«Ulrich Hegey HEC Paris Frédéric Palominoz ENSAE Armin Schwienbacherx University of Amsterdam Business School & Catholic University of Louvain ...»
Venture Capital Performance: The Disparity Between
Europe and the United States
University of Amsterdam Business School & Catholic University of Louvain
We would like to thank Catherine Casamatta, Bruno Crepon, Didier Guennoc, Michael Halling, Jan MahrtSmith, Marco Pagano, Henri Pages, and Alessandro Sembenelli as well as seminar audiences at the European Finance
Association, the First RICAFE Conference, the Journees Banque de France, the Gutmann Center Symposium in Vienna and the University of Toulouse for helpful suggestions. We gratefully acknowledge …nancial support from the Fondation Banque de France and the European Commission (Contract HPSE-CT-2002-00140) as well as support from the European Venture Capital Association.
y HEC School of Management Paris, Dept. of Finance and Economics, 1 rue de la Liberation, 78351 Jouy-enJosas, France. email@example.com, Tel. +33 1 396 772 99.
z ENSAE, 3 avenue Pierre Larousse, 92245 Malako¤ Cedex, France. firstname.lastname@example.org, Tel. +33 1 5504 0005.
x Universiteit van Amsterdam Business School, Finance Group, Roetersstraat 11, 1018 WB Amsterdam, The Netherlands. email@example.com, Tel. +31 20 5257179.
Venture Capital Performance: The Disparity Between Europe and the United States Abstract This paper compares the success of venture capital investments in the United States and in Europe by analyzing individual venture-backed companies and the value generated within the stage …nancing process. We document that US venture capitalists generate signi…cantly more value with their investments than their European counterparts. We …nd di¤erences in contracting behavior, such as staging frequency and syndication, and evidence that they help to explain the observed performance gap and we report a substantial unexplained residual. We …nd that US venture funds investing in Europe do not perform better their European peers. European Common Law and Civil Law countries exhibit comparable levels of venture performance, and di¤erences in stock market development or tax subsidies in favor of venture investments are unrelated to performance di¤erences. European IPO exits from venture investments yield returns similar to the US, while trade sale exits weakly underperform. We attribute the overall performance gap essentially to the segment of poorly performing companies.
Keywords: venture capital performance, …nancial development,stage …nancing, exit, monitoring.
JEL classi…cation: G24; G38.
1. Introduction Venture capital is an American invention, and the United States is home to the largest venture capital industry by far. Venture capital has not spread globally as easily as have other …nancial innovations. What are the necessary conditions for a successful imitation of the US model?
Europe, the world’ second most important region in terms of R&D spending, is an interesting s case to consider. While European governments have exhorted the virtues of venture capital, and designated its development as a key policy priority for more than twenty years, the sector has remained a laggard until recently. A European venture capital (VC) industry geared towards innovation and early-stage …nancing has really only taken hold in the late 1990s,with investments reaching 12 billion dollars in 1999, roughly a quarter of the US level.
This paper proposes a direct comparison between the United States and Europe, and seeks to explore the process by which VC creates economic value on the basis of company-level data.
It contributes to an emerging literature on comparative international studies of venture capital by focusing on possible obstacles to the emergence of a VC industry in developed countries with relatively high levels of R&D spending, investor protection, and law enforcement.
Trade associations and VC professionals have long asserted that realized returns of venture investments in Europe have historically been below required returns, and pointed to this underperformance as the main obstacle to the development of a strong VC industry. Also, the relative lack of venture funding in Europe has been frequently attributed to the absence of attractive and liquid markets for VC exits, in particular for IPOs (e.g., Black and Gilson, 1998). These beliefs form the starting point for our analysis. We investigate whether a measurable performance gap between the United States and Europe exists for the late 1990s when the European VC market was emerging in terms funding levels and exit opportunities. And if so, what explains it? Does the European case hold more general lessons for venture capital development?
More speci…cally, we investigate the period between 1997 and 2003, starting with the year that marks the beginning of rapidly increasing venture funding levels in Europe and ending after European venture investments had peaked. Our focus on this recent period also o¤ers an opportunity to address Black and Gilson’ hypothesis on the absence of su¢ ciently liquid primary s equity markets, since this period was characterized in Europe by a surge in high-tech IPOs as well as the creation of a number of new stock markets geared to high-tech companies markets.
To the best of our knowledge, no comparative analysis of US and European VC performance has been undertaken previously.
Research Design and Results. Based on data from the Venture Economics database, we measure the value generated by every portfolio company in our sample by the Internal Rate of Return (IRR) between the …rst …nancing round and the last round valuation (prior to exit) of the project.
Round valuation data represent an intriguing data source in VC funding as they report the deal pricing in every …nancing round, i.e. actual transactions between stakeholders with con‡icting interests that determine the fraction of equity-linked securities a¤orded to investors in exchange for their cash. Determining VC success on the basis of round valuations o¤ers a methodological advantage when compared with the alternative, analyzing performance until exit: performance measures based on exit valuations will inevitably select only projects that exit with a recorded market valuation; these investments, however, typically constitute an upwards biased sample.
Moreover, since we are interested in the economic value generated by venture investments rather than an assessment of the asset class, our method seems well-suited as it allows to compute the gross valuation e¤ect of venture investments rather than net cash disbursements to limited partners on which asset management studies focus. The focus on round valuations also allows us to perform new tests for data endogeneity based on lead-lag structures of intermediate valuations.
While our focus on early stage investments clustered in a few industries and a limited time period partially mitigates the impact of cross-sectional heterogeneity in risk, we explicitly control for risk, in particular for region-speci…c and technology-speci…c risk, but …nd results that are largely invariant with respect to the choice of risk adjustment. Our main results are the following.
Our data show a gap between the value generated by US venture capital investments and European investments that is statistically highly signi…cant and very large in economic terms. The di¤erence is independent of the choice of performance benchmark.
We observe several di¤erences in the relationship between VCs and entrepreneurial …rms and in their behavior that indicate a more active role of US VCs and a sophisticated cooperation between them. US venture capitalists invest almost twice as much in their portfolio companies, make a larger portion of funding contingent on the completion of the …rst round, organize themselves in larger syndicates, tend to involve corporate VC more frequently, and tend to be more specialized.
These di¤erences can partially explain the observed di¤erence in value creation. We …nd the positive relationship between the frequency of monitoring and performance that theory predicts for the US but a negative relationship in Europe. Also, the amount invested in the …rst round as well as the presence of corporate VC (dimensions where the US dominates) have a signi…cant positive impact on returns. Much of the performance gap, however, remains unexplained by such di¤erences.
US venture funds investing in Europe do not create more value than their European peers.
Venture investments in European Common Law and European Civil Law countries show a comparable level of value creation.
Di¤erences in the tax treatment or the legal environment for venture investments are unrelated to the gap in value creation.
For the subsample of companies with a successful venture exit, we …nd no performance di¤erence for companies exiting through IPOs and only a small di¤erence for companies exiting via trade sales. We conclude that the di¤erence in the value creation process must be primarily due to di¤erences in poorly performing investments as we …nd.
Overall, our results suggest that the United States appears to have a markedly better developed market for VC, with Europe still signi…cantly lagging behind. We test for a wide variety of possible reasons that the …nance literature mentions as possible performance drivers, but …nd that only contracting proxies can explain some of the di¤erence, while causes such as tax treatment, legal systems or stock market development, or the import of experienced VCs, seem to be unimportant. Other factors that our microlevel study based on VC transactions cannot capture may be important.1 These factors could include, among others, the presence of clusters of innovative activity, research output and spillovers from research universities and organizations, cultural attitudes towards risk and propensity to entrepreneurship.
Related Literature. Our work is related to an emerging literature on international and comparative studies. On the whole they report VC contracting practices at variance with the US role model. Lerner and Schoar (2004) look at emerging markets in Eastern Europe and elsewhere and …nd a strong reliance on straight equity and direct board control. Cumming, Schmidt and Walz (2004) show that an increase in legality accelerates the …rst investment, and facilitates syndication and board representation of venture investors for a sample drawn from North and South America, Asia and Europe. Similarly, Cumming, Fleming and Schwienbacher (2006) …nd evidence for Asia-Paci…c in support of the relevance of legal systems for exit choices. Cumming and Walz (2004) report that variables capturing VC, entrepreneur and investment characteristics can account for a substantial fraction of cross-sectional return variations in IRRs. Our …ndings are generally consistent with these studies, but we try to explain internal rates of return simultaneously by company-speci…c and country-speci…c in‡uences. This puts us in a position to argue that contracting features and the legal environment alone are unlikely to explain the distance to the United States.
The majority of empirical studies on VC in Europe are based on questionnaires. In a crosscountry analysis, Manigart et al. (2002) highlight some determinants of required returns on VC investments, and Sapienza, Manigart and Vermeir (1996) examine the impact of the VC governance structure in di¤erent countries. Bottazzi, Da Rin and Hellmann (2004) show that independent, specialized, experienced and highly educated VCs are more likely to be actively involved in the management of portfolio companies. Bottazzi, Da Rin and Hellmann (2005) present questionnaire evidence that across Europe better legal systems (measured by legal origin or rule of law) are associated with more investor involvement, more downside protection for the investors, and more corporate governance involvement of VCs. Schwienbacher (2004) analyzes the determinants of VC exit decisions in a comparison of US and European VC …rms and …nds signi…cant di¤erences, for example on the use of convertibles and syndicate size, con…rming and extending …ndings of Bascha and Walz (2001) for Germany. Da Rin, Nicodano and Sembenelli (2006) study the evolution of aggregate VC investments in fourteen European countries as a function of policy measures and …nd a signi…cant impact of the creation of stock markets geared to entrepreneurial …rms and of capital gains taxations. Mayer, Schoors and Yafeh (2005) investigate the funds raising process in various countries. Schmidt and Wahrenburg (2003) report that reputational e¤ects of venture capitalists are a major determinant of contractual relations between European VC funds and their investors. Unlike our paper, these studies do not investigate VC performance directly.
There has also been a recent literature studying the returns of private equity and VC from an asset pricing perspective, including Kaplan and Schoar (2005), Gottschalg and Phalippou (2007), Ljungqvist and Richardson (2004), and Jones and Rhodes-Kropf (2004). On the whole, this literature shows that private equity returns contain considerable systematic and idiosyncratic risk and that risk-adjusted net returns do not outperform public equity investments.2 There are considerable di¤erences between this strand of work and our study. First, these studies focus on the United States and they mix VC and private equity investments whereas our study is very careful in selecting only VC investments in start-up companies. Second, since our focus is on economic value created rather than risk-adjusted investor returns, we look at investments at the portfolio-company level and we use stage valuations to minimize selection bias. By contrast, these studies look at returns at the fund level,3 they focus on cash distributions and thus on exited investments.4 Third, we focus on an international comparative study in which the US and the European samples are selected according to consistent criteria, whereas those studies have overwhelmingly US-based samples.