,610

1,000)

Annual earnings (DKK

484

-723 2,067

-38,168

574

475

-708 2,635

-42,224

540

1,000)

Total assets; balance

484 1,305 2,135

28,002

476 1,211 2,716

23,987

(DKK 1,000)

Equity (DKK 1,000)

484

527 1,648

-3,406

17,184

476

506 2,340

-2,868

23,119

Notes: T-test: Two sided t-tests with unequal variances. KS-test: Kolmogorov-Smirnov tests*, **, ***: statistically significant at 10,

5 and 1 percent level.

3.2 Survival

Understanding survival is a crucial part of any empirical analysis on young firms. And firm exit

is indeed the prime reason for why IM’s investments need to be written off, so it is important to

have an overview of the performance of IM-firms in terms of survival.

Yet, it is important to keep in mind that policy will typically be based on potential growth

effects of a programme, which is a combination of the survival of firms, and the growth in surviv-

ing firms. If survivors show exceptionally high growth, then low survival rates themselves may

not be considered any major problem.

In the following, three different definitions of firm transitions are

considered:

Firm exits: Firms that leave the combined Experian-Statistics Denmark data before

2011. This transition category makes no effort to distinguish different reasons of

why the given firms leave the data. So ‘firm exit’ covers over, e.g., both organisa-

tional transitions like IPOs on the one end of the scale and bankruptcy on the other.

Firm closures: Firms that the Danish Business Authority registered as closed

(“ophørt” or “opløst”), liquidated (“likvidation”), or subject to enforced winding-up

(“tvangsopløsning”) or bankruptcy (“konkurs”).

Bankruptcy (“konkurs”). This is the most precise measure of business failure in the

present data.

ii)

iii)

Note the different definitions are nested: all bankruptcies are closures, and all closures are exits.

Exits and (to a lesser extent) closures can take place as the result of organisation changes. The

background information used

to define the firm closure and the bankruptcy events are from the

Business Authority’s files that are part of the Experian data delivery. Unfortunately, these data

do not have any information on firm sales; these might be assumed to be the most prominent

reason for firms exiting the data without being registered as closures.

To follow firm transitions, it is necessary to identify the year of the transition from active to

inactive. This is achieved by defining a firm as active as long as it hands in an annual financial

report to the business authority or is sampled in any of the Statistics Denmark databases. The

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

latter are having information up to 2011, thus, we can only follow transitions from active to inac-

tive up to the years 2010 to 2011.

A first look at the data reveals that up to the most recent information in 2013, 492 (55 per-

cent) of the 888 IM-firms in the combined treatment-control data are categorised as firm clo-

sures. The corresponding number for firms in the reference group is 286 (32 percent). And 180

IM-firms (20 percent) are categorised as bankruptcies, against 97 reference firms (11 percent).

All raw transition data up to 2010 is collected in TABLE A.2.1 in the appendix. FIGURES

3.1.A-C illustrates the results of the table by calculating firms’ survival shares as functions of

calendar year and firm age. For example, FIGURE 3.1.A graphs columns C/A against D/B and

FIGURE 3.2.C is depicts columns G/A and H/B.

FIGURES 3.1.A-C show that approximately 10 percent of all IM-firms leave the data as

exits per year, that approximately ten percent are registered as closures every year, and that

approximately 4 percent of firms are subject to bankruptcy. IM-firms’ experience highest transi-

tion rates in the first half of the observation period, and decreasing rates at its end. It is not pos-

sible to make statements on whether or not IM-firms’ transition rates are ‘high’ or ‘low’, howev-

er, it should be noted that they are higher than the transition rates of the reference group.

FIGURE 3.1.A: The shares of firms that leave the combined Experian-Statistics Denmark data between

year t and t+1

0,25

0,2

0,15

0,1

0,05

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Reference firms

IM-firms

In 2010, which is the last year with full coverage in the Statistics Denmark data, 445 IM-firms have left the data

as exits, against 378 of the reference firms.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 3.1.B: The shares of firms that leave the combined Experian-Statistics Denmark data between

year t and t+1 as closures

0,2

0,18

0,16

0,14

0,12

0,1

0,08

0,06

0,04

0,02

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Reference firms

IM-firms

FIGURE 3.1.C: The shares of firms that leave the combined Experian-Statistics Denmark data between

year t and t+1 as bankruptcies

0,08

0,07

0,06

0,05

0,04

0,03

0,02

0,01

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Reference firms

IM-firms

Knowing the shares of firms that leave at given firm ages allows calculating empirical survivor

functions. These are summarized in FIGURES 3.2.A-C. Approximately 30 percent of a cohort of

IM-firms is found to be expected in the data 12 years later. Again, it is not possible to make any

statements on whether this number is high or low, but, it is again the case that IM-firms have

higher transition rates compared to the firms in the reference group. This exit difference be-

comes even more apparent when considering not just firm exits from the data, but firm exits that

are categorized as closures or bankruptcies in the business authority/Experian data.

FIGURES 3.2.B and C consider firms as surviving in case of no closure or bankruptcy.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 3.2.A: Empirical survivor function: Shares of firms that stay (i.e. are not subject to exit) in the

combined Experian-Statistics Denmark data (firm age on horizontal axis)

0,9

0,8

0,7

0,6

0,5

0,4

0,3

0,2

10 11 12

Reference firms

IM-firms

FIGURE 3.2.B: Empirical survivor function: Shares of firms not subject to closure in the combined Ex-

perian-Statistics Denmark data (firm age on horizontal axis)

0,9

0,8

0,7

0,6

0,5

0,4

0,3

0,2

Reference firms

IM-firms

FIGURE 3.2.C: Empirical survivor function: Shares of firms not subject to bankruptcy in the combined

Experian-Statistics Denmark data (firm age on horizontal axis)

0,9

0,8

0,7

0,6

0,5

0,4

0,3

0,2

Reference firms

IM-firms

We conclude the elaboration on firm transitions by taking a look at the statistical robustness of

the findings of TABLE A.2.1 by the means of a simple discrete-time hazard models.

The coeffi-

This follows the simple discrete-time estimation set-up suggested by Jenkins, 1995. Note that exponentiated

coefficents of the logit models of the table approximate the increases in the predicted probabilities of any of the

exit events taking place at a given firm age.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

cient estimates of these models approximate the percentage-wise changes in the transition prob-

abilities in association with firm age and whether or not the firm is an IM-firm or not. Results

are summarized in TABLE A.2.2 and show that differences in the different transition probabili-

ties between the two groups of firms are outside the ‘normal’ variation, and, thus, statistically

significant.

So these simple statistical tests confirm that IM-firms must be regarded as having

higher exit probabilities than other firms sharing basic characteristics.

3.3 Employment growth

Employment growth is one of the most simple and robust measures of additional economic ef-

fects that can be expected to be created by the IM-programme. Also, employment growth might

be considered the most prominent success measure brought out by the advocates of Innovation

support programmes.

For the analysis, employment information is obtained by merging the Experian data and

Statistics Denmark’s FIRM statistics database.

Firms’ employment figures are following a

skewed distribution, with a large share of firms not showing any positive number of individuals

being associated with the firm in the data. For 2011, median employment is two jobs for IM-

firms and one job for reference firms, mean employment is 4.2 employees for IM-firms and 4.9

employees for reference firms, and maximum employment is 114 for IM-firms and 178 for refer-

ence firms. Recall we only consider firms with a maximum number of 30 employees in the first

year they occur in the data.

There are a few highly successful firms in terms of employment growth in both the group of

IM-firms and the group of reference firms. These firms realize employment increases of more

than a hundred employees over a period of a few years. Although impressive, these develop-

ments are still considered to be in a realistic range and not to give any reason to doubt the validi-

ty of the employment information in the data. So these ‘outliers’ are kept in the data for the in-

vestigations.

At first, it might be noted that we do not observe single firms with any significant job de-

struction taking place when they exit the data.

For IM-firms, the largest number of employees

in the last year of firm existence (before 2011) is 22 jobs. The corresponding number for refer-

ence firms is six jobs. So, spurious job destruction in association with for example firm sales is

not suggested to be any major issue in the data.

By construction, IM-firms and reference firms in the data have the same distribution of

starting years, i.e., the years in which they occur in the data for the first time. In the following,

we distinguish two kinds of job creation of IM-firms and reference firms: (a) job creation before

being in the data for the first time, and (b) subsequent job creation. The job creation before

being in the data for the first time is equivalent to the number of jobs at firm age zero, which we

defined as the first year in which the firm is in the data.

This finding is robust to including characteristics of the firms when they were started into the models. These

year zero characteristics are statistically significantly related to the exit probabilities, but do not change the esti-

mates related to the comparisons of participant and reference firms.

Strictly speaking, it is the number of staff and not the number of employees that is measured by these databases.

So employment numbers includes the entrepreneur or the owner of the firm, as far as he or she is registered in

the data.

The year 2011 is the year after which there is no employment registration by Statistics Denmark in our data,

which implies a break in the way employment data is registered for the project between 2011 and 2012.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

So the number of jobs at firm age zero is the number of jobs in the firm the first time it is in

the data, and will be referred to as job creation in association with start-up. And the job creation

after firm age zero will be referred to as job creation after start-up.

Employment developments – and the developments of all other success variables of the

report – can be followed over calendar time and by firm age. TABLES A.3.1 and A.3.2 in the

appendix of this report collect all employment information of all entry cohorts - aggregated and

accumulated over firm age and calendar year. Rather than solely concentrating on surviving

firms, these tables have the advantage of incorporating employment losses associated with firm

exit.

FIGURES 3.3.A-C illustrates the main findings of tables TABLES A.3.1 and A.3.2. First, the

development in the number of employees in the combined Experian-Statistics Denmark data

over the observation period is illustrated in FIGURE 3.3.A. There is an increase in employment

in IM-firms, approaching approximately 1,650 jobs being registered at the end of the observa-

tion period.

IM-firms have generated fewer jobs than the firms in the reference group. This is mostly

due to different increases in the early years of the observation period. The gap in the aggregate

employment figures is closing over time, indicating that later cohorts of IM-firms are character-

ised by higher employment creation compared to the reference firms.

Up to 2011, IM-firms have generated 1,258 jobs before being in the data for the first time.

Thus, until 2011, IM-firms have created (1,644-1,258=) 386 jobs on top of the job creation in

association with start-up. The corresponding number for firms in the reference group is 444.

So the most prominent share of job creation in both IM-firms and reference firms takes

place in association with firm start-up, i.e. before firms are in the data for the first time. See

FIGURE 3.3.B for a summary of job creation on top of job creation in the first year of the firms’

existence. This figure also illustrates that job creation in the more recent years of the observation

period has slowed down: in the period 2006-2011, IM-firms can be shown to have generated 40

additional jobs and reference firms have even reduced employment by 54 jobs.

Up to 2010, all newly started IM-firms of age zero years had a total number of 1,130 jobs.

And up to 2011, the same IM-firms, now of firm age 1, had a total number of 1,818 jobs. In other

words, the aggregate employment between firm age zero and firm age has increased by

(1,818/1,130=1.61) 61 percent. And up to 2010, all IM-firms of age 1 had a total number of 1,701

jobs. The year after, the same firms had a total number of 1,841 jobs. In other words, the aggre-

gate employment between firm age 1 and firm age 2 years has increased by (1,841/1,701=1.08) 8

percent.

The rightmost column of TABLE A.3.2 and FIGURE 3.3.C continue this exercise for all

firms of all ages. FIGURE 3.3.C uses the growth rates to depict the employment developments of

both IM-firms and reference firms relative to employment at age zero. It is found that employ-

ment increases on top of first-year employment are taking place in young firms. After approxi-

mately three years, the group of IM-firms starts decreasing its number of employees, and after

more than approximately nine years, IM-firms’ aggregate employment is lower than it was at the

first time they figured in the data. The developments in the reference group are more positive in

the sense of sustaining higher employment levels, but follow the same negative long-term trend.

In comparison with reference firms, IM-firms grow faster in their first years. This finding

would suggest a higher job creation in the first years of the observation period than shown in

Additional investigation suggests that this increase is not a result of single firms (outliers) or large IM-firms

started with more than five individuals.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 3.3.B. Instead, the seeming divergence of FIGURES 3.3.B and C is evidence of more

recent cohorts of IM-firms being more successful in terms of job creation than earlier ones. Yet,

the data does not allow for the conclusion that IM-firms and reference firms generate sustaina-

ble employment growth on top of the job creation in association with firm start-up.

FIGURE 3.3.A. Employment in the combined Experian-Statistics Denmark database by year

1.800

1.600

1.400

1.200

1.000

800

600

400

200

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

IM-firms

Reference firms

FIGURE 3.3.B: Employment in the combined Experian-Statistics Denmark database by year (net of

employment in the first year the firms occur in the data)

800

700

600

500

400

300

200

100

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

IM-firms

Reference firms

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 3.3.C: Employment in the combined Experian-Statistics Denmark database (aggregate em-

ployment relative to employment at firm age zero years (=100%) and firm age on horizontal axis)

200%

180%

160%

140%

120%

100%

80%

60%

40%

20%

IM-firms

Reference firms

Obviously, the aggregate job creation development is hiding a large amount of firm heterogenei-

ty in employment growth, and are combinations of some firms growing and others failing. To

shed light on this issue, the subsequent elaborations on employment increases will follow indi-

vidual firms rather than considering aggregate numbers.

These considerations allow for a higher precision with regards to establishing statistical

significance. However, it needs to be kept in mind that any conclusions based on the analysis of

individual firms only apply to the subsample of firms that survive in the data, and are thus suf-

fering from what is termed survivorship bias in the literature on evaluation design.

The first evidence on surviving individual firms’ employment growth is depicted in FIG-

URE 3.4.A, which shows that surviving IM-firms are on average increasing their number of jobs

by between four and six in the years after start-up.

These numbers are similar to those of the reference group. There is a drop in the average

number of employees in the group of IM-firms at firm age seven to eight years. As can be seen in

FIGURE 3.4.B, this is a result of a limited number of successful firms in both groups reducing

their number of employees or not being observable in the data for more than seven years.

FIGURE 3.4.A: Number of employees in participant and reference firms subtracted the number of

employees in year 0 (means and firm age on horisontal axis)

IM-firms

Reference firms

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 3.4.B: Number of employees in high-growth participant and reference firms subtracted the

number of employees in year 0 (percentiles and years after year 0 on horizontal axis)

IM-firms, 90pct

Reference firms, 90pct

IM-firms, 95pct

Reference firms, 95pct

We can conclude that the average surviving IM-firm generates approximately five jobs on top of

the job creation in association with firm foundation. The comparison of IM-firms with similar

start-ups suggests that surviving IM-firms are creating slightly more jobs than other similar

start-ups, except for (a highly reduced number of) firms above age ten years.

The distribution of job creation surviving firms is also illustrated in FIGURES 3.5.A and B.

These show the cumulative empirical distributions of employment increases between firm age 0

and firm age 5 and 10 years, respectively. These distributions are more to the right for IM-firms,

indicating relatively more firms creating additional jobs when compared to the selection of ref-

erence firms. But differences are small and there are a few firms in the reference group with

employment growth above 25 employees that find no equivalents in the group of IM-firms.

FIGURE 3.5.A: Employment increases, firm age: 5 years

FIGURE 3.5.B: Employment increases, firm age: 10 years

Empirical CDF

-40

-20

Reference firms

IM-firms

-40

Empirical CDF

-20

Reference firms

IM-firms

Regressions aiming to establish statistical significance fail to detect statistically significant dif-

ferences in IM-firms’ and reference firms’ employment growth patterns: TABLE 3.3 summarizes

the results of a model that might be considered the most simple and robust to analyse job crea-

tion differences: it simply takes the most recent year in the data, which is 2011, and compares

employment in 2011 on top of employment at firm age zero.

Model 2 and 3 control for firm age and Model 3 controls for a couple of firm characteristics

at firm age zero. Any of the models can establish any evidence on systematic employment growth

differences between surviving IM- end reference firms.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

TABLE 3.3: Firm-level regressions. Dependent variable: Employment in 2011 subtracted employment

at firm age zero years

Variable

IM-firm=1

Firm age (in years)

***

(IM-firm=1)*(Firm age (in years))

Constant

2.57 ***

0.58

762

0.000

0.3339

Conditioning

yes

variables

Notes: ***, **: statistical significance at 1 and 5 percent significance level, respectively. Heteroscedasticity-consistent standard errors.

Model 1

Coeff.

0.21

Ste.

0.75

Model 2

Coeff.

0.43

0.42

0.03

0.42

762

0.0271

Ste.

0.51

0.14

0.17

0.38

Model 3

Coeff.

0.47

1.92 **

0.09

Ste.

0.52

0.82

0.13

We conclude the section on employment growth by summarising its main results:

In total, IM-firms employ approximately 1,600 individuals in 2011. The largest share of job

creation takes part in association with firm start-up rather than employment growth in existing

IM-firms. On aggregate, IM-firms increase employment in the first years of their existence.

However, after approximately three years of firm age, the group of IM-firms is, on average, de-

creasing its numbers of jobs. No evidence can be established of IM-firms being characterised by

sustainable job creation. Surviving IM-firms increase employment by approximately 0.5 jobs per

year.

Over the entire observation period of the analysis, IM-firms cannot be shown to generate

more jobs than highly similar firms selected as a point of reference for comparisons. On aggre-

gate, IM-firms lagged behind in terms of job creation compared to firms in the reference group,

however, they were able to catch up and created more jobs than reference firms in the last years

of the observation period. Surviving IM-firms cannot be shown to be different from firms in the

reference group in terms of job creation.

3.4 Turnover developments

Both turnover and value added, to be covered in the next subsection, are the most obvious

measures of economic activity. These variables allow us to conclude whether or not any poten-

tial additional effects of the IM-programme are reflected in the bottom line results of IM-firms.

In the following it needs to be kept in mind that a share of the firms in the sample can be ex-

pected to be characterised by long product development time horizons. For this reason, there

might be a number of firms not performing very well in terms of turnover and value added in the

first years of their existence.

For the analysis, turnover is measured by Statistics Denmark’s variable GF_OMS. This

variable is collected by VAT registration, and, thus, available for all registered firms above the

minimum activity levels required for being sampled in the Statistics Denmark’s firm databases.

Our turnover information is characterised by outliers, i.e. single firms reporting turnover

figures orders of magnitude higher than the 99

percentile of the turnover distribution. There

are no arguments for dropping these firms from the analysis, as they are, for

example, not con-

centrated in the financial industry or have other common traits that would justify deleting them

from the sample. Instead, the sample is divided into different subsamples with different treat-

ments to these outliers:

All IM- and reference firms.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

ii)

All IM- and reference firms, except the largest five in terms of turnover in each

group in any given year.

Before moving to the developments in the turnover variable, it might be noted that average

turnover difference between IM-firms and reference firms

at firm age zero years is

(1324/1221=1.085) 8.5 percentage points. Further, cf. FIGURE A.2, there is no indication of

larger shares of IM-firms starting their developments in the data with zero turnover than is the

case for firms in the reference group, so there is a good basis for the comparisons.

Aggregate turnover developments are summarised in FIGURES 3.6.A-B.

It is found that

one or a couple of firms in the reference group are experiencing large increases in their turnover

up to the financial crisis, which creates a salient peak in the turnover development figures for

this group of firms.

Disregarding these very few firms (and the five highest turnover firms in the group of IM-

firms) in FIGURE 3.6.C creates a picture that looks familiar from the employment analysis: ref-

erence firms experienced greater turnover than IM-firms until the financial crisis in 2008-2009,

and similar turnover in the last years of the observation period.

Except for the peak in turnover in the reference group between 2004 and 2007, turnover

per employee is approximately DKK 1million for both the group of IM-firms and reference firms,

and, thus, is in the expected range.

FIGURE 3.6.A: Aggregate turnover all firms by year (DKK1,000)

8.000.000

6.000.000

4.000.000

2.000.000

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

IM-firms

Reference firms

This might be important to note, as earlier comparisons of firms receiving venture capital with other firms have

been met with the criticism of venture capital being directed at firms without economic sales activity in the first

years of their existence. And an earlier Danish study on venture capital fails to select firms with similar turnover

levels into the reference group for comparisons (Ernst

and Young, 2010).

For turnover, value added and annual earnings, the report does not present any equivalents to FIGURE 3.3.C.

This is because these variables are flow variables, while employment might be considered a state variable.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 3.6.B: Aggregate turnover alle firms except for top-5 turnover firms by year (DKK1,000)

2.000.000

1.500.000

1.000.000

500.000

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

IM-firms

Reference firms

FIGURE 3.6.C: Aggregate turnover on top of turnover at firm age zero except for top-5 turnover firms

by year (DKK1,000)

1.000.000

800.000

600.000

400.000

200.000

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

IM-firms

Reference firms

Turnover-statistics at the firm level are summarized in FIGURES 3.7.A and B. There is steady

increase in the mean turnover for surviving IM-firms up to firm age approximately six years,

after which we cannot find any positive development in mean turnover figures.

IM-firms and reference firms end up at similar mean turnover levels after seven to eight

years. However, there are a couple of firms in the group of reference firms that are characterised

by high turn turnover growth in their early years.

The ‘erratic’ movements in the mean of reference firms’ turnover developments are the

results of a few outliers: FIGURE 3.7.B suggests turnover developments for the largest share of

firms to follow highly similar trends, with relatively more IM-firms being characterised by high

turnover growth than reference firms.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 3.7.A: Annual turnover in IM-firms and reference firms (means) subtracted the turnover in

year 0 by firm age: years after year 0 on horizontal axis

20.000

15.000

10.000

5.000

10 11

IM-firms

Reference firms

FIGURE 3.7.B: Annual turnover in high-growth IM-firms and reference firms (percentiles) subtracted

the turnover in year 0 by firm age: years after year 0 on horizontal axis

45.000

40.000

35.000

30.000

25.000

20.000

15.000

10.000

5.000

IM-firms, 90pct

Reference firms, 90pct

IM-firms, 95pct

Reference firms, 95pct

Any potential difference in the distributions of turnover increases is investigated closer in FIG-

URES 3.8.A and B, which show the empirical cumulative distributions of turnover increases

between firm age zero and firm age five and ten years. Just like for employment, we find the IM-

firms’ distribution being slightly to the right, and just like in FIGURES 3.7.B, there are more

successful IM-firms that increase turnover by between approximately DKK12 and 20 million

over a ten-year horizon. Yet, distribution differences are, in general, small after ten years and

almost absent after five years, so there is little chance to detect statistically significant differ-

ences.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 3.8.A.: Turnover (DKK1,000) increases.

Firm age: 5 years

FIGURE 3.8.B.: Turnover (DKK1,000) increases.

Firm age: 10 years

Empirical CDF

-20000

-10000

Reference firms

10000

IM-firms

20000

-20000

Empirical CDF

-10000

Reference firms

10000

IM-firms

20000

It is tested whether or not differences in turnover increases are statistically significantly different

between IM-firms and reference firms by simple linear regressions. The results of these regres-

sions are summarized in TABLE 3.4. As expected on the basis of the visual inspections, differ-

ences are not statistically significant from zero, and small R2 statistics of the regressions of

Models 1 and 2 give evidence of IM-programme participation not being able to explain any con-

siderable share of variation in the firms’ turnover developments.

TABLE 3.4: Firm-level regressions. Dependent variable: Turnover in 2011 (DKK1,000) subtracted turno-

ver at firm age zero years

Ste.

IM-firm=1

1,468

1,040

Firm age (in years)

***

217

(IM-firm=1)*(Firm age (in years))

294

Constant

4,784 ***

931

767

717

0.0003

Conditioning varia-

bles

Notes: ***: statistical significance at 1 percent significance level. Heteroscedasticity-consistent standard errors.

Variable

Model 1

Coeff.

657

Model 2

Coeff.

600

688

156

1,264

717

0.0242

Model 3

Coeff.

443

1,206

258

717

0.2338

yes

Ste.

1,094

1,189

243

In sum, we find a significant share of IM-firms with a positive turnover in the first year they are

registered in the data and a significant increase in turnover in IM-firms in the first years of their

existence. There are a couple of highly successful IM-firms in terms of turnover growth, but, in

general, IM-firms’ turnover developments cannot be shown to be inherently different from oth-

er, non-participant firms that share a couple of their first-year characteristics.

3.5 Value added developments

We now turn to the second financial activity measure of the analysis. Value added has the ad-

vantage over turnover that it actually measures how much value the firm has added to its prod-

ucts. This is not the case for turnover, which might be generated without enhancing the products

sold.

For the analysis, value added is from the Experian variable ‘bruttofortjenes-

te/dækningsbidrag’ and defined as turnover subtracted variable costs of production, these are

mainly intermediate inputs.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

In terms of value added no extreme or unrealistic observations are observed in the data.

Just like in the previous subsections, we first take a look at the distribution of this variable at

firm age zero, i.e., the first year the firm is in the data, and compare this distribution with the

group of firms selected as a benchmark group for comparisons.

As a point of departure, we note, cf. FIGURE A.3, that the largest share of IM-firms have

negative value added the first time they are in the data: only fifteen percent of all firms generate

more value from sales than they invest into variable inputs. There is poor reporting of the value

added figures in the data, which implies that the results for this variable apply only for relatively

small shares of all IM-firms and reference firms.

The means of value added differ, see TABLE 3.2, by merely DKK35,000 between IM- firms

and reference firms. They are not statistically significantly different from each other. But the

Kolmogorov test statistic of the same table and the distribution plot of FIGURE A.3 imply that

IM-firms are starting from a weaker position in terms of value added than the reference firms:

there are higher shares of IM-firms that have negative value added and lower shares having

positive value added in their first year than firms in the reference group.

The aggregate annual value added developments are summarized in TABLE A.5.1 and A.5.2

and FIGURE 3.9.A.

It is found that IM-firms are characterised by sluggish aggregated value added growth up

to approximately 2006, after which growth picks up and ends at approximately DKK600 million

at the end of the observation period. This is even slightly higher than for firms in the reference

group, the aggregate value added of which increased faster in the beginning of the observation

period. Both groups are characterised by low value added in young firms, such that FIGURE

3.9.A and 3.9.B almost give the same aggregate results. Thus, the largest share of value added

growth is observed – in contrast to employment growth – to take place in the years after firms

occur in the data for the first time.

The observation of low initial growth – to be followed by high growth – of IM-firms is also

confirmed when looking at the aggregate value added growth in association with firm age in

FIGURE 3.9.C: IM-firms are slow starters, when it comes to value added, but are successful later

in creating aggregate value added growth. This suggests that IM-firms are characterised by long-

er product development time horizons than reference firms.

Furthermore, it is found that, although firm numbers are strongly decreasing in firm age,

aggregate value added growth is still positive. And although relative more IM-firms leave the

data than firms in the reference group, their aggregate value added increases are still higher than

those of the reference group.

Note that adding t-test statistics to treatment-control comparisons like the one in TABLE 3.2 is standard in the

literature, while adding Kolmogorov-Smirnov statistics and distributional plots is not.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 3.9.A: Aggregate value added (DKK1,000) by year

700.000

600.000

500.000

400.000

300.000

200.000

100.000

-100.000

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

IM-firms

Reference firms

FIGURE 3.9.B: Aggregate value added net of value added at firm age zero by year

700.000

600.000

500.000

400.000

300.000

200.000

100.000

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

IM-firms

Reference firms

FIGURE 3.9.C: Aggregate value added developments (firm age on horizontal axis)

800.000

700.000

600.000

500.000

400.000

300.000

200.000

100.000

-100.000

-200.000

Reference firms

IM-firms

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

When we look at the development of individual surviving firms in FIGURES 3.10.A and B, we

find the same pattern as for the aggregate numbers: low value added growth early on, and sub-

stantial value added growth later on.

Here, it is important to note that the picture is more positive for IM-firms than suggested

by FIGURE 3.10.A, that, for reasons of exposition, was truncated at 11 years of firm age: at firm

age 12 years, average value added is 8.8 million DKK in IM-firms against DKK 300,000 for firms

in the reference group.

Findings lend themselves to the interpretation of IM-firms following more patient business

models than the firms in the reference group.

FIGURE 3.10.A: Value added (DKK1,000) in participant and reference firms (means) subtracted value

added at firm age 0 (years after year 0 on horizontal axis)

4.000

3.500

3.000

2.500

2.000

1.500

1.000

500

-500

IM-firms

Reference firms

FIGURE 3.10.B: Value added (DKK1,000) in participant and reference firms (percentiles) subtracted

value added at firm age 0 (years after year 0 on horizontal axis)

70.000

60.000

50.000

40.000

30.000

20.000

10.000

IM-firms, 90pct

Reference firms, 90pct

IM-firms, 95pct

Reference firms, 95pct

The finding of surviving IM-firms showing more positive value added developments is further

confirmed by the empirical cumulative distributions summarized by FIGURE 3.11.B. However,

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Model 2 of TABLE 3.5 finds low statistical significance of this finding and Model 3 finds that

statistical significance is not robust to the inclusion of control variables. So the positive results of

the analysis of surviving firms still need to be interpreted as tentative.

FIGURE 3.11.A: Value added (DKK1,000) increases.

Between firm age 0 and 5 years.

5000

Reference firms

10000

IM-firms

15000

FIGURE 3.11.B: Value added (DKK1,000) increases.

Between firm age 0 and 10 years.

-5000

-10000

Empirical CDF

10000

Reference firms

20000

IM-firms

30000

TABLE 3.5: Firm-level regressions. Dependent variable: Value added in 2012 (DKK1,000) subtracted

vaue added at firm age zero years

Model 1

Variable

IM-firm=1

Firm age (in years)

(IM-firm=1)*(Firm age (in years))

Constant

Conditioning variables

Coeff.

941

Ste.

770

Model 2

Coeff.

-1,273

191

409

349

0.0639

Ste.

847

244

281

Model 3

Coeff.

-1,041

-476

328

349

0.2893

yes

Ste.

965

927

261

1,247

349

0.0048

***

352

Notes: ***, **, *: statistical significance at 1, 5 and 10 percent significance level, respectively. Heteroscedasticity-consistent standard errors.

3.6 Annual Earnings

We now turn to annual earnings. This is a measure of profitability and financial performance,

and can be related to the financial returns of investors like, in this case, the IMs.

Annual earnings are measured by earnings before interest and taxes (ebit). This variable is

from the Experian data. There are a couple of records in the group of IM-firms with very high

earnings in certain years, but these can be confirmed as related to transactions of highly success-

ful firms that are typically in the pharmaceutical industry.

In the first year firms are in the data average annual earnings are small in absolute value.

Firms in the reference group almost perfectly match IM-firms in terms of average net income,

but have, cf, FIGURE A.4 in the appendix, lower shares of low-income firms.

With the exception of 2012, which has witnessed a couple of successful business transac-

tions of IM-firms, on average IM-firms have making negative annual earnings. See FIGURE

3.12.A and B. Over the time period 1999 to 2012, annual earnings in IM-firms add up to an ac-

cumulated loss of approximately DKK 5 billion (cf. FIGURE 3.12.C). IM-firms and reference

firms accumulate similar losses in the first half of the observation period. After approximately

2005, IM-firms accumulate more losses and are thus indicated to be more cost-intensive relative

to value added when compared to the firms in the comparison group.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 3.12.A: Aggregate annual earnings (DKK1,000) by year

400.000

200.000

-200.000

-400.000

-600.000

-800.000

-1.000.000

IM-firms

Reference firms

FIGURE 3.12.B: Aggregate annual earnings net of earnings at firm age zero by year

600.000

400.000

200.000

-200.000

-400.000

-600.000

-800.000

IM-firms

Reference firms

FIGURE 3.12.C. Accumulated total annual earnings (mio. DKK)

1999 2001 2003 2005 2007 2009 2011

-1.000

-2.000

-3.000

-4.000

-5.000

-6.000

Reference firms

IM-firms

When we look only at surviving firms in FIGURE 3.13 A and B, we find that neither IM-firms nor

reference firms manage to make positive earnings in the first years after start-up. Recall both

IM-firms’ and reference firms’ average earnings are approximately minus DKK700,000 in year

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

It is also found that there is a share of highly successful firms that have survived for more than

nine years and make positive profits that find no equivalents in the group of reference firms. The

highly successful IM-firms’ earnings contribute to bringing average earnings (subtracted year 0

earnings) of surviving IM-firms up to a maximum of DKK 6 million for firms of age twelve years

in our data. Unfortunately, only relatively few IM-firms (N=32) can be followed for 12 years in

the analysis’ data.

FIGURE 3.13.A: Annual earnings (DKK1,000) in IM-firms and reference firms (means) subtracted earn-

ings in year 0 (years after year 0 on horizontal axis)

8.000

6.000

4.000

2.000

-2.000

-4.000

IM-firms

Reference firms

FIGURE 3.13.B: Annual earnings in IM-firms and reference firms (percentiles) subtracted annual earn-

ings in year 0 (year after year 0 on horizontal axis)

25.000

20.000

15.000

10.000

5.000

IM-firm 90pct

Reference 90pct

IM-firm 95pct

Reference 95pct

The distributional plots of FIGURE 3.14.A and B confirm what we have seen already: that there

are large shares of IM-firms with low earnings increases in the medium run at firm age 5 years,

and some IM-firms being characterised by very high annual earnings after 10 years.

This finding is further confirmed by the regression results of Models 2 and 3: IM-firms

decrease profits in their first years of existence, but, according to Model 2, catch up and end up

with higher profit increases than firms in the reference group in the long run. Yet, just as it was

There are 12 IM-firms that can be observed for 13 years, these have annual earnings of on average DKK1.05

million.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

the case for value added, this finding is tentative, and statistical significance does not prevail

after inclusion of conditioning variables into the regression of Model 3.

FIGURE 3.14.A: Annual earnings(DKK1,000) increases.

Between firm age 0 and 5 years.

-5000

Reference firms

5000

IM-firms

10000

-10000

Empirical CDF

FIGURE 3.14.B: Annual earnings (DKK1,000) increases.

Between firm age 0 and 10 years.

-10000

Empirical CDF

-5000

Reference firms

5000

IM-firms

10000

TABLE 3.6: Firm-level regressions. Dependent variable: Annual earnings in 2012 (DKK1,000) subtracted

annual earnings at firm age zero years

Model 1

Variable

IM-firm=1

Firm age (in years)

(IM-firm=1)*(Firm age (in years))

Constant

Conditioning variables

-896

486

0.0035

989

Coeff.

3,098

Ste.

2,569

Model 2

Coeff.

-3,486 **

-202

1,064

470

486

0.0036

Ste.

1.411

196

634

388

486

0.0464

yes

Model 3

Coeff.

-3,734

103

1,174

Ste.

2,622

1,587

838

Notes: **, *: statistical significance at 5 and 10 percent significance level, respectively. Heteroscedasticity-consistent standard errors.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

4. Individual mobility and

generation of start-ups

As a final exercise in this report, we add an individual-dimension to the analysis. Little is known

about how many individuals of IM-firms start new firms, i.e., whether IM-firms act as ‘hothous-

es’ for additional start-up activity. It is a fact that a large share of IM-firms fail. But our view on

this might change if there is a lot of new entrepreneurial activity in the wake of firms that have

closed down.

4.1 Data on individuals in the Statistics Denmark Databases

Matches between individuals (i.e. workers and founders of firms) and firms are identified by the

Statistics Denmark FIDA-database that samples all Danish firm-worker employment relation-

ships and up to two firm-worker relationships per individual.

Although the FIDA-database has information on entrepreneurs it should be kept in mind

that, not all entrepreneurs are in the FIDA-database. For example, approximate 25 percent of

the individual-firm matches that the Statistics Denmark Entrepreneurship database sample as

entrepreneurs are not in the FIDA-database. This might be due to the fact that a number of firms

in the Entrepreneurship database do not survive until the FIDA-databases census date at mid-

November each year.

The following section analyses the mobility of the individuals who are in the FIDA-

database. However, in cases where it is possible and where it makes sense to do so, we add indi-

vidual-firm matches from the Entrepreneurship database – which, according to its sampling

scheme only considers individual-firm matches in the data for the first year of firm existence.

For the group of IM-firms, there are 11,488 annual records of firm-worker relationships, or

jobs, in the FIDA-database over the time period 1999-2010. These records belong to 626 differ-

ent IM-firms (against 683 for the time period 1999-2011 mentioned in section 2.3) and 5,042

different individuals (and 5,631 for the time period 1999-2011). The average firm stays for ap-

proximately 3.5 years in the database, and the average individual stays with the IM-firm for

approximately two years in the database.

The average size of IM-firms in the FIDA database increases from 2.5 in 1999 to approxi-

mately six individuals in 2010, which is in line with what we have seen in the previous section of

the report. A number of individuals in IM-firms are registered as having other jobs at other

firms: for 2,716 (24 percent) of the 11,488 records, the individual has a different job somewhere

else in the FIDA-database.

4.2 Job mobility and firm creation

It is found that large shares of individuals in IM-firms leave the IM-firms every year. As FIGURE

4.1 illustrates, approximately 40 percent of total (FIDA-database) staff quit IM-firms every year,

and approximately 10 percent leave the FIDA database. A share of individuals leaving their firms

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

must be presumed to be leaving in association with firm closure; we will elaborate on this pre-

sumption further below.

FIGURE 4.1: Shares of individuals who quit IM-firms between t and t+1

0,6

0,5

0,4

0,3

0,2

0,1

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Quit IM-portfolio firm and leave FIDA

Quit IM-portfolio firm

Of all 3,488 (FIDA-database) individuals who quit to another firm in a given year t, 595, or 17

percent, are in firms that occur in the FIDA or in the IV database in year t+1 for the first time.

And 10.9 percent move to firms started in year t.

Thus, approximately one fourth of all employees who leave IM-firms are moving to new

firms or firms of age below two years. This measures the propensity of IM-firms to generate new

firms. We do not know the specific roles of these individuals in the IM-firms from which they

are originating or in the new firm. So we might not want to call these individuals for ‘serial en-

trepreneurs’.

4.3 New firm creation by the staff of IM-firms

In the combined IV-FIDA database there are a total of 537 new firms in the period 2000-2011

with individuals who were associated with IM-firms in the previous year. This number is the

maximum number of potential (first-generation) spin-off firms, and might be related to the total

number of 626 IM-firms present in the FIDA database.

This indicates that, on average, there is roughly one new firm started by or with IM-firm

employees for each IM-firm in the FIDA-database. Of course the employees of some IM-firms

are more likely to be in new firms in the following year than others, and not every IM-firm sup-

plies individuals who are moving to new firms.

The share of IM-firms in the FIDA-database in a given year for which at least one of the

firm’s employees are in a new firm in the year after is approximately 25 percent. For example,

there are 197 IM-firms in the FIDA database in 2005, and 45 newly started firms in the IV-

FIDA-database in 2006 with individuals that were associated with an IM-firm in the previous

year. So, for 2005, (45/197=)23 percent of all IM-firms have at least one individual in a newly

started firm in FIDA in 2006.

See FIGURE 4.2.A for the absolute numbers of new firms with IM-individuals and FIGURE

4.2.B for the average number of spin-offs generated per IM-firm. The latter is calculated as the

number of new firms with IM-individuals in t+1 over the number of IM-firms in year t.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 4.2.A: Number of new firms in FIDA-database in t+1 with IM-individuals (t)

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Number of new firms with former IM-firm-individuals (t+1)

Figure 4.2.B: Number of new firms in FIDA-database in t+1 with IM-individuals (t) over number of IM-

firms (t)

0,4

0,35

0,3

0,25

0,2

0,15

0,1

0,05

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Number of new firms with former IM-firm-individuals (t+1)/Number of

IM-firms (t)

FIGURE 4.2.A and B illustrate non-increasing new firm generation activity in the wake of IM-

firms. This is remarkable, as the number of individuals in IM-firms is increasing strongly from

139 in 1999 to 1,519 in 2010.

And indeed, the share of

individuals

in IM-firms in the FIDA-database that are in new

firms in the FIDA-database in the year after is strongly decreasing over time, cf. FIGURE 4.3. So

it is not the number of individuals that are related to new firm creation activity of employees, but

the number of IM-firms – which indicates that it is a small share of IM-firm-staff, most probably

the founders, that are engaged in entrepreneurship.

Another indication of this is that the approximately 600 different individuals that leave IM-firms to newly start-

ed firms are filling approximately 750 different jobs (individual-firm matches) in these newly started firms over

the period 2000-2011, i.e., many individuals are found in more than just one newly started firm after having left

their IM-firm.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

FIGURE 4.3: Share of individuals in new firms in t+1

0,18

0,16

0,14

0,12

0,1

0,08

0,06

0,04

0,02

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Number of jobs in new firms in t+1/Number of persons in IM-firms

4.4 The number of IM-firm ‘spin-offs’

To elaborate a bit further on the firm creation activity of IM-firm-individuals, we distinguish

new start-ups by their shares of individuals originating from the same previous firms. This

measures the organisational inheritance of established to new firms.

For example, if a large share of individuals in a new firm were previously employed in the

same former firm, the new firm might be considered an “offspring”, or “spin-off” of the former

firm. We follow this notion, and define a spin-off as a new firm in the IV-FIDA-data, in which

more than 50 percent of staff (for two-individual start-ups: both individuals) originates from the

same previous firm in the year before.

In total, there are 316 firm-records in the IV-FIDA data (corresponding to a 14 percent

share of a total of 2,183 firm-records) that generate a total number of 394 spin-offs over the

period 1999-2010; however, 315 (80 percent) of these spin-offs are one-individual start-ups that

are spin-offs by definition. There are 41 two-individual start-ups where one individual was pre-

viously employed in an IM-firm. And there are 38 spin-offs with at least two IM-firm-individuals

moving to the same new firm and constituting more than 50 percent of the staff in the new firm.

We conclude, therefore, that, except for one-individual-start-ups, only a minor share of the

new firms in the combined IV-FIDA-database with previous IM-firm-individuals can be shown

to have strong ties to the IM-firm in terms of worker movements.

4.5 Individual mobility in association with firm exit

The following continues on the basis of the previous subsection on spin-offs and now only con-

siders individual mobility in association with firm exit (defined as firms leaving the FIDA-data).

This is motivated by the wish to help understanding whether an exit of IM-firms is to be

interpreted as business closure, or merely organisational transitions. If, for example, large

shares of individuals were to be found in the same other workplace after exit, this exit might

arguably be a result of a firm sale or merger rather than business dissolution. This is related to

the question of whether high closure rates are ‘real’ or an indication of the presence of closures

with the purpose of getting rid of financial obligations and outside owners.

Of the 626 different firms in the FIDA-database, 420 leave FIDA before 2011. So it is found

that closure rates are higher in the Statistics Denmark employer-employee database when com-

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

pared to the combined Experian-Statistics Denmark database employed in the first section of the

report, where positive employment was not a defining condition for firm survival.

And of the 11,488 jobs in IM-firms in the FIDA-database, 1,196 are in firms that are not

found in the database in the year after. So even though we observe high firm closure rates of IM-

firms, ‘only’ 10 percent of IM-firm jobs are lost every year by firm closure. We saw earlier that

almost 40 percent of IM-firm-employees leave their firm every year, and now we can conclude

that approximately 10 percent of IM-firm-employees leave their firm because it is closing opera-

tion.

Of the 420 closing firms, 225, i.e., more than half of these are registered as having only one

single individual associated with the firm in the last year of its existence. Of these, 166 individu-

als are in other (not necessarily new) firms in the combined IV-FIDA-database in the year after.

It was not investigated what happened to the remaining 59 individuals that might have become

unemployed, moved abroad, or died.

There are a total number of 74 firms with two employees in their last year of existence. It

happens seven times that both of these two employees are moving to the same (not necessarily

new) firm in the combined IV-FIDA-database in the year after. There are 121 firms with at least

three employees in their last year of existence. Of these, in 29 cases more than fifty percent of

employees are in the same (not necessarily new) firm in the year after closure, and in 9 cases,

more than 75 percent of employees are in the same (not necessarily new) firm in the year after.

FIGURE 4.4 summarizes these numbers.

In sum, the present information in the Statistics Denmark employer-employee data does

not suggest that the closures of any significant share of IM-firms are due to organization transi-

tions and not business dissolutions.

FIGURE 4.4: Number of new firms started with individuals from closed IM-firms. By employee move-

ments in association with firm closure (number of IM-firm's last year's staff on horizontal axis)

(4,5)

(6,10)

11+

more than 50 percent of last year staff in same new firm

more than 75 percent of last year staff in same new firm

4.6 Emigration decisions of IM-staff

We conclude the study of individual mobility staff by considering emigration decisions of the

staff of IM-firms.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

First, it can be noted that IM-firm employees are characterised by high international mo-

bility, with approximately 2,000 of the 5,631 IM-employees in the FIDA-database (over the peri-

od 1999-2011), or 35 percent, occurring in the Statistics Denmark international migration data-

base. This database samples all registered migration events from 1970 onward.

The 35-percent share might be related to an 18 percent of all FIDA-employees in 2010

being sampled in the same database. Interestingly, the largest share of migration events of IM-

firm-employees is before they started becoming part of IM-firms.

Typical emigration countries of IM-firm-individuals are the U.S.A., Sweden and the U.K.,

see FIGURE 4.5 for latest emigration information of these individuals over the period 1999-

2012.

Of the 11,520 records of IM-firm-individuals in the FIDA-data, 186, or 1.6 percent, are

registered as moving abroad in the same or in the following year. So it is not emigration that lies

behind the large quit rates that were registered for the IM-individuals in the FIDA-database.

Of the 1,196 records of IM-firm-individuals in the FIDA-database in firms that leave the

data in the following year, 34, or 2.8 percent, are leaving Denmark in the last year of their firm’s

existence in the FIDA-database or the year after. So there is also no indication of any significant

share of firm closures in Denmark being motivated by simply moving activity abroad.

FIGURE 4.5: Destination countries of emigrating IM-firm individuals

140

120

100

4.7 Short summary of the analysis of individual mobility

In short, we find high mobility of IM-firms staff, with approximately 40 percent of individuals

leaving IM-firms every year. Approximately 10 percent are leaving their IM-firm to move to jobs

in other newly started firms, and approximately 10 percent of them are leaving because their IM-

firm it is closing down.

The analysis finds that, on average, roughly one new firm is started by or with IM-firm employ-

ees for each IM-firm in the FIDA-database and that approximately 25 percent of IM-firms in the

FIDA-database in a given year, have at least one of the firm’s employees in a new firm in the

following year. There are only few new firms in the data that inherit groups of individuals from

IM-firms and which may be considered ‘spin-offs’.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

It happens rarely that individuals of closed-down IM-firms are found in the same new firm

in the year after, or migrate outside Denmark in association with firm exit.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

5. References

Berger, A.N., Udell, G.F. (1998), The economics of small business finance: The roles of private

equity and debt markets in the financial growth cycle, Journal of Banking and Finance, Vol. 22,

pp. 613-673.

Bottazzi, L., and Marco Da Rin (2002), Venture capital in Europe and the financing of innovative

companies, Economic Policy, CEPR;CES;MSH, vol. 17(34), pages 229-270.

Caliendo, M. and Kopeinig, S. (2008): Some Practical Guidance for the Implementation of Pro-

pensity Score Matching, Journal of Economic Surveys, 22(1), 31-72.

CEBR (2009), Samfundsøkonomiske effekter af Vækstfondens Vækstkautioner.

Da Rin, M., T. F. Hellmann and M. Puri (2013), A survey of venture capital research. In: George

Constantinides, Milton Harris, and René Stulz (eds.) Handbook of the Economics of Finance.

Volume 2, Part A, 2013, Pages 573–648.

Ernst and Young (2010), Evaluering af Vækstfonden.

Meyer, Thomas (2010), Venture Capital Adds Economic Spice. Deutsch Bank Research.

Hirukawa, M, and M. Ueda (2011), Venture Capital and Innovation: which is first?, Pacific Eco-

nomic Review, 16:4.

Jenkins, S.P. (1995), Easy Estimation Methods for Discrete-Time Duration Models. Oxford Bul-

letin of Economics and Statistics, Volume 57, Issue 1.

Kaiser, U. & Kuhn, J. (2012), "Long-run effects of public–private research joint ventures: The

case of the Danish Innovation Consortia support scheme," Research Policy, Elsevier, vol. 41(5),

pages 913-927.

Kaplan, S. N., and P. Strömberg (2002). Financial Contracting Theory Meets the Real World: An

Empirical Analysis of Venture Capital Contracts, Review of Economic Studies, 1–35.

Kortum, S. and J. Lerner (2000), “Assessing the impact of venture capital on innovation” Rand

Journal of Economics 31, 4, 674-692.

Lerner, J. (2012), Boulevard of Broken Dreams: Why Public Efforts to Boost Entrepreneurship

and Venture Capital Have Failed--and What to Do About It. Princeton University Press

Peneder, M. (2009), The impact of venture capital on innovation behaviour and firm growth.

Venture Capital, Taylor & Francis Journals, vol. 12(2), pages 83-107.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Møen, J. (2004), When subsidized R&D-firms fail, do they still stimulate growth? Tracing

knowledge by following employees across firms, Discussion Papers 399, Research Department of

Statistics Norway.

Oxford Research (2012), Evaluering af Innovationsmiljøerne. Danish Agency for Science, Tech-

nology and Innovation.

Økonomi og Erhvervsministeriet (2010), Udvalget af Risikovillig Kapital (rapport).

Phalippou, L. and O. Gottschalg (2009), The performance of private equity funds, Review of

Financial Studies 22, 1747-1776.

Stuck, B. and M. Weingarten (2005), How Venture Capital Thwarts Innovation, IEEE Spectrum,

April 2005, 50-55.

Van Praag, C. Mirjam, and Peter H. Versloot (2007), What is the value of entrepreneurship? - A

review of recent research . Journal of Small Business Economics, Volume 29, Issue 4, pp 351-

382.

Technopolis (2001), Evaluation of innovation activities: methods and practices. Report of the

Literature Review

Timmermans B. (2010), The Danish Integrated Database for Labor Market Research: Towards

Demystification for the English Speaking Audience. Aalborg. 2010.

Ueda, M. and M. Hirukawa (2009), The tenuous relationship of venture capital and innovation,

VOX - CEPR's Policy Portal.

Vækstfonden (2006), Venturekapital som Vækstdynamo – Effekter af Venturekapital I Danmark

Vækstfonden (2009), Venture skaber Vækst.

Zucker, L. G., M.R. Darby, and M.B. Brewer (1998), Intellectual Human Capital and the Birth of

US Biotechnology Enterprises, American Economic Review 88, 290-306.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

6. Appendix

TABLE A.1: Logit model estimation results. Selected coefficients

Coefficient

Number of employees==1

Number of employees==2

Number of employees==(3,4)

Number of employees==(5,7)

Number of employees==(8,10)

Number of employees==11+

Average age

(Average age^2)/100

Share female

Share of tertiary-level educated

Share of PhDs

Turnover (DKK 1,000)

Value added (DKK 1,000)

(Value added (DKK 1,000,000))^2

Earnings (DKK 1,000)

(Earnings >0) x (Earnings (DKK 1,000,000))^2

(Earnings <0) x (Earnings (DKK 1,000,000))^2

Total assets; balance (DKK 1,000)

Equity (DKK 1,000)

0.2622661

1.222483

1.01227

1.081062

0.3204484

-0.489454

-0.0421354

0.0316906

-1.515709

1.170964

1.802984

-0.000000808

-0.0001053

0.0089117

-0.0000949

-0.0883739

-0.001083

-0.0001342

0.0001171

***

Standard error

0.1272261

0.147239

0.1683967

0.2248791

0.3034104

0.4269707

0.0280739

0.033697

0.1931771

0.1175994

0.2194382

0.00000237

0.0000388

0.0033751

0.0000421

0.0872207

0.0010552

0.0000218

0.0000219

Notes: *, **, ***: statistically significant at 10, 5 and 1 percent level. Additional coefficients: year dummies (15), industry dummies (53),

constant term, dummy variables for missing observations for all above-shown explanatory variables.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

TABLE A.2.1: Firm closures by year and firm age

Number of firms

that are in Experian

and/or DST data for

the last time (exits)

Reference

IM-

firms

362

409

Number of firms

that are in Experian

and/or DST data for

the last time and

registered as

closures

Reference

IM-

firms

217

375

Number of firms

that are in Experian

and/or DST data for

the last time and

registered as

bankruptcies

Reference

IM-

firms

133

Year

Number of firms

that are in Experian

and/or DST data

Reference

IM-

firms

218

293

322

349

370

385

418

439

436

478

510

223

302

335

338

357

369

386

372

400

443

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Total

Firm

age

Total

823

674

546

461

393

340

293

250

208

162

101

823

731

581

461

365

277

221

173

134

362

409

217

375

133

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

TABLE A.2.2: Discrete time hazard (logit) model estimation results. Sample: active firms present in either Experian or Statistics data, 1999-2010, maximum firm age

10 years

Dependent variable: The firm becomes inactive (exit)

Model 1

Model 2

Standard

Coefficients

errors

Firm age=0 (omitted

category)

Firm age=1

Firm age=2

Firm age=3

Firm age=4

Firm age=5

Firm age=6

Firm age=7

Firm age=8

Firm age=9

Firm age=10

IM-firm x Firm age=0

IM-firm x Firm age=1

IM-firm x Firm age=2

IM-firm x Firm age=3

IM-firm x Firm age=4

IM-firm x Firm age=5

IM-firm x Firm age=6

IM-firm x Firm age=7

IM-firm x Firm age=8

IM-firm x Firm age=9

IM-firm x Firm age=10

Constant

Conditioning variab-

les:

Dependent variable: firm becomes inactive as closure

Model 1

Model 2

Standard

Coefficients

errors

Dependent variable: firm becomes inactive as bankruptcy

Model 1

Model 2

Standard

Coefficients

errors

0,341

0.303

-0.269

-0.341

-0.104

-0.454

-0.848

-0.235

0.167

-1.543

-0.883

0.022

0.242

0.609

0.815

0.313

0.542

0.975

0.009

-0.628

0.006

-2.506

***

0.173

0.184

0.225

0.242

0.236

0.279

0.346

0.288

0.272

0.596

0.228

0.162

0.176

0.232

0.254

0.272

0.330

0.401

0.404

0.456

0.921

0.127

1.058

1.018

0.448

0.376

0.617

0.265

-0.126

0.486

0.891

-0.815

-0.172

0.016

0.238

0.602

0.810

0.305

0.536

0.965

-0.002

-0.646

-0.012

***

0.216

0.224

0.258

0.274

0.269

0.307

0.369

0.315

0.301

0.610

0.261

0.162

0.176

0.232

0.254

0.272

0.331

0.401

0.404

0.456

0.921

0.551

0.770

0.253

0.289

0.065

-0.332

0.154

0.640

-1.771

-0.490

0.448

0.618

0.950

1.104

0.803

0.906

1.342

0.402

-0.218

1.116

-3.388

***

0.245

0.246

0.288

0.301

0.312

0.349

0.427

0.374

0.343

1.020

0.304

0.199

0.197

0.259

0.278

0.310

0.366

0.451

0.458

0.482

1.231

0.189

0.486

0.695

0.178

0.182

0.226

-0.003

-0.401

0.090

0.584

-1.829

-0.547

0.458

0.634

0.965

1.107

0.797

0.892

1.332

0.384

-0.249

1.092

***

0.246

0.247

0.289

0.301

0.312

0.350

0.427

0.374

0.344

1.021

0.304

0.200

0.197

0.260

0.278

0.310

0.366

0.452

0.459

0.483

1.231

2.408

1.962

1.920

1.777

2.338

1.358

1.513

1.219

3.169

-11.777

0.920

0.050

0.708

0.837

1.666

0.230

1.328

1.408

0.798

-0.918

13.127

-6.094

***

0.745

0.777

0.793

0.819

0.784

0.915

1.003

0.773

574.450

0.838

0.317

0.385

0.426

0.454

0.465

0.672

0.673

0.918

0.660

574.451

0.708

2.343

1.898

1.860

1.724

2.289

1.312

1.469

1.181

3.145

-11.777

0.860

0.055

0.714

0.836

1.660

0.216

1.312

1.391

0.783

-0.959

13.072

***

0.745

0.777

0.793

0.819

0.784

0.915

0.916

1.003

0.773

598.772

0.838

0.317

0.386

0.426

0.454

0.466

0.672

0.673

0.918

0.661

598.773

***

9,702

***

9,702

***

9,702

yes

9,648

yes

9,648

yes

9,648

Notes: ***, **, *: statistically significant at the 1, 5, and 10 percent significance level. Conditioning variables (all at firm age 0 years): Number of employees, value added, annual earnings, turnover and equity.

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

TABLE A.3.1: Employment in the combined Statistics Denmark-Experian database. IM-firms and reference firms. By entry cohort and year

Entry

cohort:

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Total employ-

ment minus

accumulated

first-year

employment

First-year

employment

Accumulated

first-year

employ-ment

Total employ-

ment

Reference firms:

Year: 1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Total

IM-firms:

Year: 1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Total

111

214

219

162

110

122

112

122

115

109

1,654

197

313

230

196

179

163

183

162

123

119

122

118

2,105

110

160

149

185

129

144

151

113

104

1,408

121

109

126

118

151

172

143

130

100

1,263

272

354

289

326

239

211

177

149

2,115

105

123

106

107

123

134

839

122

183

216

233

246

271

1,342

107

149

116

120

117

109

718

135

115

121

545

284

112

152

339

117

210

128

111

197

110

107

128

1,654

111

308

418

511

609

666

737

844

921

962

1,037

1,130

1,258

111

411

642

645

674

924

1,039

1,190

1,437

1,332

1,397

1,504

1,644

12,950

103

224

134

258

302

346

516

370

360

374

386

3,438

104

161

170

122

115

138

127

107

112

105

100

1,553

237

394

370

308

277

257

232

257

225

204

213

211

3,185

197

256

307

370

456

491

529

484

363

306

348

4,107

110

143

148

182

203

217

212

189

178

1,661

119

109

126

138

114

104

973

111

122

123

691

105

127

112

107

103

707

108

133

141

209

233

879

302

128

255

161

104

237

197

119

1,231

104

341

538

617

736

800

867

922

990

1,027

1,097

1,171

1,231

104

398

761

827

959

1,111

1,275

1,420

1,665

1,567

1,442

1,458

1,675

14,662

223

210

223

311

408

498

675

540

345

287

444

4,221

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

TABLE A.3.2: Employment in the combined Statistics Denmark-Experian database. IM-firms and reference firms. By entry cohort and firm age

Entry

cohort:

1999

Reference

firms:

Accumulated

first-year

employment:

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

104

341

538

617

736

800

867

922

990

1027

1097

1171

1231

Aggregate

employ-

ment, up to

2010

Aggregate

employ-

ment, up to

2011

Firm age:

Employment

growth, by firm

age

104

237

197

119

1,171

1,231

1.395388557

Total

IM-firms:

Accumulated

first-year

employment:

Firm age

161

170

122

115

138

127

107

112

105

100

1,553

394

370

308

277

257

232

257

225

204

213

211

3,185

256

307

370

456

491

529

484

363

306

348

110

143

148

182

203

217

212

189

178

119

109

126

138

114

104

111

122

123

105

127

112

107

103

108

133

141

209

233

1,547

1,566

1,495

1,555

1,357

1,257

1,126

889

615

309

100

1,634

1,661

1,530

1,622

1,590

1,360

1,210

967

793

657

311

14,662

1.073691015

0.977011494

1.084949833

1.022508039

1.002210759

0.962609387

0.858792185

0.892013498

1.068292683

1.006472492

0.96

4,107

1,661

973

691

707

879

302

128

255

161

12,987

111

308

418

511

609

666

737

844

921

962

1037

1130

1258

Aggregate

employ-

ment, up to

2010

Aggregate

employ-

ment, up to

2011

Employment

growth, by firm

age

111

214

219

162

110

122

112

122

197

313

230

196

179

163

183

162

110

160

149

185

129

144

151

113

121

109

126

118

151

172

143

272

354

289

326

239

211

177

105

123

106

107

123

134

122

183

216

233

246

271

107

149

116

120

117

109

135

115

121

112

152

117

128

1,130

1,701

1,689

1,532

1,318

1,172

952

717

1,258

1,818

1,841

1,629

1,439

1,281

1,223

851

1.608849558

1.082304527

0.964476021

0.939295039

0.971927162

1.043515358

0.893907563

0.866108787

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Total

115

109

1,654

123

119

122

118

2,105

104

130

100

149

472

318

215

621

418

288

208

12,950

0.88559322

0.905660377

0.96744186

0.833333333

1,408

1,263

2,115

839

1,342

718

545

284

339

210

128

11,306

TABLE A.4.1: Turnover. IM-firms and reference firms. By entry cohort and year

Entry

cohort:

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Total

turnover

minus

accumulated

first-year

turnover

First-year

turnover

Accumulated

first-year

turnover

Total

turnover

Reference

firms:

Year:

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Total

IM-firms:

Year:

1999

2000

2001

85,764.3

147,446

214,466

155,921

220,239

84,706.4

85,764.3

155,921

84,706.4

85,764.3

241,685.3

326,391.7

85,764.3

303,366.8

519,410.5

61,681.53

193,018.8

97,873.7

143,117

158,406

133,320

122,406

136,477

143,609

125,510

133,731

161,883

120,039

120,516

140,795

1,737,681

339,086

629,950

472,046

417,114

512,719

300,783

272,703

242,346

283,037

257,448

264,195

333,508

4,324,935

139,279

278,548

331,092

698,359

1,387,364

2,062,969

2,877,128

548,453

819,368

339,197

460,998

9,942,755

56,420.1

120,941

199,041

229,672

115,453

216,767

325,708

117,059

168,914

223,520

1,773,496

68,523.5

133,008

139,706

126,359

153,872

195,780

143,495

97,151,3

197,159

1,255,054

49,080.7

2,839,337

4,572,259

125,025

135,002

88,955

102,981

131,416

8,044,056

33,685.4

72,387.8

102,969

120,075

106,600

121,178

145,390

702,286

39,480.5

81,528.2

83,686.7

87,044.5

110,523

211,433

613,695

58,186.4

104,778

85,662

84,666.4

789,07.7

412,200

26,285.7

23,707

33,731.4

42,505.8

126,230

33,468

63,717.7

96,323.1

193,509

61,688.2

191,900

253,588

33,515.4

97,873.67

339,086.3

139,279

56,420.09

68,523.48

49,080.66

33,685.4

39,480.45

58,186.39

26,285.71

33,468

61,688.23

33,515.39

1,036,573

97,873.67

436,960

576,239

632,659.1

701,182.5

750,263.2

783,948.6

823,429.1

881,615.4

907,901.2

941,369.2

1,003,057

1,036,573

97,873.67

482,202.8

927,635.2

940,333.7

1,060,076

1,728,685

5,074,156

7,387,121

3,991,553

1,984,688

1,882,846

1,568,459

2,287,371

29,413,001

45,242.83

351,396.2

307,674.6

358,893.9

978,422.1

4,290,208

6,563,692

3,109,938

1,076,787

941,477

565,401.3

1,250,798

19,839,930

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Total

140,526

110,335

112,673

102,197

140,565

107,603

95,916.9

101,293

133,359

128,497

1620,640

228,835

248,046

160,207

154,309

133,083

166,238

222,985

217,125

177,514

186,489

2,270,989

118,728

99,375.8

171,378

136,135

107,135

145,932

118,614

132,919

129,657

97,158.7

1,341,739

71,829.9

98,221.5

82,277.4

103,112

131,441

196,720

197,333

159,111

143,825

149,401

1,333,272

70,845.6

188,070

234,821

277,075

396,255

367,885

295,766

369,461

244,239

2,444,417

44,186.8

71,079.4

116,598

91,881.1

100,181

129,370

89,225.5

98,214.4

740,737

67,512.3

104,473

207,485

298,631

402,273

481,739

513,183

2,075,295

49,935

129,050

113,395

117,091

113,218

171,849

694,538

63,832.5

111,215

177,472

178,834

253,417

784,770

46,821.7

46,319

79,531.4

118,590

291,262

48,126

106,469

182,785

337,379

49,672.6

151,708

201,380

109,481

71,829.9

70,845.6

44,186.8

67,512.3

49,935

63,832.5

46,821.7

48,126

49,672.6

109,480.8

162,0640

398,221.6

469,067.3

513,254

580,766.3

630,701.3

694,533.7

741,355.4

789,481.4

839,154

948,634.8

559,919

626,823.4

758,791.1

869,164.8

1,060,305

1,504,997

1,672,978

1,826,866

2,052,503

2,405,011

14,245,899

161,697.4

157,756.1

245,537.1

288,398.5

429,603.8

810,463.3

931,622.2

1,037,385

1,213,349

1,456,376

6,986,888

TABLE A.4.2: Turnover. IM-firms and reference firms. By entry cohort and firm age

Entry

cohort:

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Reference

firms:

Aggregate first-

year turnover:

97873,7

436960

576239

632659

701183

750263

783949

823429

881615

907901

941369

1003057

1036573

Total

turnover,

up to 2010

97,873.7

143,117

158,406

133,320

122,406

136,477

143,609

125,510

133,731

161,883

120,039

120,516

140,795

339,086

629,950

472,046

417,114

512,719

300,783

272,703

242,346

283,037

257,448

264,195

333,508

139,279

278,548

331,092

698,359

1,387,364

2,062,969

2,877,128

548,453

819,368

339,197

460,998

56,420.1

120,941

199,041

229,672

115,453

216,767

325,708

117,059

168,914

223,520

68,523.5

133,008

139,706

126,359

153,872

195,780

143,495

9715,1,3

197,159

49,080.7

2,839,337

4,572,259

125,025

135,002

88,955

102,981

131,416

33,685.4

72,387.8

102,969

120,075

106,600

121,178

145,390

39,480.5

81,528.2

83,686.7

87,044.5

110,523

211,433

58,186.4

104,778

85,662

84,666.4

78,907.7

26,285.7

23,707

33,731.4

42,505.8

33,468

63,717.7

96,323.1

61,688.2

191,900

33,515.39

1,003,057

4,491,019

6,178,599

2,021,635

2,643,939

3,122,910

3,865,625

1,130,519

1,405,050

758,527.5

384,234.1

120,515.7

Total

turnover,

up to 2011

1,036,573

4,682,919

6,274,922

2,064,141

2,722,847

3,334,342

4,011,015

1,261,935

1,602,209

982,047.7

845,232.2

454,023.4

140,795.2

Turnover

growth, by

firm age

4.668645

1.397216

0.334079

1.346854

1.261127

1.284384

0.326451

1.417233

0.698942

1.114307

1.181632

1.168273

Firm age:

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Total

IM-firms:

Aggregate first-

year turnover:

Firm age

Total

1,737,681

4,324,935

9,942,755

1,773,496

1,255,054

8,044,056

702,286

613,695

412,200

126,230

193,509

253,588

33,515.39

27,125,630

29,413,001

85,764.3

241,685

326,392

398,222

469,067

513,254

580,766

630,701

694,534

741,355

789,481

839,154

948,635

Total

turnover,

up to 2010

Total

turnover,

up to 2011

Turnover

growth, by

firm age

85,764.3

147,446

214,466

140,526

110,335

112,673

102,197

140,565

107,603

95,916.9

101,293

133,359

128,497

1,620,640

155,921

220,239

228,835

248,046

160,207

154,309

133,083

166,238

222,985

217,125

177,514

186,489

84,706.4

118,728

99,375.8

171,378

136,135

107,135

145,932

118,614

132,919

129,657

97,158.7

71,829.9

98,221.5

82,277.4

103,112

131,441

196,720

197,333

159,111

143,825

149,401

70,845.6

188,070

234,821

277,075

396,255

367,885

295,766

369,461

244,239

44,186.8

71,079.4

116,598

91,881.1

100,181

129,370

89,225.5

98,214.4

67,512.3

104,473

207,485

298,631

402,273

481,739

513,183

49,935

129,050

113,395

117,091

113,218

171,849

63,832.5

111,215

177,472

178,834

253,417

46,821.7

46,319

79,531.4

118,590

48,126

106,469

182,785

49,672.6

151,708

109,481

839,154

1,341,308

1,554,255

1,626,573

1,550,045

1,549,832

963,536.8

953,988.7

607,331.4

442,698.8

278,806.7

133,358.8

948,634.8

1,493,016

1,737,040

1,745,163

1,803,462

1,721,681

1,476,719

1,052,203

851,569.9

592,099.8

375,965.4

319,847.3

128,497.1

1.779192

1.295034

1.12283

1.10875

1.11073

0.952825

1.092022

0.892641

0.97492

0.849258

1.147201

0.963544

2,270,989

1,341,739

1,333,272

2,444,417

740,737

2,075,295

694,538

784,770

291,262

337,379

201,380

109,481

11,840,889

14,245,899

TABLE A.4.3: Turnover. IM-firms and reference firms, except for largest 5 in terms of turnover in each group of firms in each year. By entry cohort and year

Entry

cohort:

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Total

turnover

minus

accumulated

first-year

turnover

First-year

turnover

Accumulated

first-year

turnover

Total

turnover

Reference

firms:

Year:

1999

2000

2001

2002

2003

65,778.6

143,117

158,406

133,320

122,406

155,257

360,028

306,029

276,868

139,279

241,458

234,109

56,420.1

120,941

68,523.5

65,778.61

155,257.3

139,279

56,420.09

68,523.48

65,778.61

221,035,91

360,314.91

416,735

485,258.48

65,778.61

298,373.8

657,712.4

737,226.29

822,847.18

77,337.89

297,397.49

320,491.29

337,588.7

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

2004

2005

2006

2007

2008

2009

2010

2011

Total

IM-firms:

Year:

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Total

136,477

143,609

125,510

133,731

161,883

120,039

120,516

140,795

1,705,586

267,230

300,783

219,965

188,257

230,990

168,820

158,913

333,508

2,966,647

253,655

275,203

272,845

280,807

280,624

194,294

173,379

208,447

2,554,101

199,041

115,331

115,453

149,013

153,345

74,356.4

114,811

223,520

1,322,233

133,008

139,706

126,359

153,872

195,780

143,495

97,151.3

92,424

1,150,319

49,081

122,665

160,280

125,025

135,002

88,955

102,981

131,416

915,405

33,685

72,388

102,969

76,498

73,878

70,439

74,274

504,131

39,480

81,528

83,687

87,045

110,523

85,563

487,826

58,186

104,778

85,662

84,666

78,908

412,200

26,286

23,707

33,731

42,506

126,230

33,468

63,718

96,323

193,509

61,688

93,146

154,834

33,515

49,080.66

33,685.4

39,480.45

58,186.39

26,285.71

33,468

61,688.23

33,515.39

820,648.71

534,339.14

568,024.54

607,504.99

665,691.38

691,977.09

725,445.09

787,133.32

820,648.71

1,038,492

1,130,981.7

1,132,280.2

1,273,389.4

1,448,872.1

1,093,719.1

1,192,516.5

1,634,346.1

12,526,535

504,152.82

562,957.16

524,775.21

607,697.97

756,894.98

368,274.03

405,383.14

813,697.43

5,576,648.1

70,121.1

84,078.7

114,285

114,530

72,908.3

92,217.7

58,846.9

110,275

107,603

95,916.9

101,293

133,359

128,497

1,283,932

130,179

192,853

135,813

109,575

118,565

125,553

107,952

166,238

176,020

160,276

177,514

186,489

1,787,025

84,706.4

118,728

99,375.8

171,378

105,124

107,135

145,932

118,614

132,919

129,657

97,158.7

1,310,728

53,024

98,221.5

82,277.4

103,112

80,939.2

109,500

139,961

159,111

91,916.7

90,115

1,008,174

53,265.1

132,030

159,633

131,076

150,037

172,073

139,204

101,263

133,138

1,171,718

44,187

71,079

68,943

91,881

100,181

129,370

89,225

98,214

693,082

67,512

104,473

112,193

110,854

126,199

164,283

179,642

865,156

49,935

129,050

113,395

117,091

113,218

100,976

623,665

63,832

111,215

80,412

98,759

144,460

498,677

46,822

46,319

79,531

118,590

291,262

48,126

106,469

182,785

337,379

49,673

151,708

201,380

109,481

70,121.05

130,179.4

84,706.43

53,023.97

53,265.1

44,186.79

67,512.27

49,934.97

63,832.45

46,821.68

48,126

49,672.55

109,480.8

1,283,931.8

70,121.05

200,300.45

285,006.88

338,030.85

391,295.95

435,482.74

502,995.01

552,929.98

616,762.43

663,584.11

711,710.11

761,382.66

870,863.46

70,121.05

214,258.13

391,844.43

422,094.77

433,345.59

640,654.12

690,859.91

760,728.26

1,076,265.2

1,185.051.7

1,240,321.2

1,334,866

1,721,250.1

10,181,660

13,957.68

106,837.55

84,063.92

42,049.64

205,171.38

187,864.9

207,798.28

459,502.78

521,467.59

528,611.09

573,483.31

850,386.62

3,781,194.7

TABLE A.4.4: Turnover. IM-firms and reference firms, except for largest 5 in terms of turnover in each group of firms in each year. By entry cohort and firm age

Entry

cohort:

1999

Reference

firms:

Aggregate

first-year

turnover:

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

65,778.6

221,036

360,315

416,735

485,258

534,339

568,025

607,505

665,691

691,977

725,445

787,133

820,649

Firm age:

Total

turnover, up

to 2010

Total

turnover, up

to 2011

Turnover

growth, by

firm age

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Total

IM-firms:

Aggregate

first-year

turnover:

Firm age

Total

65,778.6

143,117

158,406

133,320

122,406

136,477

143,609

125,510

133,731

161,883

120,039

120,516

140,795

1,705,586

155,257

360,028

306,029

276,868

267,230

300,783

219,965

188,257

230,990

168,820

158,913

333,508

139,279

241,458

234,109

253,655

275,203

272,845

280,807

280,624

194,294

173,379

208,447

56,420.1

120,941

199,041

115,331

115,453

149,013

153,345

74,356.4

114,811

223,520

68,523.5

133,008

139,706

126,359

153,872

195,780

143,495

97,151.3

92,424

49,081

122,665

160,280

125,025

135,002

88,955

102,981

131,416

33,685

72,388

102,969

76,498

73,878

70,439

74,274

39,480

81,528

83,687

87,045

110,523

85,563

58,186

104,778

85,662

84,666

78,908

26,286

23,707

33,731

42,506

33,468

63,718

96,323

61,688

93,146

33,515

787,133.32

1,467,333.7

1,503,619.8

1,278,766.4

1,253,566.6

1,214,292.4

1,044,203.5

765,898.05

673,826.1

504,081.8

278,951.8

120,515.7

820,648.71

1,560,479.8

1,599,942.9

1,321,272.2

1,332,474.3

1,299,855.8

1,118,477.5

897,314.35

766,250.14

727,602

4873,98.9

454,023.4

140,795.2

1.9824848

1.0903743

0.8787276

1.0419997

1.036926

0.9210941

0.859329

1.0004597

1.0798068

0.9669044

1.6276052

1.1682727

2,966,647

2,554,101

1,322,233

1,150,319

915,405

504,131

487,826

412,200

126,230

193,509

154,834

33,515

10,892,189

12,526,535

70,121.1

200,300

285,007

338,031

391,296

435,483

502,995

552,930

616,762

663,584

711,710

761,383

870,863

Aggregate

turnover, up

to 2010

Aggregate

turnover, up

to 2011

Turnover

growth, by

firm age

70,121.1

84,078.7

114,285

114,530

72,908.3

92,217.7

58,846.9

110,275

107,603

95,916.9

101,293

133,359

128,497

1,283,932

130,179

192,853

135,813

109,575

118,565

125,553

107,952

166,238

176,020

160,276

177,514

186,489

84,706.4

118,728

99,375.8

171,378

105,124

107,135

145,932

118,614

132,919

129,657

97,158.7

53,024

98,221.5

82,277.4

103,112

80,939.2

109,500

139,961

159,111

91,916.7

90,112.5

53,265.1

132,030

159,633

131,076

150,037

172,073

139,204

101,263

133,138

44,187

71,079

68,943

91,881

100,181

129,370

89,225

98,214

67,512

104,473

112,193

110,854

126,199

164,283

179,642

49,935

129,050

113,395

117,091

113,218

100,976

63,832

111,215

80,412

98,759

144,460

46,822

46,319

79,531

118,590

48,126

106,469

182,785

49,673

151,708

109,481

761,382.66

1,194,515.4

1,045,858.6

1,048,255.4

867,170.9

900,131.08

681,121.17

655,500.7

508,459.06

385,849.82

278,806.7

133,358.8

870,863.46

1,346,223.1

1,228,643.2

1,166,845.8

1,011,630.5

1,001,107.1

860,763.47

753,715.12

641,596.56

475,962.32

375,965.36

319,847.3

128,497.1

1.7681295

1.0285704

1.1156821

0.9650611

1.1544519

0.9562646

1.1065801

0.9787885

0.9360878

0.9743826

1.1472009

0.9635442

1,787,025

1,310,728

1,008,174

1,171,718

693,082

865,156

623,665

498,677

291,262

337,379

201,380

109,481

8,460,410.3

10,181,660

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

TABLE A.5.1: Value added. IM-firms and reference firms. By entry cohort and year

Entry

cohort:

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Total value

added minus

aggregate

first-year

value added

First-year

value added

Aggregate

first-year

value added

Total value

added

Reference

firms:

Year:

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

Total

IM-firms:

Year:

1999

2000

2001

2002

2003

2004

2005

2006

886.4199

3,648325

11,805.85

3,829.909

27,816.55

20,476.24

33,356.14

47,071.43

-9,038.28

-12,581.3

1,111.872

17,416.11

42,498.34

38,747.02

28,380.6

-2,367.25

-2,227.17

13,829.98

18,202.21

16,966.34

-56,054.4

-425.799

-8,927.29

2,532.597

8,574.376

17,188.7

-2,407.16

-4,222.1

6,900.109

13,081.02

-4,847.51

-14,187.8

-21,342.2

-5,906.62

-7,018.12

-5,427.51

886.4199

-9,038.277

-2,367.251

-425.7991

-2,407.159

-4,847.514

-5,906.623

-5,427.505

886.4199

-8,151.8571

-10,519.108

-10,944.907

-13,352.066

-18,199.58

-24,106.203

-29,533.708

886.4199

-5,389.952

-3,142.691

2,288.8129

47,728.188

7,4639.774

84,449.522

15,879.531

2,761.9051

7,376.4171

13,233.72

61,080.254

92,839.354

108,555.73

45,413.239

-1,443.21

18,979.66

19,842.11

22,531.96

26,540.27

34,988.95

43,052.12

44,618.34

51,867.02

55,471.13

47,014

40,939.21

30,247.12

47,677.36

482,326

-18,669.9

94,824,56

127,313.9

101,583.9

101,688.4

33,474.48

47,780.38

45,749.74

39,449.85

48,013

54,066.67

57,364.42

57,436.79

790,076.2

-2,375.44

44,399.54

58,083.89

29,720.44

72,459.28

178,993.6

163,022

58,459.8

26,950

21,127.45

91,040.38

124,194.3

866,075.4

-231.74

16,927.29

26,674.03

26,168.29

28,794.24

33,676.44

51,710.23

46,094

61,958.82

57,815.39

32,240.57

381,827.6

-17,529.1

32,976.8

35,623.23

42,043.71

54,428.27

66,548.13

68,525

28,466.67

28,500.96

25,457.55

365,041.2

-2,826.52

9,642.779

16,824.09

24,007.33

30,581.56

22,031

19,452.94

35,760.58

46,517.93

201,991.7

-2,232.36

11,159.91

28,738.22

38,413.37

28,880

34,940.2

43,112.5

40,026.41

223,038.3

-9,805.97

6,857.59

12,494.3

27,923

38,961.77

34,044.23

45,636.79

156,111.8

3,103.67

18,500.51

17,611

14,006.86

25,168.27

23,595.28

101,985.6

541.03

2,984

3,477.45

21,818.27

10,492.45

39,313.2

-7,487

997.06

-1,940.39

-5,109.43

-13,539.8

-1,639.22

7,570.19

16,028.3

21,959.28

-1,387.5

32,602.83

31,215.33

-1,443.21

-18,669.86

-2,375.44

-231.74

-17,529.08

-2,826.52

-2,232.36

-9,805.97

3,103.67

54.,03

-7,487

-1,639.216

-1,387.5

-61,983.19

-1,443.21

-20,113.07

-22,488.508

-22,720.243

-40,249.323

-43,075.843

-45,308.199

-55,114.169

-52,010.504

-51,469.471

-58,956.471

-60,595.687

-61,983.187

-1,443.21

309.8

112,291.23

194,013.66

185,606.27

223,222.1

218,187.82

360,408.3

411,450.28

372,170.22

328,538

316,755.88

429,114.43

496,797.13

3,647,421.9

20,422.87

134,779.74

216,733.91

225,855.59

266,297.94

263,496.02

415,522.47

463,460.78

423,639.69

387,494.47

377,351.57

491,097.61

558,780.31

4,244,933

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

2007

2008

2009

2010

2011

2012

Total

24,072.25

13,879.43

41,222

62,934.31

67,611.54

48,356.61

455,323.6

45,877.49

19,446.81

41,997

93,307.84

79,772.12

165,741.5

718,418.6

11,516.23

14,472.14

119,777

33,459.8

25,275

23,709.43

240,268.8

22,324.61

12,433.64

10,922

27,026.47

19,350.96

43,171.7

197,343.7

132,572.8

61,588.65

68,931

90,964.7

86,796.16

108,428.3

671,061.8

-27,738.2

-16,958.5

-6,231

-7,891.18

-12,442.3

9,366.038

-92,906.7

21,306.8

44,219.86

108,682

135,197.1

173,488.5

160,486.8

790,943.1

-18,227.2

1,276.596

-1,410

16,596.08

29,588.46

11,413.21

45,222.83

-8579,06

-34866,3

-33481

-36557,8

13814,42

34124,53

-31420,7

-4,224.92

-9,308

-1,297.06

31,608.65

51,354.71

119,488.1

-6,482

-8,147.06

-26,803.9

-23,194.3

-87,821.6

-8,768.63

-33,276

-17,293.4

-76,631.4

-14,084.6

-21,324.5

-56,733.7

-8,579.058

-4,224.924

-6,482

-8,768.627

-14,084.62

-71,672.937

-38,112.766

-42,337.69

-48,819.69

-57,588.317

-71,672.937

203,125.68

111,267.48

334,619

396,824.54

440,699.07

594,340.56

2,298,215.9

241,238.45

153,605.17

383,438.69

454,412.86

512,372.01

666,013.5

2,742,341.3

TABLE A.5.2: Value added. IM-firms and reference firms. By entry cohort and firm age

Entry

cohort:

1999

Reference

firms:

Aggregate

first-year value

added:

Firm age:

-1,443.21

-20,113.1

-22,488.5

-22,720.2

-40,249.3

-43,075.8

-45,308.2

-55,114.2

-52,010.5

-51,469.5

-58,956.5

-60,595.7

-61,983.2

Total value

added, up to

2011

-1,443.21

18,979.66

19,842.11

22,531.96

26,540.27

34,988.95

43,052.12

44,618.34

51,867.02

55,471.13

47,014

40,939.21

30,247.12

47,677.36

434,648.7

790,076.2

866,075.4

381,827.6

365,041.2

201,991.7

223,038.3

156,111.8

101,985.6

39,313.2

-13,539.8

21,959.28

31,215.33

-18,669.9

94,824.56

127,313.9

101,583.9

101,688.4

33,474.48

47,780.38

45,749.74

39,449.85

48,013

54,066.67

57,364.42

57,436.79

-2,375.44

44,399.54

58,083.89

29,720.44

72,459.28

178,993.6

163,022

58,459.98

26.950

21,127.45

91,040.38

124,194.3

-231.735

16,927.29

26,674.03

26,168.29

28,794.24

33,676.44

51,710.23

46,094

61,958.82

57,815.39

32,240.57

-17,529.1

32,976.8

35,623.23

42,043.71

54,428.27

66,548.13

68,525

28,466.67

28,500.96

25,457.55

-2,826.52

9,642.779

16,824.09

24,007.33

30,581.56

22,031

19,452.94

35,760.58

46,517.93

-2,232.36

11,159.91

28,738.22

38,413.37

28,880

34,940.2

43,112.5

40,026.41

-9,805.97

6,857.592

12,494.43

27,923

38,961.77

34,044.23

45,636.79

3,103.665

18,500.51

17,611

14,006.86

25,168.27

23,595.28

541.0334

2,984

3,477.451

21,818.27

10,492.45

-7,487

997.059

-1,940.39

-5,109.43

-1,639.22

7,570.192

16,028.3

-1,387.5

32,602.83

-61,983.187

265,819.89

344,741.97

348,217.13

407,502.06

438,697.03

436,655.17

259,149.31

208,726.65

182,426.97

192,121.05

98,303.63

30,247.12

3,150,624.8

Total value

added, up to

2012

-61,983.187

298,422.72

360,770.27

343,107.7

417,994.51

462,292.31

482,291.96

299,175.72

255,244.58

207,884.52

224,361.62

222,497.93

87,683.91

47,677.36

3,599,744.6

Value added

increase, by

firm age

360,405.91

94,950.374

-1,634.27

69,777.38

54,790.25

43,594.93

-137,479.45

-3,904.73

-842.13

41,934.65

30,376.88

-10,619.72

17,430.24

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Total

IM-firms:

Aggregate

first-year value

added:

Firm age

886.4199

-8,151.86

-10,519.1

-10,944.9

-13,352.1

-18,199.6

-24,106.2

-29,533.7

-38,112.8

-42,337.7

-48,819.7

-57,588.3

-71,672.9

Total value

added, up to

2011

Total value

added, up to

2012

Value added

increase, by

firm age

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Total

886.4199

3,648.325

11,805.85

3,829.909

27,816.55

20,476.24

33,356.14

47,071.43

24,072.25

13,879.43

41,222

62,934.31

67,611,54

48,356,61

406,967

-9,038.28

-12,581.3

1,111.872

17,416.11

42,498.34

38,747.02

28,380.6

45,877.49

19,446.81

41,997

93,307.84

79,772.12

165,741.5

-2,367.25

-2,227.17

13,829.98

18,202.21

16,966.34

-56,054.4

1,1516.23

14,472.14

119,777

33,459.8

25,275

23,709.43

-425.799

-8,927.29

2,532.597

8,574.376

17,188.7

22,324.61

12,433.64

10,922

27,026.47

19,350.96

43,171.7

-2,407.16

-4,222.1

6,900.109

13,081.02

132,572.8

61,588.65

68,931

90,964.7

86,796.16

108,428.3

-4,847.51

-14,187.8

-21,342.2

-27,738.2

-16,958,5

-6,231

-7,891.18

-12,442.3

9,366.038

-5,906.62

-7,018.12

21,306.8

44,219.86

108,682

135,197.1

173,488.5

160,486.8

-5,427.51

-18,227.2

1,276.596

-1,410

16,596.08

29,588.46

11,413.21

-8,579.06

-34,866.3

-33,481

-36,557.8

13,814.42

34,124.53

-4,224.92

-9,308

-1,297.06

31,608.65

51,354.71

-6,482

-8,147.06

-26,803.9

-23,194.3

-8,768.63

-33,276

-17,293.4

-14,084.6

-21,324.5

-71,672.937

-149,339.99

-24,160.325

71,226.075

359,176.77

245,636.71

320,214.93

196,865.45

277,118.69

108,687.19

159,804.84

142,706.43

67,611.54

-71,672.937

-170,664.52

-41,453.725

48,031.735

410,531.48

279,761.24

331,628.14

357,352.25

286,484.73

217,115.49

202,976.54

166,415.86

233,353.04

48,356.61

-98,991.582

107,886.26

72,192.06

339,305.41

-79,415.53

85,991.433

37,137.317

89,619.278

-60,003.2

94,289.35

6,611.02

90,646.61

-19,254.93

552,677,1

216,559.3

154,172

562,633.5

-102,273

630,456.3

33,809.62

-65,545.2

68,133.38

-64,627.2

-59,338

-35409,2

1,703,875.4

2,298,215.9

TABLE A.6.1: Annual earnings (DKK1,000). IM-firms and reference firms. By entry cohort and year

Entry cohort:

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

Accumulated

first-year

annual

earnings

Total annual

earnings minus

accumulated

first-year

annual

earnings

First-year

annual

earnings

Total annual

earnings

Reference

firms:

Year:

1999

2000

2001

2002

2003

2004

2005

2006

-12,142

-28,327

-11,916

-16,733

-4,540

-4,124

-58,911

-17,665

-71,673

-11,358

-175,816

-126,312

-41,786

-66,291

-43,470

-30,356

-175,975

-33,482

-77,572

-83,641

5,516

-56,595

-25,798

-56,397

-117,199

-175,985

-58,473

-12,012

-14,067

-972

-24,242

-27,934

-28,922

-7,485

-22,882

-14,077

-12,142

-71,673

-30,356

-56,595

-58,473

-24,242

-7,485

-14,077

-12,142

-83,815

-114,171

-170,766

-229,239

-253,481

-260,966

-275,043

-12,142

-100,000

-53,629

-425,119

-248,605

-216,133

-375,528

-298,457

-16,185

60,542

-254,353

-19,366

37,348

-114,561

-23,414

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

2007

2008

2009

2010

2011

2012

Total

-57,038

-54,185

-64,588

-61,927

-11,300

47,677

-408,712

-15,364

-25,854

-56,784

46,906

-34,069

57,437

-662,484

-39,372

38,536

-159,352

-15,016

-28,916

124,194

-589,044

-166,584

-155,589

-276,576

-263,983

-256,619

32,241

-1,826,621

-566

3,470

-2,535

8,588

706

25,458

-79,844

-25,753

-27,520

-52,722

-30,081

-11,650

46,518

-240,568

-11,215

-6,575

-4,857

951

7,570

40,026

-56,378

-6,831

-3,446

1,748

5,800

-11,562

45,637

-29,317

-8,369

-3,316

-15,494

-1,424

19,520

23,595

-7,301

-12,507

-2,822

-4,903

9,964

10,492

-10,387

-21,228

-34,148

-45,504

-5,109

-139,331

-8,451

-3,896

16,028

-19,166

-10,553

32,603

-10,277

-8,369

-12,507

-21,228

-8,451

-10,553

-336,150

-283,412

-295,918

-317,146

-325,597

-336,150

-331,092

-246,986

-655,210

-357,688

-376,309

-382,533

-4,079,431

-47,681

48,932

-338,064

-32,091

-40,159

-46,383

-785,435

IM-firms:

Year:

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

Total

-27,141

-45,413

-58,626

-52,631

-33,714

-32,172

-55,125

-55,552

-35,638

-73,035

-36,745

-8,585

13,548

48,357

-492,160

-61,078

-106,027

-92,304

-77,262

-77,608

-76,337

199,291

-80,676

-85,113

-54,273

15,754

-1,004

165,742

-407,222

-15,202

-56,793

-49,537

-52,683

-37,028

-125,037

-122,835

-84,383

65,044

-75,845

-34,413

23,709

-608,475

-20,990

-49,015

-38,772

-49,200

-78,382

-105,426

-143,416

-125,104

-83,511

-62,840

43,172

-269,857

-23,280

-95,232

-133,448

-204,182

-79,429

-137,443

-103,396

-37,613

-10,071

108,428

-824,861

-13,449

-48,571

-75,865

-69,476

-62,949

-68,229

-78,476

-95,349

9,366

-558,234

-15,534

-61,629

-64,214

-86,787

-57,327

-57,520

-33,990

160,487

-376,550

-61,175

-87,508

-46,282

-60,447

-42,930

-32,716

11,413

-380,986

-17,137

-82,015

-101,933

-104,664

-41,955

34,125

-375,761

-9,077

-26,873

-25,876

-22,804

51,355

-99,890

-32,342

-55,744

-91,041

-23,194

-282,170

-26,085

-75,462

-17,293

-174,966

-27,240

-21,325

-96,884

-27,141

-61,078

-15,202

-20,990

-23,280

-13,449

-15,534

-61,175

-17,137

-9,077

-32,342

-26,085

-27,240

-349,730

-27,141

-88,218

-103,421

-124,411

-147,690

-161,139

-176,673

-237,848

-254,985

-264,062

-296,404

-322,489

-349,730

-27,141

-106,490

-179,855

-222,718

-232,806

-309,916

-415,243

-462,531

-662,339

-810,502

-601,625

-581,096

-515,338

179,586

-4,948,015

-18,272

-76,434

-98,307

-85,116

-148,777

-238,570

-224,683

-407,354

-546,440

-305,221

-258,607

-165,609

529,315

-2,044,075

TABLE A.6.2: Annual earnings. IM-firms and reference firms. By entry cohort and firm age

Entry cohort:

1999

Reference

firms:

Aggregate first-

year annual

earnings:

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

-12,142

-83,815

-114,171

-170,766

-229,239

-253,481

-260,966

-275,043

-283,412

-295,918

-317,146

-325,597

-336,150

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Firm age:

Total annual

earnings, up to

2011

-12,142

-28,327

-11,916

-16,733

-4,540

-4,124

-58,911

-17,665

-57,038

-54,185

-64,588

-61,927

-11,300

-5,316

-403,396

-662,484

-589,044

-1,826,621

-79,844

-240,568

-56,378

-29,317

-7,301

-10,387

-139,331

-19,166

-10,277

-71,673

-11,358

-175,816

-126,312

-41,786

-66,291

-43,470

-15,364

-25,854

-56,784

46,906

-34,069

-40,611

-30,356

-175,975

-33,482

-77,572

-83,641

5,516

-39,372

38,536

-159,352

-15,016

-28,916

10,585

-56,595

-25,798

-56,397

-117,199

-175,985

-166,584

-155,589

-276,576

-263,983

-256,619

-275,297

-58,473

-12,012

-14,067

-972

-566

3,470

-2,535

8,588

706

-3,982

-24,242

-27,934

-28,922

-25,753

-27,520

-52,722

-30,081

-11,650

-11,744

-7,485

-22,882

-11,215

-6,575

-4,857

951

7,570

-11,885

-14,077

-6,831

-3,446

1,748

5,800

-11,562

-949

-8,369

-3,316

-15,494

-1,424

19,520

1,781

-12,507

-2,822

-4,903

9,964

-120

-21,228

-34,148

-45,504

-38,451

-8,451

-3,896

-6,819

-10,553

275

-336,150

-355,298

-401,162

-360,828

-313,575

-291,345

-322,388

-274,131

-505,521

-382,604

-46,598

-95,997

-11,300

-3,696,898

Total annual

earnings, up to

2012

-336,150

-355,023

-407,980

-399,279

-313,694

-289,564

-323,337

-286,016

-517,266

-386,586

-321,896

-85,412

-51,911

-5,316

-4,079,431

Annual

earnings

increase, by

firm age

-18,873

-52,682

1,883

47,133

24,010

-31,991

36,371

-243,134

118,935

60,709

-38,813

44,085

5,984

Total

IM-firms:

Aggregate first-

year annual

earnings:

Firm age

-27,141

-45,413

-58,626

-52,631

-33,714

-32,172

-55,125

-55,552

-35,638

-73,035

-36,745

-8,585

13,548

-61,078

-106,027

-92,304

-77,262

-77,608

-76,337

199,291

-80,676

-85,113

-54,273

15,754

-1,004

89,414

-15,202

-56,793

-49,537

-52,683

-37,028

-125,037

-122,835

-84,383

65,044

-75,845

-34,413

-19,761

-20,990

-49,015

-38,772

-49,200

-78,382

-105,426

-143,416

-125,104

-83,511

-62,840

486,799

-23,280

-95,232

-133,448

-204,182

-79,429

-137,443

-103,396

-37,613

-10,071

-767

-13,449

-48,571

-75,865

-69,476

-62,949

-68,229

-78,476

-95,349

-45,869

-15,534

-61,629

-64,214

-86,787

-57,327

-57,520

-33,990

451

-61,175

-87,508

-46,282

-60,447

-42,930

-32,716

-49,928

-17,137

-82,015

-101,933

-104,664

-41,955

-28,058

-9,077

-26,873

-25,876

-22,804

-15,259

-32,342

-55,744

-91,041

-103,043

-26,085

-75,462

-73,419

-27,240

-69,643

-349,730

-790,281

-777,898

-780,136

-511,322

-634,880

-337,948

-478,677

-149,289

-265,994

-55,405

-9,589

13,548

-349,730

-859,925

-851,317

-883,179

-526,582

-662,938

-387,876

-478,227

-195,158

-266,761

431,394

-29,350

102,962

-510,195

-61,035

-105,281

253,555

-151,616

247,004

-140,279

283,519

-117,472

697,388

26,054

112,551

-4,879

-27,141

-88,218

-103,421

-124,411

-147,690

-161,139

-176,673

-237,848

-254,985

-264,062

-296,404

-322,489

-349,730

Total annual

earnings, up to

2011

Total annual

earnings, up to

2012

Annual

earnings

increase, by

firm age

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Total

8,669

-492,160

-407,222

-608,475

-269,857

-824,861

-558,234

-376,550

-380,986

-375,761

-99,890

-282,170

-174,966

-96,884

-5,127,601

8,669

-4,948,015

FIGURE A.1: Employment, firm age: 0 years

Empirical CDF

FIGURE A.3: Value added (DKK1,000). Firm age: 0 years

Empirical CDF

Reference firms

IM-firms

-1000

-500

Reference firms

500

IM-firms

1000

FIGURE A.2: Turnover (DKK1,000), firm age: 0 years

FIGURE A.4: Annual earnings(DKK1,000). Firm age: 0 years

-3000

Empirical CDF

1000

2000

Reference firms

3000

4000

IM-firms

5000

-2000

-1000

Reference firms

IM-firms

1000

Danish Agency for Science, Technology and Innovation

UFU, Alm.del - 2015-16 - Endeligt svar på spørgsmål 168: Spm. om ministeren vil fremsende relevante evalueringer af Globaliseringspuljen 2007-12, til Uddannelses- og forskningsministeren

Danish Agency for Science, Technology and Innovation