The Relevance of Financial versus Non-Financial Information for the Valuation of Venture Capital-Backed Firms

Abstract

This study examines the relevance of financial and non-financial information for the valuation of venture capital (VC) investments. Based on a hand-collected data set on venture-backed start-ups in Germany, we investigate the internal due diligence documents of over 200 investment rounds. We document that balance sheet and income statement items capture as much economic content as verifiable non-financial information (e.g. team experience or the number of patents) while controlling for several deal characteristics (e.g. industry, investment round, or yearly VC fund inflows). In addition, we show that valuations based on accounting and non-accounting information yield a level of valuation accuracy that is comparable to that of publicly traded firms. Further analyses show that the industry-specific total asset multiples outperform the popular revenue multiples but lead to significantly less accurate results than those obtained from the more comprehensive valuation models. Overall, our findings might inform researchers and standard-setters of the usefulness of accounting information for investment companies and provide additional evidence to gauge the overall valuation accuracy in VC settings.

Acknowledgements

The authors are indebted to Margarita Tchouvakhina, Volker Zimmermann, Georg Metzger, Walter Auel, Kerstin Kiehl, Klaus Mark, and Katrin Ullrich for making this research possible. Additionally, they would like to thank Michael Steinmetzer and all of the members of the KfW department Mbf3 for their support. An earlier version of this paper is part of the second author's PhD thesis (Mokwa, 2012). Furthermore, the authors are grateful to Jeff Abarbanell, John Hand, Joachim Gassen, Erik Peek, Bill Rees, Tim Adam, and John Campbell for their valuable comments. This paper also benefited from the seminar participants at the Kenan-Flagler Business School at the University of North Carolina at Chapel Hill, the 2009 European Accounting Association Conference in Tampere, the 2010 European Accounting Association in Istanbul, and the 2010 American Accounting Association in San Francisco. We also owe thanks to two anonymous referees of this journal and Steven Monahan (associate editor). The authors gratefully acknowledge funding from the German Research Foundation and the Institute of Banking and Banking Law, University of Cologne.

Notes

In addition, recent developments, such as the upcoming launch of the ‘BX Venture Market’ segment in 2012 as a separate listing venue from the ‘NASDAQ Stock Market’, highlight the importance of further research in the VC area (see http://www.bxventure.com).

This citation is taken from the so-called ‘Red Book’ edition of the IFRS, that is, International Financial Reporting Standards, as issued on 1 January 2012, Part A, p. A851. Technically the same reference could be made to the so-called ‘Blue Book’ edition of the IFRS, that is, International Financial Reporting Standards, required for annual reporting periods beginning on 1 January 2012. While we acknowledge that IAS 27 (Blue Book) ‘Consolidated and Separate Financial Statements’ does not relieve VC organisations from producing a consolidated financial statement (see IAS 27.IN5, IAS 27.16, IAS 27.BC21 – IAS 27.BC27), they have the option to measure their investments in associates (IAS 28 Blue Book edition) at fair value through profit and loss. In addition, the same holds true with regard to investments held as joint ventures (IAS 31 entitled: ‘Interest in Joint Ventures’, specifically IAS 31.BC4 – IAS31.BC12, pp. 1667–1669).

In our study, we define management experience as a position at a rank of vice president or at a higher rank with a preceding employer.

KfW participates in start-ups jointly with private investors where, in general, the investment partners invest equal amounts. KfW faces identical legal protection and pay-off rights. Put differently, KfW and the VC investors share the same risk in every deal. Consider the case of a start-up firm that requires EUR 2 million in a financing round. This amount is financed 50% by a private VC investor and 50% by KfW. In case this company goes bankrupt, both investors (the private VC and the KfW) will lose EUR 1 million. In other words, the VC investor has at least the same amount at stake as KfW, and regarding the upside potential, the VC also will share the payoffs and other legal control rights with KfW.

BVK is the association of private equity and VC companies operating in Germany and based in Berlin. For more information regarding the statistics, see http://www.bvkap.de/privateequity.php/cat/2/title/Home.

In essence, the bias is caused by the fact that standard linear predictions produce E[log(y|x)], whereas the desired result is E[y|x]. These predictions have the desirable (economic) feature of always being positive because equity values cannot be negative: thus, the approach avoids the problems of using a standard regression (see, for example, Barth et al. (2005, p. 318), where the number of negative predicted equity market values is approximately 10% in each estimation).

The focus on log errors is motivated by the skewness of the percentage errors in the distributions of the underlying fundamentals (Kaplan and Ruback, 1995; Lie and Lie, 2002). Furthermore, Dittmann and Maug (2008) compare different valuation error metrics and confirm that log errors are the most appropriate comparison measures. We also perform our analysis based on absolute percentage errors in our robustness section.

Note that the results in Tables 5 and 6 show absolute log valuation errors (i.e., inaccuracy), whereas Figure 1 shows relative log valuation errors (i.e., bias).

Further information on KfW Bankengruppe and their programmes can be found in: http://www.kfw.de/kfw/en/I/II/Download_Center/Financial_Publications/Annual_Reports.jsp.

Regarding the ERP Start Fund initiated by the KfW Bankengruppe, see: http://www.kfw.de/kfw/en/KfW_Group/Press/Latest_News/PressArchiv/bis11.2005/Pressemitteilung22422.jsp.

Additional information

Notes on contributors

Soenke Sievers

Paper accepted by Steven Monahan.

Christopher F. Mokwa

Paper accepted by Steven Monahan.

Georg Keienburg

Paper accepted by Steven Monahan.