Cross-Border Venture Capital

Investments
Hortense Tarrade

Cross-Border Venture
Capital Investments
Why Do Venture Capital Firms
Invest at a Distance?

Foreword by Prof. Dietmar Harhoff, Ph.D

RESEARCH
Hortense Tarrade
München, Germany Bernhard Schmidt
Voestalpine Langenhagen, Deutschland
Linz, Österreich

Dissertation Ludwig-Maximilians-Universität München, 2010

D 19

ISBN 978-3-8349-3118-4 ISBN 978-3-8349-6939-2 (eBook)

DOI 10.1007/978-3-8349-6939-2

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie;

detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.

Springer Gabler
© Gabler Verlag | Springer Fachmedien Wiesbaden 2012
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole
or part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical
way, and transmission or information storage and retrieval, electronic adaptation, computer
software, or by similar or dissimilar methodology now known or hereafter developed.
Exempted from this legal reservation are brief excerpts in connection with reviews or schol-
arly analysis or material supplied specifically for the purpose of being entered and executed
on a computer system, for exclusive use by the purchaser of the work. Duplication of this pub-
lication or parts thereof is permitted only under the provisions of the Copyright Law of the
Publisher’s location, in its current version, and permission for use must always be obtained
from Springer. Permissions for use may be obtained through RightsLink at the Copyright
Clearance Center. Violations are liable to prosecution under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are
exempt from the relevant protective laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the
date of publication, neither the authors nor the editors nor the publisher can accept any legal
responsibility for any errors or omissions that may be made. The publisher makes no warranty,
express or implied, with respect to the material contained herein.

Cover design: KünkelLopka GmbH, Heidelberg

Printed on acid-free paper

Springer Gabler is a brand of Springer DE. Springer DE is part of Springer Science+Business Media.
www.springer-gabler.de
 V

Foreword
Venture capital financing is an important motor of entrepreneurship throughout the world. As
borders lose their relevance thanks to new technologies and sharply reduced transportation
costs, start-ups are increasingly international from the start. This raises the question of which
VC firm they should seek financing from – local or foreign. From the venture capitalist firms'
perspective, a similar question is posed: why would they restrict their investments to
companies in their immediate vicinity? And indeed they do not. Cross-border investments
have become more popular. But this does not come without challenges caused by the distance
between VC firm and distant portfolio company.

In her dissertation, Hortense Tarrade investigates this phenomenon that has been largely
neglected in the literature. Tarrade conducts thorough literature reviews in several relevant
fields of management as well as a series of interviews to develop her hypotheses. She then
goes on to apply econometric methods to a large longitudinal dataset recording over 200,000
VC investments around the world between 1989 and 2008.

In her econometric analysis, the author first studies the decision of VC firms to make
investments locally or at a distance. She then refines the examination by looking at the
determinants of a VC's preference for investing in particular countries.

The thesis delivered by Hortense Tarrade presents interesting new research results enhancing
our understanding of venture capital firms' investment behavior. It is a welcome complement
to the existing scholarly research on venture capital and reflects the new global conditions
which both entrepreneurs and financiers face.

Prof. Dietmar Harhoff, Ph.D.

 VII

Acknowledgements
My very first and utmost thanks go to Prof. Harhoff for his invaluable support and guidance
from day one in the dissertation process. His patience and encouragements have been
indispensable in pursuing this thesis to the final step – the publication of this book. Thank you
for selecting me among other candidates despite my unconventional profile, coming from the
industry and being a foreign student new to the German system. Over the three years (and
some) as a doctoral candidate, I am very grateful for the support on many occasions in
pursuing my interests within academia, for instance helping me obtain a position as a visiting
scholar at Columbia University, but also in broadening my horizon, making a visit to Stanford
University to experience the entrepreneurship first-hand and granting me a special leave of
absence to spend a few months at a strategy consulting firm. Thank you also to the team
around Prof. Harhoff for their relentless support and interesting discussions. In particular,
many thanks to Carolin Häussler for the countless hours spent together challenging and
polishing this thesis. I am also thankful for the instructive and inspirational courses led by my
Master of Business Research advisors, Prof. Kretschmer and Prof. Hess and for the additional
discussions and feedback with Prof. Tuschke, my second thesis advisor.

During my time at the Ludwig-Maximilians-Universität München, I had the opportunity to

work as a Project Manager at the LMU Entrepreneurship Center, gaining great pride in having
been a part of this very important endeavor. Thank you to Andy Goldstein, the LMU EC's
Executive Director for helping me balance my responsibilities at the Center and as a
university researcher. Thank you also to my colleagues at the Entrepreneurship Center,
Jeannine, Nina, Richard and Robert, for the many laughs we shared in the Giselastraße.

There are too many people who played a role in bringing this thesis to its final state to name
them all individually, be it colleagues and professors from the LMU or from Columbia, the
venture capitalists who agreed to be interviewed, or even friends and acquaintances providing
tips and ideas along the way. Many thanks to all of you.

Finally, I would like to express my deepest gratitude to my family and to my closest friends
for their patience especially during the last phase of the dissertation. You have always
believed in me and helped remind me of what is really important in life.

Hortense Tarrade
 IX

Table of contents

Foreword ..................................................................................................................................V

Acknowledgements .............................................................................................................. VII

Table of contents.................................................................................................................... IX

List of appendices ............................................................................................................... XIII

List of figures ........................................................................................................................ XV

List of tables .......................................................................................................................XVII

List of abbreviations........................................................................................................... XIX

1 Introduction ...................................................................................................................... 1
1.1 Motivation ................................................................................................................. 1
1.2 Research objectives ................................................................................................... 3
1.3 Structure of the thesis ................................................................................................ 4

2 Theoretical foundations and related literature ............................................................. 6

2.1 Benefits of cross-border venture capital investments ................................................ 6
2.1.1 Introduction .................................................................................................... 6
2.1.2 Risk reduction through international portfolio diversification ...................... 7
2.1.3 Revenue growth and strategic business development.................................... 9
2.1.3.1 Opportunism and firm growth ............................................................. 9
2.1.3.2 Expansion and creation of competitive advantage ............................ 10
2.1.3.3 Sourcing of location-specific advantages .......................................... 11
2.2 Costs of investing at a distance ............................................................................... 12
2.2.1 Introduction to agency theory ...................................................................... 12
2.2.1.1 Origins and fundamentals .................................................................. 12
2.2.1.2 Applications and findings .................................................................. 14
2.2.2 Agency theory as a framework for venture capital ...................................... 16
2.2.2.1 Principal-agent relationships between fund investors, VC firms and
portfolio companies .......................................................................................... 16
2.2.2.2 Mechanisms used by VC firms to deal with agency problems with
portfolio companies .......................................................................................... 19
2.2.3 Effects of distance on the VC-entrepreneur agency relationship ................ 23
X

2.2.3.1 Effects of distance on adverse selection and corresponding coping

mechanisms ...................................................................................................... 23
2.2.3.2 Effects of distance on moral hazard and corresponding coping
mechanisms ...................................................................................................... 24
2.2.3.3 Implications for investment performance .......................................... 27
2.3 Summary, research gaps and contribution of the thesis .......................................... 29

3 Hypotheses ...................................................................................................................... 32
3.1 Hypotheses development process ............................................................................ 32
3.2 Determinants of scope decision ............................................................................... 33
3.2.1 Local supply and demand ............................................................................ 33
3.2.1.1 Local supply of investment opportunities ......................................... 33
3.2.1.2 Local demand for investment opportunities ...................................... 34
3.2.2 VC ability and fit ......................................................................................... 35
3.2.2.1 VC fund characteristics ..................................................................... 35
3.2.2.2 VC firm characteristics ...................................................................... 38
3.2.2.3 VC investment characteristics ........................................................... 40
3.3 Determinants of target location decision ................................................................. 40
3.3.1 Target location characteristics ..................................................................... 40
3.3.1.1 Supply and demand for investment opportunities ............................. 40
3.3.1.2 Target location attractiveness of exit channels .................................. 41
3.3.2 VC ability and fit ......................................................................................... 41
3.3.2.1 Distance to target location ................................................................. 41
3.3.2.2 VC experience with the target location ............................................. 43
3.4 Summary of hypotheses .......................................................................................... 44

4 Construction and description of the data set and variables ....................................... 46

4.1 Sources of venture capital data ................................................................................ 46
4.1.1 ThomsonONE.com ...................................................................................... 46
4.1.2 VentureSource ............................................................................................. 48
4.1.3 CEPRES ....................................................................................................... 49
4.1.4 Proprietary data ............................................................................................ 50
4.2 Construction and description of the data set used in this thesis .............................. 50
4.2.1 Choice and construction of the data set ....................................................... 50
4.2.2 Description of VC firms in the data set ....................................................... 52
 XI

4.2.3 Description of VC funds in the data set ....................................................... 52

4.2.4 Description of portfolio companies in the data set ...................................... 54
4.2.5 Description of investments in the data set ................................................... 56
4.3 Creation of additional variables............................................................................... 59
4.3.1 Measurement of distance ............................................................................. 59
4.3.2 Measurement of deal flow (supply) ............................................................. 60
4.3.3 Measurement of competition (demand) ....................................................... 64
4.3.4 Measurement of VC ability and fit .............................................................. 65
4.3.4.1 VC fund-related variables .................................................................. 65
4.3.4.2 VC firm-related variables .................................................................. 67
4.4 Geography of the venture capital industry .............................................................. 68
4.4.1 Historical perspective .................................................................................. 68
4.4.2 VC champion nations................................................................................... 68
4.4.3 Cross-border VC investments ...................................................................... 69
4.4.3.1 Investment scope ............................................................................... 69
4.4.3.2 Target locations ................................................................................. 72
4.4.3.3 Cross-border investment champion VC firms ................................... 74
4.5 Summary: advantages and drawbacks of the data set ............................................. 76

5 Investment scope decision ............................................................................................. 78

5.1 Naïve scope model .................................................................................................. 78
5.2 Refined scope model ............................................................................................... 80
5.2.1 Econometric model ...................................................................................... 80
5.2.2 Estimation .................................................................................................... 82
5.2.3 Discussion .................................................................................................... 90
5.2.3.1 General remarks ................................................................................. 90
5.2.3.2 Discussion of the marginal effects..................................................... 91
5.3 Intermediary conclusion .......................................................................................... 95

6 Target location decision ................................................................................................. 97

6.1 Naïve target location model..................................................................................... 97
6.2 Refined target location model ................................................................................. 98
6.2.1 Econometric model ...................................................................................... 98
6.2.1.1 Introduction ....................................................................................... 98
6.2.1.2 Example of operationalization ......................................................... 101
XII

6.2.2 Estimation .................................................................................................. 102

6.2.2.1 Challenges and solutions ................................................................. 102
6.2.2.2 Construction of additional variables ................................................ 104
6.2.2.3 Results ............................................................................................. 105
6.2.3 Discussion .................................................................................................. 108
6.2.3.1 General remarks ............................................................................... 108
6.2.3.2 Discussion of the coefficients .......................................................... 110
6.3 Intermediary conclusion ........................................................................................ 113

7 Conclusion..................................................................................................................... 115
7.1 Summary................................................................................................................ 115
7.2 Managerial implications ........................................................................................ 118
7.3 Limitations of the thesis ........................................................................................ 119
7.4 Avenues for further research ................................................................................. 120

Appendix .............................................................................................................................. 123

Bibliography ........................................................................................................................ 153

 XIII

List of appendices

Appendix 1: Overview of interviewed VCs ........................................................................... 124

Appendix 2: Overview of the data collected from ThomsonONE.com ................................. 125
Appendix 3: Industry classification based on Venture Economics Industry Codes .............. 126
Appendix 4: Classification of nations by continents .............................................................. 127
Appendix 5: Number of observed investments by VC location (1989-2008)........................ 128
Appendix 6: Location of observed VCs (by nation) .............................................................. 129
Appendix 7: Number of VC firms participating in observed financing rounds ..................... 130
Appendix 8: Location of observed portfolio companies (by nation) ..................................... 131
Appendix 9: Number of observed financing rounds received by portfolio companies ......... 132
Appendix 10: Number of observed investments per industry ................................................ 133
Appendix 11: Specification of the random utility maximization nested logit model ............ 134
Appendix 12: Naïve scope model (base model)..................................................................... 136
Appendix 13: Naïve scope model (full model) ...................................................................... 137
Appendix 14: Descriptive statistics and correlation matrix (VC firms worldwide) .............. 139
Appendix 15: Descriptive statistics and correlation matrix (US-based VC firms) ................ 142
Appendix 16: Descriptive statistics and correlation matrix (non US-based VC firms) ......... 145
Appendix 17: Descriptive statistics and correlation matrix (Germany-based VC firms) ...... 148
Appendix 18: Target location model descriptive statistics and correlation matrix ................ 151
 XV

List of figures

Figure 1: Yearly share of cross-border VC investments (1989-2008) ....................................... 5

Figure 2: Model of the relationship between LPs, VCs, funds and portfolio companies ........ 17
Figure 3: Hypotheses development process ............................................................................. 33
Figure 4: Vintage year of observed VC funds.......................................................................... 53
Figure 5: Number of observed portfolio companies by founding year and industry ............... 56
Figure 6: Yearly number of observed VC investments (1989-2008) ....................................... 58
Figure 7: Number of financing rounds per observation year and by PC industry ................... 58
Figure 8: Share of investments at a distance by US-based VC firms (1989-2008) ................. 71
Figure 9: Naïve scope model .................................................................................................... 79
Figure 10: Refined scope model............................................................................................... 82
Figure 11: Naïve target location model .................................................................................... 98
Figure 12: Refined target location model ................................................................................. 99
Figure 13: Application of the refined target location model to German VC firms ................ 102
 XVII

List of tables

Table 1: Hypotheses overview ................................................................................................. 45

Table 2: Location of observed VC firms (by continent) .......................................................... 52
Table 3: Type of observed VC funds ....................................................................................... 53
Table 4: Summary statistics of observed VC funds ................................................................. 54
Table 5: Location of observed portfolio companies (by continent) ......................................... 55
Table 6: Industry of observed portfolio companies ................................................................. 55
Table 7: Current status of observed portfolio companies ........................................................ 56
Table 8: Investment stage of observed financing rounds ......................................................... 57
Table 9: Examples of distance measurement ........................................................................... 60
Table 10: Example of local supply measurement for German VC firms ................................. 64
Table 11: Industry specialization of observed VC funds ......................................................... 65
Table 12: Stage specialization of observed VC funds.............................................................. 66
Table 13: Evolution of the origin of observed investments (1989-2008) ................................ 68
Table 14: Observed investments by VC location ..................................................................... 69
Table 15: Scope of observed investments by VC continent..................................................... 70
Table 16: Scope of observed investments by top 10 VC nations ............................................. 71
Table 17: Scope evolution of investments by Europe-based VCs (1989-2008) ...................... 72
Table 18: Top 10 investment target locations by VC nation.................................................... 73
Table 19: VC firms with the most investments at a distance ................................................... 75
Table 20: Refined scope model (coefficients).......................................................................... 84
Table 21: Refined scope model (marginal effects at the national level) ............................ 86
Table 22: Refined scope model (marginal effects at the global level) ..................................... 88
Table 23: Intermediary results (investment scope decision) .................................................... 96
Table 24: Office presence of German VCs in foreign top target locations ............................ 105
Table 25: Target location model (coefficients) ...................................................................... 106
Table 26: Summary of alternatives in the nested logit model ................................................ 109
Table 27: Intermediary results (target location decision) ....................................................... 114
 XIX

List of abbreviations

e.g. exempli gratia

et al. et alii
EUR euro
EVCA European Private Equity & Venture Capital Association
FDI Foreign direct investment
GP general partner
HQ headquarters
i.e. id est
IPO initial public offering
IT information technology
LP limited partner
PC portfolio company
PE private equity
ROW rest of the world
SD standard deviation
TL target location
UK United Kingdom of Great Britain and Northern Ireland
US United States of America
USD US dollar
VC venture capital / venture capitalist
 1

1 Introduction

1.1 Motivation
Entrepreneurs have been considered an important engine of long-term economic growth since
Schumpeter popularized the concept of “creative destruction” (1942, 73). But the new
ventures striving to bring innovations to the market often times require large upfront
investments that cannot be borne alone by young companies still generating little or no cash
flow (Gompers / Lerner 2006). By catalyzing the development of high-tech start-ups
(Hellmann / Puri 2002), venture capitalists (VCs)1 play an important role in social and
economic growth2, innovation and value creation (Kortum / Lerner 2000; Gompers / Lerner
2001; Hege et al. 2003; EVCA 2005; Global Entrepreneurship Monitor 2009). Indeed, VCs
do not simply inject the necessary funding into promising young companies, they also take an
active role in helping to manage the ventures, advising them on strategic matters, sharing their
business knowledge and introducing the company managers to the relevant business contacts.
By providing so-called “smart money”, VCs support fledgling businesses in overcoming the
liabilities of newness and smallness (Sapienza / Gupta 1994; Baum / Silverman 2004;
Sørensen 2007).3

Arguably because of the VCs’ role as strategic advisors and due to the risks facing young
firms4, academic research mostly assumes that VC is a local business and that VCs are

1
Following the European Private Equity and Venture Capital Association’s recommendation, I define venture
capital (VC) as the subset of private equity (PE) focusing on investments in entrepreneurial ventures, as
opposed to mature businesses.
2
For example, according to the European Private Equity and Venture Capital Association’s research paper on the
employment contribution of venture capital in Europe, employment in venture-backed companies increased by
an average of 30.5% per year between 1997 and 2004 in comparison to an average annual growth rate of total
employment in the European Union of 0.7% between 2000 and 2004 (EVCA 2005).
3
72% of seed and start-up companies say they could not have existed without venture capital according to a
European Private Equity and Venture Capital Association survey of the economic and social impact of venture
capital in Europe (EVCA 2002).
4
Venture capital is sometimes called “risk capital” in English (e.g., Weidig / Mathonet 2004) and in foreign
languages (e.g., “capital risque” in French or “Risikokapital” in German), emphasizing the uncertainty of VC
investment returns.

H. Tarrade, Cross-Border Venture Capital Investments, DOI 10.1007/978-3-8349-6939-2_1,

© Gabler Verlag | Springer Fachmedien Wiesbaden 2012
2

“typically geographically proximate to their investees” (Cumming 2006, 1084).5 Indeed, the
basic principles of agency and transaction cost theory dictate that geographic remoteness and
cultural distance between the VC firm and the venture should make it more difficult for the
VC and the managers to make full use of the consulting and advisory dimensions of the VC-
entrepreneur relationship than if the venture was close to the VC, thus possibly jeopardizing
the venture’s chances of success and hereby potentially harming the VC’s investment return.

However, as the VC industry keeps growing both in terms of amounts invested and number of
investments6 (Megginson 2004), cross-border venture capital investments have become an
increasingly relevant phenomenon (Baygan / Freudenberg 2000; Deloitte / NVCA 2007;
Aizenman / Kendall 2008; Alhorr et al. 2008): from just about 5% in the early nineties, the
share of international investments7 near tripled to over 15% of all investments worldwide per
year in the recent years.8 In a recent worldwide industry survey, Deloitte and the National
Venture Capital Association (2009) find that 52% of the more than 700 VCs interviewed
currently invest outside their home countries. While it is true that the VC industry has mostly
been a local one in the US9, European VCs do not seem to focus that strongly on investments
within their own nation. For example, UK and Germany-based VC firms have made about
half of their investments in ventures outside of their own nation over the past 20 years.10 In
fact, even the share of investments made by US-based VC firms into non-domestic ventures
tripled since 1989.11

5
This may also be partly due to the fact that a large part of the research on venture capital originates from the
US, where the VC industry has indeed been relatively confined to national borders in comparison to the rest of
the world (Deloitte 2006).
6
This statement naturally excludes the so-called “dot-com bubble” and its subsequent burst (1998-2001).
7
An “international” investment is defined as an investment in which the VC firm providing the capital and the
portfolio company receiving the financing are located in different nations. In this thesis, international
investments are synonymously referred to as “cross-border” investments or investments “at a distance”.
8
See Figure 1 at the end of this chapter.
9
The US National Venture Capital Association (NVCA) analyzes this observation in a recent industry report:
“Years ago, venture capitalists rarely travelled beyond their own area codes to find new deals.” However, in
the last years, VCs “have begun to identify and nurture companies by following the trail of innovation (…)
even to regions with little or no previous venture capital history but with vibrant research communities and
ambitious entrepreneurs” (NVCA 2009, 16).
10
Own calculation based on ThomsonONE.com data retrieved, completed and analyzed by the author as
described in Chapter 4.
11
Idem.
 3

These observations raise many questions, among which:

x Why do venture capital firms invest internationally?

x What drives venture capital firms to invest in portfolio companies situated in a certain
target location rather than in others?

A possible explanation is that VCs seek to capture additional utility from investing
internationally and in certain target locations in particular which they would not be able to do
in their home market. For example, VC firms may seek international portfolio diversification.
They may also aim to circumvent lower expected returns or a weak deal flow in the domestic
markets (Meyer / Shao 1995). VCs perhaps seek to create a path for future investments (e.g.,
in an uncertain but growing market) and to test the market before scaling up investment
activities. Alternatively, there may be economies of scale at the portfolio level, in which other
portfolio companies benefit from the knowledge gained in certain distant investments via
knowledge transfer by the general partners. Alternatively, the costs of investing at a distance
may not be as high as assumed in the literature and VCs may have developed dedicated
mechanisms to minimize the additional agency costs of investing at a distance.

1.2 Research objectives

Since the choice of some VC firms to invest across borders goes against the expectations
drawn from agency theory in particular, but is still is a widespread practice in the VC
industry, it is important for researchers, practitioners (entrepreneurs and financiers) and policy
makers alike to understand the motivations of cross-border investments. Surprisingly, little
research has been done to analyze the phenomenon of cross-border VC investments. A
possible explanation for this “major research gap” (Wright et al. 2005, 135) could be the
difficulty of obtaining data on venture capital deals (Lerner 1995). This research project offers
to fill parts of this gap by investigating two research questions. The first research question
investigates why certain venture capital firms choose to invest at a distance12 while others do
not: are they reacting to macroeconomic factors in doing so? Can VC firms’ international
investment behavior be determined by characteristics intrinsic to the VC firm? Second, since
VC firms do not always invest in the same set of target locations, it is interesting to
understand what drives venture capital firms to invest in portfolio companies situated in a
certain target location rather than in others. For example, what role does distance (geographic

12
For a definition and discussion of geographic distance, see section 4.3.1.
4

or cultural) play in the VC’s target location decision? With these two research question, the
thesis contributes to the still new stream of research in international venture capital in various
ways. First, while the literature often consists in ad-hoc analyses, the empirical part of this
study relies on a triple approach combining theoretical grounds, extensive literature review
and expert interviews. Second, this dissertation covers a large range of determinants,
considering both macroeconomic factors and venture capital firms’ specificities as
recommended in the existing literature (Gupta / Sapienza 1992; Hall / Tu 2003; Wright et al.
2005). The results are therefore more granular and comprehensive than in existing studies.
Third, the longitudinal comprehensive data set used in this research project is one of the most
extensive in the field of venture capital research to my knowledge. In contrast with the current
literature, the data used here covers a long time frame (1989-2008) and considers VC
investments on a worldwide scale, which is interesting based on the introductory remarks on
the differences between nations in terms of international investment patterns (e.g., US vs.
Germany). Also, the data relies on surveys reporting the actual investment behavior of VC
firms rather than their stated preferences and intentions, thus providing a better
comprehension of VCs’ preferences in terms of investment scope13 and their choice of
investment target location (Gupta / Sapienza 1992; Wright et al. 2005).14

1.3 Structure of the thesis

In the next chapter (Chapter 2), I review related literature to investigate the possible benefits a
VC firm could have from investing internationally and examine the costs of distance on the
relationship between VCs and their portfolio companies based on an extensive study of
agency theory and related literature. In Chapter 3, I build on the previous theoretical
framework and literature review as well as on own expert interviews to build the hypotheses
for the empirical implementation. After that, I consider the construction of the data set and
highlight relevant descriptive statistics in Chapter 4. In the following two chapters, Chapter 5
and Chapter 6, I provide partial answers to each of the research questions, starting with the
scope decision, followed by the target location decision. For each research question, I develop
an econometric model, use the longitudinal data to test the models and draw conclusions from
the results of the analysis. In the final section, Chapter 7, I summarize and discuss the

13
The terminology of “scope” follows Gupta and Sapienza (1992)
14
For a more in-depth discussion of the current research gaps and of the specific contributions of this thesis, see
section 2.3.
 5

implications of the findings, identify the principal limitations of the thesis and propose
avenues for future research.

Figure 1: Yearly share of cross-border15 VC investments (1989-2008)16

20%

15%

10%

5%

0%
1990 1995 2000 2005
Observation year

Total number of observed VC investments for which both VC and PC location are known (1989-2008): 213,452
Total number of observed cross-border VC investments (1989-2008): 29,667

15
A “cross-border” investment is defined as an investment in which the VC firm providing the capital and the
portfolio company receiving the financing are located in different nations.
16
Unless otherwise specified, the graphs and tables in this dissertation are an own representation based on
ThomsonONE.com data retrieved, completed and analyzed by the author as described in Chapter 4.
6

2 Theoretical foundations and related literature

In this chapter, the decision to invest at a distance is examined from two opposite
perspectives. First, the potential rationales behind a VC firm’s decision to invest in
international ventures are discussed based on managerial finance and internationalization
literature. Second, the costs associated with investing at a distance are analyzed using agency
theory. The third subsection concludes and highlights the identified research gaps and
contribution of the thesis.

2.1 Benefits of cross-border venture capital investments

2.1.1 Introduction

Despite the common assumption that VCs are home-biased, there are several reasons why VC
firms may decide to engage in cross-border investments. First, investing internationally could
be the result of a strategy to minimize a VC firm’s costs in terms of investment risk by
pursuing geographical portfolio diversification. Second, a VC firm may invest across borders
to generate more revenue – as an opportunistic reaction to an interesting investment
opportunity that happens to be in a foreign location or as a part of a long-term strategy to
develop its own competitive advantages or to reap location-specific advantages. In the
following paragraphs, a brief literature overview provides first indications about the rationales
behind distant investments by VC firms. The part on portfolio diversification relies on
managerial finance literature and the rest of this review mostly draws from research on the
internationalization of the firm (mainly via export and foreign direct investment), since a VC
firm’s international investments (investments for which the portfolio company and the VC
firm are not located in the same nation) can be considered as a form of FDI17.

17
According to Shan and Song (1997), the “US Department of Commerce defines a foreign investment as direct
when a single investor has acquired a stake of 10% or more in a US firm”. Despite the arbitrariness of the 10%
value, the notion of a certain degree of control over the investment is important. This is typically the case with
VC investors, hence the parallel between FDI and cross-border VC investments used in this chapter.

H. Tarrade, Cross-Border Venture Capital Investments, DOI 10.1007/978-3-8349-6939-2_2,

© Gabler Verlag | Springer Fachmedien Wiesbaden 2012
 7

2.1.2 Risk reduction through international portfolio diversification

The benefits of portfolio diversification are well known (e.g., French / Poterba 1991; Norton /
Tenenbaum 1993; Lu / Beamish 2004). In particular, geographic portfolio diversification
would make sense for VCs because “since the fortunes of different nations do not always
move together, investors can diversify their portfolio by holding assets in several countries”
(French / Poterba 1991, 222).

Under the commonly accepted assumption that investors seek to maximize their wealth but
are risk averse (Markowitz 1952; Sharpe 1964)18, VCs would prefer a higher expected rate of
return to a lower one (ceteris paribus) and a lower value of risk19 to a higher one (ceteris
paribus). Therefore, the goal of a diversified portfolio strategy is to achieve “the most
favorable combination of risk and expected return” (Lintner 1965, 589). Assuming
“frictionless market conditions”20, complete information, similar expected returns and risks
for all companies, a fixed VC-specific cost of set-up to enter a new country and a similar
weight of investments in each country in which the investor is diversified, Lossen (2007)
shows with a theoretical model based on the works of Markowitz (1952), Sharpe (1964) and
Lintner (1965) that:

x an increase in the VC firm-specific cost of set-up in a new country should lead to a

decrease in the optimal number of countries the VC chooses to invest in,

x an increase in the size of the fund managed by the VC firm should lead to an increase
in the optimal number of countries the investor chooses to invest in,

x the higher the expected rate of return from each investment, the lower the optimal
number of countries the VC firm should choose to invest in and

x the larger the risk associated with each investment, the higher the optimal number of
countries the VC firm should choose to invest in.

Since the cost of set-up is likely to be higher for early-stage investments, Lossen’s result
would indicate that VC firms specializing in later investment stages are likely to diversify

18
This assumption was verified in the interviews I led with VC investors.
19
Risk being the standard deviation of the expected rate of return from the investments chosen by the investor
(Weidig / Mathonet 2004).
20
No taxes, no transaction costs, perfect divisibility of assets (Lossen 2007, 19).
8

across more countries than firms investing at earlier stages.21 In addition, syndication may be
used to lower the set-up costs and allow for geographic diversification. Taken together, the
last two findings would explain why VC firms should not invest in a single country and why
they also do not diversify in all countries over the world, which is in line with Gupta and
Sapienza (1992).

In practice, while some authors do not find evidence for a significant positive relationship
between geographic diversification and firm or fund performance (Morck / Yeung 1991;
Lossen 2007), there is also repeated empirical evidence that geographic diversification indeed
helps to increase returns and to spread investment risk over several nations and that
geographic diversification is positively correlated to firm performance (Doukas / Travlos
1988; Kim et al. 1993; Delios / Beamish 1999). Furthermore, a recent industry survey of 505
VC general partners around the world shows that diversification of geographical risk is one of
the key reasons why venture capital firms are interested in investing outside of their home
market (Deloitte 2006).

However, these conclusions raise the question of why certain investors do not make use of
geographic diversification. In their paper on how investors diversify in international equity
markets, French and Poterba (1991) investigate why investors tend to hold most of their
wealth in domestic assets despite the recognized benefits of international portfolio
diversification. First, they rule out institutional explanations of “underdiversification”, such as
different tax burdens in foreign and domestic contexts, that could limit investors’ ability to
reach their expected returns overseas. While explicit limits on cross-border investments may
slow down investments by foreign investors in certain nations 22, these factors would not be
strong enough to justify the level of observed underdiversification. Instead, French and
Poterba argue that the lack of international diversification is due to the investors’ preferences
and return expectations: they show that investors systematically overvalue domestic
investments by several hundred basis points to justify their home bias.23 For example, “British
investors must expect annual returns in the U.K. market more than 500 basis points above
those in the U.S. market to explain their 82 percent investment in domestic shares” (French /

21
For example, information on the development of technologies and market response is likely to be more
widespread and cheaper to acquire in late stages than in early stages.
22
The authors give the example that foreign investors require a ministerial authorization to hold more than 20%
of any firm in France.
23
In another stream of research, Franke, Gruber, Harhoff and Henkel (2006) highlight the similarity biases of
VCs’ evaluation of start-up teams with regards to training and professional experience.
 9

Poterba 1991, 223). In addition, the authors’ analysis of the actual portfolio holdings of
investors in the US, Japan and UK, implies that investors not only dramatically overvalue
their own domestic markets, but also overvalue markets that are geographically close (e.g.,
Canada for US investors, Germany and France for UK investors). A possible explanation of
this phenomenon is that investors perceive different risks associated with different locations
and that they typically associate higher risks with more distant markets.

2.1.3 Revenue growth and strategic business development

2.1.3.1 Opportunism and firm growth

Since the role of VC firms is to identify and grow high-potential companies (Gompers /
Lerner 2006), VCs must be sensitive to their immediate environment and must realize
whether they can fulfill their role with local investments. A VC firm’s decision to pursue
international investments could be an answer to a lack of interesting investment opportunities
in the VC’s home market or an increased intensity of competition between local VC firms,
similar to the export decision of other firms being motivated by the “awareness of specific
foreign markets as a possible solution to problems in the domestic market” (Reid 1981, 102).

Another explanation of VC firms’ decision to invest at a distance is that they may seek to
maximize their welfare by capturing investment opportunities regardless of their location.24
Applying Simpson and Kujawa’s (1974, 107) terminology25, VCs may respond to foreign
investment “stimuli” (unsolicited applications from foreign companies) in the same way firms
can respond to an “export stimulus (unsolicited order from a foreign customer)” with a
positive export decision. This possibility seems to be confirmed in Deloitte’s survey on the
globalization of venture capital investments: “[VC] firms are interested in identifying the best
opportunities wherever they may be” (2006, 7).26

However, although VC firms are motivated by profit gain, and even if the lucrative
opportunities are not always located at the doorstep of the firm, not all VCs may be able to
pursue investments at a distance. To continue the parallel with research on the strategy of
multinational firms, Caves (1971) argues that the multinational firm is able to grow

24
In fact, it seems that equity capital does flow towards locations recognized for offering high profits (Caves
1971).
25
This original study investigates the export decision process of US-based manufacturing firms.
26
Deloitte’s report concludes that VCs “explore the right business model for a more global approach to finding
opportunities and generating better returns for their investors” (2006, 21)
10

internationally by leveraging some of its firm-specific assets in order to offset its liability of
foreignness, which corresponds to what Madhok (1997, 39) calls the firm’s “exploitation of
existing competitive advantage”. One way to apply this view to the case of VC firms is to
consider those VC firms that are industry-specialized. Having accumulated extensive
experience in an industry over time, they may be better suited to finance and support the
growth of foreign companies, because they have a competitive advantage over the company’s
national VC firms, as the industry knowledge is difficult and costly to acquire, thus
compensating for their liability of foreignness in new international markets.

2.1.3.2 Expansion and creation of competitive advantage

Foreign markets are not only a potential opportunity for expansion and firm growth (Reid
1981). Firms also realize economies of scale by learning through internationalization (Zahra
et al. 2000), which is what Lu and Beamish (2004, 599) call “experiential learning”, because
they acquire new knowledge in the process which they can use to build additional value-
creating skills (Barkema / Vermeulen 1998; Caves 2007). For example, it is common that
high-tech firms seek to expand their technical knowledge through foreign direct investment
(Wilbur / Alcacer 2002), but in doing so, these firms also develop their ability to capitalize on
new dynamic markets, because they learn how to better recognize important changes
following their internationalization (Zahra et al. 2000). Applying these findings to VC firms,
it is likely that the first international investments of a VC firm are the most costly, so that
some firms will not take the step of internationalization – perhaps because their size would
not support the expansion, or because they have not yet developed enough of a competitive
advantage (e.g., deep industry knowledge and experience). However, due to the learning
curve and economies of scale, it is likely that VCs who have started investing at a distance
will continue to do so because they should be increasingly good at identifying the proper
investment opportunities and managing investments overseas (Hall / Tu 2003) – just as prior
experience in a host country increases the propensity of a firm to pursue projects in that
country relative to other investment options (Davidson 1980). It is also possible that VC firms
leverage the knowledge gained through internationalization to benefit other PCs (national or
international investments).

In addition to strengthening or developing new competitive advantage (Madhok 1997),

multinational firms increase their market power over suppliers, distributors and customers via
internationalization (Kogut 1985). A comparable phenomenon could apply to VC firms who
 11

may use their international investment experience as a seal of added value for their suppliers
(investors in VC funds seeking international diversification), their distributors (possible
syndication partners seeking introduction to a new market), for their customers (portfolio
companies seeking international growth) in order to negotiate better deal terms for
themselves. Viewed in this way, expanding internationally could be viewed as a long-term
investment for VC firms, although the associated “set-up costs” (Lossen 2007, 22) may be
steep at first.

2.1.3.3 Sourcing of location-specific advantages

National comparative advantage is an essential criteria in the target location decision of FDI
(Caves 1971). For example, access to the Asian market is the number one reason why US-
based VC firms would their expand investment focus to China (Deloitte 2006). In the early
nineties, Dunning (1990, 29) already noted: “In recent years (…) the main country specific
locational determinants have shifted to reflect the innovatory and entrepreneurial dynamism
of the recipient economy”. For example, since technological advantages are “heterogeneously
distributed among countries” (Shan / Song 1997, 267), locations with a high research and
development intensity are more likely to attract investors in research-intensive industries than
states with low research and development intensity (Wilbur / Alcacer 2002) and the rarity of
high-tech start-ups as investment targets can be an explanation for a lack of equity investment
into a nation (Teece 1992). Looking at the decision of multinational firms to acquire or merge
with a foreign firm, Friedman et al. (1992, 404) observe that the decision to invest in a
particular location is motivated by “market structure and diversification considerations” and
may not necessarily be driven by the specific location. However, location-specific
considerations are important for the establishment of a new plant: in their analysis of foreign
multinationals’ FDI decision in the US, the authors find that access to markets, labor market
conditions and state promotional efforts are significant drivers of the location decision.27 If
these findings can be ported to a VC’s decision to invest at a distance, then their decision may
be motivated by the attractiveness of a certain location. For instance locations with “a large
amount of existing activity” (Friedman et al. 1992, 409) may be considered as more attractive
by VCs for an investment, because the amount of activity signals that the favorable conditions
enumerated above are met.

27
The original analysis is carried at the US state level.
12

2.2 Costs of investing at a distance

Despite the various benefits associated with cross-border VC investments, many firms do not
invest beyond their home market. In this section, I introduce the basics of the agency theory
and demonstrate that the relationship between the venture capital investor and the
entrepreneur receiving VC funding is akin to a principal-agent set-up. I use a thorough
literature review to analyze the specificities of this interaction and draw the conclusions
relevant to this thesis. In particular, I infer first learnings on the possible effects of distance on
the VC-entrepreneur relationship.

2.2.1 Introduction to agency theory

2.2.1.1 Origins and fundamentals

In the late 18th century, Adam Smith (1776, 439) was already describing agency problems in
organizations as a consequence of the separation between ownership and control: “The
directors of such companies, however, being the managers rather of other people’s money
than of their own, it cannot well be expected, that they should watch over it with the same
anxious vigilance with which the partners in a private copartnery frequently watch over their
own. (…) Negligence and profusion, therefore, must always prevail, more or less, in the
management of the affairs of such a company.” Although its early foundations can be traced
back at least to the 1930’s, with Ronald Coase’s work on the firm (1937), agency theory
really stems from the stream of research by economists in the 1960’s and 1970’s (Arrow
1964; Wilson 1968; Arrow 1971) exploring the behavior of cooperating parties (individuals or
groups) with differing risk preferences.28 Agency theory broadens this literature by including
the agency problem (Eisenhardt 1989) and analyzing the patterns in an agency relationship.
Agency theory becomes a dominant theory of governance during the 1980’s, leading to
abundant academic research based on this theory.

An agency relationship refers to a relationship between a party (the principal) who engages
another party (the agent) to perform a task on his behalf, which involves delegating some
decision making to the agent (Ross 1973; Jensen / Meckling 1976). The simple agency model

28
“Risk adverse actors will often reject a contract in which they bear most of the risk even if it is efficient. Risk
adverse agents essentially purchase insurance by accepting fixed salary contracts with lower total value but
less variation. Therefore, taking risk into account allows economists to explain the existence of “inefficient”
contracts” (Kiser 1999, 149). For example, Wilson (1968) analyzes the case of syndicates, i.e., a group of
individuals making a common decision under uncertainty, with the perspective of a joint payoff.
 13

assumes that both principal and agent are self-interested so that both will seek to maximize
their expected utility through their actions. If they have conflicting desires and different
attitudes towards risk, it is likely that the agent’s motivation will not perfectly align with that
of the principal and that as a result, the agent will not act in the principal’s best interest. If the
principal does not fully know the agent’s activities and can only observe the outcome of the
agent’s behavior, this discrepancy leads to the agency problems of moral hazard and adverse
selection (Reuer / Miller 1997). Adverse selection is the risk that the agent may misrepresent
his skills and abilities to the principal, letting him believe that he can perform the task to be
delegated to him. Moral hazard is the risk incurred by the principal that the agent may not
engage his best efforts in realizing the task delegated to him.

Two solutions are available to the principal to mitigate agency problems (Van Osnabrugge
2000): 1) discovering the agent’s behavior and 2) incentivizing the agent to reach the
expected outcome. First, the principal can discover the agent’s behavior through monitoring
(i.e., through a behavior-based contract). However, since information is not free, monitoring
comes at a cost and investing in correcting the agent’s behavior only pays off if the gain from
the correction exceeds the cost of performing the correction (Mitnick 1997). Second, the
principal can align the agent’s preferences by contracting on the outcome of his behavior, thus
transferring the risk to the agent. The costs associated with this solution consist of the
investment required to measure the outcome (assumed to be low in the simple model) and of
the cost of transferring the risk to the agent who is assumed to be risk averse.

The problem of the agency set-up can be formalized in a utility model29: Let ‫ܩ‬ሺǤ ሻ be the
agent’s utility function and ܷሺǤ ሻ the principal’s utility function. The act chosen by the agent is
noted ܽ ‫ܣ א‬. In the random state of nature ߠ ‫ א‬π, the random payoff of the agent’s act is
‫ݓ‬ሺܽǡ ߠሻ. The fee scheduled to be paid by the principal to the agent for his services is the
function ݂, with ݂ ൌ ݂ሺ‫ݓ‬ሺܽǡ ߠሻǢ ߠሻ. Since the agent seeks to maximize his expected utility, he
will choose an act ܽ so as to:

ƒš  ሼ‫ܩ‬ሾ݂ሺ‫ݓ‬ሺܽǡ ߠሻǢ ߠሻሿሽ

௔ ఏ

with ܽ଴ being the optimal act chosen by the agent, conditional to the particular fee schedule,
i.e., ܽ଴ ൌ ܽሺ‫ۄ݂ۃ‬ሻ, where ܽሺǤ ሻ is a mapping from the space of fee schedules into ‫ܣ‬. It follows

29
The following canonical definition of the agency problem is adapted from Ross (1973).
14

that if the principal has complete information about the fee to act mapping, and since he seeks
to maximize his utility, he will now choose a fee so as to:

ƒš  ሼܷሾ‫ݓ‬ሺܽሺ‫ۄ݂ۃ‬ሻǡ ߠሻ െ ݂ሺ‫ݓ‬ሺܽሺ‫ۄ݂ۃ‬ሻǡ ߠሻǢ ߠሻሿሽ

‫ۃ‬௙‫ۄ‬ ఏ

2.2.1.2 Applications and findings

Agency theory determines the most efficient contract governing the principal agent
relationship given a set of assumptions on the people (self-interest, risk aversion), the
organizations (e.g., goal conflicts, information asymmetries) and the cost of information, and
extensions of the simple model are made by relaxing these assumptions. The context
determines which contract (behavior-based or outcome-based) is best appropriate to minimize
agency problems (Eisenhardt 1984). Findings are obtained through logical and mathematical
proofs as well as empirical qualitative research (Eisenhardt 1989).

There are two streams of research in agency theory: the positivist stream and the normative
stream (Eisenhardt 1989). The positivist stream focuses on identifying the situations in which
agency problems are likely to occur and describes the governance mechanisms solving the
problems linked to the agent’s self-serving behavior (Jensen 1983). For example, the theory
states that outcome-based contracts and information systems (e.g., efficient capital markets,
boards of directors) are effective in curbing agent opportunism (Jensen / Meckling 1976). The
normative stream focuses on how to structure the contractual relation between the principal
and the agent to optimize the principal’s welfare in a given context (Jensen / Meckling 1976;
Holmstrom 1979), varying uncertainty, risk aversion, information and other variables. For
instance, relaxing the hypothesis of the agent’s risk aversion, it follows that wealthy agents
may be less risk averse, so that outcome-based contracts are favorable to the principal
(MacCrimmon / Wehrung 1990). If one relaxes the hypothesis of goal conflicts, for instance
in clan-oriented firms (Ouchi 1979), then goal conflicts are less likely and both principal and
agent are likely to work towards the same outcome, so that behavior-based contracts would be
preferable. Other variables can be introduced too. For example, task programmability reflects
the routine aspect of the task performed by the agent: task programmability is much higher for
a cashier than for an entrepreneur (Eisenhardt 1985) and the principal should rather use a
behavior-based contract with the cashier and an outcome-based contract with the entrepreneur
(Eisenhardt 1988).
 15

Agency theory is highly relevant to business research because essentially, all contractual
arrangements contain elements of agency and principal-agent configurations are universal.
For example, the relationship between stockholder and manager or between employer and
employee are subject to agency problems (Ross 1973; Jensen / Meckling 1976). Therefore,
agency theory has been applied to many fields of business research, such as economics,
management, finance and many more. In economics, agency theory is used to study the
behavior of actors in a situation where the interests of owner and manager diverge. The theory
is particularly valuable for economists who see the organization as a “black box” (Eisenhardt
1989, 58). For example, Jensen and Meckling (1976, 8) provide an explicit agency theory of
the firm as a “nexus of contracts” and show how individuals in the firm can reach an
equilibrium allowing profit maximization despite conflicting objectives. Management
scholars have used agency approaches to explore the behavior in boards of directors (e.g.,
Fama / Jensen 1983; Kosnik 1987; Dalziel et al. 2009) and investigate topics such as
organizational control (e.g., Thompson 1967; Abrahamson / Choelsoon 1994), bargaining
(e.g., Kesner et al. 1994) and compensation practices (e.g., Eisenhardt 1984; Conlon /
McLean Parks 1990). For instance, Eisenhardt (1984) investigates the factors affecting the
design of an optimal employment contract, balancing the trade-offs between outcome and
behavior-based contracts. In organizational behavior, agency theory has been applied to the
analysis of acquisition and diversification strategies (e.g., Arrow 1975; Amihud / Lev 1981).
For example, Arrow (1975) uses the theory to justify the need for vertical integration. In the
field of finance, agency theory can be used to understand ownership and financing structures
(e.g., Jensen / Meckling 1976; Agrawal / Mandelker 1987). Applications of the agency
problems also reach to domains adjacent to business research. For instance, political scientists
use the theory to analyze subjects such as regulatory policy and bureaucratic behavior (e.g.,
Spiller 1990) or political corruption (e.g., Alt / Lassen 2003). Finally, in sociology, agency
theory is mostly applied to dyad-settings and explains phenomena such as blame (e.g.,
Leatherwood / Conlon 1987) or lying (e.g., Grover 1993) in organizations for example.

Information asymmetries between VCs and entrepreneurs have also been shown to cause
agency problems (e.g., Sapienza / Gupta 1994; Muzyka et al. 1996; Wright / Robbie 1998;
Van Osnabrugge 2000), due to the partial separation of ownership and control (Smith 1776).
In the following section, I introduce the parties involved in the venture capital system and
show in more details how agency theory can be applied to the venture capital context.
16

2.2.2 Agency theory as a framework for venture capital

2.2.2.1 Principal-agent relationships between fund investors, VC firms and portfolio

companies

A VC firm’s core competency lies in its ability to identify, retain and grow high-potential
young companies, with the ultimate goal to reach or exceed the target returns expected by its
investors.30 For the questions at stake in this thesis, it is essential to understand who are the
parties involved in the venture capital system and how they interact. The venture capital
system is depicted schematically in Figure 2 and described in greater detail below.

VCs regularly raise money for new venture funds from diverse investors (mostly institutional
investors such as banks, pension funds, insurance companies…) with the promise to achieve
high returns with smart investments in young companies (Sorenson / Stuart 2001). Investors
in VC funds are Limited Partners (LPs) and receive a share of the profits generated by the VC
fund in exchange for their capital investment.

Once it raised a fund, the VC firm is responsible for managing this fund and the LPs pay a fee
to the VC firm for this service. Also, the VC firm receives a commission on the returns
generated by the VC fund. The VC firm seeks to create value by selecting and investing
capital from the VC fund into companies with a high growth potential it takes into its
portfolio.31 But the VC firm does not only invest capital in the PCs, it also supports their
growth by providing guidance and strategic consulting to the management team (Weidig /
Mathonet 2004). Typically, one general partner at the VC firm is responsible for the
company’s proper development. In exchange for the financial investment and the consulting
services they receive from the VC firm, portfolio companies give out equity, meaning that a
share of the portfolio company belongs to the VC firm via the managed VC fund.

30
A VC I interviewed in the context of this thesis puts it bluntly: “There is only one goal: produce superb
returns”.
31
The mere fact of investing in several companies with the same fund raises the VC’s ability to generate high
returns with a mitigated exposure to risk (“naïve diversification”). For example, Schmidt shows that there is a
“high marginal diversifiable risk reduction of about 80% when the portfolio size is increased to include 15
investments” (2004, 20).
 17

Figure 2: Model of the relationship between LPs, VCs, funds and portfolio companies

INVESTORS
INVESTORS
(Limited Partners)
INVESTOR
(Limited Partners)
(Limited Partner)

Fee

Capital
Returns
investment

Management

VC FIRM
VC FUND
(General Partners)

Portion of returns

Capital
Equity
investment
Consulting

INVESTORS
INVESTORS Financial flows
(Limited Partners)
(Limited COMPANY
PORTFOLIO Partners) (capital and equity)
Service flows

VCs’ investments are illiquid because VCs receive equity in exchange for their financial and
strategic support, and this equity only becomes interesting once the ventures have grown to a
successful mature company. Only when an investment is exited (the ideal case being through
an Initial Public Offering (IPO) or a trade sale) can a VC be sure to have produced the returns
expected by the LPs. Despite having passed the in-depth due diligence process (the process of
evaluating an investment candidate), very few of the PCs succeed and many either barely
return the initial investment or must be liquidated.32 The success of a fund therefore relies on
a few ventures performing exceptionally well and producing enormous returns.33

The relationship between PC and VC is particularly interesting because on the one hand, the
VC’s earnings depend on the performance of the fund, i.e., on the performance of each PC but

32
Using VentureXpert data on 5,000 venture capital investments by US-based VC firms, Weidig and Mathonet
(2004) show that the probability of a loss is equal to 42% (the probability of a total loss being as high as 30%)
and that the average multiple is equal to 6.2 (median multiple 1.5, standard deviation 53.8).
33
Schmidt (2004, 35) observes maximal internal rates of return of up to almost 3,000% on venture capital
investments between 1990 and 2002 (Table 4.1).
18

on the other hand, the PC founders are really the ones managing the company. In other words,
the VC firm delegates the responsibility to produce returns to the managers but cannot fully
observe their behavior. Therefore, this is a classical principal-agent setting, in which the VC is
the principal and the manager of the PC is the agent.

Agency theory is a common perspective to analyze the VC-entrepreneur relationship.34 The

simple agency set-up applies to the VC investor (the principal) and to the entrepreneur (the
agent).35 Indeed, this is a situation of separation of ownership and control because the VC
owns a part of the company, but the entrepreneur has the control over his own action, without
permanent supervision of the VC. The entrepreneur is responsible for the fructification of the
VC’s investment, but potential opportunism cannot be ruled out because the VC-entrepreneur
relationship is characterized by important information asymmetries (Wright / Robbie 1998;
Cumming / Dai 2008): the VC cannot fully observe the entrepreneur’s behavior and there is
little to no public information about the invested company.

It follows that the characteristic agency problems of moral hazard and adverse selection are
likely to arise. In the VC context, adverse selection refers to the entrepreneur’s
misrepresentation of his own skills or of his start-up’s real value (Eisenhardt 1989) and to the
investor’s inability to properly assess these prior to investment (Wright / Robbie 1998). For
instance, the founder may tend to oversell the venture’s merits and viability and withhold

34
While agency theory has an intuitive appeal for VC research, its limits must also be kept in mind. First, agency
theory relies on the assumption that principals and agents are solely motivated by economic self-interest.
While it is the venture capital firm’s primary role to maximize fund returns on behalf of the LPs, entrepreneurs
may not only act in order to maximize their own welfare, but also act in pursuit of other motivations, such as
the need for achievement (Shane et al. 2003). Second, agency theory does not leave room for concepts of trust
and fairness. In pre-investment settings, it is true that trust may not be established between VCs and PCs.
However, since the relationship between VCs and PCs typically extends over several years – sometimes over
10 years (Meyer / Shao 1995) – it is likely that trust and fairness will gain importance in defining the VC-PC
relationship once the investment was made (Sapienza et al. 1996; Arthurs / Busenitz 2003). Third, principal
and agent’s goals are not likely to diverge greatly, because both VC and entrepreneur aim for success of the
venture and while the VC can diversify risk using a portfolio strategy, failure is likely to have more serious
consequences for the entrepreneur (e.g., unemployment, possible loss of private assets). Of course, it is
possible that VC and entrepreneur may have a different preferred strategy to achieve success. For example,
Mäkelä and Maula (2005) highlight the case of discrepancies in preferences for the venture’s international
expansion.
35
The principal-agent model could also be applied to the LPs as principals and to the VCs as agents (Wright /
Robbie 1998) because by investing in a VC fund, the LP basically delegates the investment decisions to the
VC. Thus, there is separation of ownership (the money in the VC fund is essentially the LPs’ money) and
control (the VC is responsible for producing the expected returns by investing smartly in portfolio companies),
while the LP cannot fully control the VC’s behavior. It is therefore not surprising that the legal contracts
between LPs and VCs are very complex and negotiated upfront of the LP’s investment into a fund. Also, there
is a post-control mechanism incentivizing VC firms to invest according to the LPs’ interest because
subsequent LP investments in the VC firm’s next fund are likely to be cut off if the previous VC fund did not
perform up to the LP’s expectations. In this section, I do not elaborate further on the LP-VC agency
relationship but rather focus on the VC-PC relationship.
 19

important information (Sahlman 1990), or present a biased perspective of his own ability to
manage the venture successfully. Once having received VC funding, the entrepreneur may
also withhold or modify information crucial to the VC’s assessment of the actual development
of the company. Second, moral hazard refers to the lack of effort on the entrepreneur’s part:
for example, the entrepreneur may use VC money to maximize his own utility rather than
adding value to the company. For instance, he could take on risky investments with the VC’s
money without exercising the same caution as if it were his own money.36

Both positivist and normative approaches have been used in the VC context. The positivist
literature describes how the problems linked to the entrepreneur’s self-serving behavior can
be solved and the normative literature makes recommendations as to what type of contract
should be used to maximize the investor’s welfare in a given situation.

2.2.2.2 Mechanisms used by VC firms to deal with agency problems with portfolio
companies

Resolving agency problems is critical to the success of VC investments (Cumming / Johan

2008) and VCs are considered “among the most sophisticated of financial intermediaries at
mitigating informational asymmetries and agency costs” (Cumming 2006, 1084). In the next
section, I present the concrete mechanisms used by VC firms to cope with the risks linked to
their agency relationship with the portfolio companies (adverse selection and moral hazard).

Mechanisms against adverse selection

VC firms already seek to decrease information asymmetries upfront of their investment in

order to reach a better negotiation position when forming the investment contract with the
entrepreneur (Van Osnabrugge 2000). To do so, classical agency theory recommends
thorough pre-investment screening and an intense due diligence process.

Pre-investment screening is commonly facilitated by the use of referrals as a signal that an

investment opportunity is worth investigating (Tyebjee / Bruno 1984). Indeed, while VCs
receive many investment proposals without introduction from a business partner for example,
they rarely invest in such companies (Fried / Hisrich 1994). Instead, they tend to focus their
work on opportunities recommended by other VC firms they are used to working with or on

36
Contrary to other types of funding such as bank loans, the company does not have to pay back the VC firm’s
investment, because the VC firm receives equity for their investment.
20

projects introduced by other contacts, such as bankers, managers of existing or other portfolio
companies.

Applicants who pass the first pre-investment screening then undergo an intense due diligence
process (Weidig / Mathonet 2004). The due diligence process involves a thorough background
check of investment projects by the VC firm and aims at reducing a priori information
asymmetries: “The VC investment decision-making process is designed to reduce the risk of
adverse selection” (Fried / Hisrich 1994, 29). The difficulty of assessing the value of an
investment opportunity is two-fold for VC firms. First, the companies can be in an early stage
of development, possibly without an established market or a real proof of concept, so that VC
firms must invest in ideas and teams rather than in mature companies (Manigart et al. 1997).
Second, the transactions are private, so that there is no public information available about the
companies requesting financing. To better understand how VCs cope with the serious adverse
selection risks, Fried and Hisrich (1994) interviewed 18 VCs located in the US and developed
a model of VC investment decision making. They find that while most proposals are rejected
with minimal time investment, it takes over three months on average for a proposal to
successfully go through the due diligence process before being funded. During this period of
extensive scrutiny and information gathering, the VC’s goal is to validate the information
contained in the business plan and to determine the value of the proposal on three dimensions:
quality of concept, ability of the management team and value of prospective returns.37 They
do so in many ways: for instance, VCs interview current or potential customers, competitors
and industry experts, they lead comprehensive market studies and they meet with the team
and former business contacts of the founders.38

Finally, once the VC’s interest is confirmed, VCs use outcome-based contracting by carefully
structuring the investment deals in a legal contract (Smith 1998) in order to enforce the
alignment of the entrepreneur’s interests with the VC’s preferences. These contracts specify
the restrictions on the rights of the agent, define performance criteria and set the payoff
functions (Van Osnabrugge 2000; Pruthi et al. 2003). Fried and Hisrich argue that these

37
These dimensions of investment evaluation are confirmed by many authors, including Muzyka et al. (1996)
and Manigart et al. (1997).
38
Interestingly, Manigart et al. (1997) find in their survey of 136 senior investment managers in the UK, France,
Netherlands and Belgium, that the weight of the different sources of information differs between European
VCs. For example, they show that French VCs place more emphasis on the curriculum vitae of the
management team than their counterparts, and that the due diligence process is likely to be less costly and
time-intensive in the UK because British VCs use less reports by independent consultants than French, Dutch
and Belgian VCs do.
 21

contracts even act as a pre-investment sorting mechanism because “by allocating much of the
risk of failure to the entrepreneur, [they discourage] entrepreneurs who are not confident of
their venture’s success” (1994, 35). In other words, contracting on the outcome of the
investment helps sorting out entrepreneurs that may misrepresent their own capabilities or the
ability of their venture to become successful.

The investment is only realized if all these steps are completed successfully, which limits the
agency risks inherent to the VC-entrepreneur relationship a priori.

Mechanisms against moral hazard

Since contracts are always incomplete and since writing a good contract is costly, the ex-post
allocation of control is also important for VCs.

First, VCs typically include a clause in the formal investment contracts giving them the power
to replace the company’s management team if it fails to reach the milestones agreed upon.
This acts as a powerful deterrent against the potential opportunistic behavior of the
entrepreneur (Sorenson / Stuart 2001).

Second, VCs use a form of outcome contracting by staging their investments (Sahlman 1990;
Jeng / Wells 2000). VCs normally do not provide the full investment sum to the start-up
upfront but define strict milestones and make subsequent money injection subject to the
venture’s performance (Wright / Robbie 1998; Cumming 2006; Gompers / Lerner 2006). This
way, moral hazard is reduced because entrepreneurs have a clear incentive not to engage in
self-serving behavior and to reach the desired milestones.

The third mechanism used by VCs to protect themselves against moral hazard is to discover
the behavior of the managers in their portfolio companies by providing intensive oversight
and active monitoring (Wright / Robbie 1998; Sorenson / Stuart 2001; Cumming 2006).
Lerner (1995, 301) points out that VCs “need to provide monitoring (…) to limit the
opportunistic behavior” of private firms. VCs spend over half of their time monitoring the
PCs they are personally responsible for (Gorman / Sahlman 1989). This involvement can take
several forms. Lerner (1995) shows that VCs control their portfolio companies via a seat on
the PCs’ board of directors: he demonstrates that VCs’ involvement as directors intensify
when the risks of deviation from welfare maximization are higher. They study the presence of
VCs on the board of directors of 271 biotechnology ventures in the US between 1978 and
1989 and show that the number of seats reserved to VCs systematically increases when the
CEO is replaced, as this often signals that the company is facing difficulties, while changes in
22

the number of seats allocated to other third parties are not significant. In another study, Barry
et al. (1990) analyze 433 IPOs by VC-backed firms over ten years (1978-1987) and find that
VCs hold one third of the board seats on the companies going public. By sitting on the board
of their portfolio companies, VCs can better control and channel the strategic development of
the start-ups (Wright / Robbie 1998). Another way to monitor portfolio companies is to
maintain regular contact with them (Pruthi et al. 2003). For instance, Gorman and Sahlman
(1989) found that VCs spend an average of 80 hours per year on-site with each of their
portfolio companies and that lead VCs visit their portfolio companies on average nineteen
times per year. They also observe that early-stage ventures receive more attention than later-
stage ventures (two hours per week on average). In a more recent study, Fritsch and Schilder
(2008) find that independent VCs have 1.5 face-to-face contacts with their ventures per month
and 8 contacts via telecommunication per month. Finally, VCs can take a more interventionist
role by consulting their invested portfolio companies on key decisions (Wright / Robbie
1998). In particular, VCs typically provide their support with top management recruitment,
introduction to potential customers and suppliers (Lerner 1995; Mäkelä / Maula 2005) and
strategic planning as a whole. Another important role of the VC is to help their PCs secure
subsequent funding (Gorman / Sahlman 1989). By assisting ventures with important decision-
making, VCs can make sure that the company’s behavior serves the VC’s ultimate goal of
producing extraordinary returns. Barry et al. (1990) show that the monitoring and guidance
provided by VCs is an essential part of their role and the quality of their support has a positive
influence on the valuation of the companies going public. Similarly, Jääskeläinen, Maula and
Seppä (2006) find that VC funds are more successful39 when more attention40 is allocated to
the portfolio companies. MacMillan, Kulow and Khoylian (1988) are more specific with
regards to what type of attention is more beneficial to the success of individual investments:
they find that monitoring operations and developing professional support have a significant
positive impact on the venture’s performance41, while more hands-on management such as

39
They measure success as the number of IPOs, controlling for the number of investments and make a
robustness check by considering all positive exits (IPOs, mergers, acquisitions and buyouts), obtaining similar
results.
40
They measure allocation of attention as the size of portfolio (number of companies in the portfolio) relative to
the number of partners in each firm.
41
They use four measurements of performance: sales, market share, profits and ROI.
 23

searching for candidates of management team and soliciting customers and distributors has a
significant negative impact on all dimensions of venture performance.42

2.2.3 Effects of distance on the VC-entrepreneur agency relationship

Based on the theoretical and empirical findings exposed in the previous section, it appears that
VCs must interact with the entrepreneurs to reduce agency risks and to improve the success
chances of the portfolio companies. For this given research project, it is now important to
understand how distance may affect the VC-entrepreneur relationship. In the next section, I
present a literature review of studies providing first elements of answers in understanding how
the intensity and the nature of agency problems may vary with the distance between the
investor and the invested company, how the mechanisms employed by VC firms are affected
by distance and what the expected effect on investment performance could be.

2.2.3.1 Effects of distance on adverse selection and corresponding coping mechanisms

VC firms cope with adverse selection by attempting to reduce information asymmetries

between themselves and the entrepreneurs seeking funding. Prior to investment, they do so by
collecting and analyzing relevant information to check the start-up’s business plan in the
process of due diligence. Intuitively, information gathering and appraisal may be more costly
when a VC screens distant companies (Johanson / Vahlne 1978; Cumming / Dai 2008).43 For
instance, VCs may not know the distant company’s market as well as they know their own,
which could result in erroneous analyses of the company’s true business potential. This
argument remains valid after the investment was made, for example, when the VC evaluates
the possibility of participating in the company’s next financing round. In addition to the
increased cost of obtaining information for the due diligence process and the monitoring
phase, distance may also mean a difference in legal environment and tax system, cultural
context and business habits between the investor and the entrepreneurs, leading to higher
agency costs in international investments in comparison to domestic VC investments.
Applying these arguments to US-based entrepreneurs seeking funding from VC firms in
emerging Asian economies, Meyer and Shao (1995) find that the associated agency costs will
be higher than if the US PCs would receive funding from domestic VC firms.

42
This result should not be interpreted as a sign of causality. It is indeed possible that VCs must involve
themselves with hands-on management tasks especially when the venture is performing badly.
43
Similarly, foreign firms are “at an information disadvantage” in the evaluation of quality of possible targets of
foreign direct investment (Shan / Song 1997).
24

Another way to reduce information asymmetries is the reliance on referrals, for instance by
other venture capital firms. Since larger VC firms (VC firms with more capital under
management) have historically developed stronger and more abundant links with other VC
firms, they are more likely to receive a wider set of referrals than smaller VC firms. Studying
US-based VC firms’ geographic scope preferences, Gupta and Sapienza (1992) show that
larger VC firms prefer investments within a broader geographic scope than smaller VC firms.
Building on this study, Hall and Tu (2003) test Gupta and Sapienza’s hypotheses using a
sample of UK-based VC firms and confirm that the willingness of a VC firm to invest
overseas is indeed significantly positively related to the size of its investment funds. They
argue that this may not only be due to the extent of networking but also to economies of scale
reducing the costs of search for information.44 From a social topography perspective,
investors who built a central position in a syndication network have access to more and better
information about distant investment opportunities. This means that first, they are in a better
position to become aware of distant investment opportunities and second, that they can better
assess the quality of such potential deals, so that in the end, they are more likely to invest in
distant ventures than isolated VC firms (Sorenson / Stuart 2001).

2.2.3.2 Effects of distance on moral hazard and corresponding coping mechanisms

VCs’ oversight of new firms creates substantial costs linked to the frequent visits and
involvement (Lerner 1995; Balcarcel 2004) and telecommunication is not a substitute to face-
to-face interaction (Fritsch / Schilder 2008). As a consequence, monitoring distant
investments may be more difficult and more costly than monitoring local investments
(Cumming / Dai 2008), which in turn could translate into less frequent interaction and less
value added by the VC.

First, the cost of monitoring (time and expense) is expected to go up with the distance
between the investor and the entrepreneur (Tyebjee / Bruno 1984; Pruthi et al. 2003; Balcarcel
2004). Despite developments in communication and transportation techniques, greater
distances are still associated with higher financial monitoring costs. Time spent in
transportation from the VC’s office to the PC’s office is time taken away from monitoring
another company (Fritsch / Schilder 2008), so that “geographic proximity reduces the time
costs of monitoring” (Sorenson / Stuart 2001, 1555). The head of a Munich-based VC firm I
interviewed also confirmed this in practice: “We will never NOT do an investment simply

44
Also see section 2.1.3.2.
 25

because it is on the wrong side of the Rheine. But if we have two identical investment
opportunities, one in Munich and one in Lyon? Yes, we would rather do the one in Munich”.
Similarly, VCs in the Silicon Valley have traditionally sought investment targets within a one-
hour drive and a one-day trip (Zook 2002). If geographic distance translates into a difference
in the language spoken by VC and PC manager, it will be more difficult and more costly for
VCs to communicate and effectively monitor distant ventures (Sapienza et al. 1996). Another
reason for higher monitoring costs could be the cultural distance between the two parties. For
instance, Meyer and Shao (1995) underline that American business persons prefer formal
written contracts while Asians traditionally rely on personal relationships. This could be a
major difficulty in the VC-entrepreneur setting since the contract plays an important role in
curbing moral hazard risks. Similarly, Pruthi, Wright and Lockett (2003) show that American
VC firms investing in Indian companies are more likely to rely on board membership to
exercise control over the distant invested venture, whereas their local counterparts are more
likely to rely on informal monitoring mechanisms.

Second, distance may lead to lower monitoring intensity. Sorenson and Stuart (2001) stress
that the sociology literature expects that interpersonal interaction declines with an increase in
geographic distance and Gupta and Sapienza hypothesize that “the extent of venture capital
firms’ monitoring and involvement is likely to be inversely proportional to the geographic
distance between them and the focal ventures” (1992, 351). Sapienza (1992) empirically
demonstrates that this is the case by surveying 51 US-based ventures and their lead investor.
In their sample, the distance between lead VC investor and the venture varies from 5 minutes
to 10 hours of travel time, with a mean of 2 hours (standard deviation of 2 hours). They show
that geographic distance has a weak but significant negative effect on the frequency and
openness of the interactions between VCs and their portfolio companies, suggesting that
distance is an impediment to timely and effective information exchange. In the follow-up
study (Sapienza et al. 1996) replicating the survey in Europe, specifically in the UK, France
and Netherlands, the authors also find weak but significant negative effect of distance on the
amount of face-to-face interaction between VCs and their portfolio companies. This statement
was verified for VCs in Germany, Austria and Switzerland, where VCs acting regionally
spend significantly more time supporting current investments vs. selecting new investments
than VC firms acting internationally (Jungwirth / Moog 2004). Applied to the VC’s
monitoring role as a board member, Lerner (1995, 302) finds in his study of biotechnology
firms having received venture capital between 1978 and 1989 that VCs “with offices within 5
26

miles of the [PC’s] headquarters are twice as likely to be board members as those more than
500 miles distant”, which implies that VCs located at a distance from their portfolio
companies are less likely to be involved in the strategic overview of their ventures than close
VCs. Using a field study in India comparing domestic and foreign VCs, Pruthi et al. (2003)
show that foreign VCs are significantly less involved in their Indian PCs’ operational business
than Indian VCs. This finding is confirmed by Jungwirth and Moog (2004) in the German-
speaking area (Austria, Germany, Switzerland).

Thirdly, the value of the monitoring and advisory services may be lessened by geographic
distance (Sorenson / Stuart 2001). For instance, the VC may not be as familiar with the distant
market as with his local market (Pruthi et al. 2003).

Investment stage seems to be a moderator of the effect of distance on the VC’s preferences in
geographic scope of investments, because younger start-ups are expected to require more
monitoring than mature companies (Sapienza et al. 1996). In their study of 169 US-based VC
firms and their preferences regarding industry diversity and geographic scope, Gupta and
Sapienza (1992) find significant evidence that VC firms investing in early stage companies
are more likely to prefer investments within a narrower geographic scope than would VC
firms investing in later stages, a result Sapienza (1992) also obtains in another survey by
finding a significant positive correlation between geographic distance between VC and
portfolio company and the development stage of the venture. This result was confirmed more
recently by Hall and Tu (2003) for UK-based VC firms. On the opposite, Fritsch and Schilder
(2008) do not find such evidence in their analysis of German VCs.

Another moderator of the effect of distance seems to be the portfolio company’s industry.
Jungwirth and Moog (2004) analyze the selection and support in VC financing using 103 VCs
in the German-speaking countries (Germany, Austria, Switzerland). They find that VC firms
with a stronger local focus are significantly less likely to invest in high-tech ideas than VC
firms with an international scope of investment. Since they also argue that low-tech
investments are associated with “hands-on” support while high-tech investments are
associated with more “hands-off” support, it seems logical that the former would require
spatial proximity while the latter would not necessarily. They also find that generalist VC
firms tend to invest more in low-tech projects, while specialized VC firms tend to invest more
in high-tech projects, so that one would expect specialized VC firms to be more open to
investing at a distance.
 27

VCs have developed dedicated mechanisms to minimize the negative effects of distance
(increased moral hazard and higher costs associated with the regular coping mechanisms).
One way VCs bridge the distance with their distant PCs is by using syndication (Sorenson /
Stuart 2001; Fritsch / Schilder 2006). In a syndicated deal, a number of VC investors co-
invest in a promising company, thus sharing their resources, the investment risk and the
upward potential (Jääskeläinen et al. 2006; Meuleman / Wright 2007; Fritsch / Schilder 2008).
In the case of distant investments, a VC firm may use syndication by partnering with a local
investor to invest in and manage the foreign company (Tyebjee / Bruno 1984). For example,
Fritsch and Schilder (2008) investigate the share of investments made at a distance (over
100 km or abroad) by Germany-based VCs and find a significant positive correlation with the
share of syndication partners of these VC firms also located at a distance (over 100 km or
abroad).45 They infer that the investor close to the venture is likely to be the lead investor,
hence the investor spending the most time with the PC (Gorman / Sahlman 1989), while co-
investors are less involved in the monitoring activities (Fritsch / Schilder 2008). Therefore,
spatial proximity may be more relevant for the lead investor than for co-investors in the
syndicated deal. Citing evidence by the European Private Equity and Venture Capital
Association (EVCA), Meuleman and Wright (2007) note that cross-border syndication is a
widespread phenomenon in Europe. In their study of syndication patterns, the authors
demonstrate that firm size (in terms of human resource base) and firm experience with the
host country increases the chances of investing on a sole basis. They also show that VC firms
use cross-border syndication to gain experience about a foreign market, which then allows
them to invest without a syndication partner. Another solution to bridge the distance between
distant VC firms and PC investments is simply for the VC firm to establish a local presence.
For example, if the firm will be seeking to invest repeatedly in a given target location, it may
be reasonable to establish a local office responsible for generating and screening deal flow
and possibly even manage the investments (Meyer / Shao 1995). Indeed, Hall and Tu (2003)
find that younger UK-based VCs and VCs operating from more offices are more willing to
invest overseas.

2.2.3.3 Implications for investment performance

Since the amount of attention allocated by VCs to PCs positively impacts their performance
(Jääskeläinen et al. 2006), one consequence of larger distances between VCs and PCs may be

45
Cumming and Dai (2008) report a similar result in their longitudinal study of US-based VC firms.
28

that the quality of the usual VC value-added support decreases, leading to lower
performances, thus harming investment performance. However, empirical evidence for this
hypothesis is contradictory.

On the one hand, several authors demonstrate weak negative effects of distance on investment
performance. For instance, Balcarcel (2004) compares the likelihood of success of
investments by US-based VCs either in domestic companies or in foreign companies and
finds weak evidence that domestic investments are more likely to succeed46. Cumming and
Johan (2008) present similar (albeit weak) evidence that Canadian entrepreneurial firms in the
same province as their investors are more likely to go public. Cumming and Dai (2008)
analyze VC investments by US-based VC firms between 1980 and 2000 and the effect of
distance on the investment performance using three measures of performance: the likelihood
of a successful exit (occurrence of an IPO or of a trade sale), the timing of the exit and the
value growth between investment and exit. They find a significant negative effect of distance
between the VC firm and the PC on the likelihood of an IPO or merger or acquisition, weak
evidence of a slower time to exit and no evidence of an effect of distance on value growth.
Giot and Schwienbacher (2007) show a significant effect of proximity on speeding up exit
through a trade sale, a result confirmed by Cumming and Dai (2008) for exits via IPO and
merger and acquisition but not for value growth between investment and exit. However, an
alternative explanation not directly linked to an effect of distance may be provided by Butler
and Goktan (2008): the authors show that inexperienced VC firms are likely to invest in
younger and smaller companies located geographically closer to them than other VC firms.
But since younger VC firms are subject to the phenomenon of grandstanding, by which they
force invested companies into faster (but lower priced) exits through IPOs in order to achieve
reputation, it is possible that distance is indirectly negatively linked to the occurrence and
timing of exits per IPOs.

On the other hand, while they do find that distant investments receive less face-to-face
support than close investments, Sapienza, Manigart and Vermeier (1996) find no relation with
the value added by VC firms in their pooled European sample (UK, France and Netherlands).
Considering VCs located in Germany and in the US, Ortgiese (2007) also does not find
evidence that VCs located in the same nation as their PC add significantly more value to their
investments than distant VCs.

46
She defines success as the occurrence of an IPO or of a trade sale.
 29

2.3 Summary, research gaps and contribution of the thesis

Based on a review of related literature, this chapter investigated the possible reasons behind a
VC firm’s decision to invest across borders. First, VCs may seek to increase their
performance and/or decrease their investment risks by pursuing a portfolio geographic
diversification strategy. Following the literature review, later-stage and syndicated
investments are expected to be more favorable to distant investments than early-stage and solo
investments, and VC firms managing larger funds are likely to invest more at a distance than
firms managing smaller funds. Since VCs systematically overvalue domestic investments,
they seem to be subject to a strong home bias, making them more likely to prefer investments
in geographically close markets over investments in distant markets. Second, VCs may invest
across borders in order to grow. International investments may be a response to unsatisfactory
conditions in the VC’s home market (e.g., lack of investment opportunities) or simply an ad-
hoc decision to pursue an interesting but unsolicited investment request from a foreign
company. Third, international investments may be part of a business strategy seeking to
exploit the VC firm’s specific advantages (e.g., industry expertise) or to develop or gain new
capabilities through a learning process, realizing economies of scale by repeatedly investing
internationally. In addition, international investment activity may be a signal used by VC
firms to increase their market power over limited partners, partner VC firms and current and
future portfolio companies. Fourth, VC firms may be interested in investing in a particular
target location to benefit from its specific advantages. In particular, it is likely that VCs are
attracted to regions showing a high level of innovatory and entrepreneurial dynamism with a
large amount of existing dynamism.

Although there are several ways a VC firm could benefit from investing across borders, it is
important to consider the costs of investing at a distance too. VC firms face the typical agency
problems of adverse selection (the problem of identifying the right candidates for a fruitful
investment) and of moral hazard (the problem that the entrepreneurs may tend towards
pursuing personal gain at the cost of the VC’s return expectations) with their investee
companies. These agency costs are likely to be aggravated by physical and cultural distance
between the VC and PC, which makes the option to invest at a distance less attractive than
local investments for VC firms. However, the additional costs associated with investments at
a distance may differ among investments (e.g., between industries or investment stages) and
VCs have developed dedicated mechanisms to effectively overcome this distance.
30

On the one hand, there is abundant literature about firms’ internationalization process and
many studies analyze the agency relationship between VC firm and portfolio companies as
well as the mechanisms used by VC firms to deal with agency problems. On the other hand,
and despite evidence that VC cross-border investments have been gaining in importance over
the past years, still little is known about how and why VC firms internationalize by investing
across borders or how the VC-PC relationship is affected by distance.

First, while there are both costs and benefits attached to cross-border investments, it is not
clear yet which conditions (internal or external to the VC firm) cause one aspect to outweigh
the other. In this thesis, I consider both VC firm characteristics and environmental conditions
to understand in which contexts the benefits of cross-border investments outweigh the costs of
investing at a distance. For example, I analyze VCs’ sensitivity to market conditions as a
possible factor influencing the benefits the firm gains from investing across borders, as well
as the impact of firm experience on the potential reduction of the costs associated with
investing at a distance.

Second, most studies are restrictive in their geographical scope of consideration. Most studies
only focus on US-based VC firms (e.g., Tyebjee / Bruno 1984; Gupta / Sapienza 1992;
Sapienza 1992; Meyer / Shao 1995; Zook 2002). Others use samples from selected European
nations such as UK, France, Netherlands, Germany, Austria or Switzerland (e.g., Sapienza et
al. 1996; Hall / Tu 2003; Jungwirth / Moog 2004; Meuleman / Wright 2007) or from the rest
of the world (India in particular) (e.g., Pruthi et al. 2003). Unfortunately, this restriction
prevents the comparison between local preferences and hinders the analysis of environmental
factors. Furthermore, since studies with different geographic focus rarely employ the same
design, it is virtually impossible to compare their results. By taking a global view of the VC
industry, this thesis enables important distinctions between nations but allows for a direct
comparison between samples. In addition, the use of a sample covering 20 years of VC
investments allows for a better understanding of how external factors have impacted
internationalization decisions of VC firms over time.

Third, existing studies often do not distinguish between venture capital and private equity, or
they focus on later stage investments (e.g., Diller / Kaserer 2005; 2006; Tykvova / Schertler
2006; Meuleman / Wright 2007). Although findings from PE research do deliver elements of
answers and inspiration for VC research, it makes sense to differentiate between the two in
order to deliver a more realistic study of the mechanisms specific to venture capital. For
example, VC firms focus on nascent or young companies and seek to build these companies
 31

up, whereas PE firms focus on companies that are already mature and help them with
financial structuring. Therefore, the challenges facing each investee organization are likely to
differ drastically, necessarily impacting the investor firm’s international strategy. VC and PE
firms also differ in their structure and organization: PE funds are much larger than VC funds,
and so is the typical size of their investments. Since larger investments are at stake, it is
possible that VC and PE firms differ in their investment strategies. Furthermore (and although
the underlying mechanisms are comparable), the sourcing and evaluation of investment
opportunities, the management of portfolio companies and the associated challenges differ
between start-ups and established companies, so that the strategies employed by VC firms and
PE firms cannot be directly compared. As a consequence, it is to be expected that the investor
firms’ attitude towards investing at a distance may differ significantly between VC and PE
firms. It follows that investment behavior analyses should consider PE firms and VC firms
separately. By focusing the analysis of VC firms and their investment strategies, this thesis
contributes in understanding the decision mechanisms of important players in social and
economic growth, innovation and value creation.

Because the mechanisms at play in the decision to invest at a distance are not well understood
and existing literature only provides limited insight, it is important to investigate this matter
more in detail. In the following chapter, I build on the present chapter and on exploratory
expert interviews to identify the likely determinants of VC’s investment scope and of their
choice of investment target location.
32

3 Hypotheses

3.1 Hypotheses development process

Several complementary approaches were used to develop the hypotheses, as summarized
graphically in Figure 3. First, the hypotheses rely on the theory foundations and findings from
the literature review presented in Chapter 2. Second, the hypotheses were validated and
completed based on expert interviews led with venture capitalists around the world.47 The
interviewed VCs were selected in order to obtain the perspectives of a diverse group of VC
firms in terms of investment strategy, both in terms of investment scope and location of
portfolio companies.

The hypotheses pertain to the factors motivating the VCs’ decision to:

1. invest at a distance, i.e., beyond their national borders (the scope decision, see section
3.2)

2. invest in companies situated in a given nation rather than others (the target location
decision, see section 3.3)

and are grouped in two thematic groups for each research question:

1. the market conditions describing the supply and demand in the relevant markets (see
section 3.2.1 for the applications to the scope decision and section 3.3.1 for the
application to the target location decision), and

2. the relevant VC characteristics and fit with potential target location, i.e., VC fund and
firm-level factors (see section 3.2.2 for the applications to the scope decision and
section 3.3.2 for the application to the target location decision).

A complete overview of the hypotheses developed in this chapter is provided at the end of this
chapter, in Table 1.

47
An overview of the interviewed VCs is available in Appendix 1.

H. Tarrade, Cross-Border Venture Capital Investments, DOI 10.1007/978-3-8349-6939-2_3,

© Gabler Verlag | Springer Fachmedien Wiesbaden 2012
 33

Figure 3: Hypotheses development process

THEORY LITERATURE EXPERT HYPOTHESES

FOUNDATIONS REVIEW INTERVIEWS FORMULATION

Internationalization and Extensive literature Series of one-on-one Hypotheses on the

financial theory applied to
venture capital firms and
funds
Agency theory and its
application to the VC-PC
relationship
review and analysis of interviews with general determinants of VC
empirical work partners at VC firms investment scope and
highlighting the potential around the world target location
benefits and costs of VC
investments at a distance
Iterations

3.2 Determinants of scope decision

3.2.1 Local supply and demand

3.2.1.1 Local supply of investment opportunities48

The shortage of quality deals in the VC’s area may justify the need for a VC to screen
potential deals at a distance. A German venture capitalist argued: “We believe there are only
15 to 20 good deals per year all over Europe and we have to try to capture as many of those as
possible”. By looking at investment opportunities at a distance, VCs can enlarge their
investment horizon and have a better pool of possible deals to choose from. In the quest for
superior returns, VCs have an incentive to receive a large number of investment proposals
(Sorenson / Stuart 2001) because having more investment opportunities to choose from raises
the quality of the better proposals, thus possibly improving fund performance (Jääskeläinen et
al. 2006). There may be enough local deal flow to fulfill the VC’s investment requirements49,
but if not, VCs may need to increase their investment scope accordingly. Fritsch and Schilder
(2008, 2128) confirm this hypothesis in their interviews of German VCs: “They would invest
in promising new companies located nearby if they were available”. On the other hand, VC
investments are mostly national in the US: “The adage that ‘venture capital is a local

48
“Supply” as a synonym for deal flow has already been used in the literature (e.g., Jääskeläinen et al. 2006).
However, some authors use “supply” to designate the availability of venture capital financing and “demand”
for the abundance of good business ideas (e.g., Baygan / Freudenberg 2000; Jeng / Wells 2000; Romain / Van
Pottelsberghe 2004; Gompers / Lerner 2006). In this study, I prefer using the first analogy because I take the
VC perspective and not the entrepreneur perspective.
49
In this section, I assume that the quality distribution of the quality of deals is similar in all areas, making the
quantity of deals the only relevant measurement because all else equal, VCs will prefer a local investment to a
distant investment (this argument is developed further in the section 3.3.2.1).
34

business’ still rings true. US VCs (…) believe there are enough quality deals here to support
their funds” (Deloitte / NVCA 2007, 2). Further, 40% of the US-based VCs (28% of non-US-
based VCs) interviewed by Deloitte (2006) stated that the existence of an adequate deal flow
was the primary reason why they are not expanding their international investment focus. The
larger the local deal flow, the more likely a VC’s investment scope remains local, whereas the
investment scope should be larger when local deal flow decreases. This hypothesis has also
been raised by researchers from the World Bank (Aizenman / Kendall 2008) and academic
researchers (Guler / Guillen 2004):

H1: A venture capital firm’s propensity to invest at a distance decreases when

the local deal flow is more abundant.

3.2.1.2 Local demand for investment opportunities

As much as the local supply of interesting investment targets may have an effect on the
propensity of a VC to invest at a distance, the local demand for quality deals may also be of
importance. On the one hand, one could argue that an area with a higher VC concentration
will automatically attract more financing requests, thus possibly providing sufficient supply
for these VCs, who may choose to limit their scope of investment to such companies. This
would mean that the local VC density is negatively correlated with the VC choice of
investment scope. On the other hand, VC firms may also see each other as competitors.
Sapienza, Manigart and Vermeir argue that the “competition for attractive investments is
heating up as economies become more globalized” (1996, 440). The scarcity of the good local
deals becomes an even stronger problem if more local VCs are competing for them (Hall / Tu
2003). This view has also recently been shared by Cumming and Dai, who observe that
“competition among VCs in the local area decreases their local bias” (Cumming / Dai 2008,
3). Based on interviews with VCs, I follow the hypothesis that VC firms view each other as
competitors. If this is true, then VCs may enlarge their scope of deal screening and their scope
of investment as more local VCs compete for the local deals:

H2: A venture capital firm’s propensity to invest at a distance increases with

the local demand for deals (i.e., with the intensity of competition among
local VC firms).
 35

3.2.2 VC ability and fit

In a given supply and demand environment, not all VC firms behave the same way because
they do not all share the same ability to invest at a distance. It is therefore necessary to
complete the first group of hypotheses with a second group focusing on intrinsic VC factors
and investment characteristics.

3.2.2.1 VC fund characteristics

Fund industry specialization

Some VC funds are generalists, others are semi-focused, with specific funds dedicated to
investments in a specific industry (e.g., a VC firm managing concurrently one life sciences
fund and one high-tech fund) and yet others may have an expertise limited to one particular
sector or industry (Jones / Rhodes-Kropf 2003). Restricting the industry in which VCs can
invest logically limits the deal flow available locally (providing the fund size remains
unchanged), which could force the VC to enlarge the geographical scope of its investments to
prevent a decrease in the quality of the portfolio companies it selects. Industry specialization
may also constitute a form of firm-specific advantage for VC firms, enabling them to compete
with distant VC firms for PCs.50 I therefore postulate that industry specialization “shrinks the
world” for VCs:

H3a: A venture capital firm’s propensity to invest at a distance is larger for

funds with an industry specialization than for funds without industry
specialization.

It is possible that the industry of specialization makes a difference in the scope of investment.
VCs may have to interact more frequently with portfolio companies evolving in a faster-paced
industry, i.e., when the environment of the PC changes fast (e.g., fast research and
development cycles, rapidly moving competitive landscape) and speed (e.g., ability to reap
first-mover’s advantages, setting up of efficient logistics) can make a vital difference between
competitors. PCs in fast industries should necessitate more frequent support and control from
their VC. For fast-paced industries, VCs should therefore prefer ventures located close to their
office, whereas distance may not necessarily be seen as an obstacle to invest in industries with
a slow product development cycle. In particular, IT investments usually offer quicker returns

50
See section 2.1.3.2.
36

than life science investments.51 The overhead cost of managing an investment located at a
distance over long periods may be easier to amortize than the cost of managing a remote
investment to be exited within months or just a few years after the initial investment.

H3b: A venture capital firm’s propensity to invest at a distance increases with

typical time to market in the industry the venture capital fund specializes
in.

Fund stage specialization

Most of the VC funds have traditionally specialized in an investment stage (e.g., seed or
expansion). Since agency problems are the most acute in the earlier phases (Hall / Tu 2003),
active involvement in the start-up is the best way to control the investment, CEO-VC
interactions are required more frequently in early-stage ventures (Sapienza / Gupta 1994; Van
Osnabrugge 2000). The establishment of a trust relationship between VC and entrepreneur
and a display of goodwill and good performance by the portfolio company will reduce the
need for monitoring (Sapienza et al. 2000), thus diminishing the costs associated with
behavior discovery over time, when PCs mature to later development stages. Also, the
benefits of geographic portfolio diversification are likely to be more prominent in later
investment stages than in earlier investment stages (Lossen 2007). It is therefore likely that
VCs will choose a smaller scope of investments if they specialize in earlier-stage ventures and
that VCs investing in later-stage ventures will chose larger scopes of investment. The
investment policy employed by Sequoia Capital for instance, a major Silicon Valley-based
VC firm, confirms this view: “For seed and early stages it is helpful if the company is close to
our office in the U.S. since these sorts of companies require very frequent contact. Growth
stage companies are another matter and we are more than happy to invest in whatever places
the founders and managements of these more mature companies have chosen for
themselves.”52

51
For instance, for pharmaceutical ventures working on a new drug, the trial phase alone can take up to seven
years (Marwaha et al. 2007) and the complete development more than a decade (DiMasi et al. 2003; Gompers /
Lerner 2006). An own analysis of first-round investments made between 1995 and 2004 shows that IT
companies went public significantly faster (on average 3.11 years after the first received investment round) than
life science companies (on average 4.21 years after the first received investment round). A t-test revealed that the
difference in time to IPO was significant at the 1% level.
52
http://www.sequoiacap.com/us/venture-capital/ (stand: July 2009)
 37

H4: A venture capital firm’s propensity to invest at a distance is larger for

venture capital funds specialized in later stages of investment than in early
stages.

Fund size

Due to the scarcity of good investment opportunities, VCs must see a lot of business plans to
find appropriate ventures to invest in. For VCs with similar investment size preferences, VCs
managing larger funds have more portfolio companies and must have seen more business
plans than those with smaller funds. In order to generate a higher intensity of deal flow, VCs
with larger fund size are therefore expected to enlarge their radius of search for business plans
in order to reach a sufficient quantity and quality of deals to choose from (Aizenman /
Kendall 2008). As a consequence, more of their portfolio companies are expected to be
located further away from their office. This view was already verified for UK-based VCs
(Hall / Tu 2003). Deloitte (2006) also confirms in its recent survey on cross-border
investment preferences that limited fund size is one of the key reasons why VCs are not
expanding their international investment focus.

H5: A venture capital firm’s propensity to invest at a distance increases with

the size of the managed venture capital fund.

Type of fund

Although most LPs do not impose constraints on the VC’s investment strategy, some do,
which may have a strong influence on the investment scope decision. For instance, some
investors require their money to be invested only in a particular geographical area, directly
affecting the VC’s investment scope. For example, the European Investment Fund was
reported by several of the VCs I interviewed as imposing such restrictions, namely, that they
ask VCs who raise from them to invest within Europe. Similarly, other investors are not yet
ready to see their money invested in high-growth but yet uncertain regions of the world such
as India or China. When LPs impose such restrictions, they possibly force VCs to reduce their
investment scope. Since it is usually not possible to access contracts between VCs and their
investors, an approximation can be made by considering the type of investor in the fund
managed by a VC firm. Following Gupta and Sapienza (1992) as well as Doran and Bannock
(2000), I propose that public investors are more restrictive in their investment scope
requirements than private investors. Indeed, investors such as governments or university
38

programs are likely to be motivated by the support of the local community, hence imposing
restrictions on the location of the investments (Fritsch / Schilder 2008), whereas private
investors are more likely to solely pursue return maximization.53 I check this effect with the
following hypothesis:

H6: A venture capital firm’s propensity to invest at a distance is lower if its

venture capital fund primarily comes from public money than from private
sources.

Fund age

An aspect other authors have not studied are possible differences in a VC firm’s investment
strategy over the lifetime of a fund. Based on my exchanges with VCs, I argue that VCs may
go with safer investments towards the end of a fund’s life than when the fund has just been
raised because they have more time at their disposal to bring risky investments to a
satisfactory level when the fund was recently raised. Assuming that all investments have the
same return potential, investments made at a distance only differ from local investments in
their risk-return ratio on the risk aspect: since investments made at a distance are riskier than
local investments (assuming similar quality), then VCs should become less likely to invest at
a distance towards the end of a fund’s life.

H7: A venture capital firm’s propensity to invest at a distance decreases with

the age of the managed venture capital fund.

3.2.2.2 VC firm characteristics

VC firm size

VC firms with more partners available to manage the portfolio companies should be better
able to afford investing at a distance because the opportunity cost (attention allocated to the
start-ups, travel time, adaptation of contracts...) of helping to manage a distant portfolio
company versus a local company decreases with the number of partners available to manage
the VC firm’s investments. Conversely, lack of partner capacity is one of the primary reasons
why VCs do not wish to expand their international investment focus (Deloitte 2006). I

53
VC general partner interviews generally confirmed this hypothesis.
 39

therefore hypothesize that VC firms with more general partners (for a given fund size) are
more likely to invest at a distance.

H8: A venture capital firm’s propensity to invest at a distance increases with

the firm’s number of general partners.

VC firm experience

Some sources find that younger VC firms tend to invest more at a distance than their more
experienced counterparts (Hall / Tu 2003), arguing that older VC firms may either be more
inclined to conservatism or constitute a significant barrier to the entry of newer VC firms on
the local market, pushing them to seek deals beyond the domestic market. However, this
contradicts the traditional stage model of internationalization (Johanson / Vahlne 1977) which
predicts that VC firms should rather enlarge their scope of investment gradually, learning first
from local investments before taking the risk to invest at a distance. It is true that this model
may have lost some of its explanatory power in the past years with the introduction of the
born-global concept, in which organizations can be international from inception (Oviatt /
McDougall 1994; Johanson / Vahlne 2003).54 But these limitations should not apply to
venture capital firms because it is not possible to establish the network necessary to a
successful VC firm instantly, and even less for distant operations. Interviews led with VC
firms investing across borders also showed that they had indeed increased their number of
cross-border investments over time. If this is true, then VC firms that have been in operation
for longer and VC firms who have more investment experience55 should be more likely to
invest at a distance.

H9: A venture capital firm’s propensity to invest at a distance increases the

longer it has been in operation.

54
This concept has been applied to start-up companies but not to VC firms.
55
VC firm reputation is another interesting possible determinant of investment scope. However, since reputation
has been measured with firm age and size (Gompers / Lerner 2006; Cumming / Dai 2008) and these
determinants are already included in hypotheses, I leave reputation out of the analysis to avoid redundancy.
40

3.2.2.3 VC investment characteristics

Investment syndication

As discussed in section 2.2.3.2, syndication can be used by VC firms to bridge distance

between the VC firm and a distant portfolio company. In this case, a VC firm could invest in
the distant PC together with a partner who is better positioned to monitor the portfolio
company:

H10: A venture capital firm’s propensity to invest at a distance is higher if the

investment is syndicated than if the VC firm is investing on its own.

3.3 Determinants of target location decision

This section formulates hypotheses for the second research question: what are the factors
determining that a VC will invest in a PC located in a certain location rather than another?
Both the market conditions in terms of supply and demand in the alternative target locations
as well as the VC’s own fit with the possible target location are expected to drive this
decision.

3.3.1 Target location characteristics

This hypothesis group includes four aspects: first, the supply of investment opportunities in
the target location, second, the demand for these potential deals (i.e., the VC competitive
landscape), third, the attractiveness of the target location for entrepreneurs and fourth, the
intrinsic attractiveness of the target location for VCs.

3.3.1.1 Supply and demand for investment opportunities

Similarly to the hypotheses introduced in section 3.2.1, the comparison of supply and demand
for investment opportunities in possible target locations should drive the propensity of a VC
to invest in a given target location. Locations with more supply, i.e., more deal flow, should
attract more VC investments. For example, Guler and Guillén (2004) find a significant
positive link between a country’s GDP and the propensity for US VCs to invest in that
country. Practitioner reports confirm this idea: “VCs are making the majority of their foreign
investments in areas with higher quality deal flow” (, 1Deloitte / NVCA 2007). Similarly,
locations with more local demand, i.e., with a stronger presence of local investors competing
for good local investment opportunities, should make it harder for other VCs to enter the
 41

market, thus lowering the probability of VCs to select this target location. I therefore propose
following hypotheses:

H1: A venture capital firm is more likely to invest in a target location

providing more deal flow than in a target location providing less deal flow.

This hypothesis reflects the projected returns of investments made in that area, which is
possibly the main determinant of VC investment as highlighted in the expert interviews. The
argument of the competition for good investment deals remains relevant for the choice of
target location of investment:

H2: A venture capital firm is less likely to invest in a target location with a
relatively higher venture capital firm density than in a target location with
a relatively lower venture capital firm density.

3.3.1.2 Target location attractiveness of exit channels

For VCs, a viable exit mechanism is essential (Cumming / Macintosh 2003). I therefore
expect that a dynamic primary market, demonstrated by the importance of IPOs (Jeng / Wells
2000) and the availability of a secondary market such as New York City’s NASDAQ and
Frankfurt’s Neuer Markt (Baygan / Freudenberg 2000) play an important role in a VC’s
decision to invest in a start-up located in the target location.

H3: A venture capital firm is more likely to invest in a target location with a
dynamic initial public offering market or a secondary market than in a
location without such exit channels.

3.3.2 VC ability and fit

In this second group of hypotheses, the effects of two factors are studied: first, the distance
between VC location and target location and second, the VC’s experience with this target
location.

3.3.2.1 Distance to target location

Geographic distance

Attention allocated to portfolio companies impacts the performance of investments: the more
attention a venture receives, the better its performance (Jääskeläinen et al. 2006). On the other
42

hand, the time spent by a venture capitalist with a venture, be it for monitoring or for added-
value assistance, has a considerable opportunity cost. Moreover, ensuring frequent contact
between the entrepreneur and the general partner managing the investment is not as easy for
start-ups located at a distance than for local investments (Tyebjee / Bruno 1984).56 The
reoccurring cost of interacting with a remote PC may deter VCs from investing in ventures
situated in farther locations because they either would not be able to monitor them properly
(Sapienza / Gupta 1994) or because the cost to manage the company would be too high
(Lerner 1995; Meyer / Shao 1995; Balcarcel 2004). For example, US-based VCs cite travel
time and effort as one of the main impediments to investing in China (Deloitte 2006). Existing
research usually demonstrates a location bias, which can be interpreted as a preference for
investments in closer target locations (Aizenman / Kendall 2008). It is therefore to be
expected that VC firms are more likely to invest in closer than remote target locations.57

H4: A venture capital firm’s propensity to invest in a target location increases

with the geographical proximity of this location with the venture capital
firm’s location.

Using a sample drawn from the 2000 Yearbook of the British Venture Capitalist Association,
Hall and Tu find that half of the UK-based VC firms claiming to be ready to invest in Europe
had at least one office outside of the UK, while 80% of the UK-based VCs ready to invest
beyond Europe had at least one office outside the UK (Hall / Tu 2003). While it is possible
that international office presence may be a result of firm size or firm experience, it is also
possible that VCs open new offices internationally in order to better access deal flow in other
nations (i.e., a subsidiary office is used to bridge the distance between VC and PC). In
particular, VCs that have an office in a certain location should be more likely to invest in
ventures situated in this target location than VCs without a subsidiary in the target location:

H5: A venture capital firm is more likely to invest in a target location where it
has an office than in a target location where it does not have an office.

56
For a thorough discussion of the expected effect of distance on agency problems between VC firms and their
portfolio companies, see section 2.2.3.
57
Distance can be defined and measured in many ways, for instance via a calculation based on the comparison of
the zip codes of the VC firms and the portfolio companies (Lerner 1995; Butler / Goktan 2008) or via a
measure based on the latitude and longitude of their offices (Cumming / Dai 2008). For the exact
implementation of the variable in this study, see section 4.3.1.
 43

Cultural distance

Cultural distance58 has been raised as a source of failure by VC partners I have interviewed. A
stream of psychology research backs this observation, showing that information flows become
more problematic as cultural distance increases (Johanson / Wiedersheim-Paul 1975; Arenius
2005). Since information flow is essential to a successful VC-entrepreneur relationship, it is
expected that the likelihood of a VC to choose an investment in a target location decreases
with the VC’s cultural distance with this target location.

H6: A venture capital firm’s propensity to invest in a target location decreases

with the cultural distance between this location and the venture capital
firm’s location.

Cultural distance can be modeled in many ways but it is not the goal of this project to study
all its facets and apply them to the VC-entrepreneur relationship. Also, culture can translate in
environmental conditions that were already discussed above (e.g., dimensions of market
attractiveness). I therefore focus on the main elements that were not covered yet but were
pointed out as relevant by previous studies or in expert interviews. In particular, I retain the
role of language and legal system heritage (Arenius 2005; Aizenman / Kendall 2008; Sousa /
Bradley 2008). A German VC I interviewed summarizes: “We had two portfolio companies in
Helsinki. It was difficult to manage: far away, the language… It may be why they both went
bankrupt.”

3.3.2.2 VC experience with the target location

An important determinant of target location choice is possibly the VC firm’s experience with
that target location (Meuleman / Wright 2007). There is evidence that when investing in
foreign markets, VCs do not follow the same guidelines as for domestic investments (Wright
et al. 2002). From this finding, it can be concluded that it is costly for a VC to enter a new
market, but that there are likely economies of scale to be realized – a hypothesis that finds
confirmation in other studies (Hall / Tu 2003) – and a possible path dependency (Aizenman /
Kendall 2008): if a VC already made an investment in a venture based in a certain location, it
is more likely to pick an investment in this target location again than in a target location in

58
In this thesis, “cultural distance” is used as a synonymous to “psychic distance” (Reid 1981; Arenius 2005;
Sousa / Bradley 2008; Dikova 2009)
44

which it never made an investment. On the opposite, lack of expertise can be an impediment
to expanding to a new market (Deloitte 2006).

H7: A venture capital firm is more likely to invest in a target location in which
it already made an investment than in a target location in which it has not
made an investment yet.

3.4 Summary of hypotheses

The hypotheses developed in this chapter are summarized in Table 1.

 45

Table 1: Hypotheses overview

Decision Hypothesis # Determinants Expected effect on

group VC’s
propensity to invest at
a distance

Supply and H1 VC firm’s local supply of deals (local deal

Scope -
demand flow)
Supply and H2 VC firm’s local demand for deals (intensity of
Scope +
demand local VC competition)
Ability and fit H3a
Scope VC fund is industry-specialized (yes=1; no=0) +
(fund)
Ability and fit H3b Time to market in the VC fund industry of
Scope +
(fund) specialization (if applicable)
Ability and fit H4 VC fund is specialized in later investment
Scope +
(fund) stage (yes=1; no=0)
Ability and fit H5
Scope VC fund size +
(fund)
Ability and fit H6 VC fund is financed by public investors
Scope -
(fund) (yes=1; no=0)
Ability and fit H7
Scope VC fund age -
(fund)
Ability and fit H8
Scope VC firm size +
(firm)
Ability and fit H9
Scope VC firm investment experience +
(firm)
Ability and fit H10
Scope Investment is syndicated (yes=1; no=0) +
(investment)

Decision Hypothesis Determinants Expected effect on

group VC’s
propensity to invest in
target location

Target Supply and H1 Target location’s supply of deals (deal flow) +

location demand
Target Supply and H2 Target location’s demand for deals (intensity -
location demand of VC competition)
Target Supply and H3 Target location’s attractiveness of exit +
location demand channels
Target Ability and fit H4 Target location’s geographic distance to VC -
location firm
Target Ability and fit H5 VC firm has office presence in target location +
location (yes=1; no=0)
Target Ability and fit H6 Target location’s cultural distance to VC firm -
location
Target Ability and fit H7 VC firm’s amount of investment experience in +
location target location
46

4 Construction and description of the data set and variables

4.1 Sources of venture capital data

A major problem with venture capital research is the difficulty of gaining access to quality
data – giving full literal meaning to the term “private” equity (Schmidt 2004; Weidig /
Mathonet 2004; Lossen 2007): since transactions are usually performed by private
organizations, no information is directly available to the public. Fortunately, a few venture
capital databases exist and are one option to overcome this difficulty. However, these
databases only cover some aspects of the venture capital cycle and may not fit all VC research
topics, so that some researchers prefer relying on proprietary data sets. Since choosing the
right data source is critical but not straightforward, this chapter starts by comparing the
principal data sources before it goes on to discuss the choices made for this study and describe
the corresponding data. The next section introduces the necessary additional variables,
followed by a brief analysis of the VC industry’s industry geography. Finally, the last section
summarizes the chapter and synthesizes the advantages and drawbacks of the data set.

4.1.1 ThomsonONE.com

ThomsonONE.com is the for-profit online database managed and marketed by the

information service companies Thomson Reuters59 for PE practitioners including VCs,
lawyers and consultants. The PE module60 is available as an add-on to ThomsonONE.com
Investment Banking and includes historical data on buyouts, PE funds, firms, executives,
portfolio companies and limited partners around the world dating back to 1969. The
information contained in ThomsonONE.com mostly stems from first-hand information
provided by VCs and PE practitioners, as well as from information collected by Thomson
reporters, government filings and public news releases. PE firms are also surveyed on a
quarterly basis in the US and semi-annually in Europe for cash flow information.

59
Thomson Corporation and Reuters Group PLC merged to form Thomson Reuters in 2008.
60
This module is also known as Venture Economics, VentureONE and VentureXpert.

H. Tarrade, Cross-Border Venture Capital Investments, DOI 10.1007/978-3-8349-6939-2_4,

© Gabler Verlag | Springer Fachmedien Wiesbaden 2012
 47

The ThomsonONE.com database can be queried in order to build reports based on given
search criteria. For example, a query could return the details of all investment rounds that took
place within a given timeframe, provide information about VC firms located in a given area
and focusing on a given industry or list all companies that went public after having received
funding from a New York-based VC. It is also possible to request aggregated financial data
(e.g., the yearly amount of expansion-stage funding provided by Germany-based VCs to non-
Germany based IT ventures over the past 10 years), but cash flow information is not available
at the investment level.

For an observed investment, it is possible to retrieve information on:

x the company receiving funding, including, but not limited to: company name (and
previous company names), location, industry, customer type, main competitors,
business description, date founded, management contact information, board members
contacts and employment history, company current status (e.g., went public,
bankrupt), exit details if applicable,

x the VC firm(s) participating in the investment round, including: firm name, VC type
(e.g., corporate VC), date founded, capital under management, preferred investment
role, total amount invested in company, partners’ contact information, former
employees, board seats history, top co-investors and more,

x the investment round, including: investment stage, estimated round total, round date,

x the fund(s) involved, including: vintage year, fund size, complete investment profile
(geographic, stage, industry and year breakdown) and more,

x the LPs invested in the fund, including: contact information, assets under
management, total amount committed in venture capital, funds invested in and more.

Contrary to other data sources on venture capital, ThomsonONE.com is a worldwide database

covering most of the deals and industry players on all continents. The VentureXpert module
contains information on about 20,000 funds, their management VC firms (and executives),
portfolio companies and limited partners, and is advertized as the source with the most
extensive global coverage of the venture markets.61

61
The database is estimated to contain 88% of all VC funds in terms of capital committed (Lossen 2007, 34). A
manual comparison of the investments observed in the database with the actual investments by selected VC
firms known to the author revealed that the coverage of the database is indeed complete.
48

Interestingly for academic researchers, queries and reports can be downloaded to a Microsoft
Excel sheet. Also, ThomsonONE.com allows complex queries while offering the possibility
to return up to 16 variables at a time. However, the number of results that can be downloaded
and the number of variables that can be queried at once are limited62.

Another advantage of this database is that several researchers have already used it in the past
(e.g., Sorenson / Stuart (2001), Jones / Rhodes-Kropf (2003), Balcarcel (2004), Guler /
Guillén (2004), Kaserer / Diller (2004), Weidig / Mathonet (2004), Kaplan / Schoar (2005),
Phalippou / Gottschalg (2005), Giot / Schwienbacher (2007), Mann / Sager (2007), Sørensen
(2007)), making it convenient to compare, verify and put in perspective new studies with
existing literature.

4.1.2 VentureSource

The second large VC database, VentureSource, is managed and marketed by the Dow Jones &
Company, Inc and is also accessible against a fee. Dow Jones tracks the portfolio companies
mostly via a primary research process that involves direct contact with their senior
management 3 to 4 times a year and confirmation of the data by the corresponding VC firms.
Changes at the venture firms are tracked via direct contact, primarily with the VC firms’ CFO,
on a quarterly basis. The rest of the information (approximately 5%) is obtained via secondary
research.

VentureSource is a tool designed for VC investment professionals, consultants, advisors and

analysts.63 The database focuses on the VC markets in the US, Europe and Israel – China and
Taiwan are available in an extra module.

VentureSource provides information about:

x the companies having received funding, including, but not limited to: contact
information, business description, main competitors, financing rounds with
valuation64, current status, executives and board members history, contacts and
employment history, current number of employees, date founded, previous company
names and service providers (bank, auditor),

62
These variables
63
According to Dow Jones’ own statements, VentureSource is used by “48 of the top 50 VC firms”
(http://fis.dowjones.com/venture-capital-databases.html, June 2009).
64
Valuation information is missing for a third of the portfolio companies (Lossen 2006) and accessing some
valuation details may be subject to additional fees, such as post-money valuation for example.
 49

x the VC firms having participated in investment rounds, including: contact

information, capital under management, other offices, primary contact name, lead
investor position, funds raised, portfolio history, board seats history, investment
preferences,

x the investment rounds, including: round amount, statistics on comparables rounds

and more,

x the funds involved, including: name, date opened and closed, amount and industry
focus.

According to the firm’s own product description, VentureSource includes about 30,000
venture-backed companies and 8,000 private capital firms, thus covering the bulk of the
relevant transactions (approximately 97-98% in the US, 92-94% in Europe and 100% in
Israel).65

The product’s analytics tool is a practical built-in graphic tool to model research results. A
drawback of VentureSource is that only a small set of key variables can be downloaded at
once.

4.1.3 CEPRES

The CEPRES Center of Private Equity Research has been established as a cooperation
between VCM Capital Management GmbH66 and the chair of banking and finance at the
Johann Wolfgang Goethe-University of Frankfurt am Main. CEPRES contains the cash flow
history of investments (gross of management fees and carried interests), allowing precise
return measurements. Although the database contains a large number of variables about
portfolio companies and funds, researchers who have worked with CEPRES have reported
that a large share of the values are missing for some variables, (e.g., Ick 2005; Cumming /
Walz 2007). For example, Ick (2005) reports that the investment stage information is missing
for about half of his sample. This makes the database inappropriate for this study.

65
This information is based on sales material and on the company’s online description of the product
(http://fis.dowjones.com/products/venturesource.html, June 2009).
66
VCM Capital Management GmbH is now known as Sal. Oppenheim Banking Group, a German PE fund-of-
funds.
50

4.1.4 Proprietary data

Some researchers have access to proprietary data sets for example through a cooperation with
a fund of funds investor. For example, Lossen (2007) benefits from a research cooperation of
the Ludwig-Maximilians-Universität München with a European fund-of-funds investor, thus
gaining access to a sample containing 227 PE funds managed by 51 PE firms. Another case is
the work of Ljungqvist and Richardson (2003), who benefit from a special cooperation
agreement with one of the largest American institutional investors. A major advantage of
these two data sets is that they contain all the cash flow records between funds and investors
as well as between funds and portfolio companies, which is relevant for a precise study of
returns and performance.

However, proprietary data sets may not be representative of the population. In Lossen’s
example, the sample only includes firms located in the US or in Europe and is biased towards
larger and older funds. In Ljungqvist and Richardson’s case, the data set is relatively small
and biased towards large buyout funds with VC funds being underrepresented.

4.2 Construction and description of the data set used in this thesis

4.2.1 Choice and construction of the data set

In light of the previous discussion, it appears that the two most comprehensive data sources
for this study are the professional databases run by Thomson Reuters and Dow Jones. In the
following section, I highlight the principal reasons for choosing to use the VentureXpert
module of the ThomsonONE.com database and discuss potential drawbacks of this choice.

The first concern is the focus and content of the databases. While VentureSource database
may be more complete, VentureXpert’s omissions do not lead to a selection bias (Kaplan et
al. 2002). ThomsonONE.com contains more information about investors and funds, while
covering all the necessary aspects of interest at the portfolio company level. VentureSource
has more detailed information about portfolio companies but no information about limited
partners. VentureSource has a better coverage of investment capital details (individual
amounts invested and valuations for example), but this data is not critical to this study.
VentureSource’s coverage is limited geographically (US, Europe and Israel) but
ThomsonONE.com covers worldwide investment activity. Finally, VentureXpert is well
adapted for scholarly research and has been used by several venture capital researchers in the
past. In contrast, VentureSource is cumbersome to use for academic research involving large
 51

data set retrievals with more than a few variables at a time. Based on this comparison, I rely
on ThomsonONE.com’s VentureXpert module as the main data source for the implementation
of this study.

I use the ThomsonONE.com database to download the relevant information about all
observed VC investments made worldwide over the 20 years between January 1st, 1989 and
December 31st, 2008 as well as the information about the corresponding VC firms, VC funds,
PCs and LPs.67 As a result, five data were formed, respectively describing 7,601 VC firms68,
15,820 VC funds69, 50,443 PCs70, 7,906 LPs and 227,441 VC investments.71

I describe these data sets in more detail and introduce the necessary additional variables in the
following paragraphs. To ensure that the data collected from the ThomsonONE.com database
is representative, I compare the descriptive statistics with the information available from
regional venture capital associations72 whenever applicable. In addition, I check the PCs listed
in the sample with the online presence of select VC firms and with interviewed VCs. I
observe a close overlap between the sample characteristics, the information published by
regional VC associations and the results of own firm research and general partners interviews,
indicating that the sample is indeed representative.

67
Downloading this information had to be done by hand. Furthermore, the number of observations as well as the
number of variables that can be downloaded at once was unfortunately limited by the provider, making the
downloading task a lengthy process. I obtained 16 files of VC firm observations, 32 files of VC fund
observations, 22 files of investment observations, 98 files of PC observations and 17 files of LP observations
that I each transformed into comma separated files (.csv format) before I saved them as STATA data sets (.dta
format). I then combined all .dta files of the same type to create one .dta file per observation type (e.g., one
single data set containing all the downloaded information on the VC firms). Since the investment data set
contained additional information about VCs (nation, founding date and current status) and PCs (date of IPO
when applicable) that was not available in the raw downloaded data sets on VCs and PCs, I merged this
additional information from the investment data set into the VC and PC data sets in order to obtain complete
VC and PC data sets (in doing so, I dropped the investment observations for which the investor is either an
individual or an undisclosed firm).
68
7,566 of these VCs made at least one observed investment.
69
15,766 of these funds participated in at least one observed investment over the observation period.
70
48,982 of these PCs received at least one observed investment over the observation period.
71
The structure of the resulting five data sets (available variables describing each observation)
is described in

Appendix 2.
72
In particular, I analyze reports by the National Venture Capital Association (NVCA, US), the European Private
Equity and Venture Capital Association (EVCA, Europe) and the German Private Equity and Venture Capital
Association (BVK, Germany).
52

4.2.2 Description of VC firms in the data set

There are 7,570 VC firms observed in the investment database and information is available on
31 additional VC firms (these additional VCs are not taken into account in the following
description, since they do not appear in the investment data set). 58.9% of these VCs are
located in North America (over 4,200 of which in the US). The rest of the VCs are located for
two thirds in Europe and one third in Asia (see Table 2).73 The top VCs with most
investments over the observation period are located in the US or in the UK.

Table 2: Location of observed VC firms (by continent)

VC firm location (HQ) Frequency Percentage

North America 4,462 58.9%
Europe 2,027 26.8%
Asia 922 12.2%
ROW 156 2.1%
Unknown 3 0.0%
Total 7,570 100%

4.2.3 Description of VC funds in the data set

15,766 funds are observed in the investment database. Information is missing for 92 of these
funds74 and is available for 142 additional funds (i.e., funds for which information was
collected but that did not participate in an investment round over the observation period). As
shown in Table 3, 66.7% of these funds are independent partnerships (i.e., the regular type of
VC fund), 19.5% are VC funds belonging to a bank, 9.9% are corporate VC funds. The rest of
the funds are public funds or funds set up and managed by private individuals doing business
angel investments.

The vast majority of the observed funds were created during the timeframe of the analysis
(1989-2008), with a peak between 1997 and 2000, when it was particularly easy to raise VC
funds (see Figure 4). Each observed fund made on average 14.4 investments (SD=29.8) in 8.8
PCs75 (SD=16.3). The average fund size is 180.9 million USD76 (see Table 4).

73
For a detailed overview of the location of the observed VCs (by nation), see Appendix 6.
74
“Undisclosed Fund” is not considered in the following statistics.
75
This number seems realistic based on the manual verification with a few VC firms. For example, TVM Capital
lists 30 current and past PCs in the technology sector, which were financed with 5 specialized funds (i.e., 6
PCs per fund).
76
This is consistent with the numbers reported by the European Private Equity and Venture Capital Association
(e.g., average size of the new funds raised in 2007 of 151.8 million EUR) (EVCA 2008). For example, TVM
Capital’s latest funds amount to 299 million USD (Life Sciences) and 121 million USD (Information
Technology).
 53

Table 3: Type of observed VC funds

VC fund type Frequency Percentage

Independent 10,515 66.7%
Bank 3,070 19.5%
Corporate 1,558 9.9%
Public 354 2.2%
Individuals 160 1.0%
Unknown 109 0.7%
Total 15,766 100%

Figure 4: Vintage year of observed VC funds

2,500

2,000

1,500

1,000

500

0
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2009
Vintage year

N=15,674
Funds for which the vintage year is unknown are not shown in this graph (N=92)
54

Table 4: Summary statistics of observed VC funds

Variable Mean SD Minimum Maximum

Number of investments by VC fund 14.38 29.83 1 1,047
(including follow-up investments)
Number of unique PCs in VC fund 8.77 16.33 1 699
Fund size (in million USD)77 180.94 600.69 0.178 20,30079

4.2.4 Description of portfolio companies in the data set

The PC data set counts 50,443 companies, 48,982 of which received at least one investment
between 1989 and 2008. The information on these companies (variables reported by
ThomsonONE) is detailed in

77
Fund size is available for 10,017 out of 15,765 funds.
78
Small fund sizes are likely to be associated with investments by individuals or small organizations.
79
The largest funds are typically VC funds managed by banks or large organizations. For example, the largest
fund in the data set is Goldman Sachs’s “GS Capital Partners VI, L.P.” (20.3 billion USD), see
http://www2.goldmansachs.com/services/investing/private-equity/gs-capital-partners/index.html (July 2010).
 55

Appendix 2. In particular, the data set includes information on the PC location (nation), the
year the PC was founded, the industry the PC is in, the PC’s current status (e.g., went public
or defunct) and its IPO date (if applicable).

Over 53.2% of the observed PCs are based in North America (mostly in the US), 27.7% in
Europe and 13.5% in Asia (see Table 580). The remaining 5.6% of the PCs are either located
in the rest of the world (984 PCs) or their location is unknown (1,747 PCs).

An overview of the share of industry representation among observed PCs is available in Table
6. Over half of the observed PCs are in the IT industry81, 14% in Life Sciences82 and about a
third in non high-tech industries. About 5% of the observations cannot be allocated to a
specific industry. While Life Sciences and Non High-Tech PCs are founded regularly
throughout the 20 years of the analysis, the so-called “IT bubble” can clearly be observed
with a strong increase in the number of IT start-ups being founded between 1994 and 1999
(from 500 new IT PCs in 1994 to 3,000 new IT PCs founded in 1999) and a sharp decline in
2000 and 2001 (down to less than 900 in 2001), as shown in Figure 5.83

Of the PCs observed in the data set, roughly half are still active today, a third went public or
was acquired by or merged with another company. 12.5% of the observed PCs are now
defunct (see Table 7).

Table 5: Location of observed portfolio companies (by continent)

PC location (HQ) Frequency Percentage

North America 26,081 53.2%
Europe 13,542 27.7%
Asia 6,628 13.5%
ROW 984 2.0%
Unknown 1,747 3.6%
Total 48,982 100%

Table 6: Industry of observed portfolio companies

PC industry Frequency Percentage

IT 25,275 51.6%
Non High-Technology 14,471 29.5%
Life Sciences 6,617 13.5%
Unknown 2,620 5.4%
Total 48,982 100%

80
A detailed overview of PC location by PC nation is available in Appendix 8.
81
Deloitte & Touche LLP. and the US National Venture Capital Association reported that US-based VC firms
were holding 51% of their investments in IT investments as of 2005 (Deloitte / NVCA 2005).
82
“Life Sciences” and Medical / Biotechnology” are used as synonymous.
83
This peak is also observed in the number of financing rounds of IT start-ups (see Figure 7).
56

Table 7: Current status of observed portfolio companies

PC status84 Frequency Percentage

Active 24,018 49.0%
Merged or acquired 10,641 21.7%
Defunct 6,707 13.7%
Went public 6,123 12.5%
Unknown 1,493 3.1%
Total 48,982 100%

Figure 5: Number of observed portfolio companies by founding year and industry

3,000

2,500
Number of observed PCs

2,000

1,500

1,000

500

0
1980 1985 1990 1995 2000 2005
PC founded year

PC industry:

IT Medical / Biotechnology Non High Technology Unknown

This figure only shows PCs founded after 1980 (N=45,813)

4.2.5 Description of investments in the data set

The data set observes 227,441 investments, with a number of observed investments steadily
increasing from each observation year to the next (with the exception to the IT bubble around
the year 2000), as shown in Figure 6.

An investment is defined as the event by which a given VC firm invests in a given PC on a

given date. Syndicated investments (in which several VC firms invest together in the same PC
on the same date) are counted as several investments. If a VC firm follows up on its initial
investment and provides further funding to the PC at a later date, these follow-on investments
are counted as additional investments too.

The data set observes 108,928 investment rounds (or “investment opportunities”). An
investment round is defined as a given PC receiving VC funding at a given date, regardless of
how many VC firms participate in the financing round. A PC can receive several rounds of

84
Status as of July 2009.
 57

VC financing over its lifetime and these are counted as that many different financing rounds.
The data set observes 48,982 unique portfolio companies having received at least one round
of financing between 1989 and 2008.85 The distinction between PC, financing round and
single investment is important to interpret the statistics and analyses in this document.

Each observed PC received an average of 2.4 financing rounds.86 Over half of the observed
financing rounds (close to 60,000) are syndicated rounds87, meaning that more than one VC
firm participated in the financing round. The majority of the syndicated financing rounds
involved 2 or 3 VC firms investing simultaneously in a common PC.88

While the number of financing rounds by PC industry is relatively steady over time (below
2,000 a year), IT financing rounds peak between 1999 and 2001, with over 10,000 financing
rounds in IT PCs in 2000 alone. Since then, IT financing rounds stabilized somewhat below
4,000 per year (see Figure 7).

31.7% of the observed financing rounds went to PCs in an early stage of development, 39.8%
in growth stage and 19.2% in late stage (see Table 8).

Table 8: Investment stage of observed financing rounds

Round investment stage Frequency Percentage

Early stage 34,541 31.7%
Growth stage 43,306 39.8%
Late stage 20,945 19.2%
Other 10,136 9.3%
Total 108,928 100%

85
See section 4.2.4 for more details on the observed PCs.
86
This is consistent with the 2.7 financing rounds per PC reported by Gompers and Lerner (2006, 179).
87
An investment round is defined as syndicated if more than two or more VC firms are observed to invest in the
same PC at the same date. Should a PC receive an investment round by two different VC firms but only one of
these investments is observed, the investment round would be defined as not syndicated. Therefore, it is
possible that the frequency of syndication is underestimated.
88
For more details on the number of VCs participating in each observed financing round, see Appendix 9.
58

Figure 6: Yearly number of observed VC investments (1989-2008) 89

30,000

20,000

10,000

0
1990 1995 2000 2005
Investment year

N=227,441

Figure 7: Number of financing rounds per observation year and by PC industry

10,000
Number of observed financing rounds

8,000

6,000

4,000

2,000

0
1990 1995 2000 2005
Financing round year

PC industry:

IT Medical / Biotechnology Non High Technology Unknown

N=108,928

89
All years are observed in their entirety.
 59

4.3 Creation of additional variables

4.3.1 Measurement of distance

Following a common method (Sorenson / Stuart 2001; Pruthi et al. 2003), geographic distance
is defined based on the location of the VC firm’s headquarters and the portfolio company’s
main office.90 I use the VC and PC nation as the unit of analysis, group nations in continents91
and use dummy variables to distinguish between three ranges of investment scope: national,
continental and global92. Table 9 shows an example of the values of each dummy variable
representing the geographic distance variable between a VC firm and its investments in
different target locations.

Alternatively, the scope unit could have been defined at the state level (e.g., California in the
US or Bavaria in Germany) or at the continent level. Other possibilities could have been to
choose a non-political unit to define scope such as the cluster level (e.g., the Silicon Valley in
the US) or even a purely geographic measure such as a given perimeter or travel time.

The advantage of defining scope at the nation level is that the nation is a unit for which a clear
definition of market conditions is possible (in particular supply and demand). This would not
be the case at the cluster level or with absolute geographic distance for example. Choosing the
nation level as the base for the definition of distance also allows for the identification of
additional influencing factors (legal systems for instance), which would not be possible at the
continent level.

However, a nation-based measure of distance also has some drawbacks. First, a nation-based
measure of distance may be problematic for VCs located in very small countries. For
example, investment opportunities may not be as abundant in a small country such as
Luxembourg (less than 3,000 km2 large with a population of just half a million) as in a large
country such as the US (over 9,000,000 km2 and over 300 million inhabitants). To counter
this problem, relative measures of supply and demand of deals are used in the regressions.
Second, the physical distance is not perfectly portrayed by the indication of common

90
The data source neither provides information on whether or where a VC firm has other offices and when these
were opened, nor on whether, from where and when a portfolio company may have relocated. Therefore, the
chosen measurement of distance may be inaccurate in some cases.
91
The classification of nations in continents is available in Appendix 4.
92
This distinction follows the model of the German Private Equity and Venture Capital Association (BVK) who
distinguishes between three types of investments made by German firms: within Germany, in Europe but not
in Germany, and outside of Europe (BVK 2006).
60

nationality. For example, constraints for a Munich-based VC investing in and monitoring a

Berlin-based start-up (one time zone, 1-hour flight) are not similar to the challenges facing a
Silicon-Valley-based VC investing in and monitoring a New-York-based company (3 time
zones, 5-hour flight), although both investments would be considered as local investments
when distance is defined based on the VC’s nation. To account for these differences, VC
location control dummies are included in the regressions.

Table 9: Examples of distance measurement

VC firm location PC location At a distance National Continental Global

Germany (Europe) Germany (Europe) 0 1 0 0
Germany (Europe) UK (Europe) 1 0 1 0
Germany (Europe) US (North America) 1 0 0 1
US (North America) Canada (North America) 1 0 1 0
US (North America) Mexico (ROW) 1 0 0 1
US (North America) China (Asia) 1 0 0 1
Location is determined by VC and PC headquarters location.
Investment at a distance: PC and VC are located in different nations.
National investment: PC is located in VC nation.
Continental investment: PC is located in VC continent but not in VC nation.
Global investment: PC is located outside of VC continent.

4.3.2 Measurement of deal flow (supply)

Ideally, one would want to know how many applications a given VC firm receives in a given
year, or how many requests for funding were emitted in a certain area over a given timeframe.
Since the actual deal flow is not observable, an approximation has to be made. In the
following paragraphs, economic growth, innovation intensity and observed investments are
discussed as possible indicators of the deal flow available to VCs in a given area over a given
timeframe.

Economic growth

VCs must provide a steady rate of return to their investors regardless of externalities such as
industry health. They must therefore ensure access to a sufficient deal flow, both in terms of
quality and quantity, even during economical downturns. There are two arguments in favor of
linking local economic growth and supply of interesting investment opportunities.

First, according to Schmookler’s empirical economic study unveiling demand-pull influences

(Schmookler 1966), technological change and innovation (the areas of interest for VCs) are
driven by a raise in demand. In other words, technological change and innovation occurs
mostly in times of economic growth. An implication of this theory is that the number of high-
value company creations increases in times of economic growth and decreases in times of
 61

recessions. Using data from the Global Entrepreneurship Monitor, Linghui Tang and Peter
Koveos (2004) demonstrate this link: they show that the total national entrepreneurial activity
of a country is directly related to this country’s GDP growth rate over the previous five years,
especially in developed countries (Linghui / Koveos 2004; Du et al. 2008).93 This means that
more high-potential start-ups are created when the economy is growing and less when the
economy is slowing down.94 Figures reported by regional VC associations also show a similar
correlation (McFadden 1974; EVCA 2007). The phenomenon can be explained by
considering that the growing demand for products and services in times of economic growth
constitute a pull effect for venture creation by raising perceived success chances for
entrepreneurs (Schmookler 1966; Scherer 1982; Bretz et al. 2009).95

Second, the quality of the start-ups created in times of economic growth is expected to be
higher than in times of recession. In times of growth, more attractive employee positions are
available in existing firms (Bretz et al. 2009). Therefore, an entrepreneurial career should be
selected by rational individuals only if they perceive it as more attractive than an employee
position, signaling the potential of the venture to become successful. On the opposite, in times
of recession or less favorable economic state, company creations can be motivated by a lack
of employment opportunity, i.e., the entrepreneurial path is preferred to unemployment
(Linghui / Koveos 2004), which could indicate entrepreneurship by necessity, with low
concern for high-quality projects.

For these two reasons, VCs should receive more good business plans from local entrepreneurs
(hence the higher probability to find interesting investment opportunities) in times of
economic boom in the VC’s area. In contrast, in times of economic downturn in the VC’s
area, they will receive less good business plans from local ventures and may have to enlarge
the radius of venture locations they consider. As a consequence, they would invest in non-
local start-ups to compensate for the lack of good local deals. This statement was supported

93
In an attempt to research the determinants of venture capital further, a panel analysis of OECD countries
shows that rather the contemporaneous GDP growth rate has a significant impact on the amount of VC
invested (Romain / Van Pottelsberghe 2004). However, it is not clear from the study whether more VC money
is invested in start-ups because there are more promising start-ups in years of economic growth or whether VC
firms tend to invest more during economic growth, speculating on following years of economic growth.
94
Studies showing a negative impact of per capita income on self-employment levels (Noorderhaven et al. 2004;
Wennekers et al. 2005; Carree et al. 2007) are not in contradiction with this analysis: most of these studies use
data on business owners, whereas venture capitalists are interested in the creation of new business, i.e.,
“nascent entrepreneurship” (Wennekers et al. 2005).
95
This trend can be balanced by the increase in the availability of attractive employee positions in existing firms
in times of economic growth, which may reduce the readiness to choose an entrepreneurial career path.
62

by the expert interviews. A New-York based investor comments: “In times of crisis you
receive fewer business plans. (…) It certainly feels to me that the quality deals that you want
to see are higher when the tech markets are buoyant”. Similarly, a Munich-based VC
explains: “People go out to start companies when the economic climate is good. If the climate
is not good, they do not want to leave their job to start new companies.”

Gross domestic product per nation and per year could be used to operationalize this variable.
This information could be obtained from the International Monetary Fund’s “World
Economic Outlook” database96. However, since a measure of economic growth common to all
nations and available for the twenty observation years could not be found, another indicator of
deal supply must be found.

Innovation intensity

Innovation intensity would be another possible proxy for the deal flow available to VC firms
based on recent research having proved that “money goes where the innovation is, not the
other way around” (Wadhwa 2008, 18), a view shared by Guler and Guillen (2004). From an
economic standpoint, it is also the VC firms’ role to finance innovation in its early stages by
investing in promising innovative companies because banks are not set up to do so, due to
different risk-return models (Guilhon / Montchaud 2006).

Amounts spent on research and development within a given timeframe could be used to
operationalize this variable. For example, the Organisation for Economic Co-operation and
Development (OECD) publishes a history of gross domestic expenditure on research and
development as a percentage of gross domestic product in selected nations 97 in its yearly
factbook.98 Unfortunately, the data is not available for all the nations and observation years,
which makes a complete analysis impossible.

Patents have been another popular measure of innovation (Furman et al. 2002), the two
having even been sometimes used as synonyms (Bessler / Bittelmeyer 2008). The common
choice of patent characteristics as an indicator of innovation can be explained by the
convenient availability of public patenting databases (Kleinknecht et al. 2002). In addition,
studies showing that VCs put a premium on start-ups demonstrating the value of their

96
The data is available on www.imf.org.
97
Data is available for 38 nations in the 2007 version.
98
The data is accessible online (www.oecd.org).
 63

innovation via patents confirm this view (Lerner 1994; Mann / Sager 2007; Häussler et al.
2008; Hsu / Ziedonis 2008). However, the main issue with proxying deal flow by patent data
is that established corporations are likely to be responsible for a large part of the patent
applications, which means that high patent application or attribution numbers in a certain
nation at a certain time does not necessarily indicate high levels of deal flow for VC firms.

Observed investments

The most straightforward way to proxy the real deal flow is to count how many deals were
actually funded within a given timeframe.99 I use relative values in percentage points. For
instance, looking at an investment by a Germany-based VC firm in 2005, I approximate the
VC firm’s local supply of deals at the time of the investment by counting how many
Germany-based PCs received a financing round in 2005100 (160 financing rounds) and
calculate relative local supply by dividing by all PCs worldwide (7,191 investment
opportunities) in 2005, i.e., 2.34% (see Table 10 for a complete example). The analyses
presented in Chapter 5 and Chapter 6 use the percentage value of deal supply.

99
I thank Professor Frank Gruber (Ecole Polytechnique Fédérale de Lausanne) for this suggestion.
100
I do not distinguish between new and follow-up investments or between investment rounds because a VC can
join (or retrieve from) an investment opportunity at any time.
64

Table 10: Example of local supply measurement for German VC firms

Investment year Supply (nation) Supply (worldwide) Relative local supply (/worldwide)

1989 4 2,246 0.18%

1990 7 1,927 0.36%
1991 8 1,823 0.44%
1992 5 2,147 0.23%
1993 14 1,930 0.73%
1994 14 2,064 0.68%
1995 21 2,654 0.79%
1996 55 3,944 1.39%
1997 39 4,322 0.90%
1998 79 5,472 1.44%
1999 190 7,664 2.48%
2000 526 13,841 3.80%
2001 403 8,842 4.56%
2002 184 5,865 3.14%
2003 127 6,889 1.84%
2004 160 7,468 2.14%
2005 168 7,191 2.34%
2006 166 6,877 2.41%
2007 206 7,879 2.61%
2008 182 7,660 2.38%

Total 2,558 108,705 2.35%

4.3.3 Measurement of competition (demand)

This variable measures the intensity of the competition by local VC firms for attractive deals.
It is measured by counting how many VC firms are active in a given year, within a certain
area, only considering VC firms existing at that time (i.e., for which the founding year
precedes the observed investment year).101 For example, looking at an investment by a
Germany-based VC firm in 2005, I indicate the firm’s local competition (local demand for
deals) at the time of the investment by counting how many VC firms located in Germany are
active in 2005 (i.e., were founded in 2005 or earlier), which is 299 VC firms. To calculate
relative demand, I divide by the number of all VC firms active worldwide in 2005 (7,033 VC
firms), which is 4%. The percentage point value is use for these analyses.

101
In this regard, the method may overestimate the intensity of the competition, since defunct or inactive VC
firms are considered as competitors in the years following their retirement from investment activities.
 65

4.3.4 Measurement of VC ability and fit

4.3.4.1 VC fund-related variables

VC fund industry specialization

A fund is considered specialized if at least 75% of its portfolio companies are in the same
industry group. The complete investment history of the fund’s new investments102 is taken
into account as observed in the sample. The industry classification of portfolio companies
follow the Venture Economics Industry Codes developed by Thomson Reuters and indicates
what a company develops, manufactures, produces, provides or resells.103 The primary
assigned industries (i.e., the industry classification that describes the main function of the
company) considered here are IT, Life Sciences and Non-High-Technology.104 If the fund is
specialized, the industry of specialization may have an effect on the VC firm’s scope of
investments as well as on its choice of investment target locations. I therefore control for the
fund’s industry of specialization, if applicable. I use the specialization in the IT industry as a
reference. As detailed in Table 11, 39.8% of the observed funds are not industry-specialized.
The remaining 60.2% are industry-specialized funds105 (39.4% are specialized in the IT
industry, 11.6% in Life Sciences and 9.2% in Non High-Technology).

Table 11: Industry specialization of observed VC funds

Fund industry of specialization Number of observed funds Share of observed funds

No industry specialization 6,270 39.8%
IT 6,213 39.4%
Life Sciences 1,835 11.6%
Non High-Technology 1,447 9.2%
N=15,765
A fund is considered specialized in a given industry if at least 75% of its PCs are in this industry.

VC fund investment stage specialization

A VC fund is considered specialized in a given investment stage if more than 50% of its new
investments are in portfolio companies in that investment stage. I use a lower percentage
definition than for industry specialization and only consider new investments because

102
Only new investments are considered, because a portfolio company’s industry does not change over time: this
way, I do not count several times a PC that received several rounds of financing from the same fund.
103
For a detailed description of each industry’s sub-groups, see Appendix 3.
104
The portfolio companies assigned to the industry category “unknown” are counted in the fund, but not
assigned to any specific industry, so that the amount of fund specialization may be underestimated.
105
This figure is close to the 54% of industry-specialized new European funds reported by the European Private
Equity and Venture Capital Association in 2007 (EVCA 2008).
66

contrary to industry classification, investment stage varies over time, so that an early-stage
investment today may become a later-stage investments a few years later if the VC firm
follows up on its initial investment. Following Schmidt (2004), I aggregate the reported stage
classification by ThomsonONE into three simple investment stages: early stage, growth
stage106 and late stage. The share of specialization in each investment stage for the observed
VC funds is depicted in Table 12.

Table 12: Stage specialization of observed VC funds

Investment stage of Number of observed funds Share of observed funds

specialization
No stage specialization 5,753 36.5%
Early stage 5,225 33.1%
Growth stage 2,713 17.2%
Late stage 2,074 13.2%
N=15,765
A fund is considered specialized in a given investment stage if over 50% of its investments are in this stage at the
time of the investment.

VC fund size

Fund size is measured in the nominal amount of capital (in USD) in the initial fund (i.e.,
amount raised by VCs from LPs). Fund size is reported in two thirds of the funds listed in the
sample.107 For the funds without a reported fund size, I assign the average fund size of the
funds of the same type.

VC fund age at the time of investment

I calculate the age of the fund in years by subtracting the year the fund was founded from the
year of the investment at stake. In the 14 instances for which the fund founding year is not
available, I manually research information. The funds for which fund size is still not available
after this search are assigned the year of the fund’s first observed investment as the fund’s
founding year (i.e., the VC fund’s age is equal to 0 for its first observed investment).

VC fund source

The information collected about LPs cannot be linked to specific funds due to the lack of a
common key variable (see

106
Schmidt (2004, 34) calls this stage “expansion stage”.
107
See section 4.2.3.
 67

Appendix 2). Therefore, it is not possible to distinguish the exact sources of financing of each
fund. An approximation is made by looking at the VC fund type reported by ThomsonONE,
which I aggregate into five simple categories: independent, bank, corporate, public and
individuals (see section 4.2.3 for details).108 I use dummy variables to distinguish between the
different types and choose the “independent” category as the point of reference, since this is
the most common VC fund type.

4.3.4.2 VC firm-related variables

VC firm experience

One way of measuring VC firm experience is to use the VC firm age at the time of the
investment. This variable is calculated by taking the difference in years between the year of
the investment round and the year the VC firm was founded (Cumming / Dai 2008).
Sometimes, the database allocates investments to the firm before the reported founding date
of the VC firm: in such cases, following Jääskeläinen et al. (2006), I set the VC founding year
to the year of the first investment. In the 84 instances for which the firm founding year is not
available, I manually research this information or assign the year of the firm’s first observed
investment as the firm’s founding year (i.e., the VC firm’s age is equal to 0 for its first
observed investment).

An alternative measure of VC firm experience is to count how many investments the VC firm
has made prior to the current investment (Sorenson / Stuart 2001; Cumming / Dai 2008).

VC firm office presence

This information is not available in the data set, but I manually collect this information for all
German VCs. I then assign the value 1 to the variable if the VC firm has an office in the PC’s
nation and the value 0 otherwise.

VC firm investment experience with target location

To model experience, I count the number of observed investments made by the VC firm in the
target location until the observation date.

108
The data set contains the following fund types: angel, business / community development program, corporate
PE / venture fund, endowment / foundation / pension fund, evergreen, fund of funds, government, independent
private partnership, individuals, investment advisory affiliate, investment bank, other, other banking / financial
institution, small business investment company (SBIC), secondary purchase, university development program.
68

4.4 Geography of the venture capital industry

4.4.1 Historical perspective

The VC industry has historically been the most developed in the US, developing around the
country’s two main innovation poles, California (in particular the Silicon Valley) and
Massachusetts, in the direct vicinity of some of the nation’s most prestigious research
universities (e.g., Stanford University, Massachusetts Institute of Technology and Harvard
University).109 Since the start of the VC industry in the mid-1970’s, investments in venture
funds grew to an all-time high 105 billion USD in the US at the height of the Internet bubble
in 2000 (Gompers / Lerner 2006). American companies such as eBay, Google, Apple or
Genentech all once received venture capital. Across the Atlantic, the European VC industry is
still not as prominent, but it is catching up: a recent report on entrepreneurship by The
Economist (Anonymous 2009) shows that European VC investments have grown on average
23% per year between 2003 and 2006, compared to just 0.3% in the US. In order to confirm
this statement, an analysis of investments made by VCs per continent over the period between
1989 and 2008 was led (Table 13): from just 4.28% of all VC investments observed between
1989 and 1993, the proportion of number of investments made by European VCs increased to
23.11% in the last 5 years of observation (2004-2008), which confirms the trend exposed by
The Economist.

Table 13: Evolution of the origin of observed investments (1989-2008)

VC location 1989-1993 1994-1998 1999-2003 2004-2008

Asia 348 (01.63%) 1,184 (03.34%) 7,542 (08.21%) 5,143 (06.54%)

Europe 916 (04.28%) 2,733 (07.72%) 18,014 (19.60%) 18,184 (23.11%)
North America 20,127 (94.07%) 31,440 (88.76%) 65,840 (71.64%) 54,924 (69.81%)
ROW 5 (00.02%) 64 (0.18%) 513 (00.56%) 431 (00.55%)

Total 21,396 35,421 91,909 78,682

Number of investments by VC location by observation tranche (N=227,407).

VC location is not available for 3 VC firms.

4.4.2 VC champion nations

Table 14 shows the total number of observed investments per VC continent and nation over
the past 20 years.110 With 168,653 observed investments, US-based VCs account for almost

109
The next top US states for venture-backed jobs and revenues are New York, Pennsylvania and Texas (NVCA
2009).
110
A detailed table of the number of observed investments per nation is provided in Appendix 5.
 69

three quarters of all the observed investments made during these twenty years. In Europe, the
UK, France and Germany are leading the industry, followed by Israel, Sweden and the
Netherlands: VCs headquartered in the UK made 11,668 investments over the past twenty
years, while French VCs made close to 6,907 and German VCs roughly two thirds of that
amount. In Asia, South Korea-based VC firms made almost as many investments as their
German counterparts, followed by Australia, Japan and India. Beside Brazil (just 400
observed investments between 1989 and 2008), nations in South America, Africa and the
Middle-East each made less than 100 observed investments over the past 20 years.

Table 14: Observed investments by VC location

VC location Frequency Percentage

North America
US 168,653 74.16%
Canada 3,678 1.62%

Europe
UK 11,668 5.13%
France 6,907 3.04%
Germany 4,464 1.96%
Israel 2,210 0.97%
Sweden 1,888 0.83%
Netherlands 1,880 0.83%
Finland 1,810 0.80%
Switzerland 1,349 0.59%
Denmark 1,185 0.52%
Belgium 1,176 0.52%
Other (Europe) 5,310 2.34%

Asia
South Korea 4,109 1.81%
Australia 2,775 1.22%
Japan 1,515 0.67%
India 1,419 0.62%
Singapore 1,172 0.52%
Taiwan 1,041 0.46%
Other (Asia) 2,186 0.96%

ROW
Other (ROW) 1,013 0.45%

N=227,407
Investments for which VC location is unknown are not reported in the table.
Nations in which national VCs made less than 1,000 investments over the observation period are reported in
aggregated form.

4.4.3 Cross-border VC investments

4.4.3.1 Investment scope

Table 15 shows that VCs’ preferences in terms of investment scope differ by continent: while
VCs in North America made less than 6.58% of their investment outside of their own
70

nation111, their European and Asian counterparts each made close to 40% of their investments
at a distance.

Table 15: Scope of observed investments by VC continent

VC location National Continental Global Total

North America 153,183 (93.43%) 2,242 (01.37%) 8,537 (05.21%) 163,962
Europe 21,977 (61.64%) 5,955 (16.70%) 7,720 (21.65%) 35,652
Asia 8,025 (62.43%) 1,345 (10.46%) 3,485 (27.11%) 12,855
ROW 600 (61.04%) 91 (09.26%) 292 (29.70%) 983
Total 183,785 (86.10%) 9,633 (04.51%) 20,034 (09.39%) 213,452
N= 213,452
Investments for which VC or PC nation is not known are not reported.
The share of investments made by VCs at a given investment scope over the observation period are indicated in
parentheses.

In those countries with low availability of domestic venture capital, start-ups may be more
likely to rely on foreign venture capitalists to support their development. For instance, in
1999, at the peak of venture capital investments, Denmark, Ireland and Finland,
comparatively small countries with regards to venture capital, received the highest share of
cross-border venture capital from all European countries (Baygan / Freudenberg 2000).

In countries where venture capital has been developing and maturing for a longer period of
time, local competition for deals and possible market saturation may be forces encouraging
domestic VCs to look for cross-border investment targets. Another possibility could be that
VCs first gain experience locally before considering engaging in cross-border investments in
more mature markets. For instance, a Munich-based VC firm founder and general partner I
interviewed argued: “We cannot simply focus on German deals if they are not going to be the
right candidates for returns”. However, US and European VCs seem to differ in their cross-
border activities. US VCs have traditionally invested largely in US American companies
rather than in foreign firms. As shown in Figure 8, US-based VC firms made just above 5% of
their investments in PCs located outside of the US during the first five years of observation.
However, this share almost tripled over the next ten years, reaching over 16% of cross-border

111
The low share of investments at a distance by VCs in North America is largely driven by US-based VCs’
focus on national PCs (US-based VC firms made only 5.90% of their investments in PCs located outside of the
US over the 20-year observation period). Although US VCs may invest in geographically distant PCs (e.g.,
VC is located on the East Coast and PC in California), such investments still fall under the category of national
investments according to the definition. However, country size cannot be the only explanation: for example,
Canada is slightly larger than the US but VCs based in Canada made 45.38% of their investments in PCs
located outside of Canada over the same observation period. Similarly, VCs located in Russia (1.8 times larger
than the US) made 35.56% of their investments in PCs located outside of Russia between 1989 and 2008.
 71

investments between 2004 and 2008 (see Figure 8).112 European VCs on the opposite made
close to 40% of their investments outside of their own nation over the complete observation
period (see Table 17). For them however, the trend seems to be that the share of national
investments is increasing (from 31.6% in 1989-2003 to 61.6% in 2004-2008). The importance
of cross-border investments differs between European nations, as shown in Table 16. For
example, French VCs made less than 20% of their investments at a distance during the 20
years of observation, while UK-based VC firms made half of their investments outside of the
UK.

Table 16: Scope of observed investments by top 10 VC nations

VC location National Continental Global Number of

investments investments investments investments
US 94.10% 0.66% 5.24% 161,157
UK 53.89% 20.13% 25.98% 10,671
France 82.10% 7.68% 10.21% 6,208
Germany 57.58% 13.16% 29.26% 4,286
South Korea 92.01% 0.85% 7.14% 3,052
Canada 54.62% 42.17% 3.21% 2,805
Australia 75.82% 4.20% 19.99% 2,667
Israel 44.29% 3.54% 52.17% 2,147
Netherlands 48.61% 30.92% 20.47% 1,656
Sweden 71.79% 17.23% 10.98% 1,439
Investments for which PC location is unknown are not reported in the table.

Figure 8: Share of investments at a distance by US-based VC firms (1989-2008)

20%

15%

10%

5%

0%
1990 1995 2000 2005
Investment round year

112
In a 2005 survey, 20% of the respondents (US based VC firms) stated that they planned to increase their
global investment activity over the next five years (Deloitte / NVCA 2005).
72

Table 17: Scope evolution of investments by Europe-based VCs (1989-2008)

Investment type 1989-1993 1994-1998 1999-2003 2004-2008 Total

(PC-VC distance)
National 280 718 11,132 9,847 21,977
31.57 % 44.43 % 62.43 % 64.28 % 61.64 %
Continental 125 326 3,015 2,489 5,955
14.09 % 20.17 % 16.91 % 16.25 % 16.70 %
Global 482 572 3,683 2,983 7,720
54.34 % 35.40 % 20.66 % 19.47 % 21.65 %
Total 887 1,616 17,830 15,319 35,652
100.00 % 100.00 % 100.00 % 100.00 % 100.00 %

4.4.3.2 Target locations

It is also interesting to look at the top destinations for VC investments per VC nation. Table
18 reports the top 10 destinations for the 10 nations having made the most investments in the
observation period. For all top 10 nations, the number one investment target location is the
VC’s nation itself. When VCs invest outside of their own nation, they primarily invest in PCs
located in the US (except US-based VCs of course).113 US VCs have the most variety in their
top investment destinations (UK, China, Canada, Israel and India as the top destinations).
European VCs have most of their PCs within Europe. UK-based VCs have 67 PCs in India
and a few French and German VCs have PCs in Canada. In some nations, VCs seem to prefer
investing in culturally close nations. For instance, Germany is the only reported nation with
Austria as a top investment destination (also a German-speaking country), Sweden has
Denmark, Finland and Norway as top investment destinations (geographically and culturally
close nations) and South Korean VCs have China as their second preferred foreign target
location after the US.

113
A similar finding is made in Deloitte and the NVCA’s survey of 545 VC firms worldwide: “the United States
remains the most attractive investment target for venture capital (VC) firms worldwide” (2005, 1). However,
the clear domination of the US both as a source of investments (VC location) and as a first foreign target
location may indicate a bias towards US observations.
 73

Table 18: Top 10 investment target locations by VC nation

VC location: Number of PCs Investment type VC location: Number of PCs Investment type

US UK
1 US 66,058 national UK 3,573 national
2 UK 950 global US 1,192 global
3 China 662 global Germany 294 continental
4 Canada 571 continental France 248 continental
5 Israel 440 global Canada 119 global
6 India 405 global Sweden 109 continental
7 France 329 global Ireland 80 continental
8 Germany 293 global Spain 76 continental
9 Japan 280 global Netherlands 60 continental
10 Netherlands 157 global India 57 global
France Germany
1 France 3,449 national Germany 1,883 national
2 US 328 global US 616 global
3 UK 76 continental UK 74 continental
4 Switzerland 45 continental Switzerland 73 continental
5 Germany 40 continental Israel 50 continental
6 Belgium 29 continental France 46 continental
7 Italy 21 continental Austria 33 continental
8 Spain 19 continental Netherlands 20 continental
9 Canada 17 global Canada 14 global
10 Israel 15 continental Sweden 13 continental
South Korea Canada
1 South Korea 2,193 national Canada 1,020 national
2 US 104 global US 725 continental
3 China 12 continental UK 17 global
4 Thailand 3 continental Japan 14 global
5 UK 2 global France 12 global
6 Canada 2 global Israel 5 global
7 Japan 2 continental Germany 4 global
8 Hong Kong 2 continental Ireland 4 global
9 Taiwan 2 continental India 4 global
10 Israel 1 global Australia 2 global
Australia Israel
1 Australia 990 national Israel 605 national
2 US 161 global US 536 global
3 New Zealand 46 continental UK 23 continental
4 UK 27 global France 21 continental
5 Singapore 8 continental China 8 global
6 Canada 6 global Germany 4 continental
7 Hong Kong 3 continental Netherlands 4 continental
8 Belgium 2 global Canada 2 global
9 China 2 continental Singapore 2 global
10 Germany 2 global Switzerland 1 continental
Sweden Netherlands
1 Sweden 676 national Netherlands 588 national
2 US 84 global US 192 global
3 Denmark 40 continental UK 89 continental
4 Finland 32 continental Germany 59 continental
5 UK 25 continental Belgium 45 continental
6 Norway 20 continental France 42 continental
7 Germany 10 continental Sweden 17 continental
8 Switzerland 6 continental Switzerland 17 continental
9 France 5 continental Israel 15 continental
10 Netherlands 4 continental Denmark 13 continental
74

4.4.3.3 Cross-border investment champion VC firms

A closer look at the VC firms with the most investments at a distance over the observation
period (i.e., continental or global investments) shows that there is no single VC nationality
dominating cross-border investments (see Table 19). However, Asian VC firms are
overrepresented (Singapore, Japan, Hong Kong, Taiwan). An explanation may be that these
nations have a growing power to invest in venture capital, but a lack of appropriate
investment targets at home, so that they seek interesting portfolio companies abroad. The
presence of a number of corporate VC firms is also worth noticing (e.g., Nortel Networks
Corporation, Mitsubishi Corporation or Siemens Venture Capital). Corporate VCs may be
better able to invest at a distance thanks to their existing global presence providing the
necessary appreciation and contacts to diverse local markets. These VC firms are also likely
to be industry-specialized and to finance either their own spin-off companies or external
companies in a field for which they identify future use in their core business.
 75

Table 19: VC firms with the most investments at a distance114

VC firm name VC location Number of Number of Percentage of
investments cross-border cross-border
investments investments
Actis Capital LLP UK 73 72 99%
Aureos Capital, Ltd. UK 109 107 98%
Sycamore Ventures Pte, Ltd. Singapore 69 67 97%
Emerald Technology Ventures (= SAM Switzerland 60 58 97%
Sustainable Asset Mngt)
Sumitomo Corporation Japan 89 86 97%
Nortel Networks Corporation (= Canada 56 54 96%
Northern Telecom Ltd.)
Vertex Management Pte, Ltd. (= Singapore 238 227 95%
Vertex Venture Holdings)
Japan/America Ventures, Inc. US 144 137 95%
Mitsubishi Corporation Japan 77 73 95%
CSK Venture Capital Co., Ltd. Japan 54 51 94%
Alpha Group (= Alpha Associates Channel 66 62 94%
Management, Ltd.) Islands
CIBC Wood Gundy Capital Canada 96 89 93%
SVM STAR Ventures Management Germany 261 240 92%
Robeco Private Equity Netherlands 61 56 92%
Cipio Partners Germany 139 126 91%
SAIF Partners Hong Kong 73 66 90%
JAFCO Investment [= Nomura/JAFCO Singapore 160 143 89%
Investment (Asia), Ltd.]
HBM BioVentures AG (= HBM Switzerland 137 122 89%
Partners AG)
Coller Capital UK 173 153 88%
Yasuda Enterprise Development Co., Japan 98 86 88%
Ltd. (= Nippon Ent.Dev)
Bio*One Capital Singapore 64 56 88%
Danish International Investment Fund, Denmark 94 82 87%
The (= IFU)
Allianz Private Equity Partners Germany 53 46 87%
Schroders PLC UK 83 72 87%
RVC Europe Limited (= The UK 99 85 86%
Greenhouse Fund)
Cedar Fund Israel 56 48 86%
Index Ventures Management SA Switzerland 184 157 85%
Siemens Venture Capital GmbH (= Germany 152 129 85%
SVC)
BPEP International (= Baring Private Hong Kong 115 97 84%
Equity Asia)
China Development Industrial Bank Taiwan 88 73 83%
(CDIB)

114
This table only reports the VC firms with more than 50 observed investments over the 20-year observation
period.
76

4.5 Summary: advantages and drawbacks of the data set

This thesis uses a comprehensive data set of 227,441 venture capital investments made by
7,601 VC firms worldwide in 50,443 portfolio companies worldwide from 15,820 VC funds
between January 1st, 1989 and December 31st, 2008. It was manually collected, primarily
using the VentureXpert module of the ThomsonONE.com database. To the best of my
knowledge, it is one of the most comprehensive data sets on VC investments available to
research. The data distinguishes itself from data sets used in previous studies through its high
level of detail (only a portion of which is used in the present thesis) with in-depth information
on PCs, investments, VC firms and VC funds, its international character (contrary to many
other data sets, the data set used in this thesis does not focus on one single nation or
continent) and its longitudinal aspect with 20 years of recorded investments, including the
birth of the VC industry in some nations, a worldwide boom of IT investments during the
“bubble” and subsequent crises of the VC industry. For the first time it is possible to draw
general conclusions on the investment decision of VCs worldwide, while comparing
differences between the investment decisions made by VC firms of different home nations,
and to investigate the role of environmental factors on VCs’ decision making process over an
extended period of time.

A first weakness of the data set is that the additional data collected on 7,906 limited partners
(investors in VC funds) cannot be linked to the data on VC funds. It is therefore not possible
to know exactly whether the VC firms’ decisions in terms of investment scope and investment
target location may be influenced by the characteristics of the LPs or not. A second weakness
is the lack of personal information on the venture capitalists (general partners) at the head of
the observed VC firms. For example, the presence of a GP of Chinese origin at the head of a
VC firm in the Silicon Valley might signal the readiness of the firm to make investments on
the Asian continent and in China in particular. As a consequence, VC firms are supposed to
be homogeneous and are considered as the entity being the decision maker. A third weakness
is that VC and PC location are determined by their headquarter location. There is no
information on potential subsidiary offices. This is an important limitation for the calculation
of distance in this thesis.115 Finally, while the investment data available in VentureXpert is not
entirely complete, the missing information does not introduce a selection bias (Kaplan et al.

115
The necessary information on subsidiary offices must be collected by hand from public sources, which is very
time consuming.
 77

2002), so that the data is still expected to be representative. However, the descriptive statistics
showed that the data may be biased towards investments involving a US PC and/or a US-
based VC. If this is the case, one important consequence would be that investment scope may
be artificially low for investments by US-based VCs (more national investments) and
investment scope may be artificially high for investments by non US-based VCs (more
investments into US-based PCs). On the other hand, this concern is mitigated by the
observation that not all non-US VCs have a high share of cross-border investments. For
example, while Israeli VCs made over half of their observed investments outside of Israel
between 1989 and 2008, South Korean VCs were observed to do less than 8% of their
investments outside of South Korea.116

116
See Table 16.
78

5 Investment scope decision

5.1 Naïve scope model

The first model aims at answering the question: How likely is a VC to invest in an
international portfolio company, as opposed to investing in a PC located in the same nation?
In the basic model, the dependent variable is the chosen investment scope: investment scope
is a binary variable taking the value “1” if the investment is an international investment (i.e.,
PC and VC firm are located in different nations), or “0” if the investment is a national
investment (i.e., PC and VC firm are located in the same nation). The basic model of the
scope decision is pictured in Figure 9. The basic scope model estimates the VC’s probability
to invest at a distance and is estimated by a logit regression.117

To account for the fact that VC firms may make several investments, but that investments
made by a given firm are not necessarily independent observations, the standard errors are
clustered by VC firm to allow for intra-group correlation.118 The decision model takes into
account external factors such as supply and demand of deals in the investor’s home market
relative to foreign markets, as well as VC-specific characteristics, such as the VC firm’s
expected ability to handle investments at a distance, following the hypotheses introduced and
developed in section 3.2.

The selected scope unit is that of the nation119: an investment is either within the VC’s nation
(in this case, the investment scope is equal to “0”) or outside of the VC’s nation (in this case,

117 ଵ
‫݌‬௜ ൌ ‫ ݕ݈݈ܽ݊݋݅ݐܽ݊ݎ݁ݐ݊݅ݐݏ݁ݒ݊݅݋ݐ݀݋݋݄݈݅݁݇݅ܮ‬ൌ షሺഁబ శሺσೖ
ଵା௘ ೕసభ ഁೕ ೣೕ ሻశഄሻ
so that
௣೔
݈‫ݐ݅݃݋‬ሺ‫݌‬௜ ሻ ൌ ݈݊ ቀ ቁ ൌ ߚ଴ ൅ ሺσ௞௜ୀଵ ߚ௝ǡ௜ ‫ݔ‬௝ǡ௜ ሻ
ଵି௣೔
with
ߚ଴ = Intercept
ߚ௝ = Regression coefficient of the independent variable ‫ݔ‬௝  
‫ݔ‬௝ = Independent variable or control variable
ߝ = Residual
‫݌‬௜ = Probability that observation ݅ is an international investment.
118
Note that clustering affects the standard errors, but not the estimated coefficients.
119
For a discussion of this choice, refer to section 4.3.1.

H. Tarrade, Cross-Border Venture Capital Investments, DOI 10.1007/978-3-8349-6939-2_5,

© Gabler Verlag | Springer Fachmedien Wiesbaden 2012
 79

Figure 9: Naïve scope model

VC INVESTMENT

NATIONAL INTERNATIONAL

the investment is considered as an investment “at a distance” and the investment scope is
equal to “1”). In this first analysis, I test whether the VC’s current environment in terms of
relative supply and demand for deals plays a role in the VC’s decision to invest at a distance: I
expect that VCs should be likelier to invest at a distance when there are less deals available
locally (i.e., in the VC’s nation) or when the competition for these deals intensifies among
local VC firms. I also test hypotheses on VC firm, fund and investment-specific
characteristics (such as industry specialization, fund type, VC’s experience investing at a
distance and syndication for example) to verify their role in determining the VC firm’s
decision to invest at a distance. The control variables included in the model comprise
dummies for the location of the VC firm at the continent level (North America, Europe, Asia,
ROW), dummies for the investment period (pre-bubble, bubble, post-bubble) 120 and dummies
for the industry of the PC receiving the investment (IT, Life Sciences, non High-Tech,
unknown).121

The regression is first run on the complete data set (excluding cases for which VC nation or
PC nation are unknown) and analyzes a total number of 213,452 observed VC investments
having taken place between 1989 and 2008, with VC firms located anywhere in the world.
Since the descriptive analysis carried in section 4.4 showed that North American VC firms
(the vast majority of which are US-based firms) differ significantly from the rest of the world
in their investment behavior (in particular in their decision to invest beyond their national
borders), a second and third regression present separate results for US-based firms only

120
The “bubble” period is defined as occurring between 1998 and 2001, during which the market was strongly
heated, especially for IT investments.
121
North America is used as the continent of reference, the pre-bubble period (1989-1997) as the period of
reference and IT as the PC industry of the reference.
80

(161,157 observed investments) and for worldwide VC firms excluding US-based VC firms122
(52,295 observed investments). In addition, the regression is run on the sample consisting of
only the investments made by German VC firms (4,286 observations) to enable comparison
with the results presented in Chapter 6.

The first model estimates a baseline model of controls only and the second model tests all
hypotheses at once. The results of the analysis are presented in Appendix 12 (baseline naïve
scope model) and Appendix 13 (full naïve scope model). The results report the marginal
effects at the mean of the continuous independent variables and for a change in binary
independent variables from 0 to 1. Marginal effects show what effect a unit change in an
independent variable has on the probability of a VC to invest at a distance (investment scope
equals 1), other predictors being held constant.

The main problem with the naïve scope model is that it solely explains the scope decision at
the national level (an observed investment is either national or international). For instance, the
model treats an investment made by a German VC firm in a French venture in the same way
as an investment by the same German VC firm in a Chinese venture. In the next section, the
refined scope model is introduced to solve this problem.

5.2 Refined scope model

5.2.1 Econometric model

The refined scope model offers a more granular differentiation between possible investment
scope alternatives available to VC firms by distinguishing between national investments and
international investments made either in PCs located within the VC firm’s continent
(continental scope) or in PCs located beyond the VC firm’s continent (global scope). As
pictured schematically in Figure 10, the refined scope model allows the dependent variable
(the chosen investment scope) to take three different values depending on whether the PC
receiving the observed investment is located:

1. in the VC’s nation (i.e., “national” scope),

2. outside of the VC’s nation but on the VC’s continent (i.e., “continental” scope),

3. or outside of the VC’s continent (i.e., “global” scope).

122
This regression shows that there is no significant difference between continents after excluding all
investments by US-based VC firms.
 81

Since there is a natural ranking between the three scope options, from narrow to wide, an
ordered logit model is used to estimate the refined scope model:

‡š’ሺ ܺ௜ ߚ௝ ሻ
”ሺ ‫ݕ‬௜ ൌ ݆ሻ ൌ
ͳ ൅ σ௃௝ୀଵ ‡š’ሺ ܺ௜ ߚ௝ ሻ

where

݆ = Possible investment scope (1=national / 2=continental / 3=global)

‫ݕ‬௜ = Observed investment scope of the ݅ th investment

ܺ௜ = Vector of explanatory and control variables

ߚ௝ = Regression coefficient when investment scope is ݆

The ordered logit model depends on the idea of cumulative probability:

௝

‫ܥ‬௜௝ ൌ ”ሺ‫ݕ‬௜ ൑ ݆ሻ ൌ ෍ ”ሺ‫ݕ‬௜ ൌ ݇ሻ

௞ୀଵ

The cumulative probability is turned in the cumulative logit:

‫ܥ‬௜௝
Ž‘‰‹–ሺ ‫ܥ‬௜௝ ሻ ൌ Ž‘‰ ቆ ቇ ൌ ߙ௝ െ ߚܺ௜
ͳ െ ‫ܥ‬௜௝

where

ߙ௝ = Cutpoint for level ݆

ߚ = Regression coefficient vector

Each ߙ௝  indicates the logit of the odds of being equal to or less than category ݆ for the
baseline group (when all independent variables are equal to zero). The ߚ indicates how a one-
unit increase in the independent variable increases the log-odds of being higher than category
݆.123

123
ߚ is not indexed by ݆, so that the model assumes that the one-unit increase affects the log-odds the same,
regardless of which cutpoint is considered.
82

Figure 10: Refined scope model

VC
INVESTMENT

NATIONAL CONTINENTAL GLOBAL

5.2.2 Estimation

In this analysis, I observe a posteriori the type of investments made by each VC firm, i.e.,
whether the investment’s scope was national (PC is located in the VC’s nation), continental
(PC is located in the VC’s continent but not in the same nation) or global (PC is located in a
different nation and a different continent). I use an ordered logit model and cluster by VC firm
to account for unobserved intrinsic strategic preference of each VC firm.

I expect that a larger relative supply of investment opportunities in the VC’s nation will
decrease the VC’s investment scope. For example, I postulate that the propensity of a VC to
invest globally decreases when relatively more deals become available in the VC’s nation
(i.e., the density of deals available gets higher in the VC’s nation, in comparison to
international locations). I expect that more intense competition (stronger demand) for
investment opportunities within a VC’s nation should drive VC’s investments towards
continental and global investments. As for the microeconomic factors, I expect VCs investing
from industry-specialized VC funds to be more likely to invest at a distance by engaging in
continental or global investments and that VC funds specialized in industries with a slower
time to market (Life Sciences) should be more likely to invest at a distance than VC funds
specialized in faster industries (IT). Since agency costs in the VC-PC relationship are likely to
be higher in early investment stages and to decrease over time, VC funds specializing in later
stage investments are expected to be more likely to choose broader investment scopes. VC
firms with more capital at their disposal are expected to invest more at a distance. The type of
VC fund (as an indicator of the funds’ sources) is also expected to have an influence on the
decision scope, with funds managed by public investors expected to invest less at a distance
 83

for example. More experienced VC firms (in terms of number of investments until the
observation year) are expected to be more likely to invest at a distance.

The model is calculated for each of the four samples (worldwide, US only, non US and
Germany only), as discussed in section 5.1. Descriptive statistics and the correlation matrix
for each of the four samples are reported respectively in Appendix 14, Appendix 15,
Appendix 16 and Appendix 17. While supply and demand are relatively highly correlated
across samples (especially in the US), the remaining correlation coefficients are relatively
low, indicating that collinearity should not be a concern.

The ordered logit model estimates one equation over all three levels of the investment scope
as discussed in section 5.2.1.
Table 20 presents the ordered logit regression coefficients. For a one unit increase in the
predictor, the investment scope level is expected to change by its respective regression
coefficient in the ordered log-odds scale (while the other variables in the model are held
constant). Table 21 and Table 22 present the marginal effects at the national and global scope
level.
84

Table 20: Refined scope model (coefficients)

(1) (2) (3) (4)

Worldwide US Non-US Germany
VC-PC distance
(1=national
2=regional
3=global)
Relative supply -0.061*** (0.003) -0.011** (0.003) -0.816*** (0.018) -0.438 (0.243)
of deals in VC
nation in
investment year
(%)
Relative -0.011*** (0.002) -0.112*** (0.008) 0.267*** (0.009) -0.426 (0.382)
demand for
deals in VC
nation in
investment year
(%)
Fund is 0.327*** (0.017) -0.038 (0.027) 0.683*** (0.025) 0.242** (0.089)
industry-
specialized (d)
Fund is 0.172*** (0.034) 0.204*** (0.056) 0.041 (0.048) 0.084 (0.157)
specialized in
Life Sciences
(d)
Fund is -0.137*** (0.039) 0.295*** (0.067) -0.292*** (0.049) -0.475 (0.315)
specialized in
non High-Tech
(d)
Fund is stage- -0.609*** (0.025) -0.166*** (0.048) -0.871*** (0.031) -1.218*** (0.125)
specialized (d)
*** *** *** ***
Fund is 0.504 (0.026) 0.614 (0.051) 0.417 (0.031) 0.483 (0.130)
specialized in
growth stage
(d)
Fund is 0.520*** (0.037) 0.143* (0.068) 0.621*** (0.048) 1.385*** (0.220)
specialized in
late stage (d)
Fund size (USD 0.000*** (0.000) 0.000*** (0.000) 0.000 (0.000) 0.001 (0.000)
billion)
Fund type = 0.745*** (0.024) 0.855*** (0.036) 0.698*** (0.036) 1.342*** (0.111)
corporate (d)
Fund type = 0.061*** (0.018) 0.621*** (0.029) -0.181*** (0.024) 0.189 (0.108)
bank (d)
*** ***
Fund type = 0.042 (0.082) 0.353 (0.104) -0.536 (0.126) 0.453 (0.299)
individual (d)
Fund type = -0.749*** (0.054) -1.013*** (0.185) -0.934*** (0.060) -2.183*** (0.471)
public (d)
* ***
Fund type = -0.113 (0.072) 0.320 (0.148) -0.514 (0.088)
unknown (d)
Fund age 0.009*** (0.001) -0.006*** (0.002) 0.016*** (0.002) 0.006 (0.008)
VC age 0.006*** (0.001) 0.003** (0.001) 0.010*** (0.001) -0.068*** (0.009)
VC number of 0.001*** (0.000) 0.001*** (0.000) 0.001*** (0.000) 0.009*** (0.001)
investments to
 85

date
Investment is 0.246*** (0.017) -0.969*** (0.025) 0.931*** (0.022) 1.622*** (0.092)
syndicated (d)
*** ***
PC is in Life -0.024 (0.022) -0.171 (0.038) -0.013 (0.030) -0.818 (0.107)
Sciences
industry (d)
PC is in non -0.043* (0.022) 0.588*** (0.032) -0.468*** (0.030) 0.033 (0.119)
High-Tech
industry (d)
PC industry is -0.087 (0.074) 1.380*** (0.108) -0.693*** (0.104) -0.303 (0.737)
unknown (d)
VC is located in 0.228*** (0.036) 0.867*** (0.040)
Europe (d)
VC is located in 0.269*** (0.040) 1.001*** (0.043)
Asia (d)
*** ***
VC is located in 0.638 (0.075) 1.591 (0.080)
ROW (d)
Date of -0.448*** (0.026) -0.080 (0.062) -0.561*** (0.039) -1.125*** (0.312)
investment
1998-2001
(bubble) (d)
Date of -0.657*** (0.026) -0.390*** (0.079) -0.624*** (0.038) -1.041* (0.478)
investment
2002-2008
(post-bubble)
(d)
Cut1 (constant) 0.540*** (0.047) -4.696*** (0.528) 1.170*** (0.059) -1.628** (0.585)
Cut2 (constant) 1.078*** (0.047) -4.564*** (0.528) 2.152*** (0.059) -0.786 (0.585)
Wald test Cut1 10,238.52 1,068.56 9,830.16 634.36
= Cut2: χ2(1)
p <0.001 <0.001 <0.001 <0.001
Observations 213,452 161,157 52,295 4,286
Log pseudo- -87,407.332 - - -
likelihood 36,342.635 42,171.937 3,190.490
χ2(26) 34,680.877 6,210.813 12,883.799 1,713.442
Pseudo-R2 0.166 0.079 0.133 0.212
p <0.001 <0.001 <0.001 <0.001
Investments by VC firms, 1989-2008
Ordered logit model (scope = 1 if investment is national, scope = 2 if investment is continental, scope = 3 if
investment is global); Coefficients; Dummy variables are marked (d)
References: Fund is specialized in IT, Fund is specialized in early stage, Fund type = independent partnership,
PC is in IT, VC is located in North America, Date of investment 1989-1997
Standard errors in parentheses, * p < 0.05, ** p < 0.01, *** p < 0.001
86

Table 21: Refined scope model (marginal effects at the national level)
(1) (2) (3) (4)
Worldwide US Non-US Germany
VC-PC distance
(1=national
2=regional
3=global)
Relative supply 0.005*** (0.000) 0.000** (0.000) 0.185*** (0.004) 0.104 (0.058)
of deals in VC
nation in
investment year
(%)
Relative 0.001*** (0.000) 0.005*** (0.000) -0.061*** (0.002) 0.101 (0.091)
demand for
deals in VC
nation in
investment year
(%)
Fund is -0.027*** (0.001) 0.002 (0.001) -0.156*** (0.006) -0.058** (0.021)
industry-
specialized (d)
Fund is -0.015*** (0.003) -0.010*** (0.003) -0.009 (0.011) -0.020 (0.038)
specialized in
Life Sciences
(d)
Fund is 0.011*** (0.003) -0.015*** (0.004) 0.064*** (0.010) 0.106 (0.065)
specialized in
non High-Tech
(d)
Fund is stage- 0.047*** (0.002) 0.007*** (0.002) 0.202*** (0.007) 0.281*** (0.027)
specialized (d)
Fund is -0.046*** (0.003) -0.032*** (0.003) -0.096*** (0.007) -0.117*** (0.032)
specialized in
growth stage
(d)
Fund is -0.051*** (0.004) -0.007* (0.003) -0.149*** (0.012) -0.331*** (0.046)
specialized in
late stage (d)
Fund size (USD -0.000*** (0.000) -0.000*** (0.000) -0.000 (0.000) -0.000 (0.000)
billion)
*** *** *** ***
Fund type = -0.079 (0.003) -0.053 (0.003) -0.168 (0.009) -0.323 (0.024)
corporate (d)
Fund type = -0.005*** (0.001) -0.034*** (0.002) 0.041*** (0.005) -0.045 (0.026)
bank (d)
Fund type = -0.003 (0.007) -0.018** (0.006) 0.110*** (0.023) -0.111 (0.075)
individual (d)
Fund type = 0.045*** (0.002) 0.029*** (0.003) 0.179*** (0.009) 0.334*** (0.033)
public (d)
Fund type = 0.009 (0.005) -0.016 (0.009) 0.107*** (0.016)
unknown (d)
*** *** ***
Fund age -0.001 (0.000) 0.000 (0.000) -0.004 (0.000) -0.001 (0.002)
VC age -0.000*** (0.000) -0.000** (0.000) -0.002*** (0.000) 0.016*** (0.002)
VC number of -0.000*** (0.000) -0.000*** (0.000) -0.000*** (0.000) -0.002*** (0.000)
investments to
date
 87

Investment is -0.019*** (0.001) 0.058*** (0.002) -0.202*** (0.004) -0.338*** (0.016)

syndicated (d)
*** ***
PC is in Life 0.002 (0.002) 0.007 (0.002) 0.003 (0.007) 0.183 (0.022)
Sciences
industry (d)
PC is in non 0.003* (0.002) -0.032*** (0.002) 0.102*** (0.006) -0.008 (0.028)
High-Tech
industry (d)
PC industry is 0.007 (0.006) -0.114*** (0.015) 0.138*** (0.017) 0.069 (0.160)
unknown (d)
VC is located in -0.020*** (0.003) -0.185*** (0.008)
Europe (d)
VC is located in -0.024*** (0.004) -0.238*** (0.010)
Asia (d)
*** ***
VC is located in -0.067 (0.010) -0.376 (0.016)
ROW (d)
Date of 0.034*** (0.002) 0.003 (0.003) 0.124*** (0.008) 0.258*** (0.068)
investment
1998-2001
(bubble) (d)
Date of 0.052*** (0.002) 0.017*** (0.003) 0.143*** (0.009) 0.243* (0.107)
investment
2002-2008
(post-bubble)
(d)
Observations 213,452 161,157 52,295 4,286
Log pseudo- -87,407.332 - - -
likelihood 36,342.635 42,171.937 3,190.490
2
χ (26) 34,680.877 6,210.813 12,883.799 1,713.442
Pseudo-R2 0.166 0.079 0.133 0.212
p <0.001 <0.001 <0.001 <0.001
Investments by VC firms, 1989-2008
Ordered logit model (scope = 1 if investment is national, scope = 2 if investment is continental, scope = 3 if
investment is global); Coefficients; Dummy variables are marked (d)
References: Fund is specialized in IT, Fund is specialized in early stage, Fund type = independent partnership,
PC is in IT, VC is located in North America, Date of investment 1989-1997
Standard errors in parentheses, * p < 0.05, ** p < 0.01, *** p < 0.001
88

Table 22: Refined scope model (marginal effects at the global level)

(1) (2) (3) (4)

Worldwide US Non-US Germany
VC-PC distance
(1=national
2=regional
3=global)
Relative supply of -0.003*** (0.000) -0.000** (0.000) -0.114*** (0.002) -0.074 (0.041)
deals in VC nation
in investment year
(%)
Relative demand -0.001*** (0.000) -0.004*** (0.000) 0.037*** (0.001) -0.072 (0.064)
for deals in VC
nation in
investment year
(%)
Fund is industry- 0.017*** (0.001) -0.001 (0.001) 0.099*** (0.004) 0.041** (0.015)
specialized (d)
Fund is 0.009*** (0.002) 0.009*** (0.003) 0.006 (0.007) 0.014 (0.027)
specialized in Life
Sciences (d)
Fund is -0.007*** (0.002) 0.013*** (0.003) -0.037*** (0.006) -0.070 (0.040)
specialized in non
High-Tech (d)
Fund is stage- -0.030*** (0.001) -0.006*** (0.002) -0.131*** (0.005) -0.204*** (0.020)
specialized (d)
*** *** *** ***
Fund is 0.029 (0.002) 0.029 (0.003) 0.060 (0.005) 0.086 (0.024)
specialized in
growth stage (d)
Fund is 0.033*** (0.003) 0.006* (0.003) 0.102*** (0.009) 0.303*** (0.054)
specialized in late
stage (d)
Fund size (USD 0.000*** (0.000) 0.000*** (0.000) 0.000 (0.000) 0.000 (0.000)
billion)
*** *** *** ***
Fund type = 0.051 (0.002) 0.047 (0.003) 0.117 (0.007) 0.282 (0.026)
corporate (d)
*** *** ***
Fund type = bank 0.003 (0.001) 0.030 (0.002) -0.025 (0.003) 0.033 (0.020)
(d)
** ***
Fund type = 0.002 (0.004) 0.016 (0.006) -0.062 (0.012) 0.086 (0.062)
individual (d)
*** *** *** ***
Fund type = -0.028 (0.001) -0.026 (0.003) -0.097 (0.004) -0.195 (0.016)
public (d)
***
Fund type = -0.005 (0.003) 0.014 (0.008) -0.060 (0.008)
unknown (d)
*** *** ***
Fund age 0.000 (0.000) -0.000 (0.000) 0.002 (0.000) 0.001 (0.001)
VC age 0.000*** (0.000) 0.000** (0.000) 0.001*** (0.000) -0.011*** (0.002)
VC number of 0.000*** (0.000) 0.000*** (0.000) 0.000*** (0.000) 0.001*** (0.000)
investments to
date
Investment is 0.012*** (0.001) -0.052*** (0.002) 0.122*** (0.003) 0.226*** (0.011)
syndicated (d)
PC is in Life -0.001 (0.001) -0.006*** (0.001) -0.002 (0.004) -0.124*** (0.014)
Sciences industry
(d)
PC is in non High- -0.002* (0.001) 0.028*** (0.002) -0.060*** (0.004) 0.006 (0.020)
Tech industry (d)
PC industry is -0.004 (0.004) 0.103*** (0.013) -0.077*** (0.009) -0.047 (0.103)
unknown (d)
*** ***
VC is located in 0.012 (0.002) 0.109 (0.005)
 89

Europe (d)
VC is located in 0.015*** (0.003) 0.164*** (0.008)
Asia (d)
*** ***
VC is located in 0.043 (0.007) 0.326 (0.020)
ROW (d)
*** ***
Date of -0.022 (0.001) -0.003 (0.002) -0.074 (0.005) -0.183*** (0.049)
investment 1998-
2001 (bubble) (d)
Date of -0.033*** (0.001) -0.015*** (0.003) -0.089*** (0.006) -0.176* (0.081)
investment 2002-
2008 (post-
bubble) (d)
Observations 213,452 161,157 52,295 4,286
Log pseudo- - - - -
likelihood 87,407.332 36,342.635 42,171.937 3,190.490
2
χ (26) 34,680.877 6,210.813 12,883.799 1,713.442
2
Pseudo-R 0.166 0.079 0.133 0.212
p <0.001 <0.001 <0.001 <0.001
Investments by VC firms, 1989-2008
Ordered logit model (scope = 1 if investment is national, scope = 2 if investment is continental, scope = 3 if
investment is global); Coefficients; Dummy variables are marked (d)
References: Fund is specialized in IT, Fund is specialized in early stage, Fund type = independent partnership,
PC is in IT, VC is located in North America, Date of investment 1989-1997
Standard errors in parentheses, * p < 0.05, ** p < 0.01, *** p < 0.001
90

5.2.3 Discussion

5.2.3.1 General remarks

The ordered logit model is statistically significant at the 1‰-level and shows the effects of the
expected investment scope drivers on the VC’s propensity to invest at a larger scope. A test of
the two cutpoints distinguishing between the three levels of investment scope (national,
continental and global) shows that the difference between national and continental and
continental and global investment scope is significant at the 1‰-level for all four samples.124
This result supports the choice of using the refined scope model over the naïve scope model,
since the naïve scope model assumes that continental and global investment scopes are similar
by considering only national vs. international investment scope.

An overview of the regression coefficients (

Table 20) shows that several hypotheses are supported in all four samples. Funds specialized
in growth or late stage are significantly more likely to invest at a distance than funds
specialized in early-stage investments. Corporate funds are significantly likelier and public
funds significantly less likely to invest at a distance than private partnerships. Finally, a VC’s
propensity to invest at a distance significantly increases with the amount of investment
experience of a VC.

Several hypotheses are supported in some samples but not all. For example, the local
availability of more abundant deal flow is significantly negatively correlated at the 1% to 1‰
significance level with the VC’s likelihood to invest at a distance for US and non-US VC
firms, but the result is not significant for Germany-based firms. Industry-specialized funds are
significantly more likely to invest at a distance than non-industry-specialized funds when
considering the full sample, non-US or German VC firms, but there is no supporting evidence
for US-based VC firms. While US-based VC firms are significantly more likely to invest at a
distance if they are specialized in the Life Sciences industry than in the IT industry, the

124
“Cut1” is the estimated cutpoint on the latent variable used to differentiate narrow scope from medium and
large scope when values of the predictor variables are evaluated at zero. Similarly, “Cut2” distinguishes a
global investment from national or continental based on the underlying latent variable when values of the
predictor variables are evaluated at zero. For example, an observed investment from the full sample
(worldwide) that had a scope value of 0.540 or less on the underlying latent variable that gave rise to the scope
variable would be classified as narrow investment scope (i.e., national investment scope), given they were
committed previously to the bubble (1989-1997), involved a PC in the IT industry receiving non-syndicated
investment from a new VC located in North America, with no previous experience (VC and fund age=0),
investing from a generalist independent partnership fund in a fictive situation in which there would be no
national supply or demand for deals (i.e., with all predictors at zero).
 91

industry of specialization of VC firms has no significant effect on their propensity to invest at

a distance. Finally, the larger the fund managed by a VC firm, the likelier the firm will make
investments at a distance if the VC is located in the US, but fund size has no significant effect
on non-US or Germany-based VC firm’s propensity to invest at a distance.

Certain hypotheses are supported in some samples and not in others. Most of the time, these
discrepancies occur between US and non-US-based VC firms. For example, VCs outside of
the US behave as expected in that they react to increased local competition (increased local
demand for deals) by increasing their propensity to invest at a distance, but US-based VCs
behave in the opposite way, i.e., US-based VC firms are less likely to invest at a distance
when local competition for deals within the US increases. Similarly, while the expected
highly significant positive correlation between investment syndication and investment scope
indicates that non-US VCs use syndication to bridge the distance when investing at a broader
scope, US VCs seem to use syndication for national investments and not for cross-border
investments.

5.2.3.2 Discussion of the marginal effects

The marginal effects at the national and global level are reported in Table 21 and Table 22
and provide information about the quantified effects of each predictor on the investment scope
decision. For example, marginal effects at the national level show the effect of a change in a
predictor by one unit on the likelihood of a VC to invest nationally at the mean, all other
predictors remaining unchanged.

The ancillaries being similar to those of the coefficient table, they do not need to be
commented again. Comparing the effect of each predictor on the VC’s likelihood to invest
nationally with its effect on the VC’s likelihood to invest beyond its continent (global
investment scope), it becomes evident that significance levels are the same and the direction
of the effects contrary. This is intuitive because a factor driving a VC firm to invest nationally
systematically decreases that VC firm’s propensity to invest at the global level, and vice-
versa. However, the amplitude of the effects need not be similar. The following paragraphs
describe the marginal effects at the national and global level.

Supply and demand conditions

The first two predictors portray the role of the VC’s environment in the VC’s decision to
invest nationally, continentally or globally. Supply and demand factors do not have a
92

significant effect on Germany-based VC firms’ investment scope decision, but significantly

impact the scope of US and non-US-based VC firms.

In the samples containing only US VC firms on the one hand and all non-US VC firms on the
other hand, it appears that the “hotter” the VC firm’s nation is in the world at the time of
investment (i.e., the more deals are available to invest in nationally relative to the rest of the
world), the likelier the VC will invest nationally rather than internationally. For example, if
the number of deals available in a given nation (outside of the US) increases from 1% to 2%
relative to all deals available worldwide in a given year, the likelihood that VCs in that nation
choose to invest nationally increases by 18.5% and the likelihood that VCs in that nation
choose to invest globally decreases by 11.4% if all other predictors are left unchanged. The
mean share of national supply available to US-based investors relative to the worldwide deal
flow amounts to 31%. For US-based VCs, an increase of relative deal flow by 1 percent point
from 31% to 32% also has a positive effect on national VCs’ propensity to invest in US-based
ventures and a significant negative effect on their propensity to invest at the global level (1%-
level significance), ceteris paribus, but the size of the effect is minimal (<0.05%). Based on
these results, it seems that VC firms react to their macroeconomic environment as expected
and act rationally in the face of changing supply conditions: VCs realize when their home
market is losing some of its attractiveness on the worldwide scene (less deals proportionally,
in favor of deals outside of the VC’s nation) and they respond by investing more outside of
their nation. Hypothesis 1 is hereby supported in the sub-sample of US-based VC firms only
and in the sample of non-US VC firms, but receives no supporting evidence in the sub-sample
of German VCs.

Hypothesis 2 stating that VCs invest at a distance as an answer to increased local competition
is supported for non-US-based VCs. All other things equal, if the share of the number of VC
firms active in a nation other than the US relative to the number of VC firms active worldwide
increases from 3 to 4%, a VC in that nation would decrease its propensity to invest nationally
by 6.1% and increase its likelihood to invest beyond its continent by 3.7% (1‰ significance
level). On the opposite, if the competition intensity in the US raises from 60% to 61%, ceteris
paribus, US VCs slightly increase their propensity to invest nationally by 0.5% (1%
significance level). It is possible that due to the constant high supply of deals in the US, US-
based VCs must dedicate even more time to compete for the good local deals when new VCs
enter the American market. This could signal that the ability of US VC firms to capture the
best local deals has a signaling effect on their reputation and that they cannot afford the
 93

opportunity cost of investing internationally rather than nationally as a consequence.

Hypothesis 2 is therefore supported for non-US VCs and evidence of the contrary is provided
for US-based VCs. There is no supporting evidence for German VCs.

VC fund characteristics

As expected under hypothesis 3a, industry-specialized VC funds outside of the US in general

or in Germany in particular are significantly likelier to invest at a distance than generalist
funds (the result is not significant for US-based firms). If a fund is industry-specialized, the
likelihood that it is used for national investments is lower by 15.6% in non-US nations (1‰
significance level) and 5.8% in Germany (1%-level significance), and its likelihood to invest
globally is higher by 9.9% in non-US nations in general and by 4.1% in Germany in
comparison to funds that are not industry-specialized. Hypothesis 3b is supported for US VCs
at the 1‰ significance level: given a VC fund is industry-specialized, it is 10% less likely to
be used for national investments and 9% likelier to be used for global investments if it is
specialized in Life Sciences than if it is specialized in the IT industry. According to these
results, industry specialization either “shrinks the world” for VC firms outside of the US, or
industry-specialized VC firms have developed a core competency they can use to exploit their
competitive advantage beyond their borders and to mitigate their liability of foreignness in
their international investments.

Hypothesis 4 receives supporting evidence at least at the 5% significance level in all samples:
funds specialized in later stage investments (growth or late stage) are significantly likelier to
invest at a distance than funds specialized in investments in early-stage ventures. The effects
of late-stage specialization are the strongest in Germany: a Germany-based VC fund
specialized in growth stage is 11.7% less likely to invest within Germany and 30.3% likelier
to invest outside of Europe than an early-stage German fund, all other things being equal. This
supports the assumption that early-stage ventures may require more intense monitoring and
consulting than later-stage ventures, so that it is more practical for VCs to be close to their
early-stage investments, but that the costs of investing at a distance may be lower for later-
stage investments.

There is only partial evidence in support of hypothesis 5 since fund size only has a significant
explanatory power in the US. Should a US fund of mean size (224 million USD) raise one
more million USD, it is significantly less likely to invest nationally and significantly likelier
to invest at the global level, but the size of the effect is very small. In the other samples, fund
94

size has no significant effect. This indicates that the investment scope strategy of non-US VCs
and German VCs are independent of the size of their funds.

Hypothesis 6 is supported in all samples: public funds (from governments, universities, etc.)
are significantly less likely to be used for cross-border investments than funds managed in
independent partnerships. For example, a public German VC fund is 33.4% likelier to be used
to invest in German ventures and 19.5% less likely to be used to invest in non-European
ventures than regular German VC funds. This underlines the assumption that public funds are
not solely motivated by profits, but also pursue the goal to support local entrepreneurship
initiatives. On a side note, corporate VCs managing VC funds are significantly more likely to
invest at the global level than regular VCs (independent partnerships), presumably thanks to
their existing extensive international networks. This corroborates the preliminary results
presented in Table 19 and discussed in section 4.4.3.3.

Support for hypothesis 7 is contradictory. The hypothesis is supported at the 1‰ significance

level in the US: should a US VC fund have been raised one year earlier than the mean (VC
fund age amounts to 6.97 years instead of 5.97 years), its propensity to invest globally
decreases significantly and its propensity to invest nationally increases significantly, but by
less than 0.05%. For non-US VC firms, the effect is the opposite, albeit also weak (lower than
0.5%). The data does not provide the expected proof that German VCs are less likely to invest
internationally as funds mature and approach their expiration date (i.e., that German VCs
become more risk averse as funds come to maturity and less time is available to fructify
investments). The corresponding hypothesis does not receive significant support for German
VCs, which means that German VC firms do not necessarily change their risk preferences in
terms of investment scope over the lifetime of a fund.

VC firm characteristics

Hypothesis 8 cannot be analyzed for lack of information on the size of the observed VC firms
in terms of number of GPs available to manage portfolio companies.

Hypothesis 9 receives partial support with both measures of experience. VCs with more
investment experience are significantly more likely (at least at the 1% significance level) to
invest at a distance than less experienced VCs, but the effects of investment experience are
very weak. For example, a German VC firm with 61 investments is 0.2% less likely to invest
within Germany and 0.1% more likely to invest outside of Europe than when it had made 60
investments, if everything else stays equal. Older VCs are also likelier to invest at a distance
 95

in all samples, except for German VCs. Measuring experience by age returns a contradictory
result for German VCs: if a VC firms was created 8.8 years before the observed investment, it
is likelier to invest in a German VC by 1.6% than if it had been created 7.8 years ago. In other
words, younger German VC firms are likelier to invest at a distance than older German VC
firms. This could indicate that younger German VCs may face fierce competition from
established national VCs and may be forced to look for investment opportunities outside of
Germany at first. Alternatively, younger German VC firms may be more likely to be set up
for cross-border investments than their older counterparts. For example, younger German VC
partners may be likelier to have received an international education than their more seasoned
counterparts, so that they are likelier to have a broader network of international contacts than
older VCs. For US and non-US VCs as a whole, higher levels of investment experience are
significantly positively correlated with the VC’s propensity to invest at a distance, which
provides some support for the hypothesis that VCs can realize economies of scale when
investing at a distance (i.e., the costs of investing at a distance are the highest at first but
decrease with time).

Investment characteristics

Tests of hypothesis 10 return contradictory results. While syndicated investments are

significantly less likely to be investments at the national level and significantly likelier to be
global investments for VC firms outside of the US and in Germany (at the 1‰ significance
level), the opposite is true for US-based VCs. Ceteris paribus, a German VC firm’s propensity
to invest at the global level is 22.6% higher if the investment is syndicated than if the VC firm
invests on its own. On the opposite, US VC firms are 5.8% likelier to invest nationally if the
invest together with another VC firm. These findings indicate that non-US and German VCs
do use syndication to bridge the distance between themselves and their distant portfolio
companies. It is likely that due to the size of their home market (area-wise), US-based VCs
already use syndication to bridge the geographic distance within the country, so that
syndication is not specific to cross-border investments but to actual geographic distance.

5.3 Intermediary conclusion

The hypotheses and the results of the investment scope analysis are summarized in

Table 23.
96

Table 23: Intermediary results (investment scope decision)

Decision Hypothesis # Determinants Expected effect Supporting

group on VC’s evidence125
propensity to
invest at a
distance

Scope Supply and H1 VC firm’s local supply of deals (local - (Yes)

demand deal flow)
Scope Supply and H2 VC firm’s local demand for deals + (Yes)
demand (intensity of local VC competition)
Scope Ability and H3a VC fund is industry-specialized (yes=1; + (Yes)
fit (fund) no=0)
Scope Ability and H3b Time to market in the VC fund industry + (Yes)
fit (fund) of specialization (if applicable)
Scope Ability and H4 VC fund is specialized in later + Yes
fit (fund) investment stage (yes=1; no=0)
Scope Ability and H5 VC fund size + (Yes)
fit (fund)
Scope Ability and H6 VC fund is financed by public investors - Yes
fit (fund) (yes=1; no=0)
Scope Ability and H7 VC fund age - (Yes)
fit (fund)
Scope Ability and H8 VC firm size + N/A126
fit (firm)
Scope Ability and H9 VC firm investment experience + Yes
fit (firm)
Scope Ability and H10 Investment is syndicated (yes=1; no=0) + (Yes)
fit
(investment)

Yes Supporting evidence in all samples

(Yes) Supporting evidence in some samples
No No supporting evidence
N/A Not applicable (hypothesis could not be tested)

125
Hypotheses are tested one four samples: (1) full sample (all observed investments worldwide), (2)
investments by US-based VCs, (3) investments by non US-based VCs and (4) investments by Germany-based
VCs.
126
Firm size (number of GPs) is not available.
 97

6 Target location decision

6.1 Naïve target location model

The target location model aims at comparing the drivers of attractiveness of the areas in which
different potential portfolio companies can be found. The VC firm is expected to choose the
target location that yields the maximum expected returns, all other things remaining equal,
because once the money is committed to the selected venture, it can no longer be used for
another venture. Therefore, the VC makes a decision in favor of a venture located in a certain
area at the expense of investments in ventures located in other areas. The naïve target location
model is pictured in Figure 11.

The naïve target location model could be implemented using the discrete choice model
developed by McFadden (1974) to investigate the importance of various regional and intrinsic
characteristics in determining the location choice made by VCs when choosing to invest in a
portfolio company. The use of a conditional logit model (an extension of the multinomial
logit model) would be appropriate because the VC’s choice behavior includes explanatory
variables characterizing both the attributes of the choice alternatives (e.g., attractiveness of
different target locations) as well as characteristics of the VC firm making the choices (e.g.,
VC experience.). This model is built in analogy to models used by researchers who study the
location decision of firms’ foreign direct investments. For example, Du, Lu and Tao (2008)
use such a model to explore the role of agglomeration economies (network externalities),
government institutions and other factors in determining the choice of location of foreign
invested enterprises in China from US, Europe, Japan and Korea. Such discrete choice models
have also been widely used in marketing analysis, transportation and urban studies (Dardis /
Hrozencik 1985; Kim et al. 2005).

H. Tarrade, Cross-Border Venture Capital Investments, DOI 10.1007/978-3-8349-6939-2_6,

© Gabler Verlag | Springer Fachmedien Wiesbaden 2012
98

Figure 11: Naïve target location model

VC INVESTMENT

TL1 TL2 TL3 ... TLn

The problem with this model is one of econometric nature: in order to implement the
conditional logit model properly, one must assume the independence from irrelevant
alternatives (McFadden 1974). In other words, error terms must be independently distributed.
If the independence of irrelevant alternatives assumption is violated, the conditional logit
parameters are biased and the results cannot be interpreted (Cheng 2007).127 Unfortunately,
this would be the case with the naïve target location model described above, because some
options are closer substitutes than others. In the case of a German VC firm for example,
investing in a venture located in Austria is a closer substitute to investing in a venture located
in Switzerland than investing in an Indian venture. Therefore, it is plausible that the errors
associated to Swiss and Austrian ventures are positively correlated, distorting the results of
the analysis. For this reason, the next section introduces a refined target location model.

6.2 Refined target location model

6.2.1 Econometric model

6.2.1.1 Introduction

The refined target location model copes with this concern by grouping together alternatives
that are likely to be close substitutes. In the case of cross-border investments, these groups (or
“nests”) can be chosen to mirror the three groups selected for the refined model (national
investments, continental investments and global investments), as shown in Figure 12.

127
In theory, other assumptions regarding the distribution of the error terms are possible, but they would lead to
very complex models that cannot easily be implemented, even with a computer-based program.
 99

Figure 12: Refined target location model

VC INVESTMENT

NATIONAL CONTINENTAL GLOBAL

TL1 TL2 TL3 ... TLcn TLcn+1 ... TLgn

This type of model is estimated by a nested logit regression. The nested logit model relaxes
the restrictive independence of irrelevant alternatives property by grouping closely
substitutable alternatives into a subnet (Train 2002). Following the example given by Heiss
(2002) for the choice of travel mode128, one can decompose the choice probabilities into two
parts. The probability of a VC choosing to invest in a certain target location is equal to the
product of the probability that he invests in a target location within that nest and the
conditional probability that he invests exactly in this target location, given this nest is chosen.
For example, the probability of a certain German VC firm to invest in a venture located in
France ”ሺ‫ ݕ‬ൌ ‫݁ܿ݊ܽݎܨ‬ሻ is equal to the probability of that VC choosing a continental
investment, ”ሺ‫݈ܽݐ݊݁݊݅ݐ݊݋ܥ א ݕ‬ሻ, times the conditional probability of investing in a French
start-up given a continental scope is chosen ”ሺ‫ ݕ‬ൌ ‫݁ܿ݊ܽݎܨ‬ȁ‫݈ܽݐ݊݁݊݅ݐ݊݋ܥ א ݕ‬ሻ. Heiss
stresses that “this decomposition is valid in general by the rules of conditional probability, but
it is especially useful for thinking about the nested logit” (2002, 233). Furthermore, as noted
by Mayer and Mucchielli in their study of plant location129, the nested logit model “does not
need the tree structure to correspond to a real sequential process involving different time
periods for each stage of the tree” (2004, 139).

According to Heiss (2002), there are two different nested logit models: the non-normalized
nested logit, which is often more difficult to interpret, and the random utility maximization

128
In this model, the first decision step defines whether the individual chooses between public and other
transportation mode, and the second decision step determines the travel mode: in the first nest, bus and train, in
the second nest, air and car.
129
In this study, the authors split the investment decision between nations on one side and regions within nations
on the other side, and they use a nested logit model to estimate the model.
100

nested logit, which he recommends as the model of choice if there are generic variables.
Indeed, using non-normalized nested logit, the parameters meant to measure the
dissimilarities between the nests are distorted by the different scaling of the coefficient of the
generic variables across nests. For these reasons, I choose to use a random utility
maximization nested logit to estimate the two-step model.

The probability that a VC firm chooses to invest in target location ݆ within scope ݇ (݇ ‫ א‬ሼͳ ൌ
݊ܽ‫݈ܽ݊݋݅ݐ‬ǡ ʹ ൌ ܿ‫݈ܽݐ݊݁݊݅ݐ݊݋‬ǡ ͵ ൌ ݈݃‫݈ܾܽ݋‬ሽ) is:130

”ሺ‫݆݊݋݅ݐܽܿ݋݈ݐ݁݃ݎܽݐ‬ǡ ‫݇݁݌݋ܿݏ‬ሻ ൌ ܲ௝௞ ൌ ܲ௝ȁ௞ ܲ௞

Each part of the equation can be specified as follows:

݁ ఉ௫ೕȁೖ
ܲ௝ȁ௞ ൌ
σ௃௝ୀଵ
ೖ
݁ ఉ௫ೕȁೖ

݁ ఊ௭ೖାఛ಼ூ௏ೖ
ܲ௞ ൌ
σ௄
௞ୀଵ ݁
ఊ௭ೖ ାఛ಼ ூ௏ೖ

where for the kth branch, ‫ܸܫ‬௞ ൌ Ž σ௝ୀଵ

௃
ೖ
݁ ఉ௫ೕȁೖ and where ‫ ݔ‬refers to attributes of the target
locations and ‫ ݖ‬refers to the characteristics of the investment scope (a VC firm makes a
simultaneous investment decision that consists of selecting an investment scope and a target
location). ‫ܸܫ‬௞ is the inclusive value. The inclusive value parameter ߬ is a measure of the
correlation among the random error terms due to unobserved attributes of target location
choices within a nest.131

Nested logit models are fairly new but have found applications in fields of research such as
transportation and internationalization research. In the latter category, nested logit models are
mainly used to identify the drivers of a location choice strategy. For example, Pusterla and
Resmini (2007) study 4,103 manufacturing foreign investments in Bulgaria, Hungary, Poland
and Romania during the 1990’s and Cheng (2007) examines the location decision structure of
Japanese investors in China, finding that foreign investors first choose a Chinese region and
then a specific place to locate their plant. An interesting study by Belderbos and Sleuwaegen
(2005) looks at the spatial configurations of overseas plants for 120 Japanese firms. The
authors use a two-step nested logit model to describe, (1) the firms’ internationalization

130
The specification is adapted from Knapp et al. (2001).
131
For a more detailed description of the specification of the random utility maximization nested logit model, see
Appendix 11.
 101

decision (Asia-bound, West-bound or global) and (2) the choice between international plant
configurations. They use three categories of drivers: competitive drivers (e.g., technology
intensity), firm characteristics (e.g., firm size and experience) and location-specific pull
variables (such as labor cost or market size).

The main disadvantage is that the estimation of model parameters can be computationally
challenging (Suzuki 2007). A second limitation of the model is that it does not give
indications as to the sequence of the two decision steps. For instance, a VC firm could expand
its investment scope because it is presented with a high-quality investment opportunity that is
usually beyond its investment area. Alternatively, a VC firm could decide to expand its
investment scope beyond its usual range (e.g., as a thought-through internationalization
strategy) and therefore now also invest in ventures located further than the majority of its
previous portfolio companies. The model does not allow to differentiate between these two
cases.

The main advantage of the two-step decision model is that it allows to differentiate between
(1) the key drivers pushing venture capital firms to invest nationally, at the continent level or
globally, and (2) the key drivers determining the decision to invest in a venture located in a
particular target location over another target location. Furthermore, studies using two-step
models and comparing them with one-step models find that two-step models fit the data
better. For example, using a two-step nested logit model to examine airline choice, Suzuki
(2007) demonstrates that it is superior to one-step choice models. He argues that travelers pick
an airline in two steps: first, they eliminate certain choices that do not fit their acceptable
standards, and second, they maximize their utility within the pre-selected alternatives. He
shows that this model is significantly better than a classical one-step model. Applying this
concept to the present analysis, the two-step decision model is expected to fit the investment
data at hand better than the naïve target location model.

6.2.1.2 Example of operationalization

Let us illustrate the new target location model for a fictive VC firm located in Germany, as
illustrated in Figure 13132. Investments by this German VC can be located in many target
locations (i.e., in many different nations) across the world. The alternatives of target locations
are grouped in three nests (i.e., branches of the logic tree). The German VC firm can invest in

132
The two-letter country codes in the figure follow the ISO 3166-1 alpha-2 code for country names. The figure
shows the top 10 target locations of German VC investors.
102

Figure 13: Application of the refined target location model to German VC firms

INVESTMENT by
Germany-based VC

CONTINENTAL GLOBAL
NATIONAL
= Europe = beyond Europe

DE UK CH IL FR AT NL SE US CA

a German start-up, in a European start-up outside of Germany, or in a start-up outside of

Europe. The example considers the top 10 target locations of investments by German VC
firms (as introduced in Table 18) and these target locations are partitioned as follows:

x branch 1 includes Germany, i.e., the target location corresponding to a “national”

investment scope, 133
x branch 2 includes the top target locations corresponding to a “continental” investment
scope, i.e., UK, Switzerland, Israel, France, Austria, Netherlands, Sweden and
x branch 3 includes the top target locations corresponding to a “global” investment
scope, i.e., US and Canada.

6.2.2 Estimation

6.2.2.1 Challenges and solutions

First, the estimation of the nested logit model requires that as many observations be created
for each investment case (i.e., for each investment decision) as possible target locations a VC
can theoretically choose from. In this thesis, 10 observations are created per observed
investment, corresponding to the top 10 target locations of the VC firms located in the nation
of interest, as introduced in section 4.4.3.2 (Table 18).134 All observations have similar
characteristics except for the variables relating to the target location. One alternative is chosen
for the investment (the actual observation) and the other 9 are discarded alternatives. The

133
In this model, the first branch is called a “degenerate” branch, because it has only one leave, i.e., choosing this
branch automatically determines the target location choice.
134
Attempts to compute the model with more target locations led to unstable results or analyses not running at
all. Indeed, there is not enough variance for all variables in the cases of “rare” target locations if more possible
target locations are taken into account.
 103

observation a posteriori of the target location choices made by the VCs allows to evaluate the
drivers of utility attached by VCs to each target location under given circumstances.

The second implementation challenge is that it is not possible to run a common model for all
VC firms at once. This becomes clear when attempting to apply the tree model to portray the
target location decision of a VC firm located in another nation than Germany. For example,
the decision tree would be a different one for US-based VC firms: US-American VCs have
the choice between branch 1, which is to invest in a US-American PC, branch 2, which
essentially means to invest in a Canadian PC (North America without the US), or branch 3,
which is the rest of the world (any location outside of North America). Instead, one must
specify a separate tree structure and run the model separately for each VC firm nation, the
largest unit for which the decision model remains consistent among observations. In this
dissertation, I focus the implementation of this third model to investments by VC firms
located in Germany. The rationale for this choice is many-fold:

x First, it makes sense to implement this model to a leading nation in terms of number of
investments, and Germany is one of the nations with the most observed investments in
the sample after the US, UK and France.

x Second, the model should work best for nations that have traditionally invested
significantly at a distance. Among the top 4 investor nations, UK and Germany are
leaders in the share of investments they have made at a distance, ahead of France.

x Third, as specified above, each investment case must be duplicated a posteriori so that
10 fictive alternatives are available to the VC firm to choose from and so that the
utilities can be calculated. For reasons of limited computational power, it is difficult to
impossible to select a nation with too many investment cases. With just over 4,000
observed investments over the analysis period, Germany is a good fit for the test of the
model.

x Fourth, facing a choice between UK and Germany, it is preferable to select the market
I have the better knowledge of, in order to enable better interpretation and discussion
of the results, which is the German market.

The refined target location model is therefore implemented to analyze investments by

Germany-based VC firms between 1989 and 2008. Further, the target locations considered in
the model are restricted to the top 10 target locations of German VC investors, to ensure
sufficient variance among observation cases.
104

6.2.2.2 Construction of additional variables

The data set is completed with additional variables. For each alternative, I indicate the supply
and demand situation in that nation at the time of investment as explained in section 4.3.2.

I add information on the VC firm’s experience in investing in that target location in non-
syndicated deals135.

Since VCs are interested in successfully exiting their grown investments – and an IPO being
the “pinnacle of a successful investment” (Draho 2004, 95), one would expect that a local
trading platform would make a certain location more attractive to investors.136 However, the
presence of a stock exchange being granted in each of the 10 target location alternatives
considered in the sub-sample, this criterion is not a differentiating factor in the target location
decision.

Geographic distance is indicated by the investment scope (national, continental or global) and
is further refined by a binary variable indicating whether the target location has a common
border with Germany.

I model cultural distance in several ways. Each alternative target location is attributed the type
of legal system it is based on (common law or civil law) as reported by the Central
Intelligence Agency’s World Factbook137. This categorization helps identify cultural ties
between options, as the origins and developments of legal systems find their roots in historical
events. It is therefore possible that German VCs exhibit a preference for target locations
following civil law like Germany. Since the sample focuses on investments by German
investors and since expert interviews highlighted that language barriers can play an important
role in the successful development of a portfolio company, I also add a variable indicating
whether the target location uses a Germanic language (such as English or Dutch for example).

Last, I research manually (using information on the VC firms’ websites) whether the VC firm
has an office located in the potential target location. Subsidiary presence in the top 10 target
locations is presented in Table 24. 20.33% of the 246 German VCs observed in the sample
have a subsidiary office outside of Germany. German VCs with a subsidiary office abroad are

135
This is to ensure that the VC firm has real experience investing in the target location and did not simply inject
capital in a deal led by another VC firm.
136
According to a recent survey, the opportunities to achieve exits is one of VC’s main concerns when investing
across borders (Deloitte 2006).
137
https://www.cia.gov/library/publications/the-world-factbook/fields/2100.html (accessed May 2010)
 105

for the most part present in the US (7.32% of the observed German VCs), in Switzerland
(4.47%) and in the UK (2.03%). A few VCs also have an office in France, Austria, the
Netherlands and Sweden. None of the observed PCs have an office in Israel or in Canada.138

Table 24: Office presence of German VCs in foreign top target locations

Subsidiary location Number of German VCs Percentage of VCs

Continental
Switzerland 11 4.47%
UK 5 2.03%
Israel 0 0%
France 3 1.22%
Austria 2 0.81%
Netherlands 1 0.41%
Sweden 1 0.41%
Global
US 18 7.32%
Canada 0 0%

6.2.2.3 Results

The results of the regression are presented in Table 25. Part 1 reports the coefficients at the
target location level (nation of the portfolio company receiving the selected investment) given
the choice of the investment scope, and Part 2 reports the coefficients at the scope level
(nests). National scope is used as the base nest. The results are discussed in the next section.

138
A few VCs more have an office outside of the top 10 target locations (not reported in Table 24). For example,
9 German VCs have a subsidiary in Asia (Singapore, Hong Kong, China, India) and 6 German VCs have an
office in another European nation (Belgium, Denmark, Italy, Spain).
106

Table 25: Target location model (coefficients)

(part 1)

Target location selected for investment (PC nation)

PC nation
Relative supply of deals in PC nation in investment year (%) 0.024** (0.009)
Relative demand for deals in PC nation in investment year (%) 0.006 (0.013)
PC nation and VC nation share a common border (d) 0.429*** (0.090)
VC has an office in PC nation (d) 0.202* (0.096)
PC nation uses a Germanic language (d) 0.155* (0.061)
PC nation follows civil law (d) -0.427*** (0.099)
VC number of solo investments in PC nation 1.058*** (0.082)
Cases 3,996
Observations 39,960
Log pseudo-likelihood -
3,745.782
χ2(45) 1,542.720
p <0.001
Standard errors in parentheses
Investments by Germany-based VC firms in top 10 target locations, 1989-2008
Dummy variables are marked (d)
References: Fund is specialized in IT, Fund is specialized in early stage, Fund type = independent partnership,
PC is in IT, Date of investment 1989-1997
*
p < 0.05, ** p < 0.01, *** p < 0.001
 107

(part 2)

Target location selected for investment

(PC nation)
National Continental Global
Relative supply of deals in VC nation (base) 0.465 (0.243) -0.611** (0.232)
in investment year (%)
Relative demand for deals in VC -0.988*** (0.231) -0.098 (0.273)
nation in investment year (%)
Fund is industry-specialized (d) 0.516*** (0.139) 0.408*** (0.111)
Fund is specialized in Life Sciences (d) -0.202 (0.232) -0.320 (0.210)
Fund is specialized in non High-Tech -0.285 (0.497) -0.719* (0.364)
(d)
Fund is stage-specialized (d) -0.741*** (0.188) -1.713*** (0.176)
Fund is specialized in growth stage (d) 0.612** (0.190) 0.789*** (0.187)
Fund is specialized in late stage (d) 0.069 (0.392) 1.311*** (0.271)
Fund size (USD billion) 0.532 (0.526) 0.512 (0.462)
Fund type = corporate (d) 0.669*** (0.175) 0.803*** (0.154)
Fund type = bank (d) -0.202 (0.179) 0.103 (0.136)
Fund type = individual (d) 0.683 (0.353) 0.103 (0.457)
Fund age 0.000 (0.013) 0.023* (0.010)
VC age -0.032* (0.014) -0.088*** (0.012)
VC number of investments to date 0.007*** (0.001) 0.011*** (0.001)
PC is in Life Sciences industry (d) -0.328* (0.163) -0.956*** (0.141)
PC is in non High-Tech industry (d) -0.437* (0.216) 0.147 (0.141)
Date of investment 1998-2001 0.172 (0.473) -0.574 (0.333)
(bubble) (d)
Date of investment 2002-2008 (post- 1.132* (0.543) -0.400 (0.450)
bubble) (d)
τ (constant) 1.000 0.515*** (0.049) 0.563*** (0.093)
Likelihood-ratio test for independence 52.52
of irrelevant alternatives (τ = 1): χ2(2)
p <0.001
Cases 3,996
Observations 39,960
Log pseudo-likelihood -3,745.782
χ2(45) 1,542.720
p <0.001
Standard errors in parentheses
Investments by Germany-based VC firms in top 10 target locations, 1989-2008
Dummy variables are marked (d)
References: Fund is specialized in IT, Fund is specialized in early stage, Fund type = independent partnership,
PC is in IT, Date of investment 1989-1997
*
p < 0.05, ** p < 0.01, *** p < 0.001
108

6.2.3 Discussion

6.2.3.1 General remarks

A nested logit model was used to test the determinants of German VC’s target location
choice. Both VC characteristics and target location characteristics (intrinsic and in relation to
the VC location) were considered, and PC and investment characteristics were controlled for.
The model is statistically significant at the 1‰ level. Part 1 shows the impact of the predictors
on the VC’s choice of a target location (PC nation) conditional to the investment scope. Part 2
shows the impact of the predictors on the choice of a nest, i.e., on the selection of the
investment scope.

A likelihood-ratio test for independence of irrelevant alternatives shows that continental and
global investment scopes are significantly different from each other at the 1‰ level. The
hypothesis that all cross-border investments can be treated similarly is therefore rejected and
the separation between continental and global investments is justified.

For each nest, the joint distribution of the error terms has an additional parameter ߬
representing a measure of the mutual correlation of the error terms of all alternatives within
this nest.139 The inclusive value parameters are clearly below 1140, which implies that the
model is consistent with random utility maximization. Furthermore, this shows that the nested
logit model is better than the multinomial logit model introduced as the naïve target location
model in section 6.1.141

The sample contains observed investments by German VC firms from 1989 to 2008 in PCs
located in the top 10 target locations for German VC investments. The sample contains 3,996
observed investments (cases), i.e., 39,960 observations.142 The summary of the alternatives
and frequencies in the estimation sample is reported in Table 26: the sub-sample of German
investments matches the results obtained from the full sample of German investments

139
For a given nest, ߬ ൌ ඥͳ െ ߩ with ߩ representing the correlation coefficient. ߬ is the inclusive value
parameter (see paragraph 6.2.1.1 for details) and is sometimes called “dissimilarity parameter” (Heiss 2002).
140
The inclusive value parameter at the national level has been manually constrained to 1 since “national” is a
degenerate branch (the choice of this nest automatically determines the target location choice).
141
The multinomial logit model assumes that the inclusive value parameter is equal to 1 in all nests.
142
Due to small number of observations or no observation for which the investment was made at a global scope
by a public investor or into a PC for which the industry is unknown, those variables were discarded and the
corresponding 290 observations for which they took the value “1” were dropped to allow the model to run.
 109

introduced in Table 18143, so that no bias was introduced when dropping certain
observations144. As detailed in Table 26, German investors in the sub-sample made 41.75% of
their investments in foreign portfolio companies (11.54% in Europe, 30.21% outside of
Europe). National investments in German PCs represent 58.26% of the observed investments.
When investing at a continental scale (in European PCs outside of Germany), German VC
investors selected investments located in Switzerland or in the UK half of the time. France
and Austria were the third and fourth-ranked target locations within Europe, followed by
Netherlands and Sweden each receiving less than 2 investments per year on average over the
observation period (1989-2008). The vast majority of global investments consist in
investments in US-based PCs (1,189 investments, 29.75% of the total observed investments)
and a few global investments went to Canadian PCs.

Table 26: Summary of alternatives in the nested logit model

Alternative Frequency selected Percent selected

Investment scope
National 2328 58.26%
Continental 461 11.54%
Global 1207 30.21%
Target location
Germany (national) 2328 58.26%
Switzerland (continental) 119 2.98%
UK (continental) 113 2.83%
Israel (continental) 71 1.78%
France (continental) 64 1.60%
Austria (continental) 40 1.00%
Netherlands (continental) 35 0.88%
Sweden (continental) 19 0.48%
US (global) 1189 29.75%
Canada (global) 18 0.45%

143
Note that Table 18 reports the number of PCs by PC nationality while Table 26 reports the number of
investments per PC nationality, hence a few minor differences in the ranking of target locations between the
two samples.
144
See footnote 142.
110

6.2.3.2 Discussion of the coefficients145

Level 1: Choice of a nest

Most of the coefficients at the nest level (equivalent to the investment scope choice per
definition of the nests) are similar to the coefficients of the ordered logit model of the German
VCs’ investment scope decision (see
Table 20 and discussion in section 5.2.3) in terms of significant positive or negative effect of
each predictor on the investment scope decision (here equivalent to the nest choice).

In addition to the results of the refined scope model, the hypothesis that an abundance of local
supply in the VC’s nation has a significant negative impact on the decision to invest at a
distance is supported at the global level (1% significance level): the more investment
opportunities are available within Germany in relation to the worldwide supply of investment
opportunities at a given time, the more German VCs will invest nationally instead of globally.
The results of the nested logit model also show that VCs specialized in non High-Tech
industries are less likely to invest at a global scope than at a national scope, in comparison to
VCs specialized in the IT industry. Finally, German VCs seem to be significantly more
inclined to invest at a global scale rather than at a national scale as their fund mature
(significance at the 5% level). This result goes against the hypothesis that VCs become more
risk averse as their funds come to maturation due to the shorter time available to grow the
ventures before the fund closes.

Level 2: Choice of a target location

Supply and demand conditions in target location

According to hypothesis 1 and 2, the supply and demand conditions in a target location are
expected to drive the VC firm’s investments towards or away from that target location.

First, an abundance of investment opportunities in a target location are expected to attract

foreign investors looking for investment targets within a scope including the target location.
This hypothesis is supported by the data (significance at 1%). For example, should a German

145
It is not possible to calculate marginal effects in this setting because the values of some variables within a
group depend on the value of these variables in other observations within this group or in other groups. For
example, relative supply of investment opportunities within the UK at a certain date depends on the supply of
investment opportunities, say, in France (due to the relative aspect of the variable definition), so that supply
within UK cannot be modified keeping all other supply values constant (neither within UK nor across target
locations), which prevents marginal effects from being calculated. For a detailed description of a similar case,
see http://www.stata.com/support/faqs/stat/mfx_unsuit.html. As a consequence, the comments in this chapter
must base on the coefficients and the size of the effects cannot be discussed.
 111

VC decide to invest on a European scale outside of Germany, its propensity to invest in a

company located in the UK significantly increases as the relative number of deals available in
the UK increases. This leads to two conclusions. First, German VCs are able to detect
variations in supply levels in real time. Since 1 out of 5 German VC firms has a subsidiary
outside of Germany146, these VC firms may use their foreign offices obtain information about
potential deals at a distance. German VCs who do not have an office abroad must obtain this
information through other channels than local presence. Perhaps, they benefit from a
knowledgeable international business network from which they can draw information on
investment opportunities abroad. Second, German VCs seem to follow the deal flow, i.e., they
invest where innovation is created and not the other way around.

Second, higher levels of competition intensity in the target location (in terms of relative
number of VC firms competing locally) are expected to decrease the likelihood of a VC to
invest in this target location. The data does not provide significant evidence that this is the
case. In other words, German VCs do not seem to be sensitive to the amount of competition in
a target location when they decide to invest or not in that target location. There are several
possible explanations for that. First, it is possible – although unlikely – that VCs are not able
to assess the amount of competition in the target locations they can choose from. Second, they
may be able to observe the competition intensity but do not take it into account, possibly
because they see local investors as potential partners. Third, the partners through whom VCs
learn about investment opportunities are able to secure the investments upfront, so that
competitive investment offers by local VC firms are unlikely to matter.

VC ability and fit with target location

Hypothesis 4 deals with the role of geographic distance between the VC’s office and the PC’s
location. With hypothesis 4, the headquarters of the VC firm are considered and it is expected
that closer target locations are more likely to be selected for an investment than target
locations located further away from the VC’s home nation. An approximation is made by
looking at the neighbor nations of the VC nation as opposed to target locations without a
common border with the VC nation. In the case of German investors, the pre-selected foreign
target locations with a common border with Germany are Switzerland, France, Austria and the
Netherlands. All other variables being held constant, the data analysis shows that a target
location with a common border with Germany is likelier to receive investments from a

146
See Table 24.
112

German VC than other target locations. This result supports the idea that distance causes
additional costs in VC investments. Either distant investment opportunities are more costly to
discover (and therefore less likely to receive investments) or they are more costly to grow
after receiving an investment.

Hypothesis 5 stipulates that VC firms can use subsidiary offices to bridge the distance
between the headquarter office and the PC office if the PC is located in a target location in
which the VC firm has a subsidiary office. It is therefore expected that all other things being
equal, the likelihood to invest in a target location is higher if the VC firm has a subsidiary
office there than if it does not. There is significant supporting evidence in favor of hypothesis
5 (result significant at the 5% level): the observed German VCs do use subsidiary offices to
invest in PCs located in the nation of the subsidiary office. For example, German VC firms
who have an office in the US are significantly likelier to invest in US-based companies than
German VC firms without an office in the US. However, chronology and causality are not
determined by the data. On the one hand, VCs may open an office in a foreign nation because
they have invested repeatedly in that target location in the past. On the other hand, VCs may
open an office abroad in order to gain a good position to identify and support local PCs in the
future.

Hypothesis 6 pertains to the cultural distance between the VC and the PC. Cultural distance is
described by two variables in the analysis: language and legal system. Language proximity is
shown to be a significant positive predictor of target location choice: given a German VC
chooses to invest in a European venture, the likelihood to choose a target location is
significantly higher (at the 5% level) if that target location uses a Germanic language (e.g.,
UK) than if it does not (e.g., France). On the other hand and contrary to expectations, a
similar legal system is not a positive predictor of the target location decision. Given a certain
investment scope, German VCs are likelier to invest in PCs located in nations following
common law than in nations following civil law like Germany (result significant at the 1‰
level). It may be that common law is more favorable to company creation and VC investments
than civil law, thus attracting investments from German investors despite the cultural
distance. Hypothesis 6 is therefore only partially supported.

Under the last hypothesis, the likelihood that a VC decides to invest in a certain target
location should increase with the amount of previous investment experience the VC has with
this target location, ceteris paribus. The multivariate analysis supports this hypothesis. Given
an investment scope, the more German VCs have gained investment experience in a certain
 113

target location147, the likelier they are to invest in this target location again. This result
supports the hypothesis that VCs can realize economies of scale by investing repeatedly in a
target location, but that the entry costs (first investments in a new target location) are high.
Alternatively, VC firms may have an unobserved affinity with one target location148, and the
VC firm may leverage this advantage by investing in the target location. Since there is no
information about the partners in the data, it is not possible to support this conclusion at this
point.

6.3 Intermediary conclusion

The hypotheses and the results of the target location analysis presented in this chapter are
summarized in Table 27.

147
In syndicated investments, German VCs may have invested in a distant target location without having actively
participated in the growth process of the distant PC. In such cases, they would not have gained significant
experience with the target location. Therefore, only solo investments are taken into account in the
measurement of investment experience in a target location.
148
For example, a general partner is originally from that nation or has studied and worked there for an extended
period of time.
114

Table 27: Intermediary results (target location decision)

Decision Hypothesis Determinants Expected effect on Supporting

group VC’s evidence149
propensity to invest in
target location

Target Supply and H1 Target location’s supply of deals + Yes

location demand (deal flow)
Target Supply and H2 Target location’s demand for deals - No
location demand (intensity of VC competition)
Target Supply and H3 Target location’s attractiveness of + N/A150
location demand exit channels
Target Ability and H4 Target location’s geographic - Yes
location fit distance to VC firm + Yes
Target location and VC
nation share a common border
Target Ability and H5 VC firm has office presence in target + Yes
location fit location (yes=1; no=0)
Target Ability and H6 Target location’s cultural distance to - (Yes)
location fit VC firm + No
Target location uses a + Yes
similar legal system151
Target location uses a
linguistically close language152
Target Ability and H7 VC firm’s amount of investment + Yes
location fit experience in target location

Yes Supporting evidence

(Yes) Supporting evidence under some implementation of the underlying determinant
No No supporting evidence
N/A Not applicable (hypothesis could not be tested)

149
Hypotheses are tested on a sub-sample of investments made by Germany-based VC firms in the top 10 target
locations for German investments.
150
The presence of a dynamic stock exchange is given in all 10 target locations and is therefore not a
differentiating determinant.
151
For the test of this hypothesis on German investments, civil law = 1 and common law = 0.
152
For the test of this hypothesis on German investments, Germanic language = 1 and other language = 0.
 115

7 Conclusion

7.1 Summary
The objective of this thesis was to improve the current state of understanding of VC firms’
behavior by investigating their decision making process in making investments. While
industry reports show that cross-border investments have gained importance over the last
years, little scholarly research has investigated the drivers of this phenomenon up to now.
This thesis positions itself as a first step towards a better understanding of why VC firms
invest a distance and provides elements of answer on two aspects of the decision to invest at a
distance. First, the dissertation examines the predictors of VC investment scope (i.e., whether
VCs invest at the national, continental or global level). Second, the thesis investigates the
drivers of VC investment target location choice.

Chapter 2 examined the potential benefits and costs of VC cross-border investments from a
scholarly perspective. On the one hand, VCs have the possibility to benefit from investing
across borders, as suggested by international management and finance theories. For example,
VC firms may seek to increase performance, reduces risk through portfolio diversification,
grow, exploit or gain new competitive advantages or benefit from the specific advantages of a
certain target location. On the other hand, VCs face additional costs when investing at a
distance, as dictated by agency theory, because they must carefully select and help their
portfolio companies successfully grow to maturity until investment exit to achieve the
expected investment returns and this process is expected to be more difficult at a distance. As
a result of this balance, VCs decide whether or not to invest at a distance and in which target
location to invest.

Based on the extensive theory and literature review as well as expert interviews, hypotheses
are developed in Chapter 3. VCs’ decision to invest at a distance or to invest in a certain target
location is expected to be motivated by external factors, in particular the supply of and
demand for investment opportunities, and by intrinsic factors characterizing the VC firm, the
VC fund, and the investment type.

H. Tarrade, Cross-Border Venture Capital Investments, DOI 10.1007/978-3-8349-6939-2_7,

© Gabler Verlag | Springer Fachmedien Wiesbaden 2012
116

The large longitudinal of worldwide VC investments used in this thesis is described in

Chapter 4. The data set is mostly drawn from ThomsonONE.com’s VentureXpert module and
contains 227,441 venture capital investments made by 7,601 VC firms worldwide in 50,443
portfolio companies worldwide from 15,820 VC funds between January 1st, 1989 and
December 31st, 2008. The descriptive statistics provide information about the composition of
the data sets as well as first indications of different preferences in terms of invest scope and
investment target location choice. The nation is selected as the unit of VC and PC location.

Drivers of VC investment scope are studied in Chapter 5. They are tested with an ordered
logit model indicating the effects of the predictors on the VC firm’s propensity to invest
nationally, continentally (at the VC firm’s continent level but outside the VC firm’s own
nation) or globally (beyond the VC firm’s continent). The model is applied to four samples:
the complete sample (investments by VCs worldwide), investments by US-based VC firms
only, investments by all non US-based VC firms and finally investments by Germany-based
VC firms. Most hypotheses receive supporting evidence (either in all samples or in selected
samples). For non US-based VC firms, the main drivers of cross-border investments were 1)
Investment syndication 2) Lack of local supply of deals 3) Funds not managed by public
investors 4) Fund industry specialization 5) Late investment stage 6) Local competition
intensity (VC’s local demand for local deals). A similar result is obtained for German VC
firms, except that local supply and demand are not significant predictors of cross-border
investments. The analysis also highlights important differences in the behavior of US and non
US VC firms, justifying the separate analysis of US and non US-based VC firms again a
posteriori. For example, while an increase in local competition increases the propensity of non
US VC firms to invest at a distance, it decreases the propensity of US VCs to invest across
borders.

The results of the investment scope model show that VCs invest at a distance to guarantee
growth as a reaction to local (national) environmental conditions: a decrease in local supply of
investments generally increases VCs’ propensity to invest in an international PC. Also, VCs
tend to invest more at a distance when they can leverage their specific industry advantage
abroad: fund industry specialization significantly increases VCs’ propensity to invest at a
distance. The analysis further provides evidence that VCs specialized in later investment
stages (growth or late stage) are significantly more likely to invest at a distance than VCs
specialized in an early investment stage: this indicates that the additional agency costs of
investing at a distance decrease as PCs gain experience and establish themselves on the
 117

market. There is also support for the hypothesis that VCs can realize economies of scale,
because higher levels of VC investment experience positively and significantly correlates with
the VC’s propensity to invest at a distance. Finally, as expected from the discussion on
agency theory, VCs can decrease the cost of investing at a distance by using syndication and
seeking the help of another VC firm located closer to the common portfolio company. US
VCs use syndication to invest nationally too, presumably due to the large size of their nation.

Drivers of target location choice are studied in Chapter 6. They are tested with a two-step
nested logit model on a sub-sample of 3,996 investments by Germany-based investors in the
top 10 investment target locations of German VC firms. Target locations were grouped in
three nests, corresponding to the investment scope analyzed in the previous chapter. The
determinants of target location choice were expected to include target location specificities as
well as measures of fit between the investor and the target location. First, I find that German
VCs are attracted by nations with a high level of investment supply, suggesting that VC
investments follow deal flow and not the opposite, which is consistent with Wadwha (2008)
and Guler and Guillen (2004). On the other hand, German VC firms were not sensitive to the
competition intensity in the target location.153 Second, the results provide evidence that VCs
seek geographic proximity to their portfolio companies: a common border with Germany
significantly increases the likelihood of a nation to be chosen for an investment. This provides
supporting evidence that investing at a distance is costly for VC firms and that investment
costs increase with distance. Having a subsidiary office abroad helps VCs bridge the distance
with PCs in the nation of the subsidiary office. Third, results pertaining to cultural distance
are contradictory. On the one hand, language proximity is found to be a significant predictor
of target location choice. On the other hand, a similar legal system is not a positive predictor
of the target location choice. On the opposite, ceteris paribus, German VC firms are likelier to
invest in target locations following a common law legal system than a civil law system.
Fourth, the analysis provides further support for the supposition that VC firms realize
economies of scale when investing at a distance: a solo investment in a target location
significantly increases the VC’s propensity to invest again in this same target location.

153
For a discussion of these results, see section 6.2.3.2.
118

7.2 Managerial implications

This thesis studies the determinants of geographical scope and choice of investment location
in VC investments using a longitudinal comprehensive data set of VC investments made
around the world over the past 20 years. I propose and test an extensive set of hypotheses
pertaining to supply and demand for attractive deals on the one hand and to intrinsic VC firm
abilities on the other hand. For these reasons, this study contributes to the yet scarce coverage
of cross-border investments in the academic context. In addition, it has managerial
implications for practitioners, in particular for LP investors, VC investors, entrepreneurs and
policy makers.

First, a detailed evaluation of VCs’ investment strategy (scope and target location) is
necessary for LPs to fit their own risk profile and diversification needs (Gupta / Sapienza
1992): the framework developed in this study can therefore help them to better analyze how
VCs differ in their selection of a geographic scope and investment target locations. For
example, LPs should realize that seeking to diversify industry risks by investing in generalist
VC firms is likely to lead to a reduction of their portfolio diversification in the geographic
dimension.

Second, VC firms should understand how to develop a sensible international strategy that fits
their own capabilities and that allows them to either increase performance or reduce
investment risk. This thesis provides elements of answer as to how they can do so using cross-
border investments. For example, syndication was identified as a means to bridge distance
between VC and PC and could be used by young VCs as a first step towards systematic
investment internationalization.

Third, entrepreneurs should be aware of VCs’ preferences in terms of scope and investment
location. The results of this study can help them develop a sound understanding of VCs’
strategic choices in order to approach the right VCs, saving time and raising their chances of
obtaining funding (Gupta / Sapienza 1992). For example, an application of the results would
be that Swiss entrepreneurs should not limit their investor search to Swiss investors, but
should also consider German investors who are open to investments in Europe. Indeed,
Switzerland is both geographically and culturally close to Germany, so that a German VC’s
propensity to invest in a Swiss venture should be higher than the VC’s propensity to invest in
another target location within Europe outside of Germany.
 119

Fourth, by understanding the determinants of VCs’ investment scope preferences and choice
of target location, policy-makers may attract VC investments to their area. For example, they
can develop programs specially targeted at attracting those VCs whose profile corresponds to
the determinants signaling readiness to invest in the policy maker’s home nation. These
implications may be particularly relevant to markets with limited availability of domestic
venture capital. For example, policy-makers in a certain nation may provide grants or tax
breaks to German VC firms ready to open a subsidiary office in their nation, since a
subsidiary office is positively correlated with investments in the subsidiary’s location.

7.3 Limitations of the thesis

An important limitation of this study is that some of the suggested determinants cannot be
directly observed and quantified (e.g., actual deal flow received by VC firms from national
ventures, attractiveness of a target location). As a consequence, proxy variables were used to
infer these unknown values. Although these were selected with care and backed by expert
interviews and existing literature, some information necessarily lacks accuracy, reducing the
reliability of the results. Taking the example of deal flow, it would be interesting to have
access to the full deal flow received by a VC firm over an extended period of time, so that the
criteria used for the selection of the target location become evident. Similarly, geographic
distance has been defined based on the headquarters of the VC firms and of the PCs. As such,
the VC-PC distance may be overestimated in the case either organization has another closer
office. The lack of information on subsidiary offices prevented a more precise definition of
distance.154 Furthermore, the definition of investment scope at the nation level also causes
some limitations (e.g., Silicon Valley – New York is treated in the same way as Munich –
Berlin in the analysis, i.e., as within national scope). It would be interesting to see if the
results are robust to another definition of distance and of investment scope, for example in
terms of travel time.155

Secondly, the study of the determinants of VCs’ target location decision focuses on the
analysis of investments made by VC firms based in Germany. It is possible that the findings
cannot be generalized perfectly to VCs based in other nations, since the study of the scope
decision already highlighted important disparities between US and Germany-based VC firms.

154
Except for German VCs, for which subsidiary offices were research by hand.
155
The rationale for defining distance and investment scope at the nation level despite these limitations is
explained in section 4.3.1.
120

In addition, the marginal effects could not be calculated, so that effect sizes could not be
discussed. As a consequence, the main levers of the target location decision could not be
identified, so that managerial implications and recommendations could not be formulated with
great precision.

Thirdly, the analysis does not take into account VC firm characteristics at the management
level or at the VC fund investor level, either because this information was not available, or
because it was not possible to match it with the investment observations. Instead, VC firms
were supposed to be homogeneous and were considered as the entity being the decision
maker. However, it is possible that a VC firm’s international investment preferences may be
influenced by LPs without it being observed in this study. Also, characteristics of General
Partners (e.g., nationality, international education, etc.) may play a role in the VC firm’s
cross-border investment activities156. The investigation of these aspects requires a dedicated
study.

7.4 Avenues for further research

Reviewing his recent survey of the global VC industry, Mark Heesen, president of the US-
based National Venture Capital Association, concludes: “We will see more globalization in
the next decade” (Deloitte / NVCA 2009). International venture capital and cross-border
investments in particular should therefore receive increased attention from scholars to follow
and assist this important evolution of the VC industry. This thesis goes a step in this direction
by focusing on the determinants of VC investment scope and investment target location
choice. A few avenues for research are already introduced in the discussion of the limitations
of the thesis in the previous section. Further, researchers are warmly invited to expand both on
the deal sourcing aspects and on the consequences of the decision to pursue cross-border
investments, for VC firms as well as for ventures receiving funding.

For instance, this study simply assumes the observation of the deal flow, but its origination –
especially for remote locations – was not investigated. While it is understood that deal flow
stems mainly from the VC’s own business network (e.g., syndication partners,

156
Studies show that individual venture capitalists are prone to biases when selecting portfolio companies. For
example, Franke, Gruber, Harhoff and Henkel (2006) show that VCs are biased in favor of ventures whose
management has a comparable academic and professional background. Entrepreneur nationality could
therefore well be another source of VC selection bias. Other papers also indicate that VC firms may have a
particular affinity to a certain target location due to characteristics of the decision makers (e.g., Mayer / Flynn
1973; Simpson / Kujawa 1974; Reid 1981).
 121

recommendation by venture partners), deal sourcing and assessment of investment projects

has been found to differ significantly between VCs of different nations, such as UK,
Netherlands, Belgium and France for example (Manigart et al. 1997). These differences are
not accounted for here but may be interesting to investigate in further detail in future studies.

Secondly, mimetic market entry behavior among VC firms was not taken into account in this
study.157 It is possible that VCs invest in a certain target location because other VCs do so. In
this case, it would be interesting to understand which firms play the role of leaders and which
firms are followers. Furthermore, such a study could also highlight the existence of a potential
early or late mover advantage in entering a foreign target location.

Thirdly, the internationalization process of VC firms has not yet been studied in detail. This
thesis already gives some insight on whether investing at a distance is appropriate for a VC
firm (scope decision, ability and fit) and what determinants indicate a VC’s readiness to invest
in a given target location based on their characteristics (target location decision, ability and
fit). However, the pace of internationalization and the organizational forms chosen by
internationalizing VC firms is beyond the frame of this study and would require a separate
study.

Last but not least, the question of the value added by distant VCs is still open: How is the
interaction between VC and portfolio company affected by distance? How can VCs support
foreign start-ups entering their market? Some elements of answers have already been
provided by a few authors (e.g., MacMillan et al. 1988; Lerner 1995; Mäkelä / Maula 2005;
Ortgiese 2007), but findings are sometimes contradictory and more research is required to
understand exactly how VCs can add value portfolio companies situated at a distance. In
particular, a study of the effects of distance on performance (at the investment and at the
portfolio level) could show how well distant investments perform and whether a performance
difference can be detected between different target locations.158 Interest for this particular

157
Greve (2000) provides an example of how to include interorganizational imitation as a determinant of market
niche entry decisions by using data on Tokyo banks between 1894 and 1936. Other interesting sources for
such a study include a studies on herding (e.g., Aizenman / Kendall 2008), studies of multinational enterprises
(e.g., Caves 2007) or studies of the export decision process (e.g., Simpson / Kujawa 1974).
158
In such a study, the authors could seek to quantify how much “better” than local venture a distant venture
should be in order to make up for the extra cost and risk incurred by the VC firm when investing at a distance.
Since data on venture application is rarely available, a conjoint analysis with partners from VC firms could
provide relevant understanding of this phenomenon – conjoint analyses have already successfully been used in
the analysis of VCs’ investment decision, e.g. by Muzyka et al. (1996) and by Franke et al. (2006; 2008).
However, the VCs interviewed in preparation to this study mentioned that such a phenomenon may not exist in
practice, and that there is no difference in the requirements imposed to distant and local PCs.
122

topic has already been shown by a few authors. For example, Balcarcel (2004) presents a first
analysis of cross-border investments by US VC firms and shows that distant investments are
less successful than local investments. Another example is given by Meyer and Shao (1995)
and Lossen (2006) who investigate the role of geographic diversification on portfolio
performance. With the data at hand, it is possible to deliver a much more granular research of
the performance differences at the investment level (i.e., between local and distant
investments or between different target locations), as well as at the portfolio level. Such an
extension of the study would not only enrich the still scarce literature on the topic but it would
also be of high relevance for LPs, VCs and entrepreneurs, the first two in their investment
strategy, the latter in their selection of the right VC investors.
 123

Appendix

H. Tarrade, Cross-Border Venture Capital Investments, DOI 10.1007/978-3-8349-6939-2,

© Gabler Verlag | Springer Fachmedien Wiesbaden 2012
124

Appendix 1: Overview of interviewed VCs

VC firm General Office location Other offices Firm geographical

Partner investment focus
Alta Partners Jean Deleage San Francisco, CA, - Worldwide
US
Atlas Venture Axel Bichara159 Boston, MA, US London, UK US and Europe
RRE Ventures Stuart Ellmann New York, NY, US - US (focus on NY-region)
Siemens Venture Dr. Ralf Munich, Germany US Worldwide
Capital Schnell160 Asia
Sofinnova Partners Chika Paris, France - Europe
Yoshinaga
Star Ventures Dr. Meir Barel Herzelia, Israel Germany Worldwide
US
Target Partners Waldemar Jantz Munich, Germany - Germany, Austria,
Switzerland
TVM Capital Dr. Alexandra Munich, Germany Boston, MA, Worldwide
Goll US
Wellington Rolf Dienst Munich, Germany London, UK Europe
Partners Palo Alto, CA,
US

159
Mr. Bichara’s experience was collected at the 2009 edition of the VC Panel Discussion organized by the
LMU Entrepreneurship Center (LMU EC, Ludwig-Maximilians-Universität München), the Center for Digital
Technology and Management (CDTM, Ludwig-Maximilians-Universität München and Technische Universität
München) and the Institute for Innovation Research, Technology Management and Entrepreneurship (INNO-
tec, Ludwig-Maximilians-Universität München) under the theme “Startup Financing – US and German
Approaches Compared” at Amerika Haus on Tuesday, May 5 th, 2009.
160
Idem.
 125

Appendix 2: Overview of the data collected from ThomsonONE.com

Each type of observation is gathered in a separate data set.

The table lists the information collected from ThomsonONE.com. 161
The variables used as identifiers are underlined, key variables linking data sets together are in italics.

VC firms (7,601 observations)

vc_name VC firm name
vc_type VC firm type
vc_city VC firm city (HQ)
vc_state VC firm state code (HQ)
vc_nation VC firm nation (HQ)
vc_capital VC firm reported capital under management (USD million)
vc_nbfunds VC firm number of funds
vc_amt Total known amount invested by VC firm in all PCs (USD thousand)
vc_nbpcs Number of PCs in which VC firm invested
vc_founded VC firm founding date
vc_status VC firm investment status
VC funds (15,820 observations)
fund_name Fund name
vc_name VC firm name
fund_type Fund type
fund_size Fund size (USD million)
fund_stage Fund stage focus
fund_yr Fund year
fund_amt Total known amount invested by fund in all PCs (USD thousand)
fund_nbpcs Number of PCs in which fund invested
Investments (227,441 observations)
pc_name PC name
vc_name VC firm name
inv_round_date Investment round date
pc_ipo_date PC IPO date, if applicable
fund_name VC fund name
inv_new New investment (or follow-on)
inv_round_nb Investment round number
inv_deal Investment deal value (USD million)
inv_round_usd Investment round amount disclosed (USD million)
inv_round_usd_e Investment round amount estimated (USD million)
inv_nbinv Number of investors in investment round
pc_stage PC stage at round date
Portfolio companies (PCs) (50,766 observations)
pc_name PC name
pc_city PC city (HQ)
pc_state PC state code (HQ)
pc_nation PC nation (HQ)
pc_ind PC industry sub-group
pc_status PC current public situation
pc_ipo_date PC IPO date, if applicable
pc_founded PC founding date
pc_descr PC business description
pc_nbinv Total number of PC investors
pc_amt Total known amount invested in PC (USD million)
pc_vcs VC firms invested in PC
Limited partners (LPs) (7,906 observations)
lp_name LP name
lp_type LP type
lp_city LP city
lp_state LP state
lp_nation LP nation
lp_assets LP total assets under management (USD million)
lp_alt_assets LP total alternative assets under management (USD million)
lp_inv_vc LP VC fund commitments (USD million)
lp_inv_lbo LP LBO fund commitments (USD million)
lp_inv_mezz LP mezzanine fund commitments (USD million)
lp_inv_fof LP fund of fund commitments (USD million)
lp_inv_other LP other PE fund commitments (USD million)
lp_inv_dir LP direct fund commitments (USD million)
lp_inv_pe LP total PE fund commitments (USD million)

161
This table only contains the raw information obtained from ThomsonONE.com and does not include the
additional variables created or researched manually by the author.
126

Appendix 3: Industry classification based on Venture Economics Industry Codes

Primary assigned industry Included industry sub-groups

IT Computer hardware
Communications
Computer software
Computer (other)
Semiconductors / Electronics
Internet-specific

Life Sciences Medical / Health

Biotechnology

Non High-Technology Business services

Agriculture / Forestry / Fish
Transportation
Construction
Industrial / Energy
Financial services
Utilities
Consumer-related
Manufacturing
Other
 127

Appendix 4: Classification of nations by continents

North America

Canada US

Europe

Austria Georgia Luxembourg Slovenia

Belarus Germany Macedonia Spain
Belgium Greece Malta Sweden
Bosnia Herzegovina Hungary Moldova Switzerland
Bulgaria Iceland Monaco Turkey
Channel Islands Ireland Netherlands Ukraine
Croatia Israel Norway UK
Cyprus Italy Poland
Czech Republic Kazakhstan Portugal
Denmark Kyrgyzstan Romania
Estonia Latvia Russia
Finland Liechtenstein Serbia and Montenegro
France Lithuania Slovakia

Asia

Australia Indonesia Pakistan Taiwan

Bangladesh Japan Philippines Thailand
China Malaysia Singapore Vietnam
Hong Kong New Zealand South Korea
India North Korea Sri Lanka

ROW

Afghanistan Colombia Kuwait Senegal

Algeria Comoros Lebanon Sierra Leone
Angola Costa Rica Macau Somoa
Argentina Cuba Madagascar South Africa
Armenia Dominica Mali St Lucia
Azerbaijan Dominican Rep. Mauritania Swaziland
Bahamas Ecuador Mauritius Tanzania
Bahrain Egypt Mexico Togo
Belize El Salvador Morocco Tonga
Benin Fiji Mozambique Trinidad and Tobago
Bermuda French Guiana Netherlands Antilles Tunisia
Bolivia Gabon Nicaragua Uganda
Botswana Ghana Niger Uruguay
British Virgin Islands Guatemala Nigeria United Arab Emirates
Brazil Honduras Oman Venezuela
Burkina Faso Ivory Coast Panama Zambia
Cameroon Jamaica Papua N Guinea
Cayman Islands Jordan Peru
Chile Kenya Saudi Arabia

Only observed nations (location of at least one VC or PC involved in an observed investment) are listed.
128

Appendix 5: Number of observed investments by VC location (1989-2008)

North America Total 178,513

US 175,109 Canada 3,404
Europe Total 37,206
UK 9,559 Czech Republic 148
France 6,726 Russia 116
Germany 5,735 Hungary 109
Finland 1,998 Luxembourg 80
Israel 1,576 Greece 72
Netherlands 1,551 Channel Islands 65
Belgium 1,523 Romania 52
Sweden 1,416 Iceland 22
Spain 1,207 Ukraine 17
Denmark 982 Slovakia 11
Switzerland 947 Turkey 10
Austria 944 Latvia 7
Ireland 628 Slovenia 7
Norway 546 Estonia 5
Italy 392 Lithuania 5
Portugal 383 Bulgaria 3
Poland 361 Cyprus 3
Asia Total 10,916
South Korea 3,033 New Zealand 129
Australia 2,230 Malaysia 98
Japan 1,178 Thailand 41
India 1,120 Philippines 29
Singapore 966 Vietnam 20
Taiwan 907 Indonesia 13
Hong Kong 863 Pakistan 5
China 280 Sri Lanka 4
Rest of the world Total 774
Brazil 364 Togo 8
South Africa 78 United Arab Emirates 8
Nigeria 52 Botswana 7
Mauritius 36 Madagascar 7
Bermuda 35 Nicaragua 7
Cameroon 33 Bahamas 6
Bahrain 29 Ghana 6
Tunisia 21 Swaziland 6
Costa Rica 18 Chile 4
Cayman Islands 14 Mexico 4
Argentina 11 Ecuador 1
Netherlands Antilles 9 Kuwait 1
Morocco 8 Trinidad and Tobago 1
Total 227,409
Investments for which the VC location is unknown are not included.
 129

Appendix 6: Location of observed VCs (by nation)

VC location (HQ) Number of observed VCs

North America
US 4,256
Canada 206
Europe
UK 482
France 271
Germany 246
Sweden 136
Israel 97
Spain 94
Netherlands 87
Switzerland 85
Italy 70
Belgium 66
Denmark 57
Finland 52
Austria 44
Norway 40
Ireland 39
Portugal 27
Poland 20
Other (Europe) 114
Asia
Japan 160
Australia 144
South Korea 114
China 111
Hong Kong 99
India 97
Singapore 69
Taiwan 63
Malaysia 24
Other (Asia) 41
ROW
Brazil 45
South Africa 28
Other (ROW) 83
Total 7,567
VCs for which the nation is unknown are not shown in this graph (N=3).
Only nations with over 10 observed VCs are listed, others are aggregated.
130

Appendix 7: Number of VC firms participating in observed financing rounds

Number of VC firms participating Number of observed Share of observed

in observed financing round investment rounds investment rounds
1 49,478 45.4%
2 23,640 21.7%
3 13,234 12.1%
4 8,381 7.7%
5 5,212 4.8%
6-10 7,866 7.2%
>10 1,117 1.0%
Total 108,928 100.0%
Financing rounds with more than one VC firm participating are syndicated financing rounds.
 131

Appendix 8: Location of observed portfolio companies (by nation)

PC location (HQ) Number of observed investments

North America
US 162,859
Canada 2,919
Europe
UK 8,460
France 6,739
Germany 3,981
Israel 1,868
Sweden 1,705
Netherlands 1,357
Denmark 1,116
Finland 1,113
Other (Europe) 7,239
Asia
South Korea 3,053
Australia 2,288
India 2,098
China 2,094
Other (Asia) 3,098
ROW
Other (ROW) 1,498
Total 213,485
Nations in which PCs received less than 1,000 investments are aggregated by world continent.
Investments for which the PC location is unknown are not included.
132

Appendix 9: Number of observed financing rounds received by portfolio companies

Number of rounds of financing received by PC Frequency Percent

1 24,918 51.30%
2 9,028 18.59%
3 4,871 10.03%
4 3,089 6.36%
5 2,061 4.24%
6 1,674 3.45%
7 968 1.99%
8 654 1.35%
9 442 0.91%
10 266 0.55%
11 208 0.43%
12 129 0.27%
13 86 0.18%
14 61 0.13%
15 46 0.09%
16 23 0.05%
17 9 0.02%
18 10 0.02%
19 5 0.01%
20 5 0.01%
21 6 0.01%
22 4 0.01%
23 3 0.01%
24 1 0.00%
25 1 0.00%
27 1 0.00%
Total 48,569 100.00%
 133

Appendix 10: Number of observed investments per industry

PC industry Number of observed Share of observed

investments investments
IT 134,938 59.33%
Non High-Technology 32,896 14.46%
Life Sciences 44,385 19.51%
Unknown 15,222 6.69%
Total 227,441 100.00%
134

Appendix 11: Specification of the random utility maximization nested logit model 162

The utility assigned by VC firm ݅ to the target location ݆ has a deterministic part and a
stochastic part:

ܷ௜௝ ൌ ܸ௜௝ ൅ ߳௜௝

The probability ܲ௜௝  that VC firm ݅ chooses target location ݆ is equal to the probability of ܷ௜௝
being the largest of all ܷ௜௝ ,…,ܷ௜௃ . With ‫ݕ‬௜ ሼͳǡ ǥ ǡ ‫ܬ‬ሽ denoting the alternative that decision
maker ݅ chooses, this probability is:

ܲ௜௝ ൌ ”ሺ‫ݕ‬௜ ൌ ݆ሻ ൌ ”൫ܷ௜௝ ൐ ܷ௜௞ ‫ ݇׊‬ൌ ͳǡ ǥ ǡ ‫ܬ‬ǣ ݇ ് ݆൯

ൌ ”ሺ߳௜௞ െ ߳௜௝ ൑ ܸ௜௝ െ ܸ௜௞ ‫ ݇׊‬ൌ ͳǡ ǥ ǡ ‫ܬ‬ǣ ݇ ് ݆

The deterministic part of the utility ܸ௜௝  can be written as:

ܸ௜௝ ൌ ߙ௝ ൅ ࢞௜௝ ߚ ൅ ࢠ௜ ߛ௝

where ࢠ௜ is the vector of VC firm-specific variables describing characteristics of the decision
maker, ߛ௝ is the parameter vector for each alternative ݆ associated with the individual-specific
variables, ࢞௜௝ is the matrix of alternative-specific variables varying both over individuals and
alternatives and ߚ is the parameter for each alternative.

The target location choices available to the German VC firm are now partitioned into 3
subsets (nests): ‫ܤ‬௠ , ݉ ‫ א‬ሼͳǡʹǡ͵ሽ, with ‫ܤ‬ଵ ൌ ƒ–‹‘ƒŽǡ ‫ܤ‬ଶ ൌ ܿ‫݈ܽݐ݊݁݊݅ݐ݊݋‬ǡ ‫ܤ‬ଷ ൌ ݈݃‫݈ܾܽ݋‬, so
that each target location belongs to exactly one nest. The nest to which alternative ݆ belongs is
denoted ‫ܤ‬ሺ݆ሻ. Considering the top 10 target locations of investments by German investors,
following nests are built:

‫ܤ‬ଵ ൌ ሼ‫ݕ݊ܽ݉ݎ݁ܩ‬ሽ,
‫ܤ‬ଶ ൌ ሼܷ‫ܭ‬ǡ ܵ‫݈݀݊ܽݎ݁ݖݐ݅ݓ‬ǡ ‫݈݁ܽݎݏܫ‬ǡ ‫݁ܿ݊ܽݎܨ‬ǡ ‫ܽ݅ݎݐݏݑܣ‬ǡ ܰ݁‫ݏ݈݀݊ܽݎ݄݁ݐ‬ǡ ܵ‫݊݁݀݁ݓ‬ሽ,
‫ܤ‬ଷ ൌ ሼܷܵǡ ‫ܽ݀ܽ݊ܽܥ‬ሽ

The probability of the VC firm ݅ choosing alternative target location ݆, ”ሺ‫ݕ‬௜ ൌ ݆ሻ, is equal to
the product of the probability to choose some alternative in nest ‫ܤ‬ሺ݆ሻ, ”ሼ‫ݕ‬௜ ‫ܤ א‬ሺ݆ሻ}, and the
conditional probability to choose exactly alternative ݆, given some alternative in the same nest
‫ܤ‬ሺ݆ሻ is chosen ”ሼ‫ݕ‬௜ ൌ ݆ȁ‫ݕ‬௜ ‫ܤ א‬ሺ݆ሻ}; that is, ܲ௝ ൌ ”ሺ‫ ݕ‬ൌ ݆ሻ ൌ ”ሼ‫ ݕ‬ൌ ݆ȁ‫ܤ א ݕ‬ሺ݆ሻሽ Ǥ ”ሼ‫א ݕ‬
‫ܤ‬ሺ݆ሻሽ. For example, the probability of a German VC firm to invest in a venture located in

162
The model specification is adapted from Heiss (2002).
 135

France ”ሺ› ൌ  ”ƒ…‡ሻ is equal to the probability of that VC choosing a continental

investment ”ሼ› ‫  א‬ଶ ሽ times the conditional probability of investing in a French start-up
given a continental scope is chosen ”ሼ› ൌ  ”ƒ…‡ȁ› ‫  א‬ଶ ሽ.

The random utility maximization nested logit model allows the alternatives within a nest to
have mutually correlated error terms. For each nest ݉ ൌ ሼͳǡʹǡ͵ሽ, the joint distribution of the
error terms has an additional parameter ߬௠ that represents a measure of the mutual correlation
of the error terms of all alternatives within this nest. ߬௠ ൌ ඥͳ െ ߩ௠ with ߩ௠ representing the
correlation coefficient.

The utilities are rescaled by the inverse of the dissimilarity parameter ߬ሺ݆ሻ for this nest163:

݁ ௏ೕ Τఛೕ
”ሼ‫ ݕ‬ൌ ݆ȁ‫ܤ א ݕ‬ሺ݆ሻሽ ൌ
σ௞‫א‬஻ሺ௝ሻ ݁ ௏ೖ Τఛೕ

The denominator in this equation represents a scaled measure of the attractiveness of the nest
‫ܤ‬ሺ݆ሻ. The log of this expression for each nest ݉ is called inclusive value ‫ܸܫ‬௠ . It corresponds
to the expected value of the utility VC firm ݅ obtains from the alternatives in nest ݉:

‫ܸܫ‬௠ ൌ ݈݊ ෍ ݁ ௏ೖ Τఛ೘
௞‫א‬஻೘

so that:

݁ ఛሺ௝ሻூ௏ሺ௝ሻ
”ሼ‫ܤ א ݕ‬ሺ݆ሻሽ ൌ
σଷ௠ୀଵ ݁ ఛ೘ ூ௏೘

163
Thanks to the normalization, the utilities in each nest are scaled by a similar factor and can therefore be
compared across nests.
136

Appendix 12: Naïve scope model (base model)

(1) (2) (3) (4)

Worldwide US Non-US Germany
PC is in Life -0.002 (0.006) -0.009 (0.005) 0.012 (0.018) -0.169* (0.080)
Sciences industry
(d)
PC is in non High- -0.014 (0.008) 0.061*** (0.012) -0.169*** (0.016) -0.191* (0.089)
Tech industry (d)
PC industry is -0.023 (0.021) 0.224*** (0.067) -0.237*** (0.023) -0.222* (0.110)
unknown (d)
VC is located in 0.341*** (0.032) -0.034 (0.052)
Europe (d)
VC is located in 0.371*** (0.033) -0.045 (0.052)
Asia (d)
VC is located in 0.414*** (0.068) 0.026 (0.078)
ROW (d)
*** ***
Date of 0.022 (0.011) 0.052 (0.009) -0.182 (0.037) -0.377*** (0.114)
investment 1998-
2001 (bubble) (d)
Date of 0.016 (0.013) 0.044*** (0.010) -0.191*** (0.049) -0.223 (0.129)
investment 2002-
2008 (post-
bubble) (d)
Observations 213,452 161,157 52,295 4,286
Log pseudo- - - - -
likelihood 72,552.486 35,096.307 33,925.769 2,777.548
χ2(8) 502.230 234.164 149.738 65.849
Pseudo-R2 0.157 0.029 0.027 0.049
p <0.001 <0.001 <0.001 <0.001
Investments by VC firms, 1989-2008
Logit model (scope = 1 if VC and PC are located in different nations, scope = 0 if VC and PC are located in the
same nation); Marginal effects; Standard errors in parentheses
(d) for discrete change of dummy variable from 0 to 1
References: PC is in IT, VC is located in North America, Date of investment 1989-1997
*
p < 0.05, ** p < 0.01, *** p < 0.001
 137

Appendix 13: Naïve scope model (full model)

(1) (2) (3) (4)

Worldwide US Non-US Germany
Relative supply -0.458*** (0.099) -0.048 (0.036) -18.019*** (1.840) -9.496 (6.400)
of deals in VC
nation (in
investment year)
Relative demand -0.183** (0.061) -0.488*** (0.078) 5.519*** (1.039) -9.510 (10.902)
for deals in VC
nation (in
investment year)
Fund is industry- 0.026*** (0.006) -0.002 (0.006) 0.148*** (0.020) 0.090 (0.075)
specialized (d)
Fund is 0.017 (0.012) 0.010 (0.010) 0.017 (0.031) 0.021 (0.106)
specialized in
Life Sciences (d)
Fund is -0.011 (0.011) 0.015 (0.012) -0.059 (0.041) -0.091 (0.157)
specialized in
non High-Tech
(d)
Fund is stage- -0.051*** (0.009) -0.007 (0.006) -0.207*** (0.031) -0.258*** (0.076)
specialized (d)
Fund is 0.054*** (0.013) 0.032** (0.011) 0.107*** (0.032) 0.121 (0.079)
specialized in
growth stage (d)
Fund is 0.055*** (0.016) 0.007 (0.009) 0.154*** (0.042) 0.288** (0.110)
specialized in
late stage (d)
Fund size (USD 0.000 (0.000) 0.000*** (0.000) 0.000 (0.000) 0.000 (0.000)
billion)
*** *** *** ***
Fund type = 0.076 (0.016) 0.052 (0.015) 0.153 (0.041) 0.302 (0.085)
corporate (d)
Fund type = 0.000 (0.012) 0.033* (0.014) -0.043 (0.026) 0.040 (0.103)
bank (d)
Fund type = 0.006 (0.039) 0.018 (0.042) -0.094* (0.047) 0.130 (0.197)
individuals (d)
Fund type = -0.048*** (0.011) -0.029* (0.012) -0.180*** (0.038) -0.332*** (0.070)
public (d)
Fund type = -0.014 (0.021) 0.016 (0.023) -0.108* (0.055)
unknown (d)
Fund age 0.001 (0.001) -0.000 (0.000) 0.002 (0.002) 0.000 (0.006)
VC age 0.001* (0.000) 0.000 (0.000) 0.003** (0.001) -0.015 (0.009)
VC number of 0.000* (0.000) 0.000* (0.000) 0.000*** (0.000) 0.002** (0.001)
investments to
date
Syndicated 0.014** (0.005) -0.058*** (0.008) 0.179*** (0.014) 0.326*** (0.039)
investment (d)
PC is in Life -0.001 (0.004) -0.007 (0.004) 0.004 (0.014) -0.177** (0.056)
Sciences industry
(d)
PC is in non -0.001 (0.006) 0.032*** (0.006) -0.085*** (0.015) -0.005 (0.096)
High-Tech
industry (d)
138

PC industry is -0.005 (0.018) 0.112*** (0.031) -0.113*** (0.032) -0.002 (0.183)

unknown (d)
VC is located in -0.012 (0.015) 0.079 (0.042)
Europe (d)
VC is located in -0.017 (0.015) 0.085 (0.050)
Asia (d)
VC is located in 0.008 (0.028) 0.203** (0.074)
ROW (d)
Date of -0.036*** (0.010) -0.003 (0.006) -0.113*** (0.026) -0.229* (0.099)
investment 1998-
2001 (bubble)
(d)
Date of -0.055*** (0.014) -0.016* (0.007) -0.131*** (0.032) -0.196 (0.130)
investment 2002-
2008 (post-
bubble) (d)
Observations 213,452 161,157 52,295 4,286
Log pseudo- - - - -
likelihood 67,553.834 33,056.136 29,089.222 2,144.944
χ2(26) 1,003.328 1,287.263 602.312 395.965
Pseudo-R2 0.215 0.085 0.166 0.266
p <0.001 <0.001 <0.001 <0.001
Investments by VC firms, 1989-2008
Logit model (scope = 1 if VC and PC are located in different nations, scope = 0 if VC and PC are located in the
same nation); Marginal effects; Standard errors in parentheses
(d) for discrete change of dummy variable from 0 to 1
References: Fund is specialized in IT, Fund is specialized in early stage, Fund type = independent partnership,
PC is in IT, VC is located in North America, Date of investment 1989-1997
*
p < 0.05, ** p < 0.01, *** p < 0.001
 139

Appendix 14: Descriptive statistics and correlation matrix (VC firms worldwide)
# Variable Mean SD 1 2 3 4 5 6 7

1 Relative supply of deals in VC nation (in investment year) 0.24 0.14 1

2 Relative demand for deals in VC nation (in investment year) 0.46 0.26 0.95 1
3 Fund is industry-specialized 0.49 0.02 0.07 1
4 Fund is specialized in IT 0.38 0.06 0.11 0.81 1
5 Fund is specialized in Life Sciences 0.07 0.00 0.02 0.28 -0.22 1
6 Fund is specialized in non High-Tech 0.03 -0.12 -0.13 0.18 -0.14 -0.05 1
7 Fund is stage-specialized 0.39 -0.28 -0.30 -0.01 -0.03 -0.05 0.13 1
8 Fund is specialized in early stage 0.10 -0.12 -0.14 0.01 -0.01 0.05 -0.01 0.42
9 Fund is specialized in growth stage 0.24 -0.20 -0.21 0.00 0.02 -0.08 0.07 0.70
10 Fund is specialized in late stage 0.05 -0.07 -0.06 -0.03 -0.08 -0.01 0.16 0.29
11 Fund size (USD million) 218.32 533.78 -0.02 0.01 -0.06 -0.04 -0.03 0.00 0.01
12 Fund type = independent 0.72 0.16 0.16 0.07 0.05 0.04 -0.01 -0.14
13 Fund type = corporate 0.07 -0.02 -0.01 0.15 0.15 0.02 -0.03 0.06
14 Fund type = bank 0.18 -0.12 -0.14 -0.15 -0.14 -0.05 0.03 0.09
15 Fund type = individuals 0.01 0.01 0.01 -0.02 0.00 -0.02 -0.01 0.01
16 Fund type = public 0.02 -0.12 -0.13 -0.05 -0.06 0.00 0.02 0.07
17 Fund type = other 0.01 -0.04 -0.04 -0.05 -0.04 -0.02 0.00 0.01
18 Fund age 5.68 6.54 0.07 0.07 -0.12 -0.08 -0.04 -0.04 -0.09
19 VC age 13.44 12.40 0.03 0.03 -0.14 -0.10 -0.06 -0.03 -0.08
20 VC number of investments to date 182.04 307.36 0.10 0.15 -0.07 -0.03 -0.04 -0.08 -0.16
21 VC relative number of national investments to date 0.83 0.26 0.55 0.58 0.05 0.07 -0.01 -0.05 -0.13
22 VC is located in North America 0.76 0.88 0.95 0.10 0.13 0.03 -0.14 -0.29
23 VC is located in Europe 0.18 -0.71 -0.76 -0.10 -0.14 -0.02 0.11 0.22
24 VC is located in Asia 0.06 -0.41 -0.45 -0.01 -0.02 -0.03 0.05 0.16
25 VC is located in ROW 0.00 -0.11 -0.12 0.01 -0.02 -0.01 0.09 0.05
26 Date of investment 1998-2001 (bubble) 0.34 -0.05 -0.01 0.11 0.17 -0.08 -0.05 0.10
27 Date of investment 1989-1997 (pre-bubble) 0.20 0.46 0.30 -0.18 -0.16 -0.06 0.00 -0.11
28 Date of investment 2002-2008 (post-bubble) 0.46 -0.33 -0.23 0.04 -0.04 0.12 0.05 0.00
29 Syndicated investment 0.78 0.19 0.23 0.10 0.13 0.06 -0.17 -0.14
30 PC in IT industry 0.59 0.09 0.12 0.24 0.47 -0.31 -0.19 -0.05
31 PC in Life Sciences industry 0.20 0.05 0.04 -0.09 -0.32 0.48 -0.08 -0.06
32 PC in non High-Tech industry 0.14 -0.08 -0.12 -0.15 -0.23 -0.10 0.36 0.10
33 PC industry unknown 0.07 -0.13 -0.14 -0.12 -0.10 -0.02 -0.01 0.05
140

Descriptive statistics and correlation matrix (VC firms worldwide) (continued)

# Variable 8 9 10 11 12 13 14 15 16 17

8 Fund is specialized in early stage 1

9 Fund is specialized in growth stage -0.19 1
10 Fund is specialized in late stage -0.08 -0.13 1
11 Fund size (USD million) -0.07 0.05 0.02 1
12 Fund type = independent -0.02 -0.14 -0.01 -0.03 1
13 Fund type = corporate -0.02 0.09 0.00 -0.07 -0.44 1
14 Fund type = bank -0.02 0.10 0.02 0.11 -0.74 -0.13 1
15 Fund type = individuals 0.03 -0.02 0.01 -0.03 -0.15 -0.03 -0.04 1
16 Fund type = public 0.15 -0.01 -0.02 -0.04 -0.22 -0.04 -0.06 -0.01 1
17 Fund type = other -0.02 0.03 -0.01 -0.01 -0.14 -0.02 -0.04 -0.01 -0.01 1
18 Fund age -0.09 -0.03 -0.01 -0.04 -0.11 -0.03 0.09 0.03 0.02 0.20
19 VC age -0.09 -0.02 -0.01 0.23 -0.01 -0.07 0.05 -0.02 -0.01 0.06
20 VC number of investments to date -0.09 -0.10 -0.03 0.39 0.10 -0.08 -0.02 -0.04 -0.06 -0.02
21 VC relative number of national investments to date 0.01 -0.14 -0.04 -0.12 0.11 -0.06 -0.09 0.02 0.00 -0.02
22 VC is located in North America -0.13 -0.22 -0.05 0.02 0.14 0.00 -0.12 0.01 -0.12 -0.04
23 VC is located in Europe 0.09 0.16 0.05 0.02 -0.06 -0.05 0.08 0.00 0.07 0.02
24 VC is located in Asia 0.09 0.12 0.00 -0.07 -0.16 0.09 0.08 -0.02 0.11 0.04
25 VC is located in ROW 0.03 0.03 0.02 -0.01 0.00 -0.01 0.01 -0.01 0.01 0.00
26 Date of investment 1998-2001 (bubble) -0.02 0.17 -0.10 -0.01 -0.06 0.09 0.03 -0.01 -0.04 0.03
27 Date of investment 1989-1997 (pre-bubble) -0.03 -0.07 -0.06 -0.09 0.06 -0.06 -0.01 -0.02 -0.05 -0.01
28 Date of investment 2002-2008 (post-bubble) 0.04 -0.10 0.14 0.08 0.01 -0.04 -0.02 0.02 0.08 -0.03
29 Syndicated investment -0.09 -0.08 -0.04 -0.05 0.05 0.05 -0.05 -0.01 -0.09 -0.01
30 PC in IT industry -0.02 0.00 -0.08 -0.02 0.04 0.08 -0.08 0.01 -0.05 -0.01
31 PC in Life Sciences industry 0.04 -0.08 -0.03 -0.02 0.02 -0.04 0.00 -0.02 0.02 -0.01
32 PC in non High-Tech industry -0.02 0.07 0.11 0.03 -0.05 -0.05 0.08 0.01 0.03 0.02
33 PC industry unknown 0.01 0.03 0.04 0.03 -0.04 -0.02 0.06 -0.01 0.02 0.01

# Variable 18 19 20 21 22 23 24 25 26 27

18 Fund age 1
19 VC age 0.36 1
20 VC number of investments to date 0.23 0.47 1
21 VC relative number of national investments to date 0.01 -0.05 0.02 1
22 VC is located in North America 0.08 0.04 0.16 0.53 1
23 VC is located in Europe -0.04 0.00 -0.10 -0.45 -0.82 1
24 VC is located in Asia -0.07 -0.06 -0.11 -0.22 -0.46 -0.12 1
25 VC is located in ROW -0.03 -0.05 -0.04 -0.06 -0.12 -0.03 -0.02 1
26 Date of investment 1998-2001 (bubble) -0.11 -0.06 -0.06 0.00 -0.01 -0.02 0.05 0.00 1
27 Date of investment 1989-1997 (pre-bubble) -0.01 -0.03 -0.13 0.12 0.19 -0.16 -0.08 -0.03 -0.36 1
28 Date of investment 2002-2008 (post-bubble) 0.11 0.08 0.16 -0.10 -0.15 0.15 0.02 0.03 -0.66 -0.47
29 Syndicated investment 0.04 0.02 0.08 0.09 0.25 -0.18 -0.13 -0.08 0.04 -0.03
30 PC in IT industry -0.03 -0.03 0.04 0.09 0.12 -0.12 -0.03 -0.02 0.18 -0.07
31 PC in Life Sciences industry 0.01 0.01 0.01 0.02 0.04 -0.01 -0.05 -0.02 -0.12 0.05
32 PC in non High-Tech industry 0.00 0.00 -0.07 -0.04 -0.14 0.11 0.06 0.06 -0.08 0.08
33 PC industry unknown 0.03 0.04 0.00 -0.16 -0.11 0.09 0.05 0.00 -0.07 -0.05
 141

Descriptive statistics and correlation matrix (VC firms worldwide) (continued)

# Variable 28 29 30 31 32

28 Date of investment 2002-2008 (post-bubble) 1

29 Syndicated investment -0.01 1
30 PC in IT industry -0.12 0.13 1
31 PC in Life Sciences industry 0.07 0.08 -0.60 1
32 PC in non High-Tech industry 0.01 -0.23 -0.50 -0.20 1
33 PC industry unknown 0.11 -0.05 -0.32 -0.13 -0.11

Pearson correlation coefficients (for two continuous variables)

Point biseral coefficients (for one continuous variable and one dummy variable)
Phi coefficients (for two dummy variables)

Correlations stronger than .4 are marked in bold.

(N=227,408)
142

Appendix 15: Descriptive statistics and correlation matrix (US-based VC firms)

# Variable Mean SD 1 2 3 4 5 6 7

1 Relative supply of deals in VC nation (in investment year) 0.31 0.06 1

2 Relative demand for deals in VC nation (in investment year) 0.60 0.04 0.82 1
3 Fund is industry-specialized 0.52 -0.22 -0.23 1
4 Fund is specialized in IT 0.42 -0.19 -0.18 0.83 1
5 Fund is specialized in Life Sciences 0.08 -0.08 -0.11 0.28 -0.25 1
6 Fund is specialized in non High-Tech 0.02 0.03 0.02 0.13 -0.11 -0.04 1
7 Fund is stage-specialized 0.31 -0.02 -0.03 0.02 0.01 -0.03 0.12 1
8 Fund is specialized in early stage 0.08 0.03 -0.02 0.02 -0.02 0.06 0.01 0.44
9 Fund is specialized in growth stage 0.18 -0.02 0.02 0.03 0.05 -0.08 0.06 0.72
10 Fund is specialized in late stage 0.04 -0.07 -0.09 -0.03 -0.06 -0.01 0.14 0.32
11 Fund size (USD million) 224.43 488.95 -0.13 -0.15 -0.04 -0.03 -0.03 0.03 0.01
12 Fund type = independent 0.76 0.01 -0.01 0.04 0.02 0.04 0.01 -0.12
13 Fund type = corporate 0.07 -0.06 -0.04 0.16 0.15 0.03 -0.02 0.10
14 Fund type = bank 0.15 0.04 0.06 -0.13 -0.11 -0.05 0.00 0.04
15 Fund type = individuals 0.01 -0.03 -0.03 -0.03 -0.02 -0.03 0.00 0.02
16 Fund type = public 0.01 -0.03 -0.05 -0.05 -0.05 0.00 0.01 0.05
17 Fund type = other 0.00 0.02 0.02 -0.04 -0.03 -0.02 0.00 0.07
18 Fund age 5.97 6.74 -0.01 -0.06 -0.11 -0.08 -0.05 -0.02 -0.05
19 VC age 13.70 11.91 -0.01 -0.07 -0.11 -0.07 -0.07 -0.01 -0.08
20 VC number of investments to date 213.49 327.20 -0.16 -0.23 -0.07 -0.04 -0.04 -0.06 -0.12
21 VC relative number of national investments to date 0.92 0.13 0.09 0.11 0.06 0.08 -0.01 -0.07 -0.07
22 VC is located in North America 1.00 . . . . . . .
23 VC is located in Europe 0.00 . . . . . . .
24 VC is located in Asia 0.00 . . . . . . .
25 VC is located in ROW 0.00 . . . . . . .
26 Date of investment 1998-2001 (bubble) 0.34 -0.14 -0.06 0.11 0.16 -0.08 -0.03 0.11
27 Date of investment 1989-1997 (pre-bubble) 0.25 0.80 0.87 -0.24 -0.21 -0.08 0.03 -0.06
28 Date of investment 2002-2008 (post-bubble) 0.42 -0.57 -0.70 0.11 0.03 0.15 0.00 -0.05
29 Syndicated investment 0.84 -0.11 -0.08 0.09 0.10 0.04 -0.13 -0.05
30 PC in IT industry 0.63 -0.10 -0.06 0.23 0.46 -0.34 -0.15 -0.01
31 PC in Life Sciences industry 0.20 0.04 0.01 -0.11 -0.35 0.49 -0.06 -0.04
32 PC in non High-Tech industry 0.12 0.12 0.11 -0.17 -0.20 -0.09 0.30 0.07
33 PC industry unknown 0.05 -0.02 -0.04 -0.07 -0.06 -0.01 -0.01 0.01
 143

Descriptive statistics and correlation matrix (US-based VC firms) (continued)

# Variable 8 9 10 11 12 13 14 15 16 17

8 Fund is specialized in early stage 1

9 Fund is specialized in growth stage -0.14 1
10 Fund is specialized in late stage -0.06 -0.10 1
11 Fund size (USD million) -0.06 0.03 0.05 1
12 Fund type = independent 0.02 -0.15 0.00 -0.09 1
13 Fund type = corporate -0.02 0.13 0.00 -0.07 -0.49 1
14 Fund type = bank -0.05 0.08 0.00 0.18 -0.73 -0.11 1
15 Fund type = individuals 0.04 -0.02 0.02 -0.04 -0.18 -0.03 -0.04 1
16 Fund type = public 0.14 -0.03 -0.01 -0.04 -0.17 -0.03 -0.04 -0.01 1
17 Fund type = other -0.02 0.10 -0.01 -0.01 -0.13 -0.02 -0.03 -0.01 -0.01 1
18 Fund age -0.08 0.00 -0.02 -0.10 -0.15 -0.02 0.13 0.04 0.02 0.24
19 VC age -0.08 -0.04 -0.01 0.17 -0.03 -0.10 0.10 -0.02 -0.02 0.10
20 VC number of investments to date -0.07 -0.09 -0.01 0.37 0.04 -0.09 0.04 -0.05 -0.05 -0.01
21 VC relative number of national investments to date 0.06 -0.10 -0.06 -0.23 0.14 -0.04 -0.15 0.00 0.03 0.00
22 VC is located in North America . . . . . . . . . .
23 VC is located in Europe . . . . . . . . . .
24 VC is located in Asia . . . . . . . . . .
25 VC is located in ROW . . . . . . . . . .
26 Date of investment 1998-2001 (bubble) -0.04 0.21 -0.09 0.01 -0.08 0.08 0.05 -0.02 -0.04 0.04
27 Date of investment 1989-1997 (pre-bubble) 0.01 -0.05 -0.05 -0.13 0.02 -0.06 0.04 -0.02 -0.03 0.00
28 Date of investment 2002-2008 (post-bubble) 0.04 -0.16 0.13 0.10 0.06 -0.02 -0.08 0.04 0.07 -0.04
29 Syndicated investment -0.05 -0.02 -0.02 -0.06 0.01 0.04 -0.02 -0.01 -0.06 0.00
30 PC in IT industry -0.03 0.03 -0.05 -0.02 0.02 0.07 -0.06 0.00 -0.04 0.01
31 PC in Life Sciences industry 0.06 -0.08 -0.02 -0.03 0.02 -0.04 0.00 -0.02 0.03 -0.02
32 PC in non High-Tech industry -0.01 0.06 0.07 0.05 -0.03 -0.05 0.06 0.02 0.02 0.01
33 PC industry unknown -0.03 0.00 0.05 0.04 -0.02 -0.01 0.03 0.00 -0.01 0.00
144

Descriptive statistics and correlation matrix (US-based VC firms) (continued)

# Variable 18 19 20 21 22 23 24 25 26 27

18 Fund age 1
19 VC age 0.35 1
20 VC number of investments to date 0.21 0.46 1
21 VC relative number of national investments to date -0.01 -0.03 -0.08 1
22 VC is located in North America . . . . .
23 VC is located in Europe . . . . . .
24 VC is located in Asia . . . . . . .
25 VC is located in ROW . . . . . . . .
26 Date of investment 1998-2001 (bubble) -0.09 -0.04 -0.05 -0.01 . . . .
27 Date of investment 1989-1997 (pre-bubble) -0.03 -0.04 -0.18 0.09 . . . . -0.41 1
28 Date of investment 2002-2008 (post-bubble) 0.11 0.08 0.21 -0.07 . . . . -0.60 -0.49
29 Syndicated investment 0.01 0.02 0.05 0.10 . . . . 0.03 -0.09
30 PC in IT industry -0.02 -0.01 0.03 0.11 . . . . 0.17 -0.11
31 PC in Life Sciences industry 0.00 -0.01 0.00 0.03 . . . . -0.13 0.05
32 PC in non High-Tech industry 0.01 0.00 -0.05 -0.07 . . . . -0.07 0.12
33 PC industry unknown 0.02 0.02 0.02 -0.18 . . . . -0.03 -0.02

# Variable 28 29 30 31 32

28 Date of investment 2002-2008 (post-bubble) 1

29 Syndicated investment 0.05 1
30 PC in IT industry -0.06 0.08 1
31 PC in Life Sciences industry 0.08 0.05 -0.66 1
32 PC in non High-Tech industry -0.04 -0.18 -0.48 -0.19 1
33 PC industry unknown 0.05 0.00 -0.29 -0.11 -0.08

Pearson correlation coefficients (for two continuous variables)

Point biseral coefficients (for one continuous variable and one dummy variable)
Phi coefficients (for two dummy variables)

Correlations stronger than .4 are marked in bold.

(N=168,653)
 145

Appendix 16: Descriptive statistics and correlation matrix (non US-based VC firms)
# Variable Mean SD 1 2 3 4 5 6 7

1 Relative supply of deals in VC nation (in investment year) 0.01 0.01 1

2 Relative demand for deals in VC nation (in investment year) 0.03 0.02 0.71 1
3 Fund is industry-specialized 0.40 -0.14 -0.13 1
4 Fund is specialized in IT 0.27 -0.08 -0.09 0.75 1
5 Fund is specialized in Life Sciences 0.06 -0.06 -0.03 0.30 -0.15 1
6 Fund is specialized in non High-Tech 0.07 -0.07 -0.07 0.33 -0.17 -0.07 1
7 Fund is stage-specialized 0.64 -0.01 -0.08 0.02 0.02 -0.07 0.07 1
8 Fund is specialized in early stage 0.18 -0.01 -0.10 0.04 0.06 0.03 -0.06 0.35
9 Fund is specialized in growth stage 0.39 -0.02 -0.03 0.00 0.03 -0.08 0.02 0.60
10 Fund is specialized in late stage 0.07 0.03 0.05 -0.01 -0.11 -0.02 0.19 0.21
11 Fund size (USD million) 200.80 645.06 0.13 0.21 -0.10 -0.09 -0.02 -0.03 0.03
12 Fund type = independent 0.59 0.04 0.12 0.09 0.07 0.03 0.02 -0.04
13 Fund type = corporate 0.07 -0.05 -0.07 0.12 0.16 0.01 -0.06 -0.01
14 Fund type = bank 0.27 0.02 -0.05 -0.15 -0.15 -0.04 0.01 0.07
15 Fund type = individuals 0.01 0.00 0.01 0.03 0.05 -0.02 -0.02 0.00
16 Fund type = public 0.05 -0.06 -0.10 -0.03 -0.04 0.02 0.01 0.03
17 Fund type = other 0.01 -0.02 -0.01 -0.07 -0.05 -0.03 -0.02 -0.12
18 Fund age 4.85 5.85 0.02 0.07 -0.18 -0.14 -0.04 -0.05 -0.11
19 VC age 12.66 13.67 0.14 0.27 -0.23 -0.19 -0.05 -0.06 -0.04
20 VC number of investments to date 91.76 217.59 0.26 0.37 -0.18 -0.13 -0.04 -0.08 -0.11
21 VC relative number of national investments to date 0.58 0.35 0.19 -0.02 -0.09 -0.11 -0.08 0.09 0.19
22 VC is located in North America 0.06 -0.12 -0.01 -0.01 0.02 0.01 -0.05 -0.07
23 VC is located in Europe 0.68 0.30 0.35 -0.07 -0.10 0.03 0.01 -0.04
24 VC is located in Asia 0.24 -0.22 -0.33 0.07 0.10 -0.03 -0.02 0.07
25 VC is located in ROW 0.02 -0.13 -0.16 0.04 0.00 -0.02 0.10 0.04
26 Date of investment 1998-2001 (bubble) 0.34 -0.02 0.00 0.11 0.20 -0.06 -0.09 0.08
27 Date of investment 1989-1997 (pre-bubble) 0.07 -0.13 0.07 -0.06 -0.07 -0.03 0.04 -0.03
28 Date of investment 2002-2008 (post-bubble) 0.59 0.08 -0.03 -0.07 -0.15 0.08 0.06 -0.06
29 Syndicated investment 0.61 0.03 0.11 0.06 0.11 0.10 -0.17 -0.13
30 PC in IT industry 0.49 0.02 0.01 0.21 0.48 -0.22 -0.23 -0.01
31 PC in Life Sciences industry 0.17 0.00 0.04 -0.04 -0.23 0.47 -0.11 -0.08
32 PC in non High-Tech industry 0.22 0.01 -0.01 -0.08 -0.26 -0.12 0.42 0.06
33 PC industry unknown 0.12 -0.03 -0.05 -0.18 -0.14 -0.05 -0.05 0.03
146

Descriptive statistics and correlation matrix (non US-based VC firms) (continued)

# Variable 8 9 10 11 12 13 14 15 16 17

8 Fund is specialized in early stage 1

9 Fund is specialized in growth stage -0.37 1
10 Fund is specialized in late stage -0.13 -0.22 1
11 Fund size (USD million) -0.08 0.11 -0.02 1
12 Fund type = independent -0.02 -0.02 0.00 0.07 1
13 Fund type = corporate -0.02 0.01 -0.01 -0.06 -0.34 1
14 Fund type = bank -0.02 0.06 0.04 -0.02 -0.73 -0.17 1
15 Fund type = individuals 0.01 0.00 -0.02 -0.02 -0.10 -0.02 -0.05 1
16 Fund type = public 0.13 -0.05 -0.04 -0.05 -0.26 -0.06 -0.13 -0.02 1
17 Fund type = other -0.03 -0.08 -0.02 0.00 -0.14 -0.03 -0.07 -0.01 -0.03 1
18 Fund age -0.08 -0.06 0.02 0.12 -0.06 -0.07 0.04 -0.02 0.04 0.16
19 VC age -0.09 0.03 0.00 0.35 0.02 -0.01 -0.01 -0.05 0.00 0.00
20 VC number of investments to date -0.09 0.00 -0.07 0.51 0.17 -0.08 -0.11 -0.03 -0.05 -0.03
21 VC relative number of national investments to date 0.16 0.04 0.03 -0.12 -0.09 -0.13 0.11 0.04 0.12 0.01
22 VC is located in North America 0.02 -0.08 0.00 -0.01 -0.08 0.01 0.10 -0.01 -0.05 0.02
23 VC is located in Europe -0.06 -0.01 0.03 0.10 0.19 -0.17 -0.08 0.04 -0.05 -0.04
24 VC is located in Asia 0.04 0.06 -0.04 -0.10 -0.17 0.19 0.04 -0.03 0.08 0.04
25 VC is located in ROW 0.03 0.00 0.03 -0.02 0.03 -0.03 -0.01 -0.01 -0.01 0.00
26 Date of investment 1998-2001 (bubble) 0.04 0.11 -0.12 -0.05 -0.03 0.09 -0.01 0.04 -0.05 0.02
27 Date of investment 1989-1997 (pre-bubble) -0.04 0.01 -0.03 -0.02 0.06 -0.03 -0.02 -0.02 -0.05 0.00
28 Date of investment 2002-2008 (post-bubble) -0.02 -0.11 0.13 0.05 0.00 -0.08 0.02 -0.02 0.07 -0.02
29 Syndicated investment -0.08 -0.05 -0.03 -0.05 0.01 0.08 -0.02 -0.02 -0.06 0.00
30 PC in IT industry 0.03 0.03 -0.13 -0.03 0.04 0.10 -0.09 0.03 -0.04 -0.01
31 PC in Life Sciences industry 0.03 -0.08 -0.04 0.01 0.01 -0.03 -0.01 -0.02 0.03 0.00
32 PC in non High-Tech industry -0.08 0.03 0.18 0.01 -0.03 -0.07 0.07 -0.01 0.01 0.01
33 PC industry unknown 0.02 0.00 0.02 0.03 -0.04 -0.03 0.05 -0.02 0.02 0.01

# Variable 18 19 20 21 22 23 24 25 26 27

18 Fund age 1
19 VC age 0.37 1
20 VC number of investments to date 0.24 0.57 1
21 VC relative number of national investments to date -0.08 -0.17 -0.18 1
22 VC is located in North America 0.04 0.02 -0.05 -0.10 1
23 VC is located in Europe 0.08 0.09 0.14 -0.01 -0.38 1
24 VC is located in Asia -0.09 -0.09 -0.11 0.06 -0.15 -0.82 1
25 VC is located in ROW -0.05 -0.08 -0.05 0.01 -0.03 -0.19 -0.07 1
26 Date of investment 1998-2001 (bubble) -0.17 -0.11 -0.09 0.02 -0.02 -0.08 0.10 -0.01 1
27 Date of investment 1989-1997 (pre-bubble) -0.04 0.00 -0.04 -0.11 0.04 -0.04 0.03 -0.02 -0.20 1
28 Date of investment 2002-2008 (post-bubble) 0.19 0.11 0.11 0.04 0.00 0.09 -0.11 0.02 -0.86 -0.34
29 Syndicated investment 0.03 0.01 0.01 -0.21 0.11 0.01 -0.05 -0.08 0.06 -0.07
30 PC in IT industry -0.09 -0.09 -0.02 -0.06 0.00 -0.05 0.07 -0.02 0.23 -0.05
31 PC in Life Sciences industry 0.04 0.04 0.03 -0.04 0.00 0.07 -0.07 -0.03 -0.07 0.01
32 PC in non High-Tech industry 0.02 0.00 -0.04 0.15 -0.08 0.02 0.00 0.07 -0.09 0.10
33 PC industry unknown 0.07 0.09 0.06 -0.06 0.09 -0.02 -0.02 -0.02 -0.14 -0.06
 147

Descriptive statistics and correlation matrix (non US-based VC firms) (continued)

# Variable 28 29 30 31 32

28 Date of investment 2002-2008 (post-bubble) 1

29 Syndicated investment -0.02 1
30 PC in IT industry -0.19 0.13 1
31 PC in Life Sciences industry 0.07 0.12 -0.44 1
32 PC in non High-Tech industry 0.04 -0.24 -0.52 -0.24 1
33 PC industry unknown 0.17 -0.05 -0.36 -0.17 -0.20

Pearson correlation coefficients (for two continuous variables)

Point biseral coefficients (for one continuous variable and one dummy variable)
Phi coefficients (for two dummy variables)

Correlations stronger than .4 are marked in bold.

(N=58,755)
148

Appendix 17: Descriptive statistics and correlation matrix (Germany-based VC firms)

# Variable Mean SD 1 2 3 4 5 6 7

1 Relative supply of deals in VC nation (in investment year) 0.01 0.00 1

2 Relative demand for deals in VC nation (in investment year) 0.03 0.00 0.49 1
3 Fund is industry-specialized 0.48 0.05 0.11 1
4 Fund is specialized in IT 0.35 0.11 0.06 0.77 1
5 Fund is specialized in Life Sciences 0.11 -0.08 0.08 0.36 -0.25 1
6 Fund is specialized in non High-Tech 0.02 -0.03 0.01 0.16 -0.11 -0.05 1
7 Fund is stage-specialized 0.49 0.15 0.07 -0.01 -0.02 -0.05 0.14 1
8 Fund is specialized in early stage 0.16 0.09 -0.01 0.04 0.05 0.01 -0.04 0.45
9 Fund is specialized in growth stage 0.28 0.14 0.06 -0.06 -0.04 -0.06 0.04 0.64
10 Fund is specialized in late stage 0.04 -0.08 0.05 0.03 -0.05 -0.03 0.32 0.22
11 Fund size (USD million) 111.01 105.83 0.00 0.13 -0.06 -0.09 0.05 -0.01 -0.20
12 Fund type = independent 0.66 -0.08 -0.12 0.08 0.00 0.12 0.00 -0.16
13 Fund type = corporate 0.12 0.02 0.08 0.17 0.27 -0.12 -0.04 -0.01
14 Fund type = bank 0.16 0.06 0.04 -0.24 -0.23 -0.06 0.04 0.14
15 Fund type = individuals 0.02 0.09 -0.02 0.11 0.15 -0.04 -0.01 0.13
16 Fund type = public 0.03 -0.06 0.07 -0.06 -0.10 0.06 0.01 0.11
17 Fund type = other 0.01 0.06 0.03 -0.09 -0.07 -0.03 -0.01 -0.09
18 Fund age 4.51 5.13 -0.11 0.07 -0.20 -0.17 -0.07 0.03 -0.19
19 VC age 7.80 7.04 -0.22 0.07 -0.10 -0.21 0.12 0.09 -0.29
20 VC number of investments to date 59.64 91.46 -0.13 0.17 0.08 -0.01 0.19 -0.08 -0.35
21 VC relative number of national investments to date 0.57 0.37 0.21 0.14 -0.12 -0.17 0.03 0.07 0.36
22 VC is located in North America 0.00 . . . . . . .
23 VC is located in Europe 1.00 . . . . . . .
24 VC is located in Asia 0.00 . . . . . . .
25 VC is located in ROW 0.00 . . . . . . .
26 Date of investment 1998-2001 (bubble) 0.44 0.67 -0.16 -0.01 0.08 -0.12 -0.05 0.11
27 Date of investment 1989-1997 (pre-bubble) 0.05 -0.46 -0.79 -0.09 -0.07 -0.04 0.00 -0.09
28 Date of investment 2002-2008 (post-bubble) 0.51 -0.47 0.49 0.05 -0.05 0.14 0.05 -0.07
29 Syndicated investment 0.71 -0.09 0.03 0.03 0.03 0.04 -0.10 -0.20
30 PC in IT industry 0.56 0.11 -0.06 0.19 0.48 -0.36 -0.15 0.00
31 PC in Life Sciences industry 0.26 -0.06 0.04 -0.05 -0.36 0.53 -0.09 -0.05
32 PC in non High-Tech industry 0.14 -0.07 0.02 -0.17 -0.20 -0.12 0.33 0.07
33 PC industry unknown 0.04 -0.04 0.04 -0.06 -0.03 -0.04 -0.02 -0.02
 149

Descriptive statistics and correlation matrix (Germany-based VC firms) (continued)

# Variable 8 9 10 11 12 13 14 15 16 17

8 Fund is specialized in early stage 1

9 Fund is specialized in growth stage -0.27 1
10 Fund is specialized in late stage -0.09 -0.13 1
11 Fund size (USD million) -0.24 -0.03 0.01 1
12 Fund type = independent 0.00 -0.17 0.00 0.01 1
13 Fund type = corporate -0.08 0.08 -0.06 -0.18 -0.52 1
14 Fund type = bank -0.08 0.19 0.08 0.23 -0.61 -0.17 1
15 Fund type = individuals 0.10 0.07 -0.02 -0.09 -0.17 -0.05 -0.06 1
16 Fund type = public 0.28 -0.09 -0.04 -0.13 -0.25 -0.07 -0.08 -0.02 1
17 Fund type = other -0.04 -0.06 -0.02 0.06 -0.13 -0.04 -0.04 -0.01 -0.02 1
18 Fund age -0.15 -0.09 0.00 0.02 0.03 -0.07 0.03 0.03 -0.08 0.08
19 VC age -0.20 -0.18 0.05 0.23 0.21 -0.19 -0.05 -0.04 -0.09 0.01
20 VC number of investments to date -0.18 -0.21 -0.08 0.34 0.23 -0.06 -0.18 -0.06 -0.05 -0.04
21 VC relative number of national investments to date 0.28 0.17 0.02 -0.19 0.01 -0.26 0.07 0.09 0.19 0.10
22 VC is located in North America . . . . . . . . . .
23 VC is located in Europe . . . . . . . . . .
24 VC is located in Asia . . . . . . . . . .
25 VC is located in ROW . . . . . . . . . .
26 Date of investment 1998-2001 (bubble) 0.10 0.11 -0.14 -0.09 -0.01 -0.03 0.04 0.13 -0.13 0.05
27 Date of investment 1989-1997 (pre-bubble) -0.04 -0.07 -0.01 -0.10 0.12 -0.07 -0.05 -0.03 -0.04 -0.02
28 Date of investment 2002-2008 (post-bubble) -0.08 -0.08 0.14 0.13 -0.04 0.06 -0.01 -0.12 0.14 -0.04
29 Syndicated investment -0.15 -0.05 -0.11 0.10 -0.01 0.05 0.02 -0.07 -0.02 -0.06
30 PC in IT industry 0.03 0.03 -0.11 -0.08 -0.07 0.19 -0.07 0.09 -0.07 -0.01
31 PC in Life Sciences industry 0.01 -0.04 -0.07 0.09 0.09 -0.17 0.02 -0.06 0.08 0.01
32 PC in non High-Tech industry -0.04 0.02 0.19 -0.02 -0.01 -0.07 0.08 -0.03 0.02 0.01
33 PC industry unknown -0.04 -0.03 0.08 0.03 0.00 0.02 0.01 -0.03 -0.03 -0.02

# Variable 18 19 20 21 22 23 24 25 26 27

18 Fund age 1
19 VC age 0.48 1
20 VC number of investments to date 0.22 0.62 1
21 VC relative number of national investments to date -0.05 -0.15 -0.43 1
22 VC is located in North America . . . . .
23 VC is located in Europe . . . . . .
24 VC is located in Asia . . . . . . .
25 VC is located in ROW . . . . . . . .
26 Date of investment 1998-2001 (bubble) -0.20 -0.30 -0.27 0.16 . . . .
27 Date of investment 1989-1997 (pre-bubble) 0.00 0.02 -0.07 -0.19 . . . . -0.20 1
28 Date of investment 2002-2008 (post-bubble) 0.19 0.29 0.29 -0.08 . . . . -0.91 -0.23
29 Syndicated investment 0.06 0.09 0.18 -0.32 . . . . -0.12 0.01
30 PC in IT industry -0.12 -0.20 -0.03 -0.13 . . . . 0.16 0.01
31 PC in Life Sciences industry 0.05 0.13 0.12 0.12 . . . . -0.07 -0.02
32 PC in non High-Tech industry 0.10 0.08 -0.12 0.07 . . . . -0.11 0.02
33 PC industry unknown 0.02 0.09 0.02 -0.05 . . . . -0.06 -0.02
150

Descriptive statistics and correlation matrix (Germany-based VC firms) (continued)

# Variable 28 29 30 31 32

28 Date of investment 2002-2008 (post-bubble) 1

29 Syndicated investment 0.12 1
30 PC in IT industry -0.17 -0.02 1
31 PC in Life Sciences industry 0.08 0.11 -0.66 1
32 PC in non High-Tech industry 0.10 -0.15 -0.46 -0.24 1
33 PC industry unknown 0.07 0.06 -0.24 -0.12 -0.09

Pearson correlation coefficients (for two continuous variables)

Point biseral coefficients (for one continuous variable and one dummy variable)
Phi coefficients (for two dummy variables)

Correlations stronger than .4 are marked in bold.

(N=4,464)
 151

Appendix 18: Target location model descriptive statistics and correlation matrix164
# Variable Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 13

1 distance 1.71 0.89 1

2 Austria 0.01 0.03 1
3 Canada 0.00 0.09 -0.01 1
4 France 0.02 0.04 -0.01 -0.01 1
5 Germany 0.58 -0.94 -0.12 -0.08 -0.15 1
6 Israel 0.02 0.04 -0.01 -0.01 -0.02 -0.15 1
7 Netherlands 0.01 0.03 -0.01 -0.01 -0.01 -0.11 -0.01 1
8 Sweden 0.00 0.02 -0.01 0.00 -0.01 -0.08 -0.01 -0.01 1
9 Switzerland 0.03 0.06 -0.02 -0.01 -0.02 -0.20 -0.02 -0.02 -0.01 1
10 UK 0.03 0.05 -0.02 -0.01 -0.02 -0.19 -0.02 -0.02 -0.01 -0.03 1
11 US 0.28 0.90 -0.06 -0.04 -0.08 -0.74 -0.08 -0.06 -0.04 -0.11 -0.10 1
12 PC in IT industry 0.58 0.16 -0.05 0.05 0.02 -0.15 0.04 0.04 0.03 -0.03 0.04 0.14 1
13 PC in Life Sciences industry 0.27 -0.12 0.07 -0.04 0.00 0.10 -0.01 -0.03 -0.02 0.05 -0.03 -0.12 -0.71 1
14 PC in non High-Tech industry 0.15 -0.07 -0.01 -0.02 -0.04 0.08 -0.05 -0.01 -0.02 -0.03 -0.02 -0.04 -0.49 -0.25
15 PC industry unknown 0.00 -0.03 -0.01 0.00 -0.01 0.02 0.03 -0.01 0.00 0.02 -0.01 -0.03 -0.06 -0.03
16 Date of investment 1998-2001 (bubble) 0.44 -0.16 -0.01 0.01 -0.03 0.17 0.00 -0.01 0.03 -0.04 -0.02 -0.14 0.15 -0.08
17 Date of investment 1989-1997 (pre- 0.05 0.16 -0.01 -0.01 0.00 -0.13 0.02 -0.01 -0.02 -0.03 -0.03 0.17 0.01 -0.02
bubble)
18 Date of investment 2002-2008 (post- 0.51 0.09 0.02 0.00 0.03 -0.11 -0.01 0.01 -0.02 0.06 0.04 0.06 -0.15 0.09
bubble)
19 Relative supply of deals in target 0.21 0.31 0.89 -0.07 -0.04 -0.07 -0.72 -0.09 -0.06 -0.04 -0.11 -0.09 0.99 0.14 -0.12
location (in investment year)
20 Relative demand for deals in target 0.19 0.25 0.90 -0.07 -0.04 -0.07 -0.73 -0.09 -0.06 -0.05 -0.12 -0.08 1.00 0.14 -0.12
location (in investment year)
21 VC experience in investing in the target 0.72 0.45 -0.38 -0.09 -0.05 -0.12 0.42 0.01 -0.05 -0.04 -0.08 -0.11 -0.26 -0.03 0.03
location alone
22 VC has an office in target location 0.74 -0.56 -0.17 -0.11 -0.21 0.70 -0.22 -0.15 -0.11 -0.24 -0.21 -0.30 -0.10 0.08
23 Target location follows civil law 0.67 -0.92 0.07 -0.09 0.09 0.83 -0.19 0.06 0.05 0.12 -0.23 -0.88 -0.17 0.13
24 Target location speaks a Germanic 0.67 -0.93 0.07 -0.09 -0.18 0.83 -0.19 0.06 0.05 0.12 0.12 -0.89 -0.16 0.12
language
25 Target location has a common border 0.64 -0.93 -0.14 -0.09 0.09 0.88 -0.18 0.07 -0.09 0.13 -0.22 -0.84 -0.16 0.12
with VC nation

# Variable 14 15 16 17 18 19 20 21 22 23 24

14 PC in non High-Tech industry 1

15 PC industry unknown -0.02 1
16 Date of investment 1998-2001 (bubble) -0.12 0.02 1
17 Date of investment 1989-1997 (pre- 0.02 -0.01 -0.20 1
bubble)
18 Date of investment 2002-2008 (post- 0.11 -0.01 -0.91 -0.23 1
bubble)
19 Relative supply of deals in target -0.04 -0.04 -0.12 0.25 0.01 1
location (in investment year)
20 Relative demand for deals in target -0.05 -0.04 -0.13 0.21 0.04 1.00 1
location (in investment year)
21 VC experience in investing in the target -0.01 0.03 0.06 -0.15 0.00 -0.28 -0.27 1
location alone
22 VC has an office in target location 0.03 0.00 0.08 -0.01 -0.08 -0.29 -0.28 0.54 1
23 Target location follows civil law 0.07 0.02 0.14 -0.16 -0.07 -0.87 -0.88 0.30 0.45 1
24 Target location speaks a Germanic 0.06 0.02 0.14 -0.16 -0.07 -0.88 -0.88 0.31 0.46 0.88 1
language
25 Target location has a common border 0.07 0.02 0.14 -0.15 -0.08 -0.83 -0.83 0.34 0.53 0.95 0.87
with VC nation

Pearson correlation coefficients (for two continuous variables)

Point biseral coefficients (for one continuous variable and one dummy variable)
Phi coefficients (for two dummy variables)
Correlations stronger than .4 are marked in bold. (N=4,245165)

164
The table relates to investments by Germany-based VC firms.
152

165
Only the top investment target locations by the main 10 investor nations are considered in this model (i.e., 27
alternative target locations per investment), hence the slightly lower number of observations.
 153

Bibliography

Abrahamson, Eric / Choelsoon, Park (1994): "Concealment of Negative Organizational Outcomes: An Agency
Theory Perspective" in: The Academy of Management Journal, Vol. 37, No. 5, 1302-1334
Agrawal, Anup / Mandelker, Gershon N. (1987): "Managerial Incentives and Corporate Investment and
Financing Decisions" in: The Journal of Finance, Vol. 42, No. 4, 823-837
Aizenman, Joshua / Kendall, Jake (2008): The Internationalization of Venture Capital and Private Equity,
UCSC and the NBER; The World Bank, Working Paper, September 2008
Alhorr, Hadi S. / Moore, Curt B. / Payne, G. Tyge (2008): "The Impact of Economic Integration on Cross-
Border Venture Capital Investments: Evidence From the European Union" in: Entrepreneurship: Theory &
Practice, Vol. 32, No. 5, 897-917
Alt, James E. / Lassen, David Dreyer (2003): "The Political Economy of Institutions and Corruption in American
States" in: Journal of Theoretical Politics, Vol. 15, No. 3, 341-365
Amihud, Yakov / Lev, Baruch (1981): "Risk Reduction as a Managerial Motive for Conglomerate Mergers" in:
The Bell Journal of Economics, Vol. 12, No. 2, 605-617
Anonymous (2009): "A special report on entrepreneurship" in: The Economist, March 14th 2009, 3-18
Arenius, Pia (2005): "The Psychic Distance Postulate Revised: From Market Selection to Speed of Market
Penetration" in: Journal of International Entrepreneurship, Vol. 3, No. 2, 115-131
Arrow, Kenneth J. (1964): "Control in Large Organizations" in: Management Science, Vol. 10, No. 3, 397-408
Arrow, Kenneth J. (1971): Essays in the theory of risk-bearing, Chicago
Arrow, Kenneth J. (1975): "Vertical Integration and Communication" in: The Bell Journal of Economics, Vol. 6,
No. 1, 173-183
Arthurs, Jonathan D. / Busenitz, Lowell W. (2003): "The Boundaries and Limitations of Agency Theory and
Stewardship Theory in the Venture Capitalist/Entrepreneur Relationship" in: Entrepreneurship: Theory &
Practice, Vol. 28, No. 2, 145-162
Balcarcel, Ana L. (2004): "Performance of cross-border venture capital investments" in: Financial Management
Association Annual Meeting 2004, Vol. 5301607, October 7-10, 2004, New Orleans, LA
Barkema, Harry G. / Vermeulen, Freek (1998): "International Expansion Through Start-Up or Acquisition: A
Learning Perspective" in: Academy of Management Journal, 41, 7-26
Barry, Christopher B. / Muscarella, Chris J. / Peavy Iii, John W. / Vetsuypens, Michael R. (1990): "The role of
venture capital in the creation of public companies: Evidence from the going-public process" in: Journal of
Financial Economics, Vol. 27, No. 2, 447-471
Baum, Joel A. C. / Silverman, Brian S. (2004): "Picking winners or building them? Alliance, intellectual, and
human capital as selection criteria in venture financing and performance of biotechnology startups" in: Journal
of Business Venturing, Vol. 19, No. 3, 411-436
Baygan, Günseli / Freudenberg, Michael (2000): "The Internationalisation of Venture Capital Activity in OECD
Countries: Implications for Measurement and Policy" in: OECD Science, Technology and Industry Working
Papers, No. 7
Belderbos, Ren / Sleuwaegen, Leo (2005): "Competitive drivers and international plant configuration strategies:
a product-level test" in: Strategic Management Journal, Vol. 26, No. 6, 577-593
Bessler, Wolfgang / Bittelmeyer, Claudia (2008): "Patents and the Performance of Technology Firms: Evidence
from Initial Public Offerings in Germany" in: Financial Markets and Portfolio Management, Vol. 22, No. 4,
323-356

H. Tarrade, Cross-Border Venture Capital Investments, DOI 10.1007/978-3-8349-6939-2,

© Gabler Verlag | Springer Fachmedien Wiesbaden 2012
154

Bretz, Michael / Irsch, Norbert / Lageman, Bernhard / Licht, Georg / Wallau, Frank (2009):
"MittelstandsMonitor 2009: Deustche Wirtschaft in der Rezession - Talfahrt auch im Mittelstand", KfW,
Creditreform, IfM, RWI, ZEW: Jährlicher Bericht zu Konjonktur- und Strukturfragen kleiner und mittlerer
Unternehmen, March 2009, Frankfurt am Main
Butler, Alexander W. / Goktan, Sinan (2008): On the Matching of Companies and Their Financial
Intermediaries: Evidence from Venture Capital, Department of Finance and Managerial Economics,
Working Paper, School of Management, The University of Texas at Dallas, September 2008
BVK (2006): "Statistik 2005", German Private Equity and Venture Capital Association (BVK): Das Jahr in
Zahlen, February 2006
Carree, Martin / van Stel, Andre / Thurik, Roy / Wennekers, Sander (2007): "The Relationship between
Economic Development and Business Ownership Revisited" in: Entrepreneurship and Regional Development,
Vol. 19, No. 3, 281-291
Caves, Richard E. (1971): "International Corporations: The Industrial Economics of Foreign Investment" in:
Economica, 38, 1-27
Caves, Richard E. (2007): Multinational enterprise and economic analysis, Third edition, Cambridge, MA
Cheng, Shaoming (2007): "Structure of Firm Location Choices: An Examination of Japanese Greenfield
Investment in China" in: Asian Economic Journal, Vol. 21, No. 1, 47-73
Coase, R. H. (1937): "The Nature of the Firm" in: Economica, Vol. 4, No. 16, 386-405
Conlon, Edward J. / McLean Parks, Judi (1990): "Effects of Monitoring and Tradition on Compensation
Arrangements: An Experiment with Principal-Agent Dyads" in: The Academy of Management Journal, Vol. 33,
No. 3, 603-622
Cumming, Douglas J. (2006): "The Determinants of Venture Capital Portfolio Size: Empirical Evidence" in:
Journal of Business, Vol. 79, No. 3, 1083-1126
Cumming, Douglas J. / Macintosh, Jeffrey G. (2003): "Venture Capital Exits in Canada and the United States"
in: University of Toronto Law Journal, Vol. 53, 101-200
Cumming, Douglas J. / Walz, Uwe (2007): Private Equity Returns and Disclosure Around the World, Working
Paper, York University - Schulich School of Business, July 2007
Cumming, Douglas J. / Johan, Sofia (2008): "Information asymmetries, agency costs and venture capital exit
outcomes" in: Venture Capital, Vol. 10, No. 3, 197 - 231
Cumming, Douglas J. / Dai, Na (2008): Local Bias in Venture Capital Investments, Working Paper, Anderson
School of Management, University of New Mexico
Dalziel, Thomas / White, Robert E. / Arthurs, Jonathan D. / Hoskisson, Robert E. (2009): "Initially Distracted:
The Influence of Boards on Agency Costs in Initial Public Offering (IPO) Firms" New Frontiers in
Entrepreneurship, New York, 11-30
Dardis, Rachel / Hrozencik, Diane (1985): "The impact of socio-economic variables and automobile stock
characteristics on the purchase of new imported automobiles" in: Advances in Consumer Research, Vol. 12, No.
1, 538-543
Davidson, W. (1980): "The location of foreign direct investment activity: Country characteristics and experience
effects" in: Journal of International Business Studies, 9-22
Delios, Andrew / Beamish, Paul W. (1999): "Geographic scope, product diversification and the corporate
performance of Japanese firms" in: Strategic Management Journal, 20, 711-727
Deloitte (2006): "Global trends in venture capital, 2006 survey"
Deloitte / NVCA (2005): "US VCs to expand global investments": Survey report, June 2005
Deloitte / NVCA (2007): "Percentage of US venture capitalists investing globally lower than expected": News
release, July 2007
Deloitte / NVCA (2009): "Global economic downturn driving evolution of venture capital industry": News
Release, June 10, 2009, Newy York, NY
 155

Dikova, Desislava (2009): "Performance of foreign subsidiaries: Does psychic distance matter?" in:
International Business Review, Vol. 18, No. 1, 38-49
Diller, Christian / Kaserer, Christoph (2005): What Drives Cash Flow Based European Private Equity Returns?
– Fund Inflows, Skilled GPs and/or Risk?, RICAFE - Risk Capital and the Financing of European Innovative
Firms, Working Paper, Centre for Entrepreneurial and Finance Studies (CEFS) and Department for Financial
Management and Capital Markets, Technische Universität München, February 2005
DiMasi, Joseph A. / Hansen, Ronald W. / Grabowski, Henry G. (2003): "The price of innovation: new estimates
of drug development costs" in: Journal of Health Economics, Vol. 22, No. 2, 151
Doran, A. / Bannock, G. (2000): "Publicly sponsored regional venture capital: what can the UK learn from the
US experience?" in: Venture Capital, No. 2, 255-285
Doukas, John / Travlos, Nickolaos G. (1988): "The Effects of Corporate Multinationalism on Shareholders'
Wealth: Evidence from International Acquisitions" in: Journal of Finance, 43, 1161-1175
Draho, Jason (2004): The IPO decision: why and how companies go public, Edward Elgar Publishing, 2005,
Cheltenham, UK
Du, Julan / Lu, Yi / Tao, Zhigang (2008): "FDI location choice: agglomeration vs institutions" in: International
Journal of Finance & Economics, Vol. 13, No. 1, 92-107
Dunning, J. H. (1990): "The globalization of firms and the competitiveness of countries" in: Globalization of
Firms and the Competitiveness of Nations, Institute of Economic Research, Lund University, Lund, 9-57
Eisenhardt, Kathleen M. (1984): "Contract Model: Theory and an Empirical Test" in: Academy of Management
Proceedings
Eisenhardt, Kathleen M. (1985): "Control: Organizational and Economic Approaches" in: Management Science,
Vol. 31, No. 2, 134-149
Eisenhardt, Kathleen M. (1988): "Agency- and Institutional-Theory Explanations: The Case of Retail Sales
Compensation" in: The Academy of Management Journal, Vol. 31, No. 3, 488-511
Eisenhardt, Kathleen M. (1989): "Agency Theory: An Assessment and Review" in: Academy of Management
Review, Vol. 14, No. 1, 57-74
EVCA (2002): "Survey of economic and social impact of venture capital in Europe", European Private Equity
and Venture Capital Association: Research paper, June 2002
EVCA (2005): "Employment contribution of private equity and venture capital in Europe", European Private
Equity and Venture Capital Association: Research paper, November 2005
EVCA (2007): "Further increase in the number of business plans received by European VCs in 2007": EVCA
Barometer
EVCA (2008): "2007 European Private Equity Activity Survey", European Private Equity and Venture Capital
Association Latest Pan-European private equity industry data, March 2008
Fama, Eugene F. / Jensen, Michael C. (1983): "Separation of Ownership and Control" in: Journal of Law and
Economics, Vol. 26, No. 2, 301-325
Franke, Nikolaus / Gruber, Marc / Harhoff, Dietmar / Henkel, Joachim (2006): "What you are is what you like -
similarity biases in venture capitalists' evaluations of start-up teams" in: Journal of Business Venturing, Vol. 21,
No. 6, 802-826
Franke, Nikolaus / Gruber, Marc / Harhoff, Dietmar / Henkel, Joachim (2008): "Venture Capitalists' Evaluations
of Start-Up Teams: Trade-Offs, Knock-Out Criteria, and the Impact of VC Experience" in: Entrepreneurship:
Theory & Practice, Vol. 32, No. 3, 459-483
French, Kenneth R. / Poterba, James M. (1991): "Investor diversification and international equity markets" in:
American Economic Review, Vol. 81, No. 2, 222
Fried, Vance H. / Hisrich, Robert D. (1994): "Toward a Model of Venture Capital Investment Decision Making"
in: Financial Management, Vol. 23, No. 3, 226-242
Friedman, Joseph / Gerlowski, Daniel A. / Silberman, Johnathan (1992): "What Attracts Foreign Multinational
Corporations? Evidence From Branch Plant Location in the United States" in: Journal of Regional Science, 32,
403
156

Fritsch, Michael / Schilder, Dirk (2006): "Is Venture Capital a regional business? The role of syndication", Max
Planck Institute of Economics, Group for Entrepreneurship, Growth and Public Policy, October 2006
Fritsch, Michael / Schilder, Dirk (2008): "Does venture capital investment really require spatial proximity? An
empirical investigation" in: Environment and Planning A, Vol. 40, No. 9, 2114-2131
Furman, Jeffrey L. / Porter, Michael E. / Stern, Scott (2002): "The determinants of national innovative capacity"
in: Research Policy, Vol. 31, No. 6, 899-933
Giot, Pierre / Schwienbacher, Armin (2007): "IPOs, trade sales and liquidations: Modelling venture capital exits
using survival analysis" in: Journal of Banking & Finance, Vol. 31, No. 3, 679-702
Global Entrepreneurship Monitor (2009): "2008 Executive Report", Global Entrepreneurship Research
Consortium (GERA)
Gompers, Paul / Lerner, Josh (2001): The money of invention: How venture capital creates new wealth, Harvard
Business School Press, Cambridge, MA
Gompers, Paul / Lerner, Josh (2006): The Venture Capital Cycle, The MIT Press, Second Edition, Cambridge,
MA
Gorman, Michael / Sahlman, William A. (1989): "What Do Venture Capitalists Do?" in: Journal of Business
Venturing, Vol. 4, No. 4, 231-248
Greve, HR (2000): "Market niche entry decisions: Competition, learning, and strategy in Tokyo banking, 1894-
1936" in: Academy of Management Journal, Vol. 43, No. 5, 816-836
Grover, Steven L. (1993): "Lying, Deceit, and Subterfuge: A Model of Dishonesty in the Workplace" in:
Organization Science, Vol. 4, No. 3, 478-495
Guilhon, Bernard / Montchaud, Sandra (2006): "The dynamics of venture capital industry" in: International
Journal of Technology Management, Vol. 34, No. 1/2, 146-160
Guler, I. / Guillen, M. (2004): "Knowledge, institutions and the internationalization of the US Venture capital
industry" in: Academy of Management Annual Meeting 2004, August 2004, New Orleans, USA
Gupta, Anil K. / Sapienza, Harry J. (1992): "Determinants of Venture Capital Firms' Preferences Regarding the
Industry Diversity and Geographic Scope of Their Investments" in: Journal of Business Venturing, Vol. 7, No. 5,
347-362
Hall, Graham / Tu, Ciwen (2003): "Venture capitalists and the decision to invest overseas" in: Venture Capital,
Vol. 5, No. 2, 181-190
Häussler, Carolin / Harhoff, Dietmar / Müller, Elisabeth (2008): To Be Financed or Not - The Role of Patents for
Venture Capital Financing, Working Paper, Ludwig-Maximilians-Universität Munich, September 2008
Hege, Ulrich / Palomino, Frédéric / Schwienbacher, Armin (2003): Determinants of Venture Capital
Performance: Europe and United States, RICAFE - Risk Capital and the Financing of European Innovative
Firms, Working Paper, HEC School of Management and CEPR, November 2003
Heiss, Florian (2002): "Structural Choice Analysis with Nested Logit models" in: The Stata Journal, Vol. 2, No.
3, 227-252
Hellmann, Thomas / Puri, Manju (2002): "Venture Capital and the Professionalization of Start-up Firms:
Empirical Evidence" in: The Journal of Finance, Vol. 57, No. 1, 169-197
Holmstrom, Bengt (1979): "Moral Hazard and Observability" in: The Bell Journal of Economics, Vol. 10, No. 1,
74-91
Hsu, David H. / Ziedonis, Rosemarie H. (2008): "Patents as quality signals for entrepreneurial ventures" in:
Academy of Management Proceedings, 1-6
Ick, Matthias (2005): "Performance Measurement and Appraisal of Private Equity Investments relative to Public
Equity Markets", http://www.cepres.de
Jääskeläinen, Mikko / Maula, Markku / Seppä, Tuukka (2006): "Allocation of Attention to Portfolio Companies
and the Performance of Venture Capital Firms" in: Entrepreneurship: Theory & Practice, Vol. 30, No. 2, 185-
206
Jeng, Leslie A. / Wells, Philippe C. (2000): "The determinants of venture capital funding: evidence across
countries" in: Journal of Corporate Finance, Vol. 6, No. 3, 241-289
 157

Jensen, M. C. (1983): "Organization Theory and Methodology" in: The Accounting Review, Vol. 58, No. 2, 319-
339
Jensen, M. C. / Meckling, W. F. (1976): "Theory of the Firm: Managerial behavior, agency costs, and ownership
structure" in: Journal of Financial Economics, Vol. 3, No. 4
Johanson, J. / Wiedersheim-Paul, F. (1975): "The internationalisation of the firm: Four Swedish cases" in:
Journal of Management Studies, Vol. October, 305–322
Johanson, J. / Vahlne, J. E. (1978): "A model for the decision-making process of the firm" in: European
Research in International Business, Elsevier North-Holland, New York, NY, 12-27
Johanson, Jan / Vahlne, Jan-Erik (1977): "The Internationalization Process of the Firm-A Model of Knowledge
Development and Increasing Foreign Market Commitments" in: Journal of International Business Studies, Vol.
8, No. 1, 23-32
Johanson, Jan / Vahlne, Jan-Erik (2003): "Business Relationship Learning and Commitment in the
Internationalization Process" in: Journal of International Entrepreneurship, Vol. 1, No. 1, 83-101
Jones, Charles M. / Rhodes-Kropf, Matthew (2003): The Price of Diversifiable Risk in Venture Capital and
Private Equity, Columbia University, Working Paper, May 2003
Jungwirth, Carola / Moog, Petra (2004): "Selection and support strategies in venture capital financing: high-tech
or low-tech, hands-off or hands-on?" in: Venture Capital, Vol. 6, No. 2/3, 105-123
Kaplan, Steven N. / Schoar, Antoinette (2005): "Private Equity Performance: Returns, Persistence, and Capital
Flows" in: Journal of Finance, Vol. 60, No. 4, 1791-1823
Kaplan, Steven N. / Sensoy, Berk A. / Strömberg, Per J. (2002): How Well do Venture Capital Databases Reflect
Actual Investments?, Working Paper, University of Chicago Graduate School of Business, September 2002
Kaserer, Christoph / Diller, Christian (2004): European Private Equity Funds – A Cash Flow Based
Performance Analysis, Working Paper, Center for Entrepreneurial and Financial Studies (CEFS) and Department
for Financial Management and Capital Markets, Technische Universität München, May 2004
Kesner, Idalene F. / Shapiro, Debra L. / Sharma, Anurag (1994): "Brokering Mergers: An Agency Theory
Perspective on the Role of Representatives" in: The Academy of Management Journal, Vol. 37, No. 3, 703-721
Kim, Jae Hong / Pagliara, Francesca / Preston, John (2005): "The intention to move and residential location
choice behaviour" in: Urban Studies, Vol. 42, No. 9, 1621-1636
Kim, W. Chan / Hwang, Peter / Burgers, Willem P. (1993): "Multinationals' Diversification and The Risk-Return
Trade-Off" in: Strategic Management Journal, 14, 275-286
Kiser, Edgar (1999): "Comparing Varieties of Agency Theory in Economics, Political Science, and Sociology:
An Illustration from State Policy Implementation" in: Sociological Theory, Vol. 17, No. 2, 146-170
Kleinknecht, Alfred / Van Montfort, Kees / Brouwer, Erik (2002): "The non-trivial choice between innovation
indicators" in: Economics of Innovation & New Technology, Vol. 11, No. 2, 109
Kogut, Bruce (1985): "Designing Global Strategies: Comparative and Competitive Value-Added Chains" in:
Sloan Management Review, Vol. 27, No. XX, 27-38
Kortum, Samuel / Lerner, Josh (2000): "Assessing the contribution of venture capital to innovation" in: RAND
Journal of Economics, Vol. 31, No. 4, 674-692
Kosnik, Rita D. (1987): "Greenmail: A Study of Board Performance in Corporate Governance" in:
Administrative Science Quarterly, Vol. 32, No. 2, 163-185
Leatherwood, Marya L. / Conlon, Edward J. (1987): "Diffusibility of Blame: Effects on Persistence in a Project"
in: The Academy of Management Journal, Vol. 30, No. 4, 836-847
Lerner, Josh (1994): "The importance of patent scope: An empirical analysis" in: RAND Journal of Economics,
Vol. 25, No. 2, 319-333
Lerner, Josh (1995): "Venture Capitalists and the Oversight of Private Firms" in: Journal of Finance, Vol. 50,
No. 1, 301-318
Linghui, Tang / Koveos, Peter E. (2004): "Venture Entrepreneurship, Innovation Entrepreneurship, and
Economic Growth" in: Journal of Developmental Entrepreneurship, Vol. 9, No. 2, 161-171
158

Lintner, John (1965): "Security Prices, Risk, and Maximal Gains From Diversification" in: Journal of Finance,
Vol. 20, No. 4, 587-615
Ljungqvist, Alexander / Richardson, Matthew (2003): "The cash flow, return and risk characteristics of private
equity" in: National Bureau of Economic Research Working Paper Series, Vol. No. 9454,
Lossen, Ulrich (2006): The Performance of Private Equity Funds: Does Diversification Matter?, Institut für
Innovationsforschung und Technologiemanagement, Working Paper, Fakultät für Betriebswirtschaft, Ludwig-
Maximilians-Universität München, June 2006
Lossen, Ulrich (2007): Portfolio Strategies of Private Equity Firms, Theory and Evidence, München
Lu, Jane W. / Beamish, Paul W. (2004): "International Diversification and Firm Performance: The S-Curve
Hypothesis" in: Academy of Management Journal, 47, 598-609
MacCrimmon, Kenneth R. / Wehrung, Donald A. (1990): "Characteristics of Risk Taking Executives" in:
Management Science, Vol. 36, No. 4, 422-435
MacMillan, Ian C. / Kulow, David M. / Khoylian, Roubina (1988): "Venture Capitalists' Involvement in Their
Investments: Extent and Performance" in: Journal of Business Venturing, Vol. 4, No. 1, 27
Madhok, A (1997): "Cost, value and foreign market entry mode: The transaction and the firm" in: Strategic
Management Journal, Vol. 18, No. 1, 39-61
Mäkelä, Markus M. / Maula, Markku V. J. (2005): "Cross-Border Venture Capital and New Venture
Internationalization: An Isomorphism Perspective" in: Venture Capital, Vol. 7, No. 3, 227-257
Manigart, Sophie / Wright, Mike / Robbie, Ken / Desbrieres, Philippe / De Waele, Koen (1997): "Venture
Capitalists' Appraisal of Investment Projects: An Empirical European Study" in: Entrepreneurship: Theory &
Practice, Vol. 21, No. 4, 29-43
Mann, Ronald J. / Sager, Thomas W. (2007): "Patents, venture capital, and software start-ups" in: Research
Policy, Vol. 36, No. 2, 193-208
Markowitz, H. (1952): "Portfolio selection" in: Journal of Finance, Vol. 7, No. 1, 77-91
Marwaha, Sam / Patil, Samir / Singh, Navjot (2007): "Using IT to speed up clinical trials" in
McKinsey&Company: McKinsey on IT, Spring 2007, New York, NY
Mason, Colin M. / Harrison, Richard T. (1995): "Closing the Regional Equity Capital Gap: The Role of Informal
Venture Capital" in: Small Business Economics, Vol. 7, No. 2, 153-172
Mayer, C. / Flynn, J. (1973): "Canadian small business abroad: Opportunities, aids and experiences" in: The
Business Quarterly, Vol. 38, 33–47
Mayer, T. / Mucchielli, J.-L. (2004): "Hierarchical location choice and multinational firms strategy: A nested
logit applied to Japanese investments in Europe" Multinational firms : the global-local dilemma, London
McFadden, Daniel (1974): "Conditional logit analysis of qualitative choice behavior" in Zarembka: Frontiers in
Econometrics, New York, 105-142
Megginson, William L. (2004): "Toward a global model of Venture Capital?" in: Journal of Applied Corporate
Finance, Vol. 16, No. 1, 89-107
Meuleman, Miguel / Wright, Mike (2007): "Extending the Staged Model of Foreign Expansion: Cross-Border
Syndication by UK Private Equity Firms in Continental Europe, 1999-2005" in: Academy of Management
Proceedings, 2007, 1-6
Meyer, J. E. / Shao, J. (1995): "International venture capital portfolio diversification and agency costs" in:
Multinational Business Review, 53-58
Mitnick, Barry M. (1997): "Agency theory" in P. Werhane & R. E. Freeman (Eds.): The Blackwell Encyclopedic
Dictionary of Business Ethics, Oxford, UK, 12-15
Morck, R. / Yeung, B. (1991): "Why investors value mulitnationality" in: Journal of Business, 64, 165-187
Muzyka, Dan / Birley, Sue / Leleux, Benoit (1996): "Trade-offs in the investment decisons of European venture
capitalists" in: Journal of Business Venturing, Vol. 11, No. 4, 273-287
 159

Noorderhaven, Niels / Thurik, Roy / Wennekers, Sander / Van Stel, André (2004): "The Role of Dissatisfaction
and per Capita Income in Explaining Self-Employment across 15 European Countries" in: Entrepreneurship:
Theory & Practice, Vol. 28, No. 5, 447-466
Norton, Edgar / Tenenbaum, Bernard H. (1993): "Specialization Versus Diversification as a Venture Capital
Investment Strategy" in: Journal of Business Venturing, Vol. 8, No. 5, 431
NVCA (2009): "Venture impact - The economic importance of venture capital-backed companies to the US
economy": IHS Global Insight, Arlington, VA
Ortgiese, Jens (2007): Value Added by Venture Capital Firms, Köln
Ouchi, William G. (1979): "A Conceptual Framework for the Design of Organizational Control Mechanisms" in:
Management Science, Vol. 25, No. 9, 833-848
Oviatt, Benjamin M. / McDougall, Patricia Philli (1994): "Toward a Theory of International New Ventures" in:
Journal of International Business Studies, Vol. 25, No. 1, 45-64
Phalippou, Ludovic / Gottschalg, Oliver (2005): "Performance of Private Equity Funds" in: EFA 2005 Moscow
Meetings, Moscow, Russia
Pruthi, Sarika / Wright, Mike / Lockett, Andy (2003): "Do Foreign and Domestic Venture Capital Firms Differ
in Their Monitoring of Investees?" in: Asia Pacific Journal of Management, Vol. 20, No. 2, 175-204
Pusterla, Fazia / Resmini, Laura (2007): "Where do foreign firms locate in transition countries? An empirical
investigation" in: Annals of Regional Science, Vol. 41, No. 4, 835-856
Reid, S. D. (1981): "The decision-maker and export entry and expansion" in: Journal of International Business
Studies, Vol. 12, No. 2, 101-112
Reuer, Jeffrey J. / Miller, Kent D. (1997): "Agency Costs and the Performance Implications of International Joint
Ventures" in: Strategic Management Journal, Vol. 18, No. 6, 425-438
Romain, Astrid / Van Pottelsberghe, Bruno (2004): The determinants of venture capital: A Panel Data Analysis
of 16 OECD Countries, Working Paper, Université Libre de Bruxelles, Solvay Business School, April 2004
Ross, Stephen A. (1973): "The Economic Theory of Agency: The Principal's Problem" in: American Economic
Review, Vol. 63, No. 2, 7
Sahlman, William A. (1990): "The structure and governance of venture-capital organizations" in: Journal of
Financial Economics, Vol. 27, No. 2, 473-521
Sapienza, Harry J. (1992): "When do venture capitalists add value?" in: Journal of Business Venturing, Vol. 7,
No. 1, 9-27
Sapienza, Harry J. / Gupta, Anil K. (1994): "Impact of agency risks and task uncertainty on Venture Capitalist-
CEO interaction" in: Academy of Management Journal, Vol. 37, No. 6, 1618-1632
Sapienza, Harry J. / Manigart, Sophie / Vermeir, W. (1996): "Venture capitalist governance and value added in
four countries" in: Journal of Business Venturing, Vol. 11, No. 6, 439-469
Sapienza, Harry J. / Koorsgaard, M. Audrey / Goulet, Philip K. / Hoogendam, Jeffrey P. (2000): "Effects of
agency risks and procedural justice on board processes in venture capital-backed firms" in: Entrepreneurship &
Regional Development, No. 12, 331-351
Scherer, F. M. (1982): "Demand-Pull and Technological Invention: Schmookler Revisted" in: The Journal of
Industrial Economics, Vol. 30, No. 3, 225-237
Schmidt, Daniel (2004): Private equity-, stock- and mixed asset-portfolios: A bootstrap approach to determine
performance characteristics, diversification benefits and optimal portfolio allocations, Center for Financial
Studies, Working Paper
Schmookler, Jacob (1966): Invention and economic growth, Cambridge, MA
Schumpeter, Joseph Alois (1942): Capitalism, Socialism and Democracy, Harper Perennial Modern Classics,
November 2008
Shan, Weijian / Song, Jaeyong (1997): "Foreign Direct Investment and the Sourcing of Technological
Advantage: Evidence From the Biotechnology Industry" in: Journal of International Business Studies, 28, 1997
2nd Quarter, 267-284
160

Shane, Scott / Locke, Edwin A. / Collins, Christopher J. (2003): "Entrepreneurial motivation" in: Human
Resource Management Review, Vol. 13, No. 2, 257-279
Sharpe, W. F. (1964): "Capital asset prices: A theory of market equilibrium under conditions of risk" in: Journal
of Finance, Vol. 19, No. 3, 425-442
Simpson, C. L. Jr. / Kujawa, D. (1974): "The export decision process: an empirical inquiry" in: Journal of
International Business Studies, 107-117
Simpson Jr, CL / Kujawa, D (1974): "The export decision process: an empirical inquiry" in: Journal of
International Business Studies, 107-117
Smith, Adam (1776): The Wealth of Nations, Digireads, 2009
Smith, D. Gordon (1998): "Venture Capital Contracting in the Information Age" in: The Journal of Small &
Emerging Business Law, Vol. 2, No. 1, 134-176
Sørensen, Morten (2007): "How Smart is Smart Money? A Two-Sided Matching Model of Venture Capital" in:
Journal of Finance, Vol. 62, No. 6, 2725-2762
Sorenson, Olav / Stuart, Toby E. (2001): "Syndication Networks and the Spatial Distribution of Venture Capital
Investments" in: American Journal of Sociology, Vol. 106, No. 6, 1546-1588
Sousa, Carlos M. P. / Bradley, Frank (2008): "Cultural distance and psychic distance: refinements in
conceptualisation and measurement" in: Journal of Marketing Management, Vol. 24, No. 5/6, 467-488
Spiller, Pablo T. (1990): "Politicians, Interest Groups, and Regulators: A Multiple-Principals Agency Theory of
Regulation, or "Let Them Be Bribed"" in: Journal of Law and Economics, Vol. 33, No. 1, 65-101
Suzuki, Yoshinori (2007): "Modeling and testing the "two-step" decision process of travelers in airport and
airline choices" in: Transportation Research: Part E, Vol. 43, No. 1, 1-20
Teece, David J. (1992): "Foreign Investment and Technological Development in Silicon Valley" in: California
Management Review, 34, Winter92, 88-106
Thompson, Jim (1967): Organizations in action: social science bases of administrative theory
Train, Kenneth E. (2002): Discrete Choice Methods with Simulation, Working Paper, University of California,
Berkeley, March 2002
Tyebjee, Tyzoon T. / Bruno, Albert V. (1984): "A model of Venture Capital investment activity" in:
Management Science, Vol. 30, No. 9, 1051-1066
Tykvova, Tereza / Schertler, Andrea (2006): Rivals or Partners? Evidence from Europe's International Private
Equity Deals, Discussion Paper No. 06-091, Working Paper, ZEW Mannheim, Dept. of International Finance
and Financial Management, December 2006
Van Osnabrugge, Mark (2000): "A comparison of business angel and venture capitalist investment procedures:
an agency theory-based analysis" in: Venture Capital, Vol. 2, No. 2, 91-109
Wadhwa, Vivek (2008): "Does Venture Capital Spur Innovation?" in: Business Week Online, 10/1/2008, 18
Weidig, Tom / Mathonet, Pierre-Yves (2004): The Risk Profile of Private Equity, European Investment Fund,
Working Paper
Wennekers, Sander / Van Wennekers, André / Thurik, Roy / Reynolds, Paul (2005): "Nascent Entrepreneurship
and the Level of Economic Development" in: Small Business Economics, Vol. 24, No. 3, 293-309
Wilbur, Chung / Alcacer, Juan (2002): "Knowledge Seeking and Location Choice of Foreign Direct Investment
in the United States" in: Management Science, 48, 1534-1554
Wilson, Robert (1968): "The Theory of Syndicates" in: Econometrica, Vol. 36, No. 1, 119-132
Wright, Mike / Robbie, Ken (1998): "Venture Capital and Private Equity: a review and synthesis" in: Journal of
Business Finance & Account, Vol. 25, No. 5, 6, 521-570
Wright, Mike / Lockett, Andy / Pruthi, Sarika (2002): "Internationalization of Western Venture Capitalists into
Emerging Markets: Risk Assessment and Information in India" in: Small Business Economics, Vol. 19, No. 1,
13-29
 161

Wright, Mike / Pruthi, Sarika / Lockett, Andy (2005): "International venture capital research: From cross-
country comparisons to crossing borders" in: International Journal of Management Reviews, Vol. 7, No. 3, 135-
165
Zahra, Shaker A. / Ireland, R. Duane / Hitt, Michael A. (2000): "International Expansion by New Venture Firms:
International Diversity, Mode of Market Entry, Technological Learning, and Performance" in: Academy of
Management Journal, 43, 925-950
Zook, Matthew A. (2002): "Grounded capital: venture financing and the geography of the Internet industry,
1994-2000" in: Journal of Economic Geography, Vol. 2, No. 2, 151-177