ADVANCES IN FINANCIAL ECONOMICS VOLUME 10
THE RISE AND FALL OF EUROPE’S NEW STOCK MARKETS EDITED BY
GIANCARLO GIUDICI Politecnico di Milano, Dipartimento di Ingegneria Gestionale, Italy
PETER ROOSENBOOM Rotterdam School of Management, Erasmus University, the Netherlands
2004
Amsterdam – Boston – Heidelberg – London – New York – Oxford Paris – San Diego – San Francisco – Singapore – Sydney – Tokyo
THE RISE AND FALL OF EUROPE’S NEW STOCK MARKETS
ADVANCES IN FINANCIAL ECONOMICS Series Editors: Mark Hirschey, Kose John and Anil K. Makhija
ELSEVIER B.V. Sara Burgerhartstraat 25 P.O. Box 211 1000 AE Amsterdam The Netherlands
ELSEVIER Inc. 525 B Street, Suite 1900 San Diego CA 92101-4495 USA
ELSEVIER Ltd The Boulevard, Langford Lane, Kidlington Oxford OX5 1GB UK
ELSEVIER Ltd 84 Theobalds Road London WC1X 8RR UK
© 2004 Elsevier Ltd. All rights reserved. This work is protected under copyright by Elsevier Ltd, and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. Permissions may be sought directly from Elsevier’s Rights Department in Oxford, UK; phone: (+44) 1865 843830, fax: (+44) 1865 853333, e-mail:
[email protected]. Requests may also be completed on-line via the Elsevier homepage (http://www.elsevier.com/locate/permissions). In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (+1) (978) 7508400, fax: (+1) (978) 7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (+44) 20 7631 5555; fax: (+44) 20 7631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of the Publisher is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier’s Rights Department, at the fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. First edition 2004 British Library Cataloguing in Publication Data A catalogue record is available from the British Library. ISBN: 0-7623-1137-1 ISSN: 1569-3732 (Series) ∞ The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed in
The Netherlands.
CONTENTS LIST OF CONTRIBUTORS
vii
PREFACE
ix
VENTURE CAPITAL AND NEW STOCK MARKETS IN EUROPE Giancarlo Giudici and Peter Roosenboom
1
PRICING INITIAL PUBLIC OFFERINGS ON EUROPE’S NEW STOCK MARKETS Giancarlo Giudici and Peter Roosenboom
25
FINANCING GROWTH AND INNOVATION THROUGH NEW STOCK MARKETS: THE CASE OF EUROPEAN BIOTECHNOLOGY FIRMS Fabio Bertoni and Pier Andrea Randone
61
MANAGERIAL INCENTIVES AT THE INITIAL PUBLIC OFFERING: AN EMPIRICAL ANALYSIS OF THE ALTERNATIVE INVESTMENT MARKET Peter Roosenboom
81
THE VALUATION OF FIRMS LISTED ON THE NUOVO MERCATO: THE PEER COMPARABLES APPROACH Lucio Cassia, Stefano Paleari and Silvio Vismara
113
VALUING INTERNET STOCKS AT THE INITIAL PUBLIC OFFERING Michiel Botman, Peter Roosenboom and Tjalling van der Goot
131
v
vi
THE ROLE OF ACCOUNTING DATA AND WEB-TRAFFIC IN THE PRICING OF GERMAN INTERNET STOCKS Andreas Trautwein and Sven Vorstius
157
THE EXPIRATION OF MANDATORY AND VOLUNTARY IPO LOCK-UP PROVISIONS – EMPIRICAL EVIDENCE FROM GERMANY’S NEUER MARKT Eric Nowak
181
UNDERPRICING OF VENTURE-BACKED AND NON VENTURE-BACKED IPOS: GERMANY’S NEUER MARKT Stefanie A. Franzke
201
THE PERFORMANCE OF VENTURE-BACKED IPOS ON EUROPE’S NEW STOCK MARKETS: EVIDENCE FROM FRANCE, GERMANY AND THE U.K. Georg Rindermann
231
THE NEUER MARKT: AN (OVERLY) RISKY ASSET OF GERMANY’S FINANCIAL SYSTEM Hans-Peter Burghof and Adrian Hunger
295
THE LONG-TERM PERFORMANCE OF INITIAL PUBLIC OFFERINGS ON EUROPE’S NEW STOCK MARKETS Giancarlo Giudici and Peter Roosenboom
329
LIST OF CONTRIBUTORS Fabio Bertoni
Politecnico di Milano, Milan, Italy
Michiel Botman
University of Amsterdam, Amsterdam, The Netherlands
Hans-Peter Burghof
University of Hohenheim, Stuttgart, Germany
Lucio Cassia
University of Bergamo, Dalmine, Italy
Stefanie A. Franzke
Center for Financial Studies and J. W. Goethe-University, Frankfurt am Main, Germany
Giancarlo Giudici
Politecnico di Milano, Milan, Italy
Adrian Hunger
University of Munich and Dresdner Bank AG, Munich, Germany
Eric Nowak
University of Lugano, Lugano, Switzerland
Stefano Paleari
University of Bergamo, Dalmine, Italy
Pier Andrea Randone
Politecnico di Milano, Milan, Italy
Georg Rindermann
Allianz Group, Munich, Germany
Peter Roosenboom
Rotterdam School of Management, Erasmus University, Rotterdam, The Netherlands
Andreas Trautwein
WHU Otto Beisheim Graduate School of Management, Vallendar, Germany
Tjalling van der Goot
University of Amsterdam, Amsterdam, The Netherlands
Silvio Vismara
University of Bergamo, Dalmine, Italy
Sven Vorstius
WHU Otto Beisheim Graduate School of Management, Vallendar, Germany vii
PREFACE With the opening of the Nouveau March´e in France in 1996, followed by the Neuer Markt in Germany in 1997 and the Nuovo Mercato in Italy in 1999, the opportunities for small companies to obtain a listing on European exchanges were growing rapidly. Other European countries with new stock markets included Belgium, Denmark, Finland, Greece, Ireland, the Netherlands, Poland, Portugal, Spain, Sweden and Switzerland. These stock markets had one common aim – to attract early stage, innovative and high-growth firms that would not have been viable candidates for public equity financing on the main markets of European stock exchanges. Of these new markets, the Neuer Markt emerged as Europe’s answer to NASDAQ. However, Europe’s new markets met with only limited success. Many markets were unable to attract sufficient numbers of listings to sustain market interest, while others suffered from inadequate rules or poor liquidity. In addition, Europe’s new stock markets were hard-hit by the bursting of the Internet bubble. The market capitalisation of new markets fell to record lows in 2001 and 2002. Insider trading scandals and accounting frauds tarnished the reputation of new markets. As a result, investor confidence quickly disappeared. The most painful consequence has been the closure of EuroNM Belgium in 2001, the German Neuer Markt in 2003 and NASDAQ Europe in 2004. What went wrong? On the one hand, markets for high-growth companies were inherently volatile. The overoptimistic valuations of the Internet bubble had to be corrected. On the other hand, there were more specific reasons for the failure of Europe’s new stock markets. These lightly regulated markets were located at the juncture between private venture capital and main stock exchanges. They could be viewed as “public” venture capital that partially substituted for deficient private venture capital markets in Europe. However, stock market financing lacked the typical provisions such as active monitoring and convenants that are implemented by venture capitalists to protect their investments against information asymmetries and entrepreneurs’ opportunism. At the same time, listing requirements imposed by the new stock markets did not protect investors from scandals and frauds. On paper, the Neuer Markt had the most stringent listing requirements in Europe. Companies had to report quarterly earnings under U.S. or international accounting standards within two months of ix
x
PREFACE
them being available and could issue only common, as opposed to preference, shares. Moreover, insiders had to agree to a six-month lock-up period following the IPO before they could sell their shares. However, the enforcement of these rules was mostly lacking. For example, insiders of some companies listed on the Neuer Markt circumvented the lock-up rules and several companies reported false annual and quarterly reports. In addition, the Neuer Markt did not have kick-out clauses comparable to NASDAQ that allowed it to strike penny stocks from listing. Although many of the Neuer Markt companies became insolvent, it was relatively difficult for these companies to be expelled from the market until October 2001. This meant that these companies continued to tarnish the reputation of the Neuer Markt. This book discusses the rise and fall of Europe’s new stock markets. The book consists of 12 chapters. We will briefly discuss each chapter in turn. Chapter 1, co-authored by Giancarlo Giudici and Peter Roosenboom, describes the development of venture capital and new stock markets in Europe. Markets for high-growth stocks offer venture capitalists a valuable exit opportunity for their investments. This allows them to re-invest their money in other start-up companies and may spur new business creation and technological innovation. They show that the private equity market in Europe today is as large as it was just before the advent of new stock markets in 1997–1999. As such, the need for stock markets that allow private equity investors to divest their equity stakes in growth companies did not disappear. In Chapter 2, Giancarlo Giudici and Peter Roosenboom examine the differences in pricing Initial Public Offerings (IPOs) on Europe’s new stock markets and on the main stock markets of European exchanges. Analyzing a large sample of 1,120 European IPOs, they find that companies that went public on new markets are significantly smaller, younger and riskier than companies that listed on the main markets. They report a 22.3 percentage point difference in the average first-day return of 578 companies that went public on new markets (34.3%) and the average first-day return of 542 companies that went public on main markets (12%). They attempt to explain this difference. Their results show that reduced incentives to control wealth losses and differences in firm and offer characteristics partially explain higher first-day returns on new markets. Their results also show that the opportunity to bundle IPO deals has been important to control underpricing costs on new stock markets. However, a large part of the difference in average first-day return cannot be explained by differences in sample characteristics. Chapter 3, written by Fabio Bertoni and Pierandrea Randone, analyses how capital is raised and employed by a sample of 28 European biotechnology companies listed on Europe’s new stock markets from 1996 to 2000. The authors analyse the financing and the investment policy of these companies, and make a
Preface
xi
comparison with a sample of U.S. biotechnology companies listed on NASDAQ. Their results show different growth patterns among European and U.S. firms. European companies primarily rely on capital raised at the time of the IPO, while U.S. companies also raise a significant fraction of capital in seasoned equity issues after the IPO. Moreover, European companies are more aggressive in investing shortly after the IPO, especially in marketing and operating expenses, than U.S. companies. In Chapter 4, Peter Roosenboom examines the role of managerial incentives in 188 small and entrepreneurial companies that went public on the Alternative Investment Market of the London Stock Exchange. Managerial incentives are measured as the increase in the amount of executive wealth (composed of shareholdings, option holdings and human capital) per £1,000 increase in shareholder wealth. He shows that managerial incentives are higher if the manager co-founded the firm, chairs the board of directors, and has been employed by the firm for a larger number of years. In addition, he identifies a trade-off relation between board monitoring and incentives that is specific to CEOs. He finds that managers with large pre-IPO shareholdings may use the IPO as a wealth diversification opportunity. These managers receive smaller stock options grants and sell more shares in the IPO than other managers. Chapter 5 is written by Lucio Cassia, Stefano Paleari and Silvio Vismara. They study the peer comparable approach used for the valuation of companies that went public on the Italian Nuovo Mercato during 1999–2002. In Italy, IPO prospectuses often report the valuation methods used by investment banks. This allows them to analyze the accuracy of “real-world” valuation estimates. They show that underwriters rely on price-to-book and price-earnings multiples. The valuation estimates generated by these multiples are closest to offer prices. Conversely, when using enterprise value ratios, comparable firms’ multiples are typically higher than those of the firms going public. They argue that underwriters have the possibility to select comparables that make their valuations look conservative. Chapter 6 is devoted to an analysis of Internet-stock IPOs written by Michiel Botman, Peter Roosenboom and Tjalling van der Goot. They investigate the relevance of accounting and other information to valuing Internet IPOs during the years 1998–2000 in Europe and the United States. The authors compare European Internet companies to U.S. Internet firms at the time of the IPO. They find that European firms tend to report a smaller amount of loss in the year before the IPO, sell more shares to the public, have more concentrated ownership by the largest owner and experience lower first-day returns than comparable U.S. Internet firms. They show that market value is negatively related to net income in the Internet bubble period before April 1, 2000 in both European and U.S. IPO markets. This
xii
PREFACE
is consistent with an Internet firm’s start-up expenditures being considered as assets, not as costs. Furthermore, for the U.S. IPO market, they find that free float is value relevant during the Internet bubble. Underwriters and issuers restricted the supply of shares at the IPO. This drove up market prices as investors were keen to buy Internet IPO shares. Chapter 7 is written by Andreas Trautwein and Sven Vorstius. This chapter looks at the value-relevance of accounting data and measures of web-traffic for Internet firms listed on the Neuer Markt at the height of the stock market bubble from October 1999 to May 2000. In doing so, the chapter contributes to the understanding of the investment behavior of market participants during that time. They show that earnings and cash flow cannot explain the valuation of Internet companies, while there is a positive association between total sales and market capitalization. In addition, sales and marketing expenses as well as research and development costs are relevant value-drivers. Furthermore, they find a positive relation between market values and a number of web-metrics such as customer loyalty, reach, page impressions, and unique visitors. The authors conclude that during the Internet bubble measures of web-traffic were at least as relevant as financial data when explaining market values of Internet companies. In Chapter 8, Eric Nowak explores the stock price impact of expirations of lock-up provisions that prevent insiders from selling their shares after the IPO. He examines 172 lock-up expirations of 142 IPOs on Germany’s Neuer Markt. He reports statistically significant negative abnormal returns and a 25% increase in trading volume surrounding lock-up expiration. He is the first to differentiate between the stock price effects of mandatory lock-up provisions and the U.S.-type private lock-up agreements between issuers and underwriters. He refers to the latter as “voluntary” lock-up agreements. He shows that the average negative price reaction is significantly stronger for the expiration of voluntary lock-up agreements than for mandatory prohibitions of disposal. He finds that the negative abnormal returns are larger for firms with high volatility, superior performance after the IPO, low free float and venture capital financed firms. Chapter 9 is authored by Stefanie Franzke. This chapter sheds further light on the role of venture capitalists and underwriters in certifying the quality of a company. She finds that many financial intermediaries are involved in IPOs at the Neuer Markt: 104 underwriters and 148 venture capitalists. In addition, she reports that venture-backed companies are less profitable compared to non venture-backed companies. The pre-IPO owners of venture-backed firms sell significantly more existing shares at the time of the IPO compared to the owners of non venture-backed firms. She finds no evidence for a trade-off between nonunderwriting costs and IPO underpricing. There is no support for the hypothesized certification role of underwriters and/or venture capitalists. It does not seem to pay
Preface
xiii
to hire a prestigious intermediary, at least as far as underpricing is concerned. On the contrary, the involvement of a prestigious venture capitalist is associated with higher underpricing. Chapter 10, written by Georg Rindermann, presents one of the first comparative empirical assessments of the role of venture capitalists in the going public process and their impact on the long-term performance of IPOs in France, Germany and the United Kingdom. His findings suggest that that there are substantial variations in the experience and sophistication of venture capitalists. In particular, international venture capitalists are on average older than national ones, back a larger number of IPOs in the sample, are more often represented on the board, invest with a higher number of syndication partners, and hold larger equity positions in portfolio firms. He reports that venture-backed IPOs do not generally outperform non venture-backed issues. Instead, only a subset of international venture capitalists appears to have positive effects on both the operating and market performance of portfolio firms. The result that venture-backed issues do not commonly outperform non venture-backed ones has an important implication for research on venture capital finance. It indicates that the findings of previous studies on the role of venture capitalists in the U.S. and their influence on the operating and long run market performance of IPO firms can generally not be transferred to European countries. In Chapter 11, Hans-Peter Burghof and Adrian Hunger present a clinical analysis of the German Neuer Markt. The authors document the initial enthusiasm of investors for the Neuer Markt. The deep crisis of the Neuer Markt is attributed to investors’ delusion, to the burst of the Internet bubble and to the numerous cases of frauds and defaults, that sank the image of the growth exchange and caused its closing. The closing of the Neuer Markt and the rebranding and restructuring of the entire Frankfurt stock market indicate the seriousness of the crisis of German public equity markets. Chapter 12 is written by Giancarlo Giudici and Peter Roosenboom. This chapter documents the long-run stock price performance of companies listed on Europe’s new stock markets. They report that the average company that went public on these markets has been a very poor long-term investment. Investors would be left with an average of only 68 cents (72 cents) compared to one euro invested in the local market index (NASDAQ Composite index). The authors test the divergence of opinion hypothesis of Miller (1977) as one possible explanation for why the average company performs so poorly. This hypothesis states that overoptimistic investors initially set market prices above fundamental values (resulting in high first-day returns) and that prices gradually decline to fundamental values over time as more pessimistic investors enter the market. Their results provide some support for the divergence of opinion hypothesis of Miller (1977). In particular, they find
xiv
PREFACE
that IPO underpricing is negatively related to long-run stock price performance. This suggests that investor overoptimism on the first trading day has a transitory effect on prices. Giancarlo Giudici and Peter Roosenboom Editors
VENTURE CAPITAL AND NEW STOCK MARKETS IN EUROPE Giancarlo Giudici and Peter Roosenboom ABSTRACT In this chapter we describe the development of venture capital and new stock markets in Europe. We argue that markets for high-growth stocks offer venture capitalists a valuable exit opportunity for their investments. This allows them to re-invest their money in other start-up companies and may spur the rate of new business creation and technological innovation. The private equity market in Europe today is as large as it was just before the advent of new stock markets in 1997–1999. As such, the need for stock markets that allow private equity investors to divest their equity stakes in growth companies did not disappear.
1. INTRODUCTION For the first time in recent history, in 2000 more companies listed in continental Europe than in the United States. In particular, 727 companies listed on the New York Stock Exchange and on the NASDAQ while about 900 firms went public on European exchanges.1 These statistics are surprising, given that the depth of the U.S. financial markets has been commonly set against the shallow capital markets in continental Europe, dominated by large mature firms, privatising companies and business groups with interlocking ownership (Faccio & Lang, 2002; Franks & Mayer, 1997). Ritter (2003) has highlighted the fast and significant evolution of The Rise and Fall of Europe’s New Stock Markets Advances in Financial Economics, Volume 10, 1–24 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3732/doi:10.1016/S1569-3732(04)10001-7
1
2
GIANCARLO GIUDICI AND PETER ROOSENBOOM
the IPO market in Europe. The reason of the high level of IPO activity in Europe may be related to three recent developments: (i) the de-mutualization of European exchanges, that favoured an aggressive marketing policy towards attracting more firms to the stock market than in the past; (ii) the growth of private equity investments, especially in technology start-ups. These start-up companies went public on stock exchanges, taking advantage of the euphoria for high tech and dot.com stock; (iii) the establishment of new stock markets for growth and technology companies (the Neuer Markt in Germany, the Nouveau March´e in France, the Nuovo Mercato in Italy, EASDAQ/NASDAQ Europe, only to cite the most important new stock markets). Among these recent developments, the third represents the most intriguing one in European financial markets. New stock markets played a crucial role in the rapid expansion between 1998 and 2000, as well as in the dramatic decline in 2001 and 2002, when the market capitalisation of new markets fell to record lows. The most painful consequence has been the closure of the German Neuer Markt in 2003 and NASDAQ Europe in 2004. Why are stock markets for small and high-growth companies important? Several studies suggest that these stock markets help to foster a vibrant venture capital industry by providing a means for venture capitalists to exit their investments. For example, Black and Gilson (1998) argue that the opportunity to exit investments through an Initial Public Offering (IPO) explains the greater vitality of venture capital in the United States. Jeng and Wells (2000) find that IPO activity is the strongest driver of venture capital investments. Increased venture capital investments may lead to a higher pace of technological innovation and business creation. Kortum and Lerner (2000) find that the amount of venture capital activity in an industry significantly increases its rate of patenting. They show that venture capital accounts for about 15% of industrial innovations. Hellmann and Puri (2000) find that the presence of a venture capitalist is associated with a significant reduction in the time taken to bring a product to market, especially for innovators. Michelacci and Suarez (2004) show that the earlier young firms go public the quicker venture capital can be redirected towards new start-ups. Hence a stock market for high-growth firms may encourage business creation. New business creation is in turn important for employment growth (Audretsch, 2002). Stock markets are also important for economic growth. Minier (2000) examines the effect of opening a first national stock exchange on economic growth. She finds that countries that opened stock markets grew faster than similar countries that did
Venture Capital and New Stock Markets in Europe
3
not open exchanges. Levine and Zervos (1998) show that for every ten percentage point increase in the value of share trading, economic growth increases by one percentage point. Rajan and Zingales (1998) show that a high level of financial development increases the rate of new business creation. Subrahmanyam and Titman (1999) argue that when a country’s stock market reaches a critical mass, the market can “snowball” with new firms deciding to list on the stock market, making the market more liquid and efficient, which in turn attracts more firms to go public. Taken together, these studies suggest that stock market development and venture capital are important to economic growth, new business creation and technological innovation. Markets for high-growth stocks offer venture capitalists a valuable exit opportunity for their investments. This allows them to re-invest their money in other start-up companies and may increase the rate of new business creation and the pace of technological innovation. Stock markets can thus serve as catalysts for economic growth and the creation of jobs. This chapter continues as follows. Section 2 discusses the evolution of venture capital, private equity and stock exchanges in Europe. This section also compares the venture capital industry in Europe to that in the United States. In Section 3 we discuss Europe’s new stock markets. Section 4 presents some concluding remarks.
2. THE EVOLUTION OF VENTURE CAPITAL, PRIVATE EQUITY AND STOCK EXCHANGES IN EUROPE This section discusses the development of European financial markets in the 1990s. The statistics reported about investments in venture capital and private equity in Europe will show that there are still strong arguments suggesting that stock markets or exchange segments specifically designed for growth firms should exist. In particular in Europe, the flow of investments in private equity is still larger than in 1996–1998, i.e. the period in which the major new markets started their operations. Therefore, the availability of an exit for such investments, one being an IPO, continues to be important. A much different fate has been experienced by new market indices, in many cases fallen to their record lows, as well as the number of companies seeking to list their shares on new markets that has come to a standstill. European financial markets underwent a major change during the 1990s. The advent of the common currency and the convergence towards homogeneous institutional settings of financial markets contributed to overcome national barriers, in particular in the field of banking and intermediation services. Nonetheless,
4
GIANCARLO GIUDICI AND PETER ROOSENBOOM
no relevant progress towards a single pan-European stock exchange has been observed (European Commission, 2001). The steady increase in cross-border financial investments increased the competition between national exchanges, as highlighted by alliances and acquisitions, sometimes only announced (such as the project of the iX exchange between the Frankfurt and the London stock markets, or the acquisition of the London Stock Exchange by the Stockholm Exchange) and sometimes implemented (such as the Euronext and Norex alliances). The advent of new markets around Europe from 1995 to 2000 is probably the most striking evidence that, in a favourable market momentum, stock exchanges preferred to grow internally instead of going towards integration. The development of new markets coincided, not by chance, with an unprecedented growth of private equity investments and in particular venture capital. Private equity is defined as the investment by professional investors (such as investment banks, closed-end funds, and business angels) in equity capital of private companies that are often owned by a small number of shareholders. Venture capital is a specific form of private equity investment, targeted at start-up companies, in particular in technology sectors. Venture capital is by no means the main source of capital for companies in industrialised countries. For example, from 1990 to 1999, $137 billion has been invested in venture capital activities in the United States. During that same period companies listed on the New York Stock Exchange and NASDAQ raised $500 billion in equity capital (NASDAQ, 2003). In Europe, the venture capital industry is even less developed, to a larger extent if we consider that in the U.S. venture capital statistics exclusively relate to start-up financing (from the seed phase to late-stage development) while European statistics refer to a more general definition that includes buy-out and replacement capital (i.e. purchases of secondary shares). Having said this, in Europe from 1990 to 2000 d85 billion has been invested in venture capital financing, while companies raised more than d200 billion on stock exchanges (European Venture Capital Association, 2003). Figure 1 reports the annual flow of venture capital investments, compared to GDP, in the major European countries, and in the United States, from 2000 to 2002. Interestingly, in 2000 (although it has been a record year for venture capital) investments represented only a small fraction of the countries’ wealth, and only in the U.S. and in the U.K., and in some Nordic countries such as Sweden and Finland, they played a significant role. The fraction has further decreased in 2001 and 2002, reflecting the negative market momentum of financial markets. Although its limited relevance in quantitative terms, venture financing has significantly contributed to the creation of small successful enterprises, in particular in technology sectors, in which the access to external finance is a necessary condition to promote innovation and R&D activity. The European Venture Capital
Venture Capital and New Stock Markets in Europe
5
Fig. 1. Venture Capital Investments in the United States and Europe. Note: Comparison between venture capital investments in the largest European countries and in the United States, as a percentage of Gross Domestic Product (GDP). Source: European Venture Capital Association (2003) and PricewaterhouseCoopers (2003).
Association (2003) underlines that between 1991 and 1995 European venturebacked companies exhibited exceptional growth rates, if compared to the EU largest 500 companies. The sales of venture-backed firms grew at a mean annual rate equal to 25%, twice the rate computed for large companies. The number of employees in venture-backed firms grew annually at a rate of 15%, compared to 2% for large companies. Capital expenditures of venture-backed firms increased every year at a rate of 25%. In 1995, R&D expenses represented 8.6% of sales in venture-backed companies, and only 1.3% of sales in large companies. In a recent survey of 351 companies in their seed, start-up or expansion stage, about 90% have created new jobs (European Venture Capital Association, 2002). Overall, these 351 companies created 16,143 new jobs (an average of 46 per company).
6
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Fig. 2. Annual Venture Capital Investments the United States and Europe. Note: Annual venture capital and private equity investments in Europe (data in d million) and in the United States (data in US$ million) from 1995 to 2002. Source: European Venture Capital Association (2003), National Venture Capital Association (2003), and PricewaterhouseCoopers (2003).
Companies that received seed or start-up capital between 1995 and 2001, have an annual growth rate of 125% over the first four years after venture capital investment, whereas companies that received expansion financing grew by 33% per year during the four year period (European Venture Capital Association, 2002). For all companies, venture funding was followed by a sharp increase in spending on R&D (European Venture Capital Association, 2002). Figure 2 exhibits the annual flow of venture capital investments from 1995 to 2002, in Western Europe and in the United States. For Europe, the total flow of all private equity investments is also reported. After a progressive growth during the 1990s, venture capital fundraising and investments reached record levels in 1999 and 2000 for both the United States and Europe. The National Venture Capital Association (2003) reports that in the United States, $85 billion has been invested in technology start-ups in 2000, +91% with respect to 1999. The flow of all venture capital investments totalled $106 billion. The maximum level has been recorded during the first six months of 2000. In Europe, during the same year, more than 10,000 new start-up companies have been financed with venture capital, totalling about d35 billion (European Venture Capital Association, 2003), +39% compared to 1999, with d19.7 billion directly attributable to pure venture capital financing (+84% compared to 1999). A steady reversal can be observed in 2001. In the United States as well in Europe, during the first quarter of that year, investments declined by 60% with respect to 2000. The halt of the venture capital industry has been confirmed in 2002: in the United States the flow of investments fell below the level of 1998, −49% compared to 2001. In Europe, the trend has been similar for pure
Venture Capital and New Stock Markets in Europe
7
venture capital. However, the total flow in 2002 has been above the level of 1998, and the total private equity investments even increased in 2002, compared to 2001. The fall in funds raised by venture capitalists and private equity investors has been more dramatic. In the United States, in 2000 venture capitalists raised resources for more than $106 billion, but only about $41 billion has been raised in 2001 (−61%) and $7.3 billion in 2002 (−93% with respect to 2000, −82% compared to 2001). In Europe, the capital raised by venture capitalists and private equity investors totalled d48 billion in 2000 (more than the total capital effectively invested in the same year, as shown in Fig. 2), while it reduced to about d38 billion in 2001 (−20%), and d19 billion in 2002 (−50% compared to 2001, −60% compared to 2000). Like the flow of investments, also the flow of fundraising has experienced a less pronounced decrease in Europe than in the United States. Figure 3 reports the evolution between 2000 and 2002 of the investments, by stage of firms’ life cycle, both in the United States (a) and in Europe (b). In the United States the distribution of investments remained rather stable, with a slight reduction in seed and start-up financing, compared to later-stage development. In Europe statistics about venture capital, as mentioned previously, take into account buyout investments (i.e. acquisitions of established firms) and replacement capital (i.e. purchases of secondary shares without subscription of new shares). In 2000, 19% of total investments have financed seed and start-up companies, compared to 8% in 1999, while 27% has been channelled to later-stage investments and 41% to buyouts. In 2001, and to a lower degree in 2002, investments in mature enterprises have been preferred to new ventures. Figure 4 compares the evolution of venture capital investments, by business sector, from 2000 to 2002, in the U.S. (a) and in Europe (b). The figure shows that in addition to a decrease in the flow also a re-allocation of capital among several business sectors has occurred. There is a strong reduction of funds allocated to the Internet business, while both in Europe and in the United States an increase in the investments in the biotechnology and life sciences business is observed. This suggests that after the burst of the dot-com bubble venture capitalists turned their attention to growth opportunities in the biotechnology sector. Figure 4 further highlights that the ratio between professional investments in technology business and traditional sectors in Europe has been different from the United States. In the United States, on average, from 2000 to 2002, more than 65% of the investments were channelled to technology companies, while the opposite is true for Europe, where about 65% of the investments have financed traditional sectors (buildings, agriculture, services, transports, chemicals, mechanics, commerce).
8
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Fig. 3. Distribution of Venture Capital Investments in the United States and Europe. Note: Distribution of venture capital investments in the United States (a) and in Europe (b) from 2000 to 2002, by lifecycle stage of the financed company. Source: European Venture Capital Association (2003) and National Venture Capital Association (2003).
Venture Capital and New Stock Markets in Europe
9
Fig. 4. Industry Composition of Private Equity Investments in the United States (a) and in Europe (b). Comparison between 2000, 2001 and 2002. Source: European Venture Capital Association (2003) and National Venture Capital Association (2003).
In Europe, technology enterprises absorbed financial resources (from venture capital and other private equity sources) for d11.5 billion in 2000, +68% compared to 1999, but seven times lower than in the U.S. (PricewaterhouseCoopers, 2003). Surprisingly, in a general scenario of decreasing investments, in 2001 and 2002 the relative incidence of high-technology investments increased in the United States, while the opposite happened in Europe.
10
GIANCARLO GIUDICI AND PETER ROOSENBOOM
The general tendency in 2001 and 2002 was to give priority to the consolidation of investments in progress, and to select opportunities in the most promising business, such as biotechnology. The relative incidence of the biotechnology sector on total investments increased significantly both in the United States and in Europe. Finally, it is interesting to consider the evolution of the divestments flow. In the United States the negative market momentum and the burst of the Internet bubble twice penalized professional investors. On the one hand, divestments through IPOs on stock exchanges fell dramatically ($4 billion in 2001 and only $2.5 billion in 2002, compared to $25 billion in 1999 and almost $28 billion in 2000). On the other hand, the lack of profits in Internet and technology companies, forced professional investors such as venture capitalists to write-off their portfolio investments in these companies. This generated sizeable losses for professional investors. In Europe in 2001 divestments totalled d12.5 billion, compared to d9.1 billion in 2000. The most frequent exit route has been the trade-sale that accounted for d4.2 billion, compared to d3 billion in 2000. Divestments through IPOs on stock exchanges accounted for d250 million vs. d570 million in the previous year. In 2002 divestments totalled d8.1 billion, of which 1% came from IPOs, 29.8% from tradesales, and 28.5% from write-offs and capital losses (compared to 23.2% in 2001). In sum, notwithstanding the negative cycle momentum following 2000, the private equity market in the United States and especially in Europe is as large as it was just before the advent of new markets (1997–1999). As such, the need for specialised exchanges that allow private equity investors to divest their equity stakes in growth companies did not disappear.
3. EUROPE’S NEW STOCK MARKETS The U.S. economy has shown strong momentum in the 1990s. Interestingly, much of the positive push derives from the contribution of small firms. According to the USA Small Business Administration,2 from 1990 to 1999 small enterprises created 12 million new jobs, versus a reduction of 650,000 jobs in large enterprises. The growth has been particularly strong in high-technology sectors (Audretsch, 2002). During the same period, continental European firms’ growth rates have been significantly lower, and employment did not receive significant impulses. Several explanations have been suggested for the lower growth rates of European companies. The less-developed capital markets in Europe can be viewed as one of the primary causes. The scope of capital markets in Europe is thought to be limited because of the poor legal protection of shareholders that reduces the willingness of investors to hold securities in European companies (La
11
Venture Capital and New Stock Markets in Europe
Porta et al., 1997). From 1992 to 1995, more than 2,200 companies listed on U.S. exchanges, while in Europe only 800 firms went public during the same period. Moreover, firms listed in Europe during that period have been essentially large and mature companies, with no need to raise capital, scarce growth opportunities, and seeking to reduce their debt burden (Pagano et al., 1998). European technology firms seeking to raise equity capital preferred to list in the United States. Pagano et al. (2002) show that from 1986 to 1997 the number of European companies listed in the U.S. has been growing, while the number of U.S. firms listing in Europe has been decreasing. In 1998, 135 EU companies had a listing on the NASDAQ (52 from the U.K., 18 from the Netherlands, 15 from Sweden, 14 from Ireland, 3 from Germany and 2 from Italy). The national governments and the European Commission quickly became aware of the need to reform financial markets, and pointed to the poor development of venture capital and equity investments as one the fundamental reasons for Table 1. Europe’s New Markets Compared with NASDAQ (January 1, 2003). Market
Country
TechMark/AIMb Neuer Markt Nuevo Mercado Nouveau March´e/Euronext Nuovo Mercato Sitech NASDAQ Europe (EASDAQ) ITEQ SWX new market KVX growth market Nya marknadend NM-list Nieuwe Markt/Euronext New market NEHA Euro-NM Belgium/Euronext Novo Mercado/Euronext
UK Germany Spain France Italy Poland Belgium Ireland Switzerland Denmark Sweden Finland The Netherlands Greece Belgium Portugal
NASDAQ
United States
Total Capitalisationa
Capitalisation/ GDP(%)
Listed Companies
391,395 9,928 9,576 6,954 6,438 5,327 3,043 834 630 598 513 290 379 122 57 –
27.52 0.54 1.64 0.53 0.59 3.02 n.a.c 0.81 0.25 0.37 0.23 0.24 0.01 0.01 0.02 –
914 240 13 147 45 24 40 8 9 10 17 15 11 5 11 –
1,994,494
19.91
3,649
Source: Federation of European Stock Exchanges (2003), NASDAQ (2003), Internet URLs of exchanges. a Data in d million. b Statistics for AIM only are as follows: 704 listed companies, total capitalisation d15,760 million. c The ratio is not significant, since the majority of listed companies are foreign firms. d Nya Marknaden is not a regulated market of the Stockholm Exchange, although it shares the same trading platform.
12
GIANCARLO GIUDICI AND PETER ROOSENBOOM
the inferior competitiveness of EU companies in comparison to their U.S. counterparts (European Commission, 1993, 1995). The strategic priority of developing private equity investments has been repeated in the report written by the Lamfalussy Committee (European Commission, 2001) and by the Green Paper on entrepreneurship (European Commission, 2003). The reform of stock exchanges has been one of the urgent topics in the reform agenda. In a few years, ad hoc exchanges for small growing companies (the new markets) opened in most EU countries, targeted particularly at new technology-based small firms, that apparently could not find their way to existing established stock markets. Some of these new markets attempted to join in a pan-European network (Euro.NM) so as to provide cross-national visibility for listed companies. However, the Euro.NM network was disbanded in 2000. Table 1 lists all new markets established in Europe since 1995, and the most relevant statistics (capitalisation and listed companies) are reported (as at January 1, 2003). The main distinctive features of new markets, compared to the main boards of the exchanges are the listing requirements and the mechanisms in order to provide liquidity. The two parts of Table 2 compare listing requirements of the major new markets in continental Europe, with those of the corresponding main boards. Generally, listing requirements of new markets are less tight than on main boards, both for company age and for profitability. In fact, new markets are targeted at young small firms that are not profitable at the time of the IPO, but that have large growth opportunities that need to be financed. In general, EU stock exchanges require a track record of three years to firms listing on main boards (two years on the French Second March´e), while in new markets the minimum age is one year. Only the German Neuer Markt and Euro.NM Belgium require three years. However, in many cases exceptions are tolerated by these two new markets and even start-up companies (with a track record of less than one year) have been admitted to listing. As far as the company size is concerned the German, French, Dutch and Italian exchanges require a book value of the equity capital lager than d1.5 million. Euro.NM Belgium requires an expected market capitalisation larger than d2 million. NASDAQ Europe sets alternative rules, combining expected capitalisation, equity capital and sales. On the main boards, requirements are much more diversified (the equity book value must be larger that d1.25 million in Germany, but larger than d15 million in Belgium and France, and more than d17 million in Switzerland). New markets belonging to the former Euro.NM network as well as the Swiss SWX New Market require IPO companies to raise new equity with the issue of primary shares (at least 50% of the offer proceeds, that must be larger than d5 million).
Market
Country
Firm Age and Size
Public Offering
Floating Capital
Lock-Up Provisions
At least half of the IPO shares must be newly issued; IPO proceeds larger than d5 million At least half of the IPO shares must be newly issued; IPO proceeds larger than d5 million No specific rule; dual offerings and listings on NASDAQ are welcome
25% or 10% if IPO proceeds are larger than d5 million (with at least 100,000 shares) 20% (with at least 100,000 voting shares)
Compulsory for 6 months
20% owned by at least 100 different shareholders
Compulsory, for 6 months on at least 80% of the shares
20% (with at least 100,000 voting shares)
Compulsory for 12 months on at least 80% of the shares
20% (with at least 100,000 voting shares)
Not compulsory
25% (in some cases 10%)
Compulsory for 12 months
20% (with at least 100,000 voting shares)
Compulsory for 6 months
(a) Listing Requirements on Europe’s New Stock Markets (December, 2002) Neuer Markt
Germany
Three years; equity book value larger than d1.5 million
Nuovo Mercato
Italy
One year; equity book value larger than d1.5 million
EASDAQ (NASDAQ Europe)
Belgium
Nouveau March´e (Euronext Paris)
France
Equity book value larger than d10 million and gross income equal to at least d1 million, or equity book value larger than d20 million, or market capitalisation larger that d20 million, with revenues larger than d50 million Equity book value larger than d1.5 million
Nieuwe Markt (Euronext Amsterdam)
The Netherlands
Equity book value larger than d1.5 million
Euro.NM Belgium (Euronext Brussels)
Belgium
Three years; expected market capitalisation no lower than d2 million
SWX new market
Switzerland
One year; equity book value larger than CHF 2.5 million (d1.7 million)
13
At least half of the IPO shares must be newly issued; IPO proceeds larger than d5 million At least half of the IPO shares must be newly issued; IPO proceeds larger than d5 million At least half of the IPO shares must be newly issued; IPO proceeds larger than d5 million At least half of the IPO shares must be newly issued
Compulsory for 12 months on at least 80% of the shares
Venture Capital and New Stock Markets in Europe
Table 2.
14
Table 2. (Continued ) Market
Country
Firm Age and Size
Public Offering
Floating Capital
Lock-Up Provisions
No specific rule
At least 25% (20% on SMAX segment); no specific rule for the Geregelter Markt
Not compulsory
No specific rule
At least 25% (35% on STAR segment) At least 10% (25% on Premier March´e)
Not compulsory
(b) Listing Requirements on Europe’s Main Stock Market Segments (December, 2002) Germany
Mercato Telematico Azionario Premier March´e Second March´e (Euronext Paris) Offici¨ele Markt (Euronext Amsterdam)
Italy France
The Netherlands
Eerste Markt (Euronext Brussels)
Belgium
SWX Swiss Exchange
Switzerland
Three years (no constraint for Geregelter Markt); expected capitalisation no lower than d1.25 million (minimum equity book value equal to d250,000 for Geregelter Markt) Three years; expected capitalisation larger than d5 million Two years; recommended expected capitalisation larger than d15 million Three years; equity book value larger than d5 million; the company must have reported profits at least three times during the last five yearsa Three years; expected capitalisation larger than d15 million
Three years; expected capitalisation larger than CHF25 million (d17.2 million)
No specific rule
No specific rule
At least 10%
No specific rule
No lower than 10% (25% recommended). Floating capital must capitalise at least d5 million At least 25%
No specific rule
Source: Internet URLs of stock exchanges. a The requirement about profitability does not apply if the firm’s expected capitalisation is larger than d150 million.
Not compulsory
Compulsory (for at least 180 days) only if the company reports losses Not compulsory
Not compulsory
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Amtlicher Handel Geregelter Markt (Deutsche B¨orse)
Venture Capital and New Stock Markets in Europe
15
As on main markets, companies that go public on new stock markets must appoint a sponsor that supports the company in its communications with the financial community. The liquidity is provided by market makers and brokers that are appointed by the companies. Their role is to continuously display bid and ask prices for the shares. The recommended floating capital, that may be potentially traded on the market, varies from 20 to 25% (in some cases, for large capitalisation companies, 10% is enough). With respect to this requirement, there are no significant differences between main and new markets. In order to reduce information asymmetries with investors, firms and sponsors are often requested to maintain a regular flow of information towards the market by organizing regular meetings with analysts and by providing research coverage. A further guarantee is provided through lock-up contracts, in which pre-IPO shareholders refrain from selling (part of) their shares after the listing, for a given number of months. The provision aims at avoiding excess in the supply of shares in the aftermarket (that could depress the share price) and above all at preventing insider trading. In fact, inside investors could be tempted to take their firm public and sell shares in order to take advantage of temporarily overoptimistic valuations. By committing to hold their shares, they signal their favourable expectations about the firm’s value in the future. Lock-up provisions are compulsory on almost all new markets, for a period varying from 6 to 12 months, on a fraction of shares held by pre-IPO owners varying from 80 to 100%. In none of the counterpart main boards lock-ups are compulsory (the only exception is the Dutch stock exchange, that requires a lock-up provision for unprofitable firms). In the following sections the main characteristics of the major European new markets will be described in detail.
3.1. AIM and the TechMark Segment (United Kingdom) In Europe the first successful stock market for small and medium size companies, albeit not necessarily technology companies, has been the AIM (Alternative Investment Market), established in 1995 by the London Stock Exchange. The special feature of AIM companies is the size, sometimes very small. In fact, no minimum capitalisation and floating capital are requested to list on this exchange. The only listing requirement is the adoption of accounting standards in accordance with the international generally accepted rules. Listing companies have to appoint an advisor, guaranteeing the quality of the company to investors, and a broker
16
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Fig. 5. AIM (Alternative Investment Market) and TechMark Segment.
providing liquidity with bid and ask prices. Shareholders are requested to lock-up their shares for at least 12 months from the listing. The AIM took the place of the Unlisted Securities Market (USM), an unregulated market that was set up in 1980. Figure 5 shows that new listings on AIM have been numerous, not only up to 2000, but also in the following months, despite the negative market momentum, described by the FTSE AIM market index performance, reported in the same figure. In 1999 the London Stock Exchange opened TechMark, a specific segment of the main board targeted at high-technology companies. This segment has been joined by companies in the following businesses: computer hardware, computer services, Internet, semiconductors, software, and telecommunications. The listing on the main board of the London Stock Exchange precedes the admission at the TechMark segment. As a rule, companies must therefore be at least three years old. However, high growth companies with an expected market capitalisation of at least £50 million and offer proceeds of at least £20 million may be exempted from this listing requirement.
3.2. Nouveau March´e (France), Euro.NM Belgium and Nieuwe Markt (the Netherlands) The Nouveau March´e started its activity in France in March 1996, alongside the main market (Premier March´e) and the second market (Second March´e). Candidate firms should exhibit a book equity value no lower than d1.5 million. The IPO proceeds should be no lower than d5 million, of which at least 50% from primary newly issued shares. The floating capital must be equal to at least 20% (divided
Venture Capital and New Stock Markets in Europe
17
in at least 100,000 shares). Pre-IPO shareholders must comply with a lock-up provision on at least 80% of their shares for a period of 12 months. Euro.NM Belgium opened in January 1997. Firms going public on this exchange must be at least three years old, and capitalise at least d2 million. In the Netherlands, the Nieuwe Markt (NMAX) opened in February 1997. Listing requirements are the same imposed by the Nouveau March´e in Paris, although lock-up provisions are not compulsory for all the listing companies. The three new markets joined to form the Euro.NM alliance in 1997, with the aim to “improve the cooperation between EU exchanges in order to create and develop a pan-European stock market for growth companies.” Afterwards, the network has been joined also by the German and Italian new markets, but the substantial failure of the objective to establish a pan-European exchange caused its abandonment, leaving full autonomy to single national new markets. A further obstacle towards the integration has been the establishment of the Euronext alliance in September 2000, grouping the exchanges in Paris, Amsterdam and Brussels, and later Lisbon. In the context of the restructuring plan of the merged exchanges, the largest companies listed on the respective new markets from January 2002 have been included in the NextEconomy segment (designed for high-tech companies) and in the NextPrime segment (designed for growth companies in traditional business). Figure 6 describes the evolution of the number of listed companies from 1996 to 2002 on the three Euronext new markets, and the market performance of the largest one, the Nouveau March´e. While the Dutch and Belgian exchanges did not flourish, the French Nouveau March´e progressively grew up to the level of 160 listed companies, although in 2002 the number has been reduced by the transfers of some companies to the main boards of the Paris Bourse.
Fig. 6. Nouveau March´e, EuroNM Belgium and Nieuwe Markt.
18
GIANCARLO GIUDICI AND PETER ROOSENBOOM
3.3. NASDAQ Europe (EASDAQ) EASDAQ (European Association of Securities Dealers Automated Quotation) was founded in 1996 by a group of financial and banking intermediates, supported by the European Venture Capital Association (EVCA), with the aim to “promote the economic development and the innovating activity of young technology firms, looking forward to financing their growth.” Listing requirements are significantly different from other European new markets. In detail, it is necessary that the firm complies with one among the following criteria: (i) accounting value of the total assets no lower than d5 million, and gross income no lower than d1 million; or (ii) equity book value larger than d20 million, or (iii) expected capitalisation no lower than d20 million with sales no lower than d50 million. In some respects, the admission requirements are tighter than on other European new markets. The floating capital must be no lower than 20%, owned by at least 100 different investors. Lock-up provisions are compulsory on 80% of the shares owned by pre-IPO shareholders, for at least 6 months. No particular conditions for the public offerings are imposed. A special feature of EASDAQ is that at least two market makers should compete in order to provide liquidity. Dual listing on the U.S. NASDAQ is encouraged. Figure 7 describes the progress of the number of companies listed on EASDAQ from 1996 onwards and of the index market performance. In 2000, after an initial expansion, the exchange experienced a continued reduction in the number of listed firms, due to the competition with national new markets. Several firms abandoned the market, because of takeovers or defaults, while only a few new companies were admitted to trading. In March 2001, with the aim to boost its activity, the market has been acquired by NASDAQ, and changed its name in NASDAQ Europe. The market index performance has been one of the worst among all European new markets.
Fig. 7. EASDAQ/NASDAQ Europe.
Venture Capital and New Stock Markets in Europe
19
NASDAQ Europe clearly failed its objective to build a cross-national exchange, and to provide visibility to foreign investors for their firms. The success of the national new markets demonstrated that a single pan-European exchange would remain a distant vision. The most striking evidence is the number of listed firms that chose to de-list from the NASDAQ Europe in 2001 and 2002 and to concentrate the trading in their national exchanges (for example, the companies AISoftw@re, Innogenetics, Ubizen, Melexis). In 2004 NASDAQ Europe closed its doors.
3.4. The Neuer Markt (Germany) The story of the German Neuer Markt, from its opening in January 1997 up to the announcement of its closure in 2003, illustrates the rise and fall of all Europe’s new markets. The Neuer Markt succeeded to attract hundreds of companies in a few years time, and became the largest new market in continental Europe. At its height, in March 2000, the capitalisation exceeded d200 billion, equal to about 10% of German GDP. Firms listing on the Neuer Markt must be three years old (but exceptions have been frequent), exhibit a book equity value larger than d1.5 million and a floating capital equal to 25% (10% if the company capitalisation is sufficiently large). Half of the IPO shares must be newly issued. Lock-up covenants on a 6-months basis are compulsory. The listing firm is requested to publish the IPO prospectus both in German and in English, and must commit to organise regular meetings with analysts. An advisor (Betreuer) and a broker acting as a market maker must be appointed. Figure 8 describes the growth of the market index performance and the increase in the number of listed companies, that grew close to 350 companies, but soon
Fig. 8. Neuer Markt.
20
GIANCARLO GIUDICI AND PETER ROOSENBOOM
declined because of defaults and distressed firms expelled from listing for their low level of capitalisation. The scandals that lead to the collapse of the Neuer Markt are numerous. Its closure in 2003 and the consequent transfer of the listed companies to the main segments of the Deutsche B¨orse have been justified by the loss of investor confidence.
3.5. The Nuovo Mercato (Italy) The debate about the incapability of the Milan Stock Exchange to attract firms, in particular small and medium size enterprises, has been long-lived in Italy. The number of privately owned industrial companies that can be potentially admitted to the listing is much larger than the number of companies actually floated on the Italian Exchange. The first initiative, with the opening of a second market (Mercato Ristretto) in 1977, to attract small capitalisation companies substantially failed. Figure 9 describes the evolution of the Mercato Ristretto, in terms of new listings, from 1995 to 2002, compared with numbers from the Nuovo Mercato. Figure 9 demonstrates the slow but consistent decline of the number of companies listed on the Italian Mercato Ristretto. The main reason of such failure is a process of adverse selection. The best companies listing on this second market, after a period of seasoning, transferred to the main board of the exchange, while their mediocre counterparts remained, causing an impoverishment of the market. In fact, the image of the Mercato Ristretto has never been associated to dynamic ambitious firms, but to small co-operative banks and local utilities strongly influenced by public administrations.
Fig. 9. Mercato Ristretto and Nuovo Mercato.
Venture Capital and New Stock Markets in Europe
21
Only in 1998, after the demutualization of the stock exchange, reforms gained new momentum. Listing requirements (that admitted to the trading only companies with a three-years track of profits) have been relaxed. In fact, in the previous years several Italian technology companies (such as Algol, AISoftw@re, Union Technology, Gruppo Formula, Instrumentiation Laboratory and Orthofix International) had decided to go public on foreign exchanges such as the EASDAQ, the NASDAQ or the French Nouveau March´e, because they could not find an appropriate listing venue on the Italian Exchange. The Nuovo Mercato opened in 1999, joining the Euro.NM network, after a debate about which choice should be better for the Italian Exchange (promote the panEuropean market EASDAQ, or promote a national new market). The development of the Nuovo Mercato coincided with the height of investors’ euphoria for Internet and high-tech stock, so that in a few months the Italian new market exceeded the capitalisation of the French counterpart, born three years before. From the second half of 2000, the Nuovo Mercato suffered the world crisis of stock markets, so that only a handful of companies listed in 2001 and none in 2002 (see Fig. 9). The admission requirements imposed by the Nuovo Mercato are equal to the ones requested by the French Nouveau March´e. However, the exchange also accepted some start-up companies, less than one year old. In this case, lock-up provisions must be extended for one year on 100% of the equity capital owned by pre-IPO shareholders and on 80% for two years. The Italian new market, despite the significant fall of the market index, did not experience cases of firms’ default. However, in some cases relevant restructuring plans have been implemented because of financial distress. In other cases, investors discovered that false information was given in the IPO prospectuses about company sales.
3.6. New Markets in Other European Countries Between 1999 and 2000, several European exchanges opened market segments for growth and technology companies, imitating the experience of the largest new markets. In Spain, the Nuevo Mercado was established in April 2000. The first 10 companies listed on the Nuevo Mercado transferred from the main board of the exchange, and were operating in high-tech sectors. The largest company by far has been Terra Networks (now Terra-Lycos). The Spanish Nuevo Mercado is the only new market, among the others, accepting the flotation only of profitable companies. In Portugal, the Bolsa Valores in Lisbon joined the Euronext alliance in 2001, opening a specific market for growth companies (the Novo Mercado), although at the moment no companies are listed on this new market.
22
GIANCARLO GIUDICI AND PETER ROOSENBOOM
In Nordic countries the Stockholm Exchange opened the Nya Marknaden in 1997, but differently from other new markets, it is an over-the-counter segment exploiting the same trading platform of the official market. In Finland there is a stock market for small capitalisation firm (I-List) but in 1999 a new market for growth companies opened (NM-List). Denmark opened its own KVX Growth Market in September 2000. Ireland followed with the ITEQ market. In Switzerland, the SWX New Market started to operate in 1999, hosting several foreign (and even large) companies especially in the biotech/pharmaceutical sector. In 2001, Greece opened the market NEHA (or NEXA) for innovative companies. In Eastern Europe, Poland opened a segment in the Warsaw Exchange for technology firms (SiTECH) in 2000. New markets have been established outside Europe as well. In November 1999 Japan established an exchange for entrepreneurial start-ups (Mothers) as well as Singapore (SESDAQ), Hong Kong (Growth Enterprise Market), Brazil (Novo Mercado) and the Vancouver Exchange in Canada (Canadian Venture Exchange).
4. CONCLUDING REMARKS In 1996, the lack of an early exit route for venture capitalists from youngtechnology based companies was described as the Achilles’ heel of Europe (Financial Times, March 5 1996). In 1997, a survey of the European Venture Capital Association showed that 70% of venture capitalists experienced difficulties in exiting their investments. Many venture capitalists considered an IPO as the ideal exit route because it was one of the most profitable exits for them and allowed incumbent management to stay in charge. But at that time it was considered not a viable option for many small companies (European Venture Capital Association, 1997). The new markets were a first attempt to offer venture capitalists an attractive exit route for their investments in early stage companies. These markets are important from a policy perspective. Previous research has shown that well-developed stock markets lead to a vibrant venture capital industry that in turn spurs technological innovation and new business creation (Black & Gilson, 1998; Michelacci & Suarez, 2004). In addition, a high level of financial development can act as a catalyst for economic growth (Levine & Zervos, 1998; Rajan & Zingales, 1998). It may also help to create new jobs. For example, Seifert (2002) estimates that the Neuer Markt in Germany has helped to create a total of 700,000 jobs. But how do you get sustainable markets for growth stocks? One possible answer may be to increase transparency and investor protection. Increasing the level and scope of disclosure increases the accuracy of asset pricing and may
Venture Capital and New Stock Markets in Europe
23
restore investor confidence (Fox, 2000; La Porta et al., 1997). Although on paper Europe’s new stock markets had stringent disclosure regimes, the enforcement of these rules left a lot to be desired. Venture capitalists and investment banks brought companies to the market without adequately informing investors about the risks. European stock exchanges have recognised the importance of effective disclosure. For example, the Frankfurt Stock Exchange has established a Prime Standard segment. Companies that list on this stock market segment have to meet additional disclosure requirements. The private equity market in Europe is as large as it was just before the advent of new markets (1997–1999). As such, the need for stock markets that allow private equity investors to divest their equity stakes in growth companies did not disappear. A recent survey of Grant Thornton (2002) shows that 10% of medium sized companies in Europe prefer going public as an option for change of ownership in the next three to five years. Among medium sized companies the biggest impediment to an IPO is their company size. Companies believe they should ideally have a value of d21–50 million before flotation. This highlights the continued need for stock markets in Europe that allow small and medium sized companies to go public.
NOTES 1. Source: Federation of European Stock Exchanges (2003). Statistics exclude the United Kingdom and are adjusted for Spain, where investment funds are considered as new listings. 2. See the statistics reported by http://www.sba.gov.
REFERENCES Audretsch, D. B. (2002). The dynamic role of small firms: Evidence from the US. Small Business Economics, 18, 13–40. Black, B. S., & Gilson, R. J. (1998). Venture capital and the structure of capital markets: Banks versus stock markets. Journal of Financial Economics, 47, 243–277. European Commission (1993). White paper on growth, competitiveness and employment: The challenge and ways forward into the 21st century. European Commission (1995). Green paper on innovation. European Commission (2001). Final report of the Committee of Wise Men on the regulation of european securities markets. European Commission (2003). Green paper: Entrepreneurship in Europe. http://europa.eu.int/comm. European Venture Capital Association (1997). Better exits. http://www.evca.com. European Venture Capital Association (2002). Survey of the economic and social impact of venture capital in Europe. http://www.evca.com.
24
GIANCARLO GIUDICI AND PETER ROOSENBOOM
European Venture Capital Association (2003). Yearbook. http://www.evca.com. Faccio, M., & Lang, L. H. P. (2002). The ultimate ownership of Western European corporations. Journal of Financial Economics, 65, 365–395. Federation of European Stock Exchanges (2003). Information and statistics. http://www.fese.be. Fox, M. (2000). The securities globalization disclosure debate. Washington University Law Quarterly, 78, 567–596. Franks, J., & Mayer, C. (1997). Corporate ownership and control in the UK, Germany, and France. Journal of Applied Corporate Finance, 9, 30–45. Grant Thornton (2002). European new markets guide 2002. Hellmann, T., & Puri, M. (2000). The interaction between product market and financing strategy: The role of venture capital. Review of Financial Studies, 13, 959–984. Jeng, L. A., & Wells, P. C. (2000). The determinants of venture capital funding: Evidence across countries. Journal of Corporate Finance, 6, 241–289. Kortum, S., & Lerner, J. (2000). Assessing the contribution of venture capital to innovation. Rand Journal of Economics, 31, 674–692. La Porta, R., Lopez-De Silanes, F., Shleifer, A., & Vishny, R. W. (1997). Legal determinants of external finance. Journal of Finance, 52, 1131–1150. Levine, R., & Zervos, S. (1998). Stock markets, banks, and economic growth. American Economic Review, 88, 537–558. Michelacci, C., & Suarez, J. (2004). Business creation and the stock market. Review of Economic Studies, 71, 459–481. Minier, J. (2000). Opening a stock exchange. Working Paper, University of Miami. NASDAQ (2003). The NASDAQ-AMEX fact book & company directory. http://www.nasdaq.com. National Venture Capital Association (2003). Venture Capital Yearbook. http://www.nvca.com. Pagano, M., Panetta, F., & Zingales, L. (1998). Why do companies go public? An empirical analysis. Journal of Finance, 53, 27–64. Pagano, M., R¨oell, A. A., & Zechner, J. (2002). The geography of equity listing: Why do companies list abroad? Journal of Finance, 57, 651–694. PricewaterhouseCoopers (2003). Money for growth – Technology investment report. http://www.pwcmoneytree.com. Rajan, R., & Zingales, L. (1998). Financial dependence and growth. American Economic Review, 88, 559–586. Ritter, J. R. (2003). Differences between European and American IPO markets. European Financial Management, 9, 421–434. Seifert, W. G. (2002). That damned economic miracle or the return of Dr. Mabuse. The Finance Foundation News, 3, 46–57. Subrahmanyam, A., & Titman, S. (1999). The going public decision and the development of financial markets. Journal of Finance, 54, 1045–1082.
PRICING INITIAL PUBLIC OFFERINGS ON EUROPE’S NEW STOCK MARKETS Giancarlo Giudici and Peter Roosenboom ABSTRACT In this chapter we investigate whether the pricing of IPOs on Europe’s new stock market differs from that of IPOs on main market segments. We report a 22.3 percentage point difference in the average first-day return of new market IPOs (34.3%) and the average first-day return of main market IPOs (12%). We show that reduced incentives to control wealth losses and different firm and offer characteristics partially explain the higher average first-day return on new market segments. We also find that the bundling of IPO deals has been more important to control underpricing costs on new market than on main market segments.
1. INTRODUCTION During the latter half of the 1990s new stock markets designed for high-growth and high-tech fledgling companies have been established around Europe: the French Nouveau March´e (first listing March 20, 1996), the German Neuer Markt (first listing March 10, 1997), Euro.NM Belgium (first listing April 11, 1997), the Dutch Nieuwe Markt (first listing March 25, 1997), the Italian Nuovo Mercato (first listing June 17, 1999) and EASDAQ (now NASDAQ Europe, first listing November 27, 1996).1 From 1996 to 2002, 675 companies listed on these
The Rise and Fall of Europe’s New Stock Markets Advances in Financial Economics, Volume 10, 25–59 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3732/doi:10.1016/S1569-3732(04)10002-9
25
26
GIANCARLO GIUDICI AND PETER ROOSENBOOM
exchanges, mostly after an IPO (Initial Public Offering). The admission criteria of new market segments are different and generally less strict than that of main market segments (for example, start-up companies are allowed to list on new markets and at least half of the shares sold in the IPO must be newly issued). These new markets are therefore meant to appeal to a different type of company than main market segments. The purpose of this chapter is to investigate whether the pricing of IPOs on Europe’s new stock markets differs from that of pricing IPOs on main market segments. We analyze 578 IPO firms listed on new stock markets and 542 IPO firms listed on main stock markets (the French Second March´e, the Geregelter Markt and Amtlicher Handel of the Frankfurt Stock Exchange, the Belgian Eerste Markt, the Dutch Offici¨ele Markt and the Italian Mercato Telematico Azionario) from January 1990 to December 2002. Our study contains three contributions to the literature. First, we analyze a comprehensive sample of 1,120 European IPO firms on both new stock markets and main stock markets. Second, we investigate pre-IPO ownership structure and secondary sales. Pre-IPO ownership structure and secondary sales have not been analyzed in previous studies on IPO pricing in Europe (but see Ljungqvist & Wilhelm, 2003, for evidence from the United States). New stock markets impose restrictions on companies regarding the composition of the IPO. At least half of the shares sold to the public must be newly issued. This implies that no more than half of the IPO shares can be sold by pre-IPO owners. In the United States, such rules regarding IPO composition do not exist. Third, our paper builds on the emerging literature examining why IPO first-day returns have increased during the 1996–2000 period in the United States (Ljungqvist & Wilhelm, 2003; Loughran & Ritter, 2003). We take a similar approach to explore if the IPO pricing process of new market IPOs differs from that of main market IPOs. We document that new market IPO firms are smaller, younger and riskier than main market IPO firms. They more frequently report losses in the year before the IPO and a larger fraction of new market IPO firms are from the Internet and technology sector. We report a 22.3 percentage point difference in the average first-day return of new market issues (34.3%) and the average first-day return of main market issues (12%). We show that reduced incentives to control wealth losses and differences in firm and offer characteristics partially explain higher first-day returns on new markets. We also show that the opportunity to bundle IPO deals has been important to reduce underpricing costs on new stock markets. However, a large part of the difference in average first-day return cannot be explained by differences in incentives to control wealth losses, bundling, firm and offer characteristics, post-pricing spillover variables or hot issue markets.
Pricing Initial Public Offerings on Europe’s New Stock Markets
27
The remainder of the chapter is organized as follows. In Section 2, we discuss prior literature. In Section 3 we describe the sample and summary statistics. In Section 4 we analyze the sample and report changes in the characteristics of European IPOs. In Section 5 we provide the empirical results. We present our conclusions in Section 6.
2. PRIOR LITERATURE 2.1. Partial Adjustment to Private Information The IPO pricing process is characterized by asymmetric information: the market is uncertain about the quality of the firm, while the issuing firm and its underwriter do not know the market demand for IPO shares. Under the bookbuilding method, the underwriter proposes a price range for the shares in the pre-offering phase. Investors place non-binding orders at different prices within the range. From these indications of interest, the underwriter can extract private information that can be used when setting the final offer price of the IPO. The problem is that investors have no incentive to truthfully reveal their private demand because they know that showing an interest to buy IPO shares will drive up the offer price. Benveniste and Spindt (1989) argue that this problem may be offset, if underwriters only partially adjust the offer price to positive information, as to reward investors with underpriced shares. The partial adjustment of the IPO price to strong demand has been widely documented for the United States. Hanley (1993) finds that issues where the offer price is above the maximum of the price range have higher first-day returns than those where the offer price is below the minimum of the price range. Ritter and Welch (2002) show that this empirical pattern has held for the U.S. IPO market throughout 1980–2001. Ljungqvist et al. (2003) provide contemporary European evidence on this partial adjustment phenomenon. But IPO pricing differs between the United States and Europe. While IPO prices are frequently set outside the price range in the United States, this is rarely the case in Europe. Aussenegg et al. (2003) report that from 1999 to 2000 only 3.5% of Neuer Markt issues have been priced below the minimum of the price range while none was priced above the maximum of this range. Jenkinson et al. (2003) argue that European underwriters commit not to set the offer price above the maximum of the range in order to extract private information from investors prior to deciding on the price range. The information asymmetry theories by Baron (1982) and Beatty and Ritter (1986) suggests that IPOs characterized by greater valuation uncertainty will tend
28
GIANCARLO GIUDICI AND PETER ROOSENBOOM
to be more underpriced to compensate investors for the larger risk. Asymmetric information and uncertainty are particularly relevant in risky start-up companies (Ritter, 1984). These type of firms are more likely to benefit from information acquisition during the pre-market period (Benveniste & Spindt, 1989; Ljungqvist & Wilhelm, 2003) and are more likely to go public on new stock market segments. This predicts higher first-day returns on new markets than on main markets.
2.2. Wealth Losses to Pre-IPO Owners When underwriters adjust offer prices only partially to private information and/or public information, wealth losses occur for shareholders of the issuing firm. If shares had been sold at the full information market price instead, the proceeds to selling pre-IPO shareholders would have been higher. Alternatively, the same proceeds could have been raised by selling fewer shares, resulting in less dilution to pre-IPO shareholders. Loughran and Ritter (2002) ask the question why pre-IPO shareholders “do not get upset about leaving money on the table” and provide a prospect theory explanation. Prospect theory assumes that the bad news of leaving money on the table and the good news of the increase in wealth related to an upward price adjustment is incorporated into a single event. The wealth impact can more than offset the amount of “money left on the table.” Habib and Ljungqvist (2001) show that the incentive to reduce first-day returns and wealth losses increases in pre-IPO shareholders’ participation ratio (the fraction of pre-IPO shares they sell in the offering) and in the dilution factor (the percentage increase in the number of shares outstanding after the IPO). This suggests that pre-IPO ownership structure and insider selling behavior are related to IPO pricing. Bradley and Jordan (2002) report that U.S. IPO firms have fewer incentives to control first-day returns when dilution is small. Ljungqvist and Wilhelm (2003) argue that in the late 1990s in the U.S. the incentives to control first-day returns changed: CEO pre-IPO stakes declined from 22.8% of pre-issue shares in 1996 to 11.5% of pre-issue shares in 2000, reducing their share of money left on the table and therefore their incentive to bargain over the IPO price with underwriters. They also report that both the frequency and magnitude of secondary sales of existing shares fell during the Internet bubble years. Because insiders, especially CEOs, were selling fewer shares at the IPO, they had fewer incentives to bargain over the IPO price. The listing requirements commonly imposed by new markets (at least half of IPO shares must be newly issued) could affect the incentives to control wealth losses and therefore IPO underpricing. This rule increases the dilution factor but
Pricing Initial Public Offerings on Europe’s New Stock Markets
29
decreases the participation ratio. This predicts higher first-day returns on new markets than on main markets.
2.3. Hot Issue Markets, Information Externalities and Bundling Lowry and Schwert (2002) study the cycles in both the number of IPO deals and the average first-day returns. They report that hot issue markets, defined as months in which the first-day return is above the median month’s average first-day return, are followed by IPO waves during which the number of firms going public increases. This suggests that IPO deals and first-day returns are clustered in time as first documented by Ibbotson and Jaffe (1975). The Internet bubble period during 1999–2000 is an example of such waves (Ljungqvist & Wilhelm, 2003; Loughran & Ritter, 2003). Ljungqvist et al. (2004) explain large first-day returns during these “hot issue” periods hypothesizing that “regular” (institutional) are assumed to hold part of their IPO shares as inventory and sell to newly arriving “sentiment” (retail) investors if the hot market persists. Higher first-day returns serve to compensate potential inventory losses incurred by “regular” investors if the hot issue market ends unexpectedly. Lowry and Schwert (2002) show that potential issuing firms learn from the experience of other IPO firms and are more likely to follow with their own IPO if the average first-day return of previous deals was high. Subrahmanyam and Titman (1999) argue that these information externalities play a more important role in new industries, such as the Internet sector. Benveniste et al. (2002) and Hoffman-Burchardi (2001) argue that potential issuing firms derive benefit of information about a common valuation factor from the IPO pricing of previous deals. If the first firm going public incurs the entire cost of compensating investors for costly information production, other firms subject to the same valuation factor can free ride on that information. This gives little incentive among firms to become the first to go public. Underwriters may address this coordination problem by “bundling” IPOs. This allows the cost of information production to be shared over multiple firms, reducing the disincentive to become the first mover. Once one of the firms decides to go public, this may trigger an IPO wave of industry-related firms in an attempt to benefit from the information externality. Benveniste et al. (2003) show that the “bundling” effect reduces IPO underpricing in the United States, because the cost of information production can be spread over more firms subject to a common valuation factor. Information externalities are expected to be more intense in new industries such as the Internet (Subrahmanyam & Titman, 1999). Underwriters may share the
30
GIANCARLO GIUDICI AND PETER ROOSENBOOM
information production cost by “bundling” IPO deals. Because these companies in new industries are more likely to list on new stock markets, we hypothesize that the clustering of IPOs should be more evident in new markets. This predicts lower first-day returns on new markets than on main markets.
3. DATA AND SAMPLE CONSTRUCTION We construct a large database of 1,399 admissions to trading on the Frankfurt Stock Exchange, Euronext Brussels, Euronext Amsterdam, Euronext Paris, the Milan Stock Exchange and NASDAQ Europe from January 1990 through December 2002. These admissions were identified from SDC Global New Issues database and information provided by the stock exchanges. In constructing our sample, we exclude 86 financial companies (SIC codes 6000–6999), 13 spin-offs, 21 privatisation issues2 and 143 companies previously listed elsewhere. We lack the prospectuses for 16 companies. The final sample therefore comprises 1,120 IPO firms that meet our sample criteria and for which we are able to obtain prospectuses from the company, Disclosure Global Access or B¨orsenzeitung (for German prospectuses before 1995). This number includes 73 foreign issuers.3 Our sample consists of 578 IPO firms listed on new stock markets (the French Nouveau March´e (151 firms), the German Neuer Markt (315), Euro.NM Belgium (13), the Dutch Nieuwe Markt (14), the Italian Nuovo Mercato (39) and EASDAQ/NASDAQ Europe (46)) and 542 IPO firms on the main markets (the French Second March´e (234 firms), the Amtlicher Handel (76) and Geregelter Markt (59) of the Frankfurt Stock Exchange, the Belgian Eerste Markt (39), the Dutch Offici¨ele Markt (59) and the Italian Mercato Telematico Azionario (75)).4 For the new markets, the sample period starts when the market commenced operations (March 1996 for the Nouveau March´e, November 1996 for EASDAQ/NASDAQ Europe, March 1997 for the Neuer Markt, April 1997 for Euro.NM Belgium and Nieuwe Markt and June 1999 for the Italian Nuovo Mercato).
4. SUMMARY STATISTICS 4.1. Firm Characteristics Table 1 reports summary statistics about the sample. Market prices are taken from Datastream. The data, collected from IPO prospectuses, have been converted in euros where appropriate. Panel A shows firm characteristics. Market capitalization, defined as the number of shares times the closing market price on
31
Pricing Initial Public Offerings on Europe’s New Stock Markets
Table 1. Summary Statistics.
Panel A: Firm characteristics Market capitalization (d thousands) Gross proceeds (d thousands) Total assets (d thousands) Sales (d thousands) EBITDA < 0 (% of firms) Firm age (years) Internet and technology (% of firms) Panel B: Offer characteristics First-day return (%) Dilution factor (%) Price revision (%) Underwriter market share (%) IPO volume (#)
All
New Markets
Main Markets
Test for Difference
345,740 106,010 70,863 26,349 105,792 21,574 128,010 27,436 16.79 26.35 14.00 36.88
284,651 125,749 54,099 28,217 28,641 10,574 30,313 11,619 29.93 12.38 9.50 59.86
410,887 88,346 88,740 23,454 188,067 52,546 231,115 70,915 2.77 41.25 26.00 12.36
1.30 4.08*** 2.30** 3.31*** 6.56*** 18.36*** 6.68*** 19.28*** 13.03*** 15.26*** 16.62*** 18.90***
23.53 6.09 27.52 25.00 67.06 100.00 7.41 4.03 13.28 8.00
34.34 10.00 32.64 30.00 68.45 100.00 5.75 2.67 18.31 14.00
12.01 4.64 22.05 13.64 64.82 75.00 9.18 5.16 7.92 4.00
7.66*** 4.41*** 2.83*** 15.87*** 0.95 3.49*** 6.39*** 5.96*** 14.31*** 14.42***
Note: This table shows summary statistics. The first column presents means and medians for the entire sample of 1,120 European IPO firms that went public from January 1990 to December 2002. The second column presents means and medians for 578 IPO firms on new stock markets, while the third column provides means and medians for 542 IPO firms on main stock markets. We test whether differences exist between the firm characteristics of new market and main market IPO firms. We use a standard t-test for difference in means and the Wilcoxon/Mann-Whitney test for difference in medians. Medians are shown in italics. We compute market capitalization as the number of shares outstanding after IPO times the closing market price on the first trading day. Gross proceeds is defined as the number of shares sold in IPO (excluding over-allotment option) times the offer price. Total assets and sales are for the most recent financial year disclosed in the IPO prospectus. EBITDA < 0 is a dummy variable that equals one if earnings before interest, taxes, depreciation and amortization is less than zero in the most recent financial year disclosed in the prospectus. Firm age is defined as the calendar year of the IPO minus the calendar year of founding as mentioned in the prospectus. Internet and technology is a dummy variable that takes on the value one if the IPO firm is active in the Internet and technology sector. We identify Internet and technology firms as described in Note 5. First-day return is measured as: (first-day closing market price – final offer price)/final offer price. The dilution factor is defined as the number of newly issued shares at the IPO divided by the number of pre-IPO shares outstanding. The price revision is measured as: (Final offer price – lower bound of price range)/(upper bound of price range – lower bound of price range). Price revisions are available for 547 new market issuers and 339 main market issuers. Underwriter market share is the sum of gross proceeds (excluding over-allotment option) in all local IPOs lead managed by bank j divided by the total proceeds raised in the local market during the sample period. IPO volume is the number of companies going public on local market from 30 trading days before to 10 trading days after the IPO’s pricing date. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
32
GIANCARLO GIUDICI AND PETER ROOSENBOOM
the IPO date, averages d345.7 million (d106 million, evaluated at the median). The average (median) gross proceeds, measured as the number of shares sold at the IPO (excluding over-allotment option) times the offer price, equals d70.9 million (d26.3 million). The average (median) firm reports d105.8 million (d21.6 million) and d128 million (d27.4 million) in total assets and sales, respectively. On average, 16.8% of IPO firms report negative earnings before interest, taxes, depreciation and amortization (EBITDA) in the fiscal year before the IPO. We define age as the difference between the IPO year and the original founding year of the company. The average (median) age is 26.4 (14) years. Internet and technology firms are identified following the approach of Loughran and Ritter (2003).5 In our sample, 36.9% of issuing firms are from the Internet and technology sector.6 We use a standard t-test for difference in means and the Wilcoxon/MannWhitney test for difference in medians in order to explore whether differences exist between the firm characteristics of new markets and main markets. We find that the average new market issuer is significantly smaller than the average main market issuer as measured by market capitalization, gross proceeds, total assets or sales. However, the findings for market capitalization and gross proceeds are driven by a handful of large IPO deals on the main markets. Evaluated at the median, market capitalization and gross proceeds are significantly larger for new market issuers than for main market issuers. On average, 29.9% of new market IPO firms report negative earnings versus 2.8% of main market IPO firms. The average new market issuer is significantly younger (12.4 years) than the average main market issuer (41.3 years). On average, 59.9% of new market issuers are from the Internet and technology sector versus only 12.4% of main market issuers.
4.2. Offer Characteristics Panel B of Table 1 shows offer characteristics. The average first-day return computed as the percentage difference between the first-day closing market price and the offer price, equals 23.5%. First-day returns are right-skewed with the median being 6.1%. During our sample period, 77 firms (6.9%) more than double in price on the first day of trading. On average, the amount of “money left on the table,” defined as first-day return times gross proceeds, equals d18.2 million (d1.2 million). The large difference between the average and median shows that the most of the money left on the table comes from a small number of IPOs.7 Next, we compute the dilution factor, defined as the number of newly issued shares divided by the number of pre-IPO shares outstanding. We find that the average (median) dilution factor equals 27.5% (25%) of pre-IPO shares.
Pricing Initial Public Offerings on Europe’s New Stock Markets
33
The majority of issues (898; 80.2%) were priced using the book building procedure. Only 124 cases were fixed-price offers and in 98 cases the IPO was auctioned, primarily on the French Second March´e. This reflects the move from fixed-price offerings and auctions to bookbuilding as the key method for pricing IPOs in Europe (Ljungqvist et al., 2003).8 We compute price revisions as (final offer price – lower bound of price range)/(upper bound of price range – lower bound of price range). The price revision variable equals 0.5 if the offer price being set at the midpoint of the price range, 0 if the offer price is set at the lower bound of the price range and 1 it is set at the upper bound of the range. If the offer price is below the lower bound of the price range, price revision is negative and if the offer price is above the upper bound of the price range, the price revision variable exceeds 1. We find that on the average price revision equals 67.1% for our sample of 886 bookbuilt IPOs.9 The median price revision equals 100%. This shows that in most cases the offer price is set at the upper bound of the price range. We find that European IPO prices are “sticky.” Only 12 (1.4%) bookbuilt IPOs have an offer price that exceeds the maximum of the price range. Generally, IPOs with offer prices above the price range obtain a dual listing on NASDAQ and have an U.S. investment bank as their lead manager. There are 40 (4.5%) bookbuilt IPOs with an offer price below the minimum of the price range. We calculate the underwriter market share as the percentage market share of the lead manager in the local market. Following Ljungqvist and Wilhelm (2002), market share is the sum of gross proceeds (excluding over-allotment option) in all local IPOs lead managed by the bank divided by the total proceeds raised in that local market during the sample period.10 We use the underwriter market share as a proxy for the underwriter’s reputation capital. The average (median) market share of the lead manager is 7.4% (4%). We determine the IPO volume as the number of companies going public in each local market during the 30 trading day before to 10 trading days after the pricing date (i.e. the date at which the final offer price of the IPO is determined). The IPO volume variable is used as a bundling measure. We end 10 days after the pricing date because underwriters may have expectations regarding bundling with deals in the pipeline. This follows the approach adopted by Ljungqvist and Wilhelm (2002). The average (median) IPO volume is 13.3 (7.9) companies. Next, we compare the offer characteristics of new market and main market issues. New market issues are significantly more underpriced than main market issues. In particular, new market issues have an average (median) first-day return equal to 34.3% (10.0%) versus an average (median) first-day return of 12% (4.6%) for main market issues. We observe that 69 (11.9%) of new market issues and only 8 (1.5%) of main market issues double in price on the first trading day.
34
GIANCARLO GIUDICI AND PETER ROOSENBOOM
The dilution factor is higher for the new market issuers than for the main market issuers. We are able to compute price revisions for 547 bookbuilt new market IPOs and 339 bookbuilt main market IPOs. We find no significant difference between the average price revisions between the two markets. However, evaluated at the median, price revisions for new market IPOs are higher than for main market IPOs. New market issues tend to be underwritten by less reputable underwriters than main market issues. We also find that IPO volume is significantly larger for new market IPOs than for main market IPOs.
4.3. Pre-IPO Ownership Structure Table 2 shows pre-IPO ownership structure. The average (median) insider stake equals 63.1% (73.4%) of pre-IPO shares. We define insiders as the CEO and his family, executive and non-executive directors and their family members and employees. The average (median) pre-IPO stakes held by the CEO is equal to 35% (27.7%) of pre-IPO shares. This suggests the presence of closely held firms that are managed by their founders. Venture capitalists (VCs) and private equity funds hold a pre-IPO stake in 435 (38.8%) sample companies. Conditional on venture capitalists owning a pre-IPO stake, the average (median) aggregate stake is 31% (26%) of pre-IPO shares. In our sample, corporations own stakes in 273 (24.4%) companies. This includes equity-carve outs from parent companies and cases where large corporations own strategic stakes in smaller industry-related firms. Conditional on corporations owning pre-IPO shares, the average (median) stake is 46.3% (37.3%). Financial investors own an average (median) pre-IPO stake of 21.5% (14%) in the 281 companies in which they own shares. Financial investors include banks, insurance companies, mutual funds and pension funds. We measure pre-IPO ownership concentration by a Herfindahl index computed as the sum of the squared pre-IPO ownership stakes of the CEO, venture capitalists, corporations and financial investors. The Herfindahl index ranges from zero to one; a value of zero indicates a fragmented ownership structure with an infinite number of shareholders, while a value of one indicates a single pre-IPO shareholder. The average (median) level of ownership concentration equals 0.40 (0.32). The average (median) participation ratio, defined as the number of existing shares sold by pre-IPO shareholders divided by the number of pre-IPO shares, is 11.9% (9.1%). Compared to contemporary U.S. IPOs, studied by Ljungqvist and Wilhelm (2003), European IPOs seem to be characterized by several differences: (i) the pre-IPO CEO stake is substantially larger; (ii) IPOs are less frequently venture
35
Pricing Initial Public Offerings on Europe’s New Stock Markets
Table 2. Pre-IPO Ownership Structure.
Pre-IPO insider stakes (%) CEO stakes (%) No. with VC stake Fraction with VC stake (%) VC stakes (%) No. with corporate stake Fraction with corporate stake (%) Corporate stakes (%) No. with financial investor stake Fraction with financial investor stake (%) Financial investor stake (%) Ownership concentration Participation ratio (%)
All
New Markets
Main Markets
Test for Difference
63.10 73.40 35.00 27.73 435 38.83 31.03 26.00 273 24.38 46.32 37.34 281 25.09 21.53 14.00 0.403 0.323 11.91 9.09
63.77 68.92 36.80 30.32 261 45.16 30.12 27.02 150 25.95 35.48 23.55 148 25.61 17.10 11.40 0.372 0.271 7.45 4.83
62.38 78.18 33.09 20.47 174 32.10 32.39 25.00 123 22.69 59.55 67.15 133 24.54 26.47 16.43 0.435 0.400 16.67 12.54
0.67 1.04 1.84* 3.69*** 4.52*** 1.02 0.10 1.27 5.80*** 5.52*** 0.41 3.51*** 3.05*** 3.28*** 2.17** 12.16*** 11.52***
Note: This table shows pre-IPO ownership structure. The first column presents means and medians for the entire sample of 1,120 European IPO firms that went public from January 1990 to December 2002. The second column presents means and medians for 578 IPO firms on new stock markets, while the third column provides means and medians for 542 IPO firms on main stock markets. We test whether differences exist between the firm characteristics of new market and main market IPO firms. We use a standard t-test for difference in means and the Wilcoxon/Mann-Whitney test for difference in medians. Medians are shown in italics. Ownership data is hand-collected from prospectuses. Insiders include the Chief Executive Officer (CEO) and his family, (non-) executive directors and their family members and employees. Venture capitalists (VCs) are providers of venture capital or private equity funds. Corporate shareholders are industrial and commercial companies. Financial investors include banks, insurance companies, mutual funds and pension funds. All equity stakes of pre-IPO shareholders are expressed as a percentage of pre-IPO shares outstanding. Ownership concentration is measured using a Herfindahl index, calculated as the sum of squared equity stakes held by the CEO, venture capitalists, corporates and financial investors. The participation ratio is defined as the number of existing shares sold by pre-IPO shareholders divided by the number of pre-IPO shares. ∗ Significant at the 10% level. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
36
GIANCARLO GIUDICI AND PETER ROOSENBOOM
backed, and the size of the VC stake is lower; (iii) pre-IPO ownership is more concentrated in continental Europe than in the United States. Table 2 shows that there is no difference between pre-IPO insider stakes in main market companies and new market companies. However, CEO stakes are higher for IPO firms on new markets than for IPO firms on main markets. More IPO firms are venture-backed on the new markets. However, there is no difference in the size of the stakes of venture capitalists, conditional on venture capitalists being pre-IPO shareholders. Pre-IPO ownership by corporations is significantly more frequent and larger for main market companies than for new market companies. This reflects the larger number of equity-carve outs from listed parent companies that occur on the main markets. Pre-IPO ownership of financial investors is larger for IPO firms on main markets than for IPO firms on new markets, conditional on financial investors being shareholders in the firm. The fraction of sample firms with a financial investor is not different between the two market types. Ownership is less concentrated (more fragmented) in new market IPO firms than in main market IPO firms. The participation ratio is significantly higher for main market issuers than new market issuers. This shows that the existing owners of IPO firms on main markets sell more secondary shares in the IPO.
4.4. Pre-IPO Shareholders’ Selling Behavior Table 3 shows the selling behavior of pre-IPO shareholders. Across the total sample, pre-IPO shareholders sell part of their shares at the IPO in 818 (73%) companies. Conditional on pre-IPO secondary sales, the average (median) size of secondary sales is 16.3% (12%) of pre-IPO shares. Insiders sell shares at the IPO in 646 (57.7%) companies. The conditional average (median) size of insider sales amounts to 9.3% (6.5%) of pre-IPO shares. The CEO sells shares at the IPO in 488 (43.6%) sample firms. Conditional on CEO sales, the average (median) CEO sells 6.1% (4.1%) of pre-IPO shares. There are 297 companies in which venture capitalists sell shares at the IPO. This constitutes 68.3% of the 435 sample firms that are venture-backed. The conditional average (median) size of venture capitalist sales equals 10.8% (6.8%). Table 3 shows that there are 131 firms (48% of sample firms in which corporation own pre-IPO shares) in which corporations sell shares at the IPO. The conditional average (median) size of corporate sales is 13.6% (6.6%) of pre-IPO shares. Across the sample, financial investors sell shares at the IPO in 175 companies. This equals 62.3% of sample firms that are financial investor backed (i.e. in which financial investors own pre-IPO shares). The conditional average (median) size of the financial investors sales amounts to 9.7% (5.7%) of pre-IPO shares.
37
Pricing Initial Public Offerings on Europe’s New Stock Markets
Table 3. Pre-IPO Shareholders’ Selling Behavior (Secondary Sales). All
New Markets
Main Markets
Test for Difference
No. with secondary sales Fraction with secondary sales Size of secondary sales (%)
818 73.04 16.31 12.00
375 64.88 11.48 9.00
443 81.73 20.40 15.46
6.46*** 10.11*** 10.66***
No. with insider sales at IPO Fraction with insider sales at IPO Size of insider sales at IPO (%)
646 57.68 9.27 6.48
312 53.98 6.60 4.75
334 61.62 11.76 9.29
2.59*** 7.40*** 7.82***
No. with CEO sales at IPO Fraction with CEO sales Size of CEO sales (%)
488 43.57 6.10 4.13
254 43.94 4.37 2.70
234 43.17 7.98 6.50
0.26 6.57*** 7.28***
No. VC backed firms with VC sales at IPO Fraction VC backed firms with VC sales (%) Size of VC sales (%)
297 68.28 10.76 6.75
158 60.53 8.21 6.24
139 79.89 13.65 8.65
4.33*** 4.20*** 3.11***
131
58
73
47.99
38.67
59.35
3.47***
13.59 6.60
5.08 2.81
20.35 15.00
5.74*** 6.09***
175
75
100
62.28
50.68
75.19
4.36***
9.70 5.66 46.99 53.16
5.95 3.62 46.16 50.41
12.51 8.82 47.88 56.12
3.67*** 4.23*** 1.06 2.13**
No. corporate backed firms with corporate sales at IPO Fraction corporate backed firms with corporate sales (%) Size of corporate sales (%) No. financial investor backed firms with financial investor sales at IPO Fraction of financial investor backed firms with financial investor sales (%) Size of financial investor sales (%) Post-IPO insider stakes (%)
Note: This table shows pre-IPO shareholders’ selling behavior. The first column presents means and medians for the entire sample of 1,120 European IPO firms that went public from January 1990 to December 2002. The second column presents means and medians for 578 IPO firms on new stock markets, while the third column provides means and medians for 542 IPO firms on main stock markets. We test whether differences exist between the firm characteristics of new market and main market IPO firms. We use a standard t-test for difference in means and the Wilcoxon/Mann-Whitney test for difference in medians. Medians are shown in italics. Secondary sales are existing shares being sold by pre-IPO shareholders at the IPO. Insiders, venture capitalists (VCs), corporate shareholders and financial investors are defined as before in Table 2. All sales are expressed as a percentage of pre-IPO shares outstanding. Post-IPO insider stakes are the number of shares held by insiders expressed as a percentage of post-IPO shares outstanding. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
38
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Post-IPO insider stakes are defined as the post-IPO ownership stake of insiders (CEOs, (non-) executive directors and employees) divided by the number of post-IPO shares outstanding. The average (median) post-IPO insider stake is 47% (53.2%). Comparing European IPOs with U.S. IPOs, as reported by Ljungqvist and Wilhelm (2003), the main differences are: (i) aggregate secondary sales are significantly larger in Europe than in the U.S.; (ii) CEO sales are significantly larger as well; (iii) the fraction of venture-backed IPOs with venture capitalist sales at the IPO is lower in the United States.11 Secondary sales are more widespread on main markets than on new markets. Pre-IPO shareholders sell shares at the IPO in 81.7% of main market companies and 64.9% of new market companies. Shareholders other than CEOs are more frequent sellers at the IPO on main markets. Moreover, the size of secondary sales at the IPO is significantly larger on main markets than on new markets. Conditional on owning shares, all types of pre-IPO shareholders sell more shares at the IPO on main markets. Although insiders are selling less at the IPO on new markets, postIPO insider stakes are lower on new markets than on main markets. This reflects that new market companies issue more new shares than main market companies.
4.5. What Has Changed Over Time? We also examine what has changed over our sample period 1990–2002. We distinguish four time periods: 1990–1995 (143 IPO firms), 1996–1998 (402 IPO firms), 1999–2000 (522 IPO firms) and 2001–2002 (53 IPO firms). We choose to investigate these four periods because during 1990–1995 none of the new markets exists, during the 1996–1998 period most new markets commence operations, during the 1999–2000 period the Internet bubble occurs and in 2001–2002 the Internet bubble bursts. We include 2001–2002 for information purposes only but do not test for differences with other periods because of the limited number of observations. Panel A of Table 4 investigates changes in firm characteristics. We report that pre-IPO assets, pre-IPO sales and firm age have significantly decreased from 1990–1995 to 1999–2000. We also find that the fraction of firms with losses in the year before the IPO and the fraction of Internet and high-technology firms have increased from 1990–1995 to 1999–2000. We attribute this development, at least in part, to the establishment of the new stock markets. New markets have less stringent listing requirements and have allowed smaller, younger, unprofitable and Internet and high-tech companies to go public on European exchanges. Panel B shows changes in offer characteristics. We find that average and median first-day returns increased from 1990–1995 to 1999–2000 and fell in the
Period
Panel A: Firm characteristics Market capitalization (d thousands) Gross proceeds (d thousands) Total assets (d thousands) Sales (d thousands) Fraction with EBITDA < 0 (%) Firm age (years) Internet and technology (% of firms) Panel B: Offer characteristics First-day return (%) Dilution factor (%) Price revision (%) Underwriter market share (%) Prior IPO volume (#)
Test for Difference
1996–1998
1999–2000
2001–2002
1990–1995 1990–1995 1996–1998 vs. vs. vs. 1996–1998 1999–2000 1999–2000
247,704 91,999 62,713 24,327 181,694 80,848 267,724 98,914 0.70 57.38 39.00 6.29
160,716 68,972 38,648 15,698 87,909 21,432 108,523 31,169 8.96 27.24 15.50 24.63
533,500 156,125 100,099 35,146 96,260 14,559 105,286 17,655 27.97 17.10 10.00 55.36
164,378 86,460 49,256 22,477 130,519 27,772 120,226 29,302 9.43 27.06 17.00 30.18
2.13** 3.23*** 2.47** 2.82*** 3.03*** 8.93*** 4.39*** 8.86*** 3.41*** 7.82*** 7.98*** 4.81***
1.46 4.05*** 1.24 3.98*** 1.89* 10.98*** 2.79*** 11.53*** 7.22*** 13.21*** 11.63*** 11.42***
3.19*** 10.75*** 3.43*** 10.63*** 0.29 4.23*** 0.09 5.88*** 7.41*** 5.50*** 6.99*** 9.86***
8.69 4.00 30.70 11.12 73.72 77.50 11.81 8.92 3.05 2.00
19.61 8.79 23.26 17.79 79.22 100.00 7.46 4.59 11.91 9.00
32.55 6.69 29.52 28.58 65.94 100.00 6.46 2.67 19.42 18.00
4.57 0.00 31.59 31.00 2.75 0.00 4.50 1.64 3.79 3.00
3.49*** 2.86*** 0.86 3.00*** 0.89 1.42 4.90*** 5.13*** 9.40*** 11.88***
4.44*** 1.76* 0.16 8.54*** 0.77 0.02 6.21*** 6.92*** 16.11*** 15.82***
3.63*** 0.23 4.33*** 8.83*** 3.51*** 2.49** 1.67* 3.51*** 9.69*** 9.97***
39
1990–1995
Pricing Initial Public Offerings on Europe’s New Stock Markets
Table 4. What Changed During the 1990–2002 Period?
40
Table 4. (Continued ) Period 1990–1995
0.419 0.378 15.24 10.48 10.77 8.42 6.97 5.00 12.22 7.34 18.30 8.43 12.98 5.98
1999–2000
0.380 0.275 8.95 6.34 7.32 4.87 4.66 2.81 9.37 6.40 8.95 3.90 6.82 4.36
2001–2002
1990–1995 vs. 1996–1998
1990–1995 vs. 1999–2000
1996–1998 vs. 1999–2000
0.375 0.256 6.59 3.75 4.91 3.70 3.92 2.26 7.30 6.80 20.16 24.57 7.45 1.97
0.87 0.49 0.06 0.77 1.72* 1.47 2.41** 1.67* 0.23 0.12 0.49 0.79 0.67 0.38
2.19** 1.26 5.91*** 5.14*** 4.89*** 5.18*** 5.45*** 4.22*** 1.70* 0.88 2.34** 3.51*** 2.00** 2.03**
1.84* 1.66* 7.21*** 6.44*** 4.64*** 6.00*** 4.13*** 4.61*** 2.04** 1.14 2.60** 2.40** 2.98*** 2.03**
Note: This table shows what has changed over the 1990–2002 period. There are 143 IPO firms during the years 1990–1995, 402 IPO firms in 1996–1998, 522 IPO firms in 1999–2000 and 53 IPO firms in 2001–2002. Price revisions are available for 36 IPO firms in 1990–1995, 296 IPO firms in 1996–1998, 503 IPO firms in 1999–2000 and 51 IPO firms in 2001–2002. We present means and medians for these periods. Medians are shown in italics. We test whether differences exist between the years 1990–1995 vs. 1996–1998, 1990–1995 versus 1999–2000 and 1996–1998 versus 1999–2000, respectively. We use a standard t-test for difference in means and the Wilcoxon/Mann-Whitney test for difference in medians. All variables are defined as in Tables 1–3. The average and median size of insider sales, CEO sales, VC sales, corporate sales and financial investor’s sales are computed conditional on these pre-IPO shareholders selling shares in the IPO. ∗ Significant at the 10% level. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Panel C: Pre-IPO ownership and secondary sales Ownership concentration 0.448 0.425 Participation ratio (%) 15.33 13.31 Size of insider sales 13.00 10.00 Size of CEO sales 9.74 6.82 Size of VC sales 12.89 7.18 Size of corporate sales 15.95 15.00 Size of financial investor sales 10.62 7.03
1996–1998
Test for Difference
Pricing Initial Public Offerings on Europe’s New Stock Markets
41
2001–2002 period. The median dilution factor has steadily increased over time. We argue that this change is related to the listing requirement of most new markets that at least half of the IPO shares are newly issued. Price revisions have decreased significantly from 1996–1998 to 1999–2000. This is largely due to a decrease in the second half of 2000 when the average price revision equals 28.75% and the bubble has burst. Underwriter market share has decreased over time. This shows that less reputable underwriters entered the continental European market for underwriting services. IPO volume has increased from 1990–1995 to 1999–2000 thus increasing the opportunity for underwriters to bundle IPO deals. Next, we investigate changes in pre-IPO ownership structure and secondary sales. Panel C shows that pre-IPO ownership structures have become more fragmented over time. In addition, the participation ratio, size of insider sales, CEO sales, VC sales, corporate sales and financial investor sales have decreased significantly from 1996–1998 to 1999–2000. This is consistent with the U.S. findings of Ljungqvist and Wilhelm (2003).
5. THE DETERMINANTS OF FIRST-DAY RETURNS 5.1. First-Day Return Regressions We start our analysis with a regression of first-day returns on the new market dummy that takes the value one if the IPO firm went public on one of the new markets.12 When we regress first-day returns on this new market dummy, we find that the coefficient is 0.223 and highly significant at the 1% level (t-value = 7.81). If changes in pre-IPO ownership structure, secondary sales, firm and offer characteristics and bundling can explain the difference in first-day return between new and main markets, the coefficient on the new market dummy should become smaller and lose statistical significance once we control for these changes in the regression model. To facilitate comparison, we use similar data definitions and research design as Ljungqvist and Wilhelm (2002, 2003). All variables used in our regression analyses are defined in Table 5. We estimate the following model: First-day returns = f(pre-IPO ownership, secondary sales, firm and offer characteristics, post-pricing spillover variables, bundling, bubble years dummy, new market dummy, country dummies)
(1)
Table 6 shows the results of cross-sectional regressions using first-day returns as the dependent variable. We use White (1980) standard errors to compute t-statistics
42
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Table 5. Definition of Variables. Variable Name
Definition
Pre-IPO ownership CEO stake Corporate stake VC stake Financial investor stake
% pre-IPO equity held by CEOs % pre-IPO equity held by non-financial partner companies % pre-IPO equity held by venture capitalists % pre-IPO equity held by other financial investors
IPO secondary sales CEO sales Corporate sales VC sales Financial investor sales Participation ratio Firm and offer characteristics Log(1 + age)
EBITDA < 0 dummy Log (total assets) Internet and technology dummy Underwriter market share
Dilution factor Price revision Log(expected proceeds)
Post-pricing spillover variables MktReturnpost-pricing
Mktpost-pricing
Shares sold by CEO/pre-IPO outstanding shares Shares sold by non-financial partner companies/pre-IPO outstanding shares Shares sold by venture capitalists/pre-IPO outstanding shares Shares sold by other financial investors/pre-IPO outstanding shares Number of existing shares sold by pre-IPO shareholders/number of pre-IPO shares Natural log one plus firm age, where firm age is measured as calendar year of the IPO minus the calendar year of founding as mentioned in the prospectus = 1 if the IPO firm reports negative EBITDA in the fiscal year before going public; = 0 in other cases Natural log of total assets for the most recent financial year disclosed in the IPO prospectus = 1 if the IPO firm is classified as “Internet” or “technology” stock; = 0 in other cases Percentage market share of the IPO lead manager in the local financial market (measured by gross proceeds raised during the sample period) Number of newly issued shares at the IPO/number of pre-IPO shares outstanding Equals (Final offer price – lower bound of price range)/(upper bound of price range – lower bound of price range) Natural log of expected proceeds, where expected range are measured as the midpoint of the price range times the number of share sold in the IPO Local market index return from the IPO’s final pricing and its first day of trading. We used Datastream Germany, France, Italy, Belgium and the Netherlands indices prior to 1992 and the MSCI Germany, France, Italy, Belgium and Netherlands indices thereafter Standard deviation of the daily local market index returns between the IPO’s final pricing and first day of trading
Pricing Initial Public Offerings on Europe’s New Stock Markets
43
Table 5. (Continued ) Variable Name m Uppost-pricing Pre-pricing spillover variables m Revisionpre-pricing Revisionpre-pricing MktReturnpre-pricing
Mktpre-pricing m Uppre-pricing Bundling IPO volume Bubble years dummy New market dummy
Definition Average first-day underpricing of local IPOs between the IPO’s final pricing and first-day of trading Average revision of all local IPOs between the setting of the price range and the IPO’s final pricing date Standard deviation of all local IPOs between the setting of the price range and the IPO’s final pricing date Local market index return from the setting of the IPO’s price range and its final pricing. We used Datastream Germany, France, Italy, Belgium and the Netherlands indices prior to 1992 and the MSCI Germany, France, Italy, Belgium and Netherlands indices thereafter Standard deviation of daily market index return from the setting of the IPO’s price range and its final pricing Average first-day underpricing from the setting of the IPO’s price range and its final pricing Number of total IPO deals in each local market from 30 days before and 10 days after the pricing date = 1 if the IPO took place in 1999 or 2000; = 0 in other cases = 1 if the IPO firm lists on a new stock market; = 0 in other cases
throughout this study. Columns [1] and [2] show regression results using the entire sample of 1,120 IPO firms. In column [1] we include pre-IPO ownership variables and the participation ratio, while in column [2] we substitute these variables with IPO secondary sales variables. Column [1] highlights that the relation between first-day returns and pre-IPO ownership variables are not statistically significant at conventional levels. We have also tried using the level of ownership concentration instead of the pre-IPO stakes as an independent variable (not reported). We find that the Herfindahl index is not significantly related to first-day returns. However, the coefficient on the participation ratio is significantly negative. This indicates that the incentives to reduce first-day returns are larger the more existing shares are sold in the IPO. Column [2] shows the results for each category of shareholder. First-day returns are a negative function of the size of both CEO sales and corporate sales at the IPO. This indicates that if CEOs and/or corporations are selling shares at the IPO they have larger wealth incentives to bargain for a higher offer price and
44
Table 6. First-Day Return Regressions. Dependent Variable
First-Day Returns [1]
−0.014 (−0.24) −0.015 (−0.27) 0.033 (0.50) −0.013 (−0.15)
[3]
[4]
−0.008 (−0.11) −0.050 (−0.72) 0.037 (0.51) 0.058 (0.63)
IPO secondary sales variables Size of CEO sales Size of corporate sales Size of VC sales Size of financial investor sales Participation ratio
−0.238 (−2.58)***
Bundling IPO volume/10
−0.073 (−4.39)***
−0.071 (−4.34)***
−0.098 (−4.91)***
−0.096 (−4.83)***
−0.012 (−0.77) −0.004 (−0.08) −0.009 (−0.79) 0.126 (2.99)*** −0.228 (−1.52) −0.029 (−1.84)*
−0.015 (−0.98) 0.001 (0.01) −0.011 (0.90) 0.124 (3.01)*** −0.228 (−1.53) −0.028 (−1.87)*
0.002 (0.11) −0.002 (−0.04) −0.008 (−0.58) 0.126 (2.81)*** −0.202 (−1.25) −0.070 (−1.11) 0.270 (5.82)***
−0.001 (−0.06) 0.002 (0.04) −0.010 (−0.76) 0.123 (2.80)*** −0.228 (−1.39) −0.057 (−0.21) 0.271 (5.86)***
Firm and offer characteristics Log (1 + age) EBITDA < 0 dummy Log (total assets) Internet and technology dummy Underwriter market share Dilution factor Price revision
−0.314 (−1.68)* −0.217 (−2.23)** −0.041 (−0.26) −0.070 (−0.38)
−0.544 (−1.88)* −0.121 (−0.91) −0.104 (−0.61) −0.105 (−0.50) −0.313 (−2.71)***
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Pre-IPO ownership variables CEO stake Corporate stake VC stake Financial investor stake
[2]
Bubble years dummy New market dummy Intercept R2 adjusted F-statistic
0.096 (0.16) −2.251 (−0.89) 0.197 (1.81)* 0.147 (3.07)*** 0.056 (1.32) 0.256 (1.86)* 11.03% 7.94***
0.094 (0.16) −2.174 (−0.87) 0.190 (1.72)* 0.140 (3.03)*** 0.064 (1.50) 0.262 (1.89)* 10.96% 8.25***
0.159 (0.21) −2.096 (−0.75) 0.201 (1.80)* 0.172 (3.31)*** 0.085 (1.57) 0.052 (0.31) 17.63% 10.02***
0.159 (0.21) −1.881 (−0.69) 0.196 (1.75)* 0.166 (3.22)*** 0.100 (2.00)** 0.065 (0.38) 17.47% 10.37***
Note: This table shows the OLS regression results using the first-day return as the dependent variable. Models [1] and [2] are estimated using the whole sample of 1,120 European IPO firms that went public from January 1990 to December 2002. Columns [3] and [4] are estimated for 886 bookbuilt IPOs. See Table 5 for variable definitions. White (1980) heteroscedastic-consistent t-statistics are within parentheses. ∗ Significant at the 10% level. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
Pricing Initial Public Offerings on Europe’s New Stock Markets
Post-pricing spillover variables MktReturnpost-pricing s Mktpost-pricing m Uppost-pricing
45
46
GIANCARLO GIUDICI AND PETER ROOSENBOOM
lower underpricing. The signs for the coefficients on the size of VC and financial investor sales are negative but not significant. These results are consistent with reduced incentives to control wealth losses. Given that pre-IPO owners of new market companies are selling less existing shares at the IPO, they experience less incentive to reduce IPO underpricing. We find a negative association between first-day returns and IPO volume.13 The result is consistent with bundling effects. When we interact IPO volume with the new market dummy we find that the coefficient is negative and significant at the 1% level (not shown). This is consistent with bundling effects being more important to control IPO underpricing on new markets. We include the log of firm age and the “EBITDA < 0” dummy as ex-ante risk proxies. However, in columns [1] and [2] we do not find a significant relationship between first-day returns and these two variables. There is also no significant relation between first-day returns and the log of total assets or underwriter market share. Other things equal, Internet and technology firms experience first-day returns that are about 12.5% higher than other IPO firms. The dilution factor is inversely related to first-day returns. This indicates that first-day returns are reduced when dilution costs are higher. Ljungqvist and Wilhelm (2002) have shown that first-day returns of European IPOs are related to post-pricing spillover variables. We also include their post-pricing spillover variables in our model as control variables. The variables are measured from the pricing date (i.e. date at which the final offer price is determined) and the first day of trading. This period averages 3.6 trading days in our sample.14 During this interval we compute daily returns on the local market index (MktReturnpost-pricing ), the standard deviation of those daily returns ( Mktpost-pricing ) and the average first-day returns of other companies that go public (m UPpost-pricing ). These variables may impact first-day returns because underwriters could not have taken this information into account when pricing the IPO. We find that only m UPpost-pricing is positively related to first-day returns. This suggests that first-day returns are higher if the average first-day return of previous deals has been higher. The coefficient on the bubble years dummy is positive and significant. This shows that first-day returns have been higher during the hot issue market of 1999 and 2000. Columns [3] and [4] show the regression results for the bookbuilding sample. We are able to retrieve data on the price range for 886 bookbuilt IPOs. In studying this sample, we include the price revision variable. As expected, we find that the price revision is positively related to first-day returns, consistently with the partial adjustment theory. When we interact the price revision variable with the new market dummy, we find that the coefficient is positive but insignificant (not
Pricing Initial Public Offerings on Europe’s New Stock Markets
47
shown). This is inconsistent with information acquisition being more important to the pricing process of new market IPOs. We also observe that the dilution factor and the size of corporate sales are no longer statistically significant for the bookbuilding sample, whereas they were significant for the entire sample. All other independent variables are of similar sign and significance as reported in columns [1] and [2] for the whole sample. What about our new market dummy? After controlling for other determinants of first-day returns, the coefficient on the new market dummy reduces in size and becomes insignificant. Across the entire sample, the new market dummy accounts for 5.6–6.4 percentage points of the 22.3 percentage point difference in first-day returns between new and main market issuers. For the bookbuilding sample, the new market dummy explains 8.5–10 percentage points of the 22.5 percentage point difference in first-day returns between bookbuilt IPOs on new and main markets. We conclude that the remaining part of the difference between the average first-day return on new and main markets can therefore be explained by changes that occurred in changes in pre-IPO ownership structure, secondary sales, firm and offer characteristics and bundling that we documented in Section 4. Our findings are robust to fixed-time effects and re-defining the sample to include main market issues only as from the date when the new market commenced operations. For example, for Germany we only include IPOs on the Amtlicher Handel and Geregelter Markt of the Frankfurt Stock Exchange as from March 1997 (the time when the Neuer Markt had its first listing). In the next subsection we focus on resolving problems with endogenous variables.
5.2. Endogeneity of Price Revisions and Underwriter Choice The results in Section 5.1 may be subject to endogeneity problems. The regressions for the bookbuilding sample in columns [3] and [4] of Table 6 treat price revisions as exogenous. The Benveniste-Spindt model suggests that price revisions and first-day returns should be estimated simultaneously, because the underwriter’s pricing decision depends on how much money he has to leave on the table to ensure that investors truthfully reveal their demand for IPO shares during the pre-market phase of the bookbuilding process. Underwriter choice may also be an endogenous rather than exogenous variable. Habib and Ljungqvist (2001) show that treating the underwriter choice endogenously can change the relation between underwriter reputation and first-day returns.
48
GIANCARLO GIUDICI AND PETER ROOSENBOOM
We start by running OLS regressions using the price revision variable as the dependent variable. We use a sample of 886 bookbuilt IPOs. We estimate the following model: Price revision = f(pre-IPO ownership, secondary sales, firm and offer characteristics, pre-pricing spillover variables, bubble years dummy, new market dummy)
(2)
Results are shown in column [1] and [2] of Table 7. We observe that price revisions are positively related to the size of CEO sales at the IPO. We find an inverse relation between price revision and the pre-IPO stake owned by financial investors. This suggests that price revisions are lower when financial investors own shares in the company. If CEOs are selling shares at the IPO, they have strong incentives to bargain for higher offer prices. More aggressive bargaining efforts result in larger price revisions. Ownership variables or sales by other pre-IPO shareholders are not significantly related to price revisions. This suggests that pre-IPO shareholders, other than the CEO, have little bargaining power vis-`a-vis the underwriter. Ljungqvist and Wilhelm (2003) find that CEO sales and VC sales are positively related to price revisions for the United States. Company age and the Internet and technology dummy are not significant in the price revision regressions. The coefficient on total assets is significantly negative. This shows that larger firms have lower price revisions compared to smaller firms. Underwriters may learn more information in the pre-market phase for smaller firms than for larger firms, which results in larger price revisions. Underwriter market share is unrelated to the price revision.15 Ljungqvist and Wilhelm (2002) have shown that pre-IPO spillover variables are key determinants of price revisions for European IPOs. We therefore control for their pre-IPO spillover variables. These variables are measured from the date at which the price range has been determined until the date at which the final offer price has been determined. This interval averages 7.3 trading days in our sample. We measure the average price revision of other IPOs (m Revisionpre-pricing ), the standard deviation of the price revisions of other IPOs ( Revisionpre-pricing ), market index returns (MktReturnpre-pricing ), standard deviation of market returns ( Mktpre-pricing ), and the average first-day returns (m UPpre-pricing ) during this period. Underwriters may use this information when deciding on the final offer price. We find that m Revisionpre-pricing is positively related to price revision. This indicates that price revisions are higher if the average price revision of previous IPOs has been higher. However, Revisionpre-pricing is negatively related to price revision. This shows that the effect of m Revisionpre-pricing is discounted in case there is noise in its measurement. MktReturnpre-pricing and m UPpre-pricing
Dependent Variable
Price Revision [1]
Pre-IPO ownership variables CEO stake Corporate stake VC stake Financial investor stake
Underwriter Market Share
First-Day Returns/2SLS
[2]
[3]
[4]
0.791 (2.38)** −0.235 (−0.96) 0.184 (1.04) 0.067 (0.30)
0.018 (0.23) 0.023 (0.42) 0.013 (0.32) −0.103 (−2.36)**
−0.782 (−1.98)** −0.045 (−0.18) −0.181 (−0.81) −0.144 (−0.50)
−0.089 (−1.32) 0.038 (0.51) −0.039 (−0.44) −0.225 (−2.19)**
IPO secondary sales variables Size of CEO sales Size of corporate sales Size of VC sales Size of financial investor sales Bundling IPO volume/10 Firm and offer characteristics Log(1 + age) EBITDA < 0 dummy Log (total assets) Internet and technology dummy Underwriter market share Dilution factor Price revision Log (expected proceeds)
−0.092 (−5.21)*** 0.018 (0.74) 0.034 (0.66) −0.042 (−2.52)** 0.012 (0.29) 0.118 (0.55)
0.014 (0.56) 0.056 (1.04) −0.032 (−2.06)** 0.019 (0.47) 0.165 (0.80)
0.001 (0.29) −0.006 (−0.64) 0.004 (1.59) −0.001 (−0.05)
0.003 (0.12) −0.005 (−0.09) −0.018 (−0.89) 0.120 (2.95)*** −0.105 (−0.12) −0.046 (−0.54) 0.494 (4.91)***
Pricing Initial Public Offerings on Europe’s New Stock Markets
Table 7. Price Revision, Underwriter Choice and First-Day Return Regressions.
0.024 (6.83)***
49
50
Table 7. (Continued ) Dependent Variable [1]
[2]
0.246 (4.28)*** −0.659 (−3.91)*** 1.829 (3.29)*** −2.727 (−1.07) 0.111 (3.50)***
0.251 (4.50)*** −0.639 (−3.79)*** 1.817 (3.29)*** −2.58 (−1.04) 0.107 (3.41)***
Underwriter Market Share
First-Day Returns/2SLS
[3]
[4]
Post-pricing spillover variables MktReturnpost-pricing Mktpost-pricing m Uppost-pricing Bubble years dummy New market dummy Intercept R2 adjusted F-statistic
0.309 (0.42) −1.567 (−0.68) 0.182 (2.97)*** 0.005 (0.14) −0.024 (−0.46) 1.004 (4.88)*** 14.43% 10.31***
0.011 (0.29) −0.016 (−0.30) 0.826 (4.56)*** 14.42% 10.32***
−0.021 (3.12)*** −0.015 (−1.77)* −0.202 (−6.47)***
0.164 (1.61) 0.085 (1.72)* 0.006 (−0.03)
13.29% 13.33***
14.41% 7.73***
Note: This table shows the regression results using the price revision as the dependent variable in columns [1], [2] and underwriter market share in column [3]. All regressions are estimated for 886 bookbuilt IPOs. See Table 5 for variable definitions. We estimate a 2SLS regression using first-day returns as the dependet variable. We treat first-day return as endogenous to price revision and underwriter market share. The 2SLS regression results are shown in column [4] and use models [2] and [3] as its first-stage. White (1980) heteroscedastic-consistent t-statistics are within parentheses. ∗ Significant at the 10% level. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Pre-pricing spillover variables m Revisionpre-pricing Revisionpre-pricing MktReturnpre-pricing Mktpre-pricing m UPpre-pricing
Price Revision
51
Pricing Initial Public Offerings on Europe’s New Stock Markets
are positively related to the price revisions. This suggests that price revisions are higher when pre-pricing market returns and the average first-day return of previous IPOs have been higher. The coefficient on the dummy for the bubble years is not significant. After controlling for other factors price revision is no different in 1999–2000 than in earlier periods. The new market dummy is not statistically significant. Column [3] of Table 7 present the results of the OLS regressions using the underwriter market share as the dependent variable. We use a sample of 886 bookbuilt IPOs. We estimate the following model: Underwriter market share = f(secondary sales, firm and offer characteristics, bubble years dummy, new market dummy)
(3)
In model [3] we investigate pre-IPO ownership stakes. Assuming that underwriters with larger market share are better able to control underpricing costs, pre-IPO owners have more incentives to choose a high reputation underwriter if they sell shares at the IPO. However, looking at the sales by pre-IPO owners, we do not find any significant positive relation with underwriter choice. There is a significant and negative association between the size of sales by financial investors and underwriter market share. Potential conflicts of interests may drive this result. If financial institutions own more shares in the IPO firm, they may convince management to select an affiliated lead manager for the issue, even if this underwriter is less prestigious. We leave this question for future research. The log of expected proceeds is positively related to underwriter market share. If the expected proceeds become larger, underpricing leads to a larger wealth loss for pre-IPO owners. This creates an incentive to hire a prestigious underwriter to lead manage the issue in an attempt to reduce IPO underpricing. The other firm and offer characteristics are not related to underwriter market share. The bubble years dummy is inversely related to underwriter market share. We infer that underwriters with lower market share are especially hired in periods when many firms want to go public. The new market dummy is significantly negative. This shows that new market companies often hire underwriters with smaller market share. This brings us to the endogeneity problems. We estimate a two-stage least (2SLS) version of model [4] of Table 6. We use the predicted values of price revisions of model [2] and underwriter market share of model [3] of Table 7 in the first-day return regression.16 Column [4] of Table 7 shows the 2SLS results. We again find that CEO sales and IPO volume are negatively related to first-day returns.17 Internet and technology firms continue to be associated with higher first-day returns. We also find that price revision and m UPpost-pricing
52
GIANCARLO GIUDICI AND PETER ROOSENBOOM
remain positively related to first-day returns. The bubble years dummy loses its significance. The new markets dummy remains significant at the 10% level and explains 8.5 percentage points of the 22.5 percentage point difference in first-day returns between bookbuilt IPOs on new and main markets.
5.3. Explaining the Difference in First-Day Return Between New and Main Markets In this section, we decompose the difference in the average first-day return of new market and main markets into components. We adopt the procedure of Loughran and Ritter (2003) and multiply the regression coefficient by the difference in sample characteristics. For example, we multiply the regresion coefficient of the participation ratio in regression model [1] of Table 6 (that equals −0.238) with the difference between the average participation ratio of new market IPO firms (7.45%, as reported in Table 2) and the average participation ratio of main market IPO firms (16.67%, as reported in Table 2). This yields −0.231 × (7.45% − 16.67%) = 2.19%. Table 8 shows the results. Table 8 shows that pre-IPO ownership variables (subtotal [1]) only account for a small part of the difference in first-day returns. The secondary sales variables account for 1.14–3.58 percentage points of the first-day return difference (subtotal [2]). This suggests that reduced incentives to control wealth losses can at least partially explain why there is a difference between the average first-day return of new market IPOs and main market IPOs. Next, we look at IPO volume. IPO volume captures bundling effects. The results indicate that bundling IPO deals has helped to lower first-day returns for new market IPOs. Differences in firm and offer characteristics are the most important. Subtotal [3] shows that these differences account for 8.1 percentage points to 10.4 percentage points of the difference in first-day returns. Table 8 also shows that differences in the price revision (part of subtotal [3]) have contributed to higher first-day returns on new markets. Differences in post-pricing spillover variables explain only a minor part of the difference in first-day returns (subtotal [4]). The bubble years dummy that captures hot issue markets, accounts for 5.1 percentage points to 6.4 percentage points of the difference in first-day returns between new and main markets. We calculate the part of the difference in average first-day returns that can be explained by differences in pre-IPO ownership variables, IPO secondary sales variables, bundling, firm and offer characteristics, post-pricing spillover variables and the bubble years dummy. We find that this ranges from 8.8 percentage points to 12.6 percentage points. We also determine the part of the difference in average
53
Pricing Initial Public Offerings on Europe’s New Stock Markets
Table 8. Decomposition of the Difference in First-Day Returns Between New and Main Market IPOs. Based on
Pre-IPO ownership variables CEO stake Corporate stake VC stake Financial investor stake Subtotal [1] IPO secondary sales variables Size of CEO sales Size of corporate sales Size of VC sales Size of financial investors sales Participation ratio Subtotal [2] Bundling IPO volume/10 Firm and offer characteristics Log (1 + age) EBITDA < 0 Log (total assets) Internet and technology dummy Underwriter market share Dilution factor Price revision Subtotal [3] Post-pricing spillover variables Postpricing market return Postpricing volatility Postpricing underpricing Subtotal [4] Bubble years dummy Explained (subtotals (1)−(4), bundling, IPO volume, bubble years dummy)
Model [1]; Table 6 (%) [1]
Model [2]; Table 6 (%) [2]
Model [3]; Table 6 (%) [3]
−0.05 0.06 0.11 0.03
−0.09 0.17 0.04 −0.25
0.15
−0.13 0.49 0.48 0.05 0.11
2.19
Model [4]; Table 6 (%) [4]
Model [4]; Table 7 (%) [5]
0.41 0.30 0.22 0.26
0.59 0.11 0.39 0.36
3.58
2.19
1.13
3.58
1.19
1.44
−6.21
−6.04
−8.45
−8.28
−7.93
1.38 −0.11 1.69 5.99
1.72 0.03 2.07 5.89
−0.23 −0.05 1.71 5.60
0.12 0.05 2.14 5.47
−0.35 −0.13 3.85 5.33
0.78 −0.30
0.78 −0.29
0.77 −0.70 0.98
0.87 −0.57 0.98
0.40 −0.46 1.79
9.42
10.19
8.07
9.05
10.44
−0.02 −0.42 0.92
−0.02 0.40 0.89
−0.06 −0.23 0.65
−0.06 −0.21 0.64
−0.11 −0.17 0.59
0.48
1.27
0.36
0.37
0.31
6.38 12.41
6.08 12.64
5.37 8.80
5.18 7.52
5.12 9.37
54
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Table 8. (Continued ) Based on
Unexplained (new markets dummy and country of listing dummies) Total difference in first-day return
Model [1]; Table 6 (%) [1]
Model [2]; Table 6 (%) [2]
Model [3]; Table 6 (%) [3]
Model [4]; Table 6 (%) [4]
Model [4]; Table 7 (%) [5]
9.93
9.70
13.75
15.03
13.18
22.34
22.34
22.55
22.55
22.55
Note: This table shows the decomposition of the difference in average first-day returns between new and main markets. We use the regression coefficients of columns [1] to [4] of Table 6 and column [4] of Table 7 to decompose the difference in average first-day returns into component causes. This difference in average first-day returns equals 22.34 percentage points when using the full sample of 1,120 IPOs (columns [1] and [2]) and 22.55 percentage points when using the sample of 886 bookbuilt IPOs (columns [3], [4] and [5]). We compute the difference in first-day returns attributable to differences in pre-IPO ownership variables, IPO secondary sales variables, firm and offer characteristics, post-pricing spillover variables and the bubble years dummy by multiplying the regression coefficients on these variables with the difference in the average value of that variable between new and main markets. For example, we multiply the coefficient of the participation ratio in regression model [1] of Table 6 (that equals −0.238) with the difference between the average participation ratio of new market IPO firms (7.45%, as reported in Table 2) and the average participation ratio of main market IPO firms (16.67%, as reported in Table 2). This yields −0.231 × (7.45–16.67%) = 2.19%. Note that we use the difference in averages of variables between new and main markets for bookbuilt IPOs in columns [3], [4] and [5]. These average values are different from the ones reported in Tables 1, 2 and 3 that relate to the full sample of 1,120 IPO firms. Explained is the part of the difference in average first-day returns that can be explained by differences in pre-IPO ownership variables, IPO secondary sales variables, bundling, firm and offer characteristics, post-pricing spillover variables and the bubble years dummy. Unexplained is the part of the difference in average first-day returns that cannot be explained. It consists of two components. The first component is the product of the (unreported) regression coefficients on the country of listing dummies and the difference in the fraction of new market IPO firms and main market IPO firms that list in these countries. The second component is the regression coefficient on the new markets dummy variable.
first-day returns that cannot be explained. It consists of two parts. The first part is the product of the (unreported) regression coefficients on the country of listing dummies for France, Belgium, the Netherlands and Italy and the difference in the fraction of new market IPO firms and main market IPO firms that list in these countries. The second part is the regression coefficient on the new markets dummy variable. We observe that we can not explain 9.7 percentage points to 15 percentage points of the difference in first-day returns.
Pricing Initial Public Offerings on Europe’s New Stock Markets
55
6. CONCLUSIONS This chapter reports what has changed in pricing European IPOs with the advent of new stock markets. The advent of these new markets belongs to one of the most important recent developments in the continental European IPO market (Ritter, 2003). We investigate three potential changes in the IPO pricing process related to new stock markets. First, we examine the role of pre-IPO ownership structure and secondary sales. We find that first-day returns are a negative function of the participation ratio and the size of CEO sales at the IPO. This may, at least in part, explain why first-day returns on new stock markets are higher than on main stock markets. Because pre-IPO owners, especially CEOs, of new market companies sell less shares at the IPO they experience less incentives to control underpricing costs. Second, we investigate whether bundling IPO deals are related to first-day returns. As predicted by theory and consistent with Ljungqvist and Wilhelm (2002), we find that first-day returns are lower when IPO volume is higher. During these periods underwriters can spread the cost of information production over multiple firms, thereby lowering underpricing costs. We find that the increased opportunity for bundling IPO deals has been especially important to control the underpricing costs of new market IPO firms. Third, we study the relation between first-day returns and price revisions. In line with the partial adjustment theory, we find that price revisions are positively related to first-day returns. Given their higher valuation uncertainty, we expect that underwriters of new market issues extract more information about the value of IPO shares from schedules of investors in the pre-market phase of bookbuilding and therefore would need to compensate these investors in the form of higher first-day returns. We find that this accounts for a part of the difference in the average first-day returns between new and main markets. Additionally, we find that differences in firm characteristics (e.g. age, company size and industry) account for part of the difference in first-day returns. The higher fraction of new market companies that went public during the hot issue market of 1999/2000 (proxied by the bubble years dummy) also contributes to the higher average first-day returns of new market IPOs. However, we conclude that a large part of the difference in average first-day return cannot be explained by differences in sample characteristics between new and main markets.
NOTES 1. The French Nouveau March´e, the German Neuer Markt (defunct as from 2003), Euro.NM Belgium, the Dutch Nieuwe Markt and the Italian Nuovo Mercato were members
56
GIANCARLO GIUDICI AND PETER ROOSENBOOM
of the Euro-NM network. However, the Euro-NM initiative failed to establish operating links between the member markets. The network was disbanded on 31 December 2000, leaving full autonomy to the single national markets. 2. We exclude privatisation IPOs because the pricing process in such offerings may be influenced by specific political objectives, such as dispersing share ownership in order to promote equity investing or currying favour with voters. 3. Foreign issuers are located in Austria (17 IPOs), Denmark (1), Hungary (1), Ireland (3), Israel (13), Liberia (1), Luxembourg (2), South-Africa (1), Sweden (1), Switzerland (5), the United Kingdom (14) and the United States (14, mostly with contemporary dual-listing on a U.S. exchange). 4. Following Derrien and Womack (2003), we do not consider the French Premier March´e. They report that most IPO activity in France took place on the Second March´e and Nouveau March´e with only a few large offerings, such as privatizations, being made on the Premier March´e. 5. High-tech companies are active in SIC codes 3571, 3572, 3575, 3577, 3578 (computer hardware), 3661, 3663, 3669 (communications equipment), 3674 (semiconductors), 3812 (navigation equipment), 3823, 3825, 3826, 3827, 3829 (measuring and controlling devices), 3841, 3845 (medical instruments), 4812, 4813 (telephone equipment), 4899 (communications services) and 7370, 7371, 7372, 7373, 7374, 7375, 7378 and 7379 (software). We collect SIC codes from COMPUSTAT Global Vantage and Worldscope Disclosure. We identify European Internet firms using the list provided by Knauff et al. (2003). They provide a list of 138 European Internet IPOs based on membership of the Bloomberg European Internet Index and talks with investment bankers. 6. Ljungqvist and Wilhelm (2003) report that 54.6% of U.S. IPO firms during 1996–2000 report negative earnings per share and 57.5% are from the Internet and technology sector. The average (median) age of U.S. IPO firms during this period equals 13.6 years (7 years). 7. Loughran and Ritter (2003) report similar findings for U.S. IPOs. They find that the average (median) amount of “money left on the table” increased from $8.3 million ($2 million) during 1990–1998 to $77.7 ($25.2) million in 1999–2000. Most of the amount comes from a minority of IPOs. During 1990–1998, 1.2% of U.S. IPOs doubled in price on the first day of trading. During the bubble years 1999–2000, the percentage grows up to 22.7%. 8. The United Kingdom is an exception. Ljungqvist and Wilhelm (2002) show that U.K. IPOs are rarely priced using the bookbuilding procedure. Instead, placings (fixed-price offerings) are a common method of bringing securities to listing. Placings are not registered for offering to the public at large but only to institutional investors or wealthy individuals. In contrast, French, German and Italian IPOs are typically priced using the bookbuilding procedure and have retail tranches for private investors. This is the major reason why we do not consider the United Kingdom in this study. 9. We could not identify the bookbuilding range for 12 IPOs reducing the bookbuilding sample from 898 to 886 firms. 10. When constructing the underwriter market share measure we combine local offices and take into account name changes of underwriters because of mergers and acquisitions in the underwriting industry. Our underwriter market share measure identifies the following top three underwriters for each local market: France: BNP Paribas, Cr´edit Lyonnais, Soci´et´e G´en´erale, Germany: Deutsche Bank, Dresdner Bank and DG Bank, Belgium/NASDAQ Europe: Fortis Bank, Goldman Sachs and Bank Brussel Lambert (part
Pricing Initial Public Offerings on Europe’s New Stock Markets
57
of ING), the Netherlands: Goldman Sachs, ABN AMRO and ING Barings, Italy: San Paolo IMI, IntesaBCI and Mediobanca. These underwriters, other than the Italian banks, are all mentioned on the list of most active underwriters by Ljungqvist et al. (2003). 11. Ljungqvist and Wilhelm (2003) show that only 27.5% of U.S. IPOs from 1996 to 2000 sold secondary shares with an average (median) size of 3.6% (0%) of pre-IPO shares. CEO sales occurred in only 9.7% of U.S. IPO firms. In the U.S., 14.5% of venture capitalists sell shares at the IPO, and the fraction of corporate-backed IPOs with corporate sales at the IPO is 19.9%. 12. One potential problem with our analysis is that the choice to go public on a new stock market could be endogenous. However, listing requirements are likely to make it exogenous to the analysis. 13. We have alternatively measured IPO volume at the industry level. The IPO industry volume variable measures the number of firms going public in the same Fama-French (1997) industry during the period of 30 days before and 10 days after the pricing date. We find an inverse relation between first-day returns and IPO industry volume that is siginifcant at the 10% level (not shown). 14. We obtain pricing dates and dates at which the price range is determined from the Commission Op´erations des Bourse (stock market regulator) for France and from (preliminary) prospectuses for Germany, Belgium, the Netherlands and Italy. 15. We test for endogeneity using the Durbin-Wu-Hausman test as augmented by Davidson-MacKinnon (1993). This test can be formed by including the residuals of each endogenous right-hand side variable, as a function of all exogenous variables, in a regression of the original model and then testing whether the coefficient on the residuals are significantly different from zero. We have tested for endogeneity of price revision to underwriter market share using the Davidson-MacKinnon test. We find that the test statistic is not significant (p-value = 0.64). We therefore treat underwriter market share as an exogenous variable in the price revision regressions. 16. We test for endogeneity using the Davidson-MacKinnon (1993) test. We find that first-day returns are endogenous to price revisions (p-value < 0.01) as well as underwriter market share (p-value = 0.02). A necessary condition for the system to be identified is that the number of exogenous variables that are not included in an equation is at least as large as the number of endogenous variables included in that equation. This necessary condition for identification is satisfied in our model. We need at least two valid instruments to identify our system. We use m Revisionpre-pricing and Revisionpre-pricing as our instruments for price revision and the log of expected proceeds as our instrument for underwriter market share. In order to be valid instruments these variables need to correlate with price revision and underwriter market share, respectively, but not with first-day returns. We find that m Revisionpre-pricing and Revisionpre-pricing are correlated with price revisions (correlation coefficients are 0.26 and −0.29, respectively) and less correlated with first-day returns (correlation coefficients are 0.16 and −0.08, respectively). We find that the log of expected proceeds is correlated with underwriter market share (correlation coefficient equals 0.31) but uncorrelated to first-day returns (correlation coefficient equals 0.03). 17. Another potential problem is that first-day returns may be endogenous to secondary sales. We have randomly selected 135 bookbuilt IPOs from our sample (about 15% of the bookbuilt IPO sample) and collect both the preliminary and final IPO prospectus. This allows us to investigate whether the number of shares is adjusted under the influence of information collected during the bookbuilding period. However, we find that pre-IPO
58
GIANCARLO GIUDICI AND PETER ROOSENBOOM
owners only very infrequently change the number of shares they sell in the IPO. In particular, we find that only 11 companies (8.1% of total) change the number of secondary shares. The change in the number of secondary shares is small with an average of 0.5% and a median of −3.6% of the number of secondary shares mentioned in the preliminary prospectus. This makes it very unlikely that first-day returns are endogenous to secondary sales in continental Europe.
ACKNOWLEDGMENTS Giancarlo Giudici acknowledges funding from CNR and from Cofinanziamento MIUR. Peter Roosenboom acknowledges funding from ERIM. We thank Andrea Randone, Mark Koevoets and Willem Schramade for their excellent research assistance. We thank Jay Ritter and seminar participants at the EFMA 2002 Annual Meeting in London, at the 2002 Tor Vergata Financial Conference in Rome, at Erasmus University Rotterdam and at the University of Bologna for helpful comments and suggestions. We are grateful to Josef Schuster, Sigrid Vandemaele and Marno Verbeek for their help. All errors are our own.
REFERENCES Aussenegg, W., Pichler, P., & Stomper, A. (2003). IPO pricing with bookbuilding and a when-issued market. Working Paper, Vienna University of Technology and Boston College. Baron, D. (1982). A model of the demand of investment banking advising and distribution services for new issues. Journal of Finance, 37, 955–976. Beatty, R. P., & Ritter, J. R. (1986). Investment banking, reputation and the underpricing of initial public offerings. Journal of Financial Economics, 15, 213–232. Benveniste, L. M., Busaba, W., & Wilhelm, W. J. (2002). Information externalities and the role of underwriters in primary equity markets. Journal of Financial Intermediation, 11, 61–86. Benveniste, L. M., Ljungqvist, A. P., Wilhelm, W. J., & Yu, X. (2003). Evidence of information spillovers in the production of investment banking services. Journal of Finance, 58, 577–608. Benveniste, L. M., & Spindt, P. A. (1989). How investment bankers determine the offer price and allocation of new issues. Journal of Financial Economics, 24, 343–361. Bradley, D. J., & Jordan, B. D. (2002). Partial adjustment to public information and IPO underpricing. Journal of Financial and Quantitative Analysis, 37, 595–616. Davidson, R., & MacKinnon, D. G. (1993). Estimation and inference in econometrics. New York: Oxford University Press. Derrien, F., & Womack, K. L. (2003). Auction vs. book-building and the control of underpricing in hot IPO markets. Review of Financial Studies, 16, 31–61. Fama, E. F., & French, K. R. (1997). Industry costs of equity. Journal of Financial Economics, 43, 153–194. Habib, M. A., & Ljungqvist, A. P. (2001). Underpricing and entrepreneurial wealth losses: Theory and evidence. Review of Financial Studies, 14, 433–458.
Pricing Initial Public Offerings on Europe’s New Stock Markets
59
Hanley, K. W. (1993). The underpricing of initial public offerings and the partial adjustment phenomenon. Journal of Financial Economics, 34, 231–250. Hoffman-Burchardi, U. (2001). Clustering of initial public offerings, information revelation and underpricing. European Economic Review, 45, 353–383. Ibbotson, R. G., & Jaffe, J. F. (1975). Hot issue markets. Journal of Finance, 4, 1027–1042. Jenkinson, T., Morrison, A., & Wilhelm, W. J. (2003). Why are European IPOs so rarely priced outside the indicative price range? Working Paper, University of Oxford. Knauff, P., van der Goot, T., & Roosenboom, P. (2003). Is accounting information relevant to valuing European Internet IPOs? In: A. Ginsberg & I. Hasan (Eds), New Venture Investment: Causes and Consequences (pp. 257–280). North Holland: Amsterdam. Ljungqvist, A. P., Jenkinson, T. J., & Wilhelm, W. J. (2003). Global integration of primary equity markets: The role of U.S. banks and U.S. investors. Review of Financial Studies, 16, 63–100. Ljungqvist, A. P., Nanda, V., & Singh, R. (2004). Hot markets, investor sentiment, and IPO pricing. Forthcoming in Journal of Business. Ljungqvist, A. P., & Wilhelm, W. J. (2002). IPO allocations: Discriminatory or discretionary? Journal of Financial Economics, 65, 167–201. Ljungqvist, A. P., & Wilhelm, W. J. (2003). IPO pricing in the dot-com bubble. Journal of Finance, 58, 723–752. Loughran, T., & Ritter, J. R. (2002). Why don’t issuers get upset about leaving money on the table in IPOs? Review of Financial Studies, 15, 413–443. Loughran, T., & Ritter, J. R. (2003). Why has IPO underpricing changed over time? Working Paper, University of Florida. Lowry, M., & Schwert, G. W. (2002). IPO market cycles: bubbles or sequential learning. Journal of Finance, 57, 1171–1200. Ritter, J. R. (1984). The “hot” issue market of 1980. Journal of Business, 57, 215–240. Ritter, J. R. (2003). Differences between European and American IPO markets. European Financial Management, 9, 1–23. Ritter, J. R., & Welch, I. (2002). A review of IPO activity, pricing and allocations. Journal of Finance, 57, 1795–1828. Subrahmanyam, A., & Titman, S. (1999). The going-public decision and the development of financial markets. Journal of Finance, 54, 1045–1082. White, H. L. (1980). A heteroscedastic-consistent covariance matrix estimator and a direct test of heteroscedasticity. Econometrica, 48, 817–838.
FINANCING GROWTH AND INNOVATION THROUGH NEW STOCK MARKETS: THE CASE OF EUROPEAN BIOTECHNOLOGY FIRMS Fabio Bertoni and Pier Andrea Randone ABSTRACT This chapter analyses how capital is raised and employed by a sample of 28 European biotechnology companies listed on Europe’s new stock markets from 1996 to 2000. We find that biotechnology companies rely heavily on IPO proceeds in order to finance their growth. We compare the behaviour of European firms to a sample of comparable U.S. firms. The analysis reveals that European companies tend to raise more capital at the IPO and to invest more aggressively in the short-run, whereas U.S. biotech firms tend to have more cash available before the IPO and invest more conservatively in the short-run.
1. INTRODUCTION Several recent studies consider venture capital as a key success factor for innovation and growth as well as one of the fundamental competitive advantages of the U.S. economy (Black & Gilson, 1998; Hellmann & Puri, 2000). Kortum and Lerner (2000) show that venture capital, despite being only a small portion The Rise and Fall of Europe’s New Stock Markets Advances in Financial Economics, Volume 10, 61–79 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3732/doi:10.1016/S1569-3732(04)10003-0
61
62
FABIO BERTONI AND PIER ANDREA RANDONE
of the overall flow of financing, is directly responsible for 15% of the innovative activity in the U.S. because of its capacity to finance risky but profitable investments. Jeng and Wells (2000) argue that efficient capital markets are a necessary condition for the “venture capital cycle” to work properly. Without efficient capital markets venture capital cannot be mobilised and invested in new ventures. During the 1990s European countries have been eager to promote private equity investments and the growth of stock exchanges for high growth firms. The efforts seem to have been successful, and in 1999 and 2000 private equity in Europe grew at significant rates. Moreover today all major European countries have developed new stock markets (e.g. EASDAQ, the Neuer Markt in Germany, the Nouveau March´e in France, the Nuovo Mercato in Italy). Hundreds of firms in emerging and high growth industries raised capital and financed their growth on these new stock markets. One of the most promising sectors is the biotechnology sector. The European Commission (1993) stated that biotech activity is one of the most crucial sources of sustainable future development and has supported such activity through the BIOMED, BRIDGE, BIOTECH and BIOMET programs totalling $500 million. Biotechnology is one of the few business sectors attracting an increasing amount of private equity investments both in the U.S. and in Europe. This increase continued in 2001 and 2002, at a time when a dramatic decrease of investment flow hit other technology investments that were previously viewed as important (e.g. Internet-related business). Moreover during the last few years several research projects, such as the human genome-mapping project, highlighted the valuable growth opportunities of biosciences, which are not limited to health applications, but also include wider applications. For example, biotechnology plays a key role in aeronautics, energy production, advanced materials, and the development of security devices (Biotechnology Industry Organization, 2000). The fact that, in 2001, 1,570 biotech companies were operating in Europe (1,273 in the U.S.), employing 61,000 people and generating $10 billion in sales ($24 billion in the U.S.) reflects the fast development of biotech in Europe. This chapter aims to explore the patterns of technology financing in Europe, focussing on the role of new markets. The analysis of biotech companies is interesting because of the growth opportunities of the business, the intensity of R&D activity and the high level of entrepreneurial risk. Access to finance is a crucial key success factor in the biotech sector. On average, biotech companies need $500 million in order to develop a new product (Biotechnology Industry Organization, 2000). Moreover, start-up biotech companies face competition from large incumbent multinational companies.
Financing Growth and Innovation Through New Stock Markets
63
We analyse a sample of 28 biotech and drug companies that went public on European new stock markets between 1996 and 2002. We investigate the capability of biotech firms to raise capital at the IPO and we examine the evolution of sales, R&D and marketing costs in the years following the IPO. We compare our sample of European companies with a sample of 38 comparable U.S. companies that went public on the NASDAQ during the same period. We show that European companies raise more capital at the IPO and invest more aggressively in the short-run in order to increase their market share. Conversely, U.S. companies own more cash resources before the IPO, are more willing to postpone the raising of new equity until after the IPO and are more capable to sustain their cash burn. The remainder of this chapter is as follows. Section 2 describes the close link between economic growth, innovation and equity financing. Section 3 describes venture capital activity. Section 4 discusses the descriptive statistics of our sample of biotech companies. Section 5 reports the empirical results. Section 6 concludes.
2. FINANCING INNOVATION Innovation requires investments in R&D and marketing. These investments need substantial financing in order to be successful. The lack of capital is often cited as one of the most important constraints to innovation, especially in high-tech and emerging industries (Giudici & Paleari, 2000a; Gompers & Lerner, 1999; Himmelberg & Petersen, 1994; Moore, 1994). Frequently the entrepreneur’s personal savings are not sufficient to cover the start-up costs and outside finance must be raised. Moreover, a risk-averse founder would not be willing to invest all his wealth in the venture even if he had a sufficient endowment. Debt is usually unavailable to finance start-ups. First, start-ups face higher uncertainty, information asymmetry and risk of failure than well-established firms. Second, high-tech investments are long-term investments. Third, only a small fraction of the initial investment can be recovered in case of liquidation since investments are highly specific and intangible (Binks et al., 1992; Westhead & Storey, 1997). Fourth, interest payments would slow down firm growth. Equity financing provided by investment professionals (e.g. venture capital, private equity and business angels) is often the only alternative to raise outside capital for entrepreneurs (Gompers & Lerner, 1999). However, equity financing is expensive. The higher level of uncertainty and information asymmetry associated with start-ups increase the costs related to external financing (Jensen & Meckling, 1976). The entrepreneur usually has superior knowledge of the future prospects
64
FABIO BERTONI AND PIER ANDREA RANDONE
of the project to be financed. He can use this knowledge to reduce his effort maximising his own utility instead of the project’s value. On the other hand, financiers are not able to discriminate among good and bad projects. Moreover, young firms have a short track record and few public documents are available to outside investors. This increases the cost of collecting information (Binks & Ennew, 1996). Even if information were symmetric, the degree of uncertainty would make valuation of high-tech investments extremely difficult. The value of new ventures stems from intangible assets and real options embedded in the technology being developed. Venture capitalists have experience in providing equity financing to new ventures. We will examine the role of venture capitalists in the next section.
3. THE VENTURE CAPITAL CYCLE According to Gompers and Lerner (1999), three main features characterise venture capital: (i) the active involvement in managing the firm; (ii) the complexity of financing contracts; (iii) the presence of exit mechanisms from the investment. Advisory activities by venture capitalists are very important and increase the likelihood of firm survival. Entrepreneurs often have no business experience at all. Baker and Gompers (2004) find that venture capital backing improves firm outcomes in the long run, significantly reducing the failure rate. Megginson and Weiss (1991) argue that venture capitalists’ equity stakes provide a sort of “certification effect” on the firm’s quality. Venture capital financing allows to separately allocate cash flow rights, voting rights, board rights, liquidation rights and other control rights, as a possible solution to conflicts of interest or agency problems between investors and entrepreneurs (Kaplan & Str¨omberg, 2003; Sahlman, 1990). Several mechanisms are often adopted by venture capitalists to reduce the risk of the investment: (i) the option to take control of the firm under some circumstances (Hellmann, 1998); (ii) the use of a combination of common equity, preferred equity and convertible bonds (Cornelli & Yosha, 2003); and (iii) staged financing where capital is infused at stages corresponding to significant developments in the life of the company (Giudici & Paleari, 2000b). Stage financing limits the venture capitalist’s losses in case of default and represents a threat of abandonment in the short-run. Redemption covenants provide the venture capitalists with the means to extract the original investment from an unsuccessful company, as well as a credible threat of withdrawal over the entrepreneur. Seed capital is the first type of financing a newly founded company might want to secure, in order to fund R&D and commercial expenditures. Start-up
Financing Growth and Innovation Through New Stock Markets
65
investments are targeted at companies gearing up to produce and market their products. At the expansion stage the company has to fund growth opportunities and external acquisitions, enlarging its manufacturing and distribution capacity. Most venture capitalists exit from their investments in one of four ways: (i) sale after the company completes an Initial Public Offering (IPO); (ii) acquisition by another company; (iii) sale-back to management; and (iv) liquidation in the case of unsuccessful ventures. A distinctive feature of venture-backed firms is that the contract that they sign with the venture capitalist often includes some clauses about when and how the firms should go public (Halloran et al., 2000). Moreover going public provides new ways of raising capital for future acquisitions and investments (Ritter, 2002). Venture funding is believed to have a positive impact both on creating jobs and boosting capital markets (Black & Gilson, 1998) and on innovation. Hellmann and Puri (2000) highlight a positive relationship between the market success of innovative firms and whether they obtain venture capital or not. Kortum and Lerner (2000) estimate that venture capital accounts for 15% of recent industrial innovation in the United States. Venture capital therefore plays a key role in the “virtuous circle” shown in Fig. 1. Venture capital distributes financial resources to young and fast-growing firms that might not be able to access other types of funding due to information asymmetries. Companies financed by venture capital tend to invest in R&D and generate innovations that create value for stakeholders. This value creation feeds the economic growth and allows venture capitalists to re-invest their money in new companies. However, other factors such as the legal system, fiscal policy and labour market efficiency can have a moderating influence on the efficiency of the circle.
Fig. 1. The Relationship Among Innovation, Economic Growth and Venture Capital.
66
FABIO BERTONI AND PIER ANDREA RANDONE
Figure 1 shows that the virtuous circle needs an effective exit mechanism for venture capitalists to benefit from their investments and re-invest their money in new entrepreneurial ideas. The efficiency of venture capital financing is correlated to the availability of an efficient and easily accessible stock market (Jeng & Wells, 2000). Traditional stock markets are not appropriate for this task because venture-backed companies are usually: (i) characterised by high expected growth rates but low profitability; (ii) too small to generate enough liquidity; (iii) riskier than firms operating in mature industries. In other words, venture-backed companies bear substantial costs to reduce asymmetries in information when they list on traditional stock markets (Maksimovic & Pichler, 2001). In many developed countries specific stock market segments devoted to high-growth companies were created to solve this problem.
4. DESCRIPTIVE STATISTICS ON BIOTECHNOLOGY COMPANIES After the end of the Internet euphoria in 2001, private equity investments have re-focussed towards industries that show huge potential of growth such as biotech and life sciences. The analysis of the life science technology industry is interesting for several reasons. Barriers to entry are higher than in other industries and the time to market is longer. Moreover, the business is risky and success is subject to highly specific R&D investment and requires long-term financing such as venture capital. In addition, the industry is rapidly growing because of the recent success in genetic manipulation. For example, alliances and partnerships among biotech companies almost doubled in number in the second half of the 1990s (Biotechnology Industry Organization, 2000). As of 2001, 1,570 biotech firms existed in Europe (compared to 1,273 in the U.S.) employing a total of 61,000 people and generating $10 billions in revenues. In this chapter, we compare the growth pattern of European biotech companies that went public on Europe’s new stock markets to that of U.S. biotech companies that went public on NASDAQ. Our sample includes 28 European biotechnology firms or firms producing chemical, pharmaceutical and health-care products that went public on European new markets between 1996 and 2002. The reason why we include both biotech and health-care firms is that the distinction between the two categories of companies is often arbitrary. For example, the German Neuer Markt sometimes classifies companies that do not strictly belong to the biotech industry as biotech (e.g. Novuspharma and Biosearch Italia). Moreover drug synthesis is
67
Financing Growth and Innovation Through New Stock Markets
Table 1. Sample Distribution. Stock Exchange
1996
1997
1998
1999
2000
2001
2002
Total
Neuer Markt Nouveau March´e NASDAQ Europe Nuovo Mercato SWX New Market NM List
– 2 1 – – –
1 1 – – – –
– 2 1 – – –
4 1 – – – –
9 – – 2 2 1
1 – – – – –
– – – – – –
15 6 2 2 2 1
Total
3
2
3
5
14
1
–
28
Note: Table shows the distribution of the sample of 28 biotechnology companies that went public on Europe’s new stock markets from 1996 to 2002.
similar to biotechnology in terms of barriers to entry, time to market and R&D investments. Table 1 presents the distribution of the sample in terms of year and country of listing. Table 1 shows that most of the companies (15) are listed on the German Neuer Markt, 6 are listed on the French Nouveau March´e, two on each of EASDAQ, Nuovo Mercato and SWX New Market, and only one on the Finnish NM List. Most of the IPOs occurred during the hot issue market of 1999 and 2000. No biotech company went public during 2002. Table 2 shows descriptive statistics. Data are collected from IPO prospectuses. Market prices come from Datastream. Biotech firms going public in Europe are 8 years old and have assets amounting to $81 million, on average. Most of the firms do not report profits and the mean loss is $10 million. This loss is substantial when compared to the mean value of assets. Consistent with the findings of Bertoni and Giudici (2003), the equity stake of professional investors such as venture capitalists declines from 28.31 to 18.78%, but their presence remains significant in absolute terms. In other words, venture capitalists realise only a fraction of their capital gain when firms go public. In most of the cases, the incumbent controlling shareholder still retains a majority stake in the company after the IPO. The firms in our sample issued primary shares at the IPO worth $62.8 million. The mean first-day share return (measured as the percentage difference between the first day closing price and the offer price) is equal to 14.03%. The average one-year share performance equals 54.29%. However, this result is due to a small number of companies with large positive returns. The median one-year share performance equals −52.16%. The results are almost the same when adjusted for a benchmark of European Biotech Firms.
68
FABIO BERTONI AND PIER ANDREA RANDONE
Table 2. Descriptive Statistics for the European Sample.
Book value of assets before IPO (million Revenues (million d)a Current EBITDA (million d)a Current net profit (million d)a Company age (years) Employees Revenues/employees (million d) Shareholdings of professional investors Before the listing (%) After the listing (%) Floating capital after the IPO (%) Offer size (million d) Issue of IPO primary shares (million d) First-day share return (%) One-year share performance (%)b
d)a
Mean Value
Median Value
81.309 18.950 −9.793 −8.921 8.0 246 0.138
61.342 9.203 −6.604 −5.790 7.0 123 0.048
28.31 18.78 31.82 66.061 62.760 +14.03 +54.29 (+55.07%)
26.52 17.12 29.10 48.262 41.025 +0.36 −52.16 (−50.41%)
Note: Table shows descriptive statistics for the sample of 28 biotechnology companies that went public on Europe’s new stock markets from 1996 to 2002. a Data reported on the last balance sheet before the IPO. b The performance in parentheses is adjusted using the MSCI Europe Biotech index return (return on the index during the twelve months after the IPO are subtracted from each firm’s raw return); two distressed firms are excluded from the computation. Shareholdings by institutional investors are the fraction of firm’s equity owned by investment funds, commercial banks, insurance and investment institutions. Floating capital after the IPO is computed as the number of shares offered divided by the number of shares outstanding after the IPO. Offer size is computed as the number of shares offered times the offer price.
As pointed out by Giudici and Paleari (2002), average IPO revenues amount to 77% of total assets of the firms compared to a mere 32% of the shares offered. As a result, IPO proceeds represent most of the post-listing assets. Biotechnology firms have special characteristics when compared to firms from other industries that went public on European new markets (Giudici & Roosenboom, 2002). Firms in our sample have lower earnings (both after and before interests, taxes and depreciation). This finding reflects the fact that biotechnology firms heavily invest in R&D. The book value of assets is significantly higher than in other industries. Finally, biotech firms raise more money at the IPO. Again, this is because of the characteristics of the industry that requires substantial investments in R&D, marketing and acquisitions. Next, we compare the characteristics of European biotech firms with their U.S. counterparts. We collect data on a sample of American companies comparable to
69
Financing Growth and Innovation Through New Stock Markets
Table 3. Descriptive Statistics for the U.S. Sample. Mean Value Book value of assets before IPO (million Revenues (million d)a Current EBITDA (million d)a Current net profit (million d)a Company age (years) Employees Revenues/employees (million d) Shareholdings of professional investors Before the IPO (%) After the IPO (%) Floating capital after the IPO (%) Offer size (million d) Issue of IPO primary shares (million d) First-day share returns (%) One-year share performance (%)b
d)a
Median Value
87.031 14.782 −13.248∗ −12.943 6.0∗ 139∗∗ 0.153
69.621 7.792 −12.504 −11.768 5.0 113 0.085
43.53∗ 31.05∗∗ 25.12∗∗ 55.466 54.700 +53.58 −23.60 (−9.79%)
40.85 30.80 22.75 46.928 46.928 +6.25 −32.71 (−26.62%)
Note: Table shows descriptive statistics for the sample of 38 biotechnology companies that went public on NASDAQ from 1996 to 2002. a Data reported on the last balance sheet before the IPO. b The performance in parenthesis is adjusted using the NASDAQ Biotech Index (return on the index during the twelve months after the IPO are subtracted from each firm’s raw return); two distressed firms are excluded from the computation. ∗ , ∗∗ , ∗∗∗ The difference between the U.S. and the European sample is statistically significant at the 10, 5, 1% level, respectively. We refer to Table 2 for variable definitions.
the ones belonging to the European sample in terms of year of listing, employees, book value of assets, revenues and earnings before interest and taxes (EBIT). We choose to analyse only companies that went public on NASDAQ because 95% of biotech companies in the U.S. are listed on that stock market and European new markets have the same characteristics and objectives as NASDAQ has. We include all biotech and health-care firms listed on the NASDAQ between 1996 and 2002 having a number of employees, assets, revenues and EBIT between the highest and the lowest values of the firms in our European sample. The U.S. sample consists of 38 companies 7 of which went public in 1996, 7 in 1997, 1 in 1999, 17 in 2000, 3 in 2001, and 3 in 2002. Descriptive statistics of the U.S. sample are shown in Table 3. On average, U.S. firms are less profitable and younger than European companies but have more employees. Furthermore, we observe that the presence of professional investors such as venture capitalists is higher both before and after the IPO for U.S. firms. This confirms that the development of the private equity market is more advanced in the U.S. than in Europe. Moreover, the floating capital
70
FABIO BERTONI AND PIER ANDREA RANDONE
after the IPO and the capital raised at the IPO are significantly lower in the U.S. compared to Europe.
5. GROWTH PATTERNS OF BIOTECHNOLOGY COMPANIES In this section we analyse the relationship between financial resources (capital raised at the IPO, liquidity available at the IPO, other cash inflows) and expenses (investments, R&D expenses, marketing costs). The comparison between the two samples shows that European and U.S. firms follow different patterns of growth. We collect company’s cash flow statement data from the annual reports after the IPO. Figure 2 shows the time line. In particular, we consider the liquidity available at the IPO, seasoned issues of capital net of capital reimbursements, dividends and interests, revenues and cash costs. We label the year of the IPO as year 0. Unfortunately, several IPOs are too recent to have long-term observations. Two companies have been excluded from the sample, Applig`ene Oncor SA (a French company listed on the Nouveau March´e in 1996) and Pharming Group N. V. (a Dutch company listed on EASDAQ in 1998) because the stock prices of these companies collapsed after the IPO. We define the following variables: S(t): liquidity available at the listing (t = 0) and in the following years (t = 1, 2, . . ., n), including also short-term financial investments; K(0): capital raised at IPO; K(t): net inflow from issues of seasoned capital, capital reimbursements, dividends and interests on debt at year t; R(t): total revenues registered during year t;
Fig. 2. Annual Cash Inflows and Outflows as from the Year of the Listing.
Financing Growth and Innovation Through New Stock Markets
71
I(t): tangible and intangible investments with the exception of R&D investments made during year t; D(t): R&D expenses registered during year t reported in the income statement or the balance sheet; E(t): other cash costs incurred during year t, obtained in residual way by the equation: E(t) = S(t) − S(t + 1) + K(t) + R(t) − I(t) − D(t). The equations above represent the financial balance from year to year: the existing liquidity at the beginning of the period plus the cash inflows during the same period (revenues and net issues of seasoned capital) is equal to the liquidity at the end of the period plus cash expenses during the period (investments, R&D expenses, other costs). We also define: C (T): cumulated cash inflows from the listing to the year T, where: C (T) = in in S(0) + K(0) + Tt=1 K(t) + Tt=1 R(t); Cout (T): cumulated cash outflows from the listing to the year T, where: C out (T) = T T T t=1 I(t) + t=1 D(t) + t=1 E(t). Cin (T) represents the total resources raised by the company from the IPO to year T. Cout (T) represents the total cash expenses from the IPO to year T. The annual difference between cash inflows and cash outflows describes firm’s capability to grow in a sustainable manner. Table 4 shows the results of the analysis for 26 European biotechnology companies from the year of the IPO up to five years afterwards. We observe that the capital raised at IPO is, on average, 12 times larger than the liquidity available at that time. The net capital raised through seasoned equity offerings during the years following the IPO is only a small fraction of the total capital raised. Especially during the year 2000, many companies have raised a significant amount of capital on stock markets even though their liquidity was poor and their operating profit was negative. The capability to generate cash improves gradually after the IPO. At the same time, R&D investments and other expenses increase. Long-term investments are less relevant compared to other items. Some patterns can be identified: (a) some companies have substantial cash liquidity before the IPO and have been able to sustain it by means of earnings and seasoned equity offers; (b) some companies raise small amounts of cash at the IPO relying on subsequent revenues and seasoned offers; most of the profitable companies in our sample belong to this group; (c) a few companies use the cash raised at the IPO to reduce debt and liabilities; in these cases the cash balance is negative after the listing;
72
FABIO BERTONI AND PIER ANDREA RANDONE
Table 4. Cash Sources and Investments after the IPO for the European Sample. Listing (t = 0)
t=1
t=2
t=3
t=4
t=5
26
26
26
14
8
6
Liquidity available at the listing S(0) Mean value 5,283 Median value 2,404
– –
– –
– –
– –
– –
Capital raised at the IPO K(0) Mean value Median value
64,802 47,281
– –
– –
– –
– –
– –
Seasoned issues of capital K(t) Mean value Median value Mean cumulated value
– – –
7,824 0,191 7,824
6,804 0,204 15,056
8,563 2,652 31,767
32,467 19,596 58,636
5,124 2,871 84,704
Revenues R(t) Mean value Median value Mean cumulated value
– – –
16,946 11,555 16,946
26,973 15,405 44,542
47,655 31,150 105,190
68,326 35,787 176,814
88,499 43,990 399,460
R&D expenses D(t) Mean value Median value Mean cumulated value
– – –
15,016 9,883 15,016
19,093 15,490 34,654
20,389 20,007 52,525
33,343 21,295 97,338
43,194 24,156 181,533
Investments I(t) Mean value Median value Mean cumulated value
– – –
11,217 8,187 11,217
9,556 2,874 21,196
7,639 4,423 32,069
21,795 4,154 49,075
15,817 3,812 103,450
Other cash costs E(t) Mean value Median value Mean cumulated value
– – –
28,679 14,889 28,679
30,346 18,723 60,156
40,347 28,726 101,522
30,959 21,112 106,957
45,407 17,060 200,771
Liquidity available at the end of the period S(t + 1) Mean value – 50,993 Median value – 36,970
36,672 30,107
32,437 26,340
41,844 39,030
37,064 35,363
Sample size
Note: Table shows the average cash sources (available liquidity, IPO proceeds, seasoned issues, revenues) and investments after the IPO. The sample consists of 26 European biotech companies. All data are in d million.
(d) different patterns of R&D investments appear: some companies develop their skills in-house while others rely upon acquisitions and partnerships; (e) some companies consider marketing expenses as strategic investments (especially to enter foreign markets) and create branches and subsidiaries in other countries;
Financing Growth and Innovation Through New Stock Markets
73
(f) some firms burn the financial resources raised at the IPO at a slow rate, while other firms exhibit very high cash burn rates because of R&D and marketing expenses. The two distressed firms (Applig`ene Oncor and Pharming Group) belong to the latter category; the shareholders of these firms perceived the lack of ability to make profits from the business and refused to infuse new capital; (g) some firms engaged in mergers and acquisitions; mergers are less common than acquisitions. Table 5 shows the same analysis for the U.S. sample. We observe that U.S. companies are endowed with more liquidity before the IPO than their European counterparts. Moreover, U.S. companies are better able to raise cash after the listing in seasoned equity offerings. In addition, U.S. companies invest more financial resources in R&D and fewer resources in fixed assets and other costs than European ones. In order to better compare the characteristics of European and U.S. companies, we construct ratios in order to avoid possible distortions caused by the fluctuation of the $/? exchange rate during the 1990s: K(0)/S(0): capital raised at the IPO compared to available liquidity before IPO; T K(t)/K(0): total net capital raised in seasoned issues compared to the capt=1 ital raised at the IPO; K(0)/C (T): capital raised at the IPO compared to the total financial sources in raised from the listing to year T; T K(t)/C (T): net capital raised in seasoned issues compared to the total in t=1 financial sources raised from the listing to year T; T R(t)/C (T): total revenues compared to the total financial sources raised in t=1 from the listing to year T; T D(t)/C out (T): total R&D expenses compared to total cash outflows from t=1 the listing to year T; T I(t)/C (T): total long term investments compared to total cash outflows out t=1 from the listing to year T; T E(t)/C out (T): other cash costs compared to total cash outflows from the t=1 listing to year T; CB(T): “cash burn rate,” i.e. the average net consumption rate of the financial from the listing to the year T, where: CB(T) = (1/T)(C out (T) − T resources T t=1 R(T) − t=1 K(t))/(S(0) + K(0)). In other words, the rate CB(T) represents the mean fraction of liquidity (available before the IPO and raised at the IPO) yearly consumed by the firm. The higher the
74
FABIO BERTONI AND PIER ANDREA RANDONE
Table 5. Cash Sources and Investments after the IPO for the U.S. Sample. Listing (t = 0)
t=1
t=2
t=3
t=4
t=5
38
38
35
18
14
14
Liquidity available at the listing S(0) Mean value 23,556 Median value 17,900
– –
– –
– –
– –
– –
Capital raised at the IPO K(0) Mean value Median value
60,002 49,429
– –
– –
– –
– –
– –
Seasoned issues of capital K(t) Mean value Median value Mean cumulated value
– – –
17,200 4,100 17,200
18,163 2,700 37,301
38,761 11,550 94,260
57,286 44,263 130,601
59,687 20,419 190,287
Revenues R(t) Mean value Median value Mean cumulated value
– – –
13,854 9,950 13,854
12,751 8,000 25,369
19,982 11,562 46,779
15,453 9,740 46,965
12,271 8,354 59,236
R&D expenses D(t) Mean value Median value Mean cumulated value
– – –
23,230 24,085 23,230
25,276 21,400 48,631
36,346 27,516 82,791
35,716 26,727 104,778
37,128 34,473 141,905
Investments I(t) Mean value Median value Mean cumulated value
– – –
7,142 4,566 7,142
9,518 3,100 18,216
7,459 2,250 23,721
7,366 2,017 24,631
5,793 2,318 30,423
Other cash costs E(t) Mean value Median value Mean cumulated value
– – –
24,689 22,054 24,689
25,905 23,942 50,939
27,518 22,022 63,676
28,763 19,544 24,544
35,652 29,897 26,522
Liquidity available at the end of the period S(t + 1) Mean value – 72,670 Median value – 52,354
60,637 33,200
86,420 43,009
79,087 52,891
107,276 59,000
Sample size
Note: Table shows the average cash sources (available liquidity, IPO proceeds, seasoned issues, revenues) and investments after the IPO. The sample consists of 38 U.S. biotech companies. All data are in $ million.
cash burn rate, the faster the company burned its cash resources. When the value is close to zero, it means that cash inflows and cash outflows are in equilibrium. If the company is able to generate free cash flows the rate becomes negative (there is a net accumulation of cash).
75
Financing Growth and Innovation Through New Stock Markets
Table 6. Cash Resources and Expenses for IPO Biotech Companies. Ratio
Sample
T=0
T=1
T=2
T=3
T=4
T=5
K(0)/S(0)
European sample U.S. sample
12.266 2.547
– –
– –
– –
– –
– –
European sample U.S. sample
– –
0.121 0.287
0.226 0.589
0.358 1.235
0.859 2.190
0.938 3.185
European sample U.S. sample
0.925 0.718
0.683 0.524
0.504 0.412
0.351 0.294
0.227 0.217
0.171 0.172
European sample U.S. sample
– –
0.082 0.150
0.114 0.243
0.125 0.363
0.195 0.474
0.160 0.548
European sample U.S. sample
– –
0.179 0.121
0.341 0.183
0.495 0.228
0.560 0.224
0.655 0.213
European sample U.S. sample
– –
0.204 0.130
0.182 0.144
0.156 0.129
0.187 0.122
0.177 0.110
European sample U.S. sample
– –
0.273 0.422
0.299 0.419
0.299 0.454
0.327 0.466
0.351 0.467
European sample U.S. sample
– –
0.522 0.448
0.518 0.437
0.545 0.418
0.486 0.413
0.471 0.422
European sample U.S. sample
– –
43.0% 28.7%
39.5% 32.2%
32.1% 26.5%
18.8% 19.6%
18.2% 17.3%
T
t=1 K(t)/K(0)
K(0)/Cin (T) T
t=1 K(t)/C in (T)
T
t=1 R(t)/C in (T)
T
t=1 D(t)/C out (T)
T
t=1 E(t)/C out (T)
T
t=1 I(t)/C out (T)
CB(T)
Note: This table compares cash resources and expenses between the European and the U.S. sample. Ratios are computed according to formulas discussed in the text.
Table 6 reports the ratios defined above, calculated both for the European and the U.S. sample, using cumulated data from Tables 4 and 5. The comparison between the two samples reinforces several findings: (a) European companies have raised more capital at the IPO than U.S. companies, even when controlling for size and initial cash: 35% of their total cash inflows three years after listing comes from IPO proceeds. Instead, U.S. companies have collected more capital in seasoned equity offerings than their European counterparts. U.S. companies raised more capital in seasoned offerings than in IPOs during the three years following the IPO (the cumulated flow is 36% of the total cash inflows); (b) sales are increasingly important as form of financing, especially for European firms; (c) R&D expenses tend to decrease over time and are higher for U.S. companies; operating and marketing expenses increase over time and are more significant
76
FABIO BERTONI AND PIER ANDREA RANDONE
Fig. 3. Cash Burn Rates CB(T) for the European and U.S. Samples.
for European firms; almost half of overall financial resources are spent on fixed investments (without significant differences between the two samples); (d) the cash burn rate is higher for European companies that more aggressively invest in the short-run than U.S. companies. Only two companies in the European sample have been able to generate positive cash flows within three years. In other words, only two companies have been able to self-finance their growth. The cash burn rate of most recently listed companies is very high. These companies could drain their financial liquidity in the near future if they are not able to raise or generate new cash. Figure 3 shows the “cash burn” rate CB(T) for the European and U.S. sample. There is a common trend. In the long run the cash burn rate falls because the ability to generate cash improves. However, in the short-run, European firms consume financial resources more aggressively. Both European and American firms tend to an equilibrium in the long run and after three years their cash burn rates are similar. The results of our analysis highlight different patterns of growth for the two samples, and prompt questions about the maturity of European biotechnology companies at the time of the IPO. The traditional source of capital for technology growth companies is venture capital, which provides money for investments and R&D to start-up companies that are unable to self-finance their future growth. Once the companies have a portfolio of products and services and exhibit significant revenues, they are ready to go public to raise further capital and finance later-stage growth. Different investors finance the firms in these two stages. Usually professional investors such as venture capitalists finance start-up firms whereas stock exchange investors finance later-stage companies (when the default risk should be lower).
Financing Growth and Innovation Through New Stock Markets
77
Our analysis shows that many companies had a poor ability to generate cash after the IPO and, often, did not succeed in increasing their revenues. Giudici and Paleari (2002) find similar results for a larger sample of IPO companies, and point out that European new markets have been markets for projects where investors ultimately financed start-up companies. From this perspective new stock markets have been a form of public venture capital. This challenges the conventional argument that stock markets are the primary way-out for venture financing and suggests that exchanges have been the real financiers for many European biotechnology companies.
6. CONCLUDING REMARKS It is not yet clear whether the U.S. model (more venture capital, more listed companies, more innovation and more economic growth) will also work in Europe. The effort to improve the European economy by promoting new stock markets has contributed to the growth in the number of firms going public. In this chapter we examine a sample of European biotechnology companies that went public on these new stock markets from 1996 to 2002. We document that the capital raised at the IPO has been the main source of capital to finance R&D, marketing expenses and investments. We find that European biotechnology companies have followed a different pattern of development as compared with U.S. biotech firms. European companies tend to raise more cash at the time of the IPO and invest more aggressively in new acquisitions in the short-run. U.S. companies – though generally younger and less profitable – show a higher availability of cash before the IPO, are better able to raise equity in seasoned issues, and are able to invest in a more sustainable manner. In this context, European new markets have partially replaced professional investors in private equity in the financing of start-up companies. We believe that this might be less efficient, as public investors are less effective than professional ones in monitoring entrepreneurs due to a lack of information and co-ordination. The future of European stock markets is still being discussed and it is not clear which path the integration of financial markets will follow. Some exchanges recently announced their plans to tighten their listing requirements in order to discourage immature firms to go public. However, there are still large differences between Europe and the U.S. regarding corporate governance, taxes, regulations, and labour laws. All these features affect the development of venture capital, so that the benefits of equity financing on economic growth might be less evident. We should not underestimate the role of European governments in
78
FABIO BERTONI AND PIER ANDREA RANDONE
supporting research and development. Such a role has been played in France (the SOFARIS facility), in Germany (through the project Beteiligungskapital f¨ur Technologieunternehmen and the Kreditanstalt f¨ur Wiederaufbau), in Israel (through the Yozma programme), in Finland (through the SITRA Finnish Fund for R&D) or through indirect incentives such as equity investment promotion or tax relieves for IPO companies (for instance in Italy). Nowadays the decline of investments in the Internet sector has lead professional investors to focus their attention on other high-tech businesses, such as wireless technologies and WAP, and the latest mobile generation, where Europe arguably has a competitive edge over the U.S. and where new chances of creating value are expected. The impressive growth in the number of new European biotechnology companies demonstrates that the demand for entrepreneurial finance continues to exist in Europe.
REFERENCES Baker, M., & Gompers, P. A. (2004). The determinants of board structure at the initial public offering. Journal of Law and Economics (forthcoming). Bertoni, F., & Giudici, G. (2003). ‘New’ stock markets in Europe: A ‘new’ exit for venture capital investment. In: A. Ginsberg & I. Hasan (Eds), New Venture Investments: Choices and Consequences (pp. 201–225). Amsterdam: North-Holland. Binks, M., & Ennew, C. (1996). Growing firms and the credit constraint. Small Business Economics, 8, 17–25. Binks, M., Ennew, C., & Reed, C. (1992). Information asymmetries and the provision of finance to small firms. International Small Business Journal, 11, 35–46. Biotechnology Industry Organization (2000). Biotechnology Investors’ Forum – Europe. http://www.bio.org. Black, B., & Gilson, R. (1998). Venture capital and the structure of capital markets: Banks vs. stock markets. Journal of Financial Economics, 47, 243–277. Cornelli, F., & Yosha, O. (2003). Stage financing and the role of convertible securities. Review of Economic Studies, 70, 1–32. European Commission (1993). White paper on growth, competitiveness and employment: The challenge and ways forward into the 21st century. Giudici, G., & Paleari, S. (2000a). The provision of finance to innovation: A survey conducted among Italian technology-based small firms. Small Business Economics Journal, 14, 37–53. Giudici, G., & Paleari, S. (2000b). The optimal staging of venture capital financing when entrepreneurs extract private benefits from their firms. Enterprise and Innovation Management Studies, 1, 153–174. Giudici, G., & Paleari, S. (2002). R&D financing and stock markets. In: M. Calderini, P. Garrone & M. Sobrero (Eds), Corporate Governance, Market Structure and Innovation (pp. 217–239). Cheltenham: Edgar Elgar. Giudici, G., & Roosenboom, P. (2002). Pricing initial public offerings in Europe: What has changed? Working Paper, Politecnico di Milano and Erasmus University Rotterdam.
Financing Growth and Innovation Through New Stock Markets
79
Gompers, P. A., & Lerner, J. (1999). The venture capital cycle. Cambridge: MIT Press. Halloran, M. J., Benton, L. F., Gunderson, R. V., Del Calvo, J., & Kintner, T. W. (2000). Venture capital and public offering negotiation. New York: Aspen Law & Business. Hellmann, T. (1998). The allocation of control rights in venture capital contracts. Rand Journal of Economics, 29, 57–76. Hellmann, T., & Puri, M. (2000). The interaction between product market and financing strategy: The role of venture capital. Review of Financial Studies, 13, 959–984. Himmelberg, C. P., & Petersen, B. C. (1994). R&D and internal finance: A panel study of small firms in high-tech industries. Review of Economics and Statistics, 76, 38–51. Jeng, L. A., & Wells, P. C. (2000). The determinants of venture capital funding: Evidence across countries. Journal of Corporate Finance, 6, 241–289. Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3, 305–360. Kaplan, S. N., & Str¨omberg, P. (2003). Financial contracting theory meets the real world: An empirical analysis of venture capital contracts. Review of Economic Studies, 70, 281–315. Kortum, S., & Lerner, J. (2000). Assessing the contribution of venture capital to innovation. Rand Journal of Economics, 31, 674–692. Maksimovic, V., & Pichler, P. (2001). Technological innovation and initial public offerings. Review of Financial Studies, 14, 459–494. Megginson, W. L., & Weiss, K. A. (1991). Venture capitalist certification in initial public offerings. Journal of Finance, 46, 879–903. Moore, B. (1994). Financial constraints to the growth and development of small high technology firms. In: A. Hughes & D. Storey (Eds), Finance and the Small Firms (pp. 64–111). London: Routledge. Ritter, J. R. (2002). Investment banking and securities issuance. In: G. Constantinides, M. Harris & R. Stulz (Eds), Handbook of the Economics of Finance (pp. 255–306). Amsterdam: North-Holland. Sahlman, W. (1990). The structure and governance of venture capital organizations. Journal of Financial Economics, 27, 473–521. Westhead, P., & Storey, D. J. (1997). Financial constraints on the growth of high-tech small firms in the UK. Applied Financial Economics, 7, 197–201.
MANAGERIAL INCENTIVES AT THE INITIAL PUBLIC OFFERING: AN EMPIRICAL ANALYSIS OF THE ALTERNATIVE INVESTMENT MARKET Peter Roosenboom ABSTRACT This chapter examines the determinants of managerial incentives at the time of an Initial Public Offering (IPO) on the Alternative Investment Market (AIM) of the London Stock Exchange. We identify a trade-off relation between board monitoring and incentives that is specific to CEOs. We also investigate the role of stock option grants and share transactions at the IPO. We find that the IPO may be used as a wealth diversification opportunity. We report that undiversified managers with large pre-IPO shareholdings receive smaller stock options grants and sell more shares in the IPO than more diversified managers.
1. INTRODUCTION At the time of an Initial Public Offering (IPO), a large part of firm value depends on managers’ investment decisions. Managers have considerable discretion on how to spend the money that has been raised in the IPO. They might invest in negative net-present-value projects to increase the size of the company beyond The Rise and Fall of Europe’s New Stock Markets Advances in Financial Economics, Volume 10, 81–112 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3732/doi:10.1016/S1569-3732(04)10004-2
81
82
PETER ROOSENBOOM
what is optimal. They might decide to spend the money on projects that increase their private benefits of control at shareholders’ expense. Equally problematic, risk-averse managers might decide to pass up positive net-present-value projects that in their view are too risky. In order to motivate managers to act in the best interest of shareholders, standard principal-agent theory advocates aligning the interests of managers with those of shareholders (Jensen & Meckling, 1976). Managers should bear the wealth consequences of their decisions by holding stock and/or options in the employing company. However, the resulting lack of ability to diversify his or her personal investment portfolio increases managers’ risk aversion. When executives become more risk averse, they place a lower value on stock options than the cost to shareholders (Hall & Murphy, 2002; Meulbroek, 2001; Murphy, 1999). This predicts that options should only be granted if the incentive effect (the increased performance created by improved managerial incentives) exceeds the difference between the company’s cost and the manager’s value. One of the main challenges that companies face at the IPO is therefore to harmonize the financial interests of managers with those of outside shareholders in a cost-effective manner. Often executive directors have invested a large part of their personal wealth in shares of the IPO firm. In the months before the IPO, managers generally receive stock options that add to the link between their wealth and shareholder wealth. In addition, managers have their human capital (i.e. employment and income opportunities) invested in the IPO firm. In this chapter we analyze managerial incentives at the IPO using a sample of 188 firms that listed on the Alternative Investment Market (AIM) of the London Stock Exchange from June 1995 to December 1999. AIM is a new stock market that offers small U.K. companies the opportunity to raise capital. We investigate the cross-sectional determinants of wealth-to-performance sensitivity of managers at the IPO. Wealth-to-performance sensitivity measures the increase in the amount of executive wealth (composed of shareholdings, option holdings and human capital) per £1,000 increase in shareholder wealth (Jensen & Murphy, 1990). Additionally, we examine the factors that influence option grants and the decisions of executive directors to buy or sell shares in the IPO. Since founders and managers with long tenure and large shareholdings are undiversified with respect to firm-specific risks, we hypothesize that these undiversified executives sell stock in the IPO for wealth diversification reasons. On the other hand, newly hired executives with small shareholdings are likely to hold a larger part of their personal wealth outside the firm. These more diversified executive directors may therefore be willing to retain their shareholdings or even to buy additional shares in the IPO firm to increase their shareholdings. We
Managerial Incentives at the Initial Public Offering
83
also analyze whether undiversified executives receive smaller option grants than more diversified managers. We hypothesize that because undiversified executives already have a large exposure to firm-specific risks, they are less willing to bear additional firm-specific risks in the form of equity-based compensation. This study adds to the existing literature in two important ways. Firstly, the analysis is organized surrounding a corporate event, the IPO. At the IPO the average company raises a large amount of capital to pursue its growth plans. Managers often have special knowledge that is crucial to successfully exploit these growth options, whereas outside shareholders are relatively uninformed about the firm’s prospects. The IPO therefore presents an interesting opportunity to analyze managerial incentives at small and entrepreneurial companies. A study of smaller firms is of interest because the factors that influence incentives in this class of firms may differ from those in large public U.K. firms that have been examined in previous studies (e.g. Conyon & Murphy, 2000; Conyon et al., 2000). Secondly, due to data constraints previous studies have largely examined the wealth-to-performance sensitivity of Chief Executive Officers (CEOs). It is not clear whether the results of those studies may be generalized to all executive directors or whether differences exist between CEOs and other executives. Tying executive wealth to shareholder wealth through share ownership and stock options may be particularly important to executive directors just below the CEO level. IPO firms need to ensure that remuneration is competitive compared with comparable publicly traded companies. Otherwise, executive directors may be recruited by other firms or voluntarily leave the IPO firm to pursue more attractive job opportunities elsewhere. As investors that buy shares in the IPO seek stability and continuity in the senior management team in the post-IPO period, stock ownership and options may enhance and maintain loyalty and commitment during this time of change and preparation for future growth and development. A senior management team that is highly motivated, competitively rewarded and has its interests aligned with shareholders will inspire confidence in IPO investors.1 Accordingly, Welbourne and Andrews (1996) show that the use of stock option programs and profit sharing at U.S. IPO firms improves the firm’s chances of survival. To investigate any potential differences between CEOs and other executive directors, we analyze three groups of 188 chief executive officers (CEOs), 137 finance directors and 280 other executive directors that are employed by the 188 sample firms.2 We find that wealth-to-performance sensitivity (WPS) is higher if the manager co-founded the firm, chairs the board of directors, and has been employed by the firm for a larger number of years. In addition, we report that the WPS of
84
PETER ROOSENBOOM
CEOs is inversely related to board monitoring by large outside shareholders and equity incentives of independent directors. These relations are less marked for finance directors and other executive directors. This finding suggests an agency costs tradeoff between incentives and board monitoring that is unique to CEOs. Consistent with an efficient contracting view, board monitoring reduces the need to tie CEO wealth to firm performance. However, board size relates negatively to the WPS of CEOs. Smaller boards force CEOs to bear more firm-specific risk. Consistent with Yermack (1996), this suggests that small board size together with high-powered CEO incentives may address the agency problem associated with going public. Growth options positively impact the WPS of CEOs. Since these firms derive a large portion of their value from assets not yet in place, they are riskier and controlling agency problems by monitoring managers’ efforts becomes difficult. CEO incentive alignment may mitigate these monitoring problems in firms that derive a large part of their value from future investments. Managers at IPO firms are typically unable to diversify away the risk associated with their wealth, since their human capital is largely invested in a single position of employment. In addition they often own large shareholdings in the IPO firm. This constraint on the executive’s ability to reduce personal risk affects their tolerance for additional risk. When we investigate the changes in managerial incentives due to share transactions at the IPO, we find that managers, other than finance directors, sell more shares at the IPO when their pre-IPO WPS is high (i.e. when they are undiversified through their human capital and substantial pre-IPO shareholdings). Another finding is that managers, other than CEOs, obtain less option grants when their WPS is already high before the IPO. These results are consistent with some managers seeking to diversify their wealth and using the IPO as a wealth diversification opportunity. The rest of the chapter is organized as follows. Section 2 describes the data. Section 3 explains the methodology and measurement of variables. Section 4 presents the empirical results. Section 5 concludes.
2. SAMPLE DESCRIPTION AND VARIABLE MEASUREMENT 2.1. Data and Sample Selection The initial sample consists of 553 companies that were admitted to trading on the Alternative Investment Market (AIM) of the London Stock Exchange from June 1995 to December 1999. We exclude companies that transferred from the Unlisted Securities Market (18 firms), Official List (24), Rule 4.2 (85) and OFEX
Managerial Incentives at the Initial Public Offering
85
(11).3 These firms were previously traded on these stock markets and can therefore not be considered IPO firms. Additionally, re-admissions following cancellation (74) are omitted. The vast majority of re-admissions involve AIM companies that change their name and subsequently resume trading under this new name. Next, we exclude financial services companies (68 firms) and non-U.K. companies (26). The financial reporting and regulatory environment of these companies is different from domestic non-financial companies. As a final step, we exclude companies that go public using the introduction method (34 firms). Introductions do not enable companies to raise new equity capital at the IPO, but only to sell existing shares to the public. Moreover, the admission documents that accompany introductions do not include sufficiently detailed information for our research purposes. The prospectuses of 25 IPO firms could not be retrieved. The final sample therefore consists of 188 domestic non-financial firms that went public on AIM from its launch in June 1995 to December 1999. There is no distinct industry clustering. The sample includes 40 different two-digit Standard Industrial Classification (SIC) industry groups. Computer programming, data processing and other computer-related services is the most important SIC industry group with 30 IPO firms. Other major SIC industry groups are amusement and recreation services (14 firms), management services (10), electronics and other electrical equipment (10), drinking and eating places (9) and printing and publishing (9). AIM requires no minimum shares to be in public hands, no trading record and no prior shareholder approval for acquisitions or disposals. In addition there is no minimum market capitalization.4 The key regulation which companies must satisfy is derived from the European Public Offers Directive. Although admission documents do not have to be pre-vetted by the London Stock Exchange, the company is required to appoint a nominated adviser – which is a firm of experienced corporate finance professionals – and a nominated broker. The nominated adviser assists the company in preparing the mandatory admission document. The document must include all relevant information about the company and its activities – including financial information and projections, together with details of all directors. We use this admission document as our primary source of information. Share prices are collected from Datastream.
2.2. Firm Characteristics Panel A of Table 1 shows summary statistics on firm size and risk. Market capitalization is measured as the number of post-IPO shares times the closing market price on the first trading day and amounts to £13 million, evaluated at the
86
Table 1. Sample Characteristics.
Panel A: Firm characteristics Market capitalization on first-day of trade (£000) Pro-forma total assets (£000) Gross IPO proceeds (£000) Net proceeds to company (£000) Industry standard deviation (%) Tangible fixed assets (% of assets) Market-to-book ratio R&D expenditure (% of sales) Firm age (years)
Standard Deviation
Minimum
Q1
Median
Q3
Maximum
17,553.88 8,183.00 4,730.21 3,620.25 36.09 17.47 5.74 6.93 16.14
15,969.52 7,996.16 5,687.31 5,354.77 14.86 21.19 6.76 33.46 24.69
764.89 509.85 300.00 –885.08 15.91 0.00 0.91 0.00 1.00
6,762.60 3,024.47 1,639.04 981.55 26.91 3.26 2.62 0.00 4.00
13,003.98 6,102.50 3,223.32 2,292.99 34.54 7.83 3.99 0.00 8.00
24,981.17 10,152.22 5,686.73 4,159.54 42.66 25.21 6.09 0.00 16.00
113,880.40 51,969.57 49,171.22 47,532.18 150.25 92.66 53.25 280.77 132.00
0.36 5.37 60.50 4.24 35.26 2.79 0.10
0.48 1.54 16.67 9.87 16.25 6.57 0.37
0.00 2.00 16.67 0.00 0.00 0.00 0.00
0.00 4.00 50.00 0.00 25.00 0.03 0.00
0.00 5.00 60.00 0.00 33.33 0.22 0.00
1.00 6.00 75.00 0.00 50.00 1.93 0.00
1.00 12.00 100.00 60.00 75.00 49.69 3.90
PETER ROOSENBOOM
Panel B: Board and ownership characteristics Executive chairman (dummy) Board size Executive directors (% of board size) Affiliated directors (% of board size) Independent directors (% of board size) Stock ownership per independent director (%) Options per independent director (%)
Mean
0.66 0.27
0.48 0.45
0.00 0.00
0.00 0.00
1.00 0.00
1.00 1.00
1.00 1.00
Note: Table 1 shows firm, board and ownership characteristics for a sample of 188 firms that went public on the Alternative Investment Market (AIM) of the London Stock Exchange from June 1995 to December 1999. Panel A shows firm characteristics. We measure market capitalization on the first day of trade as the number of post-IPO shares times the closing market price on the first day the shares starts trading on AIM. Pro-forma total assets are taken from the pro-forma statement of net assets included in the prospectus. This pro-forma statement of net assets presents the balance sheet of the IPO firm shortly after going public. Gross IPO proceeds is calculated as the number of shares sold in the IPO times the offer price. Net proceeds to the company are the gross proceeds to the company minus the costs of going public. Gross proceeds to the company are calculated as the number of newly issued shares times the offer price. Industry standard deviation equals the industry median annualized standard deviation of monthly returns for the year prior to the IPO. Industries are defined at the 4-digit SIC code level provided that 3 listed firms are available. Otherwise, we shift to a 3-digit or 2-digit SIC code level. Tangible fixed assets are expressed as a percentage of pro-forma total assets and are taken from the pro-forma statement of net assets included in the prospectus. The market-to-book ratio is the ratio of first-day market capitalization and pro-forma book value of equity. In general, pro-forma book value of equity is the sum of pre-IPO book value of equity plus the net proceeds to the company. Research and development expenditure is expressed as a percentage of sales in the year prior to the IPO. Firm age is the number of years the firm has been in existence before the IPO. Panel B presents board and ownership characteristics. Executive chairman is a dummy that takes on the value one if an executive director chairs the board. Board size is the number of executive and non-executive directors that serve on the board. Executive directors are expressed as a percentage of total board size. Affiliated directors are defined as those non-executive directors that are co-founders of the company, retired executives, part-time employees, or family of executive directors or founders. Non-executive directors that earn more than £30,000 in fees are also classified as affiliated. Independent directors are non-affiliated non-executive directors and expressed as a percentage of total board size. The number of shares and options held by independent directors is expressed as a percentage of post-IPO shares. Non-management shareholders include venture capitalists, industrial and commercial companies and institutional investors that own shares in the IPO firm. A dummy indicates whether a non-management shareholder owns more than 5% of pre-IPO shares. Another dummy indicates whether at least one non-management shareholder is represented in the board of directors. All monetary amounts are expressed in 1999 pounds using the Retail Prices Index as inflation adjustment.
Managerial Incentives at the Initial Public Offering
Outside shareholder owns more than 5% (dummy) Outside shareholder sits on board (dummy)
87
88
PETER ROOSENBOOM
median. Pro-forma total assets is taken from the pro-forma statement of net assets included in the prospectus, and its median value equals £6.1 million. In general, the pro-forma assets are determined as the sum of the pre-IPO total assets plus the net IPO proceeds to the company. In the median IPO £3.2 million worth of shares are sold to the public. On average, the shares sold in the IPO represent 32.5% of the number of post-IPO shares. In the average IPO, 85% of shares that are sold in the IPO are newly issued. The company receives the proceeds of these newly issued shares. The remaining 15% of IPO shares are being sold by pre-IPO shareholders. The net IPO proceeds to the company are substantial compared to firm size. The median company raises £2.3 million by selling newly issued shares to the public. This highlights the importance of the IPO as a corporate event. The typical company raises a large amount of capital to finance its future growth plans. IPO firms are characterized by an inherent lack of price history. This makes it difficult to measure firm risk ex ante. Following Baker and Gompers (1999), we therefore use an industry-matching procedure to assess company risk. First, we gather the names and SIC codes of all U.K. companies that were publicly traded during the sample period from Worldscope Disclosure.5 We then match IPO firms to listed firms in the same industry. We define industries at the 4-digit SIC code level provided there are at least three listed firms.6 Otherwise, we shift to a 3-digit or 2-digit SIC code level. Next, we download the monthly share prices of the industry-matched firms from Datastream for the year before the IPO date. As a final step, we calculate the annualized standard deviation of the monthly stock returns for each of the industry-matched firms. We use the median of those standard deviations as our ex ante proxy for the IPO firm’s risk. Panel A of Table 1 shows that the median industry standard deviation equals 34.5% per year. The average tangible fixed assets are equal to 17.5% of pro-forma total assets. The market-to-book ratio is calculated as the ratio of first-day market capitalization to pro-forma book value of equity. In general, pro-forma book value of equity is the sum of pre-IPO book value of equity plus the net proceeds to the company. The average market-to-book ratio equals 5.7. The typical IPO firm spends 6.9% of sales on research and development in the year prior to its IPO. The average company age is 16 years with a median of 8 years.
2.3. Board and Ownership Characteristics Panel B of Table 1 presents information on the composition of the board of directors. In 36% of the sample firms, an executive director chairs the board of directors. This does not only involve CEOs being the chairman. In 17% of the
Managerial Incentives at the Initial Public Offering
89
firms, the CEO chairs the board, in the remaining 19% other executive directors serve as chairperson. The median board has five members. On average, 60.5% of directors are executives and 39.5% are non-executives. Affiliated directors are defined as those non-executive directors that are co-founders of the company, retired executives, part-time employees, or family of executive directors or founders. Non-executive directors that earn more than £30,000 in fees are also classified as affiliated because they may act as paid advisors to the firm rather than active monitors. We assume that these affiliated directors do not act as independent monitors on account of business or family relationships. On average, 4.2% of all directors are affiliated to management. We classify the remaining 35.3% of directors as independent directors. We assume that these independent directors act as active monitors of top management. The stock ownership per independent director averages 2.8% of post-IPO shares. The mean option holding per independent director is equal to 0.1% of post-IPO shares. Two-thirds of the IPO firms have an outside non-management shareholder that owns more than 5% of pre-IPO shares.7 In 27% of IPO firms at least one non-management shareholder is represented on the board of directors.
2.4. Executives’ Firm-Specific Wealth The unit of analysis in subsequent tests is the individual executive director. In total, the 188 sample firms employ 605 executive directors. We divide these executive directors into three groups on the basis of their job complexity. We distinguish 188 chief executive officers (CEOs), 137 finance directors (comparable to Chief Financial Officers (CFOs) in the U.S.) and 280 other executive directors. Information concerning share and option holdings of each executive director is obtained from admission documents.8 The admission document also contains a description of each executive director’s service agreement. A service agreement specifies the annual base salary earned by the director and any possible bonus arrangements. Panel A of Table 2 presents data for the group of 188 CEOs. All monetary amounts are expressed in 1999 pounds using the Retail Prices Index as inflation adjustment. The average CEO owns 20.2% of post-IPO shares. The monetary value of CEO ownership averages £2.8 million, evaluated using the closing market price on the first day of trading on the stock market. The Black-Scholes value of the average CEO option holdings amounts to £94,000, which covers 0.7% of post-IPO equity (see Section 3.1 for the calculation procedure). Conyon and Murphy (2000) report that the average shareholdings and option holdings of CEOs of the 510 largest U.K. firms in 1997 are 2.13 and 0.24% (expressed as a percentage of outstanding shares). In percentage terms, CEOs at IPO firms therefore
90
PETER ROOSENBOOM
Table 2. Executives’ Firm-Specific Wealth. Mean
Standard Deviation
Minimum
Q1
Median
Q3
Maximum
Panel A: Chief executive officers (N = 188) Equity (%) 20.16 19.82 Equity (£000) 2,891.03 3,550.49 Options (%) 0.70 5.63 Options (£000) 93.81 215.22 Shares transacted (%) 1.00 3.34 Shares transacted (£000) 157.68 689.74 Base salary (£000) 82.63 33.17 Cash compensation (£000) 98.52 44.23 Age (years) 44.88 8.11 Tenure (years) 7.40 6.66
0.00 0.00 0.00 0.00 −6.96 −2,719.77 0.00 0.00 27.00 0.08
5.40 14.76 29.54 93.14 641.23 1,805.42 3,842.97 21,008.72 0.00 0.00 0.82 13.37 0.00 0.00 102.53 1,774.53 0.00 0.00 0.56 33.00 0.00 0.00 80.66 7,420.92 61.48 79.99 100.32 229.47 70.28 93.32 116.36 321.25 38.00 46.00 51.00 66.00 3.00 6.00 10.25 43.00
Panel B: Finance directors (N = 137) Equity (%) 2.65 6.96 Equity (£000) 508.25 2,114.34 Options (%) 0.55 0.96 Options (£000) 97.95 210.03 Shares transacted (%) 0.08 0.91 Shares transacted (£000) 0.65 159.54 Base salary (£000) 63.94 27.29 Cash compensation (£000) 74.95 34.81 Age (years) 42.50 7.34 Tenure (years) 3.72 4.51
0.00 0.00 0.00 0.00 −2.94 −1,310.18 0.00 0.00 28.00 0.08
0.00 0.00 0.00 0.00 −0.02 −1.70 48.00 54.63 38.00 0.75
Panel C: Other executive directors (N = 280) Equity (%) 8.37 11.92 Equity (£000) 1,574.61 3,506.28 Options (%) 0.56 1.24 Options (£000) 83.54 184.65 Shares transacted (%) 0.39 1.71 Shares transacted (£000) 66.89 352.08 Base salary (£000) 69.79 32.40 Cash compensation (£000) 82.16 39.80 Age (years) 43.41 8.60 Tenure (years) 6.42 6.65
0.00 0.00 0.00 0.00 −6.96 −2,719.77 0.00 0.00 25.00 0.08
0.26 39.28 0.00 0.00 0.00 0.00 50.00 60.00 37.00 2.00
0.23 29.71 0.16 16.24 0.00 0.00 60.10 68.28 41.00 2.00
2.16 55.20 261.92 23,270.67 0.80 6.74 109.78 1,862.34 0.00 7.74 0.00 832.52 75.00 165.00 87.91 229.47 47.00 71.00 6.00 30.00
3.67 10.56 62.89 459.60 1,550.89 26,993.03 0.00 0.73 9.46 0.00 89.04 1,774.53 0.00 0.00 16.12 0.00 0.00 3,090.18 65.56 81.95 341.47 75.84 100.17 341.47 43.00 49.00 80.00 4.00 9.00 51.00
Note: Table 2 shows firm-specific executive wealth. We distinguish three groups of executives; 188 CEOs (Panel A), 137 finance directors (Panel B) and 280 other executive directors (Panel C). Equity is expressed as a percentage of post-IPO equity. The monetary amount of equity ownership is calculated as the number of shares owned times the first-day closing market price. Options are expressed as a percentage of post-IPO shares. The monetary amount of options equals the Black-Scholes value of the option grants held by the executive director at the time of the IPO. Shares sold is also expressed as a percentage of post-IPO shares. Note that this variable is positive in case of sales and negative in case of buys. The monetary amount is evaluated at the first-day market closing price. Base salary is taken from the executives’ service agreement. In case the executive director is a part-time employee of the company, we calculate full-time equivalent base salary. Cash compensation is the sum of base salary and expected bonuses. See Note 10 for details on the computation of cash compensation. Age is the age of the executive director at the time of the IPO. Tenure refers to the number of years the executive directors has been employed by the IPO firm. All monetary amounts are expressed in 1999 pounds using the Retail Prices Index as inflation adjustment.
Managerial Incentives at the Initial Public Offering
91
own more equity and options than CEOs at large publicly traded companies. On balance, CEOs transact at the IPO by selling 1% of post-IPO shares, which adds up to £158,000, on average. Please note that buy transactions are recorded with a negative sign in determining this net transaction position of CEOs. In particular, 77 CEOs (41%) are either selling or buying shares in the IPO. This matches the 41% of CEOs that transact shares in U.S. IPOs (Baker & Gompers, 1999). If we distinguish between buying and selling of IPO shares, we observe that 56 CEOs (29.8%) are selling shares of previously owned stock and only 21 CEOs (11.2%) are buying additional shares in the firm. Conditional on selling stock in the IPO, the average CEO sells £610,000 worth of shares. Conditional on buying shares in the IPO, the average CEO buys £183,700 worth of shares. The average CEO earns an annual base salary of £83,000.9 Not surprisingly, this is far less than the base salary earned by CEOs that work for the 510 largest public firms in the U.K. Conyon and Murphy (2000) report an average base salary equal to £589,000 for the year 1997. Cash compensation is the sum of base salary and expected bonuses.10 Panel A of Table 2 shows that the mean CEO cash compensation is £99,000. The median CEO is 46 years and has been employed by the company for 6 years. Panel B of Table 2 shows that finance directors own far less equity in the IPO firm. On average, finance directors own 2.6% of post-IPO shares, which represents a monetary value of £0.5 million. The median levels are even lower. The median finance director owns 0.2% of post-IPO equity worth £30,000. Finance directors’ option holdings involve 0.6% of post-IPO shares and have a Black and Scholes value of £98,000, on average. With their small shareholdings, finance directors sell only £6,500 worth of shares in the IPO, which converts into 0.1% of post-IPO equity. In total, 51 finance directors (37.2%) are transacting at the time of the IPO. When we distinguish between sellers and buyers of shares, we find that 12 (8.8%) finance directors are selling shares in the IPO and 39 finance directors (28.5%) are buying shares in the IPO. If we condition on selling activity, the average finance director sells shares worth £239,340. Conditional on finance directors buying stock in the IPO, they buy shares worth £71,350, on average. The average base salary earned by finance director comes to £64,000 per year, whereas cash compensation equals £75,000 a year. The median finance director is 41 years old and has only been employed by the IPO firm for 2 years. Panel C of Table 2 shows that the average shareholdings of other executive directors equals 8.4% of post-IPO equity. The mean shareholdings are worth £1.6 million, while the median shareholdings amount to £460,000. The other executive directors own stock options that comprise 0.6% of post-IPO shares and are worth £83,500, on balance. The typical other executive director uses the IPO to sell 0.4% of post-IPO shares with a monetary value of £67,000. When we split the group of 101 (36.1%) other executive directors that are transacting into
92
PETER ROOSENBOOM
the subgroup of 63 (22.5%) sellers and 38 (13.6%) buyers, the following picture emerges. Conditional on selling shares in the IPO, other executive directors sell £374,000 worth of shares. Conditional on buying shares in the IPO, other executive directors buy £123,210 worth of shares. The average other executive director is paid a base salary equal to £70,000 per year and cash compensation totaling £82,000 per annum. The median other executive director is 43 years old and has been working at the IPO firm for 4 years. At this point, we observe important differences in the shareholdings of CEOs, finance directors and other executive directors. CEOs and other executive directors own larger shareholdings in the IPO firm than finance directors. This may be explained by the higher tenure among CEOs and other executive directors compared to finance directors. Finance directors are often hired in the run up to the IPO and have only two years of tenure at the median. As a consequence, finance directors have had less time to build shareholdings in the company through previous option grants. Another explanation for the differences in shareholdings has to do with founder status. A majority of CEOs is the (co-) founder of the company, whereas finance directors are hired at a later stage. More precisely, 55.3% of CEOs has (co-) founded the company against 6.6% of finance directors and 29.3% of other executive directors. When looking at the share transactions at the time of the IPO, we observe that CEOs and other executive directors are more likely to sell shares of previously owned stock in the IPO than to buy additional shares in the IPO firm. In contrast, finance directors are more likely to buy shares in the IPO firm than to sell shares. This reflects that finance directors are recent additions to the management team and have only a relatively small part of their personal wealth invested in the firm. Conversely, CEOs and other executive directors have invested a large part of their personal wealth in shares of the IPO firm. They are more interested in using the IPO as a wealth diversification opportunity by selling shares of previously owned stock. Although the reported transacting levels may seem small expressed as a percentage of total equity, they are substantial when expressed in multiples of annual salary and bonus. For example, the average CEO sells shares in the IPO worth £610,000 (this is the conditional average of 56 selling CEOs). This converts into 6.6 times his annual salary and bonus. Buying or selling shares in the IPO therefore has a substantial impact on executive wealth.
2.5. Option Portfolio Characteristics Table 3 reveals that 80 CEOs (42.6%), 86 finance directors (62.3%) and 138 other executive directors (49.3%) hold unexercised stock options at the time of the IPO.
93
Managerial Incentives at the Initial Public Offering
Table 3. Option Portfolio Characteristics. Mean Panel A: Chief executive officers (N = 80) Mean offer price-to-strike ratio 4.83 Mean time-from-grant (years) −0.37 Mean time-to-maturity (years) 6.61 Mean time-to-vest (years) 1.83 With performance criteria (%) 43.54 Performance criteria relate to Earnings per share growth (%) Profits (%) Share price (%) Unclassified (%)
16.25 7.50 3.75 16.04
Panel B: Finance directors (N = 86) Mean offer price-to-strike ratio Mean time-from-grant (years) Mean time-to-maturity (years) Mean time-to-vest (years) With performance criteria (%)
6.34 −0.37 7.13 1.92 41.67
Performance criteria relate to Earnings per share growth (%) Profits (%) Share price (%) Unclassified (%)
17.44 3.78 6.40 14.05
Panel C: Other executive directors (N = 138) Mean offer price-to-strike ratio 5.32 Mean time-from-grant (years) −0.37 Mean time-to-maturity (years) 6.93 Mean time-to-vest (years) 2.02 With performance criteria (%) 42.27 Performance criteria relate to Earnings per share growth (%) Profits (%) Share price (%) Unclassified (%)
Standard Deviation
Minimum
Q1
Median
Q3
Maximum
17.21 0.90 2.45 1.46
0.40 0.00 0.90 −1.43
1.00 −0.02 5.20 0.16
1.00 −0.05 6.90 2.89
1.75 −0.11 7.00 2.98
133.33 −5.06 10.00 3.95
20.75 0.93 2.28 1.42
0.50 0.00 0.90 −3.20
1.00 −0.02 6.10 0.89
1.00 −0.04 7.00 2.89
2.98 −0.13 9.85 2.98
135.00 −5.06 10.00 3.03
12.42 0.85 2.17 1.31
0.75 0.00 1.70 −1.38
1.00 −0.02 5.54 0.99
1.00 −0.05 6.90 2.87
2.72 −0.14 8.95 2.96
75.00 −5.06 10.00 3.95
17.39 6.16 3.62 15.10
Note: Table 3 shows option portfolio characteristics. Note we calculate summary statistics only for those executive directors that own unexercised options at the time of the IPO. Out of the 304 executive directors with unexercised option holdings, 226 (74.3%) own options from a single option grant, 61 (20%) own options from two grants and 17 (5.6%) hold options from three grants or more. We calculate the mean offer-price-to-strike ratio of the option portfolio as the simple average of the offer price-to-strike ratio of individual option grants. Offer priceto-strike ratio is the IPO firms’ offer price divided by the exercise price of an option. Similarly we determine the mean time-from-grant, mean time-to-maturity and mean time-to-vest. Time-from-grant is defined as the number of years from the option grant to the IPO date. The mean time-to-maturity is the number of years from the IPO date to the date when the option expires. The mean time-to-vest is determined as the number of years from the IPO date to the vesting date (i.e. the date at which the options become exercisable). Some options only become exercisable when performance criteria are satisfied. Performance targets may be specified in terms of earnings per share growth, accounting profits or share price. In some cases, performance targets are present, but unspecified in the prospectus.
94
PETER ROOSENBOOM
The higher incidence of options among finance directors may assist in the recruitment of these directors. The use of options as an instrument to retain and motivate key employees is more widespread than reported by Baker and Gompers (1999). Analyzing U.S. data from 1978–1987, Baker and Gompers (1999) document that only 18.3% of IPO firms grant options to the CEO. We find a frequency of stock options that is more than twice as large. This difference may be attributed to the dramatic increase in the use of stock options from the 1980s to the 1990s. Table 3 also shows the option portfolio characteristics. For each executive, we compute the mean offer price-to-strike ratio for the options in his or her portfolio.11 Evaluated at the median, stock options are granted at the offer price. No more than 7.5% of option portfolios have a mean offer price-to-strike ratio less than one (i.e. options are in-the-money). These in-the-money options tend to be granted 6 months or more in advance of the IPO. We also compute the mean time-from-grant as the number of years from the date of grant to the IPO date. We observe that most options are granted shortly before the IPO. Only 18% of all options are granted 6 months or more in advance of the IPO date. The mean time-to-maturity of the options in the executive’s portfolio is measured as the number of years from the IPO date to the date when the option expires. The mean time-to-maturity is 7 years at the median. The mean time-to-vest is determined as the number of years from the IPO date to the vesting date (i.e. the date at which the options become exercisable). At the median, it takes somewhat less than years for the options to vest. A substantial fraction of the stock options (about 42%) is subject to performance criteria, typically related to earnings per share growth. This implies that these options only become exercisable if specific performance targets are satisfied. In contrast, stock option plans in the United States rarely include performance conditions (Conyon & Murphy, 2000). Interestingly, the option portfolio characteristics are similar across CEOs (Panel A of Table 3), finance directors (Panel B) and other executive directors (Panel C).
3. MEASURING WEALTH-TO-PERFORMANCE SENSITIVITY 3.1. Valuing the Four Components of Executives’ Firm-Specific Wealth As noted before, wealth-to-performance sensitivity (WPS) is defined as the monetary change in executive wealth per £1,000 change in shareholder wealth. In this chapter, we identify 4 components of executive wealth: stock ownership, option holdings, cash compensation, and shares transacted at the IPO. Before we can compute WPS, we therefore need to value the four sources of executive wealth.
Managerial Incentives at the Initial Public Offering
95
The monetary value of post-IPO stock ownership of executives (Eex ) is evaluated at the first-day closing market price.12 We employ the Black-Scholes (1973) model to value executives’ stock options (Oex ). The model is specified as follows. √ O ex = N[P(Z) − Xe −rT (Z − T )] (1) where Z=
ln(P/X) + T(r + 2 /2) √ T
(2)
= Cumulative probability function for normal distribution. N = Number of shares covered by option grant. X = Exercise price. P = First-day closing market price. T = Time-to-maturity (in years). r = Risk-free rate. = Annualized median industry standard deviation. Note that we have not adjusted the Black-Scholes model for dividend yields, as suggested by Merton (1973). It is difficult to reliably estimate expected dividend yields for IPO firms. Additionally, young and entrepreneurial firms are less likely to pay dividends. The risk-free rate is measured as the yield on 7-year U.K. Treasury bills at the IPO date. The 7-year period corresponds to the median time-to-maturity (see Table 3). To calculate the median industry standard deviation, we define industries at the 4-digit SIC code level provided at least three industry-matched listed firms could be identified. Otherwise, we define industries at the 3-digit or 2-digit SIC code level. We calculate the annualized standard deviation of the monthly stock returns for each of the industry-matched firms. We then use the median of those standard deviations as our ex ante proxy for the IPO firm’s risk when calculating the Black-Scholes value of the option grants.13 If options are subject to performance criteria, we discount the Black-Scholes value by 20%. Although the performance conditions clearly affect the value of executive options, the 20% discount is arbitrary. However, it is widely used in the literature to discount performance-contingent pay (e.g. Conyon & Murphy, 2000). Moreover, Conyon et al. (2000) show that performance conditions in the U.K. are binding approximately 20% of the time. In calculating the value of option holdings, we consider all unexercised options held by the executive. If executives hold options from more than one option grant, we calculate the Black-Scholes value of each grant using the option specifications for that grant. We then sum the value of the separate option grants to arrive at the total value of the executive’s option holdings.14
96
PETER ROOSENBOOM
We determine the present value of cash compensation until retirement (Cex ) as a proxy for human capital. Cash compensation is defined as the sum of base salary and expected bonus (see Note 10). We discount the cash compensation until retirement by a real rate of 3%. Years-to-retirement is the highest of 65 minus the executive’s age and 3 years. The monetary value of the number of shares transacted (i.e. bought or sold) in the IPO is evaluated at the first-day closing market price. Note that a negative sign indicates that the executive director has bought shares in the IPO.
3.2. Empirical Specification of Wealth-To-Performance Sensitivity Next, we compute wealth-to-performance sensitivity (WPS). Murphy (1999) argues that WPS is the only meaningful measure of managerial incentives and the severity of the agency problem. In deciding whether or not to consume perks, the manager’s decision will depend solely on his or her percentage ownership and not on the monetary value of his or her stock ownership.15 Moreover, this measure is widely used in the empirical literature and similar to the one developed by Jensen and Murphy (1990). We compute WPS as a weighted sum of the elasticities associated with shareholdings, option holdings and cash compensation. Shareholdings and option holdings are evaluated at the first-day market closing price. It is easy to compute the elasticity of equity holdings, because stock value increases by 1% for each 1% increase in the stock price. In other words, the elasticity of shareholdings (E ) equals one. It is more complicated to determine the elasticity of option holdings (O ). Consistent with Baker and Gompers (1999), we use the partial derivative of option value with respect to first-day market closing price (the option delta). The option delta appears in Eq. (1) as (Z), but is commonly denoted as N(d1 ). Delta measures the monetary change in value of the option per £1 increase in share price. Hence we need to convert these monetary changes into percentage changes if we want to measure the option’s elasticity. The option’s elasticity is the product of delta () times the ratio of the first-day closing market price (P) to the Black-Scholes value of the call option (c). ηO =
P c
(3)
If executives hold options from more than one option grant, we compute the elasticity measure for each grant. Given that IPO firms are characterized by a lack of time-series data, we cannot determine the elasticity of cash compensation (C ). However, previous research using data for large publicly traded U.K. companies
Managerial Incentives at the Initial Public Offering
97
has shown that the elasticity of salary to shareholder wealth is about 0.1 (Conyon & Murphy, 2000). We therefore assume this level of elasticity in the analysis. Executive wealth derived from share transactions in the IPO has no sensitivity to shareholder wealth. For that reason, it is ignored when computing WPS.16 More formally, WPS is measured as expressed in Eq. (4). Recall that WPS measures the increase in executive wealth per £1,000 change in shareholder wealth. O ex E ex C ex WPS = 1000E + 1000O + 1000C E E E j
(4)
i=1
where j is the number of option grants the executive directors has received, E denotes the market value of post-IPO equity (evaluated at the first-day closing market price), Eex is the market value of stockholdings owned by the manager (evaluated at the first-day closing market price), Oex denotes the Black-Scholes value of option holdings (evaluated at the first-day closing market price) and Cex is the present value of cash compensation until retirement. Table 4 shows the decomposition of the WPS for the three groups of executives. Panel A of Table 4 shows that CEOs have an average WPS of £227 per £1,000 change in shareholder wealth. This is similar to U.S. results. Baker and Gompers (1999) find that CEOs at U.S. IPO firms increase their wealth by $212 per $1000 increase in shareholder wealth. These levels of WPS are substantial when compared to WPS reported for large public companies. For example, Conyon and Murphy (2000) report that CEOs at large U.S. companies increase their wealth by $4.18 per $1000 increase in shareholder wealth, on average. In 1997, the average CEO at the 510 largest U.K. firms increases his or her wealth by only £2.33 per £1,000 change in shareholder wealth (Conyon & Murphy, 2000). Especially due to higher stock ownership, the average WPS of CEOs is therefore more than 95 times larger in U.K. IPO firms than in large public U.K. companies. With 88.7%, equity is the largest contributor to the mean WPS, followed by a 6.1% contribution of cash compensation and 5.2% contribution of stock options. These results are somewhat different from the U.S. results of Baker and Gompers (1999). They report that 95.1% of average WPS comes from equity holdings and only 1.3% from options. We attribute this difference to their earlier sample period (1978–1987), during which options were not widely used. Panel B of Table 4 shows that the average wealth of finance directors changes by £45 per £1,000 change in shareholder wealth. Equity adds 58.8% to the mean wealth-to-performance sensitivity. Options and salary make a substantial contribution to overall incentives, namely 17.2% for options and 24% for cash compensation. Panel C of Table 4 shows that other executive directors’ wealth increases by £104 per £1,000 increase in shareholder wealth. As with CEOs,
98
PETER ROOSENBOOM
Table 4. Wealth-to-Performance Sensitivity. Percentage of Total Equity Panel A: Chief executive officers (N = 188) Equity 20.16 Options 0.70 Cash compensation 13.97 Shares transacted 1.00
Elasticity × 1,000
Contribution to Sensitivity
1,000.00 1,684.29 100.00 0.00
201.59 11.79 13.97 0.00
88.67 5.19 6.14 0.00
227.35 196.96 0.55 943.98 161.71
100.00 86.63 0.24 415.21 71.13
26.55 7.75 10.82 0.00
58.84 17.18 23.98 0.00
45.12 70.81 0.00 554.16 22.70
100.00 156.94 0.00 1,228.19 50.31
83.70 8.52 11.80 0.00
80.47 8.19 11.34 0.00
104.02 123.30 0.49 657.17 52.60
100.00 118.53 0.47 631.77 50.57
Average sensitivity Standard deviation Minimum Maximum Median sensitivity Panel B: Finance directors (N = 137) Equity 2.65 Options 0.55 Cash compensation 10.82 Shares transacted 0.08
1,000.00 1,409.91 100.00 0.00
Average sensitivity Standard deviation Minimum Maximum Median sensitivity Panel C: Other executive directors (N = 280) Equity 8.37 Options 0.56 Cash compensation 11.80 Shares transacted 0.39
1,000.00 1,521.43 100.00 0.00
Average sensitivity Standard deviation Minimum Maximum Median sensitivity
Percentage of Mean Sensitivity
Note: Wealth-to-performance sensitivity (WPS) captures the monetary change in executive wealth per £1,000 change in shareholder wealth. It is expressed by the following formula. O ex E ex C ex 1000E + 1000O + 1000C E E E i=1 j
WPS =
where E denotes first-day market capitalization of equity (i.e. the number of post-IPO shares times the firstday closing market price). Market values of shareholdings (Eex ) and shares transacted are also measured using the first-day closing market price. The Black-Scholes value of option holdings (Oex ) is calculated as described in section 3.1. The present value of cash compensation (Cex ) is calculated by discounting the cash compensation until retirement by a 3% real rate. Years-to-retirement is the highest of 65 minus the executive’s age and 3 years. The elasticity of shareholdings (E ) is equal to one. The elasticity of option holdings (O ) is equal to Black-Scholes delta times the first-day closing price divided by the Black-Scholes option value (see Eq. (3)). The elasticity of cash compensation (C ) is set equal to 0.1.
Managerial Incentives at the Initial Public Offering
99
the majority share of overall incentives (80.5%) comes from stock ownership. Options are more important than with CEOs but less important than with finance directors, and represent 8.2% of mean WPS. Cash compensation makes up the remaining 11.3% of mean WPS.
4. EMPIRICAL RESULTS 4.1. Determinants of Managerial Incentives 4.1.1. Executive-Specific Determinants In this section, we examine the cross-sectional determinants of WPS using OLS regressions. Table 5 presents the results. We start with a discussion of executive-specific determinants of managerial incentives. Older executives have incentives to choose investment projects that pay off before they retire. Because older executives tend to focus on short-term goals, it is important to provide them with high-powered incentives (Gibbons & Murphy, 1992). However, we find that director age is not a significant determinant of WPS. There is even a significant and negative relation between the wealth-to-performance sensitivity and CEO age. This result is puzzling. Older CEOs seem to have lower rather than higher WPS. One possible explanation is that older CEOs have been selling shares to other directors or investors in the pre-IPO period for consumption or diversification purposes. However, we have no data to support this conjecture. We also examine tenure is as a determinant of WPS. Executive directors often acquire firm-specific knowledge when the firm has employed them for a longer period. It is important to retain these directors in the post-IPO period through high-powered incentives. Tenure is only significantly related to the WPS of CEOs and finance directors (Panel A and B of Table 5). This may point to the importance to retain highly qualified and tenured CEOs and finance directors. CEOs are important to successfully exploit the future growth opportunities the IPO firm may have. Successful finance directors may be heavily recruited by other firms (Mian, 2001). Finance directors are important to IPO firms since they are responsible for all financial affairs of the company, including the IPO. They oversee the preparations of financial reports and serve as a point person for external communication of financial strategy. Alternatively, it is possible that more tenured directors have built larger percentage shareholdings in the company through previous option grants. Next, we investigate founder effects. We expect founder to forego diversification benefits by holding a larger amount of shares in the company. The reason is that founders derive non-pecuniary private benefits of control from their majority ownership (Denis & Denis, 1994). Accordingly, founder status is highly statistically
100
PETER ROOSENBOOM
Table 5. Determinants of Wealth-to-Performance Sensitivity.
Executive-specific determinants Age (years) Tenure (years) Founder (dummy) Chairperson (dummy) Board and ownership characteristics Board size Independent directors (%) Equity incentives per independent director Blockholder > 5% (dummy) Blockholder in board (dummy) Firm-specific determinants Log (firm size) Firm risk (%) Tangible assets (%) R&D expenditure (%) Firm age (years) Market-to-book ratio Intercept R2 adjusted F-value
Panel A: Chief Executive Officers (N = 188)
Panel B: Finance Directors (N = 137)
−2.19 (−1.68)* 4.62 (2.50)** 139.88 (6.17)*** 177.79 (5.42)***
−0.57 (−1.32) 4.90 (2.32)** 84.89 (2.46)** 27.95 (1.02)
−0.94 (−1.20) 1.28 (1.07) 65.25 (4.51)*** 171.82 (4.67)***
−19.30 (−2.66)** 130.18 (1.93)* −443.59 (−3.38)*** −85.11 (−3.19)*** −45.97 (−2.10)**
−6.29 (−1.63) 14.16 (0.25) −8.66 (−0.15) −0.17 (−0.01) 7.71 (0.40)
−3.78 (−0.98) −27.74 (−0.59) −209.35 (−2.12)** −57.69 (−3.72)*** −9.40 (–0.69)
−49.40 (−4.16)*** −88.78 (−1.43) −10.38 (−0.19) 1.11 (0.04) −0.03 (−0.08) 4.76 (2.30)** 796.83 (7.00)***
−8.27 (−0.79) 18.62 (0.60) −21.96 (−1.10) −6.42 (−0.88) −0.08 (−0.57) −0.33 (−0.60) 152.83 (1.79)*
−18.83 (−1.89)* −44.13 (−1.42) −22.80 (−0.60) −2.63 (−0.20) −0.34 (−1.01) −0.18 (−0.24) 380.82 (4.14)***
0.51 14.18***
0.16 2.74***
Panel C: Other Executive Directors (N = 280)
0.41 13.98***
Note: Table 5 shows cross-sectional regression results for 188 CEOs (Panel A), 137 finance directors (Panel B) and 280 other executive directors (Panel C). The dependent variable is the wealth-to-performance sensitivity, which measures the monetary change in executive wealth per change £1,000 in shareholder wealth. Firm size is measured as the pro-forma assets as disclosed in the prospectus. Firm risk is measured as the industry median annualized standard deviation of monthly returns for the year prior to the IPO. Industries are defined at the 4-digit SIC code level provided that 3 listed firms are available. Otherwise, we shift to a 3-digit or 2-digit SIC code level. Age refers to age of the executive director at the time of the IPO. Tenure is the number of years the executive has been employed by the IPO firm. Founder is a dummy that takes on the value one if the executive has (co-) founded the IPO firm. Chairperson is a dummy that takes on the value one if the executive director chairs the board. Board is the number of executive and non-executive directors that serve on the board. Independent directors are non-affiliated non-executive directors and expressed as a percentage of total board size (see Table 1). Equity incentives per independent director are calculated as follows. The equity incentives per independent director are measured as the sensitivity of the share and option holdings to a £1,000 change in shareholder wealth. That is, we use Eq. (4), ignoring cash compensation, to compute the equity incentives for each independent director. We then calculate the average equity incentives (WPS) per independent director employed by the particular IPO firm. The blockholder > 5% dummy indicates whether a non-management shareholder owns more than 5% of pre-IPO shares. The blockholder in board dummy indicates whether at least one non-management shareholder is represented in the board of directors. Tangible fixed assets is expressed as a percentage of pro-forma total assets. Research and development expenditure is expressed as a percentage of sales in the year prior to the IPO. Firm age is the number of years the firm has been in existence before the IPO. The market-to-book ratio is the ratio of first-day market capitalization and pro-forma book value of equity. White (1980) heteroskedastic-consistent t-statistics are within parentheses. ∗ Significant
at the 10% level. at the 5% level. ∗∗∗ significant at the 1% level. ∗∗ Significant
Managerial Incentives at the Initial Public Offering
101
significant in all regressions. If the executive has (co-) founded the firm he or she is likely to own more stock, which increases WPS. Chairperson status is important as well. If an executive chairs the board, he or she typically earns more cash compensation because the chairmanship involves additional work. In addition, the chairperson tends to own more stock in the IPO firm. Chairperson status influences WPS of CEOs and other executive directors but not the WPS of finance directors. 4.1.2. The Influence of Board and Ownership Characteristics Previous empirical studies have produced conflicting findings concerning the relationship between board characteristics and WPS. According to Ryan and Wiggins (2001) an effective board of directors may monitor the executive’s observable actions, reducing the need for costly equity-based incentives. Moreover, boards are responsible for a cost-effective use of incentive compensation. Beatty and Zajac (1994) find a negative relation between the fraction of independent directors and the use of incentive compensation in U.S. IPO firms. In contrast, other studies argue that the board of directors and the use of equity-based rewards are complementary mechanisms working together to mitigate the agency problem (Mehran, 1995; Milliron, 2000). To start, we examine the role of board size. Small boards are less subject to non-executive directors’ free riding behavior and may therefore be more vigilant in exercising their monitoring role. We observe that the size of the board of directors is negatively related to WPS of CEOs, but not to the WPS of finance directors or other executive directors. Consistent with Yermack (1996) we find that smaller board of directors tie CEO wealth more closely to shareholder wealth thereby forcing them to bear more firm-specific risk. These results suggest a complementary instead of a substituting role for small boards. Small boards together with equity-based incentives may thus address the agency problem. We also investigate the number of independent directors (expressed as a percentage of total board size). A higher fraction of independent directors on the board is expected to be associated with active board monitoring, thus reducing the need for incentive alignment. However WPS is not negatively related to the fraction of independent directors. One possible reason is that the boards of IPO firms are easily “captured” by management. These boards are often insider-dominated and CEOs of IPO firms may exert substantial influence over the appointment of “independent” board members (Baker & Gompers, 2003). Possibly, this makes the fraction of independent directors a less appropriate proxy for vigilant board monitoring. Next, we examine the incentives of independent directors. If independent directors have their incentives aligned with those of outside shareholders they
102
PETER ROOSENBOOM
may become more active monitors of top management. The equity incentives per independent director are measured as the sensitivity of the share and option holdings to a £1,000 change in shareholder wealth. That is, we use Eq. (4), ignoring cash compensation, to compute the equity incentives for each independent director. We then calculate the average equity incentives (WPS) per independent director employed by the particular IPO firm. Table 5 reports a negative relation between the equity incentives per independent director and the WPS of CEOs and other executive directors, but not for finance directors. This is consistent with a monitoring role for independent directors that own stock and/or options in the IPO firm. Core et al. (1999) also find an inverse relation between equity incentives of independent directors and CEO compensation. The presence of outside shareholders that own more than 5% of pre-IPO equity is negatively related to the WPS of CEOs and other executive directors, but not for finance directors. Outside shareholders that are represented on the board of directors have a negative impact on the WPS of CEOs, but not on the WPS of finance directors or other executive directors. We infer that large shareholder monitoring is a substitute for CEO incentives. Baker and Gompers (1999) find that venture capitalists are important outside shareholders in U.S. IPO firms. In unreported tests, we investigated whether venture capitalists play a special role in monitoring firms. However, we do not find that firms that received pre-IPO financing by venture capitalists have any different incentives than non-venture backed IPO firms. In summary, the analysis shows that the equity incentives per independent director, the presence of large outside shareholders and large shareholder board monitoring are inversely related to the WPS of CEOs. We interpret this as consistent with a trade-off between board monitoring and percentage management ownership. This suggests that the agency problem at IPO firms can be addressed either by effective board structures or by sizable percentage ownership of managers. It is important to stress that these relations are absent in case of finance directors and are less pronounced for other executive directors. Whereas boards can evaluate CEOs by looking at overall firm performance, boards may be less able to evaluate the specific contribution to performance of finance directors and other executive directors. Board size, on the other hand, is a complement to CEO incentives. Smaller boards of directors force CEOs to bear more firm-specific risk. This suggests that small boards of directors together with high-powered managerial incentives address the agency problem. The fraction of independent directors is not significantly related to WPS. Since CEOs may exert influence over the appointment of non-executive directors, we argue that the fraction of independent directors may not proxy for independent board monitoring. This contrasts with the findings of
Managerial Incentives at the Initial Public Offering
103
Beatty and Zajac (1994). They report an inverse relation between the fraction of independent directors and the use of incentive compensation in U.S. IPO firms. 4.1.3. Firm-Specific Determinants The regression models presented in Table 5 also include firm-specific determinants of managerial incentives. We use the natural logarithm of pro-forma assets as a proxy for firm size. Demsetz and Lehn (1985) argue that managers at large firms will have lower percentage levels of stock ownership because of the large monetary investment required for a larger ownership stake. Accordingly, Table 5 shows that the WPS of CEOs and finance directors is inversely related to firm size. This can be attributed to wealth constraints faced by managers at larger companies. Using a sample of U.S. IPO firms, Baker and Gompers (1999) report similar findings. Firm risk is measured by the industry median annualized standard deviation of monthly returns for the year prior to the IPO. Firm risk is expected to negatively relate to WPS. If the firm gets riskier, managers will be less willing to tie their personal wealth to shareholder wealth (Aggarwal & Samwick, 1999). We find that firm risk is not significantly related to WPS. Nonetheless, the negative sign of firm risk is consistent with a trade-off between inducing the required amount of effort from the manager and minimizing the risk he is required to bear. Tangible assets proxy for the asset structure of the IPO firm. If a large part of IPO firm’s assets is tangible, it is easier to finance with debt and avoid raising outside equity capital. This would allow managers to retain a higher level of stock ownership in the firm, and increase WPS. In contrast, we do not find a significant association between tangible assets (expressed as a percentage of pro-forma assets) and WPS. We also incorporate research and development expenditure (as percentage of sales) as a control variable. If firms spend more on investments in intangible research and development projects, they are less likely to use debt financing (Baker & Gompers, 1999). Hence, they need to raise more outside equity capital, reducing the potential stock ownership of managers and thus WPS. However, we do not find any significant relation between research and development expenditure and WPS. We include firm age to control for age differences. Young firms may be liquidity constrained and therefore reward their managers using options. Additionally, young and small firms face less binding wealth constraints, allowing managers to hold larger percentage ownership. Nonetheless, firm age is not significant in any of the regressions. Market-to-book ratios are incorporated to control for growth opportunities. Growth options usually have uncertain outcomes and impose more risk on managers. In addition, when a large part of firm value is derived from assets not yet in place, it becomes more difficult for outside shareholders to evaluate managerial actions (Baber et al., 1996; Kole, 1997; Mehran, 1995). This is expected to increase the need for managerial incentives.
104
PETER ROOSENBOOM
Panel A of Table 5 shows that the market-to-book ratio is positively related to the WPS of CEOs but not to the WPS of finance directors or other executive directors. In unreported tests, we also investigate any differences related to high-technology firms. When we incorporate a high-tech dummy in the analysis, we find that high-tech IPO firms do not have substantially higher WPS than other IPO firms.
4.2. The Role of Stock Option Grants In this section, we examine the role of stock option grants. Beatty and Zajac (1994) argue that undiversified and risk-averse executives may not be willing to accept stock options if they already own large shareholdings in the IPO firm. Moreover, boards may not grant costly stock options to executives with large shareholdings since their interests are already well aligned with those of other shareholders. Table 6 shows the empirical results of the OLS regressions. The dependent variable is the Black-Scholes value of the option holdings divided by first-day market capitalization. We define pre-IPO wealth-to-performance sensitivity (WPS) as in Eq. (4) but using pre-IPO stock ownership and excluding the option part. That is, executives’ pre-IPO WPS depends only on their pre-IPO shareholdings and cash compensation. Table 3 shows that the vast majority of stock options (82%) in the sample are granted within 6 months before the IPO date. The pre-IPO WPS therefore approximates the situation within 6 months prior to the IPO. Table 6 shows that undiversified executives receive smaller option grants, but the effect is not statistically significant for CEOs. We infer that, consistent with expectations, undiversified finance directors and other executives receive smaller option grants. It is worth noting that none of the firm-specific characteristics and other executive-specific characteristics have explanatory power. In unreported tests we also include board characteristics. We find that there is no statistically significant influence of board size and board composition on stock option grants. Stock option grants at the IPO therefore seem to be primarily driven by pre-IPO WPS. In general, this is consistent with the results of several other empirical studies. For example, Ryan and Wiggins (2001), Bryan et al. (2000) and Mehran (1995) all find an inverse relation between the use of stock options and managerial stock ownership.
4.3. Using the IPO as an Opportunity to Buy and Sell Shares As a final point, we examine the role of share transactions in the IPO. Undiversified managers may be more likely to use the IPO to diversify their shareholdings
105
Managerial Incentives at the Initial Public Offering
Table 6. Wealth Diversification and Unexercised Options. Panel A: Chief Executive Officers (N = 188)
Panel B: Finance Directors (N = 137)
Panel C: Other Executive Directors (N = 280)
Executive-specific determinants Pre-IPO wealth sensitivity/1000 Age (years) Tenure (years) Founder (dummy) Chairperson (dummy)
−0.0075 (−0.82) −0.0001 (−0.01) −0.0004 (−1.91)* −0.0018 (−0.43) 0.0029 (0.82)
−0.0267 (−2.54)** −0.0183 (−2.87)*** 0.0001 (.94) −0.0001 (−0.66) −0.0001 (−0.01) −0.0002 (−1.58) 0.0068 (1.63) 0.0023 (1.14) 0.0006 (0.23) 0.0125 (2.32)**
Firm-specific variables Log(Firm size) Firm risk (%) Tangible assets (%) R&D expenditure (%) Firm age (years) Market-to-book ratio Intercept
−0.0037 (−1.53) −0.0057 (−1.03) 0.0027 (0.68) 0.0016 (0.51) −0.0001 (−1.91)* 0.0002 (1.62) 0.0477 (1.88)*
−0.0008 (−0.65) 0.0050 (0.83) −0.0022 (−0.81) −0.0028 (−2.39)** −0.0001 (−0.67) −0.0001 (−1.76)* 0.0092 (0.68)
−0.0043 (−2.30)** −0.0003 (−0.09) −0.0066 (−2.01)** 0.0002 (0.14) 0.0001 (0.68) −0.0001 (−0.67) 0.0522 (2.55)**
0.03 1.42
0.08 3.10***
R2 adjusted F-value
0.04 1.69*
Note: Table 6 shows cross-sectional regression results for 188 CEOs (Panel A), 137 finance directors (Panel B) and 280 other executive directors (Panel C). The dependent variable is the BlackScholes value of the option holdings divided by first-day market capitalization. We define preIPO WPS as in Eq. (4) but using pre-IPO stock ownership and excluding the option part. That is, executives’ pre-IPO WPS depends only on their pre-IPO shareholdings and cash compensation. Other independent variables are defined as before. White (1980) heteroskedastic-consistent t-statistics are within parentheses. ∗ Significant at the 10% level. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
by selling shares of previously owned stock. More diversified managers may be willing to buy additional shares in the IPO to show their commitment to the IPO firm. For the period of the next 12–18 months the IPO represents the only opportunity for share transactions. Typically, managers are subject to lock-up provisions that prevent them from selling their shares in the 12–18 months after the IPO. Analyzing U.K. data, Espenlaub and Tonks (1998) report that directors’ dealing in IPO firm’s shares is generally small in the three years after the IPO. Moreover, managers cannot exercise their options until the options vest. Table 3 shows that options granted at the IPO vest after a period of three years subsequent to the IPO. Accordingly, Espenlaub and Tonks (1998) report few option-related share transactions in the three years after the IPO.
106
PETER ROOSENBOOM
Table 7. Wealth Diversification and Shares Transacted in the IPO. Panel A: Chief Executive Officers (N = 188)
Panel B: Finance Directors (N = 137)
Executive-specific determinants Pre-IPO wealth sensitivity/1000 Age (years) Tenure (years) Founder (dummy) Chairperson (dummy)
0.0887 (4.24)*** 0.0001 (0.31) −0.0004 (−1.06) −0.0048 (−1.36) −0.0026 (−0.28)
0.0213 (0.76) 0.0001 (−0.46) 0.0002 (0.74) 0.0085 (1.80)* 0.0004 (0.20)
Firm-specific determinants Log(Firm size) Firm risk (%) Tangible assets (%) R&D expenditure (%) Firm age (years) Market-to-book ratio Intercept
0.0064 (2.23)** −0.0111 (−0.73) −0.0049 (−0.47) 0.0024 (0.68) 0.0001 (1.40) −0.0001 (−0.21) −0.0641 (−2.56)**
R2 adjusted F-value
0.19 5.07***
−0.0007 (−0.52) −0.0033 (−0.69) −0.0021 (−0.61) 0.0002 (0.19) 0.0001 (0.11) 0.0002 (1.27) 0.0073 (0.44) 0.10 2.31**
Panel C: Other Executive Directors (N = 280)
0.0718 (2.98)*** 0.0001 (0.98) 0.0003 (0.95) −0.0035 (−1.53) −0.0072 (−1.72)* 0.0017 (1.71)* 0.0055 (0.70) 0.0055 (0.98) 0.0004 (0.26) 0.0001 (0.01) 0.0001 (0.96) −0.0278 (−2.56)** 0.21 7.86***
Note: Table 7 shows cross-sectional regression results for 188 CEOs (Panel A), 137 finance directors (Panel B) and 280 other executive directors (Panel C). The dependent variable is the market value of the shares transacted (i.e. bought or sold) in the IPO, divided by first-day market capitalization. Note that sales have a positive sign, whereas buys have a negative sign. We define pre-IPO WPS as in Eq. (4) but using pre-IPO stock ownership and excluding the option part. That is, executives’ pre-IPO WPS depends only on their pre-IPO shareholdings and cash compensation. Other independent variables are defined as before. White (1980) heteroskedasticconsistent t-statistics are within parentheses. ∗ Significant at the 10% level. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
Table 7 presents the results. The dependent variable is the market value of the shares transacted in the IPO, divided by first-day market capitalization. Please note that sales have a positive sign, whereas buys have a negative sign. Pre-IPO WPS is defined as before. Panels A and C show that CEOs and other executive directors sell more when their pre-IPO WPS is already high. This effect is significant at the 1% level. The results for finance directors are less clear (Panel C). One explanation could be that finance directors own only few shares and are therefore less likely to sell. This poses a more general problem to the analysis. Executive directors can only sell stock in the IPO, if they own stock pre-IPO. To
Managerial Incentives at the Initial Public Offering
107
address this problem, we re-estimate regressions using data for executive directors that own stock pre-IPO. We find similar results. Overall, our findings are consistent with the use of the IPO as a wealth diversification mechanism by some managers. Undiversified and risk-averse executives (i.e. executive directors with large pre-IPO WPS) are selling more stock in the IPO than executive directors with low pre-IPO WPS do. These more diversified executives are more likely to retain their stock or perhaps even buy additional shares in the IPO. To some extent, this result compares to the results of Ofek and Yermack (2000) and Core and Guay (1999). They find that CEOs sell shares of previously owned stock during periods in which they are granted new stock options. These effects are strongest for executives who already own many shares, whereas stock options do increase the shareholdings of managers with low ownership.
5. CONCLUSIONS At the time of the IPO, a large part of firm value depends on management’s investment decisions. It is therefore crucial to align the financial interests of managers with those of outside shareholders in a cost-effective manner. In this chapter, we study IPO management’s incentives and changes therein for a sample of 188 AIM IPO firms. The Alternative Investment Market (AIM) is a new stock market that was established in June 1995 to allow small U.K. companies to go public. The unit of analysis is the individual executive director. In total, the 188 sample firms employ 605 executive directors. We divide these executive directors into 3 groups on the basis of their job complexity. We distinguish 188 chief executive officers (CEOs), 137 finance directors (comparable to Chief Financial Officers (CFOs) in the U.S.) and 280 other executive directors. We employ wealth-toperformance sensitivity (WPS) to capture managerial incentives. WPS measures the increase in executive wealth per £1,000 increase in shareholder wealth. Firstly, we examine the cross-sectional determinants of WPS. We find that WPS is higher if the manager co-founded the firm, chairs the board of directors, and has been employed by the firm for a larger number of years. The WPS of CEOs is inversely related to board size. This suggests that smaller boards force CEOs to bear more firm-specific risk. This result is consistent with the findings of Yermack (1996) for a sample of large publicly traded U.S. companies. We find that the WPS of CEOs is negatively related to equity incentives per independent director, the presence of large shareholders that own more than 5% of pre-IPO stock and large shareholder board monitoring. We view this as evidence of a trade-off between
108
PETER ROOSENBOOM
CEO incentives and board monitoring. It is important to stress that these relations are less pronounced in case of finance directors or other executive directors. Whereas boards can evaluate CEOs by looking at overall firm performance, boards may be less able to evaluate the specific contribution to performance of finance directors or other executive directors. Growth options positively impact the WPS of CEOs. Since these firms derive a large portion of their value from assets not yet in place, they are riskier and controlling agency problems by monitoring managers’ efforts becomes difficult. CEO incentive alignment may mitigate these monitoring problems in firms that derive a large part of their value from future investments. Secondly, we investigate the role of stock options. Executive directors often own large shareholdings in the IPO firm. What is more, managers have their human capital (i.e. employment and income opportunities) invested in the IPO firm. We find that managers, other than CEOs, with large pre-IPO WPS receive smaller option grants than managers with low pre-IPO WPS. Thirdly, we report that managers with large pre-IPO WPS sell more shares in the IPO than executive directors with low pre-IPO WPS do. This allows some undiversified executives to seize the IPO as a wealth diversification opportunity and to diversify their personal wealth portfolio. This effect is absent for finance directors. In conclusion, this study contributes to the literature by investigating the role of managerial incentives at the IPO. Unlike previous studies this analysis is structured around an important corporate event – the IPO. In addition, we investigate not only CEOs but also finance directors and other executive directors. We identify a trade-off relation between board monitoring and incentives specific to CEOs. We show that the IPO may be used as a wealth diversification mechanism. We find that undiversified managers receive smaller option grants than more diversified managers. In addition, undiversified executives sell more shares in the IPO than more diversified executive directors.
NOTES 1. Also see “Directors’ Remuneration on Flotation 1999/2000,” Arthur Andersen. 2. “Other executive directors” include sales executives, technical directors, subsidiary management and managers responsible for manufacturing, etc. Unless explicitly stated otherwise, the terms “managers” or “executive directors” refer to CEOs, finance directors and other executive directors as a group. 3. OFEX is an off-exchange share matching and trading facility for shares of U.K. unquoted companies. OFEX was introduced in October 1995 as a replacement trading facility for companies that had previously been trading under Rule 4.2 of the London Stock Exchange. Rule 4.2 allowed for trading in unquoted securities by member firms until the
Managerial Incentives at the Initial Public Offering
109
end of September 1995 when the rule was changed for use in relation to trading in London Stock Exchange suspended securities only. Companies that previously traded under Rule 4.2 transferred to OFEX or AIM. Unlisted Securities Market (USM) was set up in 1980 to provide an easier route to the market for small companies. The market closed at the end of 1996 at which USM companies were given the option to move to either the Official List or AIM. 4. Companies that go public on the Official List (also called Main Market) of the London Stock Exchange need to satisfy the following admission criteria. A minimum of 25% of shares in public hands, pre-vetting of admission documents by UKLA (U.K. Listing Authority) and a minimum market capitalization. Normally, companies wanting to go public on the Official List are required to have at least a 3-year trading record. 5. Listed firms are defined as companies that are trading on AIM or the Official List of the London Stock Exchange and that did not go public in the previous 12 months. In order to obtain a sufficient number of industry-matched firms, we need to consider the Official List. Moreover, in the beginning of the sample period (June, 1995) no AIM companies are available since AIM just started operations at that time. 6. Evaluated at the median, an IPO firm is industry-matched to 8 other publicly traded firms. A total of 116 IPO firms (61.7%) are matched at the 4-digit SIC code level, 46 IPO firms (24.5%) are matched at the 3-digit SIC code level and 26 IPO firms (13.8%) are industry-matched at the 2-digit SIC code level. 7. Throughout the chapter non-management shareholders refer to shareholders, other than management, that own shares in the IPO firm. Examples of non-management shareholders include venture capitalists, industrial and commercial companies and institutional investors. 8. Option holdings include both approved and unapproved options. Inland Revenue approved option schemes cover stock options worth up to maximum of £30,000. Any gain in the value of the options once exercised is subject to capital gains tax at the time of selling the shares. Unapproved schemes involve option arrangements made in excess of £30,000. The tax treatment of unapproved options is less favorable because they are subject to income tax at the executive directors’ marginal rate when exercised rather than at the time of the ultimate sale of the shares. The majority of options (85.3%) in the sample are unapproved. There are no major differences in the personal tax regimes in the United Kingdom and the United States (Conyon & Murphy, 2000). 9. Not every executive director is employed on a full-time basis. Specifically, 4 CEOs (2.1%), 28 finance directors (20.4%) and 21 other executive directors (7.5%) are part-time employees of the company. We gather details on their part-time employment from the service agreement and calculate the full-time equivalent pay for these executives. The part-time employment of executive directors is characteristic of small start-up companies. For example, Mitsuhashi and Welbourne (1999) report that 13% of CEOs in U.S. IPO firms are employed on a part-time basis. 10. Annual bonus schemes are generally subject to performance criteria and a maximum bonus that can be awarded to the executive. We calculate the expected bonus for every director by imposing a 20% discount to the maximum bonus level, if this level disclosed in the service agreement. Although this discount of 20% is arbitrary, it is widely used in the literature to discount performance-contingent pay (e.g. Conyon & Murphy, 2000). If the maximum level for the director is not specified, but the service agreement mentions
110
PETER ROOSENBOOM
that the director qualifies to participate in the bonus plan, we add 32% to his or her base salary to arrive at the level of cash compensation. This percentage is based on the sample median of the disclosed maximum bonuses (which is equal to 40% of base salary) and the 20% discount. If a director does not qualify to receive bonuses, the cash compensation is set equal to his or her base salary. 11. Out of the 304 executive directors with unexercised option holdings, 226 (74.3%) own options from a single option grant, 61 (20%) own options from two grants and 17 (5.6%) hold options from three grants or more. 12. Following Baker and Gompers (1999), we evaluate executive wealth using first-day market closing price. To check the robustness of this approach, we calculate the correlation coefficients between the first-day market price and the market closing prices at 20, 40 and 60 trading days after the IPO. We find correlation coefficients equal to 0.96, 0.93 and 0.94, respectively. We infer that using these market prices would not materially change the conclusions. Another problem is that first-day market prices may be driven by high first-day returns (underpricing). For example, IPO firms in the sample experience an average first-day return equal to 19% (13.4% evaluated at the median). However, offer prices and first-day closing market prices are highly correlated (correlation coefficient equals 0.97). We believe that using offer prices instead of first-day market prices would not significantly change the cross-sectional regression results. 13. We checked the robustness of Black-Scholes values by using an ex-post risk measure. That is, we determined the standard deviations of daily returns for each of the IPO firms during 250 trading days after going public (excluding the initial 20 trading days). However, results of subsequent tests are qualitatively similar when using this ex-post risk measure to value stock options. 14. The use of the Black-Scholes model has several shortcomings (Conyon & Murphy, 2000; Murphy, 1999). First, the Black-Scholes formula seems to work well for short-term traded options, but less so for non-traded executive options that expire in multiple years. Second, executive options are forfeited, if the executive leaves the company before the options vest. This probability of forfeiture reduces the cost of granting the option. Third, Black-Scholes valuations are appropriate only for options held until expiration. In contrast, executive options can be exercised immediately upon vesting. Hall and Murphy (2002) demonstrate that risk-averse and undiversified executives tend to exercise early, which reduces the company’s cost of granting options. In summary, the Black-Scholes model overstates the value of options to the risk-averse executive recipient, and is, at best, a measure of the firm’s opportunity cost of granting executive options. In spite of these limitations, the Black-Scholes method is commonly used in the empirical literature (e.g. Bryan et al., 2000; Core & Guay, 1999; Ryan & Wiggins, 2001; Yermack, 1995). 15. In unreported tests, we have also performed all subsequent analyses using wealthto-performance elasticity. The wealth-to-performance elasticity measures the percentage increase in executive wealth per 1% increase in stock price. Results of these tests are roughly similar to those using WPS. 16. There are three potential differences between this measure of WPS and the one employed by Baker and Gompers (1999). Firstly, we use cash compensation to calculate the present value of future pay. Secondly, we take into account the effect of performance conditions when valuing stock options. Thirdly, our measure of incentives encompasses the entire portfolio of unexercised options.
Managerial Incentives at the Initial Public Offering
111
ACKNOWLEDGMENTS I thank Piet Moerland, Rez Kabir, Chris Veld, Uli Hege, Tjalling van der Goot and seminar participants at the Free University of Amsterdam for their valuable comments and suggestions on an earlier draft of this paper.
REFERENCES Aggarwal, R. K., & Samwick, A. A. (1999). The other side of the trade-off: The impact of risk on executive compensation. Journal of Political Economy, 107, 65–105. Baber, W. R., Janakiraman, S. N., & Kang, S. (1996). Investment opportunities and the structure of executive compensation. Journal of Accounting and Economics, 21, 297–318. Baker, M., & Gompers, P. A. (1999). Executive ownership and control in newly public firms: The role of venture capitalists. Working Paper, Harvard Business School. Baker, M., & Gompers, P. A. (2003). The determinants of board structure at the initial public offering. Journal of Law and Economics, forthcoming. Beatty, R. P., & Zajac, E. J. (1994). Managerial incentives, monitoring, and risk bearing: A study of executive compensation, ownership and board structure in initial public offerings. Administrative Science Quarterly, 39, 313–335. Black, F., & Scholes, M. (1973). The pricing of options and corporate liabilities. Journal of Political Economy, 81, 637–654. Bryan, S., Hwang, L., & Lilien, S. (2000). CEO stock-based compensation: An empirical analysis of incentive-intensity, relative mix, and economic determinants. Journal of Business, 73, 661–694. Conyon, M. J., & Murphy, K. J. (2000). The prince and the pauper? CEO pay in the U.S. and U.K. Economic Journal, 110, 640–671. Conyon, M. J., Peck, S. I., Read, L. E., & Sadler, G. V. (2000). The structure of executive compensation contracts: UK evidence. Long Range Planning, 33, 478–503. Core, J. E., & Guay, W. (1999). The use of equity grants to manage optimal incentive levels. Journal of Accounting and Economics, 28, 151–184. Core, J. E., Holthausen, R. W., & Larcker, D. F. (1999). Corporate governance, chief executive officer compensation, and firm performance. Journal of Financial Economics, 51, 371–406. Demsetz, H., & Lehn, K. (1985). The structure of corporate ownership: Causes and consequences. Journal of Political Economy, 93, 1155–1177. Denis, D. J., & Denis, D. K. (1994). Majority owner-managers and organizational efficiency. Journal of Corporate Finance, 1, 91–118. Espenlaub, S., & Tonks, I. (1998). Post-IPO directors’ sales and reissuing activity: An empirical test of IPO signalling models. Journal of Business, Finance and Accounting, 25, 1037–1079. Gibbons, R., & Murphy, K. (1992). Optimal incentive contracts in the presence of career concerns: Theory and evidence. Journal of Political Economy, 100, 468–505. Hall, B. J., & Murphy, K. J. (2002). Stock options for undiversified executives. Journal of Accounting and Economics, 33, 3–42. Jensen, M., & Meckling, W. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3, 305–360.
112
PETER ROOSENBOOM
Jensen, M., & Murphy, K. (1990). Performance pay and top management incentives. Journal of Political Economy, 98, 225–264. Kole, S. (1997). The complexity of compensation contracts. Journal of Financial Economics, 43, 79–104. Mehran, H. (1995). Executive compensation, ownership and firm performance. Journal of Financial Economics, 38, 163–184. Merton, R. (1973). Theory of rational option pricing. Bell Journal of Economics and Management Science, 4, 141–183. Meulbroek, L. K. (2001). The efficiency of equity-linked compensation: Understanding the full cost of awarding executive stock options. Financial Management, 30, 5–44. Mian, S. (2001). On the choice and replacement of chief financial officers. Journal of Financial Economics, 60, 143–175. Milliron, J. (2000). Board of director incentive alignment and the design of executive compensation contracts. Working Paper, University of Chicago. Mitsuhashi, H., & Welbourne, T. M. (1999). Chief executive officer tenure in initial public offering firms: An event history analysis of the determinants of turnover. Working Paper, Cornell University. Murphy, K. J. (1999). Executive compensation. In: O. Ashenfelter & D. Card (Eds), Handbook of Labor Economics (Vol. 3, pp. 2485–2563). North Holland, Amsterdam. Ofek, E., & Yermack, D. (2000). Taking stock: Equity-based compensation and the evolution of managerial ownership. Journal of Finance, 55, 1367–1384. Ryan, H. E., & Wiggins, R. A. (2001). The influence of firm- and manager-specific characteristics on the structure of executive compensation. Journal of Corporate Finance, 7, 101–123. Welbourne, T. M., & Andrews, A. O. (1996). Predicting the performance of initial public offerings: Should human resource management be in the equation? Academy of Management Journal, 39, 891–919. White, H. S. (1980). A heteroscedastic-consistent covariance matrix estimator and a direct test of heteroscedasticity. Econometrica, 48, 817–838. Yermack, D. (1995). Do corporations award CEO stock options effectively? Journal of Financial Economics, 39, 237–269. Yermack, D. (1996). Higher market valuation for firms with a small board of directors. Journal of Financial Economics, 40, 185–211.
THE VALUATION OF FIRMS LISTED ON THE NUOVO MERCATO: THE PEER COMPARABLES APPROACH Lucio Cassia, Stefano Paleari and Silvio Vismara ABSTRACT In this chapter we study the peer comparable approach used for the valuation of companies that went public on the Italian Nuovo Mercato. In Italy, IPO prospectuses often report the valuation methods used by investment banks. This allows us to analyze the accuracy of “real-world” valuation estimates. We show that underwriters rely on price-to-book and price-earnings multiples. The valuation estimates generated by these multiples are closest to offer prices. Conversely, when using enterprise value ratios comparable firms’ multiples are typically higher than those of the firms going public. We argue that underwriters have the possibility to select comparables that make their valuations look conservative.
1. INTRODUCTION At the time of an Initial Public Offering (IPO), firms are faced with the difficult decision of how to determine the offer price for their shares. Issuers therefore delegate the pricing decision to an investment bank that underwrites the securities issue (Baron, 1982). Today, most IPOs are priced using the book building procedure (Sherman, 2002). Under this framework, underwriters determine the final offer The Rise and Fall of Europe’s New Stock Markets Advances in Financial Economics, Volume 10, 113–129 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3732/doi:10.1016/S1569-3732(04)10005-4
113
114
LUCIO CASSIA ET AL.
price in two steps. First, they determine an initial price range for the shares using traditional valuation techniques such as the discounted cash flow (DCF) method or the comparable multiples method. Second, they collect indications of interest from institutional investors. These investors indicate how many shares they would like to buy and at what price within the price range. Investment banks use this information about investor demand to set the final offer price within or outside the initial price range. Valuation is a key issue during the first step of the bookbuilding procedure. However, we are aware of only two studies that examine the valuation of IPOs using multiples. Kim and Ritter (1999) examine the use of multiples of comparable firms to value U.S. IPOs. They consider both historical accounting numbers (i.e. book value, earnings, cash flows, and sales) and forecasted earnings, and find that price-earnings (P/E) multiples based on forecasted earnings dominate all other multiples in terms of valuation accuracy. In their study, comparable firms are selected from two sets: recent IPOs (comparable firms that went public no more than 12 months prior to the IPO firm’s offer date and that operate in the same industry) and firms chosen by a research boutique (Renaissance Capital). In another study, Purnanandam and Swaminathan (2004) investigate a sample of U.S. IPOs from 1980 to 1997. They find that the median IPO is overvalued at the offer by about 50% to its industry comparables. They report similar results using alternative sets of comparable firms selected on the basis of industry, industry and size and a combination of industry, sales and ROA. Kim and Ritter (1999) and Purnanandam and Swaminathan (2004) select comparable firms on the basis of an “algorithmic” process. The reason is that the prospectuses of U.S. IPOs do not report any information on the comparable firms’ multiples that underwriters used when setting the initial price range. It can be argued that the accuracy of valuations by investment banks differs from that of the valuations by researchers. Analysts have more information available than academics and may take into account firm-specific factors. Moreover, the selection of the set of comparable firms is up to a certain extent arbitrary and underwriters may be tempted to choose comparable firms that make the offer price look conservative (Kim & Ritter, 1999). To the best of our knowledge, no study has previously investigated the accuracy of the actual valuation process used by underwriting investment banks. In this chapter we study the valuation of companies that went public on the Nuovo Mercato in Italy using comparable firms that are selected by underwriters in the “real-world.” The peer comparables approach is the valuation methodology most frequently adopted for the pricing of firms that recently (1999–2002) went public in Italy. In this process, the choice of the multiples is critical as the median of the multiples of
The Valuation of Firms Listed on the Nuovo Mercato
115
comparable companies is used as a driver for pricing the companies going public. We examine the multiples relied on most by underwriters when pricing IPOs. We assess the valuation accuracy by comparing the IPO prices with the valuation estimates obtained for each multiple on the basis of information published in the prospectuses. We find that the prices are mainly driven by traditional multiples, such as price-earnings (P/E) and price-book value (P/BV) ratios. Multiples relative to the enterprise value do not play a significant role in the pricing process, even though these multiples are frequently cited. EV multiples for comparable firms are often higher than those of IPO firms. Thus, comparable firms reported in the prospectuses may be selected in order to make the IPO look conservative. The valuation accuracy is also tested using a regression approach with comparable firms’ multiples as independent variables (Kim & Ritter, 1999). However, the accuracy of valuations does not benefit substantially and the empirical relation between the multiples of IPO firms and their comparables does not increase using the regression approach. These results are consistent with the industry practice of using the simple multiples approach rather than the regression. This chapter continues as follows. The next section summarizes the characteristics of the peer comparable approach to valuation. Section 3 examines the valuation methods used for pricing the IPOs on the Nuovo Mercato, while Section 4 reports the results of the valuation accuracy analysis. We present our conclusions in Section 5.
2. VALUATION USING MULTIPLES Equity valuation using multiples is widely recommended by practitioner publications and valuation textbooks (e.g. Damodaran, 2002; Palepu et al., 2000). Multiples valuation methods are based on the market valuations of companies that are thought to be “comparable” to the firm that is to be priced. The basic hypothesis is that the value of the indicator/multiple for firms in the same market can be used as a driver for the valuation of a specific company, assuming that investors evaluate the results of the company in the same way in which they evaluate those of the comparable firms. For example, when using the price-earnings valuation method, the value of a company is estimated based upon how the shares of similar companies are currently priced in the market. The estimated value is found by multiplying the firm’s earnings to the average (or median) price-earnings ratio for the set of comparable firms. Thereby, the comparable method yields the risk of self-justifying an over or under valuation of the firm if the market evidences cycles of relative over or under valuation of the firms in a certain industry.
116
LUCIO CASSIA ET AL.
The methods of relative valuation have the advantage of being intuitive and simple to apply. However, they are accompanied by several problems. The first problem is that it is difficult to find two identical companies. If the established selection criteria are too strict, few comparable firms can be identified. If the selection criteria are not very strict, the risk is to obtain a sample of firms non-comparable with the one to be evaluated. The correct selection of the peer group therefore requires the solution of the trade-off between more or less strict selection criteria. Furthermore, it is probable that given a certain peer group, different valuations result depending on the type of indicator used (e.g. earnings, book value of equity). The choice of the multiples constitutes a critical decision in the valuation process. The more a performance measure refers to the “upper” part of the profit and loss account, the less it is affected by the accounting policies used in drafting the balance sheet. On the other side, the more the performance measure approaches the net profit, the better it reflects the diversity in firms’ operating efficiency. For example, the use of sales multiples does not yield any information about production, marketing and administration efficiency and interest and taxes; whereas net profit incorporates the effect of all these elements. The most popular multiple used in practice is the price-earnings (P/E) multiple. The P/E multiple is obtained dividing the market value of the comparable firm by its net profit. The latter can be this year’s profit, the previous year’s profit or next year’s expected profit. One potential problem of the P/E multiple is that it cannot be used for valuation purposes when the firm does not expect earnings in the short term. There are several studies dealing with the valuation accuracy of the multiples approach. Boatsman and Baskin (1981) study the peer comparables approach and show that the valuation process leads to better results when comparable firms are chosen in the same industry and with similar historical earnings growth, relative to when they are chosen randomly. Alford (1992) examines the valuation accuracy of earnings per share (EPS) multiples when comparable firms are selected on the basis of industry, size, earnings growth, and leverage. He finds that valuation errors decrease when the industry definition used to select comparable firms is narrowed from classification based on a single digit SIC code to two and three digits. Kaplan and Ruback (1995) compare the performance of DCF estimates to the estimates obtained from comparable firms (and comparable transactions) based valuations. With reference to a sample of 51 highly leveraged transactions, they find that estimates based on the comparable firm method underestimate the transaction value. Beatty et al. (1999) evaluate the predictive ability of a range of models and find that those based on weighted earnings and book value provided the best price estimates. Baker and Ruback (1999) use the harmonic mean estimator to calculate multiples based on EBITDA, EBIT, and sales, and report that industryadjusted EBITDA performs better than EBIT and sales. Bhojraj and Lee (2002)
117
The Valuation of Firms Listed on the Nuovo Mercato
focus on the selection of comparable firms and use of a linear regression approach that predicts a “warranted multiple” for each target company. They identify peer firms as those having the closest warranted valuation multiple.1 Liu et al. (2002) show that (forward and historical) earnings multiples provide the best estimates, followed by multiples based on cash flow measures and book value of equity. Liu et al. (2004) provide international evidence that multiples based on earnings perform best. They also show that the valuation accuracy decreased temporarily during the “bubble” of the late 1990s: it began to decline in 1997, reached a low in 2000, and improved thereafter.
3. VALUATION METHODS FOR THE IPOS ON THE NUOVO MERCATO The approach of comparable firms is the most frequently adopted valuation methodology for the pricing of 83 firms that recently (1999–2002) went public in Italy. Table 1 shows that 86.7 and 79.5% of companies use the multiples method and the DCF method, respectively. The multiples method was used more frequently in the IPOs of the Nuovo Mercato (37 IPOs out of 39). Moreover, 11 companies report to have exclusively used the multiples method for the setting of the offer price, while only 3 companies exclusively adopted the DCF technique (not reported in the table). The majority of firms (46 of which 27 on the Nuovo Mercato) choose the multiples method as principal valuation method and used the DCF method as control method; on the other hand, only 5 Nuovo Mercato companies chose
Table 1. The Use of Valuation Methods. Year
Multiples Method
DCF Method
IPO Sample
Main Market MTA
Nuovo Mercato
Main Market MTA
Nuovo Mercato
Main Market MTA
Nuovo Mercato
1999 2000 2001 2002
14 (82%) 10 (100%) 9 (69%) 2 (50%)
6 (100%) 27 (93%) 4 (100%) –
11 (65%) 8 (80%) 10 (77%) 4 (100%)
6 (100%) 23 (79%) 4 (100%) –
17 (100%) 10 (100%) 13 (100%) 4 (100%)
6 (100%) 29 (100%) 4 (100%) –
Total Sample
35 (80%) 37 (95%) 72 (86.7%)
33 (75%) 33 (85%) 66 (79.5%)
44 (100%) 39 (100%) 83 (100%)
Note: This table shows the number of firms that adopted the multiples method and/or the DCF method. Data are taken from the prospectuses of 83 non-financial companies that went public on the Main Market (MTA) or Nuovo Mercato of the Italian Exchange during 1999–2002.
118
LUCIO CASSIA ET AL.
the DCF method as their principal method and multiples for confirmation.2 This may be due to the difficulties in using the DCF technique to value Nuovo Mercato companies. At the time of the IPO, many of these firms did not report profits and had a short operating history. These findings show a significant change when compared with the procedure followed by firms that listed on the Italian main market from 1995 to 1997, where less attention was given to the multiples method (Giorgino et al., 2001). When using multiples to value firms, analysts obtain a convenient valuation without incurring extensive time and effort costs, yet lose some of the benefits of a direct valuation. Although the comparable method can reduce the probability of misvaluing a firm relative to others, it provides no safeguard against an entire sector being under- or over-valued. The peer comparables approach may therefore cause overoptimistic valuations to be self-justified. Prior studies that examine comparable firms often focus solely on P/E ratios (e.g. Alford, 1992; Boatsman & Baskin, 1981), and therefore do not consider firms with losses. This limitation is important for our sample since many high-tech firms report negative earnings. Along with the traditional ratios P/BV, P/E and P/CF, companies in our sample also used Enterprise Value multiples (i.e. EV/Sales, EV/EBIT, and EV/EBITDA). In particular, Table 2 shows that companies listed on the Nuovo Mercato often adopt Enterprise Value multiples (especially EV/Sales, 32 IPOs), while the ones listed on the MTA preferred to adopt the P/E and the P/CF multiples (30 and 28 IPOs respectively). The P/BV ratio is widely used on both the markets (30 IPOs on MTA and 35 on the Nuovo Mercato). Finally, the adoption of sales multiples (EV/Sales and P/Sales) is more common on the Nuovo Mercato. Sometimes firms are valued using non-financial multiples. For instance, valuation of the Internet-based firms referred to web traffic measures as number of users, time on the website, type of service offered and percentage of a web site’s visitors relative to the total web-surfing population (Demers & Lev, 2001; Hand, 2001; Trueman et al., 2000). Bartov et al. (2002) compared the valuation of Internet and non-Internet IPO-firms, and found that the valuation of the former group departs from conventional wisdom with earnings not being priced, and negative cash flows being viewed as investments. The analysis of the IPOs on the Italian Exchange from 1999 to 2002 points out that the number of comparable firms mentioned in the prospectus seems to be on average greater for IPO firms on the Nuovo Mercato than for IPO firms on the main market. For Nuovo Mercato firms that report the names of the peer comparables in the prospectus, there are on average 9.3 comparable companies. For the main market the average number of peer comparable firms is 5. Nevertheless, when taking into consideration only Italian comparable firms, the proportion reverses: the firms listed on the MTA on average mention 2.2 Italian comparables while
Year
EV/Sales
Main Market MTA 1999 5 2000 7 2001 6 2002 0 Total
18 (51%)
Nuovo Mercato 1999 2 2000 26 2001 4
EV/EBITDA
EV/EBIT
P/BV
P/E
P/CF
P/Sales
6 8 9 1
7 6 5 2
12 10 8 0
11 7 9 2
11 9 8 0
0 2 0 0
1 0 0 0
0 1 0 0
24 (69%)
20 (57%)
30 (86%)
30 (86%)
28 (80%)
2 (6%)
1 (3%)
1 (3%)
3 21 4
3 13 4
6 25 4
4 20 4
4 22 4
0 4 0
0 2 0
0 2 0
6 27 4
P/EBIT P/EBITDA
Total
14 10 9 2 35 (100%)
Total
32 (86%)
28 (76%)
20 (54%)
35 (95%)
28 (76%)
30 (81%)
4 (11%)
2 (5%)
2 (5%)
37 (100%)
Sample
50 (69%)
52 (72%)
40 (56%)
65 (90%)
58 (81%)
58 (81%)
6 (8%)
3 (4%)
3 (4%)
72 (100%)
The Valuation of Firms Listed on the Nuovo Mercato
Table 2. What Type of Multiples are Used to Value IPOs?
Note: This table shows the number of firms that adopt a particular multiple for the pricing of shares. The sample consists of companies that went public on the Main Market (MTA) and the Nuovo Mercato in the period 1999–2002.
119
120
LUCIO CASSIA ET AL.
Table 3. Valuation Using Multiples. EV/Sales
EV/EBITDA
EV/EBIT
P/BV
P/E
P/CF
MTA Nuovo Mercato
4.63 (2.66) 18.49 (4.10)
11.87 (8.42) 56.79 (26.48)
16.62 (13.64) 39.31 (26.13)
6.53 (4.60) 46.90 (19.65)
36.55 (19.25) 268.32 (57.33)
14.87 (12.70) 72.52 (41.56)
Sample Test on the difference
13.50 (3.59) 0.021**
36.06 (13.14) 0.013**
27.97 (18.06) 0.038**
28.26 (10.84) 0.001***
141.90 (33.57) 0.050**
43.69 (17.57) 0.005***
Note: Table shows the mean (median) values of the multiples most commonly adopted for the pricing of IPOs on the Main Market (MTA) and the Nuovo Mercato in the period 1999–2002. The multiples are calculated at the offering price. The number of observations for each multiple is reported in Table 2. The last row reports p-values of a t-test on the difference between markets. ∗∗ 5% significance level. ∗∗∗ 1% significance level.
those on the Nuovo Mercato mention only 1.1 Italian comparables. Table 3 reports the mean and median values of the most used multiples using the offer prices. Multiples are higher on the Nuovo Mercato than on MTA. This is not surprising since firms going public on this market are typically younger and riskier. Previous studies of the IPO pricing (Kim & Ritter, 1999; Purnanandam & Swaminathan, 2004) use three separate prices to compute the market value of equity in order to evaluate the accuracy of alternative valuation methods and to examine the role of accounting information in valuation. Similarly, our study considers the following values: the preliminary offer price (POP), defined as the midpoint of the range of the offer price disclosed in the prospectuses; the final offer price (OP); and the market price (MP) at the close of the first trading day. Between the filing of the prospectus and the offer date there is a so-called “waiting period,” during which the underwriter gathers information about the market demand for the issue. This additional information affects the final offer price, which may or may not be within the preliminary offer price range. Most firms of the sample (61 out of 83) are sold at a price within in the price range. In no case the offer price is higher than the range disclosed in the prospectus. On the other hand, 26.5% of the sample IPOs (10 cases on the MTA and 12 on the Nuovo Mercato) went public at a price lower than the minimum value of the range. The firms that went public during the 2000–2002 period are most likely to have an offer price below the lower bound of the price range. This roughly corresponds to the findings of Ritter and Welch (2002). They report that 25.0% of U.S. IPOs between 1995 and 1998 had a final offer price below the minimum threshold of the range, while 25.9% of the offerings had a price higher than the maximum. The percentages became respectively 18.1% and 45.1% in 1999–2000, then 25.0% and 15.0% in the year 2001. Table 4 compares the preliminary offer price, the offer price, and the market price in terms of percentage price changes. It is worthwhile to note that the average offer price is set at a lower value compared to the midpoint
121
The Valuation of Firms Listed on the Nuovo Mercato
Table 4. Price Updates and Underpricing.
MTA MTA (∗ ) Nuovo Mercato Sample Sample (∗ ) Test on the difference between markets Test on the difference from zero
Number of IPOs
OP − POP (%) IPOs
MP − POP (%) POP
44 43 39 83 82
−6.5 (−4.1) −6.9 (−4.1) 6.4 (−3.4) −0.5 (−4.0) −0.6 (−4.1)
11.7 (−4.4) −1.7 (−5.0) 33.7 (3.5) 22.0 (−3.0) 15.1 (−3.3)
p-Value
0.394
0.393
0.695
p-Value (∗ ) p-Value
0.381 0.947
0.113 0.081*
0.018** 0.007***
p-Value (∗ )
0.935
0.155
<0.001***
Underpricing MP − OP (%) OP 17.2 (0.1) 5.3 (−0.1) 22.7 (7.1) 19.8 (0.9) 13.6 (0.8)
Note: Mean (and median) values of the percentage price change: (i) from the preliminary offer price (POP) to the final offer price (OP); (ii) from the POP to the market price (MP); and (iii) from the OP to the MP. POP is defined as the midpoint of the offer price range disclosed in the prospectus, while MP is the closing price at the first trading day. The sample consists of nonfinancial companies on the Main Market (MTA) and the Nuovo Mercato in the period 1999–2002. Rows marked by (∗ ) exclude data of Finmatica that listed in 1999 on MTA. Finmatica is a high technology company that experienced excessive underpricing (+532%). The last two rows report p-values of a t-test on the difference between markets and a t-test that tests whether the average is different from zero. ∗ 10% significance level. ∗∗ 5% significance level. ∗∗∗ 1% significance level.
of the offering range (POP). However, the only significant observable change is from the offer price to the market price. Underpricing is statistically positive due to the high values of underpricing on the Nuovo Mercato.
4. ANALYSIS OF THE VALUATION ACCURACY The focus of the literature on the valuation accuracy has typically focussed on U.S. IPOs. These usually do not report information on the multiples of comparable firms adopted for the valuation process. These firms were selected as comparable on the basis of a mechanical algorithm, typically referring to industry classifications such as that proposed by Fama and French (1997) or more simply using the SIC industry codes. However, underwriters choose the comparables on the basis of a situation-specific analysis, and may be tempted to justify a high multiple. Because of this lack of information on the comparables selected by
122
LUCIO CASSIA ET AL.
underwriters, previous studies did not investigate the effective multiples used for IPO pricing. In contrast, we have access to detailed information on the valuation method used in the pricing of Nuovo Mercato IPOs. In this section we examine the valuation accuracy of multiples valuation. Following previous studies such as Kim and Ritter (1999) and Purnanandam and Swaminathan (2004), we examine how close value estimates are to IPO prices: price (IPO)i = ␥i x i,comp + i
(1)
where ␥i is the value driver (such as earnings) for IPO firm i, xi,comp is the median multiple on the driver for peer comparables of IPO firm i (such as the median price-to-earnings), and i is the pricing error. For the six most frequently used multiples, we compare the value of the multiple for each firm to that of its comparable. We argue that a specific multiple is more relevant, the more the value of the multiple for the IPO firms reflects the value of the multiple for comparable firms. For each multiple we define a multiple relevance index, expressed as the ratio between the median value of the comparables’ multiples and the value of the firm’s multiple: x i,comp (2) multiple relevance (x i ) = xi where xi,comp is the median value of the multiple x for peer comparables of IPO firm i, xi is the value of multiple x for IPO firm i. For instance, if the multiple relevance of P/E is equal to 1, the P/E of the firm going public is exactly equal to the median value of the P/E ratios of the comparable firms selected by the underwriter. If the multiple relevance is higher than 1, the multiple is higher for the median comparable firm than for the company going public. Table 5 reports selected statistics of the multiple relevance indices. P/BV and P/E are the only multiples that are not statistically different from 1. In other words, the offer price of the average IPO is in line with the median values of these multiples for the firms chosen as comparables. We conclude that underwriters rely mostly on P/BV and P/E multiples in IPO valuation. These results are robust with respect to the different stages of the IPO pricing: P/BV and P/E are the “dominant” multiples either on the basis of preliminary offer price, offer price, or market price. As expected, the multiple relevance indices calculated on the basis of the preliminary offer price are predictably closer to 1 than those of the offer price or market price. This suggests that the additional information about investor demand collected during the “waiting period” gets incorporated into the offer price. It is unlikely that the historical accounting data of the IPO firm or its comparable firms’ market multiples subsumes this incremental information (Kim & Ritter, 1999). If we consider all six multiples at once, we conclude that the mean multiple relevance index is statistically higher than 1 regardless of which
EV/Sales
EV/EBITDA
EV/EBIT
P/BV
P/E
P/CF
6 Multiples
34
28
15
23
22
22
144
Preliminary offer price (POP) Test on the difference from one
3.48 (1.67) 0.051*
1.53 (1.38) <0.001***
2.08 (1.72) 0.024**
1.40 (0.61) 0.365
1.44 (0.91) 0.346
1.38 (1.37) 0.026**
1.99 (1.36) 0.002***
Offering price (OP) Test on the difference from one
3.68 (1.82) 0.031**
1.85 (1.57) <0.001***
2.55 (1.70) 0.011**
1.71 (0.95) 0.149
1.78 (1.04) 0.270
1.78 (1.51) 0.008***
2.31 (1.48) <0.001***
Market price (MP) Test on the difference from one
3.29 (1.66) 0.065*
1.73 (1.40) 0.001***
2.34 (1.47) 0.028**
1.64 (0.90) 0.158
1.73 (0.97) 0.300
1.71 (1.53) 0.016**
2.15 (1.37) <0.001***
Number of cases
The Valuation of Firms Listed on the Nuovo Mercato
Table 5. Multiple Relevance Indices.
Note: Table shows the mean (median) values of the multiple relevance indices. This index is defined as the ratio between the median of the multiples for the comparable firms and the value of the multiple for the IPO firm. The sample consists of 39 Italian IPOs on the Nuovo Mercato during the period 1999–2002. We refer to Table 4 for the definition of POP, OP, and MP. We report p-values for a t-test that tests whether the average differs from one. ∗ 10% significance level. ∗∗ 5% significance level. ∗∗∗ 1% significance level.
123
124
LUCIO CASSIA ET AL.
of the three prices is used (1.99 for POP, 2.31 for OP, and 2.15 for MP). We argue that the selection of comparables is done in a way that makes the underwriters’ valuations look conservative. We now determine the valuation accuracy by examining the distribution of valuation errors, expressed as the natural logarithm of the estimated value relative to the transaction value (Finnerty & Emery, 2004; Kaplan & Ruback, 1995; Kim & Ritter, 1999): valuation error (x i ) = ln(x i,comp ) − ln(x i )
(3)
where xi,comp is the median value of the multiple x for peer comparables of IPO firm i, xi is the value of multiple x for IPO firm i. We find that valuation errors are smaller for the multiples based on preliminary offer prices than those on offer prices or market prices. At the level of the single multiple, we show that P/BV and P/E are the most accurate: the mean of their valuation errors is not statistically different from zero. This finding is consistent with the multiple relevance analysis (see Table 5), and it is to some extent consistent with the widespread rate of adoption of these multiples (see Table 2).3 However, as also reported by Kim and Ritter (1999), there is a significant variation in the values of multiples, especially in the case of P/E multiples. When looking at the other multiples we find that the POP valuation error of P/CF multiples is on average not different from zero, but this finding is not robust for OP and MP valuation errors. Enterprise Value (EV) multiples are on average statistically positive, and their mean and median values of valuation errors are higher than 30% regardless of the price used (POP, OP or MP). We argue that either these ratios are not really taken into account during the IPO pricing process or that the analysts select comparables that will not make the IPO look overpriced. We also report the percentage of predicted valuations within 15% of the actual multiple. These percentages are generally between 10 and 25%, and are sensitive to the price used in the calculation. If we refer to POP, EV multiples show the highest percentage within 15% (26.7% of valuation errors are less than 15% for EV/EBIT). On the other hand, when referring to the offer price, P/E and P/BV multiples are the most accurate. Finally, EV/EBIT and P/E show the highest percentage of valuation errors within 15% for market price-based multiples. This evidence is broadly consistent with Kim and Ritter (1999) although their results are insensitive to the price used.4 Finally, we investigate absolute valuation errors, measured as the predicted price5 minus the actual price, divided by the actual price (Alford, 1992):6 x i,comp − x i absolute valuation error (x i ) = (4) xi
125
The Valuation of Firms Listed on the Nuovo Mercato
where xi,comp is the median value of the multiple x for peer comparables of IPO firm i, xi is the value of multiple x for IPO firm i. The absolute error analysis confirms the accuracy of P/BV and P/E; only the absolute errors of these multiples are on average not statistically different from zero. The analysis of the valuation relevance and of the valuation errors leads to the conclusion that P/BV and P/E multiples are the most relevant multiples for the pricing of Italian IPOs in the period 1999–2002. These tests of the accuracy of multiple-based valuations compare the IPO-firms’ multiples to the multiples of their peer comparables. The implied assumption is that the price of a firm is directly proportional to the observed value driver. As an alternative to this “simple multiple approach,” a more general regression approach may be used: (5)
x i = ␣ + x i,comp + i
where xi is the value of multiple x for IPO firm i, ␣ and  are the regressions coefficients (intercept and slope coefficient), xi,comp is the median multiple for peer comparables of IPO firm i, and i is the pricing error. The regression approach expressed in Eq. (5) is a generalization of the simple approach of Eq. (1) where we forced the intercept and slope coefficients in Eq. (5) to be zero and one, respectively.
Table 6. Regression Results. R 2ADJ (%)
x i = ␣ + x i ,comp + i ␣
Valuation Error Mean

Coeff.
t-Stat.
Coeff.
t-Stat.
P/BV POP OP MP
3.38 2.74 3.15
3.21 2.43 2.79
0.581 0.631 0.593
3.02 2.49 2.73
43.0 43.7 40.4
8.3 10.4 10.3
P/E POP OP MP
32.25 20.65 24.00
3.15 2.46 2.74
0.428 0.532 0.485
3.45 3.45 3.63
21.3 40.7 33.7
37.0 22.8 23.1
Median
Percentage Within 15% KR 1999
Pure
Simple
−8.5 −10.0 −9.6
56.5 47.8 43.5
13.0 17.4 17.4
13.0 17.4 8.7
25.4 0.5 5.3
18.2 22.7 27.3
13.6 22.7 9.1
13.6 18.2 22.7
Note: Table shows the results of OLS regressions with IPO multiples as the dependent variables and using comparable firm multiples as explanatory variables. The sample consists of companies that are valued using the P/BV ratio (23 firms) and the P/E ratio (22). Valuation errors are defined as the natural logarithm of the ratio of the median comparable firms’ multiple divided by the IPO multiple. The percentage of predicted valuations within 15% is the proportion of IPOs in which the valuation error is less than or equal to 15%. The column “pure” refers to the regression approach without constraints: x i = ␣ + x i ,comp + i . The column “KR 1999” refers to regressions on data with all IPOs and comparable firms median P/BV constrained to be no greater than 10 and P/E no greater than 100 (as in Kim & Ritter, 1999). The column “simple” reports previous results without using regression predictions (see Table 6). We refer to Table 4 for the definition of POP, OP, and MP. t-Stat. for alpha (beta) refers to test of hypothesis that alpha (beta) equals zero (one).
126
LUCIO CASSIA ET AL.
The null hypothesis is that  equals one (a firm going public with comparables showing high multiples should also be capitalized at high multiples). Table 6 reports the results from the regressions using the most relevant multiples (i.e. P/BV and P/E) as the dependent variables. For both multiples, the slope coefficients are less than one; they range from 0.428 for P/E at preliminary offer prices to 0.631 for P/BV at offer prices (Kim and Ritter’s coefficients are between 0.126 and 0.275). Further, as in Kim and Ritter (1999), the highest values of  are Table 7. Valuation Errors. Valuation error (%) Number of IPOs
EV/Sales
EV/EBITDA
34
28
EV/EBIT 15
P/BV
P/E
P/CF
23
22
22
6 Multiples 144
POP Mean Median Standard deviation Interquartile range Percentage within 15% Percentage positive Mean absolute error
42.5* 51.1 136.9 95.2 14.7 79.4 247.9*
33.3*** 32.4 44.4 47.4 17.9 89.3 52.6***
53.9*** 54.2 59.5 80.7 26.7 73.3 107.7**
−16.9 −48.9 93.1 101.4 13.0 43.5 39.6
−36.0 −9.6 130.0 150.7 13.6 50.0 43.6
16.8 31.3 60.5 72.4 9.1 68.2 38.4**
16.5* 30.5 101.8 84.8 15.3 68.8 98.8***
OP Mean Median Standard deviation Interquartile range Percentage within 15% Percentage positive Mean absolute error
54.6** 59.6 132.2 96.0 8.8 79.4 268.1**
50.2*** 45.0 47.9 35.4 0.0 92.9 84.7***
70.5*** 52.9 67.5 84.7 13.3 86.7 155.0**
6.4 −5.1 92.0 100.7 17.4 47.8 71.0
−15.8 3.5 124.1 115.6 18.2 54.5 77.8
37.5** 41.1 65.7 57.5 4.5 77.3 77.9***
34.3*** 38.9 99.5 86.1 9.7 73.6 131.0***
MP Mean Median Standard deviation Interquartile range Percentage within 15% Percentage positive Mean absolute error
43.7* 50.8 126.8 114.9 17.6 73.5 229.1*
42.3*** 34.0 48.8 28.8 10.7 89.3 73.1***
57.8*** 38.2 71.0 75.4 26.7 86.7 134.0**
−0.3 −10.3 99.6 103.2 8.7 47.8 64.1
−21.5 −3.0 125.4 133.9 22.7 45.5 73.5
31.4** 42.3 67.9 58.7 18.2 77.3 70.7**
26.0*** 31.5 99.3 83.6 16.7 70.1 114.5***
Note: Table shows statistics on valuation errors for the sample of 39 companies that went public on Nuovo Mercato in the period 1999–2002. Valuation errors are defined as the natural logarithm of the ratio between the median of comparable firms multiple and the IPO multiple (Finnerty & Emery, 2004; Kaplan & Ruback, 1995; Kim & Ritter, 1999). The interquartile range is a measure of dispersion calculated as the difference between the 75th percentile and the 25th percentile. The percentage of predicted valuations within 15% is the proportion of IPOs where the valuation error is less than or equal to 15%. percentage positive is the proportion of valuation errors that is positive. The absolute valuation error is measured as the ratio of predicted price less actual price over actual price (Alford, 1992). For the definition of POP, OP, and MP refer to Table 4. A t-test on the difference from zero is carried out both for valuation errors and for absolute valuation errors. ∗ 10% significance level. ∗∗ 5% significance level. ∗∗∗ 1% significance level.
The Valuation of Firms Listed on the Nuovo Mercato
127
generally found for P/BV, and for offer prices. The empirical relationship between IPO firm multiples and comparables’ multiples is therefore tenuous; nevertheless it is higher than in other studies where the comparable firms are selected in a mechanical way (Kaplan & Ruback, 1995; Kim & Ritter, 1999). The accuracy of valuations does not benefit substantially by using the regression approach instead of the simple one. The percentage of valuations within 15% of the actual multiples when a regression approach is used (Table 6), is of the same order of magnitude of the percentages reported in Table 7, where the simple multiples approach is used.7 We conclude that IPO prices are not set mechanically on the basis of their comparables multiples (functional fixation hypothesis), but rather that analysts do take into account firm-specific factors when pricing IPOs.
5. CONCLUSIONS In this study we examine the valuation methods that underwriters use to value firms that went public on the Nuovo Mercato during 1999–2002. We document that the peer comparable approach is the most frequently used valuation method. According to this method, IPOs are priced with reference to other companies multiples. To be precise, the firm’s value is estimated by multiplying a ratio relative to a performance measure for comparable firms (e.g. comparables’ median price-to-earnings) times the firm’s performance measure (e.g. firm’s earnings). Valuation by comparables therefore relies on the assumption that the performance measure has the same “proportionality to value” for comparable firms as for the company being valued. In this way, the methodology builds in errors that the market might be making in valuing comparable firms. In the IPO market, it may generate a “run-up effect”: the rise in prices of a firm in a specific sector that listed first makes it convenient for follower companies (who list subsequently) to use comparables. We show that underwriters rely on price-to-book (P/BV) multiples and priceearnings (P/E) multiples when pricing Nuovo Mercato IPOs. These multiples generate valuation estimates that are closest to IPO prices. On the contrary, when using Enterprise Value ratios (i.e. EV/Sales, EV/EBIT, and EV/EBITDA), comparable firms multiples are typically higher than the multiples for the firms going public. We argue that either these ratios are not really taken into account during the IPO pricing, or that the analysts select comparables that will make their valuations look conservative. The arbitrariness of using comparable firm multiples for valuing IPOs gives great responsibility to investment bankers. They have an incentive to avoid IPO misvaluations and to build a reputation in valuing IPOs. If the firm is
128
LUCIO CASSIA ET AL.
undervalued, its existing shareholders do not appreciate giving up part of the offer value (“money left on the table”); if the firm is overvalued, the risk is that of compromising the success of the offer and of displeasing the investors that will be cautious in subscribing to future IPOs underwritten by the same investment banks. These reputation incentives apply as far as underwriters deal repeatedly in the IPO market (i.e. act as “repeat players”). However, the effects of these reputation-based incentives may be imperfect in markets where the number of firms going public each year is small. In these markets the underwriter’s reputation incentive might not be sufficient to reduce the information asymmetry between the firm going public and the investors.
NOTES 1. Specifically, Bhojraj and Lee estimate a series of annual cross-sectional regressions of a given valuation multiple on various explanatory variables that drive cross-sectional differences in this ratio (e.g. expected profitability, growth, and cost-of-capital as explanatory variables for enterprise-value-to-sales ratio). The estimated coefficients from previous year’s regressions are then used, in conjunction with each firm’s current year information, to generate a prediction of the firm’s current and future ratio. This prediction is referred to as a firm’s “warranted multiple” and becomes the basis for the identification of comparable firms. 2. In 11 cases, companies did not explicitly mention in their prospectuses which of the two methods was used as the principal method. 3. Table 2 shows that P/BV, P/E, and P/CF are the most cited multiples in the pricing section of prospectuses. Thus, it is predictable that these multiples are somewhat more significant than the others. However, the analysis of valuation errors is completely independent from the rate of adoption. Further, the valuation errors analysis considers only those IPOs that use the comparable company valuation method and that mention which firms are assumed as peer comparables. 4. Kim and Ritter (1999) analyse three multiples (P/E, P/BV, and P/Sales) and find that P/BV is the multiple with the highest number of valuation errors within 15%. 5. Note (see Eq. (1)) that the predicted price is estimated as the IPO firm’s value driver multiplied by the median multiple on the driver for peer comparables. 6. Alford (1992) evaluates the accuracy of the P/E for a sample of about 1,500 firms listed on the U.S. markets in 1986 and reports that the median absolute valuation error equals 24.5% when comparable firms are selected by three-digit SIC industrycodes and 29.4% when all sample firms are used as comparables (ignoring industry membership). 7. In Table 6, the percentage of valuation errors within 15% is nearly the same for the simple approach and the “pure” regression approach (i.e. without any adjustment on the multiples). On the contrary, the percentage does increase if the multiples are adjusted like in Kim and Ritter (1999): all P/BV above 10 are set equal to 10 and all P/E above 100 are set equal to 100. However, this increase in percentage of valuation errors within 15% is almost entirely due to the setting of a maximum value for the multiples.
The Valuation of Firms Listed on the Nuovo Mercato
129
REFERENCES Alford, A. (1992). The effect of the set of comparable firms on the accuracy of the price-earnings valuation method. Journal of Accounting Research, 30, 94–108. Baker, M., & Ruback, R. (1999). Estimating industry multiples. Working Paper, Harvard University. Baron, D. (1982). A model of the demand of investment banking advising and distribution services for new issues. Journal of Finance, 37, 955–976. Bartov, E., Mohanram, P., & Seethamraju, C. (2002). Valuation of internet stocks – An IPO perspective. Journal of Accounting Research, 40, 321–346. Beatty, R. P., Riffe, S. M., & Thompson, R. (1999). The method of comparables and tax court valuations of private firms: An empirical investigation. Accounting Horizons, 13, 177–199. Bhojraj, S., & Lee, C. M. C. (2002). Who is my peer? A valuation-based approach to the selection of comparable firms. Journal of Accounting Research, 40, 407–439. Boatsman, J., & Baskin, E. (1981). Asset valuation with incomplete markets. Accounting Review, 56, 38–53. Damodaran, A. (2002). Investment valuation: Tools and techniques for determining the value of any asset. New York: Wiley. Demers, E., & Lev, B. (2001). A rude awakening: Internet shakeout in 2000. Review of Accounting Studies, 6, 331–359. Fama, E. F., & French, K. R. (1997). Industry costs of equity. Journal of Financial Economics, 43, 153–194. Finnerty, J. D., & Emery, D. R. (2004). The value of corporate control and the comparable company method of valuation. Financial Management, 33, 91–99. Giorgino, M., Giudici, G., & Paleari, S. (2001). Nuove quotazioni e IPOs: l’esame alle matricole. Bancaria Editrice, Collana Banca e Mercati, 29. Hand, J. (2001). The role of economic fundamentals, web traffic, and supply and demand in the pricing of U.S. Internet stocks. European Financial Review, 5, 295–317. Kaplan, S. N., & Ruback, R. S. (1995). The valuation of cash flow forecasts: An empirical analysis. Journal of Finance, 50, 1059–1093. Kim, M., & Ritter, J. R. (1999). Valuing IPOs. Journal of Financial Economics, 53, 409–437. Liu, J., Nissim, D., & Thomas, J. (2002). Equity valuation using multiples. Journal of Accounting Research, 40, 135–172. Liu, J., Nissim, D., & Thomas, J. (2004). Price multiples based on forecasts and reported values of earnings, dividends, sales, and cash flows: An international analysis. Working Paper, Yale University. Palepu, K. G., Healy, P. M., & Bernard, V. L. (2000). Business analysis and valuation. Cincinnati: South-Western College Publishing. Purnanandam, A. K., & Swaminathan, B. (2004). Are IPOs really underpriced? Review of Financial Studies (forthcoming). Ritter, J. R., & Welch, I. (2002). A review of IPO activity, pricing, and allocation. Journal of Finance, 57, 1795–1828. Sherman, A. E. (2002). Global trends in IPO methods: Bookbuilding vs. auctions. Working Paper, Notre Dame University. Trueman, B., Wong, F., & Zhang, X. J. (2000). The eyeballs have it: Searching for the value in Internet stocks. Journal of Accounting Research, 38, 137–162.
VALUING INTERNET STOCKS AT THE INITIAL PUBLIC OFFERING Michiel Botman, Peter Roosenboom and Tjalling van der Goot ABSTRACT This chapter investigates the relevance of accounting and other information to valuing Internet IPOs during the years 1998–2000 in Europe and the United States. We show that market value is negatively related to net income in the Internet bubble period before April 1, 2000 in both European and U.S. IPO markets. This is consistent with an Internet firm’s start-up expenditures being considered as assets, not as costs. Furthermore, for the U.S. IPO market, we find that free float is value relevant during the Internet bubble. Underwriters and issuers restricted the supply of shares at the IPO. This drove up market prices as investors were keen to buy Internet IPO shares.
1. INTRODUCTION The rise and fall of Internet stocks has by now generated a substantial body of academic research. Before the decline in Internet stock prices (the Internet “shakeout”) of spring 2000, academics generally attempted to explain the relatively high valuations of Internet firms (Hand, 2001; Schultz & Zaman, 2001; Trueman et al., 2000). After the Internet shakeout, studies concentrate on the
The Rise and Fall of Europe’s New Stock Markets Advances in Financial Economics, Volume 10, 131–155 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3732/doi:10.1016/S1569-3732(04)10006-6
131
132
MICHIEL BOTMAN ET AL.
implications and the origin of this event (Demers & Lev, 2001; Johnston & Madura, 2002; Ljungqvist & Wilhelm, 2003; Ofek & Richardson, 2003). In this chapter we examine the valuations of Internet firms at the Initial Public Offering (IPO) in Europe and the United States both before and after the Internet shakeout. Most Internet companies report losses because they heavily invest in customer base creation and brand development. These investments are either immediately expensed in income statements or capitalized and amortized. Hand (2000, 2001) reports an inverse relationship between market value and earnings for U.S. Internet firms before the Internet shakeout. Losses therefore appear to enhance, not reduce market value because investors recognize that losses arise because of strategic expenditures by management, not poor performance. In this chapter, we examine whether a similar relationship can be found between market value and losses for Internet companies at the time of the IPO. Additionally, we investigate whether the negative pricing of losses persists beyond the Internet shakeout in April 2000. Our contribution to the existing literature is threefold. First, prior empirical research on the value relevance of accounting information for Internet firms has been dominated by U.S. studies (Core et al., 2003; Hand, 2000, 2001). In this study we analyze both European and U.S. Internet companies. In particular, we analyze a sample of 138 European and 292 U.S. Internet companies that went public during the years 1998–2000. The European countries included are Belgium, Finland, France, Germany, Italy, the Netherlands, Norway, Spain, Sweden, Switzerland, and the United Kingdom. We are aware of only two studies that have compared the European and U.S. IPO markets. Ritter (2003) has written a short survey on the differences between these two IPO markets and Aaij and Brounen (2002) have investigated the initial returns and long-term performance of high-tech IPOs in Europe and the United States. Second, we analyze the relevance of accounting information to valuing Internet firms at the time of the Initial Public Offering (IPO), both before and after the Internet shakeout in April 2000. The IPO is an important event for Internet companies. Schultz and Zaman (2001) find that Internet firms go public to raise money and to grab market share through takeovers and strategic alliances. Going public thus yields an important advantage to Internet firms vis-`a-vis competitors and potential entrants in the Internet sector. In addition, there are large differences in the amount and type of information available for IPO firms vs. publicly traded Internet companies. For example, reliable web traffic measures are generally unavailable for start-up Internet firms at the time of their IPO. Investors therefore need to rely on the information disclosed in the prospectus when valuing Internet IPOs. Following Bartov et al. (2002), we examine the value drivers underlying Internet IPOs from two perspectives: the offer price and the stock price at the end
Valuing Internet Stocks at the Initial Public Offering
133
of the first trading day. The underwriter and the issuing firm set the offer price. This allows us to examine the determinants of the offer price that are considered to be key by the most informed parties in the IPO market. The stock price at the end of the first trading day indicates how small investors perceive the value drivers of Internet IPO firms. Third, we extend our analysis to non-accounting information that is available at the time of the IPO. We examine free float (i.e. the percentage of shares being sold in the IPO) and the percentage of post-IPO shares held by the largest shareholder. We expect to find that free float is negatively related to market value. If fewer shares are sold at the IPO, the supply of shares is restricted. This may drive up market prices as investors rush for Internet IPO shares (Bartov et al., 2002; Hand, 2001). This effect is expected to be especially important before the Internet shakeout. We expect to find a positive relationship between market value and the percentage of post-IPO ownership of the largest shareholder. The largest owner is the party closest to the Internet firm. If the largest owner decides to retain a large fraction of the post-IPO shares, this may signal positive news to investors about the value of the Internet firm (Schultz & Zaman, 2001). The remainder of the paper is organized as follows. In Section 2 we discuss prior research. Section 3 presents descriptive statistics. Section 4 contains the methodology. Section 5 presents empirical results. Section 6 concludes the paper.
2. LITERATURE OVERVIEW The main empirical finding of papers on the valuation of U.S. Internet firms is that typical accounting information still plays a dominant role in explaining the crosssectional variation in market valuations (Core et al., 2003; Hand, 2000, 2001). Nonetheless, web traffic (e.g. number of visitors to the web site, page views) has additional explanatory power beyond that of standard accounting measures for specific types of Internet companies, such as e-tailers and content and portals (Demers & Lev, 2001; Rajgopal et al., 2003; Trueman et al., 2000). In this section, we discuss several U.S. studies on the valuation of Internet firms. We refer to Ofek and Richardson (2002) for a survey of market efficiency in the Internet sector. Hand (2000) finds that basic accounting data is value relevant in a non-linear way. Analyzing 167 pure-play Internet firms from 1997 to mid-1999, he finds that Internet firm’s market values are linear and increasing in the book value of equity, but concave and increasing (decreasing) in negative (positive) net income. This implies that the larger losses, the higher market values. Hand (2000) argues that investors seem to recognize that losses reflect strategic expenditures for customer base creation and market share, not poor performance.
134
MICHIEL BOTMAN ET AL.
In a companion study, Hand (2001) investigates the importance of supply and demand forces above economic fundamentals. Hand (2001) finds that Internet market values are correlated with proxies for demand and supply forces in the form of public float. After controlling for accounting data and web traffic, his results show that Internet firm’s market values are negatively related to public float. The negative relation between free float and market value indicates that stock prices are higher if the supply of freely traded shares in the hands of the general public is restricted. Ofek and Richardson (2003) document that high Internet stock prices in initial trading can be attributed to short sale constraints. They argue that the expiration of the lockup presents a possible loosening of this short sale constraint. Consistent with expectations Internet stocks perform particularly poorly after the lockup expires. Core et al. (2003) examine the explanatory power and stability of a regression model of market values on traditional accounting variables for a large sample of firms over the past 25 years. They investigate how equity valuation changed in the recent New Economy sub-period of 1996–1999. Overall, their results suggest that traditional explanatory variables such as earnings and book value of equity, remain applicable to firms in the New Economy period, but that there is greater variation remaining to be explained by uncorrelated omitted factors. This shows that accounting information remains important in the New Economy period. Trueman et al. (2000) are unable to find a significant positive association between bottom-line net income and the stock price in their sample of e-commerce companies and portals/communities. However, they do find a positive and significant association between gross profits and stock prices. This finding can be explained by the fact that bottom-line net income is often reduced by large non-recurring costs or costs that analysts and investors consider being investments rather than expenses. They argue that gross profit, in contrast, reflects the firm’s current operating performance and is viewed as a more stable benchmark for future profitability. Several studies have investigated the value relevance of web traffic. Hand (2001) reports that market values are reliably related to only one out of four measures of web traffic – the number of unique visitors. Web traffic only explains a small fraction of the cross-sectional variation in market values. Instead, accounting variables such as book value of equity and earnings overwhelmingly determine the market value of Internet firms. Trueman et al. (2000) also investigate web traffic measures. They find that web traffic measures (i.e. the number of visitors to the site and page views) have incremental explanatory power above accounting information such as earnings. The importance of web traffic measures declines when the components of earnings are examined instead of bottom-line net
Valuing Internet Stocks at the Initial Public Offering
135
income, suggesting that some of the value implications of web traffic measures are already captured by the components of earnings. Moreover, the significance of web traffic does not mean that accounting information is irrelevant to valuing Internet firms. Rajgopal et al. (2003) also find a positive relation between market-to-book ratios and web traffic after controlling for accounting information. They analyze a sample of 86 Internet firms from the sub-sectors ISPs, content/portals and e-commerce. They assume that web traffic is driven by managerial choices. While other studies see web traffic as an exogenous variable, they use it as an endogenous variable. They show that web traffic increases if the firm engages in strategic alliances, media visibility and marketing expenditure. The value-relevance of web traffic therefore does not come from fundamental links between traffic and revenues but possibly from future growth potential through network effects and customer relationships. In another study, Rajgopal et al. (2002) conduct an event study using a sample of 57 Internet firms that are engaged in Business-to-Business e-commerce. They report that managerial actions such as alliances, acquisition of new customers and promotions are associated with positive abnormal returns. This explains a large part of the cross-sectional variation in the post-IPO stock returns beyond that explained by accounting earnings. This is consistent with investors not only looking at accounting earnings when setting the prices of Internet stocks but also to managerial actions. While the aforementioned studies examine possible value drivers before the Internet shakeout, Demers and Lev (2001) compare the role of various types of information in explaining stock prices before and after the Internet shakeout in the spring of 2000. They study a sample of 84 Internet companies over the period February 1999 to May 2000. In 1999 marketing expenditures and product and development costs are positively related to price-to-sales ratios, implying that the market sees marketing and R&D costs as intangible assets instead of current expenditures. In 2000, however, this relation disappears suggesting that the stock market is no longer willing to capitalize extraordinary expenditure as intangible assets. Demers and Lev (2001) also investigate the importance of web metrics (the attraction of new visitors to the site, retention of visitors at the site and the ability to generate repeat visits). Contrary to recent doubts about the relevance of web traffic caused by the Internet shakeout, they report that web metrics are still relevant to valuing Internet firms in 2000. Bartov et al. (2002) examine the valuation of 98 U.S. Internet IPOs from 1996 to June 1999. They investigate the difference between the underwriter-determined final offer price and the market-determined stock price at the end of the first trading day. They find that investors infer a higher firm value from positive cash flows, sales growth, R&D expenditures, high-risk warnings and relative
136
MICHIEL BOTMAN ET AL.
offering size than do underwriters. In the next section, we describe our sample of Internet IPOs.
3. DATA AND DESCRIPTIVE STATISTICS 3.1. Sample Selection Following Hand (2000) we define an Internet company as a company which obtains the majority of its revenues (>50%) through or because of the Internet. We use three sources to collect an initial sample of Internet companies that went public on a European stock exchange between January 1, 1998 and December 31, 2000. Internet companies are selected from: (i) all traded companies on European stock markets for young, high growth companies in Belgium (Euro.NM Belgium), France (Nouveau March´e), Germany (Neuer Markt), Italy (Nuovo Mercato) and the Netherlands (NMAX), and from the London Stock Exchange (Alternative Investment Market and Official List) and EASDAQ (now NASDAQ Europe); (ii) all members of the Bloomberg European Internet Index; and (iii) through several interviews with merchant bankers from Kempen & Co. We used two sources to identify U.S. Internet firms that went public on NASDAQ National Market during the years 1998–2000. Our first source is a list of 527 Internet-related offerings used by Loughran and Ritter (2003), who obtained their data by merging and amending Internet identifications of Securities Data Corporation (SDC), Dealogic and IPOmonitor.com. Next, we matched this list against the firms marked as Internet-related by our second source, www.edgar-online.com. We include Internet companies documented by both sources in our initial sample, which contains 382 firms. In order to be included in the final sample, firms have to meet several additional criteria. First, the final prospectus must be available and include annual accounts covering a full year. Second, unit offerings are excluded from the sample. Third, financial institutions and mutual funds are excluded from the initial sample because of their less comparable financial statement data. After applying these criteria, we are left with a sample of 138 European Internet companies. Table 1 shows that 11 European countries are represented in the sample. Three U.S. firms with a listing on EASDAQ are included as well. A large portion of our European sample is listed on the German Neuer Markt. From our initial U.S. sample, 13 firms are excluded because they were issued at an exchange other than the NASDAQ National Market. After applying the sample criteria we are left with 328 U.S. Internet firms. For 36 companies we are
Valuing Internet Stocks at the Initial Public Offering
137
Table 1. Geographical Distribution of Internet IPOs. Country
N
Percentage
Belgium Finland France Germany Italy Netherlands Norway Spain Sweden Switzerland United Kingdom United States EASDAQ
2 3 27 64 9 1 1 1 5 3 19 292 3
0.5 0.7 6.3 14.9 2.1 0.2 0.2 0.2 1.2 0.7 4.4 67.9 0.7
Total
430
100.0
unable to obtain the first-day closing market price from SDC or Datastream. Our final sample thus consists of 292 U.S. Internet IPOs. 3.2. Sub-Sectors of the Internet Industry We divide the Internet industry in 6 sub-sectors: ISPs, content/portals, e-commerce, IT-infrastructure, Internet software and Internet services (see Fig. 1). Table 2 shows
Fig. 1. Sub-Sectors of the Internet Industry.
138
MICHIEL BOTMAN ET AL.
Table 2. Distribution of Internet IPOs Over Sub-Sectors. Sub-Sector Internet service provider Content/portals E-commerce IT-infrastructure Internet software Internet services Total
Europe
United States
t-Test for Difference
15 (10.9%) 19 (13.8%) 17 (12.3%) 4 (2.9%) 39 (28.3%) 44 (31.9%)
27 (9.2%) 23 (7.9%) 76 (26.0%) 33 (11.3%) 94 (32.2%) 39 (13.4%)
0.528 1.925* 3.255*** 2.922*** 0.822 4.650***
138 (100%)
292 (100%)
Note: This table shows the distribution of Internet IPOs over sub-sectors. We use a t-test to test whether this distribution differs between Europe and the United States. ∗ Significance at the 10% level. ∗∗∗ Significance at the 1% level.
the distribution of the companies in the sample over these sub-sectors of the Internet industry.1 We briefly discuss each sub-sector. Internet Service Providers (ISPs) are network service providers supplying IP-based connections to its network and access to the public Internet to consumers and/or business users. Our sample contains 15 European and 27 U.S. Internet IPOs from the ISP sector. Content companies offer free information to users about products and services through their web sites. Portals are major starting sites for users when they get connected to the Internet or that users tend to visit as an anchor site. A portal is essentially a content website because it offers free information to navigate the Internet. Because the sources of revenue are the same for these two sub-sectors (primarily advertising revenues), we take content and portals as one sub-sector of Internet industry. There are 19 European and 23 U.S. IPO firms that operate in the content/portals sector. E-commerce is the activity where vendors of goods and services and a buyer of such goods and services enter into a commercial transaction over a digital infrastructure. Our sample contains 17 European and 76 U.S. companies from the e-commerce sector. Together with ISPs and content/portals firms, e-commerce firms are web-based companies that are expected to earn revenues directly or indirectly by attracting web traffic to their sites. IT-infrastructure companies provide the physical hardware used to interconnect computer and users. Infrastructure includes the transmission media, including telephone lines, cable television lines, and satellite and antennas, and also the router, aggregator, repeater, and other devices that control transmission paths. The sample consists of 4 European and 33 U.S. IT-infrastructure companies. Internet software companies provide the various kinds of programs used to operate Internet-related computers and devices. Our sample is composed of 39
Valuing Internet Stocks at the Initial Public Offering
139
European and 94 U.S. Internet software companies. Companies in the Internet services market can be divided in companies that provide product support and maintenance for software and hardware, and companies that provide professional services. Professional services include consulting, development and integration, education and training, management services, and business management services. Our sample contains 44 European and 39 U.S. Internet services firms. We also perform a standard t-test to test the significance of differences in sub-sector distribution between Europe and the United States. We find that a significantly higher fraction of European Internet IPO firms are active in the content/portals and Internet services sub-sectors while a significantly higher fraction of U.S. Internet IPO firms operate in the e-commerce and IT-infrastructure sub-sectors. 3.3. Summary Statistics Figure 2 shows the number of IPOs per quarter during the period January 1, 1998 until December 31, 2000. We observe that Internet IPOs are not equally distributed over time. In the first two quarters of the year 2000 the number of firms going
Fig. 2. Number of Internet IPOs During 1998–2000.
140
MICHIEL BOTMAN ET AL.
Table 3. Summary Statistics of European Internet IPOs.
Offer value ($ million) Market value ($ million) Sales ($ million) EBIT ($ million) Net income ($ million) Total assets ($ million) Book value of equity ($ million) Number of employees Free float (%) Ownership of largest shareholder (%) First-day return (%)
Mean
Median
Standard Deviation
Minimum
Maximum
N
889.166 1,241.757
204.803 243.737
3,376.763 4,672.922
8.613 11.261
30,190.995 41,921.937
138 138
36.940 −2.558 −3.325 42.694 16.172
5.652 −1.076 −1.311 9.854 2.394
132.157 13.438 10.261 133.926 45.889
0.000 −109.995 −92.201 0.268 −56.997
931.890 59.668 25.022 1,113.958 298.279
138 138 138 138 138
140.907 24.877 39.639
60.000 23.592 33.875
411.012 7.251 22.407
4.000 8.204 10.000
4,512 49.242 92.000
138 133 118
51.606
14.099
95.689
−84.894
444.444
138
Note: This table shows summary statistics for the total sample of 138 European Internet IPOs. Offer value is determined as the number of post-IPO shares outstanding times the final offer price. Market value is computed as the number of post-IPO shares outstanding times the market closing price on the first day of trading. Sales, earnings before interest and taxes (EBIT), net income and the number of employees are for the last fiscal year before the IPO. Total assets and the book value of equity are taken from the balance sheet for the fiscal year prior to the IPO. Free float is defined as the number of shares sold in the IPO divided by the number of post-IPO shares outstanding. Ownership of the largest shareholder equals the number of shares retained by the largest owner scaled by the number of shares outstanding after the IPO. The first-day return (underpricing) is calculated as the percentage difference between the first day closing market price and the offer price.
public in Europe is higher than in other quarters. In the United States the number of Internet firms going public reached its high in the third quarter of 1999. After the Internet shakeout in April 2000, the number of Internet IPOs in both Europe and the United States declined rapidly in the second half of 2000. The major part of the accounting data is obtained through Bloomberg for Europe and IPO prospectuses available at www.sec.gov for the United States. Missing data are obtained through Datastream or are hand-collected from annual reports and prospectuses. We convert all European monetary data into U.S. dollars using exchange rates from Datastream. Tables 3 and 4 present descriptive statistics of the variables used in this study for Europe and the United States, respectively. The first row of Tables 3 and 4 shows that the offer value, measured as the number of post-IPO shares times the final offer price, equals $889.2 million in Europe and $559.9 million in the United States. However, the difference is not
141
Valuing Internet Stocks at the Initial Public Offering
Table 4. Summary Statistics of U.S. Internet IPOs.
Offer value ($ million) Market value ($ million) Sales ($ million) EBIT ($ million) Net income ($ million) Total assets ($ million) Book value of equity ($ million) Number of employees Free float (%) Ownership of largest shareholder (%) First-day return (%)
Mean
Median
Standard Deviation
Minimum
Maximum
N
559.871 859.990 19.370 −0.379 −14.600 35.859 17.161
360.236 333.809 6.955 0.018 −8.445 14.079 5.533
800.048 1,506.638 60.092 3.217 40.769 144.532 99.412
68.107 3.000 0.000 −43.263 −645.397 0.094 −30.291
10,713.043 14,416.290 706.466 12.202 9.235 2,343.132 1,668.106
292 292 292 292 292 292 292
234.599 19.177 28.787
154.000 18.048 22.210
297.706 7.301 17.884
13.000 5.352 5.980
3,557.000 48.786 86.400
292 292 292
78.576
57.321
114.176
−94.318
525.000
292
Note: This table shows summary statistics for the total sample of 292 U.S. Internet IPOs. Offer value is determined as the number of post-IPO shares outstanding times the final offer price. Market value is computed as the number of post-IPO shares outstanding times the market closing price on the first day of trading. Sales, earnings before interest and taxes (EBIT), net income and the number of employees are for the last fiscal year before the IPO. Total assets and the book value of equity are taken from the balance sheet for the fiscal year prior to the IPO. Free float is defined as the number of shares sold in the IPO divided by the number of post-IPO shares outstanding. Ownership of the largest shareholder equals the number of shares retained by the largest owner scaled by the number of shares outstanding after the IPO. The first-day return (underpricing) is calculated as the percentage difference between the first day closing market price and the offer price.
statistically significant using a standard t-test ( p-value = 0.12). The median offer value is much lower at $204.8 million in Europe and $360.2 million in the United States. We conduct a Wilcoxon/Mann-Whitney test to evaluate whether these median values are statistically different (not tabulated). We find that evaluated at the median, the offer values are significantly larger in the United States than in Europe ( p-value < 0.01). The average (median) market value equals $1,241.8 million ($243.7 million) in Europe and $860 million ($333.8 million) in the United States. However, these differences in the average ( p-value = 0.21) and median market value ( p-value = 0.11) between Europe and the U.S. are not statistically significant. The average (median) sales in the last fiscal year before the IPO equal $36.9 ($5.7 million) in Europe and $19.4 million ($7 million) in the United States. This indicates that market value and offer values are high when measured against current revenues. The difference in sales is significant ( p-value = 0.06) when
142
MICHIEL BOTMAN ET AL.
evaluated at the mean but not when evaluated at the median (p-value = 0.63). The average (median) earnings before interest and taxes equals –$2.6 million (–$1.1 million) in Europe and –$0.4 million ($0 million) in the United States. The differences in the average ( p-value < 0.01) and median earnings before interest and taxes ( p-value < 0.01) are statistically significant. We also look at net income before the IPO. The average (median) net income in the year before the IPO is −$3.3 million (−$1.3 million) in Europe and −$14.6 million (−$8.4 million) in the United States. These differences are highly significant, both at the mean ( p-value < 0.01) and median ( p-value < 0.01). The majority of European Internet firms (67.4%) and U.S. Internet IPO firms (93.2%) report negative net income in the fiscal year before the IPO. This difference between Europe and the United States in the fraction of loss reporting Internet companies is highly significant ( p-value < 0.01). Total assets in the year before the IPO average $42.7 million in Europe and $35.9 million in the United States. Total assets equal $9.9 million in Europe and $14 million in the United States, evaluated at the median. The difference in average total assets is not significant ( p-value = 0.64). However, the difference in median total assets between Europe and the U.S. is statistically significant ( p-value = 0.02). The average (median) book value of equity is $16.2 million ($2.4 million) in Europe and $17.2 million ($5.5 million) in the United States. These differences are not significant at the mean ( p-value = 0.91) or at the median ( p-value = 0.20). The average (median) company employs 141 (60) people in the year before the IPO in Europe and 235 (154) people in the United States. These differences are statistically significant at both the mean ( p-value < 0.01) and the median (p-value < 0.01). Tables 3 and 4 show that the free float, defined as the number of shares sold at the IPO scaled by the number of post-IPO shares, is 24.9% with a median of 23.6% in Europe, and 19.2% with a median of 18% in the United States. Again, these differences are significant both at the mean ( p-value < 0.01) and the median ( p-value < 0.01). The average (median) ownership by the largest shareholder equals 39.6% (33.9%) of post-IPO shares in Europe and 28.8% (22.2%) of post-IPO shares in the United States. This difference in the average ( p-value < 0.01) and median ( p-value < 0.01) ownership by the largest shareholder is highly significant. We also calculate first-day returns. The first-day closing market price is typically higher than the offer price, so that first-day returns are positive. The positive first-day offer-to-close return is commonly known as underpricing. The average (median) first-day return equals 51.6% (14.1%) in Europe and 78.6% (57.3%) in the United States. This difference in the average ( p-value < 0.01) and median ( p-value < 0.01) first-day returns between Europe and the U.S. is statistically significant. There are 24 European Internet IPOs (17.4%) that
Valuing Internet Stocks at the Initial Public Offering
143
double in price on the first trading day (i.e. have a first-day return of 100% or higher) vs. 100 U.S. Internet IPOs (34.2%). This difference is highly significant ( p-value < 0.01).
4. METHODOLOGY We build our empirical tests on the well-known residual-income or Ohlson (1995) model. This model states that the market value of equity is a function of the book value of equity and residual earnings: P t = BVEt +
∞ E(REt+i ) i=1
(1 + r)t
(1)
where Pt is the firm’s market value at the end of the current period t, BVEt is the book value of equity at that time, and REt+i is the firm’s residual earnings for period t + i (defined as the period’s earnings available to shareholders less a charge applied to beginning-of-period book value), r is the firm’s required rate of return on equity capital, and E(·) is the expectation operator. Following Trueman et al. (2000), we take an empirical application of the Ohlson model as the base for empirical tests. In particular, we regress market values on an intercept, book value of equity, current earnings and other explanatory variables. We use offer values as an alternative dependent variable. Table 5 provides the Table 5. Variable Definitions. Variable Name
Definition
Dependent variables MVE Market value of equity; number of post-IPO shares times the market price on the first day of trading OVE Offer value of equity; number of post-IPO shares times the final offer price Independent variables EBIT Earnings before interest and taxes; last fiscal year before IPO NI Net income; last fiscal year before IPO ISP Internet Service Provider; indicator variable that takes on value of one if IPO firm is classified as Internet Service Provider REV Sales; last fiscal year before IPO BVE Book value of equity; last fiscal year before IPO FLOAT % Free float; number of shares sold at the IPO scaled by the number of post-IPO shares TOPSH % Ownership of largest shareholder; number of shares held by largest owner scaled by number of post-IPO shares
144
MICHIEL BOTMAN ET AL.
definitions of dependent and independent variables that we use in regression analyses. It is important to note that we do not deflate market values and offer values by book value of equity or sales. Other studies use market-to-book ratios (Core et al., 2003; Trueman et al., 2000) or price-to-sales ratios (Demers & Lev, 2001) as a dependent variable. We do not adopt this approach because market-to-book ratios and price-to-sales ratios do not have the same economic interpretation as they have in a cross-section of more established and profitable firms. Book values and sales tend to be small or absent for Internet firms. These ratios therefore tend to “blow up” because of a small denominator problem. For example the market-to-book ratio in our sample would range from −27,100 to 507,820. We use log-linear regression methodology advocated by Hand (2000, 2001). There are three important advantages to this methodology. First, Hand (2000) shows that the log-linear regressions yield lower pricing errors for Internet stocks than do regressions using per-share or non-logged data. Second, log-linear regressions moderate the impact of anomalous or outlier observations in accounting data. Tables 3 and 4 show that the accounting data is highly skewed. Using log-transformed accounting data may thus help to resolve this problem. Third, Hand (2001) argues that log-linear regressions generally achieve greater homoscedasticity in regression residuals. We log transform each dependent and independent accounting variable Z (where Z is defined in $ millions) in the following way: LZ = log [Z + 1] if Z >= 0,
and
− log [−Z + 1] if Z < 0
(2)
We add $1 million to Z so that LZ is defined when Z is at or nearly zero. The log-linear regression model may incorporate concavity, linearity or convexity between the dependent and independent variables. Consider the relation between the log-transformed non-negative values of X and Y: log(Y + 1) = ␣ +  log(X + 1) ⇔ LY = ␣ + LX
(3)
This implies that the non-logged and unscaled relation between X and Y is given by: Y = e ␣ (X + 1) − 1
(4)
The coefficient  measures the degree and type of non-linearity between X and Y. We refer to Hand (2000, 2001) for a more detailed discussion of the non-linear regression model. For non-negative values of X, the relation between X and Y in Eq. (4) is concave if 0 <  < 1, linear if  = 1 and convex if  > 1. When X is negative and log-transformed using Eq. (2), the relation between X and Y is concave if −1 <  < 0, linear if  = −1, and convex if  < −1. If  = 0 then X and Y are unrelated regardless of the sign of X. If log (Y + 1) is a function of several
Valuing Internet Stocks at the Initial Public Offering
145
independent variables, suppose X and W, then  captures the marginal concavity, linearity or convexity of X (i.e. holding constant W). The next section describes our log-linear regression results.
5. EMPIRICAL RESULTS 5.1. Log-Linear Regressions for Market Values Table 6 presents the log-linear regression results using logged market value as the dependent variable for our European sample. Table 7 shows the results for the U.S. sample. Market value is defined as the number of post-IPO shares times the market closing price on the first day of trading. We assume that market values capture the value assessment of small investors. We put the letter “L” in front of the variable name to indicate that the variable has been log-transformed using Eq. (2). In the first model we regress logged market values on logged earnings before interest and taxes (LEBIT), an indicator variable for Internet Service Providers (ISPs), logged sales and logged book value of equity. We include an indicator variable for ISPs in the regression model because these firms have large market values by comparison.2 We use sales as a proxy for marketplace acceptance and market share. For young, fast growing firms it could be more important to aim for revenues than profits (Bowen et al., 2002). Book value of equity directly comes from the residual income model that we discussed in the previous section. Moreover, book value of equity is a standard control variable in valuation regressions (Collins et al., 1999). The first column of Tables 6 and 7 show the results of model (1). There is a significant and negative relation between logged net income (NI) and logged market values. Investors therefore value losses at the IPO in both European and U.S. IPO markets. The larger the loss, the higher the market value. The coefficient on the indicator variable for ISPs is positive and significant. This reflects that ISPs have higher market values in comparison to other Internet companies. The indicator variable for ISPs is significantly positive in all subsequent regressions. Logged sales are positively associated with logged market values for the European sample. We argue that firms with higher sales have more marketplace acceptance and market share, which is valued by investors. However, this result is not robust to alternative specifications of the regression model, as we will discuss next. The log of the book value of equity is not related to logged market value in Europe, but is positively related to logged market value in the United States. Next, we distinguish between profit and loss reporting Internet firms (see model (2)). We include a variable LNI POS that equals the logged net income
146
Table 6. Log-Linear Regression Results for Market Values of European Internet IPOs. (1) LNI LNI POS LNI NEG LEBIT LEBIT POS LEBIT NEG FLOAT % TOPSH % ISP LREV LBVE CONSTANT R2 adjusted (%) F-test N
−0.386 (−5.307)
(2)
(3)
(4)
***
(5)
(6) Bubble
(7) Post-Bubble
0.489 (0.954) −0.594 (−3.503)***
−0.265 (−0.778) −0.430 (−1.582)
***
−0.363 (−4.171) −0.131 (−0.542) −0.471 (−4.655)*** −0.348 (−4.510)*** −0.090 (−0.411) −0.468 (−4.613)***
0.701 (4.259)*** 0.195 (1.995)** 0.051 (0.917) 2.047 (19.080)***
0.718 (4.331)*** 0.154 (1.309) 0.034 (0.573) 2.032 (19.573)***
0.731 (4.303)*** 0.256 (2.508)** 0.048 (0.841) 2.023 (18.563)***
0.754 (4.428)*** 0.180 (1.427) 0.018 (0.317) 2.018 (18.386)***
35.952 20.226*** 138
36.089 16.472*** 138
35.802 20.101*** 138
36.193 16.542*** 138
−0.838 (−1.002) 0.214 (1.024) 0.686 (4.148)*** 0.121 (1.016) 0.087 (1.172) 2.230 (7.402)*** 39.919 13.513*** 114
−1.502 (−1.535) 0.041 (0.137) 0.543 (2.409)** −0.105 (−0.571) 0.114 (1.306) 2.597 (7.020)*** 36.915 7.019*** 73
0.767 (0.652) 0.269 (1.056) 0.911 (4.025)*** 0.249 (1.662) 0.061 (0.431) 1.464 (3.385)*** 61.802 10.245*** 41
MICHIEL BOTMAN ET AL.
Note: The table shows the log-linear regression results using logged market values (LMVE) as the dependent variable. Independent variables are defined as in Table 5. We put the letter “L” in front of the variable name to indicate that the variable has been log-transformed using Eq. (2). LNI POS equals LNI if net income is positive, 0 otherwise. LNI NEG equals LNI if net income is negative, 0 otherwise. LEBIT POS and LEBIT NEG are defined similarly but then using earnings before interest and taxes. We define the bubble period as the period from January 1, 1999 to March 31, 2000. The post-bubble period is from April 1, 2000 to December 31, 2000. White (1980) heteroskedastic-consistent t-statistics are within parentheses. ∗∗ Significance at the 5% level. ∗∗∗ Significance at the 1% level.
(1) LNI LNI POS LNI NEG LEBIT LEBIT POS LEBIT NEG FLOAT % TOPSH % ISP LREV LBVE CONSTANT R2 adjusted (%) F-test N
(2)
(3)
(4)
−0.227 (−2.770)***
(5)
(6) Bubble
(7) Post-Bubble
−0.119 (−1.579) 0.577 (1.990)** −0.418 (−3.608)***
0.247 (0.823) −0.210 (−1.72)*
0.741 (1.176) −0.437 (−1.123)
−3.463 (−5.873)*** −0.526 (−2.239)** 0.247 (1.915)* −0.010 (−0.109) 0.046 (0.779) 3.090 (15.075)***
−1.661 (−0.886) 0.315 (0.281) 0.504 (1.806)* −0.377 (−1.500) 0.237 (1.712)* 2.194 (4.363)***
22.581 11.750*** 259
20.031 2.145* 33
−0.027 (−0.168) 0.537 (1.870)* −0.277 (−1.325)
0.237 (2.036)** 0.050 (0.662) 0.173 (3.385)*** 2.115 (22.637)***
0.210 (1.772)* −0.030 (−0.359) 0.146 (2.735)*** 2.004 (19.470)***
0.300 (2.423)** 0.052 (0.621) 0.216 (4.488)*** 2.284 (27.583)***
0.270 (2.133)** 0.008 (0.091) 0.209 (4.283)*** 2.274 (27.546)***
−3.291 (−5.993)*** −0.451 (−1.927)* 0.328 (2.798)*** 0.003 (0.035) 0.086 (1.588) 3.063 (17.064)***
10.042 9.121*** 292
11.744 8.745*** 292
7.127 6.582*** 292
7.955 6.030*** 292
21.100 13.970** 292
Valuing Internet Stocks at the Initial Public Offering
Table 7. Log-Linear Regression Results for Market Values of U.S. Internet IPOs.
Note: The table shows the log-linear regression results using logged market values (LMVE) as the dependent variable. Independent variables are defined as in Table 5. We put the letter “L” in front of the variable name to indicate that the variable has been log-transformed using Eq. (2). LNI POS equals LNI if net income is positive, 0 otherwise. LNI NEG equals LNI if net income is negative, 0 otherwise. LEBIT POS and LEBIT NEG are defined similarly but then using earnings before interest and taxes. We define the bubble period as the period from January 1, 1999 to March 31, 2000. The post-bubble period is from April 1, 2000 to December 31, 2000. White (1980) heteroskedastic-consistent t-statistics are within parentheses. ∗ Significance at the 10% level. ∗∗ Significance at the 5% level. ∗∗∗ Significance at the 1% level.
147
148
MICHIEL BOTMAN ET AL.
if it is positive, and a variable LNI NEG that equals the logged net income if it is negative. We find a positive relation between LNI POS and logged market values in the United States, but not for Europe. We also find a strongly negative and thereby concave relation between LNI NEG and logged market values for both Europe and the United States. This reinforces that investors are negatively pricing losses. The larger the loss, the higher the market value. Investors appear to recognize that losses reflect strategic expenditures for customer base creation and market share, not poor performance. The coefficient on logged sales is no longer statistically significant at conventional levels in the European regressions. We include earnings before interest and taxes (EBIT) in model (3) as an alternative earnings measure. We find that logged earnings before interest and taxes is negatively related to logged market values in Europe, but not in the United States. In model (4) we split between Internet firms that report positive earnings before interest and taxes (LEBIT POS) and negative earnings before interest and taxes (LEBIT NEG). The coefficient on LEBIT POS is insignificant in Europe and positively significant in the United States. The coefficient on LEBIT NEG is negative and highly significant in Europe but insignificant in the United States. We attribute these differences between Europe and the United States to accounting differences. Our U.S. findings largely correspond to Bartov et al. (2002). They find no relation between earnings per share or cash flow per share and the closing market price on the first trading day for U.S. Internet IPOs. Model (5) is an augmented version of model (1). We add two non-accounting variables; free float and the post-IPO percentage ownership of the largest shareholder. Due to data availability, the number of observations is reduced to 114 firms for the European sample. We expect to find that free float is negatively related to market value. If fewer shares are sold at the IPO, the supply of shares is restricted. This may drive up market prices as investors compete for few Internet IPO shares. The largest owner is the party closest to the Internet firm. If the largest owner decides to retain a large fraction of the post-IPO shares, this may signal positive news to investors about the value of the Internet firm. Contrary to expectations, we do not find a significant association between logged market values and free float or post-IPO percentage ownership of the largest owner in Europe. For the U.S. sample, we find that free float is negatively related to logged market values. Again, contrary to our expectations, we do find a negative relationship between logged market values and the post-IPO percentage ownership of the largest owner. The negative and significant association between logged net income and logged market values remains for European but not for U.S. Internet IPOs. Have things changed since the Internet shakeout in April 2000? To answer this question, we estimate model (6) for the 73 European and 259 U.S. firms that went public before the April 1, 2000 (the bubble period). We observe that in the
Valuing Internet Stocks at the Initial Public Offering
149
bubble period, investors priced losses negatively in both IPO markets. The larger the loss, the higher the market value. In addition, free float and ownership by the largest owner are negatively associated with logged market values in the bubble period for the United States. This shows that U.S. stock prices were driven by supply and demand. The fewer shares were sold at the IPO, the higher the market value. Model (7) is identical to model (6), but now we estimate the non-linear regression for the sub-sample of 41 European and 33 U.S. Internet companies that went public after April 1, 2000 (the post-bubble period). We find that investors no longer negatively price losses in the post-bubble period. Both European and U.S. Internet firms with larger losses are no longer valued more by investors. This finding corresponds to the findings of Demers and Lev (2001) for the United States. They report that investors are no longer willing to capitalize marketing and R&D costs as intangible assets in 2000. In contrast to the bubble period, logged sales are positively related to logged market values in the post-bubble period in Europe. Investors seem to attach a higher value to European Internet firms that have earned more revenues before going public. Taken together, our regression results suggest that accounting information is relevant to valuing Internet IPOs. This is consistent with the findings of Hand (2000, 2001) for the United States. However, the relationships between specific accounting variables and market value appear to have changed after the Internet shakeout in April 2000. In the bubble period, investors negatively price losses. The larger the loss, the higher the market value. In the post-bubble period, this relation has disappeared.
5.2. Log-Linear Regressions for Offer Values Tables 8 and 9 show the log-linear regression results using logged offer value as the dependent variable for Europe and the United States, respectively. Offer value is defined as the number of post-IPO shares times the final offer price. We assume that offer values capture the value assessment of the underwriter and the issuing firm. This allows us to examine the determinants of the offer price that are considered to be key by the most informed parties in the IPO market. For reasons of brevity, we focus on the results that are different from the results that we discussed earlier for market values. The regression results for model (1) are presented in the first column. As with market values, there is a negative relation between logged net income (LNI) and the logged offer value. We infer that underwriters and issuing firms also price losses negatively. The larger the loss, the higher the offer value. The indicator variable for Internet Service Providers (ISPs) loads up with a highly significant
150
Table 8. Log-Linear Regression Results for Offer Values of European Internet IPOs. (1) LNI LNI POS LNI NEG LEBIT LEBIT POS LEBIT NEG FLOAT % TOPSH % ISP LREV LBVE CONSTANT R2 adjusted (%) F-test N
−0.407 (−5.926)
(2)
(3)
(4)
***
(5)
(6) Bubble
(7) Post-Bubble
0.384 (0.674) −0.605 (−3.495)***
−0.215 (−0.649) −0.466 (−1.896)*
***
−0.366 (−4.195) −0.128 (−0.526) −0.500 (−5.327)*** −0.355 (−4.806)*** −0.080 (−0.353) −0.483 (−5.091)***
0.716 (5.083)*** 0.247 (2.625)*** 0.002 (0.030) 1.905 (19.278)***
0.734 (5.186)*** 0.202 (1.772)* −0.016 (−0.285) 1.889 (19.922)***
0.753 (4.923)*** 0.306 (3.111)*** −0.008 (−0.016) 1.888 (18.882)***
0.777 (5.081)*** 0.225 (1.794)* −0.031 (−0.549) 1.882 (18.830)***
41.200 24.998*** 138
41.553 20.480*** 138
40.132 23.959*** 138
40.751 19.845*** 138
−1.328 (−1.574) 0.161 (0.862) 0.692 (5.250)*** 0.164 (1.436) 0.032 (0.471) 2.249 (7.534)*** 45.622 16.801*** 114
−1.816 (−1.870)* 0.101 (0.375) 0.551 (3.125)*** −0.033 (−0.188) 0.008 (0.083) 2.484 (6.660)*** 39.739 7.783*** 73
0.461 (0.388) 0.216 (0.867) 0.870 (4.359)*** 0.246 (1.758)* 0.042 (0.299) 1.530 (3.569)*** 62.209 10.407*** 41
MICHIEL BOTMAN ET AL.
Note: The table shows the log-linear regression results using logged offer values (LOVE) as the dependent variable. Independent variables are defined as in Table 5. We put the letter “L” in front of the variable name to indicate that the variable has been log transformed using Eq. (2). LNI POS equals LNI if net income is positive, 0 otherwise. LNI NEG equals LNI if net income is negative, 0 otherwise. LEBIT POS and LEBIT NEG are defined similarly but then using earnings before interest and taxes. We define the bubble period as the period from January 1, 1999 to March 31, 2000. The post-bubble period is from April 1, 2000 to December 31, 2000. White (1980) heteroskedastic-consistent t-statistics are within parentheses. ∗ Significance at the 10% level. ∗∗∗ Significance at the 1% level.
(1) LNI LNI POS LNI NEG LEBIT LEBIT POS LEBIT NEG FLOAT % TOPSH % ISP LREV LBVE CONSTANT R2 adjusted (%) F-test N
(2)
(3)
(4)
−0.173 (−4.535)***
(5)
(6) Bubble
(7) Post-Bubble
0.121 (1.307) −0.135 (−3.131)***
0.135 (0.482) −0.334 (−2.041)*
−2.725 (−11.430) *** 0.149 (2.083)** 0.153 (2.823)*** 0.027 (0.820) 0.047 (2.346)** 2.879 (35.781)***
−1.607 (−1.914) * 0.470 (1.243) 0.354 (2.344)** −0.132 (−1.137) 0.142 (2.885)*** 2.351 (9.862)***
58.681 53.345*** 259
50.537 5.671*** 33
−0.104 (−3.503)*** 0.311 (1.917)* −0.289 (−5.182)*** −0.046 (−0.573) 0.355 (1.491) −0.224 (−2.080)**
0.147 (2.331)** 0.125 (3.427)*** 0.125 (5.376)*** 2.228 (50.728)***
0.130 (2.135)** 0.077 (1.903)* 0.108 (4.630)*** 2.161 (46.278)***
0.190 (2.559)** 0.124 (2.912)*** 0.160 (6.780)*** 2.359 (53.566)***
0.167 (2.155)** 0.092 (1.939)* 0.154 (5.822)*** 2.352 (54.180)***
−2.609 (−10.914)*** 0.187 (2.401)** 0.205 (3.942)*** 0.035 (1.118) 0.073 (3.417)*** 2.846 (32.518)***
28.848 30.495*** 292
31.571 27.851*** 292
22.035 21.562*** 292
24.112 19.492*** 292
55.738 62.076*** 292
Valuing Internet Stocks at the Initial Public Offering
Table 9. Log-Linear Regression Results for Offer Values of U.S. Internet IPOs.
Note: The table shows the log-linear regression results using logged offer values (LOVE) as the dependent variable. Independent variables are defined as in Table 5. We put the letter “L” in front of the variable name to indicate that the variable has been log transformed using Eq. (2). LNI POS equals LNI if net income is positive, 0 otherwise. LNI NEG equals LNI if net income is negative, 0 otherwise. LEBIT POS and LEBIT NEG are defined similarly but then using earnings before interest and taxes. We define the bubble period as the period from January 1, 1999 to March 31, 2000. The post-bubble period is from April 1, 2000 to December 31, 2000. White (1980) heteroskedastic-consistent t-statistics are within parentheses. ∗ Significance at the 10% level. ∗∗ Significance at the 5% level. ∗∗∗ Significance at the 1% level.
151
152
MICHIEL BOTMAN ET AL.
coefficient. These types of Internet companies have higher offer values. The indicator variable for ISPs is significant in all subsequent regressions. Logged sales are positively related to logged offer values in both European and U.S. IPO markets. Market place acceptance appears to be a value driver of the offer price set by underwriters and issuers. In contrast to the market value regressions, the positive relation between sales and offer value does not lose significance in subsequent regressions. As before, the log of the book value of equity is not related to the logged offer value in Europe, but highly significant in the United States. In model (2) we include a variable LNI POS that equals the logged net income if it is positive, and a variable LNI NEG that equals the logged net income if it is negative. We find no significant association between logged offer value and LNI POS in Europe and a marginally significant association between logged offer value and LNI POS in the United States. We report a negative and significant relation between logged offer value and NI NEG for both Europe and the United States. Because the coefficient of LNI NEG lies between −1 and 0, there is a concave relation between losses and offer value. This reinforces that underwriters and issuers are negatively pricing losses in both European and U.S. IPO markets. The larger the loss, the higher the offer value. A difference with the market value regressions is that logged sales remain positively related to offer values. Underwriters and issuers set higher offer prices if the firm earns more revenues in Europe as well as the United States. Model (3) and model (4) use earnings before interest and taxes (EBIT) as an alternative accounting variable. Results are similar as for the regressions using logged market values as the dependent variable. The only difference relates to the American sample. The variable LEBIT NEG is negatively related to logged offer values, whereas it was not related to logged market values for the sample of U.S. Internet IPO firms. Again, logged sales remain significant in these regressions. In model (5) we include non-accounting variables. We find that free float, defined as the number of shares sold at the IPO divided by the number of post-IPO shares, is negatively related to logged offer values in both European and U.S. IPO markets. Underwriters and issuers seem to anticipate that investors will bid up prices in the light of limited supply. This leads them to set higher offer prices. The percentage post-IPO ownership by the largest shareholder is not significantly related to logged offer values in Europe but positively related to logged offer values in the United States. Compared to our previous analysis of logged market values, U.S. underwriters and issuers are therefore positively pricing the ownership of the largest owners, whereas small investors are negatively valuing higher levels of ownership by the largest owner. We also investigate whether changes have occurred by splitting the sample into a group of 73 European and 259 U.S. Internet IPOs that went public during
Valuing Internet Stocks at the Initial Public Offering
153
the Internet bubble and 41 European and 33 U.S. IPOs that went public after the Internet shakeout in April 2000. Model (6) is estimated for the bubble samples. We find that underwriters and issuers negatively price losses and that float is negatively related to logged offer values in both European and American IPO markets. There is no relation between logged sales and logged offer value. Model (7) shows the regression results for the post-bubble sample. The coefficient on LNI NEG is statistically significant and negative. This indicates that both European and American underwriters continue to price losses negatively. There is no relation between float and logged offer value in Europe. In the United States, float remains negatively related to logged offer values. This indicates that in the post-bubble period, European underwriters and issuers were no longer expecting that they could set higher offer prices when they restricted supply of shares at the IPO. Conversely, U.S. underwriters and issuers counted on restricted supply of shares to justify higher offer prices. In short, there are only few differences between the way underwriters value Internet IPOs and the way small investors value these firms at the first day of trading (values based on first-day closing prices). One important difference is that underwriters and issuers have continued to negatively price losses after the Internet shakeout in April 2000. Investors, on the other hand, do no longer value losses negatively in that period.
6. CONCLUSIONS This paper examines the relevance of accounting information to valuing Internet IPOs in Europe and the United States, both before and after the Internet shakeout in April 2000. We document several differences between European and U.S. Internet IPO firms. For example, we find that a significantly higher fraction of European Internet IPO firms are active in the content/portals and Internet services sub-sectors while a significantly higher fraction of U.S. Internet IPO firms operate in the e-commerce and IT-infrastructure sub-sectors. We also show that European IPO firms tend to report a smaller amount of loss in the year before the IPO, sell more shares to the public, have more concentrated ownership by the largest owner and experience lower first-day returns than their U.S. counterparts. We first use market value as our proxy for firm value. We assume that market values capture the value assessment of small investors. We find that accounting variables such as net income are relevant to valuing Internet IPOs. In particular, we document a non-linear relation between earnings and market values. We find a negative pricing of losses in the period before April 1, 2000. The larger the loss, the higher the market value. But the negative pricing of losses does not extend
154
MICHIEL BOTMAN ET AL.
beyond the Internet shakeout. The inverse relationship between market value of equity and earnings is consistent with an Internet firm’s start-up expenditures being considered by the market as assets, not as costs. These findings hold for both the European and the U.S. IPO market. The free float is negatively related to market values in the bubble period, but only for the U.S. IPO market. This shows that if fewer shares were sold in the IPO, the limited supply together with heavy demand for IPO shares drove market values higher. In the post-bubble period after April 1, 2000 the relation between free float and market values is absent. Contrary to expectations, we do not find a positive relation between logged market value and the post-IPO percentage ownership of the largest shareholder. Next, we perform all analyses from the viewpoint of the underwriter and issuing firm (values based on final offer price). Overall, we find qualitatively similar results. One important difference is that logged sales remain positively related to offer values in most regression specifications. Underwriters and issuers set higher offer prices if the firm earns more revenues. This holds true for both Europe as the United States. Another difference is that underwriters and issuers have continued to negatively price losses after the Internet shakeout in April 2000. Investors, on the other hand, do no longer value losses negatively during that period. Finally, underwriters and issuers are positively pricing the ownership of the largest owner, whereas small investors are negatively valuing higher levels of ownership. This finding is specific for the United States.
NOTES 1. Some companies operate in more than one sub-sector. For example, one of our sample companies, World Online, offers access to the Internet, Internet services and content/portal services. In those cases we have chosen the sub-sector which represents the major part of the firm’s revenues. 2. Our sample includes 15 European Internet Service Providers (ISPs), such as T-Online International AG, World Online International N. V. and Terra Networks, S. A. The average (median) market value for ISPs equals $7,740 million ($1,284 million) versus $449 million ($213 million) for other Internet firms. Similarly, the U.S. sample includes 27 ISPs with an average (median) market value equal to $1,401 million ($491 million) compared to $805 million ($324 million) for other Internet firms. Because we use unscaled market values, it is important to control for this effect in the regressions.
REFERENCES Aaij, S., & Brounen, D. (2002). High-tech IPOs: A tale of two continents. Journal of Applied Corporate Finance, 15, 87–94.
Valuing Internet Stocks at the Initial Public Offering
155
Bartov, E., Mohanram, P., & Seethamraju, C. (2002). Valuation of internet stocks – An IPO perspective. Journal of Accounting Research, 40, 321–346. Bowen, R. M., Davis, A. K., & Rajgopal, S. (2002). Determinants of revenue-reporting practices for internet firms. Contemporary Accounting Research, 19, 523–562. Collins, D. W., Pincus, M., & Xie, H. (1999). Equity valuation and negative earnings: The role of book value of equity. The Accounting Review, 74, 29–61. Core, J. E., Guay, W. R., & Van Buskirk, A. (2003). Market valuations in the new economy: An investigation of what has changed. Journal of Accounting and Economics, 34, 43–67. Demers, E., & Lev, B. (2001). A rude awakening: Internet shakeout in 2000. Review of Accounting Studies, 6, 331–359. Hand, J. R. M. (2000). Profits, losses and the non-linear pricing of internet stocks. Working Paper, Kenan-Flagler Business School, UNC Chapel Hill. Hand, J. R. M. (2001). The role of book income, web traffic, and supply and demand in the pricing of U.S. internet stocks. European Finance Review, 5, 295–314. Johnston, J., & Madura, J. (2002). The performance of internet firms following their initial public offering. The Financial Review, 37, 525–550. Ljungqvist, A., & Wilhelm, W. (2003). IPO pricing in the dot-com bubble. Journal of Finance, 58, 723–752. Loughran, T., & Ritter, J. (2003). Why has IPO underpricing increased over time? Working Paper, University of Florida. Ofek, E., & Richardson, M. (2002). The valuation and market rationality of internet stock prices. Oxford Review of Economic Policy, 18, 265–287. Ofek, E., & Richardson, M. (2003). Dotcom mania: The rise and fall of internet stock prices. Journal of Finance, 58, 1113–1138. Ohlson, J. (1995). Earnings, book values and dividends in security valuation. Contemporary Accounting Research, 11, 661–687. Rajgopal, S., Kotha, S., & Venkatachalam, M. (2003). The value relevance of network advantages: The case of E-commerce firms. Journal of Accounting Research, 41, 135–162. Rajgopal, S., Venkatachalam, M., & Kotha, S. (2002). Managerial actions, stock returns and earnings: The case of Business-to-Business internet firms. Journal of Accounting Research, 40, 529–556. Ritter, J. R. (2003). Differences between European and American IPO markets. European Financial Management, 9, 421–434. Schultz, P. H., & Zaman, M. (2001). Do the individuals closest to internet firms believe they are overvalued? Journal of Financial Economics, 59, 347–381. Trueman, B., Wong, M. H. F., & Zhang, X. (2000). The eyeballs have it: Searching for the value in internet stocks. Journal of Accounting Research, 38, 137–170. White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica, 48, 721–746.
THE ROLE OF ACCOUNTING DATA AND WEB-TRAFFIC IN THE PRICING OF GERMAN INTERNET STOCKS Andreas Trautwein and Sven Vorstius ABSTRACT This study looks at the value-relevance of accounting data and measures of web-traffic for Internet firms listed on the Neuer Markt. We show that earnings and cash flows cannot explain the valuation of Internet companies, while we report a positive association between total sales and market capitalisation. In addition, sales and marketing as well as research and development expenses are relevant value-drivers. Furthermore, we find a positive relation between market values and a number of web-metrics such as customer loyalty, reach, page impressions, and unique visitors. We conclude that during the Internet bubble, measures of web-traffic provided at least as much explanatory power for market values as financial statement information.
1. INTRODUCTION During the last quarter of 1999 and the first of 2000, stock prices of companies in the Internet industry increased sharply, a development that was of considerable importance to investors and researchers alike. It could be observed however, that financial performance measures, such as earnings and cash flows, did not compare to rising stock prices in the New Economy. Therefore, the reliability and sustainability of the The Rise and Fall of Europe’s New Stock Markets Advances in Financial Economics, Volume 10, 157–180 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3732/doi:10.1016/S1569-3732(04)10007-8
157
158
ANDREAS TRAUTWEIN AND SVEN VORSTIUS
astronomical market valuations observable during that time were questioned. For example, Laderman and Smith (1998, p. 121) write that it is: “difficult to develop a valuation model that explains why an Internet firm is trading at a P/E multiple exceeding 300.” In this context, the so-called Internet bubble and its possible bursting became the object of much discussion in academic circles and in the media. As a central element of this discussion, the question arose what else, if not financial performance, could be driving stock prices. Consequently, academic research focused on the developments at NASDAQ and in particular investigated the value-relevance of financial information in comparison to alternative industryspecific measures (e.g. Demers & Lev, 2001; Hand, 2001b, 2003; Rajgopal et al., 2003; Trueman et al., 2000). One such alternative metric relates to measures of web-traffic – web-metrics – which met with increased interest on the part of analysts and investors. Corresponding developments on Germany’s Neuer Markt have, however, not been investigated to the same extent.1 The objective of this chapter therefore is to identify value-drivers of German Internet companies during the period from October 1999 to May 2000. As such, our intention is not to develop a formal valuation model for firms in the Internet industry, but rather to establish factors that are highly associated with the market values of Internet companies. In doing so, this study attempts to contribute to the understanding of the investment behaviour of market participants during the Internet bubble. One might assume that in such a market environment investors would shift their focus away from financial measures and towards alternative metrics when evaluating investment opportunities. As this chapter targets an evaluation of the value-relevance of financial information compared to web-metrics, it allows conclusions to be drawn with respect to the disclosure of value-relevant information in such a market environment.2 Strict compilation provisions ensure availability of financial data in a standardised manner and hence guarantee comparability. Industry-specific information such as web-traffic measures, however, have not been included in reporting regulations and at present, a standardised reporting system and framework does not exist. In particular, this chapter examines the value-relevance of web-metrics, viz. page impressions, unique visitors, reach, stickiness, and customer loyalty vs. financial variables such as total sales, earnings before interest and taxes, operating cash flow, sales and marketing expenses, and research and development expenses. Previous studies have investigated value-drivers of listed companies in general (e.g. Francis & Schipper, 1999; Lev & Sougiannis, 1996; Lev & Thiagarajan, 1993). Several studies focus on the valuation of Internet companies in particular. Schwartz and Moon (2000) develop a real option model which describes the valuation of Internet stocks based on a variety of input variables. Demers and Lev (2001), Hand (2001b, 2003), Rajgopal et al. (2003) and Trueman et al.
The Role of Accounting Data and Web-Traffic
159
(2000) examine value-relevant factors of stock market prices of U.S. Internet companies. While Demers and Lev (2001), Hand (2001b, 2003) and Trueman et al. (2000) focus more on the value-relevance of accounting data, Rajgopal et al. (2003) concentrate on the value-relevance of web-metrics. For Germany, D’Arcy and Leuz (2000) analyse corporations listed on the Neuer Markt with regard to financial reporting and give a detailed description of accounting practices and characteristics, but do not seek to determine value-drivers. This study investigates the value-relevance of a series of firm-specific variables of Internet companies, both financial and non-financial, on the Neuer Markt in Germany. The structure of this chapter is as follows: Section 2 presents our hypotheses regarding the value-relevance of Internet firm-specific information. Section 3 explains the sampling procedure and the data collection while Section 4 covers the statistical analysis and discusses the empirical results. Section 5 concludes and compares our findings to the results of recent U.S. studies.
2. HYPOTHESIS DEVELOPMENT 2.1. The Relative Value-Relevance of Financial Information and Web-Metrics Financial information refers to data derived from financial statements, i.e. accounting data.3 Traditionally, financial information has been widely used in company valuation and therefore can be expected to be value-relevant. For example, investors typically base their investment decisions on the evaluation of financial statement data. Moreover, analysts also look, among other variables, at financials when recommending individual shares or entire industry segments. However, during the Internet bubble, it seemed that market valuations could no longer be explained solely by financial information and the application of traditional valuation models, as has been shown by recent academic research.4 This suggests that the value-relevance of financial measurements has eroded, while alternative metrics have increased in importance. This applies to Internet companies in particular (Lev, 2000). Therefore, the value-relevance of alternative measures in contrast to financial information should be assessed, in order to contribute to our understanding of the high market valuations during the Internet bubble.5 Web-metrics could be regarded as one such alternative measure, given that analysts increasingly employed measures of web-traffic in valuation models.6 However, information on web-traffic is not included in accounting standard frameworks and no standard definition exists; nor is there an obligation to publish such information. While the purpose of this study is to test potential
160
ANDREAS TRAUTWEIN AND SVEN VORSTIUS
non-accounting value-drivers of Internet firms, i.e. web-metrics, during the period between the second half of 1999 and the first half of 2000, it also aims to investigate the value-relevance of accounting data in comparison to these web-metrics. We hypothesise: Hypothesis 1. Web-metrics (WM) provide at least as much value-relevance for an Internet company as traditional financial measures (FI). The following sections introduce sub-hypotheses regarding the value-relevance of financial information and web-metrics.
2.2. The Value-Relevance of Financial Information Earnings have always played a crucial role in valuation models. However, current empirical evidence suggests that over the past 50 years, an increasing number of companies report losses (e.g. Hayn, 1995). Especially Internet companies reported huge losses, but nevertheless market valuations were high during the time period of this study. We therefore predict that earnings do not provide much value-relevance and hence do not drive market value.7 This translates into the following: Hypothesis 1a. Earnings before interest and taxes (EBIT) are not significantly positively associated with the market valuation of an Internet company. Business operations of Internet firms were often accompanied by high cash expenditures, targeted for instance at acquiring customers and capturing market share. Little if any cash could be expected to be generated from operations. As a result, operating cash flow is not expected to be closely associated with market capitalisation.8 We test the following hypothesis: Hypothesis 1b. Operating cash flow (CF) is not significantly positively associated with the market valuation of an Internet company. Returns to scale and the ability to set an industry-wide standard, along with the impact of network effects, are believed to be among the critical success factors of Internet companies. Growth in total sales may indicate an increase in the ability to generate future economic value and can hence be regarded as a potential value-driver. We therefore predict a positive relationship between total sales and market value and formulate the following hypothesis: Hypothesis 1c. Total sales (TS) are significantly positively associated with the market valuation of an Internet company.
The Role of Accounting Data and Web-Traffic
161
Sales and marketing as well as research and development expenses may be regarded as investments in intangibles such as the acquisition of customers and market share. This is believed to be of critical importance within the Internet industry, due to an inherent winner-takes-all business model.9 Hand (2001a, p. 2) writes that E-commerce firms are known for their substantial advertising expenditures and extensive marketing campaigns aimed at extending and fostering market penetration, brand perception and brand awareness. In contrast, firms selling E-commerce enabling technologies (Enabler) tend to spend more money on research and development. Both expenses may eventually translate into future value and are thus likely to be reflected in market values. The findings of Demers and Lev (2001), Hand (2003), and Trueman et al. (2000) support the prediction of a positive relationship between sales and marketing expenses and market value. Accordingly, we expect a significant value-relevance of sales and marketing expenses for the E-commerce firms and of research and development expenses for the Enabler firms: Hypothesis 1d. Sales and marketing expenses (SME)/research and development expenses (RND) are significantly positively associated with the market valuation of an Internet company.
2.3. The Value-Relevance of Web-Metrics Particularly within the environment of the Internet, customers are believed to be a valuable asset, as benefits from network effects can only be enjoyed once brand awareness and critical market share have been successfully established. In this context, web-metrics represent a chance to learn about the behaviour of Internet users and thus the customer base of a company. An overview and definition of all web-metrics used in this study is provided in Table 1. Page impressions measure the number of visits to a particular website and capture the overall frequency of usage and popularity. This is an important measure for a company in the E-commerce business and has been empirically verified as a potential value-driver by Demers and Lev (2001), Hand (2001b), and Trueman et al. (2000). Hence, we hypothesise: Hypothesis 1e. Page impressions (PI) are significantly positively associated with the market valuation of an Internet company. Unique visitors refer to the number of individuals that visit a web domain and hence approximates customer awareness. It allows conclusions to be drawn on the ability of the specific company to attract customers. It follows:
162
ANDREAS TRAUTWEIN AND SVEN VORSTIUS
Table 1. Definition of Web-Metrics. Variable
Definition
Page impressions (PI)
Number of unique visitors multiplied by the average number of pages visited per unique visitor and day over a specific period of time Number of individual visits to a web-domain Number of unique visitors divided by the population of Internet users in Germany Average time spent on visits to a web-domain in minutes per day and per unique visitor Average number of visits to a web-domain in days per month and per unique visitor
Unique visitors (UV) Reach (RE) Stickiness (ST) Customer loyalty (CL)
Hypothesis 1f. Unique visitors (UV) are significantly positively associated with the market valuation of an Internet company. Reach represents the number of individual visits to a particular web-domain as a percentage of the population of Internet users in Germany. It measures the awareness of a website among Internet users and may therefore be value-relevant. Consequently, we predict the following: Hypothesis 1g. Reach (RE) is significantly positively associated with the market valuation of an Internet company. Stickiness refers to the time an individual visit to a certain web-domain lasts. This measure could potentially be value-relevant, but is extremely dependent on the business model that in turn determines the average duration of a website visit. The significance of stickiness as a value-driver of Internet firms’ stock prices has also been empirically tested by Demers and Lev (2001) and Hand (2001b) leading to the following hypothesis: Hypothesis 1h. Stickiness (ST) is significantly positively associated with the market valuation of an Internet company. Customer loyalty offers insights into the number of repeat visits to a web-domain. Customer loyalty measures the ability of a company to generate repeated business and to bind customers. This is regarded as a critical success factor as switching costs for customers are extremely low. This leads to the following hypothesis: Hypothesis 1i. Customer loyalty (CL) is significantly positively associated with the market valuation of an Internet company.
The Role of Accounting Data and Web-Traffic
163
3. SAMPLE SELECTION AND DATA COLLECTION 3.1. Sample Companies As a standard definition for Internet companies does not exist and differing approaches can be found in the literature, this study relies on a broadly accepted definition published by internet.com.10 A total of 60 companies originating from the Internet Index of the Neuer Markt are considered in the sample. These are subdivided into two sample groups consisting of 23 E-commerce and 37 Enabler companies.11 This distinction accounts for the heterogeneity of the business models and existing strategic positions among Internet firms. Tables A1 and A2 in the Appendix show a complete list of all firms included in each group. At this point it is important to stress that an analysis of the value-relevance of web-metrics is only feasible for the E-commerce sample. The Enabler sample serves as a control group that allows us to compare the relationship between financial variables and market valuations across both groups. Financial data is collected for the period from July 1998 to May 2000, while web-metrics data covers October 1999 to May 2000 because of restricted data availability.
3.2. Data Sources and Data Collection This study is based on accounting information published in corporate reports and comprises a total of 229 firm-quarters.12 Data for the following accounting items is collected: total sales, earnings, operating cash flow, sales and marketing expenses, and research and development expenses. Two statistical adjustments are made to the data. First, we log-transform the data, which requires a linear transformation of earnings and cash flow to ensure that all observations are positive. See Section 4.1 for details on this log transformation. Second, we eliminate outliers and unusual observations in order to improve statistical robustness. Unusual observations are identified using Cook’s distance measure and a visual inspection of scatter plots. See Table A5 in the Appendix for a list of observations that are excluded. Furthermore, this study relies on web-metrics supplied by Media Metrix Europe (MMXI Europe) and the Gesellschaft f¨ur Konsumforschung (GfK).13 Information on web-traffic was available only for 16 companies in the E-commerce sample.14 Overall, information on five variables is considered, viz. page impressions, unique visitors, reach, stickiness, and customer loyalty. The statistical adjustment of the observations follow the same procedure as applied to the accounting data. However, no outliers or unusual observations can be identified.15 Thomson Datastream serves as a source for the market values that are
164
ANDREAS TRAUTWEIN AND SVEN VORSTIUS
taken on the day at which the company report was published.16 Again the data is log-transformed. It is important to note that the empirical assessment of the value-relevance of the financial information is divided into two stages. In a first step, it concentrates on the assessment of the value-relevance of accounting data across the two samples. In a second step, the value-relevance of accounting data is compared to the value-relevance of web-metrics for the E-commerce sample. Consequently, two different sets of accounting data for the E-commerce sample are subject to the analysis. Due to restricted availability of web-traffic data, these two sets differ in both the time span and the in the number of companies included. Firstly, web-metrics were not available for all companies included in the E-commerce sample and secondly, web-metrics were only available starting from October 1999. We do not discuss the empirical results concerning the second set of accounting data separately because they do not substantially differ from the first set.
4. EMPIRICAL ANALYSIS AND RESULTS 4.1. Methodology We follow Hand (2000, pp. 15–16) and apply a log-linear regression model to examine the relationship between firm-specific measures and market value. Hand (2000) presents two major econometric arguments in favour of a log linear regression model.17 Firstly, a transformation using the logarithm function can reduce the number of possible outliers, which is especially important when dealing with small samples. And secondly, the results of a log-linear regression are generally less subject to heteroscedasticity. The procedure is applied by taking the natural logarithm of all variables included in the analysis, which is reflected in the regression model shown below.18 To test the formulated hypotheses we estimate univariate as well as multivariate regression models.19 All regressions are estimated using pooled cross-sectional and time-series data and as the number of data points in some of the samples is small, level data is used.20 Consequently, the univariate regression model takes the following form: ln (Market value) = c 0 + c 1 ln (Variable)
(1)
and accordingly the multivariate regression model is denoted as: ln (Market value) = c 0 + c 1 ln (Variable1 ) + c 2 ln (Variable2 ) + · · · + c n ln (Variablen )
(2)
The Role of Accounting Data and Web-Traffic
165
4.2. Evidence of the Value-Relevance of Financial Information Table 2 summarises the results for the E-commerce and Table 3 for the Enabler sample.21 The results of the hypotheses testing are as follows. Hypothesis 1a predicts that earnings before interest and taxes (EBIT) are not positively associated with the market value of Internet firms listed on the Neuer Markt. The univariate regression results support this hypothesis for both samples. The regression coefficient is significantly negative at a 5% level for the Enabler sample. This suggests that there is a negative log-linear relationship, i.e. higher losses are associated with higher market values. We conclude that in 1999 and 2000, market participants did not seem to rely on earnings when valuing Internet companies. We find a negative rather than positive relation between earnings and market valuation for both E-commerce and Enabler firms. Hypothesis 1b states that cash flow (CF) is not positively related to the market valuation of an Internet company. Based on the empirical results obtained from the regression this hypothesis cannot be rejected for both the E-commerce and the Enabler sample. The regression coefficients are not significant on a 5% level and, in addition, an adjusted R2 of less than 2% in the Enabler sample does not suggest much value-relevance either. In conclusion, components and input variables of traditional valuation methods such as profit measures and cash flows do not provide significant explanatory power regarding the market valuation of an Internet firm in 1999 and 2000. Furthermore, we find a significantly positive association between total sales (TS) and market value. This finding is consistent with Hypothesis 1c. For the E-commerce (Enabler) sample, the regression model results in an adjusted R2 of 32% (48%), with both the constant and the coefficients being highly significant. Therefore, total sales are identified as a value-driver for both E-commerce and Enabler firms. The explanatory power, however, is slightly higher for the Enabler than for the E-commerce sample. We argue that the market recognises turnover as a proxy for market penetration and hence success of the business model. Moreover, turnover is a reliable measure in the sense that it aggregates the performance over the business activities and is not influenced by the choice of the accounting system or managerial decisions. Hypothesis 1d states that sales and marketing expenses (SME) for the E-commerce sample and research and development expenses (RND) for the Enabler are positively related to market value. Our results confirm that sales and marketing expenses and research and development expenses are value-drivers. Investors were thus prepared to pay more for a stock if the company chose to devote more resources to marketing or to research and development. For the E-commerce (Enabler) firms the model ascribes 27% (16%) of variations in market values to variations in sales and marketing expenses and research and development expenses respectively. More explanatory power is attributed to
166
Table 2. Summary Regression Statistics of the Relationship Between Market Values and Financial Information for the E-Commerce Sample. n
Constant
EBIT
H1a H1b H1c H1d
52 36 54 32
54.16 (1.97)* 7.37 (0.19) 12.03 (8.15)*** 13.36 (7.70)***
−1.97 (−1.26)
CF
TS
SME
0.67 (0.31) 0.50 (5.09)*** 0.40 (3.32)***
Adj. R2 (%) 1 0 32 27
Note: Regression coefficient and, in brackets, t-statistic. EBIT: Earnings Before Interest and Taxes, CF: Cash Flow, TS: Total Sales, SME: Sales and Marketing Expenses. ∗ Significance level at 10%. ∗∗∗ Significance level at 1%.
ANDREAS TRAUTWEIN AND SVEN VORSTIUS
Hypothesis
Hypothesis
H1a H1b H1c H1d
n
Constant
EBIT
106 103 105 48
56.96 (3.85)*** −12.60 (−0.69) 6.35 (4.57)*** 11.87 (4.67)***
−2.10 (−2.50)**
CF
TS
RND
1.80 (1.77)* 0.88 (9.77)*** 0.57 (3.17)***
Adj. R2 (%) 5 2 48 16
The Role of Accounting Data and Web-Traffic
Table 3. Summary Regression Statistics of the Relationship Between Market Values and Financial Information for the Enabler Sample.
Note: Regression coefficient and, in brackets, t-statistic. EBIT: Earnings Before Interest and Taxes, CF: Cash Flow, TS: Total Sales, RND: Research and Development Expenses. ∗ Significance level at 10%. ∗∗ Significance level at 5%. ∗∗∗ Significance level at 1%.
167
168
ANDREAS TRAUTWEIN AND SVEN VORSTIUS
sales and marketing expenses in the E-commerce sample than to research and development expenses in the Enabler sample. We conclude that investors regard sales and marketing expenses as more of an investment in the generation of future economic value than research and development expenses.
4.3. Evidence of the Value-Relevance of Web-Metrics We now investigate the sample of E-commerce companies for which data on web-metrics is available. The regression results show that web-metrics are value-relevant and positively related to the market value of an Internet firm. Consequently, investors regard web-traffic as being important when valuing E-commerce companies. Table 4 summarises the results of the univariate log-linear regressions. The results support Hypotheses 1e, 1f, 1g and 1i, but do not support Hypothesis 1h. The findings show that variations in customer loyalty, page impressions, reach and unique visitors explain 62, 40, 40, and 39% of the variation in market value of E-commerce firms during the Internet bubble, respectively. Investors seem to have attached a high degree of relevance to measures of web-traffic when evaluating an investment opportunity during that time and were prepared to pay a higher price if a company was generating more traffic. The strongest log-linear association with market values can be identified for customer loyalty, which approximates the firm’s ability to bind customers and generate repeated business. Page impressions are also important. Reach captures market penetration and brand awareness, and is positively related to market values of Internet firms. By the same token, unique visitors are a value-driver for E-commerce firms. Stickiness, however, could not be identified as a value-driver, which is probably due to remaining heterogeneity in business models within the E-commerce group. In this context, it seems plausible that customers on average spend more time when shopping for items than when carrying out a search visiting a portal.
4.4. Evidence of the Value-Relevance of Financial Information and Web-Metrics As formulated in Hypothesis 1, an objective in this study is to compare the value-relevance of financial information to the value-relevance of web-metrics. Table 5 summarises the results of the multivariate regression models. Model 1 (M1) contains all financial data, i.e. earnings (EBIT), cash flow (CF), total sales
Hypothesis
n
Constant
PI
H1e H1f H1g H1h H1i
66 87 87 66 66
11.94 (9.03)*** 9.37 (6.36)*** 19.06 (88.20)*** 20.44 (37.27)*** 18.47 (75.38)***
0.57 (6.70)***
UV
RE
ST
CL
0.87 (7.54)*** 0.93 (7.62)*** 0.19 (0.61) 2.10 (10.38)***
Adj. R2 (%) 40 39 40 0 62
The Role of Accounting Data and Web-Traffic
Table 4. Summary Regression Statistics of the Relationship Between Market Values and Web-Metrics in the E-Commerce Sample.
Note: Regression coefficient and, in brackets, t-statistic. PI: Page Impressions, UV: Unique Visitors, RE: Reach, ST: Stickiness, CL: Customer Loyalty. level at 1%.
∗∗∗ Significance
169
170
Table 5. Summary Regression Statistics of the Relationship Between Market Values and Financial Information as well as Web-Metrics in the E-Commerce (EC) and Enabler (EN) Sample. n
Constant
EBIT
CF
TS
SME/RND
M1/EC
29
M2/EN
46
−2.19 (−1.81)* −0.87 (−1.06)
−4.68 (−1.78)* 1.81 (1.63)
0.40 (1.92)* 1.04 (7.39)***
−0.14 (−0.8) 0.01 (0.05)
M3/EC
66
M4/EC
50
136.46 (3.03)*** −12.40 (−0.75) 17.20 (3.31)*** 12.27 (10.49)***
PI
RE
ST
CL
Adj. R2 (%) 51 67
−0.79 (−3.49)*** 0.45 (5.40)***
UV
0.89 (1.57)
0.37 (0.68)
0.44 (20.6)**
2.52 (8.00)*** 1.51 (6.94)***
70 73
Note: Regression coefficient and, in brackets, t-statistic. EBIT: Earnings Before Interest and Taxes, CF: Cash Flow, TS: Total Sales, SME: Sales and Marketing Expenses, RND: Research and Development Expenses, PI: Page Impressions, UV: Unique Visitors, RE: Reach, ST: Stickiness, CL: Customer Loyalty. ∗ Significance level at 10%. ∗∗ Significance level at 5%. ∗∗∗ Significance level at 1%.
ANDREAS TRAUTWEIN AND SVEN VORSTIUS
Model/ Sample
The Role of Accounting Data and Web-Traffic
171
(TS), and sales and marketing expenses (SME) as independent variables. The model explains about 51% of the variation in market values for E-commerce companies. The results show that market values of firms in the E-commerce group are negatively associated with earnings as well as cash flows, whereas the opposite holds true for total sales. All coefficients apart from sales and marketing expenses are significant at a 10% level. It follows that higher losses are associated with a higher market valuation of equity. Investors were willing to pay a higher price irrespective of larger losses incurred. The results also suggest that higher sales are associated with higher market values. However, the coefficients for sales and marketing expenses as well as cash flow are negative. Hence, taking into account the multivariate evidence, market participants seem to positively value less cash flow. The multivariate results conflict with the univariate regression results that showed that neither cash flows nor earnings explain market values and that sales and marketing expenses were positively associated with market values. Only total sales are positively related to market values in both the multivariate and univariate regression models.22 Comparing M1 (E-commerce) to M2 (Enabler), the results are as follows. The coefficient for total sales is significant in both samples and indicates a positive relationship with market values. Earnings, cash flow and research and development expenses on the other hand are not significant in the Enabler sample, while earnings and cash flow are significant at a 10% level in the E-commerce sample. The sign of the coefficient for earnings is negative for the univariate and multivariate regression models across both samples, with different levels of significance. In the Enabler sample the coefficients of research and development expenses and cash flow are positive, but are not significant. This contrasts with the univariate results. The explanatory power of M1 and M2 results in an adjusted R2 of 51% for the E-commerce sample and 67% for the Enabler. The results seem to support the notion that some financial information was indeed value-relevant for Internet firms in both samples during that time. Model (M3), containing all web-metrics, provides an explanatory power of 70%. This is comparable to the degree of association that was found for financial information. Customer loyalty and page impressions are highly significant at a 1% level, which is identical to the results obtained in the univariate models. The coefficients of reach and customer loyalty are not significant, while stickiness is significant on a 5% level. This contrasts with the results obtained in the univariate regressions. All coefficients apart from page impressions are positive. The multivariate results confirm the earlier finding that customer loyalty is crucial to the valuation of an Internet firm. While an interpretation of the results and especially a comparison to univariate regression results must be carried out with caution, we believe it is safe to conclude
172
ANDREAS TRAUTWEIN AND SVEN VORSTIUS
that web-traffic data is value-relevant for Internet firms. Furthermore, our results suggest that non-financial data is of equal significance to investors as accounting data. Combining financial data and web-metrics in one model is not statistically feasible, as the sample size in relation to the number of independent variables becomes too small. Nevertheless, we estimate model M4 that includes a combination of the most significant financial data and web-metrics. In this model, both coefficients are highly significant and a higher explanatory power of the model is obtained. This finding is consistent with Hypothesis 1. Web-metrics provide at least as much value-relevance for investors as traditional financial measures.
5. CONCLUDING REMARKS This study aims to relate firm-specific measures to the market values of Internet firms listed on the Neuer Markt in Germany during the Internet bubble in 1999 and 2000. In summary, our results suggest that measures of web-traffic provide information which is at least as relevant as financial data when explaining market values of Internet companies in this specific setting. As such, our study arrives at conclusions similar to Demers and Lev (2001), Rajgopal et al. (2003) and Trueman et al. (2000) for the United States. However, there are differences between our results and those obtained in earlier U.S. studies. Unlike Hand (2003), this study finds measures of profit to be associated less strongly with market value than total sales. In a second study, Hand (2001b) shows that financial information performs significantly better than web-traffic in explaining market values of Internet firms. This contrasts with the evidence obtained in this study that suggests a similar value-relevance instead. Moreover, our results show that measures of web-traffic are value-relevant. Trueman et al. (2000) and Rajgopal et al. (2003) also assess the additional value-relevance of web-metrics combined with accounting data, and find these measures to be highly significant, too. Trueman et al. (2000) show that unique visitors and page impressions are important value-drivers while Rajgopal et al. (2003) investigate reach and unique visitors, but do not further differentiate between the two. Hand (2001b) considers unique visitors, page impressions, reach and stickiness, but only reports visitors as a highly significant value-driver. Demers and Lev (2001) concentrate on the value-relevance of reach, unique visitors, page impressions, stickiness, and customer loyalty. In contrast to this study, their results do not show the strongest association with market valuation for the variables reach and page impressions. Our findings are of potential interest to investors and standard-setters, because web-metrics, although they prove highly value-relevant, are not subject to
The Role of Accounting Data and Web-Traffic
173
strict and standardised reporting regulations. Furthermore, generally accepted definitions for web-metrics are lacking. This brings up the question if and how industry-specific information, such as web-metrics for the Internet industry, should be disclosed. In conclusion, our findings suggest that web-metrics are relevant to the market valuation of Internet companies. More research, however, is needed to determine whether and how web-metrics could be incorporated into financial reporting regulations for Internet companies.
NOTES 1. The Neuer Markt was established in 1997 as a stock market segment of Deutsche B¨orse AG aimed at providing growth companies with an entrance possibility to the capital market. Its regulatory requirements were rather strict by German standards. In 2003 and following a major restructuring process initiated by the Deutsche B¨orse AG concerning its markets segments, the Neuer Markt ceased to exist as a market segment. In part it was replaced by the Prime Standard, which is now subject to a comparable set of regulatory requirements. 2. Analysing the value-relevance of accounting information as opposed to factors not incorporated into standardised reporting regulations has become an important topic in empirical accounting research. See Barth et al. (2001) and Holthausen and Watts (2001) for a discussion. 3. Companies listed on the Neuer Markt were required to publish quarterly and annual financial statements that comply with international accounting standards (IAS or U.S.-GAAP). This information is readily available in a standardised form, guaranteeing comparability across different companies. According to the regulatory framework of the Neuer Markt, quarterly financial statements did not need to be audited. 4. Such studies mainly investigate and discuss the value-relevance of financial data in contrast to other firm-specific information. Empirical evidence indicates that financial information has lost some of its significance in explaining movements in stock prices. See Francis and Schipper (1999), Ittner and Larcker (1998), Lee (1999), Lev (2000), and Shevlin (1996). 5. In this context it is especially important to note that the majority of Internet companies considered in this study only existed for a relatively short period and in a rapid growth environment. This significant difference in comparison with well-established companies operating in mature markets will be important when discussing and interpreting the results of the empirical analysis. A generalisation of the findings is not feasible for this reason. Furthermore, it should be noted that we do not intend to question the usefulness of financial statements. 6. In the United States, three companies provided such information. These were Nielsen NetRatings, Media Metrix and PCData, whereas in Germany a joint venture between the Gesellschaft f¨ur Konsumforschung (GfK) and Media Metrix Europe was the only provider. Nielsen NetRatings began observing private Internet use in the summer of 2000. 7. It is important to note, however, that some of the heavy losses are due to high marketing expenses. 8. A strict definition of cash flow is lacking, because of differences in computing. To establish comparability across the sample, this study relies on the measure of operating cash
174
ANDREAS TRAUTWEIN AND SVEN VORSTIUS
flow according to IAS and U.S.-GAAP. See Baetge (1998, pp. 316–341) and Pellens (1999, pp. 303, 471–476). 9. The winner-takes-all business model states that spending on intangibles leads to increasing profitability returns-to-scale. See Hand (2001a, p. 2). 10. This classification serves as a basis for the Internet Stock Index (ISDEX), published by internet.com. According to the definition, firms are further subdivided into seven major categories, which are as follows: (1) E-tailer and E-commerce; (2) Software; (3) Enabler; (4) Security; (5) Content Provider and Portals; (6) High-Speed and Infrastructure; and (7) Access and Internet-Service-Provider. See Hand (2001b, p. 302). 11. The E-commerce group includes the following segments: (1) E-tailer and E-commerce; (2) Content Provider and Portals; and (3) Access and Internet-ServiceProvider, while the Enabler group contains the following segments: (1) Software; (2) Enabler; (3) Security; and (4) High-Speed and Infrastructure. 12. The information is based on different accounting standards: approximately a quarter of our sample companies published financial statements according to IAS, while the remainder complied with U.S.-GAAP. Note that the number of observations depends on the availability of data, resulting in differences in sample size across variables. Despite Deutsche B¨orse AG’s extensive requirements with respect to financial reporting and the timeliness of such reports for companies listed on the Neuer Markt, companies did not always follow these rules in practice. See Tables A3 and A4 in the Appendix for descriptive statistics on the financial variables and Tables A7 and A8 for a correlation matrix. 13. Information on web-traffic was generated in a standardised procedure and available in a monthly report at a certain cost. Therefore, it can safely be assumed that in a (semi-strong) efficient market such information could easily be incorporated into stock prices as analysts and investors do have access. 14. See Table A1 in the Appendix for information on companies included in the analysis. 15. See Table A6 in the Appendix for a correlation matrix and Table A3 for descriptive statistics on the web-metrics. 16. Testing a two-week window surrounding this date did not show any deviation in the results. 17. In addition, such models allow statements about the nature of the relationship between the dependent and the independent variables, i.e. whether the relationship is linear, concave or convex. The coefficient can be interpreted as elasticity and thus determines the degree of non-linearity between the dependent and the independent variables. 18. The logarithmic function is only defined for positive values. Thus earnings as well as cash flows had to be linearly transformed prior to the transformation by adding an amount in excess of the largest negative value in the data. 19. The correlations between the independent variables are tested regarding possible multicollinearity. Applying the variance inflation factor as a measure to test for multicollinearity (see Gujarati, 1995, p. 328), we find that only reach and unique visitors show exposure. This has to be taken into account when interpreting multivariate regressions that include both web-metrics. 20. Alternatively, a difference or a return model could be employed. For a discussion of various methods see Easton (1999, p. 402) and Kothari and Zimmerman (1995, p. 165). 21. Eviews statistic software was used to estimate all regressions. All tests apply a White-corrected covariance matrix to minimise heteroscedasticity, see White (1980).
The Role of Accounting Data and Web-Traffic
175
22. As statistical problems due to possible omitted variables and multicollinearity may be present, the results should be interpreted with caution. The same applies to comparing univariate and multivariate regression models.
ACKNOWLEDGMENTS We would like to thank Christian Leuz, J¨urgen Weber, Barbara Weißenberger and Peter Roosenboom for valuable comments and suggestions. Additionally, we would like to thank seminar participants at the 2001 congress of the European Accounting Association in Athens and an anonymous referee for helpful feedback. The financial support from the Schmalenbach-Gesellschaft f¨ur Betriebswirtschaft and Deloitte & Touche Berlin is much appreciated. We are also grateful to Media Metrix Europe for supplying data on web-traffic. Any remaining errors are our responsibility.
REFERENCES Baetge, J. (1998). Bilanzanalyse. D¨usseldorf: IDW-Verlag. Barth, M., Beaver, W., & Landsman, W. (2001). The relevance of the value-relevance literature for financial accounting standard setting: Another view. Journal of Accounting and Economics, 31, 77–104. D’Arcy, A., & Leuz, C. (2000). Rechnungslegung am Neuen Markt – Eine Bestandsaufnahme. Der Betrieb, 8, 385–391. Demers, E., & Lev, B. (2001). A rude awakening: Internet shakeout in 2000. Review of Accounting Studies, 6, 331–359. Easton, P. D. (1999). Security returns and the value relevance of accounting data. Accounting Horizons, 13, 399–412. Francis, J., & Schipper, K. (1999). Have financial statements lost their relevance? Journal of Accounting Research, 37, 319–352. Gujarati, D. (1995). Basic econometrics. New York: McGraw-Hill. Hand, J. (2000). Profits, losses and the non-linear pricing of Internet stocks. Working Paper, Kenan-Flagler Business School, UNC Chapel Hill. Hand, J. (2001a). Evidence on the winner-takes-all business model: The profitability returns-to-scale of expenditures on intangibles made by U.S. Internet firms, 1995–2001. Working Paper, Kenan-Flagler Business School, UNC Chapel Hill. Hand, J. (2001b). The role of book income, web traffic, and supply and demand in the pricing of U.S. internet stocks. European Finance Review, 5, 295–317. Hand, J. (2003). Profits, losses and the non-linear pricing of Internet stocks. In: J. Hand & B. Lev (Eds), Intangible Assets – Values, Measures, and Risks (pp. 248–268). Oxford: Oxford University Press. Hayn, C. (1995). The information content of losses. Journal of Accounting and Economics, 20, 125–153. Holthausen, R., & Watts, R. (2001). The relevance of the value relevance literature for financial accounting standard setting. Journal of Accounting and Economics, 31, 3–75.
176
ANDREAS TRAUTWEIN AND SVEN VORSTIUS
Ittner, C., & Larcker, D. (1998). Are non-financial measures leading indicators of financial performance? An analysis of customer satisfaction. Journal of Accounting Research, 36, 1–36. Kothari, S. P., & Zimmerman, J. L. (1995). Price and return models. Journal of Accounting and Economics, 20, 155–192. Laderman, J., & Smith, G. (1998). Internet stocks: What’s their real worth. Business Week, December 14, 120–122. Lee, C. (1999). Accounting based valuation: Impact on business practices and research. Accounting Horizons, 13, 413–425. Lev, B. (2000). New accounting for the New Economy. Working Paper, Stern Business School, New York University. Lev, B., & Sougiannis, T. (1996). The capitalization, amortization, and value-relevance of R&D. Journal of Accounting and Economics, 26, 107–138. Lev, B., & Thiagarajan, S. R. (1993). Fundamental information analysis. Journal of Accounting Research, 31, 190–215. Pellens, B. (1999). Internationale Rechnungslegung. Stuttgart: Sch¨affer-Poeschel Verlag. Rajgopal, S., Kotha, S., & Venkatchalam, M. (2003). The value relevance of network advantages: The case of E-commerce firms. Journal of Accounting Research, 41, 135–162. Schwartz, E., & Moon, M. (2000). Rational pricing of Internet companies. Financial Analyst Journal, 56, 62–75. Shevlin, T. (1996). The value-relevance of nonfinancial information: A discussion. Journal of Accounting and Economics, 22, 31–42. Trueman, B., Wong, M. H. F., & Zhang, X.-J. (2000). The eyeballs have it: Searching for the value in Internet stocks. Journal of Accounting Research, 38, 137–162. White, H. (1980). A heteroscedasticity consistent covariance matrix estimator and a direct test for heteroscedasticity. Econometrica, 48, 817–838.
177
The Role of Accounting Data and Web-Traffic
APPENDIX Table A1. Companies Included in the E-Commerce Sample and Availability of Data. No. EC 1 EC 2 EC 3 EC 4 EC 5 EC 6 EC 7 EC 8 EC 9 EC 10 EC 11 EC 12 EC 13 EC 14 EC 15 EC 16 EC 17 EC 18 EC 19 EC 20 EC 21 EC 22 EC 23
Company
FI
artnet.com AG buch.de internetstores AG buecher.de AG ConSors Discount-Broker AG CTS EVENTIM AG Direkt Anlage Bank AG ebookers.com PLC Endemann!! Internet AG ENTRIUM DIRECT BANKERS AG fluxx.com AG FortuneCity.com Inc. freenet.de AG Gigabell AG Jobs & Adverts AG Lycos Europe N. V. Musicmusicmusic inc. OnVista AG ricardo.de Aktiengesellschaft TOMORROW Internet AG T-Online International AG Travel24.com AG United Internet AG WEB.DE AG
√ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √
Note: FI: Financial Information, WM: Web-Metrics.
WM ∅ √ √ √ ∅ √ √ ∅ √ ∅ √ √ ∅ √ √ ∅ √ √ √ √ ∅ √ √
178
ANDREAS TRAUTWEIN AND SVEN VORSTIUS
Table A2. Companies Included in the Enabler Sample. No.
Company
EN 1 EN 2 EN 3 EN 4 EN 5 EN 6 EN 7 EN 8 EN 9 EN 10 EN 11 EN 12 EN 13 EN 14 EN 15 EN 16 EN 17 EN 18 EN 19 EN 20 EN 21 EN 22 EN 23 EN 24 EN 25 EN 26 EN 27 EN 28 EN 29 EN 30 EN 31 EN 32 EN 33 EN 34 EN 35 EN 36 EN 37
ABIT AG AdLINK Internet Media AG Adori AG antwerpes ag ARTICON Information Systems AG BroadVision, Inc. BROKAT Infosystems AG Concept! AG Cybernet Internet Services International Inc. DataDesign AG DCI Database for Commerce and Industry AG digital advertising AG FANTASTIC CORP. Gauss Interprise AG GEDYS Internet Products AG GFT Technologies AG I-D Media AG i-FAO AG INFOMATEC AG Integra S. A. InternetMediaHouse.com AG Internolix AG INTERSHOP COMMUNICATIONS AG ISION Internet AG Kabel New Media AG Met@box AG net AG Netlife AG Openshop Holding AG Pixelpark AG PopNet Internet AG QS Communications AG SinnerSchrader AG teamwork information management AG TRIA software AG Trintech Group PLC WWL Internet AG
179
The Role of Accounting Data and Web-Traffic
Table A3. Descriptive Statistics for the E-Commerce Sample. Variable
N
Min
Max
Mean
Median
Std.
MVa
54 54 32 53 49 87 87 66 87 66 66
39.38 0.14 0.05 −27.83 −11.69 51.48 0.08 0.29 1.0 1.6 1.1
46,610 173.80 31.60 4.64 12.33 46,200 4.89 73.81 57.1 25.8 6.2
1,351 10.77 3.70 −3.76 −2.36 2,625 0.63 13.52 7.8 6.3 3.4
263 3.62 1.94 −2.64 −2.21 744 0.23 4.69 3.5 5.6 3.6
6,322 25.04 5.69 5.11 4.12 7,496 0.97 19.98 11.1 3.8 1.5
TSa SMEa EBITa CFa MVa,e UVb,e PIb,e REc,e STd,e CLd,e
Note: Std: Standard Deviation, MV: Market Value, TS: Total Sales, SME: Sales and Marketing Expenses, EBIT: Earnings Before Interest and Taxes, CF: Cash Flow, UV: Unique Visitors, PI: Page Impressions, RE: Reach, ST: Stickiness, and CL: Customer Loyalty. a Million Euro. b Million. c Percent. d Minutes. e Adjusted E-commerce sample (see Table A1).
Table A4. Descriptive Statistics for the Enabler Sample. Variable MV TS RND EBIT CF
N
Min
Max
Mean
Median
Std
106 105 48 106 104
60.00 0.52 0.14 −23.73 −55.83
13,310 67.45 9.72 10.08 31.86
1,051 7.41 1.84 −2.94 −2.98
389 4.00 1.31 −1.80 −2.02
2,138 9.22 1.82 5.41 8.26
Note: Std: Standard Deviation; MV: Market Value, TS: Total Sales, RND: Research and Development Expenses, EBIT: Earnings Before Interest and Taxes, and CF: Cash Flow. Figures in Million Euro.
Table A5. Removed Outliers and Unusual Observations. Outlier and Unusual Observations Company
Data Set
Lycos Integra Trintech
EBIT CF CF
Note: EBIT: Earnings Before Interest and Taxes, CF: Cash Flow.
Quarter 1/00 1/00 3/99
180
ANDREAS TRAUTWEIN AND SVEN VORSTIUS
Table A6. Correlation Matrix Web-Metrics (E-Commerce Sample). Variable Market Value (MV) Page Impressions (PI) Unique Visitors (UV) Reach (RE) Stickiness (ST) Customer Loyalty (CL)
MV
PI
UV
RE
ST
CL
1 0.642 0.633 0.637 0.076 0.792
1 0.895 0.857 0.044 0.760
1 0.977 −0.223 0.548
1 −0.264 0.522
1 0.158
1
Table A7. Correlation Matrix Financial Information (E-Commerce Sample). Variable Market Value (MV) Total Sales (TS) Earnings (EBIT) Sales & Marketing Exp. (SME) Operating Cash Flow (CF)
MV
TS
EBIT
SME
CF
1 0.576 −0.176 0.518 0.053
1 −0.130 0.608 −0.226
1 −0.620 −0.094
1 0.120
1
Table A8. Correlation Matrix Financial Information (Enabler Sample). Variable Market Value (MV) Total Sales (TS) Earnings (EBIT) Operating Cash Flow (CF) Research & Develop. Exp. (RND)
MV
TS
EBIT
SME
CF
1 0.694 −0.238 0.174 0.423
1 −0.195 0.073 0.494
1 0.277 −0.290
1 0.105
1
THE EXPIRATION OF MANDATORY AND VOLUNTARY IPO LOCK-UP PROVISIONS – EMPIRICAL EVIDENCE FROM GERMANY’S NEUER MARKT Eric Nowak ABSTRACT This chapter explores the stock price impact of expirations of lock-up provisions that prevent insiders from selling their shares after the Initial Public Offering (IPO). We examine 172 lock-up expirations of 142 IPOs floated on Germany’s Neuer Markt. We detect significant negative abnormal returns and a 25% increase in trading volume surrounding lock-up expiration. The negative abnormal returns are larger for firms with high volatility; superior performance after the IPO, low free float, and venture capital financed firms. The negative price reaction is significantly stronger for the expiration of voluntary lock-up agreements than for mandatory prohibitions of disposal.
1. INTRODUCTION On January 11, 2001, the German Schutzgemeinschaft der Kleinaktion¨are – an association for the protection of the interests of small shareholders – announced a warning that the lock-up provision of Letsbuyit.com, an E-commerce firm, would expire on January 21. This warning was published in all major newspapers, The Rise and Fall of Europe’s New Stock Markets Advances in Financial Economics, Volume 10, 181–200 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3732/doi:10.1016/S1569-3732(04)10008-X
181
182
ERIC NOWAK
expressing concern about the fact that most of the incumbent shareholders of the trouble-shaken firm would probably sell their shares upon expiration at the prevailing market price of d0.30, given that some of them had an initial investment per share of only d0.01.1 Although this information was public ex ante, the share price of Letsbuyit.com declined by almost 50% on the first trading day after the lock-up expiration, and the trading volume was the highest for all shares on the German XETRA stock exchange system on that day. This chapter explores the stock price impact of expirations of lock-up provisions that prevent insiders from selling their shares after the Initial Public Offering (IPO). We examine 172 lock-up expirations of 142 IPOs floated on Germany’s Neuer Markt. This chapter provides two contributions to the literature on IPOs and lock-up provisions. First, it documents further evidence on downward-sloping demand curves and costly arbitrage for a capital market outside the United States. We find statistically significant negative abnormal returns and a 25% increase in trading volume surrounding lock-up expiration. The negative abnormal returns are larger for firms with high volatility, superior performance between the IPO date and the lock-up expiration date, and low free float. Second, and more important, we can differentiate between the effects of mandatory lock-up provisions and the U.S.-type private lock-up agreements between issuers and underwriters. The latter we refer to as “voluntary” lock-up agreements that serve as a commitment device to reduce information asymmetry at the IPO. We show that the average negative price reaction is significantly stronger for the expiration of voluntary lock-up agreements than for mandatory prohibitions of disposal. Furthermore, we find that venture-capital financed firms experience more negative abnormal returns than non-venture backed firms, on average. The remainder of this chapter is organized as follows: Section 2 reviews other studies that have investigated lock-up agreements. Section 3 describes the nature of mandatory and voluntary lock-up provisions in Germany. Section 4 presents the data description and the sample selection. Section 5 presents the event study methodology. Results on abnormal returns surrounding the time of the lock-up expiration for both types of provisions and on abnormal volume are analyzed in Section 6. Section 7 investigates the relation between certain firm characteristics and the price reaction. Section 8 concludes.
2. LITERATURE REVIEW Field and Hanka (2001) examine the expiration of IPO share lockups in the United States. They find a significantly negative three-day abnormal return of minus 1.5% and a permanent 40% increase in trading volume upon expiration of the
The Expiration of Mandatory and Voluntary IPO Lock-Up Provisions
183
lock-up period for 1,948 firms in the period 1988–1997. In another study, Keasler (2001) finds negative abnormal returns prior to the lock-up releases and shows that unrestricted investors liquidate positions prior to the scheduled lock-up release. He finds that negative abnormal returns are more robust for firms that are not influenced by SEC Rule 144 than for firms that are. Cao et al. (2004) test the hypothesis that insider trading impairs market liquidity by analyzing intraday trades and quotes around 1,497 IPO lock-up expirations in the period 1995–1999. They find that, while lock-up expirations are associated with considerable insider trading activity for some IPO firms, they have little effect on effective spreads. Thus, they argue that blockholding insider traders can enter a market from which they had previously been absent, and substantially change trading volume and share price without impairing market liquidity. Aggarwal et al. (2002) develop a model in which managers strategically underprice IPOs to maximize personal wealth from selling shares at lock-up expiration. They test the model on a sample of IPOs in the 1990s and find – consistent with their model – that higher ownership by managers is positively correlated with first-day underpricing and underpricing is positively correlated with research coverage. Finally, research coverage is positively correlated with stock returns and insider selling at the lock-up expiration. Brav and Gompers (2003) focus on the role of lock-ups as a commitment device to alleviate moral hazard problems in IPOs. They find that investment banks impose longer lock-ups on their IPO firms, when moral hazard in the aftermarket is higher. On the other hand, they show that venture-backed firms and firms going public with high-quality underwriters are more likely to have early releases of insider lock-up restrictions. Ofek and Richardson (2000) investigate volume and price patterns when the lock-up period ends, and document that there is a 3% drop in the stock price, and a 40% increase in volume. They argue that the evidence is consistent with a downward sloping demand curve for shares. Harper et al. (2004) look at follow-on offerings and how these alter firm value above and beyond the typical lock-up effects, and whether the effects are conditioned by firm-specific variables. They find that follow-on offerings elicit an average market response of minus 3.21% over a three-day period surrounding the filling date. In their sample, the offerings experience adverse effects as of lock-up expiration that are about 3.75% worse than other IPOs, after considering other factors. Overall, their research suggests that follow-on offerings benefit some insiders who can circumvent the lock-up expiration date, at the expense of other investors. There are only a few studies that examine capital markets other than the United States. Surprisingly, to the opposite of studies on U.S. data, Espenlaub et al. (2001) do not find significant abnormal returns around the expiry for a sample of
184
ERIC NOWAK
IPO lock-up agreements in the United Kingdom. Goergen et al. (2004) compare the characteristics of lock-up agreements in German and French firms that went public on the Neuer Markt and the Nouveau March´e during the years 1996/1997 to 2000. They find that the level of uncertainty about the firm’s prospects and venture backing have a major influence on the characteristics of the lock-up contracts. In addition, shareholder characteristics explain the diversity of contracts that exist within the same firm. However, their paper does not look at price reactions upon the lock-up expiration day. This chapter aims to fill this gap.
3. MANDATORY AND VOLUNTARY LOCK-UP PROVISIONS IN GERMANY In March 1997, Deutsche B¨orse established the Neuer Markt, a trading segment for innovative growth stocks, who had to meet international standards of transparency and publicity.2 Trading on the Neuer Markt took place in the Regulated Unofficial Market (Freiverkehr) under private law, but all companies admitted to the Neuer Markt also had to be admitted to the Regulated Market (Geregelter Markt). Organized under private law, Deutsche B¨orse formally imposed strict admission and disclosure requirements for the Neuer Markt. In theory, the legal framework of the Neuer Market was comparable to and, in some respects, even stricter than the admission requirements and post-listing duties under the SEC regime in the United States. In practice, however, the system has been hampered by inconsistent enforcement by Deutsche B¨orse. The Neuer Markt rules were purely private agreements between Deutsche B¨orse and issuers (who were also its customers). The German stock market regulator – the BAWe now BAFin – did not have a mandate to supervise these. A total of 342 companies had listed on the Neuer Markt by July 2001. Although a number of other European growth markets opened,3 these had been significantly less popular with issuers. The Neuer Markt quickly became Europe’s biggest exchange for securities of innovative growth companies. In the end, the Neuer Markt was severely hit by the collapse of share prices following the bursting of the bubble and was finally shut down, because of the irreparable loss in investor confidence. Deutsche B¨orse required all issuers to sign and comply with the so-called “Undertaking Concerning the Prohibition on Disposal,” as stated in the Neuer Markt Rules and Regulations: Prohibition on Disposal (1) The issuer shall be obligated, subject to the applicable provisions of the national corporate law, to refrain, within a period of six months from the date of admission of the shares to the Neuer Markt, from offering or selling shares directly or indirectly, or announcing such action,
The Expiration of Mandatory and Voluntary IPO Lock-Up Provisions
185
or taking other measures economically equivalent to a sale. Further, the issuer shall inform Deutsche B¨orse without delay should it become aware of any factors indicating a breach of the prohibition on disposal on the part of an existing shareholder (Part 2, 2.2).
The prohibition of disposal, although legally only a private contract between the issuer and Deutsche B¨orse, was effectively a mandatory lock-up rule, since it was a listing requirement applying to all firms on the Neuer Markt and (at least in theory but less so in practice) enforceable by law. Furthermore, a number of issuers stated in the offering prospectus that their shareholders had agreed not to sell shares for a longer period without the consent of the underwriter under a voluntary lock-up agreement. These voluntary lock-up agreements were not mandated by the stock exchange; hence they could only be enforced if the underwriter undertook legal actions in case of deviation (which basically never happened). Typically, while the mandatory prohibition applied to all existing shareholders holding stock before the offering, only management and the largest incumbent shareholders were locked by a voluntary non-selling agreement. For example, while usually small incumbent shareholders and venture capitalists were allowed to sell six months after the IPO, the founding members and/or the top management of the firm often agreed to lock their shares for an additional period of six-to-30-months. Table 1 gives an overview on the insider trading regulation rules that applied to German firms at the time of the Neuer Markt, as compared to those for U.S. IPO firms. Mandatory lock-up rules exist only in Germany, whereas there are more general disclosure rules and restrictions concerning insider sales in the United States.
Table 1. Insider Trading Regulation for IPOs on Germany’s Neuer Markt versus the United States.
Lock-up provisions Mandatory rules
Voluntary agreements
Legal insider selling restrictions and disclosure rules
German Rules
U.S. Rules
Prohibition on disposal (Paragraph 7.2.9 Rules and Regulations Neuer Markt) Complementary lock-up contracts between underwriter and issuer Pool contracts among incumbent shareholders of the firm Section 21 WpHG (German Securities Law) Section 13 WpHG
Non-existent
Private lock-up contracts between underwriter and issuer
Rule 144 Section 16 Securities Exchange Act (SEA)
186
ERIC NOWAK
In this paper we are interested in mandatory and voluntary lock-up provisions. Technically, the two types of lock-up provisions are different in nature. However, given the severe adverse selection problem in the going public process, both serve as a commitment device to induce the public to buy shares at the offering (Brav & Gompers, 2003). The mandatory prohibition of disposal was to signal to the public that the Deutsche B¨orse would be committed to enforce this device. The (second) voluntary lock-up agreement signals not only the commitment of the issuing firm, but may also reflect the quality of the underwriter. Venture capitalists typically do not lock their investments for more than the mandatory six-month period in Germany. Their business model forces them to unwind their equity stakes in portfolio investments that successfully go public. On the one hand, one would therefore expect that venture backed firms have a larger number of shares coming to market at lock-up expiration (Brav & Gompers, 2003). On the other hand, venture capitalists may want to maintain a reputation of financing high-quality IPOs. Thus, they could force management of their portfolio firms to agree upon a further voluntary lock-up provision, and they may want to retain their own shares for signaling reasons. Or, as Barry et al. (1990, p. 461) put it: “By retaining their share ownership, the venture capitalists can provide assurance of continued monitoring and can credibly signal their belief in the firms’ prospects.” Both arguments have conflicting implications for abnormal price reactions and the contractual structure of lock-up provisions.
4. DATA DESCRIPTION AND SAMPLE SELECTION We investigate all IPOs on the Neuer Markt segment from its inception in 1997 until October 1999. For these 194 firms, we identify all lock-up events and hand-collect the dates from the offering prospectuses. In some cases we have to contact the issuers to clarify the exact date. One firm drops out of the sample, because it has a short lock-up of only three months. Another 26 firms Table 2. Sample Selection. Initial public offerings and first trading of shares on the Neuer Markt from 03/97 to 12/99 − Firm with lock-up less 6 months − Data restrictions − Confounding news one week before and after the event = Final sample of firms + Of which have complementary lock-up agreement = Final sample of events
194 firms 1 firm 26 firms 25 firms 142 firms 30 firms 172 events
187
The Expiration of Mandatory and Voluntary IPO Lock-Up Provisions
Table 3. Descriptive Statistics. Min
Volatility Post-IPO performance (log) Free float in percent Trading volume Underpricing (%) Market value of equity (DM millions)
25th Median 75th Percentile Percentile
1.62 −339.14
3.33 −83.16
4.00 −30.28
4.80 48.65
18.60 0.10 −14.11 5.62
25.03 0.48 4.24 32.76
30.50 0.83 23.47 56.50
40.00 1.44 57.01 110.68
Max
Mean
Standard Deviation
10.09 4.37 240.73 −24.46
1.65 100.76
100.00 6.18 140.65 876.00
16.72 1.07 37.70 129.66
36.39 1.16 34.70 97.39
Note: Volatility is the standard deviation in the estimation period (between IPO and unlock day). PostIPO performance is the log of the total return from the IPO until the unlock day. Free float is taken as reported by Deutsche B¨orse and checked against the offering prospectuses. Trading volume is order book turnover as reported by Deutsche B¨orse (excluding OTC trades). Underpricing is the first day return against the offering price. Market value of equity is number of shares issued multiplied by the issue price, as reported by Deutsche B¨orse.
are excluded from the sample because we could not retrieve price data. We control for confounding news one week before and after the event day of the lock-up expiration. In order to identify an information-clean event, 25 firms with confounding news (e.g. earnings announcements) one week before and after the
Fig. 1. Length of Lock-Up Agreements. Note: (∗ ) Other Lock-up-agreements involve three firms, of which two have a 6 plus 3 month lock-up, and one company that shortened the length to three months.
188
ERIC NOWAK
event day of the lock-up expiration are eliminated from the sample. The remaining sample consists of 142 IPO firms floated on the Neuer Markt. Of those 142 firms 30 have an additional voluntary lock-up agreement as stated in the offering prospectus. The final full sample therefore consists of 172 lock-up expiration events. Tables 2 and 3 provide descriptive statistics for 142 sample IPO firms. For the empirical analysis, we could take into account stock market data until June 30, 2000. The event window ends 30 trading days subsequent to the IPO. Daily stock price and trading volume data are directly provided by the Deutsche B¨orse, and are adjusted for dividend payments and capital changes. Information on free float and venture capital financing are obtained directly from prospectuses. Figure 1 shows the distribution of lock-up length for the sample of lock-up provisions. Most of the IPO firms do not have a voluntary lock-up agreement complementing the prohibition of disposal. The majority is only locked for six months.
5. EVENT STUDY METHODOLOGY We calculate abnormal returns for each IPO over the event window (t−10 ; t30 ) as the difference between the actual return and the expected return. We benchmark the expected return by market returns as well as by estimating a market model, using a simple OLS regression. The estimation window for the market model is the 90-day period (t−100 ; t−11 ). We employ the value-weighted NEMAX All-Share Performance Index as proxy for the market return. Thus the abnormal return (ARit ) is calculated as: ARit = R it − [␣i + i E(R mt )]
(1)
with Rit actual return of stock i at time t, E(Rmt ) expected return of the (NEMAX) market at time t, ␣i constant return component, i sensitivity of firm i’s stock returns to the market index return. For testing the statistical significance of the abnormal returns we employ a set of parametric as well as non-parametric tests. We have chosen the simple t-test and the modified t-test proposed by Brown and Warner (1985). In order to check for the influence of non-normal distribution of residuals in small samples, we apply the non-parametric rank test of Corrado (1989). We also compute a potentially more powerful test proposed by B¨ohmer et al. (1991) that takes heteroskedasticity into account, but explicitly employs information from the estimation period. Calculation of abnormal trading volume is done using a simple constant mean methodology. First, we calculate the average trading volume for each sample firm
The Expiration of Mandatory and Voluntary IPO Lock-Up Provisions
189
in the estimation period. We then compute an abnormal volume index (AVI) as follows: V it − 1 with AVIit = Vi
1 Vi = V it , 90 t t−10
(2)
−100
where V it is shares traded in firm i at time t, and V i is the average trading volume in the estimation period. Finally, the abnormal volume index is averaged across firms in the sample: 1 (1 + AVit ) N N
AVit =
(3)
i=1
6. EVENT STUDY RESULTS This section presents the event study results. Since the date of the lock-up expiration is common knowledge at the time of the IPO, we do not expect to find abnormal returns surrounding the event day, assuming that markets are informationally efficient. Figure 2 presents a time series plot of the average cumulative abnormal return and shows that the share price declines sharply around the lock-up expiration day. For the period from ten days before the unlock day through 30 days after, the cumulative abnormal return is significantly negative at −7.95%. Cumulative
Fig. 2. Cumulative Abnormal Returns Around Unlock Day.
190
ERIC NOWAK
Table 4. Event Study Results: Full Sample. All Events (N = 172) Event Window
CAR (%)
% Negative
Median CAR
t−10 to t−1 t−2 t−1 t0 t1 t2 t−2 to t2 t−1 to t0 t−10 to t10 t1 to t15 t−1 to t30
−0.84 0.05 −0.82** −0.19 −0.03 −0.21 −1.18rr −1.01* −3.76** −3.30* −7.95***
51.10 54.65 59.88 59.30 55.81 56.98 57.33 59.59 56.20 55.47 56.78
−0.41 −0.51 −0.92 −0.61 −0.59 −0.72 −0.61 −0.61 −0.46 −0.45 −0.48
Denote significance of the simple t-test; ttt, tt, t denote significance of the Brown and Warner t-test; and rrr, rr, r denote significance of the non-parametric rank test according to Corrado. ∗ Significance of all test metrics at the 10% level. ∗∗ Significance of all test metrics at the 5% level. ∗∗∗ Significance of all test metrics at the 1% level.
Note:
TTT, TT, T
abnormal returns over various event windows are tabulated in Table 4. Sixty percent of the daily abnormal returns on the unlock day are negative. The results are robust to different specifications of event window, benchmark, calculation of abnormal returns, and the test statistic employed. Figure 3 and Table 5 and show the results for mandatory versus voluntary lock-up provisions. Both experience significantly negative abnormal returns on the unlock day. However, those stocks with a complementary lock-up expiration underperform the benchmark by more than 20%. Thus, there is a significantly negative abnormal return upon lock-up expiration, which is stronger for voluntary lock-up agreements. An explanation of this finding is that founding entrepreneurs, managers, and other corporate insiders are more likely to be subject to the longer voluntary lock-up period. These insiders are assumed to have higher equity stakes. When these insiders sell, more shares will therefore enter the market, on average, than at the first mandatory lock-up expiration that restricts other investors and friends and family from selling. However, since we have 30 complementary lock-up agreements versus 142 mandatory prohibitions of disposal, one should be careful when interpreting this result. Finally, since the cumulative abnormal return is still negative after 30 trading days following the unlock day, we can reject a price pressure hypothesis. This price pressure hypothesis would predict only a temporary effect (Scholes, 1972).
191
The Expiration of Mandatory and Voluntary IPO Lock-Up Provisions
Fig. 3. Cumulative Abnormal Returns Based on the Length of the Lock-Up Period.
Table 5. Cumulative Abnormal Returns Around Mandatory vs. Complementary Lock-Up Expirations. Event Window
t−10 to t−1 t−2 t−1 t0 t1 t2 t−2 to t2 t−1 to t0 t−10 to t10 t1 to t15 t−1 to t30
Mandatory Prohibition of Disposal (N = 142)
Complementary Lock-Up Agreements (N = 30)
CAR (%)
% Negative
Median CAR
CAR (%)
% Negative
Median CAR
−0.73 0.25 −0.54 0.00 0.00 −0.22 −0.52 −0.54 −3.36** −2.67T −5.35***
56.83 53.52 59.15 57.76 57.04 55.63 56.36 58.45 59.36 55.45 56.16
−0.45 −0.48 −0.82 −0.27 −0.55 −0.53 −0.51 −0.53 −0.47 −0.44 −0.44
−1.38 −0.88 −2.14*** −1.06 −0.14 −0.12 −4.34** −3.20** −5.62t −6.25** −20.24***
52.67 60.00 63.33 66.67 62.31 63.33 60.67 65.00 54.60 55.56 59.69
−0.29 −0.92 −1.88 −0.81 −0.06 −1.04 −0.99 −1.14 −0.40 −0.49 −0.74
Denote significance of the simple t-test; ttt, tt, t denote significance of the Brown and Warner t-test; and rrr, rr, r denote significance of the non-parametric rank test according to Corrado. ∗∗ Significance of all test metrics at the 5% level. ∗∗∗ Significance of all test metrics at the 1% level.
Note:
TTT, TT, T
192
ERIC NOWAK
Fig. 4. Abnormal Trading Volume Around Unlock Day.
Figure 4 plots the sample mean of the daily abnormal volume as defined in Eq. (2). Figure 4 shows that, for the full sample, volume increases temporarily to 25% above average on the day after the unlock day, and remains at that level throughout the event window. Thus, unlocking the shares of the incumbent shareholders seems to result in a permanent increase in trading volume.
7. CROSS-SECTIONAL DETERMINANTS OF ABNORMAL RETURNS Tables 6–11 present the mean and median abnormal returns for various subsamples, and Table 12 presents pooled OLS regressions of the cumulative abnormal return on several control variables. Table 6 shows that firms with a high standard deviation in the estimation period (between IPO and unlock day) experience significantly negative abnormal returns of −20.41% in the thirty-day event window. On the other hand, firms with volatility below the median experience significantly positive abnormal returns of +9.72% during the same time interval. Although the causality is not clear-cut, this supports the risk-diversification hypothesis proposed by Meulbroek (2001) that states that insiders of risky high-growth firms have to reduce their stakes in order to decrease the suboptimal risk inherent in their portfolios. The significantly negative slope of volatility in the cross-sectional regressions supports this risk diversification argument (Fig. 5).
193
The Expiration of Mandatory and Voluntary IPO Lock-Up Provisions
Table 6. Cumulative Abnormal Returns Partitioned by Residual Standard Deviation. Event Window
t−10 to t−1 t−2 t−1 t0 t1 t2 t−2 to t2 t−1 to t0 t−10 to t10 t1 to t15 t−1 to t30
Firms with Volatility < Median (N = 71)
Firms with Volatility > Median (N = 71)
CAR (%)
% Negative
Median CAR
CAR (%)
% Negative
Median CAR
0.80 −0.30 −0.23 −0.26 0.73** 0.25 0.20 −0.48 3.54** 4.79*** 9.72***
58.03 60.56 59.15 63.38 50.70 52.65 56.90 61.27 54.93 51.17 52.99
−0.41 −0.57 −0.50 −0.24 −0.15 −0.03 −0.34 −0.27 −0.40 −0.08 −0.22
−2.25 0.80 −0.85 0.25 −0.74t −0.70 −1.24 −0.60 −10.26*** −10.13*** −20.41***
55.27 46.48 59.15 52.11 63.38 60.56 56.34 55.63 58.15 59.72 59.33
−0.56 −0.76 −1.53 −0.59 −0.87 −1.39 −0.94 −1.06 −0.56 −0.86 −0.81
TTT, TT, T Denote significance of the simple t-test; ttt, tt, t denote significance of the Brown and Warner t-test; and rrr, rr, r denote significance of the non-parametric rank test according to Corrado. ∗∗ Significance of all test metrics at the 5% level. ∗∗∗ Significance of all test metrics at the 1% level.
Note:
Table 7. Cumulative Abnormal Returns Partitioned by Post-IPO Performance. Event Window
t−10 to t−1 t−2 t−1 t0 t1 t2 t−2 to t2 t−1 to t0 t−10 to t10 t1 to t15 t−1 to t30
Firms with Performance < Median (N = 71)
Firms with Performance > Median (N = 71)
CAR (%)
% Negative
Median CAR
CAR (%)
% Negative
Median CAR
0.60 −0.49 −0.24 0.82T −0.14 −0.33 −0.38 0.58 2.09t 2.02tt 4.26ttt
57.04 56.34 59.15 53.52 57.75 50.70 55.49 56.34 54.73 52.11 53.83
−0.37 −0.57 −0.71 −0.11 −0.21 −0.33 −0.40 −0.21 −0.34 −0.23 −0.28
−2.05 0.98TT −0.84T −0.82T 0.14 −0.12 −0.66 −1.66** −8.81*** −7.37*** −14.96***
56.62 50.70 59.15 61.97 56.34 60.56 57.75 60.56 58.35 58.78 58.49
−0.54 −0.68 −1.07 −0.93 −0.60 −0.93 −0.67 −1.00 −0.57 −0.80 −0.71
TTT, TT, T Denote significance of the simple t-test; ttt, tt, t denote significance of the Brown and Warner t-test; and rrr, rr, r denote significance of the non-parametric rank test according to Corrado. ∗∗ Significance of all test metrics at the 5% level. ∗∗∗ Significance of all test metrics at the 1% level.
Note:
194
ERIC NOWAK
Table 8. Cumulative Abnormal Returns Partitioned by Tradable Shares after the IPO (Free Float). Event Window
t−10 to t−1 t−2 t−1 t0 t1 t2 t−2 to t2 t−1 to t0 t−10 to t10 t1 to t15 t−1 to t30
Firms with Free Float < Median (N = 71)
Firms with Free Float > Median (N = 71)
CAR (%)
% Negative
Median CAR
−2.19 −0.06 −0.48 −0.31 −0.64r −0.48 −1.97rr −0.79 −8.95*** −7.39*** −13.19***
57.61 52.11 63.38 59.15 70.42 59.15 60.85 61.27 59.09 58.78 58.93
−0.49 −0.58 −1.07 −0.49 −1.02 −1.84 −0.92 −0.82 −0.49 −0.79 −0.72
CAR (%) 0.74 0.56 −0.60 0.31 0.64 0.03 0.94 −0.29 2.23t 2.05tt 2.49tt
% Negative
Median CAR
56.06 54.93 52.57 56.34 43.66 52.11 52.39 55.63 53.99 52.11 53.39
−0.43 −0.58 −0.18 −0.18 0.00 −0.31 −0.16 −0.18 −0.40 −0.15 −0.24
TTT, TT, T Denote significance of the simple t-test; ttt, tt, t denote significance of the Brown and Warner t-test; and rrr, rr, r denote significance of the non-parametric rank test according to Corrado. ∗∗∗ Significance of all test metrics at the 1% level.
Note:
Table 9. Cumulative Abnormal Returns Partitioned by Abnormal Trading Volume. Event Window
Firms with Abnormal Trading Volume < Median (N = 71) CAR (%)
t−10 to t−1 t−2 t−1 t0 t1 t2 t−2 to t2 t−1 to t0 t−10 to t10 t1 to t15 t−1 to t30
***
−5.27 −0.01 −1.04** −0.44 −1.41** 0.04 −2.87tt −1.49** −11.14*** −6.71* −12.98***
% Negative 59.01 50.70 63.38 57.75 61.97 53.52 57.46 60.56 57.88 56.34 56.90
Firms with Abnormal Trading Volume > Median (N = 71)
Median CAR ***
−0.56 −0.68* −1.07** −0.13 −0.57 −0.43 −0.56** −0.68** −0.56*** −0.46** −0.52***
CAR (%)
% Negative
Median CAR
3.82 0.51 −0.03 0.44 1.39 −0.49 1.83 0.40 4.41 1.36 2.29*
54.65 56.34 54.93 57.75 52.11 57.75 55.77 56.34 55.20 54.55 55.50
−0.29 −0.54 −0.16 −0.34 −0.27 −1.39** −0.50* −0.29 −0.31* −0.41* −0.45
TTT, TT, T Denote significance of the simple t-test; ttt, tt, t denote significance of the Brown and Warner t-test; and rrr, rr, r denote significance of the non-parametric rank test according to Corrado. ∗ Significance of all test metrics at the 10% level. ∗∗ Significance of all test metrics at the 5% level. ∗∗∗ Significance of all test metrics at the 1% level.
Note:
195
The Expiration of Mandatory and Voluntary IPO Lock-Up Provisions
Table 10. Cumulative Abnormal Returns of Venture-Backed versus Non-Venture-Backed IPOs. Event Window
t−10 to t−1 t−2 t−1 t0 t1 t2 t−2 to t2 t−1 to t0 t−10 to t10 t1 to t15 t−16 to t30 t−1 to t30
Non-Venture-Backed IPOs (N = 67)
Venture-Backed IPOs (N = 75)
CAR (%)
% Negative
Median CAR
CAR (%)
% Negative
Median CAR
2.46 1.53** 0.32 0.52 0.83 −0.15 3.05** 0.84 2.78 −0.81 0.54 3.36
55.67 46.27 55.22 56.27 53.73 59.70 54.32 55.97 55.37 54.43 55.82 55.18
−0.41 0.22 −0.56 −0.05 −0.50 −0.90 −0.33 −0.14 −0.36 −0.37 −0.36 −0.39
−3.57t −0.89* −1.30** −0.48 −0.74 −0.29 −3.71*** −1.78** −8.86*** −4.34* −4.52* −10.64***
57.87 60.00 62.67 58.67 61.57 56.38 58.67 60.67 57.59 56.36 57.24 57.04
−0.41 −1.04 −1.05 −0.59 −0.80 −0.31 −0.77 −0.62 −0.55 −0.47 −0.50 −0.49
denote significance of the simple t-test; ttt, tt, t denote significance of the Brown and Warner t-test; and rrr, rr, r denote significance of the non-parametric rank test according to Corrado. ∗ Significance of all test metrics at the 10% level. ∗∗ Significance of all test metrics at the 5% level. ∗∗∗ Significance of all test metrics at the 1% level.
Note:
TTT, TT, T
A similar line of reasoning applies to the post-IPO performance (until the unlock day). Those firms that experience superior returns prior to the lock-up expiration seem to have significantly negative abnormal price decreases, while those firms whose stocks performed with below median performance do not have any abnormal price reactions at all. The coefficient on the post-IPO returns until the unlock is significantly negative in the cross-sectional regressions. Investors seem to be more eager to sell when the price of their shares has risen than when it has fallen (O’Dean, 1998) (Fig. 6). If arbitrage were costly, proxies for the amount of shares that come to market at the expiration of the lock-up would be positively related to the price decline. Firms with a larger fraction of their shares locked up (i.e. firms with lower free float) would have a greater number of shares brought to market at the unlock day, and hence should experience larger price declines (Brav & Gompers, 2003). We find that firms with a free float below the median have significantly negative abnormal returns, while firms with high free float do not experience abnormal returns on the unlock day. However, free float is not significant in the cross-sectional regressions. This could be due to the fact that free float is only an imperfect proxy for the amount of shares that come to market on the unlock day (Fig. 7).
196
ERIC NOWAK
Table 11. Z-Statistics of the Mann-Whitney-U-Test. Event Window
t−10 to t10 t1 to t15 t−1 to t30
Volatility < Median 3.54** 4.79*** 9.72*** Post-IPO Performance < Median
t−10 to t10 t1 to t15 t−1 to t30
2.09t 2.02tt 4.26ttt Free Float < Median
t−10 to t10 t1 to t15 t−1 to t30
−8.95*** −7.39*** −13.19*** Abnormal Trading Volume < Median
t−10 to t10 t1 to t15 t−1 to t30
−11.14*** −6.71* −4.79tt −12.98***
Volatility > Median
Mann-WhitneyU-Test
−10.26*** −10.13*** −20.41***
−3.17*** −4.52*** −5.44***
Post-IPO Performance < Median
Mann-WhitneyU-Test
−8.81*** −7.37*** −14.96***
−3.15*** −3.10*** −3.19***
Free Float > Median
Mann-WhitneyU-test
2.23t 2.05tt 2.49tt Abnormal Trading Volume > Median 4.41* 1.36 0.52 2.29*
−2.35*** −2.50*** −2.10** Mann-WhitneyU-Test −2.40*** −2.80*** −1.95** −1.18
denote significance of the simple t-test; ttt, tt, t denote significance of the Brown and Warner t-test; and rrr, rr, r denote significance of the non-parametric rank test according to Corrado. ∗ Significance at the 10% level. ∗∗ Significance at the 5% level. ∗∗∗ Significance at the 1% level.
Note:
TTT, TT, T
Interestingly, for firms with abnormal trading volume larger than the median, we cannot find statistically significant negative abnormal returns. This is puzzling, since we would expect a positive relation between the price drop and trading volume, if the abnormal price reaction is driven by downward-sloping demand curves (Shleifer, 1986). Trading volume is not significant in the cross-sectional regressions. This finding can be attributed to either a very noisy proxy for trading volume or support for a liquidity story. For those stocks that have low liquidity, there is not sufficient demand to absorb the sell orders upon lock-up expiration. Then trading in these stocks “dries out,” which leads to the abnormal price decrease. However, we have no direct evidence to support his claim. Future
197
The Expiration of Mandatory and Voluntary IPO Lock-Up Provisions
Table 12. Regression Results for Cumulative Abnormal Returns Around Lock-Up Expirations. Dependent Variable CAR (t−1 to t30 ) Constant Volatility Return since IPO Free float Abnormal trading volume Underpricing Venture capital-backing Market value of equity Number of observations Adjusted R2 F-statistic
27.178* (1.922) −7.723*** (−4.125) −0.101*** (−3.015) 0.161 (0.786) 1.605 (0.635) −0.123 (−0.010) −4.910 (−0.902) −0.057** (−2.045) 134 0.326 6.495***
t-Statistics are in parentheses. ∗ Significance at the 10% level. ∗∗ Significance at the 5% level. ∗∗∗ Significance at the 1% level.
research would have to take examine better proxies for liquidity, such as bid-ask spreads, which are unavailable at the time of the investigation. One of the most intriguing results is the empirical observation that only venturebacked IPOs experience significantly negative abnormal returns of –10.64%, while
Fig. 5. Cumulative Abnormal Returns Based on Residual Standard Deviation.
198
ERIC NOWAK
Fig. 6. Cumulative Abnormal Returns Based on Post IPO-Performance.
non-venture backed firms experience positive if any abnormal returns. However, when included in the cross-sectional regressions, the venture-capital dummy is negative but not statistically different from zero. This may be due to the fact that we cannot take the exact amount of venture financing and the reputation of the venture capitalist into account. In any case, this puzzling result is similar to the finding of Brav and Gompers (2003) that the presence of venture capital investors is
Fig. 7. Cumulative Abnormal Returns Based on Fraction of Tradable Shares (Free Float) After the IPO.
The Expiration of Mandatory and Voluntary IPO Lock-Up Provisions
199
associated with larger price declines in U.S. IPOs. Their explanation for this result is that VC-backing means a greater number of shares brought to the market, since venture capitalists distribute shares to their investors upon the lock-up expiration date (who then sell these shares directly to the market, if they have an automatic selling policy). Thus, on average, a larger number of shares will come to the market for VC-backed companies than for other companies. These results have been confirmed by a study of Kraus and Burghof (2003) who show that venture-backed IPOs seem to perform significantly better before than after the expiration of lock-up periods in Germany.
8. CONCLUSIONS This chapter explores the stock price impact of expirations of lock-up provisions that prevent insiders from selling their shares after the initial public offering (IPO). We examine 172 lock-up expirations of 142 IPOs floated on Germany’s Neuer Markt. Using an event-study methodology. We detect statistically significant negative abnormal returns and a 25% increase in trading volume surrounding lock-up expiration. This adds further evidence to the existing U.S. studies showing downward-sloping demand curves and costly arbitrage (Scholes, 1972; Shleifer, 1986). For the first time, we can differentiate between the effects of mandatory lock-up provisions and the U.S.-type private lock-up agreements between issuers and underwriters. We refer to the latter as “voluntary” lock-up agreements. We show that the average negative price reaction is significantly stronger for the expiration of voluntary lock-up agreements than for mandatory prohibitions of disposal. We investigate several control variables and find that the negative abnormal returns are larger for firms with high volatility, superior performance after the IPO, and low free float. Furthermore, we find that venture-capital financed firms experience more negative abnormal returns than non-venture backed firms, on average. A puzzling finding is the fact that abnormal trading volume seems to be negatively related to the price decline upon lock-up expiration. Unfortunately, due to data restrictions, we can not differentiate between liquidity effects and information effects, and must leave the explanation of this result for future research.
NOTES 1. www.sdk.org. 2. Neuer Markt Rules and Regulations.
200
ERIC NOWAK
3. For example the Nouveau March´e (Paris), the Nuovo Mercato (Milan), the SWX New Market (Z¨urich), the Alternative Investment Market (AIM) in London and NASDAQ Europe in Brussels (EASDAQ).
REFERENCES Aggarwal, R. K., Krigman, L., & Womack, K. L. (2002). Strategic IPO underpricing, information momentum, and lock-up expiration selling. Journal of Financial Economics, 66, 105–137. Barry, C. B., Muscarella, C. J., Peavy, J. W., III, & Vetsuypens, M. R. (1990). The role of venture capital in the creation of public companies: Evidence from the going public process. Journal of Financial Economics, 27, 447–471. B¨ohmer, E., Musumeci, J. J., & Poulsen, A. B. (1991). Event study methodology under conditions of event-induced variance. Journal of Financial Economics, 30, 253–272. Brav, A., & Gompers, P. A. (2003). The role of lockups in initial public offerings. Review of Financial Studies, 16, 1–29. Brown, S. J., & Warner, J. B. (1985). Using daily stock returns: The case of event studies. Journal of Financial Economics, 14, 3–31. Cao, C., Field, L. C., & Hanka, G. (2004). Does insider trading impair market liquidity? Evidence from IPO lockup expirations. Journal of Financial and Quantitative Analysis, 39, 25–46. Corrado, C. J. (1989). A nonparametric test for abnormal security-price performance in event-studies. Journal of Financial Economics, 23, 358–395. Espenlaub, S., Goergen, M., & Khurshed, A. (2001). IPO lock-in agreements in the UK. Journal of Business Finance & Accounting, 28, 1235–1278. Field, L. C., & Hanka, G. (2001). The expiration of IPO share lockups. Journal of Finance, 56, 471–500. Goergen, M., Renneboog, L., & Kurshed, A. (2004). Lock-in agreements in French Nouveau Marche and German Neuer Markt IPOs. Working Paper, University of Manchester and Tilburg University. Harper, J. T., Johnston, J., & Madura, J. (2004). Follow-on offerings. Journal of Banking & Finance, 28, 251–264. Keasler, T. (2001). Underwriter lock-up releases, initial public offerings and after-market performance. Financial Review, 37, 1–20. Kraus, T., & Burghof, H.-P. (2003). Post-IPO performance and the exit of venture capitalists. University of Munich, Munich Business Research, Working Paper 2003–2001. Meulbroek, L. (2001). The efficiency of equity-linked compensation: Understanding the full cost of awarding executive stock options. Financial Management, 30, 5–44. Neuer Markt Rules and Regulations, 04/2000 (www.neuermarkt.de). O’Dean, T. (1998). Are investors reluctant to realize their losses? Journal of Finance, 53, 1175–1798. Ofek, E., & Richardson, M. (2000). Large the IPO lock-up period: Implications for market efficiency and downward sloping demand curves. Working Paper, Stern School of Business. Scholes, M. (1972). The market for securities: Substitution vs. price pressure and the effects of information on share prices. Journal of Business, 45, 179–211. Shleifer, A. (1986). Do demand curves for stocks slope down? Journal of Finance, 41, 579–590.
UNDERPRICING OF VENTURE-BACKED AND NON VENTURE-BACKED IPOS: GERMANY’S NEUER MARKT Stefanie A. Franzke ABSTRACT This chapter investigates whether non venture-backed, venture-backed and bridge financed companies going public on Germany’s Neuer Markt differ with regard to issuer characteristics, balance sheet data or offering characteristics. Moreover, this chapter contributes to the underpricing literature by focusing on the role of venture capitalists and underwriters in certifying the quality of a company. Companies backed by a prestigious venture capitalist and/or underwritten by a top bank are expected to show less underpricing at the Initial Public Offering (IPO) due to reduced ex-ante uncertainty. This analysis provides evidence to the contrary: VC-backed IPOs appear to be more underpriced than non VC-backed IPOs.
1. INTRODUCTION Venture capitalists are described as experts in the field of high-risk company funding (see for example Fenn et al., 1997; Lerner, 1995; Sahlman, 1990). They not only specialize by concentrating on certain industry sectors and specific The Rise and Fall of Europe’s New Stock Markets Advances in Financial Economics, Volume 10, 201–230 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3732/doi:10.1016/S1569-3732(04)10009-1
201
202
STEFANIE A. FRANZKE
stages of a company’s development, but also actively engage in monitoring and consulting activities. Since they often serve as members on the “Aufsichtsrat”1 and frequently invest their capital conditional on whether intermediate goals have been reached, they are able to influence the behavior and corporate strategy of the company. Their incentive to improve corporate governance is on the one hand due to the finite life of the partnership and – since their compensation is linked to the firm’s performance – to the maximization of the exit price.2 On the other hand, being repeat players who regularly have to raise new funds, venture capitalists face reputational risk. One would therefore expect that, much like prestigious underwriters or auditors, venture capitalists certify the quality of a company when going public. Within the extensive underpricing literature some empirical studies examine whether the market honors the presumed monitoring activities of venture capitalists. Since this control benefit may reduce the ex-ante uncertainty for potential investors, it should lead to lower underpricing, defined as the spread between the initial offering price and the opening price on the first day of trading. However, empirical evidence is mixed. Among others, Barry et al. (1990), Megginson and Weiss (1991) and Lin and Smith (1998) confirm the certification role of venture capitalists for the U.S. market. They show that venture capital (VC)-backed IPOs are less underpriced than non VC-backed IPOs. On the other hand, Francis and Hasan (2001) and Smart and Zutter (2003), who also analyze U.S. data, find that the average initial return of venture-backed IPOs is higher than that of non venture-backed IPOs. Ljungqvist (1999) demonstrates that the finding of venture-backed IPOs appearing less underpriced has to be attributed to the incentives of the pre-IPO shareholders to reduce underpricing and not to venture-backing. Pre-IPO shareholders will care for the pricing of an issue or for the choice of an underwriter only to the extent that such decisions affect their wealth. Studies by Habib and Ljungqvist (2001) and Ljungqvist (1999) show that underpricing-induced wealth losses increase with the number of shares sold in the IPO. As a consequence, pre-IPO owners selling a substantial fraction of their shares at the IPO should have strong incentives to control underpricing. This study contributes to the literature in several ways. It analyzes the certification role of venture capitalists and underwriting banks3 at the IPO, exploring a unique German data set of companies going public on Neuer Markt. Moreover, this study examines the incentives of the pre-IPO shareholders to control underpricing. Similar to Habib and Ljungqvist (2001) and Ljungqvist (1999), who model underpricing as endogenous to the pre-IPO shareholders’ problem of minimizing the total wealth loss in an IPO, we estimate a two-stage least square regression model. The analysis of the German market is of special interest, since
Underpricing of Venture-Backed and Non Venture-Backed IPOs
203
it offers the opportunity to analyze a market, in which banks have considerable importance serving as both underwriters and financiers of VC companies. This is different from the Anglo-American markets, where pension funds are the main suppliers of VC, while banks play a minor role. Moreover the German VC market has only recently gained some importance within the financial services industry. As a consequence only few empirical studies are available4 about the players on the German VC market, their investments and divestment activities and their impact on their portfolio companies. This chapter has two objectives. First, to enlarge the level of knowledge with respect to venture capital financing in Germany and second to compare the results with those of international studies. The main result of this study is that venture-backed IPOs on the Neuer Markt experienced considerably more underpricing than non-venture backed IPOs. We do not find support for the conjecture that the prestige of the venture capitalist lowers underpricing. This results holds, irrespective of whether we control for the selling behavior of pre-IPO shareholders or for conflicts of interest due to an affiliation of the venture capitalist and the underwriting bank. This chapter is organized as follows. Section 2 defines the term venture capital as it is used in Germany and discusses why it is essential for venture capitalists to have the option to exit their portfolio companies by means of an IPO. Section 3 outlines the impact of the Neuer Markt on the primary equity market in Germany. Moreover, it provides an analysis of the costs of going public on the Neuer Markt. In Section 4 we formulate testable hypotheses. Section 5 describes the data set and the design of the empirical analysis. Sections 6 and 7 present descriptive statistics and empirical results. Section 8 discusses extensions and Section 9 concludes this chapter.
2. DEFINITION OF VENTURE CAPITAL AND THE IPO AS EXIT MECHANISM The definition of venture capital differs in the literature.5 In the Anglo-American understanding venture capital is often used in the context of early-stage (such as seed and start-up financing) and expansion financing. In Germany, venture capital is more comprehensive, since it also includes later-stage capital (such as bridge-, buy out-, and turnaround financing).6 While the former types of investments are crucial for the development and implementation of business ideas by young growth companies, the latter types of investments are important for capital structure reasons of more mature, small to medium-sized companies. To be aware of venture capital’s different meanings is important when interpreting (German) figures and in particular when comparing empirical results of various international studies.
204
STEFANIE A. FRANZKE
For the success of a transaction in the VC market the venture capitalist’s exit decision, i.e. the divestment of the portfolio company is essential. It is important not only for the relationship between the venture capitalist and the portfolio company but also for the relation between the venture capitalist and VC funds investors. Such investors evaluate the skills of the venture capitalist, which acts as their agent, based on his track record, i.e. the return on investment. At the time of the exit an objective performance measure for investors of VC funds is generated. Such a benchmark is important for future allocation decisions of VC funds investors and can thus be decisive for the ongoing existence of a venture capitalist (e.g. Black & Gilson, 1998). Among the different exit channels venture capitalists prefer either a trade sale, i.e. the selling of the portfolio company to an industrial or strategic investor, or an IPO. Black and Gilson (1998) argue that a unique advantage of the IPO is that this exit channel allows the entrepreneur to regain the control over the portfolio company. Already at the time of the venture capitalist’s initial investment in the portfolio company the implicit contract over future corporate control aligns the interests of the venture capitalist and entrepreneur and reduces the costs of contracting. Cumming and MacIntosch (2003) highlight other motives for selling the portfolio company by means of an IPO, such as high sales proceeds or building up a reputation for financing young companies. However, since the market for IPOs is highly cyclical with regard to volume and pricing, venture capitalists are more inclined to take trade sales into consideration as an exit vehicle in bear markets, in particular if the date of the termination of the invested VC funds is imminent.
3. GERMANY’S NEUER MARKT AND THE COSTS OF GOING PUBLIC The Neuer Markt was Germany’s trading segment for innovative growth companies. It was launched on March 10, 1997 as a subsidiary of the Deutsche B¨orse, with the objective to attract small- to medium-sized, young technology firms and has been dissolved in 2003. As Fig. 1 indicates, the number of companies that have gone public in Germany increased dramatically during the IPO boom period. A large part of this increase is due to the Neuer Markt. From March 1997 through March 2000, over 200 companies went public on the Neuer Markt, while at the same time new listings on the first segment (Amtlicher Handel) and second segment (Geregelter Markt) stayed close to their previous levels. In total about 320 new listings were recorded at Frankfurt Stock Exchange during that period.7
Underpricing of Venture-Backed and Non Venture-Backed IPOs
205
Fig. 1. New Issues at the Frankfurt Stock Exchange. Note: On the SMAX (Small Cap Exchange) – introduced in April 1999 – second market stocks were traded. The listing requirements of the SMAX followed – apart from minor differences – those of the Neuer Markt.
The increasing importance of IPOs as an exit vehicle for German venture capitalists is reflected in the BVK Statistics (2002) that is shown in Table 1. From 1998 to 2000 the Neuer Markt covered on average about 68% of the volume of all venture-backed IPOs by members of the BVK. The remaining 32% can be split into IPOs on other German stock markets (14%) and listings on foreign stock exchanges (18%) such as the NASDAQ. After the stock market bubble had burst in 2000 it became much more difficult and finally nearly impossible for companies to go public. This is reflected in the numbers for the period March 10, 2001–March 10, 2002 in Fig. 1. Table 1 reports that the issuing activity of venture-backed companies came to a near stand still in 2001. At the same time the percentage of companies that had to be written off increased to more than 35%. All but one Neuer Markt company (TRIUS AG) chose book building to price their IPO shares.8 The final issue price was never fixed above the initial price range9 despite two out of three issues being oversubscribed.10 Only 2.7% of all Neuer Markt IPOs11 were priced below the minimum price limit. This has also been reported by Aussenegg et al. (2003). The major potential benefit of book building, to update the offer price to strong investor demand, does not seem to be exploited fully in Germany. Ljungqvist et al. (2003) conjecture that local regulations, the costs caused by price revisions or the market power of domestic investors could serve as explanations for the unwillingness to set the final offer price outside the initial price range.
206
Table 1. Volume and Percentage of Exit Vehicles as Stated by BVK Statistics (2002). 1997
1998
1999
2000
d Million
%
d Million
%
d Million
%
d Million
2001 %
d Million
2002 %
d Million
%
IPOs total IPOs NM Divestment after IPO Trade sale Buy back Write off Other
19.9 NA NA
2.7 NA NA
75.2 47.0 NA
14.0 8.8 NA
97.1 44.5 48.6
12.6 5.8 6.3
116.3 105.8 40.4
9.2 8.4 3.2
7.3 2.1 138.5
0.4 0.1 7.5
0.0 0.0 122.3
0.0 0.0 5.7
248.5 157.0 109.4 193.3
34.1 21.6 15.0 26.6
119.6 161.1 91.5 90.0
22.3 30.0 17.0 16.8
192.8 159.5 161.1 110.0
25.1 20.7 21.0 14.3
492.4 215.7 232.1 164.6
39.0 17.1 18.4 13.1
379.1 333.8 673.8 322.6
20.4 18.0 36.3 17.4
651.6 78.1 941.2 338.7
30.6 3.7 44.1 15.9
Total
728.1
100.0
537.4
100.0
769.1
100.0
1,261.5
100.0
1,855.1
100.0
2,131.8
100.0
STEFANIE A. FRANZKE
Note: The classification “Other” contains, among other things, selling to a financial investor or a venture capitalist (i.e. secondary purchase). The abbreviation “NA” stand for “Not Available.”
207
Underpricing of Venture-Backed and Non Venture-Backed IPOs
Table 2. Costs of Going Public at the Neuer Markt (March 1997 to March 2002). In d Thousand Respectively %
Mean
Median
Std. Dev.
90%-Quantile
10%-Quantile
Obs.
Direct flotation costs Underwriting fees Indirect costs: money left on the table Gross issue proceeds Relative direct costs Relative underwriting fees Initial return
4,723 3,138 27,916
3,139 1,981 7,735
7,317 5,746 52,912
8,527 5,210 75,432
1,498 809 −267
300 300 300
69,438 8.89% 5.27%
38,713 8.32% 5.20%
180,000 3.30% 1.21%
116,852 12.43% 6.23%
15,750 5.78% 4.05%
300 300 300
49.81%
18.51%
73.13%
160.00%
−0.97%
300a
Note: Money left on the table is calculated by multiplying the total volume of issues with the difference of the initial offering price and the opening price at the first day of trading. The first-day return (underpricing) is the spread between the opening price at the first day of trading and the initial offering price. Relative direct costs are measured as the ratio of direct flotation costs and the gross proceeds of an issue. The relative underwriting fee is defined as the underwriting fee paid at IPO normalized by the gross proceeds of the issue. a 31 of the 300 observations are overpriced as indicated by the negative initial return. 34 observations have an initial return of 0%.
In order to analyze the issuing costs for companies at Neuer Markt in detail, Table 2 distinguishes between direct and indirect costs. The direct costs include auditing and consulting fees, underwriting fees, marketing costs, and costs for the printing of the prospectus and for services provided by Deutsche B¨orse.12 The direct costs are calculated from information taken from the IPO prospectuses. During the period of March 1997 through March 2002, total direct flotation costs averaged 8.89% of gross proceeds. As part of these costs the average underwriting fee amounted to 5.27% of gross proceeds. To compare, Kaserer and Kraft (2000) report that the average total direct flotation costs amounted to 7.77%, while the average underwriting fees amounted to 5.01% for German IPOs during the period 1993–1998. Much higher than the direct costs are the indirect costs in the form of underpricing. The indirect costs average 49.81%. If shares had been sold at the market price and not at the issuing price, the average issuing company could have raised about d 28 million more. In the literature (e.g. Loughran & Ritter, 2002) this cost factor is discussed as “money left on the table.” It is calculated by multiplying the total number of shares sold in the IPO with the difference of the initial offering price and the opening price at the first day of trading.
208
STEFANIE A. FRANZKE
Earlier studies on the German IPO market, such as Wasserfallen and Wittleder (1994) and Ljungqvist (1997) report lower level of underpricing. Wasserfallen and Wittleder (1994) analyze 92 IPOs from 1961 to 1987 and find an underpricing of 17.58%, on average. Ljungqvist (1997) reports an average underpricing of 9.2% for a sample of 180 IPOs from 1970 to 1993. However, the finding of a high level of underpricing during the dot-com bubble of the late 1990s is documented for other markets, as well. For instance, Ritter and Welch (2002) report that U.S. IPOs experienced an average underpricing of 65% during the years 1999 and 2000.
4. RELATED LITERATURE AND HYPOTHESES Carter and Manaster (1990) and Booth and Smith (1986) emphasize the signaling and certification-of-quality role fulfilled by prestigious underwriters, auditors and venture capitalists. Given that outside investors believe in the information advantage of a third party (underwriter or venture capitalist), this party is able to certify the quality of a company going public if it has reputational capital at stake. This reputational capital “must be greater than the largest possible one-time wealth transfer or side payment which could be obtained by certifying falsely” (Megginson & Weiss, 1991, p. 881). Furthermore it must be costly for the issuing firm to purchase the service of the certifying agent. Underwriters and venture capitalists should be able to carry out the role of a certifying agent, as they often have insider information. The underwriting bank’s information results from the involvement in due diligence activities and a potential lending relationship prior to the IPO. Underwriters have incentives to examine the quality of the firm in detail as they are liable for statements made in the IPO prospectus.13 Since venture capitalists belong to the group of owners that are actively involved in the company, they are likely to have private knowledge about the company’s history, quality of management and financial situation. In addition, venture capitalists tend to have large ownership stakes and therefore have incentives to engage in monitoring activities after the IPO. Both underwriters and venture capitalists have reputational capital at stake as their future success is closely linked to their current reputation and they repeatedly bring firms public. The better the reputation, the easier the attention of trading partners can be caught. Underwriters regularly have to attract issuers and venture capitalists frequently have to raise new funds from institutional investors. One can therefore argue that the involvement of a prestigious underwriter or venture capitalist should certify and credibly signal the quality of the issuing
Underpricing of Venture-Backed and Non Venture-Backed IPOs
209
company to the market. We thus assume that it should pay to hire a prestigious intermediary, as it leads to a higher offer price, which in turn implies lower underpricing. We formulate the following hypotheses: Hypothesis 1. The more prestigious the underwriter (UWrank) involved in the IPO, the lower the underpricing. Hypothesis 2a. The more prestigious the venture capitalist backing the company before the IPO (VCrank), the lower the underpricing. The informative value of the venture-backing signal depends in particular on the venture capitalist’s equity holdings prior to the IPO. A venture capitalist owning, say, 50% of the shares is expected to have more information than a venture capitalist that owns only 5% of a company. Thus Hypothesis 2a should be refined: Hypothesis 2b. The more prestigious the venture capitalist and the bigger the venture capitalist’s equity holdings of the issuer prior to the IPO (VCstake), the lower the underpricing. Barry (1989) argues that a focus on underpricing alone is misleading. Underpricing per se is uninformative when not controlling for the former shareholders’ incentives to influence underpricing. They will influence the pricing of an issue if their wealth is negatively affected by the price setting. In other words, entrepreneurs and venture capitalists will not care for the wealth loss occurring through underpricing when selling a single share, but they will care the higher their participation in the offering, i.e. the more shares they sell at the IPO.14 We therefore introduce Hypothesis 3: Hypothesis 3. The higher the participation ratio (participation) of former shareholders (e.g. venture capitalists or managers, respectively) the lower the underpricing. Habib and Ljungqvist (2001) and Ljungqvist (1999) extend this idea and model underpricing as endogenous to the problem of minimizing the former shareholder’s total wealth loss when going public. They assume that the wealth loss of former shareholders at the IPO is a function of (i) underpricing, when selling existing shares, (ii) the dilution of the value of retained shares,15 and (iii) costs arising in connection with activities that reduce underpricing and wealth losses, such as extensive marketing efforts prior to the IPO. They conclude that there is a trade-off for such shareholders between investing in costly actions to reduce underpricing and tolerating higher underpricing. In order to take the endogenous relation between the direct non-underwriting costs (exp) (normalized by the number of shares sold in the IPO) and underpricing into account, we adopt a two-stage least squares approach.16
210
STEFANIE A. FRANZKE
5. DATA AND METHODOLOGY My original data set contains 353 companies that went public on the Neuer Markt from March 10th, 1997 to March 10th, 2002. The final sample of 300 IPOs excludes 28 companies that changed their market tier or had already been listed at a foreign stock exchange before going public on the Neuer Markt. In addition, four companies from the financial services industry were excluded because of differences in reporting and accounting environment. Finally, another 21 companies could not be taken into consideration because the total flotation costs were not available. Given the differences in the definition of venture capital in the U.S. and Germany, we establish comparability of the empirical studies by dividing the Neuer Markt data set into three groups (Table A.1 in the Appendix): 160 non venture-backed IPOs (53.33%), 79 venture-backed IPOs (26.33%) and 61 companies (20.33%) that received bridge financing by investors. For 119 cases, the distinction between bridge financing and venture capital financing is based on the classification made by the (lead) financier17 itself in company reports or on the Internet page. In the remaining 21 cases, we distinguish between bridge financing and venture capital financing based on the history of the portfolio company and the type of financing the company received before the IPO. As bridge financiers typically have not invested seed, start-up and expansion capital next to bridge financing and therefore engage themselves at a rather late stage of the development of a company, we assume that bridge financing is associated with less monitoring activities and certification ability than venture capital financing. In order to find support for this assumption the monitoring skills of venture capitalists in comparison to those of bridge financiers are examined in more detail using proxies such as: the fraction of the issuing firm’s shares owned by the venture capitalist/bridge financier or the length of time that a venture capitalist/bridge financier has served on the supervisory board. We compare the venture-backed group and the sub sample of companies that received bridge financing to the non-venture backed group using tests for equality of means (t-test) and equality of median (Kruskal-Wallis). Because of the focus on venture capitalists and their certification role, we concentrate on the venture and non venture-backed sub samples when testing the hypotheses. We have collected detailed information for each IPO on the number of shares issued at the IPO, the issuing procedure, the offering expenses, the number of shares outstanding, the age of the company, the number of employees, the ownership structure, the supervisory board (“Aufsichtsrat”), the identity of venture capitalists or rather private equity companies and underwriters. To identify the VC-firms and private equity companies and their age, Internet pages and company
Underpricing of Venture-Backed and Non Venture-Backed IPOs
211
reports (if available), as well as the list of the full members of the Bundesverband Deutscher Kapitalbeteiligungsgesellschaften e. V. (BVK) – German Venture Capital Association – and the European Venture Capital Association (EVCA) were used. We have also collected data from the financial statements included in the IPO prospectus. Additional information was obtained from financial newspapers such as the first day of trading, the book building spread, the initial offering price, the opening price for the first trading day and the closing price for 20 trading days after the IPO and information on the greenshoe (over-allotment option) exercise. For the construction of the underwriter’s ranking we collected information on the lead managers of all Frankfurt stock market segments since 1990. This information was provided by Deutsche B¨orse. A total of 104 different underwriters (45 different lead underwriters) have been involved in IPOs at Frankfurt stock exchange from March 1997 to March 2002.18 We construct a rating for each year.19 This implies that the rating of an underwriter can change over time. The data of banks that merge during the sample period (such as Bankhaus Gontard and Metallbank or Bayerische Vereinsbank and Bayerische Hypotheken- und Wechselbank) are aggregated in order to avoid major changes in the rating. The ratings of the years 1998, 1999, 2000 and 2001 are constructed using the track record of each underwriter as measured by the relative share of lead managed IPO deals at all Frankfurt stock market segments since 1990 and the relative volume of proceeds of IPOs on the Neuer Markt as reported on December 31st of the precedent year. The relative share of a lead manager for each year is calculated by cumulating the number of IPO deals that were lead managed since 1990 and dividing this number by the cumulated number of IPOs that took place since 1990. In order to calculate the relative gross proceeds at the Neuer Markt for each bank we cumulate the volume of gross proceeds in d million each bank has underwritten (as lead- or co-underwriter) since 1997 and divide it by the total volume of gross proceeds in d million of all IPOs at the Neuer Markt since 1997. Due to the lack of a track record of the relative volume of gross proceeds at the Neuer Markt for the year 1997, the rating of 1997 is solely based on each bank’s relative share of lead management at all Frankfurt stock market segments since 1990. The rating scale for each of the two factors mentioned above is ranging from one (very experienced) to four (non experienced). Banks which have not underwritten any IPO are rated four. A rating of three is assigned to banks, which have a relative share that is equal or lower than the average share of all underwriting banks. Banks with a share higher than the average but lower than half of the share of the markets leader are rated two. The best rated banks (i.e. banks with a rating of one) have a relative share that is at least half of the share of the market leader. The final rating is estimated
212
STEFANIE A. FRANZKE
weighting the two individual ratings for the factors “lead managed IPO deals” and “proceeds of IPOs” equally. Prestigious lead underwriters with a rating of one have underwritten a total of 102 out of 300 companies. In the regression analysis a dummy variable is used which has the value one in case the underwriter’s rating is equal to one and zero in case the underwriter’s rating is lower than one. We also rate the quality of venture capitalists and private equity companies. This rating is mainly based on the age of the venture capitalist or private equity company. VC and private equity companies founded before 1983 receive a very good rating (equivalent to one), companies founded during the period 1983 to 1996 receive a mediate rating (equivalent to two). Companies founded after 1996 get the lowest rating (equivalent to three). For some companies it was impossible to find information regarding their age. In these cases we assume the lowest rating (equivalent to three). We use age as a proxy for reputation because younger venture capitalists and private equity companies are less experienced. This is reflected in a total of 148 venture funds/companies or private equity companies backing 140 IPO firms: 99 of these (67%) back only one IPO firm, 34 (23%) back up to 4, and only 15 (10%) back more than 5 IPOs during the time period March 1997 to March 2002. Only in six cases20 a relative high backing activity (i.e. a backing activity that was at least half as frequent as that of the market leader in that year) leads to an upgrade in rating during the period under consideration. In analogy to the underwriters’ rating, the information concerning the quality of the lead venture capitalist is condensed to a dummy in the regressions. The dummy is equal to one if the financier’s rating is very good (this is the case for 29 out of the 79 venture-backed IPOs and for 10 IPOs out of 61 backed by bridge financing) and zero in any other case. Table A.3 in the Appendix presents the twelve best rated venture capitalists and private equity companies. In total, 81 of the 148 venture funds/companies or private equity companies act as lead financier, i.e. they are at least once the largest investor of a portfolio company. The remaining 67 only hold small stakes in a portfolio company, i.e. they only invest in a company together with investors that own larger stakes. In order to measure the ex-ante uncertainty concerning the value of an IPO company three different proxies are used. Following Ritter (1984), Wasserfallen and Wittleder (1994) and Prabhala and Puri (1998) we use the annualized volatility of the 20 daily returns from day 2 to 21 (vola) as a proxy for uncertainty. Theory predicts a positive relation between uncertainty and underpricing. Since this proxy might be distorted due to underwriter price support in the aftermarket (see Ljungqvist, 1997), we use the log of the number of employees (empl) as well. Large companies that go public and employ many people should be less underpriced than small companies.
Underpricing of Venture-Backed and Non Venture-Backed IPOs
213
Following Ljungqvist et al. (2003) and Loughran and Ritter (2002), we analyze to what extent the underwriter updated the price during the book building procedure (bookb). This is calculated by dividing the difference between the issuing price and the minimum price limit of the book building range by the width of the book building range. Issues priced at the maximum price limit, should be more underpriced compared to IPOs with an issue price that falls within the book building range or below the minimum price limit. In this context underpricing can be interpreted as a way to reward investors for releasing private information about their demand for IPO shares. Besides we use the market trend, a proxy also used by L¨offler (2000) and Uhlir (1989). The market trend is estimated using the NEMAX, i.e. the stock market index of Neuer Markt at Frankfurt stock exchange, for the period forty days before the IPO (nemax). The only company of the data set analyzed for which we are not able to calculate the market trend prior to the IPO is Mobilcom, which went public on March 10th, 1997. In this special case we assume a market trend of zero. As L¨offler documents, there seem to exist (psychological/market) factors that lead to a significant positive relation between the trend of the NEMAX and the degree of underpricing. In addition, we introduce a dummy (bear-market) that is equal to one if the IPO took place after the 10th of March 2000 and zero otherwise. One can either apply a dummy for venture capital backing or less condensed information, i.e. the percentage of the venture capitalists’ equity holdings prior to the IPO (VCstake) and the reputation of the venture capitalist (VCrank). We will use the latter two variables. Moreover, in line with Habib and Ljungqvist (2001) and Ljungqvist (1999) we control for the participation ratio (participation), that is the fraction of shares that pre-IPO shareholders sell in the offering and the dilutions factor (dilution), which is determined as the number of new shares divided by the number of shares outstanding before flotation. The determinants of underpricing are examined applying a two-stage least square approach. With reference to the hypotheses discussed in Section 4 and taking the trade-off between investing in costly actions (exp) that reduce underpricing and tolerating underpricing into account, this leads to the following system of regressions: underpr = ␣0 + ␣1 vola + ␣2 bookb + ␣3 empl + ␣4 nemax + ␣5 bear-market + ␣6 UWrank + ␣7 VCrank + ␣8 VCstake + ␣9 participation + ␣10 dilution + ␣11 exp + 1
(1)
214
STEFANIE A. FRANZKE
exp = 0 + 1 participation + 2 dilution + 3 ln volume + 4 underpr + 2
(2)
When analyzing the trade-off between underpricing and the non-underwriting costs of the going public process (exp), one has to care for the problem of endogeneity of the regressors underpr and exp. For each endogenous regressor at least one instrumental variable is needed. For exp we use the log of the number of shares sold in the IPO (ln volume) as an instrument. This variable should be correlated with exp but not with underpr. This instrument seems to be reasonable, since the correlation coefficient for ln volume and exp is –39%, while it is –16% for ln volume and underpr.
6. DESCRIPTIVE STATISTICS Table 3 reports issuer characteristics. We find that venture-backed companies differ from non-venture backed ones with regard to EBIT (earnings before interest and taxes) per employee and profit on sales. Both ratios are on average significantly smaller for VC-backed firms: −11 vs. 24, and −44 vs. −5. This shows that VC-backed IPO firms are less profitable at the time of the IPO. There are no significant differences in the number of employees, age, and total assets. Moreover, no statistical difference can be found for the uncertainty or risk of venture- and non venture-backed IPO companies measured by the annualized volatility of the 20 daily returns from day 2 to 21. At first glance, the findings concerning offerings characteristics are in line with the results of Ljungqvist (1999). Venture-backed companies sell significantly more existing shares when going public than non venture-backed companies. This is reflected by an average of 20.49% vs. 13.20% of secondary sales of the total shares sold in the IPO, and by an on average higher participation ratio (10.33% vs. 5.22% of the total shares outstanding before the IPO). But the average and median participation ratio of managers in venture-backed IPOs is significantly lower. This underlines that it is important to differentiate between different groups of former stockholders such as venture capitalists, managers and underwriters. Furthermore, the univariate analysis presented in Table 4 shows that venturebacked companies are not less underpriced, and are not more likely to hire an expensive underwriter compared to non venture-backed firms. For both of the samples, the average underpricing is about 50% and the average gross spread equals 5%. Next, in Table 5 we examine the differences between venture-backed companies and firms that have received bridge financing. On average about two thirds of
215
Underpricing of Venture-Backed and Non Venture-Backed IPOs
Table 3. Issuer and Offering Characteristics of Venture-Backed, Bridge Financed and Non Venture-Backed Companies.
Employees NVC VC BF
Obs.
Mean
p-Value
160 79 60
243 223 132
0.6422 0.0182
Age of company NVC 160 VC 79 BF 61
Median
130 113 64
p-Value
0.8214 0.0004
11.5 10.5 9.5
0.5458 0.1613
9.5 8.0 7.0
0.4023 0.1735
Balance sheet total, in million d NVC 160 34.3 VC 79 29.2 BF 61 13.8
0.5359 0.0057
15.0 12.4 8.5
0.5200 0.0009
EBIT in thousand d per employee NVC 160 23.9 VC 79 −10.8 BF 59 −11.9
0.0009 0.0048
8.4 −7.3 0.6
0.0000 0.0001
Profit on sales in % = EBIT in thousand d per sales revenues in thousand d NVC 158 −5.14 6.59 VC 77 −44.40 0.0010 −7.39 BF 59 −38.31 0.0086 0.64
0.0000 0.0000
Issuing proceeds incl. greenshoe option in million d NVC 160 82.67 VC 79 62.55 0.4646 BF 61 43.64 0.2107
38.24 47.60 30.26
0.1510 0.0334
Existing stocks sold in % of total volume of shares issued NVC 160 13.20 VC 79 20.49 0.0013 BF 61 12.85 0.8764
7.33 18.35 9.96
0.0016 0.8639
Participation old stockholders NVC 160 VC 79 BF 61 Participation managers NVC 160 VC 78 BF 61
5.22 10.33 5.15
0.0000 0.9373
2.39 7.49 3.95
0.0002 0.6874
3.39 1.26 1.42
0.0243 0.0596
1.78 0.24 0.00
0.0042 0.0026
216
STEFANIE A. FRANZKE
Table 3. (Continued ) Obs.
Mean
Median
p-Value
Annualized volatility of 20 daily returns from day 2 to 21 in % NVC 160 87.56 VC 79 87.41 0.9794 BF 61 92.99 0.4128
78.28 81.84 81.85
0.5415 0.4621
40 day log return of NEMAX before IPO in % NVC 160 2.56 VC 79 5.77 BF 61 0.55
−3.63 0.36 −5.44
0.3489 0.4934
p-Value
0.3501 0.5973
Note: The data set consists of 160 non venture-backed IPOs (NVC), 79 venture-backed IPOs (VC) and 61 companies (BF) that received bridge financing. The participation ratio (for instance of the manager) is calculated by dividing the number of old shares sold (by the manager) by the total number of shares outstanding before flotation. NEMAX is the stock market index of the Neuer Markt at Frankfurt stock exchange. The test for differences in means is a standard t-test, allowing for unequal variance. The test for differences in medians is the Kruskal-Wallis test.
Table 4. Issuing Costs of Venture-Backed, Bridge Financed and Non Venture-Backed Companies. Obs.
Mean
p-Value
Median
p-Value
Underpricing in % NVC 160 VC 79 BF 61
48.38 52.44 50.17
17.50 24.00 18.43
0.6962 0.8738
0.2528 0.5411
Relative direct costs NVC 160 VC 79 BF 61
8.82 8.68 9.35
8.09 8.08 8.94
0.7573 0.2849
0.9287 0.0597
Relative underwriting fees (gross spread) NVC 160 5.31 VC 79 5.12 BF 61 5.30
5.17 5.13 5.37
0.2741 0.9469
0.4970 0.0988
Note: The data set consists of 160 non venture-backed IPOs (NVC), 79 venture-backed IPOs (VC) and 61 companies (BF) that received bridge financing. Underpricing is measured as the spread between the initial offering price and the opening price at the first day of trading. The test for differences in means is a standard t-test, allowing for unequal variance. The test for differences in medians is the Kruskal-Wallis test.
217
Underpricing of Venture-Backed and Non Venture-Backed IPOs
Table 5. Characteristics of Financial Intermediaries and Offering Characteristics of Venture-Backed and Bridge Financed Companies. Obs.
Mean
Median
p-Value
Number of venture capitalists or bridge financiers forming a syndicate VC 79 2.68 0.0000 BF 61 1.36
2.00 1.00
0.0000
Stake of lead venture capitalist/lead bridge financier before IPO, in % VC 78 27.10 0.0000 BF 61 11.32
22.55 9.40
0.0000
Stake of venture capitalists/bridge financiers before IPO, in % VC 79 40.31 0.0000 BF 61 12.80
36.00 10.00
0.0000
Stake of lead venture capitalist/lead bridge financier after IPO, in % VC 78 15.54 0.0000 BF 61 6.87
12.90 4.89
0.0000
Stake of venture capitalists/bridge financiers after IPO, in % VC 79 23.14 0.0000 BF 61 7.91
21.36 6.30
0.0000
Participation venture capitalists/bridge financiers VC 79 21.44 BF 61 17.89
16.67 8.24
0.0689
p-Value
0.3934
Seats on the “Aufsichtsrat” held by venture capitalists or bridge financiers, in % VC 79 26.52 0.0000 33.33 BF 61 13.46 0.00
0.0001
Duration of financial relationship in months VC 78 30 BF 59 7
0.0000
0.0000
Dummy rating of lead venture capitalist/bridge financier = 1 VC 79 36.71 0.0076 BF 61 16.39
23 6 0.00 0.00
0.0081
Note: The data set consists of 79 venture-backed IPOs (VC) and 61 companies (BF) that received bridge financing. The participation ratio (of the lead venture capitalist or bridge financier, respectively) is calculated by dividing the number of existing shares sold (by the lead venture capitalist or bridge financier, respectively) by the total number of shares outstanding before flotation. The test for differences in means is a standard t-test that allows differences in variance. The test for differences in medians is the Kruskal-Wallis test.
218
STEFANIE A. FRANZKE
the VC-backed companies have been financed by a syndicate before the IPO,21 whereas issuing companies that received bridge financing dealt with more than one bridge financier only in one out of three cases (not reported). Table 5 shows that the average stake of the lead venture capitalist is significantly higher before (27.10% vs. 11.32%) and also after the IPO (15.54% vs. 6.87%) than the stake held by the lead bridge financier. The average stake of the syndicate of venture capitalists compared is also higher than the stake of the group of bridge financiers (before the IPO 40.31% vs. 12.80%; after the IPO 23.14% vs. 7.91%). In addition, both groups of financial intermediaries sell on average about 20% of their pre-IPO stake at the IPO.22 Furthermore, venture capitalists are more likely to have more inside information than bridge financiers, since the former hold an average of 26.52%23 vs. 13.46% of the seats on the “Aufsichtsrat.” In addition, venture capitalists have been involved with the firm for two years longer than bridge financiers. Finally, about 37% of the VC-backed sample are backed by a venture capitalist with a very good rating, while only 16% of the bridge financing sample are financed by private equity companies with a very good rating. Taking the proportion of ownership and degree of insider knowledge into account the bridge financiers’ certification ability seems to be modest. Therefore we do not analyze the bridge financed sample in subsequent regression analyses.
7. EMPIRICAL RESULTS Table 6 shows the regression results. There is no evidence for a trade-off for issuers on the Neuer Markt between investing in costly actions to reduce underpricing (exp) and tolerating higher underpricing (underpr). In both regressions the variables underpr and exp lack significance. However, pre-IPO shareholders seem to spend more on non-underwriting costs the more shares they sell, i.e. the more they participate in the offering (see column 4 and 5 of Table 6). We find that the normalized non-underwriting costs are significantly negatively related to the ratio of new shares divided by the number of shares outstanding before flotation. The explanation for this finding may be similar to that for the log of the number of shares sold in the IPO (ln volume). The higher the number of shares sold in the IPO the lower the amount of non-underwriting expenses per unit of issuing proceeds. Our findings do not support the findings of Barry (1989), Habib and Ljungqvist (2001), and Ljungqvist (1999) (see Hypothesis 3) that former shareholders selling large fractions of their pre-IPO have more incentives to control underpricing.
219
Underpricing of Venture-Backed and Non Venture-Backed IPOs
Table 6. Test of the Certification Hypothesis 1. Variables
(1) Underpr
(2) Underpr
(3) Underpr
Constant
0.389 (0.3292) 0.295 (0.0049) −0.083 (0.0795) 0.165 (0.0349) 1.288 (0.0000) −0.137 (0.1157) 0.048 (0.5917) 0.310 (0.0350) −0.395 (0.0350) 0.022 (0.9674) 0.241 (0.4398)
0.403 (0.3098) 0.294 (0.0053) −0.083 (0.0766) 0.152 (0.0476) 1.298 (0.0000) −0.156 (0.0753) 0.044 (0.6086) 0.768 (0.0105) −0.284 (0.1275) 0.054 (0.9166) 0.245 (0.4302)
0.464 (0.2130) 0.299 (0.0078) −0.083 (0.0755) 0.152 (0.0502) 1.303 (0.0000) −0.150 (0.0718) 0.047 (0.5940) 0.708 (0.0183) −0.250 (0.2854)
Vola Empl Bookb Nemax Bear-market UWrank = 1 VCrank = 1 VCstake Participation Old Dilution Old Participation VC
ln volume Underpr 0.057 (0.7683)
0.044 (0.8194) −1.278 (0.0309)
0.059 (0.7720) −1.190 (0.0422)
32.13% 238
33.08% 238
32.73% 238
Interaction term: VCrank = 1 × VCstake Adj. R2 Number of observations
(5) Exp
(6) Exp
4.428 (0.0000)
4.412 (0.0000)
4.110 (0.0000)
1.349 (0.0527) −0.938 (0.0013)
1.348 (0.0526) −0.939 (0.0013)
−0.238 (0.0000) 0.074 (0.4779)
−0.237 (0.0000) 0.079 (0.4387)
0.762 (0.4316) −0.059 (0.1470) −0.232 (0.0000) 0.092 (0.3682)
14.35% 238
14.25% 238
8.46% 238
−0.091 (0.8829) 0.023 (0.6427)
Dilution VC
Exp
(4) Exp
Note: In the following a two-stage least square approach is estimated. The analysis is based on 160 non venturebacked IPOs (NVC) and 79 venture-backed IPOs (VC). The dependent variables are underpricing (underpr) and the normalized non-underwriting costs (exp). The variable vola is equivalent to the annualized volatility of the 20 daily returns from day 2 to 21, empl represents the log of the number of employees, bookb reflects the extent to which the book building range was utilized, nemax incorporates the market trend forty days before the IPO. The dummy bear-market is equal to unity if the IPO took place after March 10, 2000. The variables UWrank = 1 and Vcrank = 1 are dummies for underwriters and venture capitalists rated very good. VCstake presents the venture capitalist’s equity holding prior to the IPO, participation and dilution are explained in Notes 14 and 15, respectively. Throughout, t-tests are based on White’s heteroskedasticity-consistent standard errors. p-Values are reported in parentheses.
220
STEFANIE A. FRANZKE
We control for the incentives of the group of the pre-IPO shareholders as a whole, for the managers (not reported) and venture capitalists (see column 3 in Table 6) separately. But the variables participation and dilution lack significance – irrespective of the identity of the pre-IPO owners. In addition, we do not find support for the certification role of venture capitalists and underwriters (Hypotheses 1 or 2a). On the contrary, companies that are backed by a prestigious venture capitalist experience more not less underpricing: the coefficient on the dummy variable VCrank is positive and significant at the 5% level. Remarkably, there is no significant outcome when controlling for venture capitalists with a lower rating (not reported). However, a negative effect can be observed for the variable VCstake: the bigger the venture capitalist’s equity holdings before the IPO, the lower the underpricing. But this negative effect can only be found for companies that have been financed by a prestigious venture capitalist. This conclusion can be drawn from the second column in Table 6. It shows that the interaction term that interacts the dummy for the rating of prestigious venture capitalists with the percentage of the venture capitalists’ equity holdings prior to the IPO, is negative and significant. Moreover, the second column shows that the positive coefficient for prestigious venture capitalists increases in size, while the variable VCstake looses significance. What follows, is that except for companies financed by a prestigious venture capitalist that owns a large stake in the issuing company, issues backed by prestigious venture capitalists appear to be more underpriced. This is in line with the results of Ljungqvist (1999) for the 1990s and those of Francis and Hasan (2001) and Smart and Zutter (2003), but in contrast with the results of Lin and Smith (1998) or Barry et al. (1990). The latter empirical studies show that the higher the venture capitalist’s reputation, the lower the underpricing. We have re-estimated the regression using other factors that usually serve as proxies for the monitoring or backing-quality of venture capitalists, such as the natural logarithm of the age of the lead venture capitalist at IPO, the length of the time a venture capitalist has served on the “Aufsichtsrat” and the age of the financial relationship. However, none of these alternative proxies is negatively related to underpricing. We find that the coefficient of the age of the lead venture capitalists and the length of the membership on the “Aufsichtsrat” is positive and significant. The other variable is not statistically significant. With regard to the marginal effect of underwriter reputation we do not find significance. This suggests that companies that have hired a prestigious lead underwriter when floating stocks are not better off than others. This result corresponds to earlier findings of Kaserer and Kempf (1995) for the German market. We obtain the same results when adding a term to the regression that interacts the rating of the underwriter with that of the venture capitalist (not reported).
Underpricing of Venture-Backed and Non Venture-Backed IPOs
221
However, we find that all parameter estimates that represent the degree of ex-ante uncertainty (vola, bookb) or size (empl) show the predicted signs on a significant level. The smaller the issuing company and the larger the annualized volatility of the 20 daily returns from day 2 to 21, the higher the underpricing. In addition, the more the price was updated towards the upper bound of the initial price range, the higher the underpricing. These results are in line with earlier studies on the German market, such as Wasserfallen and Wittleder (1994). The highly significant coefficient for the market trend (nemax) supports the findings of L¨offler (2000): the initial return increases by about 1.30% with each percentage point increase in the log return of the Nemax prior to the IPO. Moreover, we find that companies going public in bear-markets, i.e. after March 10, 2000, are less underpriced than companies that went public before that time. In summary, we do not find a certification effect at the IPO for venture capitalists or underwriters. Furthermore, there is no evidence that pre-IPO stockholders selling shares at the IPO are particularly concerned about wealth loss and thus attempt to control underpricing costs.
8. EXTENSIONS The question is why issues backed by prestigious venture capitalists appear to be more underpriced. It seems to be puzzling, but similar results have been found before. Francis and Hasan (2001) analyze a data set of companies going public in the United States during the period 1990–1993 using a stochastic frontier model. They show that VC-backed IPOs suffer higher underpricing due to greater pre-market pricing inefficiencies, which are to a significant part deliberate and should compensate investors for information production. The study by Smart and Zutter (2003) examines dual- and single-class IPOs and also reports that underpricing is more pronounced among VC-backed companies. They attribute this result to the circumstance that an increasing number of IPO companies have been financed by younger VC companies, that possibly engage in grandstanding (see Gompers, 1996) by taking their companies to the market earlier and at a larger discount than do established VCs. Ljungqvist (1999) finds evidence that top underwriters are associated with significant increases in underpricing for a sample of U.S. IPOs from the 1990s. This effect is particularly concentrated among venture-backed IPOs. But why do venture capitalists choose to work with prestigious investment banks whose pricing is so much worse? Ljungqvist offers an explanation. There are situations that are characterized by a conflict of interest between entrepreneurs and venture capitalists. He considers the case that the entrepreneur sells some shares at the IPO
222
STEFANIE A. FRANZKE
but the lead venture capitalist does not sell any share. Under these circumstances the venture capitalist is not concerned about hiring a prestigious underwriter that underprices more than average. In this situation, wealth losses are borne by the entrepreneur rather than by the venture capitalist. In our sample, in particular IPOs backed by prestigious or rather older venture capitalists are considerably more underpriced than other IPOs. On average, they are underpriced by 75.32% compared to 39.16% when backed by a less prestigious, younger venture capitalist or 48.38% when non venture-backed. This result contradicts the idea of grandstanding and thus the explanation offered by Smart and Zutter (2003). It seems interesting to analyze whether the significant differences in underpricing can be explained by the selling behavior of venture capitalists. In total, 17% of the prestigious and 24% of the lowest rated venture capitalists do not sell at the IPO. In the first column of Table 7, we re-estimate the previous regression including a dummy for venture capitalists that do not sell at the IPO (nosal VC). Non-selling behavior of venture capitalists leads to a significant increase in underpricing (regardless of the rating of the venture capitalist). However, this result is not robust. Moreover it does not solve the original puzzle, since the coefficient for the dummy of IPOs backed by prestigious venture capitalists remains significant and positive, although the size of the coefficient is smaller. A further explanation why VC-backed IPOs are more underpriced is offered by Hamao et al. (2000). These authors examine IPOs in Japan. In Japan, venture capital funds are often affiliated with major financial institutions. This circumstance can lead to potential conflicts of interest, since the underwriting bank, if an owner of the issuing company, is interested in setting a higher offer price than it would if it did not own any shares. Furthermore, these banks have increased incentives to overstate the company value to investors. Given that IPO investors anticipate this conflict of interest, they will, according to theory, demand more underpricing as compensation. In line with this, Hamao et al. find higher initial returns for IPOs in which the lead venture capitalist is also the lead underwriter. Although affiliations between venture capitalists and underwriting banks exist in Germany, they are not as common as in Japan.24 We have tried to control for this phenomenon of affiliation for the German market, though we have only nineteen observations. The introduction of the variable conflict in the regression analysis has two results shown in the second column of Table 7. The variable nosal VC looses significance and the variable conflict is positive and significant at the 5% level. Even though we find that IPO investors seem to be compensated for potential conflicts of interest of the underwriting banks, the appearance of IPOs backed by prestigious venture capitalists being more underpriced deserves further attention.
223
Underpricing of Venture-Backed and Non Venture-Backed IPOs
Table 7. Test of the Certification Hypothesis 2. Variables
(1) Underpr
(2) Underpr
Constant
0.351 (0.3654) 0.304 (0.0039) −0.088 (0.0583) 0.185 (0.0177) 1.248 (0.0000) −0.150 (0.0845) 0.047 (0.5830) 0.719 (0.0052) −0.446 (0.0267) 0.396 (0.4617) 0.263 (0.3922)
0.386 (0.3174) 0.307 (0.0026) −0.085 (0.0507) 0.158 (0.0427) 1.256 (0.0000) −0.174 (0.0476) 0.070 (0.4151) 0.654 (0.0025) −0.533 (0.0078) 0.189 (0.7075) 0.241 (0.4182)
0.058 (0.7562) 0.337 (0.0347)
Vola Empl Bookb Nemax Bear-market UWrank = 1 VCrank = 1 VCstake Participation Old Dilution Old ln volume Underpr Exp
Interaction term: VCrank = 1 × VCstake
−1.159 (0.0260)
0.023 (0.9070) 0.244 (0.1110) 0.414 (0.0430) −0.993 (0.0156)
Adj. R2 Number of observations
33.75% 238
35.95% 238
Nosal VC Conflict
(3) Exp
(4) Exp
4.429 (0.0000)
4.358 (0.0000)
1.350 (0.0527) −0.938 (0.0013) −0.238 (0.0000) 0.074 (0.4623)
1.342 (0.0515) −0.939 (0.013) −0.234 (0.0000) 0.098 (0.3437)
14.35% 238
13.86% 238
Note: In the following a two-stage least square approach is estimated. The analysis is based on 160 non venturebacked IPOs (NVC) and 79 venture-backed IPOs (VC). The dependent variables are underpricing (underpr) and the normalized non-underwriting costs (exp). The variable vola is equivalent to the annualized volatility of the 20 daily returns from day 2 to 21, empl represents the log of the number of employees, bookb reflects the extent to which the price was updated during the book building procedure, nemax incorporates the market trend forty days before the IPO. The dummy bear-market is equal to unity if the IPO took place after the March 10, 2000. The variables UWrank = 1 and VCrank = 1 are dummies for underwriters and venture capitalists rated very good. VCstake presents the venture capitalist’s equity holding prior to the IPO, participation and dilution are explained in Notes 14 and 15, respectively. Throughout, t-tests are based on White’s heteroskedasticityconsistent standard errors. p-Values are reported in parentheses.
224
STEFANIE A. FRANZKE
9. CONCLUSIONS The main contribution of this empirical study is to shed further light on the growing importance of venture capital in Germany after the introduction of the Neuer Markt in March 1997. In particular, this chapter examines the role of venture capitalists and underwriters in certifying the quality of a company. Carter and Manaster (1990) and Booth and Smith (1986) argue that venture capitalists reduce ex-ante valuation uncertainty resulting in lower levels of underpricing for venture-backed IPOs compared to non venture-backed IPOs. However, to focus only on underpricing might be misleading. According to Barry (1989), Habib and Ljungqvist (2001), and Ljungqvist (1999) the selling behavior of pre-IPO shareholders at the IPO is essential. The more they participate in the offering, that is the more shares they are selling at the IPO, the more incentives they have to reduce underpricing. When running the regressions to test the hypotheses that venture-backed IPOs are less underpriced compared to non venture-backed IPOs, we control for ex-ante uncertainty, market conditions, the venture-capitalists’ stake in the company prior to the IPO, and the incentives of pre-IPO shareholders to reduce underpricing. We found that many financial intermediaries are involved in IPOs at the Neuer Markt: 104 underwriters and 148 venture capitalists and private equity companies. In addition, we report that VC-backed companies are less profitable compared to non venture-backed companies. The pre-IPO owners of venture-backed firms sell significantly more existing shares at the time of the IPO compared to the owners of non venture-backed firms. On average, the group of venture capitalists sells 20% of their pre-IPO stake at the IPO. More than two thirds of the VC-backed companies have been financed by a syndicate of venture capitalists. They seem to have considerable influence, since they hold on average a stake of about 40% of the company before the IPO and about 26.5% of the seats on the supervisory board (“Aufsichtsrat”). We estimate a system of equations using the two-stage least square technique. However, there is no evidence for a trade-off between non-underwriting costs and underpricing. There is strong evidence that higher ex-ante uncertainty is associated with higher underpricing. Furthermore, the market trend is positively related to underpricing. There is no support for the hypothesized certification role of underwriters and/or venture capitalists. Except for companies financed by a prestigious venture capitalist that owns a large share of the issuing company, it does not seem to pay to hire a prestigious intermediary, at least as far as underpricing is concerned. On the contrary, the involvement of a prestigious venture capitalist is associated with higher underpricing. This finding holds regardless of whether we control for venture capitalist not selling at the IPO or for conflicts of interest
Underpricing of Venture-Backed and Non Venture-Backed IPOs
225
due to an affiliation of the venture capitalist and the underwriting bank. The finding that prestigious venture capitalists are associated with more underpricing, warrants further research.
NOTES 1. The “Aufsichtsrat” is similar to the supervisory board. However, German stock companies are governed by two boards. The supervisory board on the one hand is elected by and represents shareholders and employees. Moreover, it appoints the company’s executive board. The executive board on the other hand comprises firm managers and oversees day-to-day operations. 2. When selling at the time of the initial public offering (IPO), this price is equivalent to the offer price. 3. There are two empirical studies analyzing the certification role of underwriters in Germany (see Kaserer & Kempf, 1995 and Wasserfallen & Wittleder, 1994). 4. Of recent date are the studies by Franzke et al. (2004), Rindermann (2003), Sch¨afer et al. (2003), Tykvova (2003), Bottazzi and Da Rin (2002), Engel (2002a) and Engel (2002b). 5. For a deeper discussion see e.g. Bygrave and Timmons (1992), Stedler (1987), Betsch et al. (2000) and Balzer (2000). 6. This broader expression is comparable to the American understanding of private equity. 7. According to Johnson (2000), from 1949 through 1996 a total of only 356 companies went public in Germany. 8. Until 1995 it was common to use the fixed-price method in Germany. One of the main problems of this method in comparison to book building is that underwriting banks do not receive any information concerning demand (e.g. through bids by institutional and retail investors) before the price fixing. 9. 72.8% of the IPOs of the sample have been fixed exactly at the upper price limit. 10. This is in the interest of management because an oversubscription of the offering enables them to influence the allotment (see among others Brennan & Franks, 1997). 11. These companies have been biolitec AG, e.multi Digitale Dienste AG, Euromed AG, Gericom AG, MSH International Services AG, Neue Sentimental Film AG, nexus AG and Paragon AG. 12. Strictly speaking the value of the greenshoe has to be added to these costs. The greenshoe consists of a portfolio of a long call option and a short position in a forward contract on the stock. This instrument is granted (free of charge) by the issuer to the underwriter with the aim to encourage the underwriter to offer price support on the secondary market. The underwriter has an incentive to buy shares on the secondary market and thus to provide price support, if the stock trades below the issuing price during the life of the option. The underwriter can then exercise the call option with a profit equal to the difference between the issuing price and the lower market price. For an in-depth analysis of the value and the use of the greenshoe on the Neuer Markt see Franzke and Schlag (2002). 13. This liability exposes underwriters to litigation. Tinic (1988) suggests that underpricing may serve as an insurance against such securities litigation. 14. The participation ratio (participation) is calculated dividing the ratio of existing shares sold by pre-IPO owners and the number of shares outstanding before flotation.
226
STEFANIE A. FRANZKE
15. The dilution factor (dilution) is determined dividing the number of new shares sold at the time of the IPO by the number of shares outstanding before flotation. In line with Ljungqvist (1999) the dilution factor is taken into account, as well, when running the regressions. However, the predicted sign of this parameter is unclear. 16. The estimation method solves the problem of the ordinary least square approach that “least square estimates are inconsistent estimates of a structural equation precisely because they are consistent estimates of a mixture of all the equations in the model included” (see Green, 1997, p. 736). 17. In line with Ljungqvist (1999), the lead financier is defined as the venture capitalist or private equity company with the largest equity stake. 18. Wasserfallen and Wittleder (1994) stress the dominant role of Deutsche Bank in the underwriter market during the time period 1961–1987, since Deutsche Bank has functioned as lead manager for almost 60% of the issues. This has changed during the time period 1990–2000. Although Deutsche Bank still belongs to the top issuers, its supremacy in underwriting has decreased. 19. Table A.2 presents the twelve best-rated underwriters serving as lead underwriter at Neuer Markt during the period 1997–2002. 20. These financial intermediaries have been Advanced European Technologies N. V., Commerz Unternehmensbeteiligungs AG, TFG Venture-Capital AG & Co. KGaA Unternehmensbeteiligungsgesellschaft, Gold Zack AG, TVM Techno Venture Management GmbH and Schroders Ltd. 21. On average a venture-backed company is financed by three different venture firms/funds (see Table 5). 22. Barry et al. (1990) report that U.S. venture capitalists own on average 34.3% prior and 24.6% after the IPO. They sell on average only about 6.6% of their pre-IPO stake at the IPO. 23. This number is lower than the one reported by Barry et al. (1990). 24. Examples are Deutsche Venture Capital Gesellschaft and Deutsche Bank, Beteiligungsgesellschaft f¨ur die Deutsche Wirtschaft and Dresdner Bank AG, TFG Venture Capital and Concord Effekten AG or Commerz Unternehmensbeteiligungs AG and Commerzbank AG.
REFERENCES Aussenegg, W., Pichler, P., & Stomper, A. (2003). IPO pricing with bookbuilding and a when-issued market. Working Paper, Vienna University of Technology and Boston College. Balzer, K. (2000). Die Bedeutung des Venture Capital f¨ur innovative Unternehmen. Aachen: Shaker Verlag. Barry, C. B. (1989). Initial public offering underpricing: The issuer’s view – A comment. Journal of Finance, 44, 1099–1103. Barry, C. B., Muscarella, C. J., Peavy, J. W., III, & Vetsuypens, M. R. (1990). The role of venture capital in the creation of public companies. Journal of Financial Economics, 27, 447–471. Betsch, O., Groh, A. P., & Schmidt, K. (2000). Gr¨undungs- und Wachstumsfinanzierung innovativer Unternehmen. M¨unchen: Oldenbourg. Black, B. S., & Gilson, R. J. (1998). Venture capital and the structure of capital markets: Banks vs. stock markets. Journal of Financial Economics, 47, 243–277.
Underpricing of Venture-Backed and Non Venture-Backed IPOs
227
Booth, J. R., & Smith, R. L. (1986). Capital raising, underwriting and the certification hypothesis. Journal of Financial Economics, 15, 261–281. Bottazzi, L., & Da Rin, M. (2002). Europe’s new stock markets. Working Paper, University of Bocconi and University of Turin. Brennan, M. J., & Franks, J. (1997). Underpricing, ownership and control in initial public offerings of equity securities in the UK. Journal of Financial Economics, 45, 391–413. Bundesverband deutscher Kapitalbeteiligungsgesellschaften – German Venture Capital Association e. V. (BVK), 2003. BVK Statistics 2002. Bygrave, W. D., & Timmons, J. A. (1992). Venture capital at the crossroads. Boston: Harvard Business School Press. Carter, R., & Manaster, S. (1990). Initial public offerings and underwriter reputation. Journal of Finance, 45, 1045–1067. Cumming, D. J., & MacIntosch, J. G. (2003). A cross-country comparison of full and partial venture capital exits. Journal of Banking & Finance, 27, 511–548. Engel, D. (2002a). Firm level implications of early stage venture capital investment – An empirical investigation. ZEW Discussion Paper No. 02–82. Engel, D. (2002b). The impact of venture capital on firm growth: An empirical investigation. ZEW Discussion Paper No. 02–02. Fenn, G. W., Liang, N., & Prowse, S. (1997). The private equity market: An overview, Financial Markets. Institutions and Instruments, 6, 1–106. Francis, B. B., & Hasan, I. (2001). The underpricing of venture and nonventure capital IPOs: An empirical investigation. Journal of Financial Services Research, 19, 93–113. Franzke, S. A., Grohs, S., & Laux, C. (2004). Initial public offerings and venture capital in Germany. In: J. P. Krahnen & R. H. Schmidt (Eds), The German Financial System (pp. 233–260). Oxford: Oxford University Press. Franzke, S., & Schlag, C. (2002). Over-allotment options in IPOs on Germany’s Neuer Markt – An empirical investigation. CFS Working Paper No. 2002/16. Gompers, P. A. (1996). Grandstanding in the venture capital industry. Journal of Financial Economics, 42, 133–156. Green, R. C. (1997). Econometric analysis. London: Prentice-Hall. Habib, M. A., & Ljungqvist, A. P. (2001). Underpricing and entrepreneurial wealth losses in IPOs: Theory and evidence. Review of Financial Studies, 14, 433–458. Hamao, Y., Packer, F., & Ritter, J. R. (2000). Institutional affiliation and the role of venture capital: Evidence from initial public offerings in Japan. Pacific-Basin Finance Journal, 8, 529–558. Johnson, S. (2000). Private contracts and corporate governance reform: Germany’s Neuer Markt. Working Paper, Massachusetts Institute for Technology. Kaserer, C., & Kempf, V. (1995). Das Underpricing-Ph¨anomen am deutschen Kapitalmarkt und seine Ursachen. Zeitschrift f¨ur Bankrecht und Bankwirtschaft, 1, 45–69. Kaserer, C., & Kraft, M. (2000). The cost of raising capital and issue size effects – The case of initial public offerings in Germany. Working Paper, University of Fribourg and University of Wuerzburg. Lerner, J. (1995). Venture capitalists and the oversight of private firms. Journal of Finance, 50, 301–318. Lin, T. H., & Smith, R. L. (1998). Insider reputation and selling decisions: The unwinding of venture capital investments during equity IPOs. Journal of Corporate Finance, 4, 241–263. Ljungqvist, A. P. (1997). Pricing initial public offerings: Further evidence from Germany. European Economic Review, 41, 1309–1320. Ljungqvist, A. P. (1999). IPO Underpricing, wealth loss and the curious role of venture capitalists in the creation of public companies. Working Paper, Oxford University.
228
STEFANIE A. FRANZKE
Ljungqvist, A. P., Jenkinson, T., & Wilhelm, W. J. (2003). Global integration in primary equity markets: The role of U.S. banks and U.S. investors. Review of Financial Studies, 16, 63–100. L¨offler, G. (2000). Zeichnungsrenditen am Neuen Markt: Gleichgewicht oder Ineffizienz? Working Paper, Frankfurt University. Loughran, T., & Ritter, J. R. (2002). Why don’t issuer get upset about leaving money on the table in IPOs? Review of Financial Studies, 15, 413–444. Megginson, W. L., & Weiss, K. A. (1991). Venture capitalist certification in initial public offerings. Journal of Finance, 46, 879–903. Prabhala, N., & Puri, M. (1998). How does underwriter price support affect IPOs? Empirical evidence. Working Paper, Yale University. Rindermann, G. (2003). Venture capital participation and the performance of IPO firms: Empirical evidence from France, Germany and the UK. Working Paper, University of Muenster. Ritter, J. R. (1984). The hot issue market of 1980. Journal of Business, 57, 215–240. Ritter, J. R., & Welch, I. (2002). A review of IPO activity, pricing and allocations. Journal of Finance, 57, 1795–1828. Sahlman, W. A. (1990). The structure and governance of venture-capital organizations. Journal of Financial Economics, 27, 473–521. Sch¨afer, D., Werwatz, A., & Zimmermann, V. (2003). The determinants of debt and (private) equity financing in young innovative SMEs: Evidence from Germany. Working Paper, German Institute for Economic Research (DIW) Berlin and KfW Group. Smart, S. B., & Zutter, C. J. (2003). Control as a motivation for underpricing: A comparison of dualand single-class IPOs. Journal of Financial Economics, 69, 85–110. Stedler, H. (1987). Venture capital und geregelter Freiverkehr: Eine empirische Studie. Frankfurt am Main: Fritz Knapp Verlag. Tinic, S. M. (1988). Anatomy of initial public offerings on common stock. Journal of Finance, 43, 789–822. Tykvova, T. (2003). Is the behaviour of German venture capitalists different? Evidence from the Neuer Markt. CFS Working Paper No. 2003/24. Uhlir, H. (1989). Der Gang an die B¨orse und das Underpricing-Ph¨anomen. Zeitschrift f¨ur Bankrecht und Bankwirtschaft, 1, 2–16. Wasserfallen, W., & Wittleder, C. (1994). Pricing initial public offerings: Evidence from Germany. European Economic Review, 38, 1505–1517.
229
Underpricing of Venture-Backed and Non Venture-Backed IPOs
APPENDIX Table A.1. Data Set Neuer Markt – March 10, 1997 to December 31, 2001.a Year
Number of VentureBacked IPOs
VC in %
Number IPOs Backed by Bridge Financing
BF in %
Number of Non VentureBacked IPOs
NVC in %
Total
1997 1998 1999 2000 2001
3 10 30 34 2
37.50 27.03 24.79 27.42 20.00
1 6 23 29 2
12.50 16.22 19.01 23.39 20.00
4 21 68 61 6
50.00 56.76 56.20 49.19 60.00
8 37 121 124 10
Total
79
26.33
61
20.33
160
53.33
300b
a There
were no IPOs in 2002 until March 26.
b The original sample consisted of 353 issues, 53 of which had to be deleted due to either data problems,
extreme values for issue size, or because the issue merely represented a change of market segment.
Table A.2. The Best Rated Lead Underwriters. Underwriter
Deutsche Bank AG Bayerische Hypo- und Vereinsbank AG (Bayerische Hypotheken- u. Wechselbank/Bayerische Vereinsbank) Commerzbank AG DG BANK AG Dresdner Bank AG BHF-Bank AG/ING Group Goldman Sachs West LB Girozentrale HSBC Trinkaus & Burkhardt KGaA Sal. Oppenheim jr. & Cie. Credit Suisse First Boston Gontard & MetallBank AG (Heinrich Gontard & Co. OHG/Metallbank GmbH)
Rating 1997
Rating 1998
Rating 1999
Rating 2000
Rating 2001
1.0 1.0
1.0 2.0
1.0 1.0
1.0 1.0
1.0 1.0
1.0 2.0 2.0 2.0 3.0 3.0 3.0 4.0 4.0 4.0
2.0 2.0 1.5 2.5 3.0 2.5 3.0 3.5 2.0 3.0
2.0 1.0 1.5 2.0 2.0 2.0 3.0 2.0 2.5 2.5
1.0 1.0 1.0 2.0 1.5 2.0 2.0 1.5 2.5 2.0
1.0 1.0 1.0 2.0 1.5 1.5 1.5 2.0 2.5 2.0
Note: This table contains the twelve best rated underwriters serving (more than five times) as lead underwriter at the Neuer Markt during the time period 1997–2001. The underwriter rating of the year 1997 is based on the relative market share of IPO deals on all Frankfurt stock market segments since 1990. Ratings of the years 1998, 1999 and 2000 are based on the track record of each underwriter concerning the relative market share of IPO deals on all Frankfurt stock market segments since 1990 and the relative share of proceeds of IPOs on the Neuer Markt since 1997. A top rating is equivalent to one, the lowest rating equals the value of four (non rated).
230
STEFANIE A. FRANZKE
Table A.3. The Best Rated Venture Capitalists/Private Equity Companies. VC/Private Equity Companies
3i Group Plc./3i Deutschland Apax Partners & Co. Beteiligungsberatung AG Atlas Venture Germany Deutsche Beteiligungs(gesellschaft) AG BdW Beteiligungsgesellschaft f¨ur die deutsche Wirtschaft mbH & Co. KG VC Baden-W¨urttemberg GmbH WestKB Westdeutsche Kapitalbeteiligungs mbH Gold-Zack AG Advanced European Technologies N. V. TVM Techno Venture Management GmbH Commerz Unternehmensbeteiligungs AG TFG Venture-Capital AG & Co. KGaA
Founded in
Backed IPO Companies (as Lead VC)
Rating
1945 1969
21 (16) 5 (2)
1 1
1980 1965 1969
5 (2) 3 (2) 2 (1)
1 1 1
1970 1969
2 (2) 2 (1)
1 1
1990 1995 1983 1987 1994
13 (12) 7 (6) 6 (3) 6 (2) 5 (4)
2 upgrade to 1 in 1999 2 upgrade to 1 in 1999 2 upgrade to 1 in 2000 2 upgrade to 1 in 1998 2 upgrade to 1 in 2000
Note: The rating representing the quality of the venture capitalists and private equity companies is mainly based on the age of the company. Venture capitalists and private equity companies founded before 1983 received a very good rating (equal to one), companies founded in the period 1983–1996 received a mediate rating (equivalent to two). In six cases (Advanced European Technologies N. V., Commerz Unternehmensbeteiligungs AG, TFG Venture-Capital AG & Co. KGaA Unternehmensbeteiligungsgesellschaft, Gold Zack AG, TVM Techno Venture Management GmbH and Schroders) there is an upgrade in rating due to high backing activity during the sample period.
THE PERFORMANCE OF VENTURE-BACKED IPOS ON EUROPE’S NEW STOCK MARKETS: EVIDENCE FROM FRANCE, GERMANY AND THE U.K. Georg Rindermann ABSTRACT This chapter investigates the impact of venture capitalists on the operating and market performance of firms going public on the French Nouveau March´e, the German Neuer Markt and the British techMARK. Considering different variables that reflect the quality of venture-backing, the findings suggest that venture-backed firms do not generally outperform those without venture-backing. However, a subgroup of internationally operating venture capitalists has positive effects on the performance of portfolio firms. The outcome is interpreted as evidence of heterogeneity among venture capitalists in the European market.
1. INTRODUCTION Venture capitalists are specialized financial intermediaries that provide more than capital to young firms (Kortum & Lerner, 2000). Besides high-risk funding they The Rise and Fall of Europe’s New Stock Markets Advances in Financial Economics, Volume 10, 231–294 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3732/doi:10.1016/S1569-3732(04)10010-8
231
232
GEORG RINDERMANN
can add value through pre-investment screening, monitoring and management support (Gompers & Lerner, 1999, 2001; Hellmann & Puri, 2002; Kaplan & Str¨omberg, 2003). Moreover, they tend to concentrate on those industries and stages in the development of companies where their expertise and monitoring ability adds greatest value. Finally, venture capitalists frequently exit their investments through capital markets (Lerner, 1994). Empirical evidence suggests that they choose the exit channel strategically and build up reputation primarily through successful IPOs (Gompers, 1996). Venture capitalists backing firms going public can also push firm performance by building relationships with top-tier financial institutions that at least partly mitigate informational asymmetries at the time of the IPO. Since venture capitalists tend to hold significant ownership and board positions (Barry et al., 1990), and continue to be involved in the firm after going public (Megginson & Weiss, 1991), they might also be able to provide better access to capital in the post-IPO period. Finally, venture capitalists tend to put effective management structures in place, which help firms to perform better in the long run (Brav & Gompers, 1997). Given these characteristics venture capitalists should be able to select high quality firms. Accordingly, the involvement of venture capitalists in IPO firms is conjectured to have a positive influence on their post-issue operating and market performance (Barry et al., 1990; Lerner, 1994; Megginson & Weiss, 1991). Prior U.S. studies report that venture-backed IPOs outperform non venture-backed issues in terms of operating and long run performance (Brav & Gompers, 1997; Jain & Kini, 1995). However, the empirical evidence with respect to the influence of venture capital on the performance of European IPOs is scarce. Furthermore, the results of existing studies on operating performance are not consistent with prior U.S. contributions (Bottazzi & Da Rin, 2002a, b). The mixed evidence indicates differences across Europe with respect to the quality of venture-backing and emphasizes the necessity for a refined assessment, taking into account the heterogeneity of venture capital in different European countries. This chapter investigates the role of European venture capitalists. The main question is whether or not venture capitalists in Europe have a positive impact on the operating and long-run market performance of firms they bring public. To examine this issue, we use a hand-collected data set of venture- and non venture-backed IPOs on the French Nouveau March´e, the German Neuer Markt, and the British techMARK dating from 1996 to 1999.1 The analysis focuses on differences in the issuer and offering characteristics as well as in balance sheet data. Moreover, we assess the influence of venture capitalists via a number of variables reflecting the quality of venture-backing, such as the pre- and post-issue shareholdings, board membership, age, syndication, organizational form, and overall participation in sample firms. The involvement
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
233
of venture capitalists in IPOs across countries is considered as an additional proxy for the experience of investors. Overall, this chapter provides a contribution to the growing number of international studies in the field of venture capital finance as well as to the literature on the operating and market performance of IPOs. The main findings suggest that there are substantial variations in the experience and sophistication of venture capitalists. In particular, international venture capitalists are on average older than national ones, back a larger number of IPOs in the sample, are often represented on the board, invest with a higher number of syndication partners, and hold larger equity positions in portfolio firms. The results of the multivariate analysis suggest that venture-backed IPOs do not generally outperform non venture-backed issues, regardless of the performance measure used. However, a subset of international venture capitalists appears to have positive effects on both the operating and market performance of portfolio firms. The results are interpreted as evidence for the heterogeneity among venture capitalists operating in the European market and emphasize the limited maturity of the young European venture capital industry that is currently undergoing a consolidation process. The remainder of the chapter is structured as follows. Section 2 discusses characteristics of the data set of IPOs, providing details on sample criteria, sources of data, and sample composition. Section 3 describes the methodology and measurement of variables. Section 4 presents the empirical results. Section 5 summarizes the main findings of the chapter and provides concluding remarks and recommendations for future research.
2. DATA 2.1. Data Collection and Data Sources Our sample consists of IPOs from France, Germany and the United Kingdom. We limit our attention to the stock markets of the largest European economies with a sufficient number of IPO firms to make comparisons meaningful. The analysis uses data from non-financial corporations going public in the three countries between 1996 and 1999.2 The data include IPOs on the Nouveau March´e in Paris, the Neuer Markt in Frankfurt, as well as the Official Listing and the Alternative Investment Market (AIM) of the London Stock Exchange (LSE). The growth segments of the Paris and Frankfurt stock exchanges started their activities in 1996 and 1997, respectively. They were created with the common objective to attract listings of innovative companies in high-growth industries. The LSE, however, does not have a separate growth segment.3 In order to obtain a homogenous
234
GEORG RINDERMANN
sample of “venturable” high technology companies, we only include IPOs of U.K. firms included in the techMARK sector. The final sample consists of 303 IPOs in France, Germany, and the United Kingdom. It includes only IPOs, and does not contain re-admissions, transfers from other stock markets or market tiers or mergers and demergers of previously listed firms. IPO firms that have been delisted are included until the date of delisting to avoid any survivorship bias. We exclude all firms without consistently reported accounting data for the periods around the IPO, i.e. one fiscal year before and two fiscal years after the IPO.4 The final sample used in the empirical analysis includes only firms with available data for at least two consecutive financial years, i.e. one year before and after the going public. The data is collected from various sources. Information about the offering terms and characteristics of each IPO, such as the size and offer price of the offering as well as the name of the underwriters, are gathered from Deutsche B¨orse, Euronext Paris, and the LSE. The IPO data are checked and, if necessary, completed with information from the offering prospectuses and the IFR Platinum New Issues database. The aftermarket stock prices of the firms going public, such as the first-day closing price and the monthly closing prices over different time intervals after the IPO, as well as the values of different benchmark indices, were obtained from Thomson Financial Datastream. We use annual statistics on the numbers of IPOs and IPO proceeds of investment banks in France, Germany and the U.K. to measure the reputation of underwriters. This data comes from the annual league tables of IFR Thomson Financial. We collect firm specific information, such as the date of incorporation, pre- and post-IPO accounting data, ownership structure and board membership from IPO prospectuses, annual reports, and Internet websites of the firms going public. For companies listed at the Neuer Markt most of the IPO prospectuses are available as downloads at the website of the stock exchange. Prospectuses of firms listed on the Nouveau March´e are obtained from the firms’ web sites or photocopied at the French stock exchange regulator in Paris. Prospectuses of firms in the U.K. sample that are not available online are sent by the companies or obtained via investment banks. The sample is divided into venture- and non venture-backed issues. Panels A–C in Appendix A list the names and origins of the venture capital firms that are the most frequently involved in IPOs in each country sample. To characterize the quality of venture capital organizations, we determine their age at the time of the IPO as well as their international involvement in firms outside their country of origin. Moreover, we collect the share of equity held by venture capitalists before and after the IPO to explore whether the quality of monitoring or certification is related to the size of the equity stakes held by venture investors. In addition,
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
235
the percentage of equity sold by venture capital is gathered to compute the selling intensity. Appendix B presents a detailed description of the variables used throughout the analysis.
2.2. Sample Characteristics Panel A of Table 1 provides an overview of the distribution of the sample by country and year, both in terms of the number of IPOs and gross proceeds. It shows that the number and value of IPOs are not evenly distributed over the sample period. The time series figures illustrate that the annual number of going publics increased continuously over time and achieved a peak with a total of 152 IPOs in 1999. Thus, 50.2% of all IPOs and 62.1% of the proceeds come from 1999. The statistics also reveal that more than half of the IPOs, both in terms of the number of issues and gross proceeds, took place on Germany’s Neuer Markt. Panel B shows that the overall composition of the sample, with regard to venture- and non venture-backed IPOs is relatively equal. In total, 154 firms or 50.8% of the IPO sample are venture-backed. The patterns in the different country samples appear fairly balanced as well. The average portion of venture-backed IPOs varies between 43% in Germany and 60% in France. The figures by year reveal that about 75% of the IPOs taking place at the British techMARK in 1998 were venture-supported. Likewise, a relatively high portion of Nouveau March´e IPOs in 1999 was venture-backed (74.1%).5 The industry distribution figures in Panel C show that the majority of both venture- and non venture-backed IPO operate in information technology (IT), software or Internet-related industries. Firms from the technology sector form the second largest industry in both groups, representing about one fifth of each subsample. Interestingly, a relatively high portion (15.6%) of firms going public at the Nouveau March´e belongs to traditional industries.6 At the Neuer Markt, only 8 firms representing 5.3% of the German sample are related to traditional businesses. The British IPO sample does not contain any firm with traditional activity background. This is due to the focus of the techMARK that favors firms from non-traditional industries. Finally, the comparison between the venture- and non venture-backed IPOs reveals that in all three countries the majority of firms from the biomedical sector is venture-backed. Panel D displays the frequency distribution of the venture-backed IPOs by the type of the lead venture capitalist. The overall figures show that 59.7% of the lead venture capitalists backing IPOs are independent, whereas 26% of them are captive organizations. Only 14.3% of the venture-supported firms received funding from a public lead venture capitalist.7 A closer look at the figures on the country level
236
Table 1. Distribution of IPOs in the Sample. Panel A: Distribution of IPOs by Year France
Germany
No. of IPOs
Proceeds in m d
1996 1997 1998 1999
13 (14.4) 13 (14.4) 37 (41.1) 27 (30.0)
All
90 (100.0)
U.K.
No. of IPOs
Proceeds in m d
197.7 (15.4) 109.1 (8.5) 427.5 (33.3) 549.5 (42.8)
8 (5.3) 32 (21.2) 111 (73.5)
373.7 (5.1) 1,361.3 (18.4) 5,658.9 (76.5)
1,283.8 (100.0)
151 (100.0)
7,393.9 (100.0)
All IPOs
Proceeds in m d
No. of IPOs
Proceeds in m d
18 (29.0) 18 (29.0) 12 (19.4) 14 (22.6)
530.5 (12.3) 811.3 (18.8) 1,118.6 (25.9) 1,866.2 (43.1)
31 (10.2) 39 (12.9) 81 (26.7) 152 (50.2)
728.2 (5.6) 1,294.2 (10.0) 2,907.3 (22.4) 8,074.5 (62.1)
62 (100.0)
4,326.6 (100.0)
303 (100.0)
13,004.2 (100.0)
No. of IPOs
Panel B: Time Series Distribution of Venture- and Non Venture-Backed IPOs France
Germany
U.K.
All IPOs
NVC
VC
NVC
VC
NVC
VC
NVC
1996 1997 1998 1999
8 (61.5) 6 (46.2) 20 (54.1) 20 (74.1)
5 (38.5) 7 (53.8) 17 (45.9) 7 (25.9)
– 5 (62.5) 16 (50.0) 44 (39.6)
– 3 (37.5) 16 (50.0) 67 (60.4)
9 (50.0) 10 (55.6) 9 (75.0) 7 (50.0)
9 (50.0) 8 (44.4) 3 (25.0) 7 (50.0)
17 (54.8) 21 (53.8) 45 (55.6) 71 (46.7)
14 (45.2) 18 (46.2) 36 (44.4) 81 (53.3)
All
54 (60.0)
36 (40.0)
65 (43.0)
86 (57.0)
35 (56.5)
27 (43.5)
154 (50.8)
149 (49.2)
GEORG RINDERMANN
VC
Biomed ITSINT Media Techno Telecom Traditional
7 (13.0) 18 (33.3) 4 (7.4) 11 (20.4) 6 (11.1) 8 (14.8)
5 (13.9) 13 (36.1) 6 (16.7) 5 (13.9) 1 (2.78) 6 (16.7)
7 (10.8) 31 (47.7) 6 (9.2) 17 (26.2) 2 (3.1) 2 (3.1)
3 (3.5) 41 (47.7) 13 (15.1) 19 (22.1) 4 (4.7) 6 (7.0)
8 (22.9) 22 (62.9) 1 (2.9) 3 (8.6) 1 (2.9) 0 (0.0)
4 (14.8) 13 (48.2) 0 (0.0) 3 (11.1) 7 (26.9) 0 (0.0)
22 (14.3) 71 (46.1) 11 (7.1) 31 (20.1) 9 (5.8) 10 (6.5)
12 (8.1) 67 (45.0) 19 (12.8) 27 (18.1) 12 (8.1) 12 (8.1)
All
54 (100.0)
36 (100.0)
65 (100.0)
86 (100.0)
35 (100.0)
27 (100.0)
154 (100.0)
149 (100.0)
Note: Percentages in parentheses. Panel D: Venture Capital Type Distribution of Venture-Backed IPOs France
Germany
U.K.
All IPOs
No.
%
No.
%
No.
%
No.
%
Independent Captive Public
18 23 13
33.3 42.6 24.1
40 16 9
61.5 24.6 13.9
34 1 0
97.1 2.9 0.0
92 40 22
59.7 26.0 14.3
All
54
100.0
65
100.0
35
100.0
154
100.0
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
Panel C: Industry Distribution of Venture- and Non Venture-Backed IPOs
237
238
GEORG RINDERMANN
reveals that the frequency distribution by venture capital type varies across the national subsamples. In the U.K., almost all venture capital firms that back IPO firms are independent organizations. By contrast, independent venture capitalists backed 61.5% in Germany and only one third of the venture-funded IPOs in France. Instead, captive venture capitalists play a more important role in both countries. They are involved in 42.5 and 24.6% of the venture-backed transactions at the Nouveau March´e and Neuer Markt, respectively. Finally, the statistics show that the portion of public venture capital firms involved in IPOs varies significantly across countries. In France, their contribution is the largest (24.1%), whereas in Germany, public venture capitalists are involved in 13.9% of all venturebacked IPOs. In the U.K., however, public venture capitalists do not participate in any IPO.
2.3. Selection Issues There are at least two potential sources of bias in the data to worry about. First, looking at venture-backed IPO firms has the limitation of ignoring those remaining private, being sold in a trade sale, or failing. In Europe, divestments via an IPO traditionally represent a relatively small fraction of the exits chosen by venture capitalists (EVCA, 2000). Instead, most European venture capitalists prefer trade sales to IPOs as an exit vehicle. Venture-backed firms going public are typically among the most successful ones since the rewards for venture capitalists are the highest both in reputation and monetary terms (Gompers, 1995; Gompers & Lerner, 1997). As a consequence, the impact of venture-backing on performance might be overestimated by looking only at the firms that go public. Yet, given that firms considering an IPO are in general the most promising and profitable ones, the potential bias towards the more successful firms may also apply to the control sample of non venture-backed firms. Second, venture-backing is not randomly distributed but typically represents an endogenous choice by entrepreneurs and venture capitalists. Not all entrepreneurs desire venture capital financing and not all entrepreneurs receive it. Hence, there might be a selection bias in the receipt of venture capital funding (Lee & Wahal, 2002). The endogenous choice in providing financing is also reflected in the eventual exit from the entrepreneurial venture, and might lead to a non-random distribution of venture-backed IPOs. In particular, the preference of venture capitalists to finance specific types of firms from a particular range of industries might be reflected in both firm and IPO characteristics, such as the size and age of firms (Gompers & Lerner, 2001; Lee & Wahal, 2002). However, the composition of our sample indicates that the stock market segments and IPOs provide
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
239
appropriate control samples for the purpose of a comparison between ventureand non venture-backed firms. The figures show that the overall frequency distribution of venture- and non venture-backed firms appears relatively well balanced with respect to the represented industries. To capture effects, which are due to the preferences of venture capitalist for particular firm types and industries, the multivariate analysis employs controls for age, size and industry of the IPO firm.
3. METHODOLOGY AND VARIABLE MEASUREMENT 3.1. Accounting Performance Measures In the analyses, we use different performance measures as dependent variables. First, we explore potential differences in the accounting performance between venture- and non venture-backed IPOs. We employ two different kinds of profitability ratios that are widely used as accounting performance measures: (1) operating return on assets; and (2) operating cash flow return on assets. Both ratios are efficiency measures on how a firm is being run and provide information on how much returns are generated by each unit of assets. In addition, both variables measure flows on a pre-tax and pre-interest basis, and avoid the mechanical effect of leverage on the results. This controls for the effects of differences in the capital gain taxes across the three countries. The operating return on assets is defined as the operating income before interest, taxes, and extraordinary items divided by total assets. Accordingly, the operating cash flow return on assets is defined as operating cash flow before interest, taxes and extraordinary items normalized by total assets, where the operating cash flow is computed as the operating income plus depreciation, amortization and provisions. The possible advantage of the cash flow-related performance ratio is that it eliminates several accruals. Therefore, the operating cash flow return should be less sensitive to manipulation by managers and exhibit more variability than the first measure. However, operating cash flows can be defined and computed in many different ways. For robustness, we employ different cash flow definitions to calculate the operating cash flow return on assets. Since an IPO is typically accompanied by a substantial increase of total assets, the accounting profitability measures scaled by assets might impart a downward bias after the IPO. Therefore, the operating income and cash flow scaled by sales are examined as alternative accounting profitability ratios. Analogous to the previously defined performance ratios, the return on sales equals the operating
240
GEORG RINDERMANN
income before interest, taxes and extraordinary items divided by total sales. Likewise, the operating cash flow to sales ratio is computed as the operating cash flow before interest, taxes and extraordinary items divided by total sales. All accounting performance ratios are computed annually at three different reference points in time, i.e. at the end of the fiscal year prior to the IPO (t−1 ) and two fiscal years subsequent to the IPO (t1 , t2 ) to assess the pre- and post-issue accounting performance, respectively. Details on the individual accounting performance measures are reported in Panel A of Appendix B. Based on Mikkelson et al. (1997), the following empirical model specification is used to test the association between operating performance and the involvement of venture capitalists in IPOs: CFROAi = 0 + 1 VCi + 2 ln Sizei + 3 ln Agei + 4 Parti + 5 URanki + i (1) where the dependent variable CFROA is the operating cash flow return on assets. The explanatory variables include a dummy that equals one if an IPO firm is venture-backed (VC), the natural logarithm of assets (ln Size), the natural logarithm of firm age (ln Age), calculated in years at the time of the IPO, the participation ratio of pre-IPO shareholders (Part), defined as number of secondary shares sold in the IPO divided by the number of shares outstanding before flotation, and the rank of the underwriter reputation (URank).8 Size and age are expected to have a positive influence on operating performance, since small and young firms usually display low accounting performance due to their small scale of operations, one-time start-up costs and high initial operating costs, high production and selling costs and low volumes of sales. Moreover, they often price products at a smaller margin over cost than large established firms to attract new clients. The portion of secondary sales in the IPO is measured by the participation ratio, a proxy that is often used in the IPO literature (Barry, 1989; Habib & Ljungqvist, 2001). Mikkelson et al. (1997) argue that firms undertaking an IPO including secondary sales must have favorable prospects. We therefore expect the participation ratio to be positively correlated with the post-offering operating performance. Assuming that underwriters certify firm quality, we expect to find a positive relation between post-IPO performance and underwriter prestige. 3.2. Market Performance Measures The operating performance of venture-backed firms tends to be low or negative in the first years of corporate history. They often have negative cash flows at the time of an IPO, and the (partial) exit from a venture is typically the primary way
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
241
for venture capitalists to realize a positive return. Consequently, a mere focus on accounting profitability ratios might be misleading in assessing the performance of venture-supported IPOs. Since the value of high-tech firms is related to growth opportunities and expectations of future profitability, it seems appropriate to examine whether the market recognizes the value-added potential of venture capital participation in firms going public. Therefore, we analyze whether the market has higher expectations of future earnings performance from venture- than from non venture-backed firms. 3.2.1. Tobin’s Q In the analysis, we use an approximation of Tobin’s Q as a measure of the stock market valuation at the end of the first trading day. Tobin’s Q captures the expectations of investors in the stock market. Numerous empirical studies in the financial economics literature apply Tobin’s Q to categorize firms according to their relative performance (Himmelberg et al., 1999; Lang et al., 1991; Lindenberg & Ross, 1981; Mørck et al., 1988). The measure is used as a proxy for companies’ future investment opportunities and as an indicator of intangible value. Hence, a high value of Tobin’s Q indicates that investors place a high valuation on the future growth opportunities of the company. Tobin and Brainard (1968) and Tobin (1969) define Tobin’s Q as the ratio of the market value of outstanding financial claims to the current replacement costs of assets. This definition assumes that replacement costs are a logical measure for the value of the alternative use of assets. Since Tobin’s Q represents the present value of future cash flows divided by the replacement costs of intangible assets, no risk adjustment is necessary to compare the ratio across firms (Lang & Stulz, 1994). Moreover, as the numerator of Tobin’s Q contains market values, it is less sensitive to discretion of managers and superior to pure accounting-based measures. In addition to pure accounting-related information, long-term improvements are taken into account, such as growth opportunities and other events that do not immediately affect the accounting statements or cash flows of a firm. On the basis of the collected data, the precise Tobin’s Q measure of equity at replacement costs is not available. Therefore, an approximation of the ratio is computed, defined as the market value of equity plus the book value of debt divided by the book value of total assets. This way to calculate Tobin’s Q, which is sometimes referred to as “simple Q,” has been used in previous studies (Loderer & Martin, 1997). There is evidence that approximations to the measurement of the Q ratio tend to yield similar values for Tobin’s Q (Chung & Pruitt, 1994). In the present analysis, we compute an approximation of the “simple Q” ratio at the first trading day for each IPO firm. The calculation method is explained in more detail in Appendix C.
242
GEORG RINDERMANN
Based on the model specifications and control variables employed in previous studies on Tobin’s Q, such as Mørck et al. (1988) and Cho (1998), the following regression equation is estimated: Q i = 0 + 1 VCi + 2 ln Sizei + 3 EqRi + 4 S Ai + 5 Alphai + i
(2)
where the dependent variable is the approximated value of Tobin’s Q at the first trading day. On the right-hand side of the equation, we use a dummy variable (VC) for the involvement of venture capitalists in the IPO. Firm size (ln Size), calculated as the natural logarithm of the market capitalization at the end of the first trading day, and the equity ratio (EqR), defined as the book value of equity divided by the book value of total assets at the fiscal year prior to the IPO, are included as controls. The equity ratio represents a widely used proxy for financial risk and has been used in prior studies as control variable of Tobin’s Q (Cho, 1998; Holderness et al., 1999). Additionally, a variable controlling for the importance of the firms’ soft capital, quantified by the ratio of sales over assets (S A), is included in the regression model. Himmelberg et al. (1999) argue that firms with higher sales to capital ratios are less easily monitored. To account for ownership, the fraction of the firms’ equity retained by the old shareholders immediately after the going public (Alpha) is employed as a control variable. Ideally, the regression should also control for the firms’ R&D and advertising expenditures, but for a majority of firms in our sample these figures are not available. 3.2.2. Buy and Hold Returns and Wealth Relatives In addition to Tobin’s Q, the buy and hold returns and wealth relatives are calculated for each IPO over different time intervals in the aftermarket period. These measures have been employed in many studies on the long-term performance of IPOs (Brav & Gompers, 1997; Loughran & Ritter, 1995; Ritter, 1991). The aftermarket performance covers the period from the first day of trading onwards, and the buy and hold return (BHR) for each IPO firm is calculated by compounding monthly returns up to 36 months after the IPO date: T BHRi,T = (1 + R i,t ) − 1 (3) t=1
where Ri,t denotes the monthly return of firm i in month t over the time interval T. If a firm was delisted before the end of the 36-month holding period, the return is compounded until the delisting date. The monthly closing prices, which are used to calculate the returns, are adjusted for dividends, stock splits, and new share issues. Given the movements of the general market over time, the raw buy and hold returns of IPOs have to be adjusted by using an appropriate benchmark. Wealth
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
243
relatives present one way to compare the buy and hold returns of IPOs to different benchmarks, e.g. the market index of the corresponding growth market segment. Based on Ritter (1991), Loughran and Ritter (1995), and Brav and Gompers (1999), wealth relatives (WR) are computed as follows: T (1 + R i,t ) (1 + BHRi,T ) = Tt=1 WR = (1 + BHRm,T ) t=1 (1 + R m,t )
(4)
Wealth relatives greater than one can be interpreted as IPOs outperforming the overall market, whereas wealth relatives smaller than one indicate underperformance to the benchmark. Several contributions, such as Barber and Lyon (1997), and Lyon et al. (1999), advocate the use of portfolios of the same firm size and book-to-market ratios to compute wealth relatives. However, given the relatively small number of IPO firms per country, the construction of such portfolios is beyond the scope of the present study. Instead, the analysis employs the stock market indices of the different domestic growth markets as benchmark portfolios, i.e. the (1) Nouveau March´e index; (2) Nemax All Share index; and (3) techMARK All Share index, for IPOs on the Nouveau March´e, Neuer Markt, and techMARK, respectively. There are two concerns with respect to the consistency and adequacy of these market indices. First, the techMARK All Share index also includes firms going public prior to the sample period, and therefore differs from the composition of the Nemax All Share and Nouveau March´e indices. To examine the effect of index constitution, we calculate synthetic equally weighted market indices for each country on the basis of the sample firms and compare these indices to the official growth market indices. Figure 1 shows both types of indices in Panels A–C. In line with expectations, the synthetic indices of the French (Panel A) and German sample (Panel B) closely follow the official market indices of the respective growth markets, albeit with slightly higher volatilities. This difference might be due to the relatively lower number of firms considered in the sample as well as different weights used for the index calculation. As shown in Panel C, the synthetic market index for the U.K. clearly outperforms the techMARK All Share index between August 1999 and March 2001. Again, this observation might be related to the relatively lower number of firms in the sample than in the stock market index as well as the use of different weights. Second, in the French and German sample, there are concerns for a simultaneity bias related to the use of the official growth market indices as benchmark. Since the calculation of the Nemax All Share and Nouveau March´e indices started simultaneously with the launch of the respective growth market segments, they
244
GEORG RINDERMANN
Fig. 1. Comparison of Market Indices and Sample-Based Indices Over Time. (A) Panel A: France (Nouveau March´e Index vs. Sample-Based Index). (B) Panel B: Germany (Nemax All Share Index vs. Sample-Based Index). (C) Panel C: U.K. (techMARK All Share Index vs. Sample-Based Index).
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
245
Fig. 2. Evolution of Different Benchmark Indices Over Time.
include almost all IPOs of the country sample. To make sure that the calculated wealth relatives of IPOs are not due to a simultaneity bias or systematic differences in the composition of the market indices, we use the NASDAQ Composite index as an alternative benchmark to adjust for the overall market development. Although the NASDAQ Composite index does not include any of the sample firms, Fig. 2 shows that it is highly correlated with the indices of the analyzed growth markets and therefore represents an adequate proxy for the overall valuation levels of technology stocks over the sample period. To assess the stock price performance of the IPO firms in the cross-section, we use the following empirical model specification based on Brav and Gompers (1997) and Carter et al. (1998): WRi = 0 + 1 VCi + 2 Q i + 3 ln Sizei + 4 URanki + i
(5)
where the dependent variable WRi denotes the natural logarithm of the 3-year wealth relative, using the respective country’s growth market index or the NASDAQ Composite index as benchmark. The explanatory variables include a dummy variable (VC), taking on the value of one if a firm is venture-backed, the approximated value of Tobin’s Q (Q) and the natural logarithm of the market value of equity (ln Size), both measured at the end of the first trading day, as well as the reputation rank of the underwriter (URank). By controlling for firm size, the analysis follows previous studies, such as Fama and French (1992), who document that size has a negative effect on stock returns. The rationale behind this expected relationship is that investors require a discount, or higher stock returns, for small firms to compensate for higher portfolio risks related to liquidity, information access and other factors. Brav and Gompers (1997) argue that asymmetric information is more likely to be a problem for small firms, as it might not pay for sophisticated investors to do research on these firms.
246
GEORG RINDERMANN
Fig. 3. Stylized Relationship Between Tobin’s Q and Market Performance of IPOs. Note: Q1 (P1 ) and Q2 (P2 ) denote the Tobin’s Q (market performance) of venture- and non venture-backed IPOs, respectively.
The approximated value of Tobin’s Q at the end of the first trading day controls for the relation between firm valuation at the time of the IPO and the long-term performance in the stock market. Previous studies, such as Fama and French (1992) or Brav and Gompers (1997), use the book-to-market ratio after the IPO as explanatory variable in the analysis of stock returns. They argue that the observed explanatory power of book-to-market for stock returns is related to the risk of financial distress. In particular, firms with higher book-to-market ratios tend to have higher returns because their risk of financial distress is higher. However, as the inverse of the book-to-market ratio is by definition highly correlated with Tobin’s Q, only the latter is included in the analysis. Investors may value venture-backed issues higher than non venture-backed firms in terms of Tobin’s Q at the time of the IPO due to the certification role of venture capitalists. We show this hypothesized relationship graphically by the difference Q between Q1 and Q2 in Fig. 3. As a consequence of the relatively lower initial valuation of non venture-backed issues and the declining information asymmetry regarding the firm quality after the IPO, the performance angle of non venture-backed firms (␣2 ) is expected to be steeper in the aftermarket than the one of venture-backed issues (␣1 ). This relationship implies that the stock price performance of venture-backed firms in the aftermarket (P1 ) rests below the performance of non venture-backed firms (P2 ), and that it is necessary to consider Tobin’s Q in the overall performance analysis. To control for the conjectured relationship between Tobin’s Q and the stock price performance in the aftermarket, as well as for significant differences between venture- and non venture-backed issues, the intercept variable Q and an interaction term VC × Q are included in the estimated regressions in Section 4.2. Finally, the quality of the underwriting investment bank is considered in the regressions. Carter et al. (1998) show that the long-term underperformance of
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
247
IPOs handled by more prestigious underwriters is less severe. Accordingly, the prestige of the underwriter, measured by the underwriter rank, is expected to have a positive effect on the performance of IPOs in the long run. More details on the methodology used to assign ranks for the reputation of underwriters are outlined in Section 3.4. The following section, discusses the variables that we use to capture the quality of venture-backing.
3.3. Quality of Venture-Backing There might be differences in the effectiveness of venture-backing. We use several proxies for the quality of venture capital monitoring as controls. These include the representation of venture capitalists on the board of directors (Board) as well as the pre- and post-issue equity shares held by all venture investors (PrEq, PosEq). The number of venture capitalists having equity positions in the IPO firm (NVC) may also have an impact on the quality of venture-backing because the use of syndication typically involves complementary skills of additional venture capitalists and provides a second opinion in the process of the project selection (Sah & Stiglitz, 1986). If these monitoring proxies are related to the quality of oversight, we expect a positive sign of the coefficients in the regressions on the operating and market performance of IPO firms. Similar to the study of Gompers (1996), the age of the lead venture capitalist at the time of the IPO (VC Age) is taken into account as a proxy for reputation and certification.9 If established venture capitalists are able to provide stronger certification than younger ones, the variable is expected to be positively related to IPO valuation in terms of Tobin’s Q. Moreover, we employ the number of IPOs in the sample (No IPOs) in which a venture capitalist is involved as financier as a proxy for experience and reputation. To control for differences in the experience between internationally and nationally operating venture capital investors, a dummy variable (Int VC) is codified for venture capitalists being involved in IPOs in at least two countries of the sample. Finally, we include the percentage of venture shareholdings sold in the IPO (Sale) to find out whether venture capitalists cash out at the time of the IPO. We expect a negative relation between the fraction of equity sold at the time of the IPO and post-issue performance. According to Wang et al. (2002), there might be differences in the quality of venture-backing across different organizational forms of venture capitalists. To control for systematic variations between the quality of venture-backing provided by independent and captive venture capitalists, a dummy variable for firms backed by independent venture capitalists (Indep) is included in the analysis. Based on the distinction made by the EVCA (2000), venture capitalists
248
GEORG RINDERMANN
are defined as independent if less than 20% of their equity is held by a single shareholder. Given the conjecture that public venture capitalists tend to employ staff with little investment skills and to have few incentives to be profitable, an additional dummy variable (Public) is employed for firms backed by public venture capital organizations. A venture capitalist is characterized as being public, if the ownership structure indicates that public bodies have a direct or indirect influence on the investments of the firm.10 Finally, we include dummies for the national origins of venture capital firms to control for the effects of different management styles across countries (Sapienza et al., 1996).
3.4. Control Variables The empirical analysis includes a number of control variables to avoid specification errors due to omitted variables. Following previous studies on IPO performance, we control for underwriter reputation. Carter and Manaster (1990), Carter et al. (1998), and Loughran and Ritter (2003) account for the quality of underwriters by using a classification methodology that is based on their relative positions in the Tombstone of an offering.11 However, the assigned ranks following this methodology, as listed by Loughran and Ritter (2003), only consider IPOs in the U.S., and do not adequately reflect the reputation of prestigious investment banks operating in European markets. Taking into account this shortcoming, we use a different measure of underwriter reputation. We measure the market share of the lead management in the domestic capital markets of France, Germany and the U.K. from 1995 to 1999, based on the underwritten IPO proceeds.12 On a scale of 1–3, the top rating is 1, and the lowest rating equals the value of 3. We assign the best value of 1 to investment banks with a market share of equal or higher than 5%, the value of 2 to those with a market share between 1 and 5%, and the value of 3 to those with a market share of lower than 1%. To account for the reputation of internationally operating high quality underwriters, which are more active in the U.S. than in the European IPO markets, investment banks classified as prestigious underwriters by Loughran and Ritter (2003) are assigned the value of 1. For an overview, the ranks of the top investment banks involved in IPOs of each country sample are outlined in the Panels A–C of Appendix D. The panels also compare the ranking scores of the method applied in the present analysis to the rankings calculated by Carter et al. (1998) and Loughran and Ritter (2003). We include industry dummy variables to account for industry-related effects. We account for six different industry sectors, based on the classification of the Financial Times. In addition, we include calendar-year dummies to check whether
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
249
IPO markets, and in particular the German Neuer Markt, attracted firms of low quality and provided ineffective screening in the late years of the sample period. To control for effects related cross-country differences, country dummy variables are included in the analysis. We also include dummy variables for the different accounting standards that are used by the firms going public, such as IAS, French, U.K. or U.S. GAAP. Information on the applied accounting standards is typically provided in the offering prospectuses and annual reports of the firms going public. Finally, we use an additional dummy variable for firms that originally went public at the Alternative Investment Market (AIM) before changing to the techMARK of the Official Market. An overview of all control variables used in the empirical analysis is provided in Panel B of Appendix B. In the next section, we discuss the methodology used to detect significant differences between venture- and non venture-backed firms.
3.5. Testing for Differences Between Venture- and Non Venture-Backed Firms In the univariate analysis, the means and medians of the group of venture-backed and of the group of non venture-baked firms are compared. Since the means are sensitive to extreme values, the reported medians help control for this shortcoming. To identify statistically significant differences between the means and medians of the two groups, we conduct Student’s t-tests and non-parametric Mann-Whitney-tests, respectively. In the multivariate analysis, a qualitative explanatory variable for the involvement of venture capitalists in IPOs, i.e. a venture capital dummy (VC), is employed. First, we include a simple intercept dummy variable for the participation of venture capitalists in the regressions to detect potential effects of venture-backing. Second, we use different interaction dummy variables in order to reveal slope effects of venture-backing on firm performance. The interaction dummy variables are the product of the original regressor of each explanatory variable and the binary venture capital dummy variable. To avoid a bias in the estimates of interaction terms, it is vital to use a venture capital intercept dummy in the regressions as well. Alternatively, the Chow-test is applied to identify whether the regressions of venture- and non venture-backed firms are structurally different (Chow, 1960). The Chow-test is a popular method of testing for differences between two or more regressions, representing a special application of the restricted least-squares method. It compares the results of a pooled regression over all observations with those of separate regressions for different subgroups (Jobson, 1991,
250
GEORG RINDERMANN
pp. 317–321). The significance of the Chow-test is assessed by comparing the impact of the pooled regression with individual estimations for each subgroup by an F-test statistic: F=
(RSSpooled − RSSindividual )/k RSSindividual /(n 1 + n 2 − 2k)
(6)
where k is the number of the estimated parameters, n1 and n2 are the numbers of observations of each subgroup, and RSSpooled and RSSindividual are the residual sums of squares of the pooled and individual regressions, respectively. The next section discusses the main empirical findings and provides interpretations regarding the research questions.
4. EMPIRICAL RESULTS 4.1. Univariate Analysis 4.1.1. Firm and Offering Characteristics at the Time of the IPO Before testing the impact of venture capital on different performance measures, we investigate the firm and offering characteristics at the time of the IPO in more detail. Panel A of Table 2 provides the descriptive statistics for the IPOs in the different country samples, comparing the means and medians of the characteristics of venture- and non venture-backed firms. Medians are reported in parentheses. The offering characteristics include the market capitalization at the end of the first trading day (Mkt Cap), the gross proceeds of the offering (IPO Size), the participation of pre-IPO shareholders in the IPO (Part), measured as the shares sold in the IPO by pre-IPO shareholders relative to the number of shares outstanding before flotation, the portion of equity held by old shareholders immediately after the IPO (Alpha) and the approximated Tobin’s Q at the first day of trading (Q). In addition, we investigate firm characteristics, such as total sales (Sales), assets (Assets), net income (Income), equity ratio (EqR) and the number of employees (Empl) at the end of the fiscal year prior to the IPO as well as firm age (Age) at the time of the IPO. Comparing the participation ratios between the two groups of firms indicates that the old shareholders of venture-backed firms in all three countries sold significantly higher portions of shares in the IPO than their non venture-backed counterparts. Accordingly, in France and Germany, the pre-IPO shareholders of venture-backed firms held lower equity stakes immediately after the IPO than firms without venture-backing, as shown by the significant differences in the alpha ratio. However, the remaining firm and offering characteristics do not indicate
Panel A: Comparison of Venture- and Non Venture-Backed IPOs by Countrya Variable
France VCBacked
Mkt Cap IPO Size Part Alpha Q Sales Assets Income EqR Empl Age
U.K.
Non VCBacked
t-Test (z-Score)
VCBacked
Non VCBacked
t-Test (z-Score)
VCBacked
Non VCBacked
t-Test (z-Score)
46.9 (26.0) 10.7 (7.4) 5.8 (6.3) 79.0 (79.4) 2.6 (2.3) 18.4 (9.8) 11.8 (8.4) 0.4 (0.4) 26.6 (23.3) 121 (94) 9.0 (7.5)
0.5 (−0.7) 1.8*** (−1.9*** ) 2.2** (−1.9*** ) −3.4* (−3.2* ) −1.4 (−1.1) 0.1 (−0.5) 1.2 (−1.6*** ) −0.9 (−1.3) 0.9 (−1.5) 0.7 (−0.1) 0.3 (−0.4)
223.2 (164.3) 46.7 (36.3) 15.4 (11.7) 62.2 (66.1) 4.7 (3.9) 25.0 (9.2) 29.0 (11.4) −0.3 (0.0) 30.2 (32.2) 162 (82) 13.0 (10.0)
251.1 (184.9) 50.7 (35.2) 9.3 (6.7) 68.6 (70.7) 4.9 (4.0) 42.3 (16.1) 26.5 (12.4) 0.2 (0.5) −10.6 (21.4) 218 (214) 14.5 (10.0)
−0.8 (−0.9) −0.6 (−0.3) 3.2* (−2.4** ) −3.2* (−2.3** ) −0.5 (−0.3) −1.2 (−1.7*** ) 0.2 (−1.5) −0.3 (−3.1* ) 1.7*** (−1.0) −1.2 (−1.0) −0.7 (−0.9)
231.1 (132.0) 75.4 (35.1) 20.6 (18.5) 62.0 (65.1) 3.8 (3.4) 87.4 (15.7) 39.6 (13.7) 1.6 (1.1) 18.5 (33.6) 349 (113) 9.8 (8.0)
166.6 (58.6) 62.5 (13.6) 10.7 (9.0) 68.4 (70.0) 4.2 (4.1) 33.5 (7.3) 43.2 (5.1) −1.2 (0.4) −32.3 (20.7) 162 (53) 11.6 (5.0)
0.9 (−1.7*** ) 0.4 (−2.3** ) 2.4** (−2.1** ) −1.9*** (−1.4) −0.9 (−1.0) 0.9 (−1.1) −0.1 (−2.4** ) 0.8 (−0.2) 1.5 (−1.3) 1.4 (−2.1** ) −0.5 (−0.8)
251
53.3 (37.2) 16.6 (10.7) 10.0 (9.6) 72.9 (72.3) 2.3 (2.3) 18.9 (10.4) 16.4 (10.7) −0.1 (0.3) 30.3 (26.7) 148 (85) 9.4 (7.0)
Germany
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
Table 2. Firm and Offering Characteristics at the Time of the IPO.
252
Table 2. (Continued ) Panel B: International Comparisonb Variable
Mkt Cap IPO Size Part ratio Alpha Q Sales Assets
EqR Empl Age
Germany vs. U.K.
VCBacked
Non VCBacked
VCBacked
−3.9* (−5.3* ) −4.1* (−5.0* ) −3.6* (−2.5** ) 4.4* (−3.5* ) −5.2* (−4.8* ) −1.7*** (−0.7) −1.9*** (−1.5) −1.3 (−1.5) 1.2 (−0.5) −2.1** (−1.1) −0.2 (−0.5)
−3.1* (−2.6* ) −2.3** (−2.5** ) −1.9*** (−1.0) 4.6* (−4.0* ) −4.4* (−3.5* ) −1.2 (−0.7) −1.5 (−1.3) 0.5 (−0.1) −1.9*** (−0.4) −0.8 (−1.3) −0.8 (−0.8)
−0.2 (−1.0) −1.9*** (−0.3) −1.5 (−1.2) 0.1 (−0.4) 1.9*** (−1.8*** ) −1.7*** (−0.6) −0.6*** (−1.2) −1.6 (−1.5) 1.7 (−0.1) −2.2** (−1.2) −1.4 (−1.2)
Non VCBacked 1.6 (−3.1* ) −0.7 (−3.0* ) −0.7 (−0.3) 0.1 (−0.2) 1.1 (−0.7) 0.4 (−1.8*** ) −1.1 (−2.3** ) 0.5 (−0.5) 0.5 (−0.1) 0.8 (−1.9*** ) 0.9 (−2.4** )
Germany vs. France VCBacked
Non VCBacked
6.3* (−7.4* ) 4.8* (−6.7* ) 2.3** (−1.9*** ) −4.5* (−3.9* ) 7.1* (−6.7* ) 1.1 (−0.1) 1.2 (0.0) −0.5 (−0.9) 0.0 (−0.6) 0.4 (−0.2) 1.7*** (−1.8*** )
4.8* (−7.4* ) 5.2* (−7.6* ) 2.2** (−1.7*** ) −6.0* (−5.8* ) 4.7* (−5.8* ) 1.3 (−2.7* ) 2.1** (−2.0** ) −0.1 (−0.9) −1.1 (−0.7) 1.7*** (−0.7) 2.3** (−2.3** )
GEORG RINDERMANN
Income
France vs. U.K.
for venture- and non venture-backed firms are expressed as means (medians in parentheses). To detect significant differences between the means [medians] of the two firm groups, Student’s t-tests [non-parametric Mann-Whitney tests] were conducted. Bold values indicate statistically significant differences at the 1% or 5% confidence level. The market capitalization of equity (Mkt Cap) is expressed in million d and related to the first day of trading. IPO Size represents the gross proceeds of the IPO incl. exercised overallotments. Part is the participation ratio (in %), defined as number of old shares sold divided by the number of shares outstanding before flotation. Alpha is the percentage of the firm’s equity retained by old shareholders after the IPO. Tobin’s Q (Q) is the approximated value at the end of the first trading day, defined as the sum of the market value of equity and book value of debt, divided by the book value of total assets. Total sales (Sales), assets (Assets), and net income (Income) are expressed in million d and related to the end of the fiscal year prior to the IPO, as well the number of employees (Empl). The equity ratio (EqR) is expressed in percent, defined as the book value of equity divided by the book value of total assets at the end of the fiscal year prior to the IPO. The firm age (Age) is expressed in years at the time of the IPO. b Values for venture- and non venture-backed firms are Student t-tests statistics (non-parametric Mann-Whitney z-scores in parentheses) to detect significant differences between the means (medians) of firm groups in different countries. Bold values indicate statistically significant differences at the 1% or 5% confidence level. The market capitalization of equity (Mkt Cap) is expressed in million d and related to the first day of trading. IPO Size represents the gross proceeds of the IPO incl. exercised overallotments. Part is the participation ratio (in percent), defined as number of old shares sold divided by the number of shares outstanding before flotation. Alpha is the percentage of the firm’s equity retained by old shareholders after the IPO. Tobin’s Q (Q) is the approximated value at the end of the first trading day, defined as the sum of the market value of equity and book value of debt, divided by the book value of total assets. Total sales (Sales), assets (Assets), and net income (Income) are expressed in million d and related to the end of the fiscal year prior to the IPO, as well the number of employees (Empl). The equity ratio (EqR) is expressed in percent, defined as the book value of equity divided by the book value of total assets at the end of the fiscal year prior to the IPO. The firm age (Age) is expressed in years at the time of the IPO. ∗ Significance level is 10%. ∗∗ Significance level is 5%. ∗∗∗ Significance level is 1%.
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
a Values
253
254
GEORG RINDERMANN
any significant differences between venture- and non venture-backed issues that apply to all three countries. It is worth mentioning that venture-backed IPO firms in Europe are older than venture-backed IPO firms in the United States. Gompers (1996) reports that the average venture-backed IPO firm in the United States is between 3 and 4 years old at the time of going public. In contrast, we find that the median age ranges from 7 years in the French sample to 10 years in the German sample. The (unreported) refined univariate analysis that distinguishes between IPOs backed by national and international venture capitalists (defined as venture capitalists backing firms in at least two of the three countries) leads to a slightly different pattern. It indicates that the median firm backed by international venture capitalists has significantly larger IPO proceeds than its equivalent backed by national venture investors. This finding holds for all three countries and suggests that venture investors classified as international are able to raise higher amounts of equity in the capital market than purely national venture investors are. In order to detect significant differences across countries, Panel B of Table 2 outlines the Student’s t- and non-parametric z-scores for international differences in means and medians of the firm characteristics, respectively. The comparison shows that venture- as well as non venture-backed firms in the French sample are significantly smaller than those in Germany and the U.K., both in terms of market capitalization and issue size. In addition, the test-scores of the participation and alpha ratios indicate that the old shareholders of venture-backed firms in France sold significantly lower portions of shares in the IPOs and held more equity shares after flotation than their counterparts in Germany and the United Kingdom. The average Tobin’s Q of firms going public on the Nouveau March´e is significantly lower than in the two other countries. Since the French accounting standards are rather conservative and do typically not yield higher book equity valuations than IAS, U.K. or U.S. GAAP, this finding suggests that firms in the French sample have systematically lower market valuations than issues in the two other countries. On the one hand, the significant differences in valuation between IPOs in France and Germany may be due to the generally larger “hype” at the Neuer Markt over the period 1997–1999, as shown in Fig. 2, leading to generally higher valuation levels than at the Nouveau March´e. On the other hand, the differences in IPO valuation between France and the U.K. could be explained by the higher liquidity available in the relatively larger British capital markets. Next, we examine the different accounting profitability ratios of venture- and non venture-backed firms at the time of the IPO. Panel A of Table 3 compares the operating performance of the two groups, measured by the operating return on assets (ROA) and the operating cash flow return on assets (CFROA). Moreover, the profitability ratios normalized by total sales (ROS, CFROS) are investigated.
Panel A: Comparison of Venture- and Non Venture-Backed IPOsa Variable
ROA CFROA ROS CFROS
France
Germany
U.K.
VCBacked
Non VCBacked
t-Test (z-Score)
VCBacked
Non VCBacked
t-Test (z-Score)
VCBacked
Non VCBacked
4.3 (7.4) 10.4 (14.7) −121.7 (6.5) −98.6 (11.9)
14.0 (9.7) 21.7 (17.6) 35.0 (8.0) 44.8 (13.3)
−2.6* (−2.0** ) −3.1* (−2.3** ) −1.1 (−1.1) −1.1 (−1.7*** )
−4.3 (1.3) 5.3 (7.3) −22.3 (1.4) −8.8 (6.3)
−2.0 (11.6) 9.1 (20.7) −53.3 (7.5) −25.9 (15.3)
−0.1 (−3.8* ) −0.3 (−3.9* ) 0.5 (−3.6* ) 0.4 (−3.3* )
−15.8 (8.4) −9.7 (13.7) −2,874.4 (4.8) −2,763.6 (8.2)
−34.7 (16.8) −29.9 (20.6) −394.8 (9.2) −386.6 (10.4)
t-Test (z-Score) 0.6 (−1.4) 0.7 (−1.2) −0.8 (−1.0) −0.8 (−0.9)
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
Table 3. Accounting Profitability at the Time of the IPO.
255
256
Table 3. (Continued ) Panel B: International Comparisonb Comparison
France vs. U.K.
Variable
VCBacked
ROA
2.3** (−0.4) 2.3** (−0.5) 1.2 (−1.3) 1.2 (−1.8*** )
CFROA ROS CFROS
Germany vs. U.K.
Germany vs. France
Non VCBacked
VCBacked
Non VCBacked
VCBacked
Non VCBacked
1.8*** (−0.2) 2.0*** (−0.3) 1.5 (−0.5) 1.5 (−1.6)
1.3 (−0.5) 1.6 (−0.2) 1.4 (−0.1) 1.4 (−0.6)
1.1 (−0.4) 1.4 (−0.6) 1.7*** (−0.4) 1.9*** (−1.8*** )
−1.8*** (−1.9*** ) −1.0 (−1.4) 0.9 (−2.1** ) 1.0 (−1.3)
−0.8 (−0.2) −0.6 (−0.2) −1.1 (−0.1) −1.2 (0.0)
a Values
GEORG RINDERMANN
for venture- and non venture-backed firms are expressed as means (medians in parentheses). To detect significant differences between the means [medians] of the two firm groups, Student’s t-tests [non-parametric Mann-Whitney tests] were conducted. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. All accounting profitability ratios are expressed in percent and based on accounting figures at the end of the fiscal year prior to the IPO. The operating (cash flow) return on assets is defined as the operating income (cash flow) divided by total assets. The operating (cash flow) return on sales is defined as the operating income (cash flow) divided by sales. b Values for venture- and non venture-backed firms are Student t-tests statistics (non-parametric Mann-Whitney z-scores in parentheses) to detect significant differences between the means (medians) of firm groups in different countries. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. All accounting profitability ratios are expressed in percent and based on accounting figures at the end of the fiscal year prior to the IPO. The operating (cash flow) return on assets is defined as the operating income (cash flow) divided by total assets. The operating (cash flow) return on sales is defined as the operating income (cash flow) divided by sales. ∗ Significance level is 10%. ∗∗ Significance level is 5%. ∗∗∗ Significance level is 1%.
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
257
However, we detect no systematic differences between the two groups that apply to all three countries. The differences between the means and medians of the operating performance measures indicate that the sample contains a number of extreme values. In fact, some firms exhibit particularly large returns on sales ratios caused by very low sales figures, which are not unusual for young technology-oriented firms during the first years of operations. Given the observed skewness of the profitability ratios scaled by total sales, we use the operating performance measures normalized by assets in the multivariate analysis. The international comparison in Panel B of Table 3 does not identify any significant differences across countries applying to both means and medians of the considered performance ratios. Likewise, the comparison between the accounting profitability of firms backed by international and national venture capitalists, which is for brevity not stated, does not indicate any systematic differences between the two groups. In summary, the univariate analysis of the IPO and firm characteristics does not reveal many systematic differences between the two groups of firms. First, it shows that venture-backed IPOs in all three countries are characterized by higher secondary sales than non venture-backed issues. This finding can be explained by the need of venture capitalists to return capital to investors and to raise new funds, even though prestigious venture capitalists typically do not sell many shares in IPOs so as to avoid the impression of cashing out (Gompers & Lerner, 1999). Second, the refined analysis, focusing on firms backed by different kinds of venture capitalists, indicates significant variations with respect to the issue size. The findings support the idea that international venture capitalists can bring larger issues to the market than national ones. Moreover, it points out the necessity to account for the heterogeneity of venture capitalists when testing for performance differences in the multivariate analysis. In the following section, we present the descriptive statistics of the post-issue operating and market performance. 4.1.2. Post-Issue Operating and Market Performance Panel A of Table 4 compares the average accounting profitability ratios of the two post-issue fiscal years between venture- and non venture-backed IPOs. The overall results do not indicate any systematic differences between the means and medians of the two groups of firms that apply to all three countries. Likewise, the refinement of the analysis in Panel B, distinguishing between firms backed by different types of venture capitalists, does not indicate any significant differences in post-issue performance. Although the comparison of the pre- and post-issue profitability ratios indicates a potential decline of the operating performance after the IPO, the data that we have does not allow a validation of this finding.13 We therefore focus on the market performance of IPOs in the subsequent analyses.
258
Table 4. Post-IPO Accounting Profitability Ratios (Two-Year Averages). Panel A: Comparison of Venture- and Non Venture-Backed IPOsa Variable
France VCBacked
ROA CFROA ROS
3.0 (3.2) 8.3 (6.3) −60.9 (3.8) −17.1 (8.3)
t-Test (z-Score) −1.9*** (−0.8) −2.0** (−0.7) 0.17 (−0.9) −0.6 (−0.5)
VCBacked −15.7 (−0.5) −1.5 (5.5) −43.9 (−2.1) −14.5 (5.7)
Non VCBacked −3.3 (3.5) 4.6 (8.3) −15.0 (3.8) 0.2 (8.8)
U.K. t-Test (z-Score) −2.2** (−1.8** ) −1.8*** (−1.8*** ) −2.3** (−2.0** ) −1.6 (−1.8*** )
VCBacked −7.0 (5.9) 1.2 (10.0) −1,237.5 (5.2) −1,148.1 (7.7)
Non VCBacked −18.8 (−1.1) −13.1 (4.9) −193.4 (−0.9) −179.0 (5.9)
t-Test (z-Score) 1.0 (−0.2) 1.4 (−0.6) −0.8 (−0.6) −0.8 (−0.6)
GEORG RINDERMANN
CFROS
−5.6 (1.6) −0.3 (7.1) −50.5 (0.8) −38.8 (8.3)
Non VCBacked
Germany
Variable
ROA CFROA ROS CFROS a Values
France Int VCBacked
Nat VCBacked
−0.6 (2.8) 4.1 (8.7) −50.0 (1.8) −41.4 (9.5)
−11.1 (1.1) −5.0 (6.1) −51.0 (0.8) −36.0 (5.5)
Germany t-Test (z-Score)
Int VCBacked
Nat VCBacked
1.5 (−0.7) 1.5 (−0.9) 0.0 (−0.1) −0.1 (−0.5)
−5.9 (1.4) 2.9 (6.0) −37.2 (−1.5) −7.9 (6.3)
−20.1 (−2.4) −3.4 (4.7) −46.9 (−2.7) −17.4 (5.7)
U.K. t-Test (z-Score)
Int VCBacked
Nat VCBacked
t-Test (z-Score)
1.1 (−0.7) 1.0 (−0.3) 0.4 (−0.2) 0.6 (−0.1)
−9.9 (7.7) −0.9 (7.7) −1,726.2 (5.2) −1,605.1 (7.4)
0.3 (6.2) 6.4 (12.7) −15.9 (6.0) −5.5 (10.5)
−0.8 (−0.4) −0.9 (−0.6) −0.7 (−0.7) −0.7 (−0.8)
for venture- and non venture-backed firms are expressed as means (medians in parentheses). To detect significant differences between the means [medians] of the two firm groups, Student’s t-tests [non-parametric Mann-Whitney tests] were conducted. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. All accounting profitability ratios are averages of two fiscal years following the IPO and expressed in percent. The operating (cash flow) return on assets is defined as the operating income (cash flow) divided by total assets. The operating (cash flow) return on sales is defined as the operating income (cash flow) divided by sales. b The values for firms backed by international and national venture capitalists are expressed as means (medians in parentheses). To detect significant differences between the means [medians] of the two firm groups, Student’s t-tests [non-parametric Mann-Whitney tests] are conducted. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. All accounting profitability ratios are expressed in percent and based on accounting figures at the end of the fiscal year prior to the IPO. The operating (cash flow) return on assets is defined as the operating income (cash flow) divided by total assets and the operating (cash flow) return on sales is defined as the operating income (cash flow) divided by sales. ∗∗ Significance level is 5%. ∗∗∗ Significance level is 1%.
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
Panel B: IPOs Backed by International and National Venture Capitalistsb
259
260
GEORG RINDERMANN
Fig. 4. Mean Wealth Relatives of IPOs by Country. Note: WR – Wealth Relative.
To examine the long-term market performance of the IPOs, buy and hold returns as well as wealth relatives are calculated by compounding monthly returns for the 3-year post-issue period. Figure 4 plots the mean cumulative wealth relatives of all IPOs by country to the respective growth market index (figure on the left-hand side) and the NASDAQ Composite index (figure on the right-hand side). The wealth relatives on the left-hand side do not suggest that the IPO firms underperform the market in the long run, as documented by prior studies (Loughran & Ritter, 1995; Ritter, 1991). Yet, the comparison of the figures on the left- and right-hand side reveals that the results are sensitive to the benchmark. The use of the NASDAQ Composite index as alternative benchmark leads to generally lower wealth relatives three years after the IPO, even though the ranking of the portfolios by country does not change when compared to the figures on the left-hand side. Most strikingly, the Neuer Markt IPOs underperform the benchmark toward the end of three years, ending up with a wealth relative of only 0.75. A closer look at the distribution of the wealth relatives reveals that they are highly skewed and that outliers drive the means of the different country portfolios. Therefore, Fig. 5 plots the medians of the wealth relatives that are less sensitive to outliers. The graphs illustrate that the median IPO in all three countries underperforms seasoned stocks, independent of the applied benchmark, with wealth relatives ranging from 0.2 to 0.8. The German IPOs perform worse than the market and the IPOs from the two other countries. Next, we examine the returns and 3-year wealth relatives of venture- and non venture-backed IPOs. Table 5 reports the buy and hold returns as well as wealth relatives of both groups against different benchmarks. We report both means and medians. The results show that the venture-backed IPOs in Germany underperform their non venture-backed equivalents in terms of raw returns and wealth relatives, independent of the benchmark. However, the differences are not
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
261
Fig. 5. Median Wealth Relatives of IPOs by Country. Note: WR – Wealth Relative.
significant at conventional statistical levels. Most striking is the fact that both groups of firms report highly negative median returns and relatively low median wealth relatives. The cross-country comparison of the returns and wealth relatives underlines the relatively poor performance of the German venture-backed IPO firms. Given the skewness of the returns and wealth relatives, Table 6 reports the Mann-Whitney test scores to discover significant cross-country differences between the medians of the groups. It shows that venture-backed firms in Germany perform significantly worse than their counterparts in France and the U.K., irrespective of the benchmark. But also the non venture-backed firms in Germany display worse performance figures than their equivalents in the two other countries in terms of buy and hold returns as well as wealth relatives against the NASDAQ Composite index. Finally, we test the conjecture that there are differences in the quality and experience of venture capitalists. Table 7 compares the average wealth relatives of firms backed by international and national venture capitalists. The results show that IPOs backed by international venture capitalists display significantly higher 3-year wealth relatives than issues backed by national venture investors, irrespective of the benchmark. Moreover, the median firm backed by international venture investors does not underperform seasoned stocks, if wealth relatives are calculated against the index of the respective growth market. Overall, these findings support the notion that international venture capitalists select better firms and provide venture-backing of higher quality than national venture investors. Summarizing the main findings of the present section, the market performance figures indicate that IPOs in the German sample perform particularly poorly vis-`a-vis to issues in the two other countries. However, as the differences between venture- and non venture-backed firms are statistically insignificant, multivariate analysis is warranted. Second, there is evidence pointing out considerable differences in the experience of different types of venture capitalists and the quality of
262
Table 5. Three-Year Post-IPO Returns and Wealth Relatives vs. Different Benchmarks. Benchmarks
France (Nouveau March´e) Nouveau March´e NASDAQ Composite
Venture-Backed IPOs IPO Return (%)
Benchmark Return (%)
80.5 (4.0) 80.5 (4.0)
Non Venture-Backed IPOs Wealth Relative
IPO Return (%)
Benchmark Return (%)
Wealth Relative
27.9 (−12.6) 28.8 (4.9)
2.06 (1.02) 1.79 (0.73)
76.8 (−36.1) 76.8 (−36.1)
55.3 (−4.2) 48.4 (11.7)
1.39 (0.64) 1.32 (0.55)
47.1 (−86.3) 47.1 (−86.3)
−54.3 (−73.9) −27.5 (−38.0)
1.28 (0.51) 0.98 (0.20)
100.9 (144.9) 162.6 (231.1)
2.12 (0.65) 1.71 (0.52)
Germany (Neuer Markt) Nemax All Share NASDAQ Composite
−61.2 (−88.3) −61.2 (−88.3)
−28.7 (−72.9) −17.0 (−32.1)
0.78 (0.35) 0.39 (0.15)
U.K. (techMARK) techMARK All Share NASDAQ Composite
194.3 (7.3) 194.3 (7.3)
94.6 (134.0) 151.8 (195.1)
1.18 (0.66) 0.94 (0.53)
447.8 (36.2) 447.8 (36.2)
GEORG RINDERMANN
Note: Three-year equal weighted returns on IPOs are compared with alternative benchmarks (medians in parentheses). The IPO and benchmark returns are calculated by compounding monthly returns for 36 months subsequent to the IPO date. Wealth relatives are computed by taking the ratio of one plus the IPO portfolio return over one plus the return on the chosen benchmark over the same period.
263
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
Table 6. International Comparison of the Post-IPO Returns and Wealth Relatives. Comparison Variable
IPO return WR to GM WR to NASDAQ
France vs. U.K. VCBacked −0.6 −1.1 −1.4
Non VCBacked −1.4 −0.4 −0.1
Germany vs. U.K.
Germany vs. France
VCBacked
Non VCBacked
VCBacked
Non VCBacked
−5.4* −2.2** −3.9*
−4.5* −1.0 −2.9*
−6.2* −3.6* −5.5*
−5.3* −0.6 −3.5*
Note: Values for venture- and non venture-backed firms are non-parametric Mann-Whitney z-scores to detect significant differences between the medians of the firm groups in different countries. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. Three-year equal weighted buy and hold returns on IPOs are compared with alternative benchmarks. GM denotes the index of the country’s growth market, i.e. Nouveau March´e, Nemax All Share, and techMARK All Share index for Nouveau March´e, Neuer Markt, and techMARK, respectively. NASDAQ denotes the NASDAQ Composite index. ∗ Significance level is 10%. ∗∗ Significance level is 5%.
Table 7. Wealth Relatives of International and National Venture Capitalists. Variable
WR to GM WR to NASDAQ
IPOs Backed by International Venture Capitalists
IPOs Backed by National Venture Capitalists
t-Test (z-Score)
2.10 (1.03) 1.56 (0.58)
0.70 (0.39) 0.57 (0.22)
2.7* (−3.8* ) 2.7* (−3.7* )
Note: The table compares the average 3-year wealth relatives of sample firms backed by international and national venture capitalists. Wealth relatives (WR) are computed by taking the ratio of one plus the IPO portfolio return over one plus the return on the chosen benchmark over the same period. GM denotes the index of the country’s growth market. The values are expressed as means (medians in parentheses). To detect significant differences between the means [medians] of the two firm groups, Student’s t-tests [non-parametric Mann-Whitney tests] were conducted. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. ∗ Significance level is 10%.
their venture-backing. To shed more light on the characteristics of different kinds of venture capitalists, the following section examines the experience of venture capitalists and their involvement in portfolio companies. 4.1.3. Experience and Involvement of Different Types of Venture Capitalists The present section explores different types of venture capitalists and their involvement in portfolio firms to highlight potential differences in the quality of
264
GEORG RINDERMANN
Fig. 6. Participation of Different Types of Venture Capitalists in the Sample.
venture-backing. As shown in the previous section, a promising method to capture differences in the experience of venture capitalist consists in distinguishing between nationally and internationally operating investors. To examine the participation of different types of venture capitalists in firms of the sample, Fig. 6 illustrates the composition of venture-backed IPOs in those backed by national and international venture capital investors for each country. It shows that international venture capitalists are the most active in the British market and the least involved in German IPOs. In Fig. 7 the composition of different lead venture capitalists involved in sample firms is shown by nationality. It reveals that most of the national venture capitalists backing IPOs are German (50%) and French (33%). By contrast, the majority of the international venture capitalists are British (65%). The finding that the more experienced international venture capitalists are mostly British and more frequently involved in IPOs in the UK than in the two other countries indicates that the British venture capital market is more sophisticated than its equivalents in France and Germany, where venture capital developed later and national venture capitalists play a more important role. Figure 8 shows the distribution of national and international venture capitalists by their founding dates. It shows that the majority of the internationally classified venture capitalists (70%) were founded before 1980. Moreover, none of the international investors started its activities after 1995. This finding supports the notion that international venture capitalists tend to have extensive business
Fig. 7. National Origin of Different Types of Venture Capitalists.
265
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
Fig. 8. Founding Dates of National and International Venture Capitalists.
experience in backing portfolio firms. National venture capitalists, by contrast, reveal a higher variance concerning the number of years in business. Only 15% of the national venture investors were established before 1980, whereas about two thirds of them began their operations between 1980 and 1985. Finally, about one fifth of the national venture capitalists has only little business experience and has started investing only after 1995. This result indicates that considerable variations exist in the backing experience among different types of venture capitalists. Table 8 compares a number of alternative measures for the experience and sophistication of venture capitalists. The measures include the: (1) age of the lead venture investor; (2) board representation at portfolio firms; (3) number of syndication partners; and (4) number of backed IPOs in the sample. First, the results show that the internationally operating lead venture capitalists are significantly older than lead financiers of nationally venture-backed firms at the time of the IPO. Second, both types of venture investors are represented on the boards of portfolio Table 8. Experience and Involvement of Venture Capitalists in IPOs of the Sample. Variable VC Age Board NVC No IPOs
International VCs 31.6 (29.0) 76% 3.3 (3.0) 11.4 (7.0)
National VCs 12.2 (12.0) 60% 2.1 (2.0) 3.1 (2.0)
t-Test (z-Score) 10.0* (−7.7* ) 2.1** 4.6* (−4.1* ) 8.4* (−6.9* )
Note: The table compares international and national venture capitalists with respect to the age of the lead venture capitalist at the time of the IPO (VC Age) and the percentage of sample firms, in which they were represented on the board (Board). Moreover, the total number of venture capitalists having invested in the same portfolio firm (NVC) and the number of backed IPOs in the sample (No IPOs) are compared between the two types of venture capitalists. The values are expressed as means (medians in parentheses). To detect significant differences between the means [medians] of the two firm groups, Student’s t-tests [non-parametric Mann-Whitney tests] were conducted. Bold values indicate statistically significant differences at the 1% or 5% levels. ∗ Significance level is 10%. ∗∗ Significance level is 5%.
266
GEORG RINDERMANN
Table 9. Shareholdings of Venture Capitalists in IPO Firms. Variable AVC PrEq AVC PosEq AVC Sale LVC PrEq LVC PosEq LVC Sale
International VCs
National VCs
t-Test (z-Score)
44.8 (40.7) 26.7 (24.8) 40.0 (35.2) 24.9 (22.3) 14.4 (12.8) 40.6 (37.6)
27.3 (22.0) 15.3 (11.8) 40.1 (33.3) 18.5 (14.4) 10.0 (8.5) 40.0 (32.2)
5.0* (−4.9* ) 5.0* (−4.4* ) 0.0 (−0.2) 2.4** (−3.7* ) 2.9* (−3.2* ) 0.1 (−0.6)
Note: The table compares national and international venture capitalists with respect to their involvement in IPO firms. AVC (LVC) PrEq and PosEq denote the percentage of equity owned by all venture capitalists (the lead venture capitalist) before and immediately after the IPO, respectively. AVC (LVC) Sale denotes the percentage of shares sold by all venture capitalists (the lead venture capitalist) at the IPO. The values are expressed as means (medians in parentheses). To detect significant differences between the means [medians] of the two firm groups, Student’s t-tests [non-parametric Mann-Whitney tests] were conducted. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. ∗ Significance level is 10%. ∗∗ Significance level is 5%.
firms, even though European venture capitalists are said to employ a hands-off management policy (Gompers & Lerner, 2001). However, board representation is used significantly more often by international venture capitalists than by national ones. Third, although syndication is a common feature applied by both groups of venture investors, international venture capitalists tend to invest with a higher number of partners than their national equivalents. Finally, international venture capitalists back a significantly higher number of IPOs in the sample than national ones, suggesting that the former are more experienced in bringing firms to the capital markets. Next, we investigate the venture shareholdings in the portfolio firms to detect any differences in the involvement between different kinds of venture capitalists. Table 9 outlines the pre- and post-issue equity stakes held by international and national venture investors at the time of the IPO. The figures show that both types of venture capitalists hold substantial pre- and post-issue equity stakes in IPOs. In line with prior U.S. studies, these findings underline that both groups of venture capitalists have significant incentives to participate in the corporate governance of portfolio firms. However, the comparison of different venture capitalists reveals statistically significant differences in their involvement in IPOs, i.e. international venture capitalists systematically hold more equity in firms they bring public than their national counterparts. Like all venture capitalists as a group, the international (national) lead manager, defined as the one with the largest equity position, represents a significant
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
267
stockholder, owning on average 24.9 (18.5)% before and 14.4 (10.0)% after flotation. The comparison of the different types of venture investors exposes considerable variations in the involvement of lead venture capitalists. In line with the previous findings for venture capitalists as a group, the lead financier of internationally venture-backed firms holds a significantly larger equity stake than its matching part in nationally venture-backed IPOs. Although the equity stakes of venture capitalists decrease considerably at the time of the IPO, the aggregate post-issue equity holdings of both the venture capital syndicate and lead financier are still large. The comparative analysis of the selling intensity14 does not reveal any significant differences between different types of venture capitalists. Nonetheless, it is worth mentioning that the IPO selling intensity in the sample is generally higher than that of U.S. venture capitalists (Barry et al., 1990). This difference suggests that, by selling a higher portion of shares in IPOs, European venture capitalists might care less about reputation than their U.S. equivalents. All in all, this section provided a number of important insights with regard to differences in the experience and involvement of different types of venture capitalists. First, international venture capitalists participated with different frequencies in the IPO markets that we consider. They were the most active in the British IPO market and the least active in the German Neuer Markt. Second, the results show substantial variations in the business experience and sophistication of different types of venture capitalists. On average, international venture capitalists are older than national venture capital investors, back a larger number of IPOs in the sample, hold board representations more often and invest with more syndication partners. Finally, the degree of involvement in IPOs suggests that both types of venture capitalists can be considered as active investors, holding large equity positions. However, international venture capitalists hold larger preand post-issue equity stakes in portfolio firms than national investors, on average. Overall, the findings support the view that international venture capitalists are more experienced and involved in ventures than their nationally operating equivalents. The following section analyzes whether venture capitalists have an effect on the choice of underwriters in IPO firms. 4.1.4. Underwriter Quality Previous studies argue that venture capitalists are likely to influence the choice of underwriters given their active participation in the operations of firms they bring public. There is evidence that venture-backed firms are underwritten by higher quality underwriters than non venture-backed issues (Megginson & Weiss, 1991). Therefore, we consider different reputation proxies for the lead underwriters, as described in Section 3.4, and compare these between venture-backed and non
268
GEORG RINDERMANN
venture-backed IPOs. As shown in Panel A of Table 10, only in the British sample, both the average market share and reputation rank of lead underwriters bringing venture-backed firms to the market are significantly higher than those of non venture-backed issues. This might be due to the fact that venture capital in the UK is more developed and has a longer tradition than in the two continental European countries, allowing venture capitalists to establish relationships with prestigious underwriters. In the French and German sample, however, no significant differences between the qualities of lead underwriters being involved in venture- and non venture-backed IPOs can be detected. There are two possible explanations for this finding. First, venture capitalists might be indifferent to the reputation of underwriters because they do not need them to certify the IPO firm’s quality and financial prospects (Doukas & Gonenc, 2001). Non venture-backed firms, by contrast, might rely to a larger extent on certification of investment banks for a successful IPO. Second, as the findings of the previous section revealed that a considerable number of IPOs in France and Germany is backed by national venture capitalists, it might be that these typically younger and less experienced investors have not yet built up relationships with top-tier investment banks. In order to check whether the mixed results concerning the reputation of underwriters are due to the heterogeneity of venture capital, we compare the lead underwriter reputation of IPOs backed by international and national venture capitalists in the refined analysis. The results in Panel A of Table 10 indicate that international venture capitalists in all countries attract more prestigious lead underwriters than their national venture capitalists in terms of the underwriter rank. However, in the French and German sample, the observed differences in the ranking scores are only weakly significant at the 10% level. In the UK, however, the results based on the market shares and reputation ranks of lead underwriters in the IPO market reveal highly significant differences between nationally and internationally venture-backed firms. In summary, there is evidence that German IPOs perform significantly worse than issues in the two other countries. Apart from significant differences in the selling behavior of existing shareholders, the findings do not reveal systematic disparities between the firm and offering characteristics of venture- and non venture-backed IPOs. The refined analysis considering different types of venture capitalists shows that IPOs backed by international venture-capitalists are larger in terms of the offering proceeds and achieve higher wealth relatives than those backed by national venture investors. Also, international venture capitalists are on average older than their national equivalents, more often represented on the board, back more IPOs, syndicate with more investment partners, and hold larger pre- and post-issue equity stakes in firms they bring public. Finally, there is weak evidence that IPOs backed by international venture capitalists use more
Panel A: Venture- and Non Venture-Backed IPOs Variable
Mkt. Share URank
France
Germany
VCBacked
Non VCBacked
t-Test (z-Score)
6.3 (1.3) 1.9 (2.0)
6.5 (0.6) 2.1 (3.0)
−0.1 (−0.2) −1.4 (−1.4)
VCBacked 6.6 (2.9) 1.9 (2.0)
U.K.
Non VCBacked
t-Test (z-Score)
VCBacked
Non VCBacked
t-Test (z-Score)
5.3 (2.9) 2.0 (2.0)
0.9 (−0.2) −0.5 (−0.6)
4.6 (1.3) 1.9 (2.0)
1.7 (0.4) 2.6 (3.0)
2.1** (−2.7* ) −3.0* (−2.8* )
Panel B: IPOs Backed by International and National Venture Capitalists Variable
Mkt. Share URank
France
Germany
U.K.
Int VCBacked
Nat VCBacked
t-Test (z-Score)
Int VCBacked
Nat VCBacked
t-Test (z-Score)
Int VCBacked
Nat VCBacked
t-Test (z-Score)
9.0 (1.3) 1.6 (1.0)
5.1 (1.3) 2.0 (2.0)
1.3 (−0.8) −1.7*** (−1.7*** )
7.1 (2.9) 1.6 (1.5)
6.5 (2.9) 2.0 (2.0)
0.2 (−1.0) −1.7*** (−1.7*** )
6.6 (6.1) 1.6 (1.0)
1.1 (0.6) 2.5 (3.0)
2.8* (−2.2** ) −3.2* (−2.8* )
269
Note: Values for venture- and non venture-backed firms are expressed as means (medians in parentheses). To detect significant differences between the means [medians] of the two firm groups, Student’s t-tests [non-parametric Mann-Whitney tests] are conducted. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. Mkt. Share denotes the relative market share of proceeds brought to the IPO markets of France, Germany, and the UK by the lead underwriter between 1995 and 1999. URank is the underwriter reputation rank on a scale of 1–3, where 1 is the most and 3 the least prestigious underwriter. The statistics consider only IPO firms backed by domestic lead venture capitalists. ∗ Significance level is 10%. ∗∗ Significance level is 5%. ∗∗∗ Significance level is 1%.
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
Table 10. Lead Underwriter Reputation of IPOs.
270
GEORG RINDERMANN
prestigious underwriters than those backed by national venture capitalists. It appears fundamental to account for the heterogeneity of venture capital firms in the multivariate analyses of the following sections.
4.2. Regression Analysis 4.2.1. Full Sample Results Performance differences between firms with and without venture-backing might be due to other factors than the participation of venture capitalists. To explore to what extent the observed operating and market performance is related to the presence of venture capitalists, we estimate cross-sectional regressions including various explanatory and control variables, as outlined in Section 3. The full sample based regressions also include country dummies and variables for different types of venture capitalists. Table 11 reports the cross-sectional estimates of the relationship between the average operating cash flow return on assets of the two fiscal years following the IPO and various explanatory variables, as specified in Section 3.1. The first and second column of Table 11 explore the constant and slope effects of venture capital. The results do not indicate any significant impact of venture involvement on the post-issue profitability. To account for the heterogeneity of venture capitalists, we extend the regression setup by the inclusion of an additional variable for the involvement of international venture capitalists (Int VC) in columns III and IV. While the use of the dummy as an intercept (column 3) does not lead to significant changes regarding the effect of venture-backing, the consideration of slope effects (column 4) shows that IPOs backed by international venture capitalists exhibit significantly higher post-issue performance than non venture-backed issues and firms backed by national venture investors. The results also indicate that the postIPO operating performance of firms backed by international venture capitalists is less dependent on firm size. In line with this finding, the significant Chow-test statistic indicates structural differences between the individual regressions of the two groups. Next, we focus on the determinants of the initial market valuation, measured by the approximated value of Tobin’s Q at the first trading day to examine the certification role of venture capitalists in IPOs. The equations in the first and second column of Table 12 control for constant and slope effects of venture capital involvement, respectively. To account for country-related effects, dummy variables for France and the UK are included in the model specifications. The findings suggest that larger firms experience higher market valuations than smaller ones at the first trading day. However, the outcome does not show any significant
271
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
Table 11. Full Sample Regressions on Post-IPO Operating Performance. Variables VC Int VC Size Size × VC Size × Int VC Age Age × VC Age × Int VC Part Part × VC Part × Int VC URank URank × VC URank × Int VC France U.K. Constant Adj. R2 F-test Chow-test Sample size
(1)
(2)
(3)
(4)
−0.05 (−1.0)
−0.03 (−0.1)
0.48* (7.2)
0.54* (6.3) −0.30 (−1.3)
−0.07 (−1.1) 0.03 (0.5) 0.47* (7.0)
0.11** (2.1)
0.05 (0.7) 0.16 (0.9)
0.11** (2.2)
0.27* (4.8)
0.23** (2.3) 0.05 (0.4)
0.26* (4.7)
0.18* (3.2)
0.13 (1.5) 0.11 (0.7)
0.19* (3.2)
−0.50 (−1.2) 0.78** (2.0) 0.54* (6.2) −0.07 (−0.2) −0.43*** (−1.8) 0.05 (0.7) 0.26 (1.2) −0.14 (−0.7) 0.23** (2.3) 0.01 (0.0) 0.04 (0.3) 0.13 (1.5) 0.28 (1.5) −0.22 (−1.4) 0.08 (0.7) −0.31** (−2.5) −0.53* (−3.5)
0.12 (1.0) −0.29** (−2.4) −0.58* (−4.9)
0.11 (0.9) −0.29** (−2.4) −0.55* (−3.7)
0.28 7.5* 1.4 303
0.28 6.4* 1.4 303
0.11 (0.9) −0.30* (−2.5) −0.58* (−4.8) 0.28 7.1* 2.6** 303
0.28 5.4* 2.6** 303
Note: The dependent variable is the average operating cash flow return on assets of the two fiscal years subsequent to the IPO. The independent variables include a dummy variable that equals 1 if a firm is venture-backed (VC), the natural logarithm of the firm’s 2-year average post-IPO assets (Size), the natural logarithm of the firm age in years (Age), the participation ratio of old shareholders in the IPO (Part), defined as number of old shares sold divided by the number of shares outstanding before flotation, and the underwriter reputation rank on a scale of 1–3 (URank), where 1 is the most and 3 the least prestigious underwriter. France and U.K. are country dummy variables, which equal 1 if a firm went public in the respective country. Int VC is a dummy variable equal to 1 if an IPO was backed by a venture capitalist that backed firms in at least two countries of the sample. All regressions employ an ordinary least squares specification. The estimations include industry, calendar year and accounting standard dummy variables, which are not reported. tStatistics are in parentheses. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. ∗ Significance level is 10%. ∗∗ Significance level is 5%. ∗∗∗ Significance level is 1%.
influence of venture-backing on Tobin’s Q. In keeping with the results of the univariate analysis, the regressions indicate that IPOs at the French Nouveau March´e undergo significantly lower market valuations in terms of Tobin’s Q than issuing firms in the two other country samples.
272
GEORG RINDERMANN
Table 12. Full Sample Regressions on Tobin’s Q at the First Trading Day. Variables VC Int VC Size Size × VC Size × Int VC EqR EqR × VC EqR × Int VC SA S A × VC S A × Int VC Alpha Alpha × VC Alpha × Int VC France U.K. Constant Adj. R2 F-test Chow-test Sample size
(1)
(2)
(3)
0.01 (0.2)
0.50 (1.2)
0.42* (6.3)
0.43* (5.2) −0.05 (−0.2)
0.00 (0.1) 0.01 (0.2) 0.42* (6.1)
−0.02 (−0.3)
−0.02 (−0.3) 0.03 (0.5)
−0.02 (−0.3)
0.08 (1.5)
0.07 (1.1) 0.02 (0.2)
0.08 (1.5)
0.23* (4.3)
0.35* (3.7) −0.47* (−1.6)
0.23* (4.3)
−0.22*** (−2.0) −0.08 (−0.7) −3.21* (−2.8)
−0.24** (−2.1) −0.08 (−0.7) −4.75* (−3.0)
0.35 10.2* 0.8 303
0.35 8.4* 0.8 303
−0.23*** (−2.0) −0.08 (−0.7) −3.20* (−2.8) 0.35 9.6* 1.8*** 303
(4) 0.41 (0.9) 0.25 (0.6) 0.43* (5.1) 0.00 (0.0) −0.11 (−0.5) −0.02 (−0.3) 0.05 (0.6) −0.07 (−0.8) 0.07 (1.1) 0.05 (0.5) −0.12 (−1.0) 0.35* (3.7) −0.46 (−1.4) −0.02 (−0.1) −0.24** (−2.0) −0.09 (−0.8) −4.71* (−2.9) 0.34 6.8* 1.8*** 303
Note: The dependent variable is the approximated value of Tobin’s Q at the first day of trading, defined as the market value of equity and book value of debt divided by the book value of total assets. The independent variables are a dummy variable that equals 1 if a firm is venture-backed (VC), the natural logarithm of the firm’s market value of equity at the first days of trading (Size), the equity ratio (EqR), defined as the book value of equity divided by the book value of assets at the fiscal year prior to the IPO, the inverse of capital intensity (S A), defined as the ratio of sales over assets at the fiscal year prior to the IPO, and the fraction of the equity retained by old shareholders after the IPO (Alpha). France and UK are country dummy variables, which equal 1 if a firm went public in the respective country. Int VC is a dummy variable equal to 1 if an IPO was backed by a venture capitalist that backed firms in at least two countries of the sample. All regressions employ an ordinary least squares specification. The estimations include industry, calendar year and accounting standard dummy variables, which are not reported. t-Statistics are in parentheses. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. ∗ Significance level is 10%. ∗∗ Significance level is 5%. ∗∗∗ Significance level is 1%.
273
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
Table 13. Full Sample Regressions on Three-Year Wealth Relatives. Variables VC Int VC Q Q × VC Q × Int VC Size Size × VC URank Size × Int VC URank × VC URank × Int VC France U.K. Constant Adj. R2 F-test Chow-test Sample size
(1) −0.08 (−1.3)
(2) 0.39 (1.1)
−0.15** (−2.1)
−0.15*** (−1.8) 0.02 (0.1)
−0.01 (−0.1)
0.13 (1.1) −0.64** (−2.1) −0.11 (−1.2)
−0.05 (−0.8)
(3) *
−0.18 (−2.7) 0.22* (3.1) −0.15** (−2.1)
−0.04 (−0.4) −0.03 (−0.4)
0.13 (0.7) 0.21 (1.6) 0.14 (1.1) −0.93 (−1.1) 0.09 2.7* 1.2 303
0.18 (1.3) 0.14 (1.1) −1.38 (1.4) 0.11 2.8* 1.2 303
0.15 (1.1) 0.07 (0.6) −0.74 (−0.9) 0.12 3.2* 3.0* 303
(4)
(5)
0.09 (0.2) 0.41 (1.5) −0.15*** (−1.8) 0.00 (0.0) 0.02 (0.1) 0.09 (0.8) −0.41 (−1.1) −0.12 (−1.2) −0.36 (−0.9) 0.13 (0.6) 0.14 (0.8) 0.12 (0.9) 0.10 (0.8) −1.03 (−1.1)
0.27 (1.0) 0.63** (2.4) −0.18** (−2.2) 0.02 (0.1) 0.07 (0.4) 0.15 (1.5) −0.49 (−1.4)
0.14 3.0* 3.0* 303
−0.49 (−1.5)
0.13 (1.0) 0.10 (0.7) −1.76** (−2.5) 0.14 3.4* 3.1* 303
Note: The dependent variable is the natural logarithm of the 3-year wealth relative using the respective country’s growth market index as benchmark. The independent variables are a dummy variable that equals 1 if a firm is venture-backed (VC), the approximated value of Tobin’s Q (Q) at the first day of trading, defined as the market value of equity and book value of debt divided by the book value of total assets, the natural logarithm of the IPO firm’s market capitalization at the end of the first day of trading (Size), and the underwriter reputation rank on a scale of 1–3 (URank), where 1 is the most and 3 the least prestigious underwriter. France and UK are country dummy variables, which equal 1 if a firm went public in the respective country. Int VC is a dummy variable equal to 1 if an IPO was backed by a venture capitalist that backed firms in at least two countries of the sample. All regressions employ an ordinary least squares specification. The estimations include industry, calendar year and accounting standard dummy variables, which are not reported. t-Statistics are in parentheses. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. ∗ Significance level is 10%. ∗∗ Significance level is 5%. ∗∗∗ Significance level is 1%.
In columns III and IV, the model setup includes an additional variable for the involvement of international venture capitalists. However, the regressions accounting for intercept and slope effects of international venture-backing do not yield any new insights concerning the impact of venture-backing on firm valuation. The insignificant Chow-test statistic confirms this diagnosis by rejecting the existence of structural differences between the groups. For robustness, we estimate alternative model specifications that include intercept and slope dummy variables for different nationalities of lead venture investors, but these models do not generate materially different outcomes. Table 13 reports the cross-sectional estimates of the determinants of 3-year wealth relatives using the country’s growth market index as benchmark. While the
274
GEORG RINDERMANN
results in the first column do not indicate any particular effect of venture capital participation, the results in the second column indicate a negative size-related slope effect of venture-backing. The score of the Chow-test, though, does not indicate overall significant structural differences between venture- and non venture-backed IPOs. Consistent with the relationship between initial firm valuation and post-issue stock price performance, Tobin’s Q turns out to be negatively related to market performance. This result suggests that IPOs with higher valuations in terms of Tobin’s Q at the first trading day perform worse than firms with a lower Tobin’s Q. However, the relationship between Tobin’s Q and stock price performance of venture-backed IPOs does not systematically differ from non venture-backed issues, as the interaction term Q × VC remains insignificant. In columns III and IV, we introduce a variable for the involvement of international venture capitalists. The refinement of the model specification in column 3 shows that IPOs backed by international venture investors perform significantly better than the average. IPOs backed by national venture capitalists, by contrast, tend to perform worse than the remainder of the sample. Although the significant score of the Chow-test indicates structural differences between the regressions of the different groups, the variable for international venture capitalist involvement becomes insignificant in the full model specification accounting for slope effects of venture-backing (column 4). A possible explanation for this is the high multicollinearity related to the similar use of dummy and interaction variables, on the one hand, and partly correlated variables, such as venture capital involvement, size and underwriter rank, on the other. Even though the underlying simple correlations are not very large, it is possible that a strong linear relationship exists among several explanatory variables. As a result, the standard errors of the individual variables can become large and the t-values remain insignificant when all variables are included at once (Jobson, 1991, pp. 320–321). Multicollinearity can be a result of too few observations for too many variables. Thus, additional data is often the best way of reducing the undesired effects. Since it is not possible in the present analysis to generate additional data for the sample period, an alternative is to eliminate variables on a theoretical basis to reduce collinearity. However, the omission of relevant variables might equally yield biased estimators of the coefficients for included variables. As prior studies on the market performance of IPOs, such as Brav and Gompers (1997), consider merely firm size and the book-to-market ratio15 in regressions on wealth relatives, column 5 reports the estimates of wealth relatives without the underwriter rank as explanatory variable. Consistent with the findings of column 3 and the significant Chow-test statistic, the results of the reduced model specification confirm the positive significant effect of international venture capitalists on the market performance of IPOs. The values of the adjusted R2 and the F-statistic
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
275
of the setup in column 5 indicate that the omission of the underwriter rank as explanatory variable does not negatively affect the overall regression fit of the model. All in all, the main findings of this section do not provide general support for the conjecture that venture-backed IPOs outperform issues without venture capital participation. However, the refined analysis that distinguishes between the backing of international and national venture capitalists shows that firms supported by international venture capitalists outperform the average IPO in the sample both in terms of operating and market performance. Since the use of dummy variables for venture capital involvement might be too coarse to adequately reflect differences in the quality and effectiveness of venture-backing, the next section performs a further refinement of the analysis. It takes into account the characteristics of venture capitalists and their involvement in the firms they bring public. Also, we investigate firms backed by various organizational types of venture capital firms with respect to their post-issue performance. 4.2.2. Venture-Backed IPOs The present section focuses on the subsample of venture-backed firms to assess the certification and monitoring quality of venture capitalists. Based on the approach followed by Barry et al. (1990), we examine several variables reflecting the reputation and involvement of venture capitalists in IPOs, as discussed in Section 3.3. As in the earlier analysis, regressions on the different performance measures are run against each of the monitoring and certification proxy variables as well as the previously used explanatory variables. The variables are included one by one in the regressions, due to the multicollinearity between the individual certification and monitoring proxies. Nevertheless, in none of the models the monitoring and certification proxies turn out to be significant at conventional levels. Therefore, the individual regression results are not tabulated. We introduce dummy variables for IPOs backed by independent and public lead venture capitalists to explore the effects of different types of venture capitalists on the performance of firms going public. Moreover, using the insights of the earlier analysis, a distinction is made between firms backed by nationally and internationally operating venture capital investors. Finally, performance differences between firms financed by bank-affiliated and non bank-affiliated venture capitalists are examined in more detail. The results shown in the first column of Table 14 indicate that firms backed by public lead venture capitalists underperform the remaining venture-backed IPOs of the sample in terms of wealth relatives. Issues backed by independent venture capitalists, however, do not perform significantly different from the average venture-backed firm, as claimed by Wang et al. (2002). The same results
276
GEORG RINDERMANN
Table 14. Wealth Relatives of IPOs Backed by Different Types of Venture Capitalists. (1) Indep Public Int VC Bank × France Bank × Germany Q Size URank France U.K. Constant Adjusted R2 F-test Sample size
−0.04 (−0.4) −0.20** (−2.1)
−0.09 (−0.8) −0.19 (−1.3) −0.04 (−0.4) 0.29 (1.6) 0.19 (1.1) −0.49 (−0.4) 0.17 2.7* 154
(2) −0.18***
(−1.7) −0.21** (−2.4) 0.32* (3.7)
−0.09 (−0.9) −0.23 (−1.6) 0.03 (0.3) 0.14 (0.8) 0.10 (0.6) 0.40 (−0.4) 0.24 3.5* 154
(3)
(4)
0.01 (0.0) −0.20** (−2.3) 0.35* (4.0) 0.28*** (1.8) 0.08 (0.5) −0.08 (−0.8) −0.24*** (−1.8) 0.02 (0.3) 0.05 (0.3) 0.09 (0.5) −0.90 (−0.8)
0.35* (4.0) 0.23** (2.1) 0.01 (0.1) −0.09 (−0.9) −0.18 (−1.3) 0.03 (0.4) 0.05 (0.3) 0.10 (0.6) −1.04 (−0.9)
0.25 3.4* 154
0.23 3.4* 154
Note: The dependent variable is the natural logarithm of the 3-year wealth relative using the respective country’s growth market index as benchmark. The independent variables are the approximated value of Tobin’s Q (Q) at the first trading day, defined as the market value of equity and book value of debt divided by the book value of total assets, the natural logarithm of the firm’s market capitalization at the end of the first day of trading (Size), and the underwriter reputation rank on a scale of 1–3 (URank), where 1 is the most and 3 the least prestigious underwriter. Indep and Public are dummy variables equal to 1 if a venture capitalist is independent or owned by a public shareholder, respectively. Int VC is a dummy variable equal to 1 if an IPO was backed by a venture capitalist that backed firms in at least two countries of the sample. Bank is a dummy variable equal to 1 if an IPO is backed by a bank-affiliated venture capitalist. All regressions employ an ordinary least squares specification. The estimations include industry, calendar year and accounting standard dummy variables, which are not reported. t-Statistics are in parentheses. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. ∗ Significance level is 10%. ∗∗ Significance level is 5%. ∗∗∗ Significance level is 1%.
are obtained if a dummy variable for the involvement of international venture capitalists (Int VC) is included in the regressions (column 2). In the third column, we extend the regression setup by including variables for the involvement of bank-affiliated venture capitalists in France and Germany, where venture capital firms are often affiliates of major universal banks. Even though the results in column 4 reveal a significant positive impact of bank-affiliated venture-backing on firm performance in France, this finding is only weakly significant if the dummy variable for international venture-backing is included (column 3). Moreover, as the coefficient of the variable Bank × Germany remains
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
277
Table 15. The Determinants of the Backing by International Venture Capitalists. Variable
Coefficient (B)
S.E.
Wald
p-Value
Exp(B)
2.14*
0.57 0.59 0.21 0.13 0.02 0.01 1.99
14.31 0.66 1.30 16.37 2.10 1.19 24.90
0.00 0.42 0.25 0.00 0.15 0.27 0.00
8.52 1.61 1.27 1.67 1.03 0.99
VC Age Board NVC No IPOs AVC PrEq AVC Sale Constant
0.47 0.24 0.52* 0.03 −0.02 −9.95*
−2 ln L LR 2 Cox & Snell’s R2 Nagelkerke R2
89.40 114.73* 0.54 0.72
Note: S. E. – standard error. The table estimates the determinants of backing by international venture capitalists using a logit model. The dependent variable is the dummy variable Int VC, indicating whether a venture capitalist is involved in IPOs in more than one country of the sample. The independent variables include the age (in years) of the venture capitalist at the time of the IPO (VC Age), a dummy variable for board representation (Board), the total number of venture capitalists having invested in the same portfolio firm (NVC), the number of backed IPOs in the sample (No IPOs), the percentage of equity owned by all venture capitalists before the IPO (AVC PrEq), and the percentage of shares sold by all venture capitalists in the IPO (AVC Sale). The Wald-test that equals the square of the ratio of the coefficient to its standard error is reported to test the significance of the individual model coefficients. Bold values indicate statistically significant differences at the 1% confidence level. ∗ Significance level is 10%.
insignificant in both model specifications, there is no uniform evidence indicating systematic performance differences between issues backed by bank-affiliated venture capitalists and non bank-affiliated venture-backed firms. As the previous analysis has shown that the dichotomous categorical variable for the international involvement of venture capitalists is a good indicator for their experience in backing IPOs, a logistic regression is used to identify the determinants of the underlying but unobservable experience of venture investors. The logistic approach estimates the probability that a venture capitalist is backing IPOs in at least two countries of the sample, by using internationality (Int VC) as dependent variable and different characteristics of venture capital investors, such as their age, board representation, the number of syndication partners and backed IPOs, as well as the pre-issue shareholdings in firms and the selling intensity, as explanatory variables. Table 15 outlines the logistic estimates of the distinct variables as predictors for the internationality of venture capitalists. The Wald-test statistic of the coefficients suggests that the age and the number of IPOs backed by a venture
278
GEORG RINDERMANN
capitalist are significant determinants of international involvement. In order to interpret the effects of changes in the logistic coefficients on the dependent variable, the estimates are transformed back into odds ratios in the last column. The transformation reveals that each additional year of the age increases the odds of a venture capitalist being international by the factor of 8.52, holding all other variables constant. Likewise, for each additional IPO backed by venture investors the odds for the involvement in international activities are increased by the factor 1.67. The overall robustness of the estimates is underlined by the statistically significant likelihood ratio (LR) of the 2 -test statistics. Going back to the conjecture that the dummy variable of internationality represents an adequate indicator for the unobservable experience of venture capitalists, the results of the logistic regression suggest that the age and number of backed IPOs should equally serve as proxies for the experience of venture investors. However, in the previous analysis on firm performance both variables turned out to be insignificant. There are at least two possible explanations for the observed insignificance. First, it could be that the distinct variables VC Age and No IPOs do not sufficiently reflect the experience of venture capitalists. Second, as the logistic regression is based on a non-linear estimation technique, it might be that a non-linear relationship exists between firm performance and the explanatory variables of venture capital experience that is not effectively considered by using linear OLS regressions. To explore the potential non-linear relationship between firm performance and the proxies for the experience of venture capitalists, such as their age and the number of backed IPOs, we run separate regressions, using the fitted values of the logit estimation in Table 15 for the internationality of venture capitalists instead of the dummy variable. However, as the influence of the predicted internationality remains insignificant, the regressions are not reported. Alternatively, regressions employing the probability values of the logit estimates for international involvement of venture capitalists are run. However, as shown in Table 16, there is weak support for the positive influence of the probability values of venture capital internationality on the firms’ market performance. The coefficient is significant only in the reduced model specification (column 3). Furthermore, the results indicate that the previously documented negative relationship between Tobin’s Q and market performance loses in importance, the higher the probability that a venture capitalist operates internationally. Hence, the market performance of firms backed by venture capitalists, which are more probable to be international, reveals a tendency to follow a random walk. Finally, the negative sign of the weakly significant interaction variable Size × Pr(VC Int) suggests that the stocks of smaller firms are more profitable in the aftermarket, the higher the probability that they are backed by an international venture capitalist.
279
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
Table 16. Regressions Using Probabilities for International VC Involvement. Variables Pr(Int VC) Q Q × Pr(Int VC) Size Size × Pr(Int VC) URank URank × Pr(Int VC) France U.K. Constant Adj. R2 F-test Sample size
(1) 0.06 (0.7) −0.10 (−0.9)
(2)
(3) 0.62***
(1.7) −0.24 (−1.7) 0.32 (1.5) 0.06 (0.3) −0.91*** (−2.0)
0.30*** (1.7) 0.19 (1.1) −0.86 (−0.8)
0.48 (0.9) −0.23 (−1.5) 0.30 (1.3) 0.03 (0.2) −0.82 (−1.5) −0.03 (−0.2) 0.08 (0.3) 0.32*** (1.8) 0.27 (1.5) −1.53 (−1.1)
0.16 2.7* 154
0.17 2.5* 154
0.18 2.8* 154
−0.17 (−1.1) 0.01 (0.1)
0.32*** (1.8) 0.27 (1.5) −1.69 (−1.5)
Note: The dependent variable is the natural logarithm of the 3-year wealth relative using the respective country’s growth market index as benchmark. Pr(Int VC) is the probability that an IPO is backed by an international venture capitalist, using the estimation approach of the logistic regression. The remaining independent variables include the approximated value of Tobin’s Q (Q) at the first day of trading, defined as the market value of equity and book value of debt divided by the book value of total assets, the natural logarithm of the IPO firm’s market capitalization at the end of the first day of trading (Size), and the underwriter reputation rank on a scale of 1–3 (URank), where 1 is the most and 3 the least prestigious underwriter. France and UK are country dummy variables, which equal 1 if a firm went public in the respective country. All regressions employ an ordinary least squares specification. The estimations include industry, calendar year and accounting standard dummy variables, which are not reported. t-Statistics are in parentheses. Bold values indicate statistically significant differences at the 1% or 5% confidence levels. ∗ Significance level is 10%. ∗∗∗ Significance level is 1%.
In conclusion, the findings of this section indicate that firms backed by public lead venture investors underperform firms backed by private venture capitalists. However, the evidence does not support the conjecture that IPOs backed by independent venture capitalists outperform issues backed by captive venture investors. Likewise, firms supported by bank-affiliated venture capitalists in countries with major universal banks do not outperform issues backed by non bank-affiliated venture investors. Finally, assuming that international venture-backing is a good indicator for the experience of venture capitalists, the findings suggest that the age and backing frequency of venture investors in IPOs of the sample are good predictors for their experience. In the following section, we perform several robustness checks to underline the stability of our empirical results.
280
GEORG RINDERMANN
4.2.3. Robustness Checks We test the robustness of the results using a number of alternative model specifications. For example, instead of using 2-year averages of the operating performance measures, the regressions are run with the individual values of each year without leading to qualitatively different results. Also, different cash flow definitions are used to compute the operating profitability measures, and pure accounting-based measures such as the operating return on assets are employed instead of cash flow-related profitability measures as dependent variables in the regression analyses. As an alternative for the computation of wealth relatives, the NASDAQ Composite index is used as a benchmark instead of the indices of the countries’ growth markets. In general, the results confirm the previously reported findings. Aside from the full-sample based estimates, we estimate additional regressions piecewise by each country, using intercept and interaction dummy variables, to detect fixed and marginal effects of venture-backing as well as cross-country differences. This does not generate materially different outcomes. Although the regression results of the German sample suggest that venture-backed IPOs underperform non venture-backed issues in the long-run, this finding does not hold when slope effects are taken into account. Given the relatively high market valuations toward the end of the sample period, regressions based on a reduced sample without the IPOs of 1999 are estimated without yielding any significant findings with respect to the influence of venture-backing on Tobin’s Q. If the data contains outliers, the least squares fit might be affected. A measure for the influence of outliers on the regression results is Cook’s D. The Cook’s D statistic measures the change in the parameter estimates caused by deleting each observation. Jobson (1991) argues that a Cook’s D above one indicates the presence of outliers. Although the value of Cook’s D is not higher than one in any of the equations, the observations with the highest Cook’s D values are taken out of the individual regressions. However, the results of the modified estimates widely confirm the earlier coefficient signs and significance levels, suggesting no influence of the potential outliers on the inferences from the regression analyses. The next section provides a brief discussion of the main empirical findings. Moreover, it points out the limitations of the study and addresses alternative explanations for the results.
4.3. Discussion, Limitations, and Alternative Explanations The empirical results suggest that only a subgroup of internationally investing venture capitalists is able to spur the performance of firms going public in France, Germany and the United Kingdom. By contrast, firms backed by nationally
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
281
operating venture capitalists do not outperform non venture-backed IPOs in the sample. This finding is inconsistent with the results of prior U.S. studies. An explanation for the variations in the performance figures of firms backed by different types of venture capitalists might be related to the lack of experience of many players in the relatively young and immature European venture capital market (Hege et al., 2003; Schwienbacher, 2002). Still, there are also several limitations to the approach of the present study, requiring a cautious interpretation of the results. First, it is possible that the investigation does not take into account the quality and intensity of venture backing sufficiently. Although various types of venture capital firms and monitoring proxies are considered, the intensity of venture investors’ support is typically related to the length of their involvement in the portfolio firms. However, the data set does not allow us to address this issue. The analysis does neither include the specialization of venture capitalists in specific investment stages or industries, even though this aspect might be a determinant of their advisory capacities. Second, the IPO sample covers only the 3-year period after the IPO and may not allow detecting whether or not venture-backed firms outperform their non venture-backed counterparts in the longer-term. Previous studies focus on a 5-year post-issue time horizon to investigate the long-run performance of venture-backed firms (Brav & Gompers, 1997). Therefore, it seems useful to extend the analyzed post-IPO period beyond the 3-year time interval to check the robustness of the results. Unfortunately, due to the relatively recent opening of the three European stock market segments that we investigate, such an analysis could only be performed for a small number of firms at the time of this study. Third, our findings related to the market performance may be strongly influenced by the Internet bubble period of 1999–2000 with excessively high stock market valuations. The unsustainable high valuations of firms listed on European stock markets at the end of the 1990s may have distorted true firm values and the marketbased performance measures of the sample firms. Although for robustness reasons the regressions were also estimated without the IPOs of 1999, an “overshooting” phenomenon in the venture capital market during the sample period may have led to a situation, where some venture firms ended up funding too many opportunities (Gompers & Lerner, 2001). Hence, the “overshooting” argument may account in part for the observed performance of venture-backed IPOs in general.
5. CONCLUSIONS This chapter provides stylized facts on the role of venture capitalists in Europe and their impact on the performance of firms going public. Given the recent rise in the
282
GEORG RINDERMANN
importance of IPOs in the continental European stock markets and the relatively recent phenomenon of venture capital in many European countries, this study represents a first comparative empirical assessment of the role of venture capitalists in the going public process and their impact on the long-term performance of IPOs in France, Germany and the United Kingdom. A further contribution lies in the international dimension of the analysis that allows for cross-country comparisons and for testing the robustness of theoretical frameworks in venture capital finance in different financial systems. Our findings suggest that that there are substantial variations in the experience and sophistication of venture capitalists. In particular, international venture capitalists are on average older than national ones, back a larger number of IPOs in the sample, are more often represented on the board, invest with a higher number of syndication partners, and hold larger equity positions in portfolio firms. We report that venture-backed IPOs do not generally outperform non venture-backed issues, irrespective of the performance measure used. Instead, only a subset of international venture capitalists appears to have positive effects on both the operating and market performance of portfolio firms. The result that venture-backed issues do not commonly outperform non venture-backed ones has an important implication for research on venture capital finance. It indicates that the findings of previous studies on the role of venture capitalists in the U.S. and their influence on the operating and long-run market performance of IPO firms can generally not be transferred to European countries. This outcome is consistent with the results of related empirical studies, documenting considerable differences in the impact of venture capital on portfolio firms between the U.S. and Europe. The overall findings are interpreted as evidence for the heterogeneity of venture capital in France, Germany and the UK and might be related to the limited maturity and size of the relatively young European venture capital industry. Given the boom in the IPO markets at the end of the 1990s, many young venture capitalists entered the market, often lacking the necessary skills to assess the quality of portfolio firms. Accordingly, the Financial Times reports: “Venture capitalists bear a huge responsibility in what happened. They did not try to understand the companies they invested in, they only had initial public offerings in mind, and now we see the result” (Financial Times, August 16, 2001). The downturn and consolidation process revealed that many European venture capitalists are relatively inexperienced and cannot guide their portfolio firms during the difficult market situations. This applies to Germany in particular: “The German fling with venture capital has been giddier than elsewhere in Europe. As the Neuer Markt surged towards its peak, the country counted well over 200 groups calling themselves venture capitalists or incubators. Now advisers have difficulty in
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
283
counting 20 serious sources of venture capital for young technology companies” (Financial Times, June 21, 2002). As the venture capital markets of the three countries are undergoing a fundamental consolidation process, more research is required to explore the role performed by venture capitalists in Europe. Future contributions might extend the length of the post-issue period and rely on a larger number of European countries and IPOs per country to check the robustness of the findings. Future work should also include IPOs of the post Internet bubble period to provide more insights with respect to the “overshooting” argument and examine whether the consolidation process in the European venture capital market has reduced the heterogeneity in the quality of venture-backing.
NOTES 1. The Neuer Markt and the Nouveau March´e are growth market segments of the Frankfurt and Paris stock exchanges, respectively. The techMARK is a tracking instrument and comprises technology firms being part of other indices of the London Stock Exchange. 2. Financial institutions were excluded from the sample due to their different balance sheet structures. 3. Although the AIM was established in order to allow small companies to have their shares listed before acquiring a full listing, it cannot be compared to other growth markets in Europe because it is a mixed segment that also contains pubs and football clubs besides technology shares. 4. We exclude 21 firms from the sample because they changed the level of consolidation during this period, or presented financial statements on periods with varying time-spans, limiting the comparability of successive accounting measures. We exclude 19 companies because of liquidations, mergers, and acquisitions by other companies, as well as missing data. 5. This finding is consistent with the timing argument of Lerner (1994) that venture capitalists tend to bring firms public when stock market valuations are relatively high. 6. For example, some of the French IPO firms operate in the beverage, grocery or textile industries. 7. The term “public venture capitalist” refers to State sponsored venture capitalists that tend to follow different objectives and to behave in a different manner when compared to private investors (e.g. Cumming & MacIntosh, 2002). 8. Details on the rank assignment for the reputation of underwriters are outlined in Section 3.4 and in Appendix D. 9. In line with Ljungqvist (1999), the lead venture capitalist is defined as the venture capitalist with the biggest equity stake. Franzke (2001) argues that the venture capitalist with the biggest equity stake tends to be the one that has been involved in the portfolio company the longest. 10. As pointed out by Bascha and Walz (2001), banks and financial institutions, which are controlled by public authorities, tend to be interested in the promotion of regional business structures and employment. 11. The resulting rankings are on a scale of 0–9, whereby a higher number denotes a higher ranking. Generally, a ranking of 8 and above is considered to be a high rank.
284
GEORG RINDERMANN
12. Methodologies that assess the quality of underwriters based on the basis of their market shares as lead underwriter in the IPO market are also used by Megginson and Weiss (1991) and Franzke (2001). See Carter et al. (1998) for a comparative analysis of different proxies for underwriter reputation. 13. The post-issue accounting profitability might be biased by IPO costs, which are frequently treated as an expense in the fiscal year following the issue. The treatment of these costs as expenses creates a considerable bias in the post-issue accounting profitability, and thus limits the comparability of the performance figures over time. 14. Following Barry et al. (1990), the selling intensity of venture capitalists at the IPO is defined as the number of shares sold divided by the pre-IPO holdings. 15. The book-to-market ratio is highly correlated with the inverse of Tobin’s Q.
ACKNOWLEDGMENTS I am grateful to Abolhassan Jalilvand, Allan Zebedee, Armin Schwienbacher, Carsten Sørensen, Carsten Wolferink, Christian Bender, Christian Harm, Douglas Cumming, Hergen Frerichs, Klaus R¨oder, Mark Wahrenburg, Michael Kruse, Nicole Neunh¨offer, Richard Sweeney, Torsten Engel and Uwe Walz for their comments. Moreover, I acknowledge Anne-Lucie Verdem, Jennifer Vandermosten, Jenny Kunz, Martin Bisicky and Stefanie Franzke for providing data and IPO prospectuses. The chapter has benefited from presentations at the University of M¨unster, the Center for Financial Studies Workshop on Venture Capital and New Markets in Frankfurt/Glash¨utten, the 2003 EFMA Annual Meeting in Helsinki and the Doctoral Student Seminar of the 2002 FMA European Conference in Copenhagen. Any errors are mine.
REFERENCES Barber, B. M., & Lyon, J. D. (1997). Detecting long-run abnormal stock returns: The empirical power and specification of test statistics. Journal of Financial Economics, 43, 341–372. Barry, C. B. (1989). Initial public offering underpricing: The issuer’s view – A comment. Journal of Finance, 44, 1099–1103. Barry, C. B., Muscarella, C., Peavy, J., III, & Vetsuypens, M. (1990). The role of venture capital in the creation of public companies: Evidence from the going-public process. Journal of Financial Economics, 27, 447–471. Bascha, A., & Walz, U. (2001). Financing practice in the German venture capital industry: An empirical assessment. Working Paper. University of T¨ubingen. Bottazzi, L., & Da Rin, M. (2002a). Venture capital in Europe and the financing of European innovative firms. Economic Policy, 34, 229–262. Bottazzi, L., & Da Rin, M. (2002b). Financing entrepreneurial firms in Europe: Facts, issues, and research agenda. Working Paper, CESifo Conference on Venture Capital, Entrepreneurship, and Public Policy. Munich.
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
285
Brav, A., & Gompers, P. (1997). Myth or reality? The long-run underperformance of initial public offerings: Evidence from venture- and non-venture-capital-backed companies. Journal of Finance, 52, 1791–1822. Carter, R., & Manaster, S. (1990). Initial public offerings and underwriter reputation. Journal of Finance, 45, 1045–1067. Carter, R., Manaster, S., & Singh, A. (1998). Underwriter reputation, initial returns, and the long-run performance of IPO stocks. Journal of Finance, 53, 285–312. Cho, M.-H. (1998). Ownership structure, investment, and the corporate value: An empirical analysis. Journal of Financial Economics, 47, 103–121. Chow, G. (1960). Tests of equality between sets of coefficients in two linear regressions. Econometrica, 28, 591–605. Chung, K. H., & Pruitt, S. (1994). A simple approximation of Tobin’s Q. Financial Management, 23, 70–74. Cumming, D. J., & MacIntosh, J. G. (2002). Crowding out private equity: Canadian evidence. Working Paper, CESifo Conference on Venture Capital, Entrepreneurship, and Public Policy. Munich. Doukas, J., & Gonenc, H. (2001). Long-term performance of new equity issues, venture capital, and reputation of investment bankers. Working Paper, New York University. EVCA (2000). Yearbook. Zaventem. Fama, E., & French, K. (1992). The cross-section of expected stock returns. Journal of Finance, 47, 427–465. Franzke, S. (2001). Underpricing of venture-backed and non venture-backed IPOs: Germany’s Neuer Markt. Working Paper No. 2001/01, Center for Financial Studies. Gompers, P. A. (1995). Optimal investment, monitoring and the staging of venture capital. Journal of Finance, 50, 1461–1489. Gompers, P. A. (1996). Grandstanding in the venture capital industry. Journal of Financial Economics, 43, 133–156. Gompers, P. A., & Lerner, J. (1997). Venture capital and the creation of public companies: Do venture capitalists really bring more than money. Journal of Private Equity, 1, 15–32. Gompers, P. A., & Lerner, J. (1999). Conflict of interest in the issuance of public securities: Evidence from venture capital. Journal of Law and Economics, 42, 1–28. Gompers, P. A., & Lerner, J. (2001). The money of invention: How venture capital creates new wealth. Boston: Harvard Business School Press. Habib, M. A., & Ljungqvist, A. P. (2001). Underpricing and entrepreneurial wealth losses in IPOs: Theory and evidence. Review of Financial Studies, 14, 433–458. Hege, U., Palomino, F., & Schwienbacher, A. (2003). Determinants of venture capital performance: Europe and the United States. Working Paper, HEC School of Management, Tilburg University and University of Namur. Hellmann, T., & Puri, M. (2002). Venture capital and the professionalization of start-up firms: Empirical evidence. Journal of Finance, 57, 169–197. Himmelberg, C., Hubbard, G., & Palia, D. (1999). Understanding the determinants of managerial ownership and the link between ownership and performance. Journal of Financial Economics, 53, 353–384. Holderness, C. G., Kroszner, R. S., & Sheehan, D. P. (1999). Were the good old days that good? Changes in managerial stock ownership since the Great Depression. Journal of Finance, 54, 435–469. Jain, B. A., & Kini, O. (1995). Venture capitalist participation and the post-issue operating performance of IPO firms. Managerial and Decision Economics, 6, 593–606.
286
GEORG RINDERMANN
Jobson, J. D. (1991). Applied multivariate data analysis (Vol. 1). New York: Springer. Kaplan, S., & Str¨omberg, P. (2003). Financial contracting theory meets the real world: An empirical analysis of venture capital contracts. Review of Economic Studies, 70, 281–315. Kortum, S., & Lerner, J. (2000). Assessing the contribution of venture capital to innovation. Rand Journal of Economics, 31, 674–692. Lang, L., & Stulz, R. (1994). Tobin’s q, corporate diversification, and firm performance. Journal of Political Economy, 102, 1248–1280. Lang, L., Stulz, R., & Walking, R. (1991). Managerial performance, Tobin’s q and the gains from successful tender offers. Journal of Financial Economics, 24, 137–154. Lee, P. M., & Wahal, S. (2002). Grandstanding, certification and the underpricing of venture capital backed IPOs. Working Paper, Emory University. Lerner, J. (1994). Venture capitalists and the decision to go public. Journal of Financial Economics, 35, 293–316. Lindenberg, E., & Ross, S. (1981). Tobin’s q ratio and industrial organization. Journal of Business, 54, 1–33. Ljungqvist, A. P. (1999). IPO underpricing, wealth loss and the curious role of venture capitalist in the creation of public companies. Working Paper, Oxford University and CEPR. Loderer, C., & Martin, K. (1997). Executive stock ownership and performance. Journal of Financial Economics, 45, 223–255. Loughran, T., & Ritter, J. R. (1995). The new issues puzzle. Journal of Finance, 50, 23–52. Loughran, T., & Ritter, J. R. (2003). Why has IPO underpricing changed over time? Working Paper, University of Notre Dame and University of Florida. Lyon, J. D., Barber, B. M., & Tsai, C.-L. (1999). Improved methods for tests of long-run abnormal stock returns. Journal of Finance, 54, 165–201. Megginson, W. L., & Weiss, K. A. (1991). Venture capitalists certification in initial public offerings. Journal of Finance, 46, 879–903. Mikkelson, W., Partch, M., & Shah, K. (1997). Ownership and operating performance of companies that go public. Journal of Financial Economics, 44, 281–307. Mørck, R., Shleifer, A., & Vishny, R. (1988). Management ownership and market valuation. Journal of Financial Economics, 20, 293–315. Ritter, J. R. (1991). The long-run performance of initial public offerings. Journal of Finance, 42, 365–394. Sah, R. K., & Stiglitz, J. E. (1986). The architecture of economic systems: Hierarchies and polyarchies. American Economic Review, 76, 716–727. Sapienza, H., Manigart, S., & Vermeir, W. (1996). Venture capitalist governance and value added in four countries. Journal of Business Venturing, 11, 439–469. Schwienbacher, A. (2002). An empirical analysis of venture capital exits in Europe. Working Paper, University of Namur. Tobin, J. (1969). A general equilibrium approach to monetary theory. Journal of Money, Credit, and Banking, 1, 15–29. Tobin, J., & Brainard, W. (1968). Pitfalls in financial model building. American Economic Review, 59, 99–102. Wang, K., Wang, C., & Lu, Q. (2002). Differences in performance of independent and finance-affiliated venture capital firms. The Journal of Financial Research, 25, 59–80.
287
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
APPENDIX A: PARTICIPATION OF VENTURE CAPITALISTS IN THE IPO MARKET
Origin Panel A: France (Nouveau March´e) BNP Paribas France D´eveloppement Galileo Partners France Financi`ere Natexis France Banques Populaires Sofinnova Partners France/Int. 3i Group/3i Gestion U.K./Int. Banexi Ventures Partners France CDC Participations France/Int. Innovacom (France France/Int. Telecom) ABN Amro Venture NL/Int. France Apax U.K./Int. Panel B: Germany (Neuer Markt) 3i Group/3i Germany U.K./Int. Goldzack Germany Apax U.K./Int. Atlas Venture Germany U.K./Int. Commerzbank Germany Unternehmensbeteiligung TVM Techno Venture Germany Management Deutsche Effekten- und Germany Wechselbeteiligungsgesellschaft (DEWG) Innovacom (France France/Int. Telecom) Schroders U.K./Int. TFG Venture Capital Germany
1996 1997 1998 1999 Total
1
0
3
6
10
0 0
2 0
5 4
1 2
8 6
2 2 1 0 1
0 1 2 1 –
2 1 1 2 1
2 1 1 2 3
6 5 5 5 5
0
0
2
2
4
0
0
2
2
4
0 0 0 0 0
1 1 0 0 3
3 5 3 0 1
4 2 2 4 0
8 8 5 4 4
0
1
0
3
4
0
0
0
3
3
0
0
1
2
3
0 0
1 0
0 0
2 3
3 3
288
GEORG RINDERMANN
APPENDIX A (Continued ) Origin Panel C: U.K. (techMARK) 3i Group Schroders Apax Mercury NatWest Prelude Rothschild Advent Alta Berkeley Baronsmead Charterhouse JAFCO Quester Sofinnova Thompson Clive & Partners
U.K./Int. U.K./Int. U.K./Int. U.K. U.K. U.K. France/Int. U.K./Int. U.S./Int. U.K. U.K. Japan/Int. U.K. France/Int. U.K./Int.
1996 1997 1998 1999 Total
5 2 1 3 1 1 3 1 0 1 1 1 0 1 0
7 1 0 1 1 2 0 1 2 0 1 1 1 0 0
2 1 1 0 1 0 0 0 0 0 0 0 1 1 1
1 1 2 0 0 0 0 0 0 1 0 0 0 0 1
15 5 4 4 3 3 3 2 2 2 2 2 2 2 2
Note: “Int.” denotes international venture capitalists being involved in IPOs of at least two countries of the sample.
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
289
APPENDIX B: DEFINITION OF VARIABLES
Panel A: Dependent Variables Operating/accounting performance variables ROA Operating return on assets, defined as the operating income divided by total assets CFROA Operating cash flow return on assets, defined as the operating income plus depreciation, amortization and provisions divided by total assets ROS Operating return on sales, defined as the operating income divided by total sales CFROS Operating cash flow return on sales, defined as the operating income plus depreciation, amortization and provisions divided by sales Market performance variables Q
BHR
BHAR
WR
Tobin’s Q, i.e. the approximated value of simple Q at the first day of trading, defined as the market value of equity and book value of debt divided by the book value of total assets, where the market value of equity equals the share price times the number of shares outstanding Buy and hold return of an IPO firm, calculated by compounding monthly returns for a given period after the first day of trading (adjusted for dividends, stock splits, and share issues) Buy-and-hold abnormal return for a given period, defined as the difference between the buy and hold return of an IPO firm and the buy and hold return of a benchmark portfolio for the same period Wealth relative, defined as the ratio of one plus the IPO portfolio return over one plus the return on the chosen benchmark over the same period
290
GEORG RINDERMANN
APPENDIX B (Continued ) Panel B: Independent Variables Firm variables Age Alpha Assets Empl EqR IPO Size Mkt Cap Part
SA Sales URanks
Industry dummy variables Biomed ITSINT Media Techno Telecom
Number of years since the firm was founded at the time of the IPO Fraction of the firm’s equity retained by old shareholders immediately after the IPO Total assets in million s Number of employees of the firm (at the end of the specified fiscal year) Equity ratio, defined as the book value of equity divided by the book value of total assets Gross proceeds of the IPO in million s, incl. exercised overallotments (“greenshoe”) Market capitalization of equity at the end of the first trading day in million s Participation ratio of old shareholders in the IPO, defined as number of old shares sold divided by the number of shares outstanding before flotation Inverse of capital intensity, or sales turnover, defined as sales over assets Total sales in million s IPO underwriter reputation rank on a scale of 1–3, where 1 is the most and 3 the least prestigious underwriter Equal to 1 if the IPO firm belongs to the biomedical sector Equal to 1 if the IPO firm belongs to the IT, software and Internet sector Equal to 1 if the IPO firm belongs to the media and entertainment sector Equal to 1 if the IPO firm belongs to the technology sector Equal to 1 if the IPO firm belongs to the telecom sector
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
291
APPENDIX B (Continued ) Panel B: Independent Variables Accounting standard dummy variables IAS Equal to 1 if the IPO firm’s statements are reported under IAS U.S.GAAP Equal to 1 if the IPO firm’s statements are reported under U.S. GAAP Country dummy variables France Germany U.K. VC variables VC Int VC
Nat VC
VC Age
Bank Indep Public Board No IPOs NVC
Equal to 1 if a firm went public in France Equal to 1 if a firm went public in Germany Equal to 1 if a firm went public in the UK Dummy variable equal to 1 if an IPO is venture-backed Dummy variable equal to 1 if an IPO is backed by an international venture capitalist that backed firms in at least two countries of the sample Dummy variable equal to 1 if an IPO is backed by a national venture capitalist that backed firms in one country of the sample Number of years since the lead venture capitalists’ foundation at the time of the IPO Dummy variable equal to 1 if a venture capitalist is a bank-affiliated organization Dummy variable equal to 1 if a venture capitalist is independent Dummy variable equal to 1 if a venture capitalist is owned by a public shareholder Dummy variable equal to 1 if a venture capitalist is on the board of an IPO firm Number of IPOs in the sample backed by a venture capitalist Number of venture capitalists backing a firm at the time of the IPO
292
GEORG RINDERMANN
APPENDIX B (Continued ) Panel B: Independent Variables AVC (LVC) PrEq AVC (LVC) PosEq AVC (LVC) Sale
Percentage of equity owned by all (lead) venture capitalists at the time of the IPO Percentage of equity owned by all (lead) venture capitalists immediately after the IPO Percentage of shares sold by all (lead) venture capitalists at the IPO
Note: The panel describes the different dependent variables that are employed throughout the empirical analysis.
APPENDIX C: CALCULATION OF TOBIN’S Q An approximation of Tobin’s Q that is sometimes referred to “simple Q” is calculated in the following way:
Q=
BV(Assets) − BV(Equity) + MV(Equity) BV(Assets)
where BV and MV denote the book and the market value, respectively. In the present analysis, an approximated value of the “simple Q” at the first day of trading is calculated as follows: Q IPO =
BV(Assetsold ) − BV(Equityold ) + MV(Equityold ) + MV(Equitynew ) BV(Assetsold ) + BV(Assetsnew )
where the sum of MV(Equityold ) and MV(Equitynew ) equals an IPO firm’s market capitalization at the end of the first day of trading, and BV(Assetsnew ) is identical with BV(Equitynew ), i.e. the number of new shares issued multiplied with the offer price per share.
The Performance of Venture-Backed IPOs on Europe’s New Stock Markets
293
APPENDIX D: PARTICIPATION OF THE TOP UNDERWRITERS IN THE IPO MARKET Proceedsa,b
Panel A: France (Nouveau March´e) BNP Paribas 10,882.2 Cr´edit Lyonnais 5,086.8 Soci´et´e G´en´erale 3,758.2 Credit Suisse First 2,332.7 Boston ABN AMROd 717.8 Cr´edit Agricole 518.0 Indosuez Lehman Brothers 499.3 Natexis Banques 373.2 Populaires Groupe Oddo 162.0 Pinatton Proceedsa,b
Panel B: Germany (Neuer Markt) Dresdner Kleinwort 12,860.0 Wasserstein Deutsche Bank 5,027.3 Goldman Sachs & 3,886.6 Co. UBS Warburg 3,878.3 Commerzbank 1,339.7 1,256.8 HypoVereinsbanke Credit Suisse First 1,227.0 Boston DZ-Bankf 1,162.3 ING BHF-Bank 890.8 HSBC Trinkaus 520.1
CM/LRc
No. of Issuesa
Market Sharea
Ranking
26 25 40 6
37.8 17.9 12.7 7.9
1 1 1 1
9.0/8.1 9.0/7.1 9.0/7.1 9.0/9.1
4 25
2.4 1.8
1 2
9.0/8.1 9.0/7.1
3 55
1.7 1.3
1 2
9.0/9.1 –/–
21
0.6
3
–/–
No. of Issuesa
Market Sharea
Ranking
LR/CMc
44
32.3
1
9.0/7.1
40 13
12.6 9.8
1 1
9.0/9.1 9.0/9.1
12 18 25 12
9.8 3.4 3.2 3.1
1 2 2 1
9.0/9.1 9.0/7.1 –/– 9.0/9.1
32 9 9
2.9 2.2 2.2
2 2 2
–/– 9.0/7.1 9.0/8.1
294
GEORG RINDERMANN
APPENDIX D (Continued ) Proceedsa,b
Panel C: UK (techMARK) UBS Warburg 10,874.5 Cazenove & Co. 10,751.9 Merrill Lynch & 9,658.0 Co. Goldman Sachs & 3,961.1 Co. Dresdner Kleinwort 3,818.7 Wasserstein Morgan Stanley 3,238.6 Robert Fleming/JP 2,698.7 Morgan INGg 897.3 Deutsche Bank 874.7 Credit Suisse First 780.8 Boston
LR/CMc
No. of Issuesa
Market Sharea
Ranking
32 38 27
17.5 17.3 15.5
1 1 1
9.0/8.1 9.0/8.1 9.0/9.1
10
6.4
1
9.0/9.1
17
6.1
1
9.0/7.1
17 12
5.2 4.3
1 2
9.0/9.1 9.0/7.1
15 12 9
1.4 1.4 1.3
2 1 1
9.0/7.1 9.0/9.1 9.0/9.1
Source: IFR Thomson Financial. a Between 1995 and 1999. b In million USD. c CM: Carter et al. (1998) ranking; LR: Loughran and Ritter (2003) ranking. d Incl. Rothschild and Banque de Neuflize Schlumberger Mallett. e incl. Bayerische Vereinsbank and Bayerische Hypobank before their merger. f Former DG-Bank. g Incl. Barings and Williams de Broe.
THE NEUER MARKT: AN (OVERLY) RISKY ASSET OF GERMANY’S FINANCIAL SYSTEM Hans-Peter Burghof and Adrian Hunger ABSTRACT In this chapter, we describe the rise and fall of Germany’s Neuer Markt from its promising start to its ultimate failure. We show that the Neuer Markt was designed to serve the special needs of small and medium sized growth firms. However, some regulatory flaws, insufficient means to enforce the rules, the IPO frenzy and the bursting of the stock market bubble destroyed its reputation beyond recovery. The closing of the Neuer Markt and the rebranding and restructuring of the entire Frankfurt stock market indicate the seriousness of the crisis of German public equity markets.
1. THE CREATION OF THE NEUER MARKT From 1997 to 2002, the German equity market consisted of four different market segments. The Neuer Markt as segment for small and medium sized innovative growth firms was the youngest of these four. Other segments were the Official Trading (Amtlicher Handel), the Regulated Market (Geregelter Markt)1 and the Unofficial Regulated Market (Freiverkehr). Even before the founding of the Neuer Markt, the concept of market segmentation and other measures in capital market reforms were introduced. These measures aimed to give a wide range of firms The Rise and Fall of Europe’s New Stock Markets Advances in Financial Economics, Volume 10, 295–327 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3732/doi:10.1016/S1569-3732(04)10011-X
295
296
HANS-PETER BURGHOF AND ADRIAN HUNGER
access to public equity markets and thereby to enhance the possibilities for German firms to raise equity capital. However, these earlier efforts were not very successful. The Neuer Markt was a new attempt to achieve a broader representation of German firms on stock markets. In this chapter, we describe this new strategy in the context of German financial markets reform, its temporary success and its ultimate failure. The natural starting point is the public discussion of financial markets reform that took place in Germany since the early 1980s.
1.1. Debates and Reforms of the German Financial System in the 1980s and in the 1990s At the beginning of the last decade, the state of German public equity markets was poor, both with regard to the number and market capitalization of quoted companies and with regard to the number of IPOs (Theissen, 1998). In 1993, market capitalization was only 24% of GDP, which was less than in other bank dominated countries such as France (36%) or (then bubble driven) Japan (71%), and far less than in market oriented financial systems such as the U.S. (82%) or the U.K. (140%) (Barth et al., 1997). However, why should German decision makers care for the development of public equity markets at all, if most German firms were seemingly able to meet their financing needs without? The German financial system can be classified as bank based and relationship oriented. This has certain disadvantages such as a lack of flexibility and an inactive market for corporate control. However, it also has some advantages such as a higher degree of long-term thinking, better conditions for long-term and specific investments and better support of firms in financial distress through their relationship lender, the so called “Hausbank.”2 Seemingly, this system is a viable alternative to the market oriented financial systems of Anglo-Saxon countries. Why change this? In the beginning, public debate in Germany focused on the so-called “Eigenkapitall¨ucke,” i.e. “equity gap.” German firms held, by international comparison, a relatively low ratio of equity to total assets. At the end of the 1970s, this rate was about 50% in the U.S. and the U.K., whereas this rate was about 20% in Germany (Claussen, 1984; Deutsche Bundesbank, 1981, 1982). It was assumed that this capital structure was not endogenously chosen but a result of exogenous restrictions in the possibilities of German firms to raise equity capital. In the years after the war this high leverage was not necessarily harmful because, due to persistent growth, almost all companies were successful. However, this changed during the economic crisis of the 1970s when many companies faced financial difficulties. In fact, the number of bankruptcies rose from values below 3,000
The Neuer Markt
297
at the beginning of the 1960s to more than 9,000 in 1980.3 Equity rather than debt from the “Hausbank” was now understood as the main instrument to avoid bankruptcy. Thus, it seemed an urgent task to improve firms’ access to equity. A renewal of the Stock Exchange Act aimed to lower the entry barriers to capital markets. Up to 1987, the German Stock Exchange Act distinguished between an official quotation and a non-official quotation. The Stock Exchange Act set the listing requirements for a listing in the Official Trading only. Official brokers were responsible for a price quotation, so that the investor could be sure that his order was executed. In contrast to the Official Trading, the Stock Exchange Act did not set rules for the non-official quotation. The stock exchanges were only authorized to “permit a regulated unofficial market for securities [. . .] provided that the proper conduct of trading and settlement appears to be assured by trading guidelines.” (§78, Stock Exchange Act; translation by Krause, 2001, p. 66). In this segmentation concept, the Official Trading achieved a sufficient degree of liquidity and trading activity in large stocks, which can be attributed to the relatively stringent listing requirements and stricter regulatory environment. However, the low liquidity and relatively high transaction costs made the non-official quotation not very attractive for either the issuers or for investors (Schrader, 1993). Only a comparatively small number of specialized investors became active in this market segment. Seemingly, the ordinary investor needed a much stronger protection of his rights to be willing to invest his money in rather small and opaque firms. The Official Trading, on the other hand, could have provided such an environment, but was, due to its strict requirements, regarded as much too expensive for small and medium-sized companies. In short, the potential capital market reform suffered from the dilemma that the new laws could not at the same time lower entry barriers through lower listing requirements and at the same time increase market liquidity through stricter regulations. The compromise was to introduce a new market segment placed in-between the Official Trading and Unregulated Market. Simultaneously, the legislator was forced to implement three directives of the European Community of the years 1979, 1980 and 1982.4 The objective of these directives was a Europe-wide harmonization of security trading and a strengthening of the listing requirements to guarantee a higher protection of investors. These intentions were contrary to the compromise that made listing easier for small and mid-sized companies. Consequently, to implement the directives of the European Community, the listing requirements were heightened for the Official Trading when the Stock Exchange Admission Regulation took effect in 1986. Additionally, to facilitate the capital raising by small- and mid-caps, the Regulated Market was established, featuring less strict listing requirements concerning size and age of the company and the contents of the issuing prospectus.
298
HANS-PETER BURGHOF AND ADRIAN HUNGER
At first sight, this development lead to an increase in the use of public equity markets to raise equity capital. The issued capital of domestic companies in the 1980s was as high as in the past 35 years combined (von Rosen, 1995). However, to a great extent this was due to secondary offerings of established quoted companies. Very few new companies joined public markets. From 1987 to 1993, no less than 1,292 companies went public on the New York Stock Exchange,5 1,106 on the London Stock Exchange and 384 on the Tokyo and Osaka Stock Exchanges. In Germany, the number of IPO companies was 146 (von Rosen, 1995). Moreover, the age of companies going public in Germany was 41–49 years on average (depending on the time period). This was three to four times higher than in the U.S. (Neumann, 1997) and four to five times higher than in the U.K. (Goergen & Renneboog, 2003). If at all, the reform had only been a partial success. Seemingly, the German capital markets suffered from regulatory flaws impeding, in particular, young and innovative growth firms to enter public equity markets. This contrasted sharply with the development of NASDAQ in the United States. Since its foundation in 1971, NASDAQ successfully attracts young growth companies. It provided the financing for thousands of high growth and high tech firms, some of which are of great significance today. The failure of German financial markets was more and more understood as a severe barrier to the innovation process.6 However, young and innovative firms did not fit into the Official Trading, where firms of many times their size were listed. On the other hand, the Regulated Market provided an insufficient degree of liquidity, and the Stock Exchange Act did not protect the Unofficial Regulated Market that remained to be “an opaque and illiquid inter-broker telephone market with low listing requirements” (Ljungqvist, 1997, p. 1311). In effect, both market segments were not sufficiently attractive for issuers. Therefore, some German firms like Digitale Telekabel AG or iXOS Software AG went public on NASDAQ, and more planned to do so.7 Both efficiency arguments and the growing international competition between different stock exchanges brought up the urge for the Frankfurt stock exchange to introduce a special market segment with a strong law-protected background to attract small- and mid-cap firms (B¨uschgen, 1997; Hopt et al., 1997). The discussion in the mid-1990s was back where it had been in the early 1980s. However, the lessons of the failed reform from 1986 had to be incorporated into the design of the new structure of the German equity markets and should have led to a rather different structure than the earlier compromise. To achieve this new structure through legal means would have made it necessary to completely modify the existing Stock Exchange Act containing rules for the Official Trading, Official Regulated Market and Unofficial Regulated Market. However, this was not needed when introducing an additional market segment. Thus, to speed up the process and achieve a higher degree of flexibility and efficiency, the Deutsche B¨orse (as the
299
The Neuer Markt
responsible body for the Frankfurt Stock Exchange) chose a mezzanine legal status for the new market segment: a private organization of trading in the context of a public stock exchange. If insufficient investor protection was at the heart of failure of the Regulated Market, new rules were required to gain the trust of the public. Whereas large firms often attain some degree of publicity and transparency through their business activities, a market segment for small and medium sized firms had to supply sufficient information for investors through comprehensive and strictly enforced disclosure rules. It had to require relevant and reliable information in the prospectus as well as a continuous flow of such information after the IPO to get and keep the public informed and interested. Among the potential investors, foreigners were not of the least importance, and disclosure requirements had to be designed to be in line with international standards. Furthermore, it could be argued that young growth firms are more flexible and less bureaucratic compared to larger firms. The inherent tendency of large bureaucracies to adhere to the rules set up by the authorities might offer an additional margin of investor protection that is not available in smaller and less law-abiding firms.8 Thus, the new concept could not be a compromise. It had to contain more and not less information requirements and investor protection than the rules for the Official Trading, even if this would limit the number of firms that can comply with such stricter rules. The general acceptance of the new market segment relied on investors’ trust, i.e. on the reliability of the information and the fairness and efficiency of market transactions. German investors are said to be cautious and skeptical concerning stock markets. Thus, the development of a better understanding and trust into equity investments at stock markets, of so called “Aktienkultur” (“equity culture”), was understood as a main prerequisite for a better development of the German stock markets. The new market segment had to pass the test of trustworthiness. If it succeeded, the structure of the German financial system might change altogether. Capital market access for young and innovative firms might spur innovation and give new growth incentives to the somewhat ailing German economy. If it failed, even just to some degree, positive expectations could turn into a negative attitude towards the new market segment and share investments in general. The consequences for the German financial system and economy could be lasting stagnation. Thus, stakes were high when the new market segment was introduced.
1.2. The Design of the Neuer Markt On March 10th, 1997, the Deutsche B¨orse founded the Neuer Markt as a privately organized market. However, privately organized markets are not regulated markets
300
HANS-PETER BURGHOF AND ADRIAN HUNGER
in the sense of the Investment Services Directive of the European Community. According to this directive, a market is regarded as a regulated market if, for instance, the Stock Exchange Supervisory Authority enacts the trading conditions and listing requirements.9 In particular international investors see a regulated market as an important qualification of a market segment. Furthermore, some institutional investors are allowed to invest in regulated markets only (Potthoff & Stuhlfauth, 1997). To avoid the problem of not being a regulated market, the admission to the Neuer Markt required an admission to the Geregelter Markt with a simultaneous waiver of being listed at the Geregelter Markt in favor of a listing at the Neuer Markt. Thus, the Neuer Markt was, despite its private organization, a regulated market (Kersting, 1997). The main purpose of the admission requirements was not to facilitate going public but to gain investors trust. At least on paper, the listing requirements of the Neuer Markt were even stricter than for an admission to the Official Trading.10 One requirement was the need to appoint at least two designated sponsors. The designated sponsors were obliged to post price indications or spreads continuously (Part 3, No. 2 Rules and Regulations Neuer Markt). As such, designated sponsors were responsible for liquidity and marketability of the shares. This lowered the transaction costs for investors. However, the general public expected the designated sponsors to do more than that. At least some investors wanted them to guarantee fair transactions and a good conduct of the firms they were sponsoring. The Rules and Regulations of the Neuer Markt required an equity capital of the issuer of at least d1.5 million (Part 2, No. 3.1).11 The aggregated proceeds had to amount to at least d5 million, and the minimum nominal value of the issue had to be at least d250,000, with a minimum number of shares of 100,000 (No. 3.7). The issuer was required to have a track record of, at least, three years (No. 3.2), and the free-float of the issue had to be at least 25% of the aggregate nominal volume. Preferred allotments were not to be included in the shares that were regarded as being widely held by the public (No. 3.10).12 Moreover, the issuer had to submit a prospectus in accordance with the Securities Prospectus Regulation (Verkaufsprospekt-Verordnung), which included additional information concerning sources and applications of funds (No. 4.1.10), affiliated enterprises (No. 4.1.11), profits, losses and dividends per share (No. 4.1.12), consolidated financial accounts (No. 4.1.13) and information about risk factors. Risk factors are “any factors that could have a substantial negative influence on the financial condition of the issuer or that could endanger the issuer’s business success” (No. 4.1.16). Information about risk factors would be crucial for investors to assess the idiosyncratic risk of the issue. After going public, quarterly reports, financial statements and management reports according to IAS or U.S.-GAAP had to be published in both German and
301
The Neuer Markt
English (No. 7.1). Furthermore, the issuer was required to hold an analyst meeting at least once a year (No. 7.1–7.3) and to publish an annual corporate timetable and detailed information concerning the convocation of the shareholders’ meeting. Existing owners have to lock-up their shares for a six-month period (No. 2.2). According to the original version of the Rules and Regulations of the Neuer Markt the issuer has to enforce the lock-up. It was later decreed that the issuer was required to obtain written confirmations from the existing shareholders that they would honor the six-month lock-up period. These statements had to be enclosed with the application form for an admission.13 Seemingly, there were some doubts that everybody was complying with the rules. As from March 1st, 2001, Deutsche B¨orse also required firms to notifiy the public of every share transaction of the issuer or management (No. 7.2). However, Deutsche B¨orse did not introduce the obligation to inform the public about the intended transactions of owners or managers in shares of the company ex ante. This information would have been of great value to investors, because it could have contained valuable information about the true prospects of the firm. Also, it could have made potential insider trading more transparent for the public. However, for the same reasons it might have made going public less attractive. The late changes in the regulations concerning insider transactions signal that these posed a severe problem and that the original rules were inadequate. To emphasize the credibility of the extensive listing requirements, the issuer was obliged to pay a fine up to d100,000 to the Deutsche B¨orse for each nonperformance, delayed or incomplete performance of the obligations resulting from an admission to the Neuer Markt (No. 2.1.4). It might be argued that this fine was insufficient. However, the private status of the Neuer Markt restricted the means of enforcement, in particular with respect to non-monetary sanctions that are particularly valuable in distress when monetary penalties are no longer credible. Some times before the introduction of the Neuer Markt, new laws for ad hoc-publicity and insider trading had been implemented for the whole German capital market, and a new regulatory agency, the Bundesaufsichtsamt f¨ur den Wertpapierhandel, had been installed to enforce these rules. Together with the special requirements of the Neuer Markt, this seemingly powerful set of rules and regulations gave support to the high expectations of the Neuer Markt as a segment for young and innovative growth firms.
1.3. The Marketing The Neuer Markt intended (and for some time succeeded) to be more than just another trading segment for new shares. The idea of the Neuer Markt was to
302
HANS-PETER BURGHOF AND ADRIAN HUNGER
facilitate the research of banks and other information intermediaries through strict listing and disclosure rules, to reduce transaction costs by designated sponsors and to increase the investors’ readiness for trading (Francioni & Gutschlag, 1998). Thereby it seemed possible to duplicate, although on a much smaller scale, the success of NASDAQ in the United States. To reach these ambitious targets, the Deutsche B¨orse promoted the new market segment actively by a number of different activities. For instance, the Deutsche B¨orse presented the Neuer Markt and its listed companies at fairs, congresses and meetings of analysts and investors at home and abroad. Furthermore, the Deutsche B¨orse developed a high media presence for the Neuer Markt by presentations of listed companies and IPO candidates in special publications, advertising campaigns, a separate section in the stock quotations of financial newspapers and intensive public relations activities.14 In its design of the Neuer Markt, the Deutsche B¨orse stressed the importance of indices as a marketing device and as an underlying for potential derivatives. Until the founding of the Neuer Markt, the Official Trading was the only market segment represented through several separate indices. The 30, 70 and 100 most important blue chips of the Official Trading were united in the DAX 30-Index, the MDAX and the DAX-100.15 Among these, the DAX 30-Index received most attention because turnover concentrated on the shares represented by this index, whereas the liquidity of the other so-called Nebenwerte was decisively smaller. However, the CDAX-Index as a broad market index referred to all stocks listed in the Official Trading and the Regulated Market. With the foundation of the Neuer Markt on March 10th, 1997, the Deutsche B¨orse created the NEMAX-All-Share-Index as a special index for the new market segment. This index was computed both as a performance-index and a price index and comprised all domestic and foreign stocks listed at the Neuer Markt. Since July 1st, 1999, the NEMAX 50-index was computed from the prices of the 50 most important firms at the Neuer Markt, and since May 15th, 2000, another 10 indices for industrial sectors extended the index family. The latter ones were also computed as price- and as performance-indices and referred to market capitalization and equity turnover.16 Due to the fast growing significance of the Neuer Markt, various derivatives were constructed. Apart from equity warrants and discount certificates, several investment banks created special certificates on Neuer Markt blue chips like Aixtron AG, STADA Arzneimittel AG, T-Online AG and many others. Dresdner Kleinwort Wasserstein, for instance, issued index participations on NEMAX-50 index, open-end certificates on NEMAX-All-Share and NEMAX 50 indices and sector-index certificates on the NEMAX-All-Share for Biotechnology, IT-Services and Media & Entertainment. The German-Swiss derivatives exchange EUREX
The Neuer Markt
303
also provided the market with futures and options on the NEMAX 50 and options on eight individual Neuer Markt firms. The efforts of the Deutsche B¨orse to establish a high reputation for the Neuer Markt met with great resonance on the side of potential IPO candidates. In 1997, 11 of 16 listed companies at the Neuer Markt declared that the good reputation of the Neuer Markt was the main motive for a listing in this segment in favor of a listing in another market segment (Theissen, 1998). The Neuer Markt became very prestigious soon, and firms could enter this prestigious market segment with a rather short track record and low turnover. For example, EM.TV & MERCHANDISING AG went public on October 30th, 1997, especially in this segment because its turnover volume was not sufficient to apply for a listing at the Official Trading or the Regulated Market. A listing at the Unofficial Regulated Market would have been a valid alternative, but the concept of the Deutsche B¨orse and the intensive marketing of the Neuer Markt promised more gains from the IPO particularly with respect to publicity (Kempkes & Haffa, 1998). Thus, despite a rather small number of IPOs during its first year, the Neuer Markt was successfully set on track to become a huge success.
2. THE IMPRESSIVE (BUT TEMPORARY) SUCCESS STORY OF THE NEUER MARKT 2.1. Underpricing, Investor Behavior and IPO Frenzy The initial success of the Neuer Markt becomes more stunning when comparing its development with that of other public equity markets in Germany in the preceding years. During the 14 years before the founding of the Neuer Markt, there were, on average, 16 IPOs per year. From 1985 to 1996, the aggregated nominal gross proceeds amounted to about d2.2 billion as a yearly average. However, without counting the IPO of the Deutsche Telekom AG in 1996 – by far the largest IPO in the history of German financial markets- the aggregated nominal gross proceeds per year are only d1.4 billion on average. After the foundation of the Neuer Markt in 1997, both the number of IPOs and the aggregated gross proceeds increased sharply – from 36 IPOs in 1997 to 67 IPOs in 1998 and to 168 IPOs in 1999 and 159 IPOs in 2000. The aggregated nominal gross proceeds increased from d2.536 billion in 1997 to d25.609 billion in 2000. Table 1 provides an overview of this development. It was not only the Neuer Markt that profited from this hot issue phase. Even firms that did not fit into the concept of the Neuer Markt did consider going public more seriously now, and some decided to use the favorable IPO climate to enter
304
HANS-PETER BURGHOF AND ADRIAN HUNGER
Table 1. Going Public in Germany 1977–2002. Year
1977–1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Numbers of IPO
10 (in all) 11 24 11 27 19 14 23 23 20 8 8 15 20 14 36 67 168 159 22 5
Nominal Gross Proceeds (in million d) n.a. n.a. n.a. 922 2,364 905 418 1,164 1,639 1,609 411 426 637 3,551 12,684a 2,536 3,292 12,964 25,609 3,179 245
Source: Titzrath (1995); Deutsche B¨orse (2000, January). Deutsche Telekom with proceeds of d10,055 million.
a Including
other market segments. Thus, the number of quoted shares increased markedly in all market segments, although to a lesser degree than at the Neuer Markt. Furthermore, the additional IPOs at the Official Trading, although small in number, mobilized high aggregated gross proceeds because, on average, larger firms went public in this market segment than in the other market segments. The success of the Neuer Markt was fuelling a new enthusiasm of German investors for share investments. This increase in the public interest for shares was well prepared by the massive marketing campaign for the IPO of the Deutsche Telekom AG that started trading on the Official Trading November 18th, 1996. For the first time in Germany, an IPO was promoted through all the channels of the mass media and with all techniques of modern marketing. The campaign succeeded in selling d10 billion worth of shares, i.e. more than seven times the average aggregated IPO proceeds of the years before, in one single IPO. The initially positive trend of the Deutsche Telekom stock price assured many unsophisticated
The Neuer Markt
305
investors that they had made the right decision to turn to equity investments for the first time in their life. When, four months later, the Neuer Markt started trading, the good taste of the Telekom IPO was still fresh. Furthermore, the high level of underpricing of some of the first IPOs stimulated the demand of investors for new shares. In particular, the first IPO at the Neuer Markt, that of Mobilcom AG, was also accompanied by a successful marketing campaign very much along the lines of Deutsche Telekom AG (they even employed the same famous actor in their TV spots). Mobilcom AG had a first-day return of more than 50%. Such gains in a nick of time appealed to investors’ greed. And, as high levels of first-day returns proved to be persistent at the Neuer Markt, more and more investors were keen on such would-be gains and were buying IPO shares. Many investors participated in all IPOs without any personal knowledge about the firms and without being able to draw relevant information from the issue prospectuses. The IPO frenzy becomes more understandable when looking at the first-day gains that were, on average, much higher for IPOs on the Neuer Markt than for IPOs on the Official Trading. In particular during the second year of the Neuer Markt, investors could have gained a profit of 80.24% on average by selling new shares on their first trading day, given that they received the same amount of new shares in all the issues. In comparison to the average first-day return (underpricing) of 10.42% in the Official Trading, these profits were highly attractive for private investors, even if these had been aware of the greater risk and potential adverse selection effects giving uninformed investors larger allotments in less profitable IPOs.17 However, many private investors were surprised and embarrassed that they usually received more shares of issues that had lower first-day returns in comparison to issues that were more underpriced. They alleged that the methods of distribution of new shares were not fair and triggered a public debate that led to some banks making the allocation process more transparent. Table 2 provides an overview of the development of the underpricing on the Neuer Markt in comparison to the Official Trading for the years 1997–2002.18 As might have been expected, in many IPOs (and these were the most attractive ones), private investors did not receive as many new shares as they intended to buy. To avoid disappointment, some opened accounts at different banks to multiply their orders. This behavior gave banks additional incentives to promote IPOs and thereby to attract new investors hoping for an additional share in the issues. However, because the Neuer Markt was aiming at small growth firms, the issues were too small to satisfy general demand. However, a positive secondary market trend made many private investors believe that the Neuer Markt was a riskless way to earn money even for investors buying the shares in the secondary market. Many investors
306
HANS-PETER BURGHOF AND ADRIAN HUNGER
Table 2. The Development of IPO Underpricing at Neuer Markt and Official Trading. Official Trading
1997 1998 1999 2000 2001 2002
Neuer Markt
Number of IPOs
Underpricing %a
Number of IPOs
Underpricing %a
9 15 26 14 5 1
7.26 10.42 15.68 11.58 −1.10 −4.55
10 39 117 120 11 1
47.66 80.24 54.84 48.69 10.89 −0.34
Source: Deutsche B¨orse AG, Fact Books, Hoppenstedt B¨orsenforum, B¨orsenzeitung. a Initial return is adjusted by the DAX-100 index.
postponed reflection about the market itself, the issuers and the potential risk of each investment in favor of the chase for easy money. Institutional investors often remained skeptical and saw in the Neuer Markt a market for gamblers. Many of them abstained from investing their own money. However, the general public was more impressed by the short-term gains. Thus, market participation soon spread to less sophisticated investors who became active on their own account after observing their neighbors’ success, or were an easy prey for financial advisors from banks and other financial institutions being less cautious with the money of their customers than their own. A visible consequence of this development was a strong increase in the number of shareholders in Germany. This number almost doubled from 3.7 million on average for the period before the foundation of the Neuer Markt to 6.2 million in the year 2000 (see Fig. 1). Whereas these developments to some degree influenced all market segments, the Neuer Markt set the pace. The relative importance of the Neuer Markt is shown by the distribution of the IPOs over the market segments. More than 68% of all IPOs took place on the Neuer Markt, which accounted for more than 43% of the total gross proceeds. However, under the special market conditions of the late 1990s the Official Trading, with an average of d366.41 million per IPO, raised most of the gross proceeds. The data in Table 3 show more details concerning the distribution of the IPOs over the different market segments. It is interesting to note which firms went public at the Neuer Markt and which went public on the other market segments of the Frankfurt stock exchange during the hot issue phase. Burghof and Fischer (2002) show that firms that went public usually grew faster and had higher earnings than privately held firms. However, for Neuer Markt firms this effect was stronger with respect to growth, and for
307
The Neuer Markt
Fig. 1. Number of Shareholders in Germany 1988–2002. Source: Deutsches Aktieninstitut e. V. 2002.
Table 3. Numbers of IPOs, Initial Capital and Gross Proceeds 1997–2002. Official Trading Number of IPOsa Number of IPOs (in percent of total) IPO underpricing (mean in %)b Gross proceedsc (in d millions) Gross proceeds (in percent of total) Gross proceeds on average (in d millions) Initial capitald (in d millions) Nominal capital (in percent of total) Initial capital on average (in d millions)
Regulated Market
Unofficial Regulated Market
Neuer Markt
Total
70 16.09
45 10.34
22 5.06
298 68.51
11.16 25,648.64 54.58
17.30 784.10 1.67
39.65 145.18 0.31
53.64 20,415.55 43.44
42.34 46,993.48 100
366.41
17.42
6.60
68.51
108.03
5,754.28 55.99
272.87 2.66
66.49 0.65
4,183.16 40.70
10,276.80 100
82.20
6.06
3.02
14.04
23.62
435 100
Source: Hunger (2002). a The total number of IPOs was 457. For 22 IPOs no sufficient data was available; the data refer to the remaining 435 IPOs. b Initial return, market-adjusted with the DAX-100 index. c Gross proceeds are measured as the number of issued shares mulitplied with the issue price per share. d Initial capital refers to the registered share capital.
308
HANS-PETER BURGHOF AND ADRIAN HUNGER
Table 4. Characteristics of IPO Firms on the Neuer Markt and Other Market Segments. One Year Before Flotation
Employees Total assets (millions of DM) Sales (millions of DM) Shareholders’ equity (millions of DM) Capital expenditure (%) Growth (%) Leverage (%) Intangible assets (%) ROA (%) Age Issue proceeds (millions of DM) Percentage of high R&D intensity companies (%)
Neuer Markt IPOs
IPOs on Other Market Segments
(N = 114)
(N = 35)
Mean
Median
Mean
Median
227 51.8 86.0 10.5 66.4 54.3 69.5 26.7 17.6 13.5 81.1 71.9
114 22.7 34.9 4.8 76.7 34.9 73.3 13.7 17.9 11.0 58.5
3,297 723.7 1,346.7 185.1 35.7 16.5 72.5 25.7 20.9 53.9 323.2 5.7
781 164.2 223.5 30.6 22.0 16.9 73.1 13.1 17.6 30.0 109.8
Note: Capital expenditure is gross investment in intangible assets and property, plant, and equipment over the sum of end of the year property, plant, and equipment plus intangible assets. Growth is the annual rate of growth of sales. Leverage is book value of debt divided by book value of debt and equity. Intangible assets are “immaterielles Anlageverm¨ogen.” ROA is EBITDA (earnings before interest, taxes, depreciation, amortization, and change of provisions) over total assets. Age is years since foundation. R&D intensity is measured as average R&D expenditure over sales in the respective industry. High R&D intensity means a percentage larger than 8.5.
non-Neuer Markt firms with respect to earnings. IPO firms in other market segments were generally older and had no strong financial needs. The IPO was therefore mainly a device to change the ownership structure of such firms.19 By contrast, firms that went public at the Neuer Markt were, by comparison, very young, grew faster and had strong financial needs that access to public equity markets was intended to meet. Ownership was concentrated in the hands of managers and founders, both before and after the IPO. Overall, the Neuer Markt was able to attract young and innovative growth firms. Table 4 gives an overview of the characteristics of German IPO firms from March 1997 (the opening of the Neuer Markt) until September 1999. The data comes from Burghof and Fischer (2002). We refer to Fischer (2002) for further information on this data. The successful establishment of the Neuer Markt also had a strong influence on the venture capital market because it offered an attractive exit channel (Deutsche Bundesbank, 2000). In 1995, only 8% of the divestments of venture capitalists
309
The Neuer Markt
were done through an IPO.20 Seemingly, at the time venture capitalists planned to divest, most portfolio firms were still too small for an IPO on the Official Trading, whereas the other market segments provided no attractive conditions for an IPO. Thus, the Neuer Markt was an ideal exit channel. It gave additional prestige to young growth firms and venture capitalists an attractive exit price. Consequently, venture capitalists had a significant share in more than 40% of the IPO firms at the Neuer Markt (Kraus & Burghof, 2003). In 1998 and 1999, the IPO, mostly taking place at the Neuer Markt, was the exit channel for almost 20% of all terminated venture capital investments of German venture capitalists (BVK). The new earnings from IPOs also changed the relative attractiveness of investments in venture capital funds that realized spectacular returns for some years. Investments in these funds soared, and many institutional investors set up new entities for such investments.
2.2. Secondary Markets: Bubble and Bust At the end of the 1990s, both the large number of IPOs and a strong increase in share prices at the Neuer Markt and the other market segments lead to a strong increase in market capitalization of the whole market. From $355 billion in 1990, market capitalization of the main markets and the parallel markets of the Deutsche B¨orse rose to a maximum of $1,432 billion in 1999, with a growth rate of more than 30% in 1998 and 1999. The development of the market capitalization of the whole German public equity market is shown in Fig. 2. The most striking point in the success story of the Neuer Markt is the development of share prices. Stock markets all around the world experienced a strong bull market at the end of the nineties. Nonetheless, the Neuer Markt strongly outperformed any benchmark. During the bullish period until March 10th, 2000, the NEMAX-All-Share index increased by about 1,636% in three years and one month only. Meanwhile, the DAX-100 index grew by comparatively modest 117%.21 The strong focus of the Neuer Markt on innovation and new technologies seemed to guarantee both higher risk and return. However, the Neuer Markt even outperformed the technology oriented NASDAQ. The NASDAQ-100 index increased only by 459%. Thus, other explanations than outstanding expected earnings from technological progress might have been responsible for the increase of stock prices at the Neuer Markt. The Neuer Markt was the first candidate for any speculation on the existence of a stock price bubble. The following graph shows the developments of DAX-100, NASDAQ-100 and NEMAX-All-Share indices (see Fig. 3).
310
HANS-PETER BURGHOF AND ADRIAN HUNGER
Fig. 2. Market Capitalization of Shares of Domestic Companies. Source: FIBV (2002). Figure shows the market capitalization of companies listed on the main and parallel markets of Deutsche B¨orse.
Fig. 3. Market Performance of Neuer Markt, Official Trading and NASDAQ from 1997 to 2002. Source: Dresdner Bank AG.
311
The Neuer Markt
However, seen from an ex post-perspective, all this is bubble economics. Neither the prices at the Neuer Markt nor at the Official Trading nor at NASDAQ could persist when confronted with realistic economic outlooks and data. At the Neuer Markt, price per earnings ratios were meaningless because many of the IPO firms did not earn any money. Thus, any price calculation had to rely on expected earnings. However, the aggregated expected earnings of the firms at the Neuer Markt that could justify the stock market valuations were beyond any sensible comprehension. The bubble had to burst when a sufficient quantity of investors became aware of this fact. Both DAX-100 and NEMAX-All-Share index reached a maximum on March 10th, 2000, and then fell back sharply to levels of 1998 or 1997. The Neuer Markt, that had experienced the strongest increase before the bubble, consequently suffered most. In early October 2002, the NEMAX-All-Share index reached its alltime low of 349.02, scarcely 4% of its maximum in March 2000. After that date, the Neuer Markt recovered somewhat. However, this slight increase was a dividend on the ruin of the Neuer Markt, because in October 16th, 2002, the Exchange Council of the Frankfurt Stock Exchange finally decided to abandon the old segmentation concept and in particular the Neuer Markt. The next section will discuss to what degree this is a particular failure of the Neuer Markt. Many equity markets all over the world experienced both the bubble and the following market crash in 2000, although not in such an extreme manner as the Neuer Markt did. Before the market crash, the Neuer Markt had been the leading market for young and innovative growth firms in a family of new competing markets in Europe aiming at this segment. The Neuer Markt did not loose this leading position after the crash, because the competing exchanges suffered as well. However, the importance of all these market segments was greatly reduced. Tables 5 and 6 compare the relative importance of the different new markets in Europe at different times. They demonstrate that all the markets lost heavily, but that the Neuer Markt lost more than his most important competitors. Table 5. European Growth Markets in Comparison (August 2001). August 2001
Neuer Markt AIM Nuovo Mercato Nouveau March´e NASDAQ Europe
Market Capitalization (in d Millions)
Equity Turnover (in d Millions)
58,553 20,579 12,909 13,000 9,380
2,628 512 1,051 381 62
Source: Deutsche B¨orse AG (2001, September).
312
HANS-PETER BURGHOF AND ADRIAN HUNGER
Table 6. European Growth Markets in Comparison (September 2002). September 2002
Market Capitalization (in d Millions)
Equity Turnover (in d Millions)
21,625 14,812 6,483 6,000 2,807
3,116 333 439 211 13
Neuer Markt AIM Nuovo Mercato Nouveau March´e NASDAQ Europe Source: Deutsche B¨orse AG (2002, November).
3. THE BREAKDOWN OF PRICES AND REPUTATION 3.1. The Economic Failure of the Market and its Participants The rise and fall of the Neuer Markt is best described by the development of the number of listed companies and their market capitalization. During the growth period of the market the number of quoted companies increased from two at the start on March 10th, 1997, to more than 300 in 2000. The market capitalization reached its highest level on March 10th, 2000, with d234 billion. At that time, the bubble burst and just two and a half years later the total market capitalization had decreased to d29 billion. This is a reduction in market value of more than d200 billion, or 87.5%. At the same time, many firms had to leave the Neuer Markt, while the primary market suffered an almost complete breakdown. From the five IPOs in Germany in 2002, only one company (REpower AG) applied for a listing at the Neuer Markt (on March 26th, 2002). Thus, with some time lag, the number of firms began to decline (see Fig. 4). The crisis of the Neuer Markt began in early 2000, when several companies had to confess that they could not meet their earnings forecasts from the prospectuses and other publications. Similar announcements are observed internationally from many firms in other stock markets at that time, marking a turning point in the world’s business cycle. However, stock market segments with older firms showed more resilience than did the Neuer Markt. Per se, this is no surprise. As mentioned above, the stock prices of many firms at the Neuer Markt could be justified, if at all, only through high expected – not current – earnings. A change in the general economic outlook should thus hit the stock prices of such firms particularly hard. However, it became obvious soon that investors had to expect worse than just a sharp correction of unrealistic bubble prices. It proved that for many firms at the
313
The Neuer Markt
Fig. 4. The Way Up and Down of the Neuer Markt.
Neuer Markt strong economic growth was a precondition for their survival. Soon after reaching the turning point of stock prices, rumors spread through market reports and popular stock exchange magazines that several companies listed at the Neuer Markt were threatened by bankruptcy. Blacklists were sent around, containing many well-known names. The management of the firms vigorously denied such threats. Nevertheless, the rumors contributed to the general downward trend of stock prices at the Neuer Markt. As we will see, they proved to be still too optimistic in the end. In September 2000, the first company declared itself bankrupt. Gigabell AG started trading on the Neuer Markt on August 11th, 1999, at d38 and reached its highest price level of d123 on July 3rd, 2000, only two months before it went bankrupt. The Deutsche B¨orse expelled Gigabell AG from the listing on February 23rd, 2001, to avert further damage to the reputation of the market.22 Furthermore, the Deutsche B¨orse intensified the listing requirements with effect from March 1st, 2001. In particular, from that day on the members of the board of directors as well as the members of the supervisory board had to publish their own sales of stocks to protect the other shareholders from the possible disadvantages of insider trading. However, market participants commented that these measures were both too late and too little. Further insolvencies followed, for instance, Micrologica AG, Infomatec AG, Met@box AG, Teldafax AG and Refugium AG, and the plunge of stock continued. From July 2001 onwards, several of the distressed or insolvent companies
314
HANS-PETER BURGHOF AND ADRIAN HUNGER
applied for a listing at the Regulated Market to relinquish the listing at the Neuer Markt.23 The first to do so (on July 6th, 2001) was Sunburst Merchandising AG. Simultaneously, the Deutsche B¨orse AG prepared for the case of a continuing reduction of the quotations and announced to strike penny-stocks and insolvent companies from the stock list with effect from October 1st, 2001. At the end of August, the NEMAX 50-index fell short of 1,000 points for the first time – the base level of ultimo 1997. Meanwhile, the insolvent Kabel New Media AG was the first company that was expelled from the stock list due to the new delisting regulations. With Biodata Information Technologies AG, Brokat Infosystems AG, Management Data Media Systems AG, mb Software AG, and Lipro AG a new wave of insolvencies hit the Neuer Markt. Biodata Information Technology AG and Lobster AG were expelled. However, on April 23rd, 2002, the Deutsche B¨orse had to revoke its delisting regulations because several companies went to court to fight their delisting and won their case. To conclude the sad story where it began, in September 2002 Mobilcom AG, the first and one of the largest listed companies at the Neuer Markt and for many the symbol of its initial success, ran into severe difficulties when its major shareholder (France T´el´ecom) withdraw its support.24 As a result, market performance hit an all-time low. Subsequently, the Deutsche B¨orse announced the dissolution of the Neuer Markt at the end of 2002. At this date, the NEMAX 50-index ended up at around 380 points and the NEMAX-All-Share index ended at around 420 points, i.e. at about 5% of their all-time high.
3.2. Fraud and Reputation However, it is not the lack of economic success alone that enraged investors and tarnished the reputation of the Neuer Markt. In November 2000, the managers of Infomatec were imprisoned for alleged stock price manipulations and insider trading.25 In December 2001, public prosecutors investigated potential price manipulation concerning one analyst and the chairman of CPU AG. In February 2002, it was rumored that most of the turnover of Comroad AG did not exist. Six months later the founder of the company was arrested and the Deutsche B¨orse expelled Comroad AG from the stock list. In October and November 2002, the founder of Kinowelt Medien AG was arrested because of alleged embezzlement and protraction of insolvency. The founders of the former media-star EM.TV & MERCHANDISING AG were also brought to court because of fraud. However, these are only some of the most notorious cases. Altogether, more than 40 companies suffered from alleged insider trading or counterfeited ad-hoc
The Neuer Markt
315
announcements or have applied for insolvency (partly because of fraudulent behavior).26 We identify certain typical patterns from the individual fraud cases that show a combination of wrong incentives, lack of supervision and immoral behavior. To begin with, founders usually planned to sell some of their shares after the IPO to reduce their idiosyncratic exposure. To achieve an attractive price for their shares, they undertook special efforts in investor relations. However, because managers were not obliged to inform investors about their selling intentions, investors could not distinguish between the regular information policy on behalf of the firm, and measures undertaken with the intent to increase the proceeds of sales by insiders. Since 1996, the German stock exchange law required management to inform about new developments that might have a substantial influence on share prices investors in ad-hoc announcements. However, the regulators soon realized that managers used these ad-hoc announcements for marketing purposes in the context of investor relations. To make a positive impression on capital markets, managers were quick to inform investors about important new strategic alliances or important contracts. Sometimes they did so although the contract only contained intentions and was not sufficiently binding. If the final agreement was not accomplished, they had to correct the earlier ad-hoc announcement. Sometimes, they forgot to do so or informed investors much later. In any case, investors might have bought shares under the false assumption that the contract was binding or had been sufficiently fixed in the meantime. In some cases, insiders sold some shares while investors were wrongly informed in this way. However, it is not straightforward about what and in which way managers should inform. Likewise, it is hard to prove that capital markets did not anticipate the “marketing” character of some ad-hoc announcements and that these ad-hoc announcements were able to influence the share price. In many cases, such anticipation and the strong upward trend of the Neuer Markt could have reduced any informational content of ad-hoc announcement to mere noise. Although some lawyers intend to specialize on such cases and collected a high number of claims (and fees) of individual investors, up to now no such case has been brought to court successfully.27 Compared with this type of potential fraud, cases in which managers cooked the books are much easier to prove. The most spectacular case of this kind was Comroad AG. It was found that only 1.4% of its sales existed, whereas most orders came from a nonexistent firm. The founder of Comroad, Bodo Schnabel, explained to the judges that he had been sure that he would be able to substitute the “pre-drawn” sales volume with true sales volume. In November 2002, he was sentenced to seven years in jail and a payment of d20 million. The severe damage inflicted by him and his like on the German financial system was to some degree
316
HANS-PETER BURGHOF AND ADRIAN HUNGER
responsible for the rigor of this conviction that has no precedent in similar cases.28 It remains an open question why auditors and other control instruments were not able to protect the German financial system. Cases of fraud do not occur only at the Neuer Markt, and it is just a plausible assumption that there was more fraud in Neuer Markt firms than in firms that stayed private or are traded on other market segments. However, the high degree of fraud met a prick-eared public. Thousands of investors had lost money and were looking for causes, and scapegoats.29 Many of the firms at the Neuer Markt ran into financial distress. In situations of financial distress, outsiders are able to scrutinize the behavior of the (former) management much closer than would be possible if the management remained in complete charge. Finally, the new legislation on ad-hoc publicity seemed to open new ways to fight insider trading, that likewise is more apparent in the rather illiquid markets for Neuer Markt stocks than for blue chip companies where such activities might have occurred in the context of the numerous profit warnings from Spring 2000 onwards. Fraud at the Neuer Markt attracted much interest, was prosecuted with vigor and became very visible to the general public, at least by comparison with cases of fraud in non-Neuer Markt firms. Altogether, the huge public interest increased the reputation damage for the Neuer Markt. Thus, the value of the brand “Neuer Markt” became negative for both the firms listed at the Neuer Markt and the Deutsche B¨orse itself. The closure of this market segment was a logical consequence.
3.3. Causes and Consequences The Neuer Markt served a purpose that, at least from a theoretical perspective, creates new value. It enabled the founders and owners of young and innovative firms to raise equity at the public equity markets while protecting their relationship specific investments through dispersed outside equity.30 So why did it fail? The basic idea of the Neuer Markt was to allow young and innovative growth firms to go public through the guarantee of a high degree of transparency and strong regulatory background. However, as the stepwise tightening of rules in the course of the crisis demonstrates, the original set up contained some elements of compromise. Some of these, like too loose listing and delisting regimes, lacking (ex ante-) publicity for insider sales and insufficient penalties in the case of violations of the rules, might be responsible for some of the adverse incentives, although the general stock price bubble is the main culprit. However, introducing new rules ex post to close the gaps created an ambivalent signal. It could have demonstrated that the Deutsche B¨orse was actively protecting investors. It was
The Neuer Markt
317
understood as a signal for the defectiveness of the total system that was seemingly constructed to deceive investors. The Deutsche B¨orse intended to use the general push of the primary market at the end of the 1990s to firmly establish the new market segment and to compete with other new markets in Europe. To achieve this, a large number of firms had to be brought to the market in a very short period of time. Investment banks and other participants generated substantial fees, and equity was never cheaper for the owner of potential IPO candidates. Furthermore, some of the listing requirements such as the three years track record or the minimum free float of 25% were non-compulsory. As a consequence, among the completed IPOs there were many that can, with hindsight, be seen as premature. Other IPO candidates lacked a sustainable economic concept or a sufficient degree of transparency, which should have been a precondition for any outside investment. Some analysts even say that most of the firms that went public on the Neuer Markt were not really prepared for this step. However, a different economic development might have led to other conclusions. Like the destruction of investors’ trust after the burst of the bubble, some other potential causes for the failure of the Neuer Markt did likewise trouble the whole financial system, but hit the Neuer Markt particularly strong. Examples are the ineffectiveness of auditing firms in protecting the public against accounting frauds, or the irresponsible behavior of investment consultants and stock analysts and the conflicts of interest inside banks. However, other stock markets or market segments faced these problems with a reputation built over decades, or are, due to their size, vital to the economy at large. To some degree, the failure of the Neuer Markt was even helpful to the other stock market segments in Germany because it drew attention away from the difficulties they had. What are the consequences of this failure? Some straightforward observations are that, for the few IPOs that took place on the Neuer Markt or other market segments, underpricing seemed to have disappeared. Adding to this observation the low level of share prices, we can conclude that going public is no longer attractive for both the supply and the demand side. That does not mean that the concept in itself has lost its attractiveness. Many firms that had prepared to go public are now waiting for more favorable market conditions (see Wieselhuber & Partner, 2002). However, it is doubtful that such conditions will turn up very soon. The opinion of the general public has turned against stock markets, and the number of shareholders has decreased sharply. At least for the time being, the candidates for an IPO at the Neuer Markt (or, from 2003 onwards, an equivalent stock market segment) have lost valuable alternatives in financing. If the promoters of the Neuer Markt were right about the importance of the opportunity to go public for innovation and growth, the failure of the Neuer Markt should contribute
318
HANS-PETER BURGHOF AND ADRIAN HUNGER
to the lasting cloudy perspectives of the German economy. Even though we can only speculate about the quantitative impact, it is an urgent task to overcome the standstill.
4. GERMAN STOCK MARKETS AFTER THE NEUER MARKT – WHERE DO WE STAND? The decision of the Deutsche B¨orse to terminate the Neuer Markt was aided by a set of modifications of the existing laws to strengthen the competitiveness of the German financial system and to improve the functioning of the capital market. The so-called Viertes Finanzmarktf¨orderungsgesetz (Fourth Financial Markets Advancement Law) is in effect since July 1st, 2002. This law intends to improve investor protection, extend freedom of action for the market participants and prevent money laundering. One key element of the reform is the modification of the Stock Exchange Act, which offers the stock exchanges more flexibility with regard to the organization of stock trading so that they are able to react immediately to changes of the market situation.31 Thus, legal barriers that restricted the institutional design of the Neuer Markt are now lifted. Through the modifications of the Stock Exchange Act relating to the Viertes Finanzmarktf¨orderungsgesetz, the requirements for stock trading were altered. A “first segment” (official market) takes the place of the former Official Trading. In addition, in a “second segment” the requirements of the former Regulated Market are still in force as a minimum standard. Both segments meet the requirements as a regulated market in the sense of the Directive on Investment Services. Furthermore, the stock exchanges were allowed to demand further requirements for parts of these market segments. Thus, the Neuer Markt could have been easily adjusted to the new legal situation. However, the Deutsche B¨orse AG was looking for modifications that might help to regain investors’ trust, or at least to make it easier for investors to forget about past losses and scandals. Thus, ideas on both marketing and institutional design were needed. Obviously, the brand “Neuer Markt” had to be abandoned because it threatened to harm even sound firms traded in this market segment. However, it was necessary that investors should get the impression that the Deutsche B¨orse did more than just change the name to achieve a true rebranding. Thus, since January 1st, 2003, a totally new segmentation concept is in force, containing only two segments, called General Standard and Prime Standard from now on.32 Both Prime and General Standard represent regulated markets in the sense of the Investment Services Directive. As a third element the Unofficial Regulated
The Neuer Markt
319
Market (Freiverkehr) keeps on trading. However, this is not a regulated market in the sense of the Investment Services Directive and might remain what it was: an illiquid market that is not interesting for most companies. The listing requirements for an admission of shares to the “second segment,” i.e. the formerly Regulated Market, are raised to the requirements of the “first segment,” the former Official Trading, and were called General Standard. Moreover, an issuer can apply for a listing at the Prime Standard if the shares are admitted to the General Standard and the issuer meets additional requirements. Interestingly, these requirements are mainly drawn from the Neuer Markt and aim in particular at international investors, as do the English names of both segments. Prerequisite for the Prime Standard is an admission to the General Standard. Moreover, the issuer is obliged: (i) (ii) (iii) (iv) (v)
to prepare consolidated financial statements according to IAS or U.S.-GAAP, to publish quarterly reports containing certain specifications, to publish a corporate action timetable, to hold an analyst’s conference annually, and to publish ad-hoc announcements in German and English.
Already listed companies entered the General Standard automatically and have to apply for a listing at the Prime Standard. Altogether 390 firms entered the Prime Standard. This should have been particularly easy for the firms at the Neuer Markt that already complied with very similar rules. The Neuer Markt itself was terminated at December 31st, 2003. With the exception of the CDAX-index covering all traded stocks in the regulated markets, integration in one of the indices of the Frankfurt Stock Exchange will require an admission to the Prime Standard.33 These indices are the DAX containing the 30 largest German blue chips, the MDAX containing 50 smaller blue chips traded in Frankfurt and the SDAX describing the performance of the 50 medium sized firms below the MDAX. However, MDAX and the SDAX comprise less technology-oriented companies. The 30 largest high-tech companies are in integrated into the TecDAX. This index mainly contains NEMAX-50 firms and thereby links the new segmentation to the Neuer Markt. The NEMAX-50 will be calculated until the end of 2004 to guarantee continuity in derivatives trading. Furthermore, a Prime All-Share index and 18 industrial sector indices will be computed. All indices (with the exception of the DAX) contain German and non-German firms traded in Frankfurt. This new index family is calculated as from March 24th, 2003. The rebranding not only of the Neuer Markt but also of the total German (Frankfurt) stock market is a signal for the seriousness of the crisis. However, as the case of the Neuer Markt tells us, it is rather difficult and somewhat paradox
320
HANS-PETER BURGHOF AND ADRIAN HUNGER
to build up reputation through discrete measures in a short period of time. Thus, the new segmentation and rebranding could be rather costly, whereas gains can be expected only in the long run. Investors need time to learn from their own experience about the reliability and trustworthiness of the new market segments. Generally higher listing requirements and a strict enforcement of the stock exchange rules could help to speed up this process, although this will not win back the trust of a generation of private investors who often lost a fortune. Moreover, such strict rules cannot protect investors from idiosyncratic and general market risk and are thus no guarantee against a further or repeated loss of investors’ confidence and trust. Thus, while efficient stock markets might support economic growth, reciprocally the recovery of the German stock markets from their deep and structural crisis might need the help of a general recovery of the economy. Economic stagnation and the inability of the German government to reform crucial elements of the public financial system makes it rather improbable that such help is forthcoming. Thus, the debate is somehow back again where it was ten and twenty years ago: the experiment of the Neuer Markt might have enlarged our knowledge about what kind of stock markets Germany needs, but did, after the closing of a window of opportunity, not provide access to public equity markets for a wide range of firms looking for such access. Meanwhile, German stock markets lost a lot of time, their reputation, and the option to develop under favorable economic conditions. Although a circular development in itself is not a bad thing, one might get the disturbing feeling that this time the German financial system stumbled down the spiral staircase.
NOTES 1. Here, Regulated Market is the translation of its German name “Geregelter Markt,” without any inference with regard to regulated markets in the sense of the Investment Services Directive of the European Community. In fact, the Official Trading, the Regulated Market and the Neuer Markt are regulated markets in this sense. 2. For a critical perception of the German financial system, see Neuberger (2000) and Edwards and Fischer (1994), among many others. 3. However, despite all efforts of reform the number of bankruptcies tripled again in the next 20 years. Information on the development of the number of bankruptcies in Germany can be found at the Statistisches Bundesamt (Federal Statistics Bureau). 4. Council Directive 79/279/EEC of 5 March 1979 coordinating the conditions for the admission of securities to official stock exchange listing, the Council Directive 80/390/EEC of 17 March 1980 coordinating the requirements for the drawing up, scrutiny and distribution of the listing particulars to be published for the admission of securities to official stock exchange listing and the Council Directive 82/121/EEC of 15 February 1982 on information
The Neuer Markt
321
to be published on a regular basis by companies the shares of which have been admitted to official stock exchange listing. 5. This number looks still more impressive when taking into account that the NYSE is not the main market for IPOs in the United States. Most firms went public on NASDAQ. 6. The findings of Barry et al. (1990), Kortum and Lerner (2000) and Myers (2000) support this virtuous circle between innovation, venture capital and going public. 7. Digitale Telekabel AG went public on NASDAQ on December 31st 1996, iXOS Software AG started trading on October 7th, 1998 with a dual listing on NASDAQ and Neuer Markt as did QS Communications AG on April 19th, 2000. 8. We thank Bernd Rudolph for this comment. 9. See Council Directive 93/22/EEC of 10 May 1993 on investment services in the securities field. 10. The table in the Appendix compares the listing requirements of the Neuer Markt with the listing requirements of the other market segments. 11. The original Rules and Regulations Neuer Markt of 1997 did not require a special amount for the equity capital of the issuer. See Neuer Markt – Regelwerk, Deutsche B¨orse, February, 1997. 12. The original version of the Rules and Regulations Neuer Markt did not differentiate between ordinary and preferred allotments by computing the free-float. See Neuer Markt – Regelwerk, Deutsche B¨orse, February 1997. 13. See Part 2, No. 2.2 Rules and Regulations Neuer Markt. 14. See a leaflet of the Deutsche B¨orse on the foundation of the Neuer Markt in March 1997; Ideen suchen Kapital – Kapital sucht Ideen. Neuer Markt; Deutsche B¨orse AG, March 1997. 15. The admission to these indices depended on the market capitalization and equity turnover. See Leitfaden zu den Aktienindizes der Deutschen B¨orse, Deutsche B¨orse, Frankfurt, August 2001. 16. Leitfaden zu den Aktienindices der Deutsche B¨orse, Version 4.2. Deutsche B¨orse, Frankfurt, August 2001. 17. See Rock (1986) for a model of adverse selection and IPO pricing. 18. Hunger (2002) shows that the differences in first-day returns are statistically significant. 19. However, Goergen and Renneboog (2003) show that the disentanglement of the old major shareholders took a long time, at least in the German firms that went public during the years 1981–1988. 20. According to Bundesverband Deutscher Kapitalbeteiligungsgesellschaften (BVK). 21. Data refer to the period from March 14th, 1997 to March 10th, 2000 (March 24th, 2000 for the NASDAQ) and relates to the price indices of the NEMAX-All-Share index, the DAX-100 index and the NASDAQ-100 index respectively (computed at the end of a week). Source: Dresdner Bank AG. 22. The formal reason for the expulsion was that Gigabell did no longer provide the required quarterly reports. 23. Before this date, changes to other market segments or exits from the public equity markets were mainly due to takeovers. The exceptions were the closely related L¨osch Umweltschutz AG and Sero Entsorgung AG. These companies were accused of fraud. They were forced to transfer to the Regulated Market in April 1999. Although this case
322
HANS-PETER BURGHOF AND ADRIAN HUNGER
contained many of the elements of the crisis of many firms at the Neuer Markt about three years later, it received little attention at that time. 24. In November 2002, an intervention of the Federal Government prevented the firm from immediate bankruptcy. 25. However, the first to go to (pre-trial) confinement were, already in 1998, the brothers Johannes and Dieter L¨obbert (and two other managers) of L¨osch Umweltschutz AG and Sero Entsorgung AG. 26. Up to the end of September 2002, there were more than 30 insolvencies of companies listed at the Neuer Markt. See the online-issue of Die Aktiengesellschaft at http://www.dieaktiengesellschaft.de/news/news19.html. 27. In autumn 2003, the prosecutor was able to force a deal on one of the two accused managers of Infomatec AG. However, many observers thought that the penalty was a trifle. The second accused did not join the deal and is still on trial. Recent developments in the prevailing German case law enhance the chance that investors might recuperate some of their losses directly from the responsible managers if they bought the respective shares shortly after a manipulated ad hoc-announcement. 28. Note that even the earliest case of fraud at the Neuer Markt is not concluded by the courts today. 29. Among this general public, we might also find members of the law enforcement agencies and their families, according to hearsay. 30. See Burghof and Fischer (2002) for an overview of the reasons to go public. 31. REGIERUNGOnline: Fortentwicklung des Finanzplatzes Deutschland: Viertes Finanzmarktf¨orderungsgesetz, Stand June 2002. November 2002. 32. Deutsche B¨orse, Rundschreiben Listing 06/02; Neusegmentierung des Aktienmarktes. November 2002. 33. Deutsche B¨orse AG; What’s New – Deutsche B¨orse stellt neue Indexsystematik vor, Frankfurt, 31.10.2002.
REFERENCES Barry, C. B., Muscarella, C. J., Peavy, J. W., III, & Vetsuypens, M. R. (1990). The role of venture capital in the creation of public companies. Journal of Financial Economics, 27, 447–471. Barth, J., Nolle, D., & Rice, T. (1997). Commercial banking structure, regulation and performance: An international comparison. Working Paper, Comptroller of the Currency, Washington. Burghof, H.-P., & Fischer, C. (2002). Why do companies go public? Empirical evidence from Germany’s Neuer Markt. Working Paper, University of Munich. B¨uschgen, H. E. (1997). B¨orsenm¨aßiges Eigenkapital f¨ur kleine und mittlere Unternehmen. ¨ Osterreichisches Bankarchiv, 45, 94–104. BVK: Statistiken, http://www.bvk-ev.de/bvk.php/cat/67/title/Statistiken+Archiv. Claussen, C. P. (1984). Der Neue Zweite Markt. Zeitschrift f¨ur das gesamte Rechnungswesen, 1, 1–22. Deutsche B¨orse AG (1997, February). Regelwerk Neuer Markt. Frankfurt a. M. Deutsche B¨orse AG (1997, March). Ideen suchen Kapital – Kapital sucht Ideen. Frankfurt a. M. Deutsche B¨orse AG (2000, January). Historical statistics. Frankfurt a. M. Deutsche B¨orse, AG (2001, October). Rules and regulations Neuer Markt. Frankfurt a. M. Deutsche B¨orse AG (2001, August). Leitfaden zu den Aktienindices der Deutschen B¨orse. Version 4.2, Frankfurt a. M.
The Neuer Markt
323
Deutsche B¨orse AG (2001, September). Facts & figures Neuer Markt. Frankfurt a. M. Deutsche B¨orse AG (2002, November). Facts & figures Neuer Markt. Frankfurt a. M. Deutsche B¨orse AG. Fact books. Frankfurt a. M., several issues. Deutsche Bundesbank (1981). Ertragslage und Finanzierungsverh¨altnisse der Unternehmen im Jahre 1980. Monatsbericht, 11, 17–26. Deutsche Bundesbank (1982). Ertragslage und Finanzierungsverh¨altnisse der Unternehmen im Jahre 1981. Monatsbericht, 11, 14–25. Deutsche Bundesbank (2000, October). Der Markt f¨ur Wagniskapital in Deutschland. Monatsbericht. Deutsches Aktieninstitut e. V. 2002 (2002). Factbook 2002, Frankfurt a. M. Edwards, J., & Fischer, K. (1994). Banks, finance and investment in Germany. Cambridge: Cambridge University Press. European Community (1979). EU Directive 79/279/EWG. European Community (1980). EU Directive 80/390/EWG. European Community (1982). EU Directive 82/121/EWG. European Community (1993). EU Directive 93/22/EWG. FIBV (2002). Market capitalization of shares of domestic companies. December 2002. Fischer, C. (2002). Motive des B¨orsengangs am Neuen Markt. Berlin: Duncker & Humblot. Francioni, R., & Gutschlag, T. (1998). Der Neue Markt. In: G. Volk (Ed.), Going Public (pp. 27–41). Stuttgart. Goergen, M., & Renneboog, L. (2003). Why are the levels of control (so) different in German and UK companies? Evidence from initial public offerings. Journal of Law, Economics and Organization, 19, 141–175. Hopt, K. J., Rudolph, B., & Baum, H. (1997). Dokumentation – Empfehlungen zur B¨orsenreform in Deutschland. WM – Wertpapiermitteilungen/Zeitschrift f¨ur Wirtschafts- und Bankrecht 34, 1637–1640. Hunger, A. (2002). Market segmentation and IPO-underpricing: The German experience. Working Paper, University of Munich. Kempkes, M., & Haffa, F. (1998). Der Neue Markt als Chance f¨ur junge Wachstumsunternehmen. In: G. Volk (Ed.), Going Public (pp. 177–190). Stuttgart. Kersting, O. (1997). Der Neue Markt der Deutsche B¨orse AG. Die Aktiengesellschaft, 5, 222–228. Kortum, S., & Lerner, J. (2000). Assessing the contribution of venture capital to innovation. Rand Journal of Economics, 31, 674–692. Kraus, T., & Burghof, H. P. (2003). Post-IPO performance and the exit of venture capitalists. Working Paper, Munich Business Research 2003–01. Krause, H. (2001). German securities regulation. M¨unchen: Verlag C. H. Beck. Ljungqvist, A. P. (1997). Pricing initial public offerings: Further evidence from Germany. European Economic Review, 41, 1309–1320. Myers, S. C. (2000). Outside equity. Journal of Finance, 55, 1005–1037. Neuberger, D. (2000). Evolution of financial systems: Convergence towards higher or lower stability? In: A. Karmann (Ed.), Financial Structure and Stability (pp. 11–33). Berlin: Springer/Physica. Neumann, M. (1997). Wagniskapital – Gutachten des Wissenschaftlichen Beirats beim Bundesministerium f¨ur Wirtschaft. In: Bundesministerium f¨ur Wirtschaft (Ed.), BMWI-Studienreihe (Nr. 95, pp. 1–27). Bonn. Potthoff, V., & Stuhlfauth, J. (1997). Der Neue Markt: Ein Handelssegment f¨ur innovative und ¨ wachstumsorientierte Unternehmen – kapitalmarktrechtliche Uberlegung und Darstellung des Regelwerkes. WM-Wertpapiermitteilungen/Zeitschrift f¨ur Wirtschafts- und Bankrecht, 26, Sonderbeilage Nr. 3/1997, 1–24.
324
HANS-PETER BURGHOF AND ADRIAN HUNGER
Rock, K. (1986). Why new issues are underpriced. Journal of Financial Economics, 15, 187–212. von Rosen, R. (1995). Going public: Defizite und Perspektiven. Zeitschrift f¨ur das gesamte Kreditwesen, 48, 374–385. Schrader, T. (1993). Geregelter Markt und geregelter Freiverkehr – Auswirkungen gesetzgeberischer Eingriffe. Wiesbaden: Deutscher Universit¨ats-Verlag. Theissen, E. (1998). Der Neue Markt: Eine Bestandsaufnahme. Zeitschrift f¨ur Wirtschafts- und Sozialwissenschaften (ZWS), 118, 623–652. Titzrath, A. (1995). Die Bedeutung des going publics. Zeitschrift f¨ur Betriebswirtschaft, 54, 133–155. Wieselhuber & Partner (2002). IPO Klima 2002. mimeo, Munich.
Market
(i) Unofficial Regulated Market (Freiverkehr)
(ii) Regulated Market (Geregelter Markt)
(iii) Official Trading (Amtlicher Handel)
(iv) Neuer Markt
Admission
Prospectus in accordance with securities prospectus regulation (Verkaufsprospektverordnung)
(i) Plus additional documents for the admission committee (Zulassungsausschuss) (e.g. statements concerning patents and legal disputes, annual financial statements, management reports for three business years, . . .)
(ii) Plus additional information included in the prospectus (e.g. source and application of funds, affiliated enterprises, profits, losses and dividends per share, consolidated financial accounts, . . .)
(iii) Plus information about risk factors included in the prospectus Only common shares Application for a listing in the regulated market and waiver in favor of a listing at the Neuer Markt
d250,000
d1,250,000
d1,500,000
Minimum capital
equity
–
The Neuer Markt
APPENDIX: LISTING REQUIREMENTS OF MARKET SEGMENTS OF THE FRANKFURT STOCK EXCHANGE DURING THE EXISTENCE OF THE NEUER MARKT (AS OF DECEMBER 2002)
325
326
APPENDIX (Continued ) –
Equal treatment of all securities holders
(ii)
(ii) Plus regular analyst meeting, at least once a year, six month lock-up period for incumbent owners
Regular disclosure after going public
–
Publication of the annual report/annual business reports
(ii) But more detailed business plan
Publication of quarterly and annual reports in German and English language, annual reports in accordance with IAS or U.S.-GAAP Publication of convocation details, of an annual corporate action timetable and any additional information Notification of any share transaction of the issuer or the management board
Publication of information about the general meeting, balance sheet and modification of statutes and securities
HANS-PETER BURGHOF AND ADRIAN HUNGER
Rules of conduct after going public
–
–
3 years
3 years (non-compulsory)
Minimum free float
–
–
25% of the aggregate nominal value of shares
In general 25% of the aggregate nominal value of shares
Additional quantitative requirements for the share issue
–
d500,000 minimum nominal value of issue
d2,500,000 minimum aggregated market price of issue
d5,000,000 minimum aggregated market price of issue, d250,000 minimum nominal value of issue, minimum 100.000 shares, 50% of all shares sold must be new shares
The Neuer Markt
Minimum age
Source: Deutsche B¨orse (2001, October).
327
THE LONG-TERM PERFORMANCE OF INITIAL PUBLIC OFFERINGS ON EUROPE’S NEW STOCK MARKETS Giancarlo Giudici and Peter Roosenboom ABSTRACT In this chapter we examine the determinants of the long-run stock price performance of Initial Public Offerings (IPOs) on Europe’s new stock markets. We report that the average company that went public on these markets has been a very poor long-term investment. We find that the stock price performance during a three-year window is inversely related to first-day returns. We also find that the long-term underperformance of IPO firms begins after the lock-up agreement has expired and insiders start trading in the firm’s shares. These findings are consistent with the divergence of opinion hypothesis of Miller (1977).
1. INTRODUCTION In 2003 the Neuer Markt, at one time Europe’s largest market for growth stocks, closed its doors amid scandals and corporate governance failures.1 Many Neuer Markt companies had seen their stock prices plummet after issuing profit warnings or missing their earnings targets. The stocks that once traded like gold nuggets at the height of the stock market bubble were suddenly sinking like lead (The Independent, 15 November 2000). The series of insolvencies and insider trading The Rise and Fall of Europe’s New Stock Markets Advances in Financial Economics, Volume 10, 329–354 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 1569-3732/doi:10.1016/S1569-3732(04)10012-1
329
330
GIANCARLO GIUDICI AND PETER ROOSENBOOM
scandals that followed, tarnished the Neuer Markt’s reputation and added to the decision to shut the market down. However, this collapse of stock prices was not at all limited to Neuer Markt companies. Many firms that went public on the French Nouveau March´e, the Italian Nuovo Mercato, the Dutch Nieuwe Markt, Euro.NM Belgium and NASDAQ Europe lost nearly all their value. To date, there has been no in-depth study investigating the long-run stock price performance of Initial Public Offerings (IPOs) on these markets. This chapter aims to fill this gap. What explains this poor long-run stock price performance of the average IPO firm? One plausible explanation is offered by the divergence of opinion hypothesis of Miller (1977). Given the inherent lack of price history and limited accounting information, the opinion about the firm’s prospects might range from “this is the next Microsoft” to “this dog will not see its second birthday” (Houge et al., 2001). Miller (1977) argues that in a market with restricted short selling, such as the IPO market, market prices might exceed fundamental values because they are determined by (a minority of) overoptimistic investors who want to believe that the company is “the next Microsoft”. Over time, information flows increase and short sale constraints are relaxed allowing more pessimistic investors to enter the market. As a result, the divergence of opinion narrows and prices will gradually converge to fundamental value. Despite its intuitive appeal, there is currently only one paper that directly tests the divergence of opinion hypothesis. Houge et al. (2001) show that a large opening bid-ask spread, a late opening trade and a high flipping ratio (their proxies for divergence of opinion) are associated with poor long-term performance of U.S. IPO firms. In this chapter we test the divergence of opinion hypothesis. We expect that this hypothesis may partially explain the poor performance of IPO firms on Europe’s new stock markets. New issues on these markets were often accompanied by advertising campaigns, including mainstream television, making subjective claims going far beyond the content of the IPO prospectus. In addition, many press officers of companies briefed the press without worrying too much if their claims would prove true in the future (Financial Times, 15 January 2001) and stock analysts and TV stock market pundits praised dot.coms to high heaven before prices came crashing down to earth. This is likely to have fuelled overoptimism leading to market prices that initially exceeded fundamental values. Our results provide some support for the divergence of opinion hypothesis of Miller (1977). In particular, we find that IPO underpricing is negatively related to long-run stock price performance. This suggests that investor overoptimism on the first trading day has a transitory effect on prices. We also find that the long-term underperformance of IPO firms begins after the lock-up agreement has expired and insiders start trading in the firm’s shares. It is likely that these
The Long-Term Performance of Initial Public Offerings
331
insiders have a different information set than outside investors and that this additional information gets incorporated into stock prices once they are allowed to trade in their firm’s shares. Additionally, we find that Internet and technology companies that went public during the bubble period of 1999–2000 performed very poorly. This captures the “Tech Wreck” that occurred after the stock market bubble burst. We report that IPOs completed in 1999–2000 that were underwritten by reputable investment banks performed significantly better than other IPOs. This chapter continues as follows. In section two we discuss the existing literature. Section three describes the data. Section four discusses the measurement of long-term returns. Section five presents our results. Section six concludes.
2. LITERATURE REVIEW Long-run underperformance is a well-documented anomaly associated with IPOs. For the United States, Ritter (1991) shows that firms going public during 1975–1984, on average, underperform a sample of matching firms over a three-year period by 29%. Loughran and Ritter (1995) test the robustness of this finding and confirm that U.S. IPOs during 1970–1990 have been poor long-term investments for investors. For the United Kingdom, Levis (1993) shows that companies that went public during 1980–1988 underperform market indices by an average of 8–23% (depending on the market benchmark used) for a period of three years after their IPO. Espenlaub et al. (2000) report that UK companies that went public during 1985–1992 show substantial negative abnormal returns after the first three years irrespective of the benchmark used. For Germany, Ljungqvist (1997) reports that companies going public during 1970–1990, on average, underperform the market by 12% in three years. However, Stehle et al. (2000) show that German firms that went public during 1960–1992, on average, underperform a portfolio of stocks with a similar market capitalization by 6% in three years. This prompts the question why the average IPO firm performs so poorly over the long run. Miller (1977) develops a theory to explain this anomaly. Miller (1977, p. 1156) writes: “the prices of new issues are set not by the appraisal of the typical investor, but by the small minority who think highly enough of the investment merits of the new issue to include it in their investment portfolio.” It is this minority of overoptimistic investors that initially sets market prices above fundamental values in a market with restricted short selling. Over time, additional information on the firm becomes available and short sale constraints are loosened. This allows more pessimistic investors to enter the market resulting in a stock price decline.
332
GIANCARLO GIUDICI AND PETER ROOSENBOOM
It is difficult to test the divergence of opinion hypothesis directly because it is difficult to measure the divergence of opinion among investors. However, Houge et al. (2001) develop three opening-day proxies for divergence opinion: opening bid-ask spread, the time of first trade and the flipping ratio (defined as sell-signed block volume divided by total volume on the first trading day). They find that a large opening spread, a late opening trade and a high flipping ratio have significant predictive ability to explain long-run IPO returns in the United States. Miller’s argument relies in part on the presence of short sales restrictions in IPO markets. Gezcy et al. (2002) find that although investors with good access to the equity lending market can short most IPOs, investors without access to specials (i.e. expensive-to-borrow stocks) can short none of them. Ofek and Richardson (2003) argue that lockup agreements represent the most stringent form of short sale constraint because it prevents insiders from trading in the firm’s stock. The expiration of the lockup agreement can be thought of as relaxing this short sale constraint. Accordingly, Ofek and Richardson (2003) report that the daily abnormal returns after lock-up expiration are significantly lower than the daily abnormal returns before lock-up expiration. This finding is consistent with the model of Aggarwal et al. (2002) where managers strategically underprice IPOs to maximize wealth by selling stocks at lock-up expiration. In the model, underpricing creates positive momentum that arguably lasts until lock-up expiration. A number of other studies are broadly consistent with Miller’s divergence of opinion hypothesis. Rajan and Servaes (1997) show that more firms go public at times when analysts are overoptimistic. They report that U.S. firms perform poorly over the first four years after their IPO when analysts are more optimistic about their long-term growth potential. Jaggia and Thosar (2004) examine the 6-month aftermarket performance of high-tech U.S. IPOs completed in the late 1990s. They find a pattern of short-run positive momentum followed by a gradual reversal in post-IPO returns. Dorn (2003) examines a unique dataset of clients of a large German retail broker. He reports that retail investors were willing to overpay for IPOs during August 1999 to May 2000. New issues that were more aggressively sought after by retail investors had higher first-day returns and lower aftermarket returns during the first year after the IPO. Dorn (2003) concludes that sentiment drives retail purchases of IPO stocks and has a transitory effect on prices. Loughran and Marietta-Westberg (2001) report strong underperformance of U.S. IPO firms following either a positive or negative extreme one-day price movement of at least ±15%. They argue that investors appear optimistic about a variety of news announcements that coincide with one-day extreme returns. Investors seem to overreact to good news events and underreact to bad news events. Jain and Kini (1994) examine the long-term operating performance of U.S. IPOs. They report that the cash flow-to-assets ratio decreases substantially
The Long-Term Performance of Initial Public Offerings
333
between the year prior to going public and three years later. Mikkelson et al. (1997) show that cash flows of U.S. IPO firms did not increase sufficiently to justify high valuations at the time of the IPO. Teoh et al. (1998) show that U.S. IPO firms attempt to mislead investors about the operating performance of the company through earnings management. They show that IPO firms that manipulate their earnings in the first fiscal year as a public company experience poor stock price performance over the next three years. Other studies have examined the relation between long-term IPO returns and underwriter reputation. Carter et al. (1998) report a negative relationship between long-run returns and underwriter reputation. In particular, they find that the average IPO underwritten by a reputable investment bank underperforms the market by 13% whereas the average IPO underwritten by a non-reputable investment bank underperforms the market by 34%. Conversely, Logue et al. (2002) do not find a relationship between underwriter reputation and long-term returns. However, they find that aftermarket underwriter activities such as the decision to exercise the overallotment option and price stabilization, are positively related to longer-run returns. Brav and Gompers (1997) report that U.S. IPO firms that were backed by venture capitalists outperformed IPO firms that did not receive venture backing. However, they conclude that this underperformance is not an IPO effect. Companies with similar size and book-to-market ratios but that did not issue equity performed just as poorly as the IPO firms in their sample did.
3. DATA AND SAMPLE DESCRIPTION We identify admissions to trading from the SDC Global New Issues database and information provided by the stock exchanges. In constructing our sample, we exclude financial companies (SIC codes 6000–6999), spin-offs, privatisation issues and companies previously listed elsewhere. Our final sample includes 555 nonfinancial companies that went public on the German Neuer Markt (303 firms), the French Nouveau March´e (144), the Italian Nuovo Mercato (35), the Dutch Nieuwe Markt (14), Euro.NM Belgium (13) and NASDAQ Europe (46) during 1996–2000. We obtain prospectuses from the company or Disclosure Global Access. Table 1 shows descriptive statistics. The average (median) company has a market value of d291.4 million (d129 million) on the first day of trading. There is a small number of large companies that go public during this period. The largest company in our sample is T-Online International AG with a market value of d13.2 billion. There are 25 companies that have a market value higher than d1 billion. The market-to-book ratio is defined as the ratio of market value and the post-issue book value of equity. The average (median) company has a market-to-book value of 6.02
334
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Table 1. Descriptive Statistics.
Market value (d thousands) Market-to-book ratio Firm age (years) EBITDA < 0 dummy Internet and technology dummy Dilution (%) Participation (%) Underwriter market share VC backing dummy First-day return (%) High-low spread (%) Bid-ask spread (%) Volume ratio Wealth relative (local index) Wealth relative (NASDAQ index)
Mean
Median
Maximum
Minimum
Std. Dev.
Obs.
291,445 6.02 12.25 0.305 0.605
129,319 4.27 9.00 0.00 1.00
13,252,500 115.07 130.00 1.00 1.00
9,054 0.01 1.00 0.00 0.00
833,118 7.91 12.88 0.461 0.489
555 555 555 555 555
32.55 7.58 5.83 0.456 35.56 13.06 1.89 0.718 0.678 0.722
30.00 5.05 3.11 0.00 10.25 11.11 1.54 0.349 0.217 0.261
167.00 63.67 46.01 1.00 433.33 85.71 10.15 20.82 44.08 51.78
0.00 0.00 0.03 0.00 −25.00 −10.16 0.00 0.00 0.001 0.001
16.70 9.35 7.68 0.498 61.84 10.73 1.62 1.39 2.54 2.72
555 555 555 555 555 465 465 465 555 555
Note: This table shows descriptive statistics. We compute market value as the number of shares outstanding after the IPO times the closing market price on the first trading day. The market-to-book ratio is measured as the ratio of market value and the post-issue book value of equity. Firm age is defined as the calendar year of the IPO minus the calendar year of founding as mentioned in the prospectus. EBITDA < 0 is a dummy variable that equals one if earnings before interest, taxes, depreciation and amortization is less than zero in the most recent financial year disclosed in the prospectus. Internet and technology is a dummy variable that takes on the value one if the IPO firm is active in the Internet and technology sector. We identify Internet and technology firms as described in Note 2. The dilution factor is defined as the number of newly issued shares at the IPO divided by the number of pre-IPO shares outstanding. The participation ratio is defined as the number of existing shares sold by pre-IPO shareholders divided by the number of pre-IPO shares. Underwriter market share is the sum of gross proceeds (excluding over-allotment option) in all local IPOs lead managed by bank j divided by the total proceeds raised in the local market during the 1990–2002. VC backing dummy is a dummy variable if one or more venture capitalists are pre-IPO shareholders. First-day return is measured as: (first-day closing market price final offer price)/final offer price. The high-low spread is defined as the difference between the highest and lowest price achieved on the first trading day divided by the average of the highest and lowest price on that day. Bid-ask spread equals the difference between the ask and bid price at the close of market on the first trading day divided by the average of the ask and bid price at the close of market on that day. Volume ratio is computed as the ratio of the number of shares traded on the first trading day and the number of shares sold to the public in the IPO. The wealth relative is defined as the ratio of the three-year holding period return on the IPO stock and the three-year holding period return on the local market index or NASDAQ composite index.
The Long-Term Performance of Initial Public Offerings
335
(4.27). There is only one company (Netlife AG) that has a negative post-issue book value of equity. In this case we set the market-to-book ratio equal to 0.01. Age is measured as the difference between the calendar year of the IPO and the calendar year of founding. We find that the average (median) firm age is 12 years (9 years). Table 1 shows that 30.5% of the companies that go public report losses and 60.5% of companies are active in the Internet and technology sector.2 Dilution is defined as the number of newly issued shares sold at the time of the IPO divided by the number of pre-IPO shares outstanding. The company raises a significant amount of new equity capital at the time of the IPO. The average (median) company has a dilution factor of 32.6% (30%). The participation ratio is measured as the ratio of the number of existing shares sold by pre-IPO owners at the time of the IPO and the number of pre-IPO shares outstanding. The participation ratio has an average value of 7.6% and a median value of 5.1%. Underwriter market share is calculated as the sum of gross proceeds of all IPOs lead managed by the underwriter divided by the sum of gross proceeds of all IPOs in the local market during 1990–2002.3 This adopts the approach of Ljungqvist and Wilhelm (2002). The average (median) market share is 5.8% (3.1%). We use underwriter market share as our measure of underwriter reputation in subsequent tests. We find that 45.6% of sample firms are backed by a venture capitalist. Venture capitalists and underwriters are repeated players in the IPO market that may certify the quality of the firm going public (Carter et al., 1998; Megginson & Weiss, 1991). First-day returns are measured as the percentage difference between the offer price and the first-day closing market price. Table 1 shows that first-day returns average 35.6% with a median of 10.3%. There are 69 companies (12.4% of our sample) that more than double in price on the first day of trade. This shows that first-day returns are highly skewed. First-day returns range from a maximum of 433.33% (Biodata Information Technology AG) to a minimum of −25% (Neue Sentimental Film AG). We use first-day returns as one of our proxies for divergence of opinion. We argue that higher first-day returns point to a higher degree of investor overoptimism. We follow Houge et al. (2001) and define three other proxies for the divergence of opinion: the high-low spread, the bid-ask spread and the volume ratio. We obtain information about the highest and lowest intraday price, bid and ask price at the close of the first-trading day and the number of shares traded on the first trading day from Datastream. We are able to collect this data for 465 sample firms. The high-low spread is defined as the difference between the highest and lowest price achieved on the first-trading day divided by the average of the highest and lowest price on that day. The average high-low spread equals 13.1% and is 11.1% evaluated at the median. We conjecture that the higher the high-low spread on the first trading day, the higher the divergence of opinion among investors.
336
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Table 2. Comparison of the Pre-Bubble and Bubble Period.
Market value (d thousands) Market-to-book ratio Firm age (years) EBITDA < 0 dummy Internet and technology dummy Dilution (%) Participation (%) Underwriter market share VC backing dummy First-day return (%) High-low spread (%) Bid-ask spread (%) Volume ratio Wealth ratio (local market index) Wealth ratio (NASDAQ index)
Pre-Bubble Period 1996–1998
Bubble Period 1999–2000
Test for Difference
130,010 63,148 7.02 4.34 14.25 11.00 0.201 0.470 33.09 30.83 10.47 7.43 6.33 3.66 0.469 29.93 10.00 7.51 4.18 2.00 1.55 0.744 0.306 1.401 0.460 1.289 0.390
359,156 166,400 5.59 4.22 11.42 9.00 0.348 0.662 32.33 26.85 6.37 4.19 5.62 2.67 0.450 37.92 10.34 14.73 13.09 1.85 1.52 0.701 0.358 0.374 0.175 0.484 0.220
2.97*** 8.58*** 1.94* 0.39 2.37** 2.33** 3.46*** 4.30*** 0.49 2.54** 4.81*** 3.05*** 1.00 1.14 0.42 1.39 0.24 6.29*** 7.54*** 0.86 0.73 0.23 1.02 4.42*** 6.89*** 3.21*** 4.81***
Note: This table compares the pre-bubble period of 1996–1998 and the bubble period of 1999–2000. The first column presents means and medians for 164 initial public offerings during the prebubble period and the second column provides means and medians for 391 initial public offerings during the bubble period. We test whether differences exist between the two periods. We use a standard t-test for difference in means and the Wilcoxon/Mann-Whitney test for difference in medians. Medians are shown in italics. We have data on the high-low spread, bid-ask spread and volume ratio for 107 initial public offerings during the pre-bubble period and 358 initial public offerings during the bubble period. ∗ Significant at the 10% level. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
The Long-Term Performance of Initial Public Offerings
337
We measure percentage bid-ask spreads as the difference between the ask and bid price at the close of the first trading day divided by the average of the ask and bid price. The average (median) percentage bid-ask spread equals 1.9% (1.5%). We argue that market makers quote wider spreads when they face more uncertainty such as divergence of opinion. The volume ratio is calculated as the ratio of the number of shares traded on the first trading day and the number of shares sold to the public at the time of the IPO. The average (median) volume ratio equals 0.72 (0.35). This shows that a substantial fraction of the shares sold to the public is traded on the first trading day. We argue that a high level of trade on the first day indicates divergence of opinion. Next, we investigate whether descriptive statistics differ between the group of 164 companies that went public during the pre-bubble period (1996–1998) and the group of 391 companies that went public during the bubble period (1999–2000). Ljungqvist and Wilhelm (2003) show that the average U.S. company that went public during the bubble period has a higher first-day return, is more likely to report a loss in the year before the IPO, is more likely to be from the Internet and technology sector and sells less existing shares at the time of the IPO than the average company that went public during 1996–1998. We report similar findings. Table 2 shows that companies going public during 1999–2000 have a significantly higher market value on the first trading day and are significantly younger than companies going public during 1996–1998. Additionally, we find that companies that went public during the bubble period are more likely to report a loss during the financial year before going public and are more likely to be from the Internet and technology sector. They also have a significantly lower participation ratio (i.e. sell less existing shares at the time of the IPO) and have a higher high-low spread than companies that went public during the pre-bubble period. We report on these two subperiods throughout this chapter.
4. MEASURING LONG-RUN STOCK PRICE PERFORMANCE We calculate cumulative abnormal returns (CARs) with monthly portfolio rebalancing (Levis, 1993; Loughran & Ritter, 1995; Ritter, 1991).4 Post-IPO returns are measured for an aftermarket period of 36 months where a month is defined as a 21-trading-day interval. We exclude the first 21 trading days after the IPO date to avoid a potential bias from the price stabilization of underwriters during that period. For IPO firms that are delisted before the 36 month holding period, the aftermarket period is truncated, ending with the delisting date.5 We
338
GIANCARLO GIUDICI AND PETER ROOSENBOOM
use the corresponding return on two different benchmarks: the local Datastream market index and the NASDAQ Composite index. We measure the abnormal return for stock i in month t as (assuming beta equals one): ARi,t = r i,t − r b,t
(1)
where ri,t is the raw return on firm i in month t and rb,t is the raw return on the particular benchmark over the same period. The average abnormal return on a portfolio of n stocks for month t is the equally-weighted average of the abnormal returns, defined as: 1 ARi,t n n
ARt =
(2)
i=1
The cumulative abnormal returns (CARs) from the beginning of the first month of trading to month s is calculated as: CAR1,s =
s ARt
(3)
t=1
Fig. 1. Cumulative Abnormal Returns Using the Local Market Index as the Benchmark. Note: This figure shows cumulative abnormal returns for an equally weighted portfolio of 555 initial public offerings during the full period of 1996–2000, 164 initial public offerings during the pre-bubble period of 1996–1998 and 391 initial public offerings during the bubble period of 1999–2000. We use the local market index as our benchmark to compute abnormal returns.
The Long-Term Performance of Initial Public Offerings
339
Figure 1 plots CARs over the 36 month period using the local market index as the benchmark. We find that our sample firms underperform the local market index by 30% in three years. However, companies that went public in the pre-bubble period 1996–1998 outperform the market by 10% in three years. This finding is due to a small number of IPOs that outperform the market. Companies that went public during the bubble period of 1999 and 2000 perform much worse. They underperform the market portfolio by more than 50% in three years. Figure 2 plots CARs using the NASDAQ Composite index as the benchmark. We find similar results. The use of two different benchmarks makes it harder to dismiss our long-horizon tests as consequence of test misspecifications. Barber and Lyon (1997) show that test statistics based on abnormal returns may be misspecified because of new listing, rebalancing, and skewness bias.6 To correct for these biases, they advocate computing a benchmark portfolio by matching the sample firms to control firms of similar sizes and book-to-market ratios. However, we cannot compute returns on portfolios of control firms because there are too few control firms that belong to the same size and book-to-market class as our sample firms. We attempt to address this problem by adding firm size and market-to-book ratios as independent
Fig. 2. Cumulative Abnormal Returns Using the NASDAQ Composite Index as the Benchmark. Note: This figure shows cumulative abnormal returns for an equally weighted portfolio of 555 initial public offerings during the full period of 1996–2000, 164 initial public offerings during the pre-bubble period of 1996–1998 and 391 initial public offerings during the bubble period of 1999–2000. We use the NASDAQ composite index as our benchmark to compute abnormal returns.
340
GIANCARLO GIUDICI AND PETER ROOSENBOOM
variables in our regressions and using the NASDAQ Composite index as our alternative benchmark. The NASDAQ Composite index captures the performance of Internet and technology stocks in the United States (that are not part of our sample). Next, we calculate wealth relatives (Ritter, 1991). Wealth relatives are the ratio of the 36 month buy-and-hold returns on the IPO stock and the buy-and-hold returns of the market index (local market index or NASDAQ Composite index) during the same period. min(T,delist) (1 + r i,t ) t=0 (4) WRi = min(T,delist) (1 + r ) b,t t=0 where min [T,delist] is the earlier of its delisting date or month 36, ri,t is the raw return on firm i in month t and rb,t is the raw return on the particular benchmark over the same period. Wealth relatives indicate how much money the investor is left with after 36 months compared to one euro invested in the market portfolio. We will focus on wealth relatives throughout this chapter. The last two rows of Table 1 show that the mean (median) wealth relative equals 0.68 (0.22) using the local market index as the benchmark and 0.72 (0.26) using the NASDAQ Composite index as the benchmark. This indicates that an investor would be left with an average of only 68 cents (72 cents) compared to one euro invested in the local market index (NASDAQ Composite index). Note that the median wealth relative is much lower than the average wealth relative (i.e. wealth relatives are highly skewed). For example, there are 167 companies (30% of our sample) that have a wealth relative less than 0.1. We therefore use the natural logarithm of the wealth relative in our regressions. Table 2 shows that the companies that went public during 1996–1998 outperform companies that went public during the bubble period of 1999–2000. However, the median company that went public during 1996–1998 underperforms the market portfolio as reflected in a median wealth relative below one.
5. EMPIRICAL RESULTS 5.1. Cross-Sectional Determinants of Three-Year Wealth Relatives We regress the natural logarithm of the three-year wealth relative on firm and offer characteristics, certification variables (underwriter reputation and venture backing dummy) and proxies for divergence opinion (first-day returns, high-low spread,
The Long-Term Performance of Initial Public Offerings
341
Table 3. Definition of Variables. Variable Name Divergence of opinion variables First-day return High-low spread
Bid-ask spread
Volume ratio
Firm and offer characteristics Log (market value)
Log (market-to-book ratio)
Log(1 + age)
EBITDA < 0 dummy Internet and technology dummy Dilution factor Participation ratio Certification variables Underwriter market share
VC backing dummy
Definition
The difference between the first-day closing market price and the final offer price divided by the final offer price. The difference between the highest and lowest price achieved on the first trading day divided by the average of the highest and lowest price on that day. Data is taken from Datastream. The difference between the ask and bid price at the close of market on the first trading day divided by the average of the ask and bid price at the close of market on that day. Data is taken from Datastream. The ratio of the number of shares traded on the first trading day and the number of shares sold to the public at the IPO. Data is taken from Datastream. Natural log of market value (in d thousands). Market value is computed as the number of shares outstanding after IPO times the closing market price on the first trading day. Natural log of the market-to-book ratio at issue. The market ratio is computed as the ratio of the first-day market value of equity and the post-issue book value of equity. The post-issue book value of equity equals the sum of the primary offering proceeds (i.e. the number of newly issued shares times the offer price) and the book value of equity from the last pre-IPO financial statement or when available from a later interim statement as disclosed in the prospectus. Natural log one plus firm age, where firm age is measured as calendar year of the IPO minus the calendar year of founding as mentioned in the prospectus. =1 if the IPO firm reports negative EBITDA in the fiscal year before going public; =0 in other cases. =1 if the IPO firm is classified as “Internet” or “technology” stock; = 0 in other cases. Number of newly issued shares at the IPO / number of pre-IPO shares outstanding. Number of existing shares sold by pre-IPO shareholders divided by the number of pre-IPO shares. percentage market share of the IPO lead manager in the local financial market (measured by gross proceeds raised during 1990–2002 and including main market segments). = 1 if one or more venture capitalists are pre-IPO shareholders of the company; = 0 in other cases.
342
GIANCARLO GIUDICI AND PETER ROOSENBOOM
bid-ask spread and volume ratio). Table 3 gives an overview of the independent variables that we include in our regressions. We first estimate regressions using the first-day return as our only proxy for divergence of opinion. The first column of Table 4 shows the results for the full period 1996–2000 when we use the local market index as the benchmark to compute the wealth relative. We find that first-day returns are negatively related to the three-year wealth relative. Companies that experience a high return on the first trading day perform worse in the long-run. A one standard deviation increase in first-day return lowers the three-year wealth ratio from the sample average of 0.68 to 0.46, other things equal. In unreported tests, we also include a dummy variable that takes on the value one if the company more than doubles in price on the first trading day. We find that this dummy variable has a coefficient of −0.84 (t-value = 3.12). These findings suggest that first-day returns have a transitory effect on market prices. We interpret this finding as consistent with Miller’s divergence of opinion hypothesis. We infer that high first-day returns are due to overoptimistic investors that set market prices above fundamental values. Over time more pessimistic investors enter the market and prices gradually converge to fundamental values. We also include firm and offer characteristics in the regression model. We find an inverse relation between market capitalization on the first trading day and the three-year wealth relative. However, we do not find any relationship between market-to-book ratios and long-term performance. Our results show that threeyear wealth relatives are a positive function of firm age and a negative function of the dummy that indicates whether the company reports a loss in the year before its IPO. This suggests that older and profitable companies perform better in the long-run. Internet and technology companies underperform companies from other industries. This reflects the steep decline of stock prices of high-technology companies around the world (the so-called “Tech Wreck”). We also include the dilution factor and participation ratio in our regression model. However, we do not find any significant relation between these two variables and the three-year wealth relative. We also include two certification variables. We find that the three-year wealth relative is a positive function of underwriter market share. A one standard deviation increase in underwriter market share increases the three-year wealth relative from its sample average of 0.68 to 0.78, other things equal. New issues that are underwritten by more prestigious underwriters thus perform better in the long-run. This finding is consistent with the U.S. results reported by Carter et al. (1998). We do not find a relation between the wealth relative and the venture capital backing dummy.
Full Period 1996–2000 Divergence of opinion variables First-day return
Pre-Bubble Period 1996–1998
Bubble Period 1999–2000
−0.636 (−3.67)***
−0.206 (−0.58)
−0.688 (−3.77)***
−0.398 (−1.93)* −1.624 (−1.73)* −0.958 (−0.72) 0.038 (1.11)
0.091 (0.25) −1.140 (−0.56) −6.138 (−0.82) 0.083 (0.97)
−0.450 (−1.95)** −0.979 (−0.60) −1.356 (−1.00) 0.016 (0.45)
−0.214 (−2.88)*** 0.024 (0.10) 0.157 (1.66)* −0.416 (−2.41)** −0.323 (−2.49)** −0.168 (−0.38) 0.948 (1.36)
−0.033 (−0.23) −0.343 (−0.80) −0.040 (−0.21) −0.478 (1.22) −0.154 (−0.60) 0.156 (0.25) 1.354 (1.28)
−0.132 (−1.45) −0.114 (−0.38) 0.187 (1.71)* −0.364 (−1.95)* −0.352 (−2.35)** −0.443 (−0.68) −0.336 (−0.34)
−0.214 (−2.43)** −0.087 (−0.29) 0.145 (1.42) −0.354 (−1.93)* −0.326 (−2.43)** −0.177 (−0.32) 0.797 (1.00)
−0.092 (−0.45) −0.048 (−0.10) −0.061 (−0.26) −0.275 (−0.53) −0.414 (−1.31) 0.731 (0.99) 0.953 (0.62)
−0.124 (−1.21) −0.439 (−1.05) 0.157 (1.35) −0.302 (−1.56) −0.350 (−2.22)** −0.861 (1.15) −0.341 (−0.33)
High-low spread Bid-ask spread Volume ratio Firm and offer characteristics Log (market value) Log (market-to-book ratio) Log (1 + age) EBITDA < 0 dummy Internet and technology dummy Dilution factor Participation ratio
Full Period 1996–2000
Pre-Bubble Period 1996–1998
Bubble Period 1999–2000
The Long-Term Performance of Initial Public Offerings
Table 4. Determinants of Long-Term Returns Using the Local Market Index as a Benchmark.
343
344
Table 4. (Continued ) Full Period 1996–2000
VC backing dummy Intercept R2 adjusted F-statistic Observations
Bubble Period 1999–2000
Full Period 1996–2000
1.770 (2.06)** 0.168 (1.20) 0.809 (0.82)
−1.710 (−1.45) 0.536 (2.02) 0.030 (0.01)
2.358 (1.97)** 0.073 (0.44) 0.030 (0.02)
3.083 (2.98)*** 0.098 (0.76) 1.043 (0.93)
16.64% 12.06*** 555
2.20% 1.37 164
18.11% 9.63*** 391
17.95% 8.81*** 465
Pre-Bubble Period 1996–1998
Bubble Period 1999–2000
−1.765 (−0.75) 0.325 (1.06) 0.282 (0.14)
3.431 (2.98)*** 0.103 (0.61) 0.666 (0.47)
1.29% 0.82 107
17.91% 6.99*** 358
Note: Table shows the OLS regression results using the log of three-year wealth ratio (local market index) as the dependent variable. See Table 3 for variable definitions. White (1980) heteroscedastic-consistent t-statistics are within parentheses. ∗ Significant at the 10% level. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Certification variables Underwriter market share
Pre-Bubble Period 1996–1998
The Long-Term Performance of Initial Public Offerings
345
Next, we split the sample into two groups: 164 companies that went public on new markets during the pre-bubble period 1996–1998 and 391 companies that went public during the stock market bubble in 1999–2000. The results are shown in columns 2 and 3 of Table 4. We find that none of the explanatory variables are significant in the earlier part of our sample period. However, with the exception of market capitalization, the independent variables that were significant in the full period regressions are again significant in the regression for the bubble period. We conclude that most of our findings are specific to the companies that went public during the bubble period of 1999–2000. We now include our other three proxies for divergence of opinion (high-low spread, bid-ask spread and volume ratio on the first day of trading). This data is available for 465 companies. We report full period results in column 4 of Table 4. We find that the coefficient of the high-low spread is significantly negative (at a 10% level). A one standard deviation increase in the high-low spread decreases the three-year wealth relative from the sample average of 0.68 to 0.57, other things equal. The negative association between first-day returns and long-run performance weakens but remains significant at the 10% level. However, the coefficients for bid-ask spread and volume ratio are not significant. One problem might be that our divergence of opinion variables are highly correlated. In unreported tests we find that the highest correlation coefficient (0.44) is that between first-day returns and high-low spreads (all other correlations between the divergence of opinion variables are below 0.10). We therefore estimate regressions that only include one divergence of opinion variable at a time (not tabulated). We find that the high-low spread is highly significant (coefficient −2.34; t-value = 2.74) but that the bid-ask spread (coefficient −0.31; t-value = −0.26) and the volume ratio (coefficient 0.03; t-value = 0.83) remain insignificant. We investigate the two subperiods in columns 5 and 6 of Table 4. Results show that the coefficient on first-day returns is statistically significant only for bubble period IPOs. All other divergence of opinion variables are insignificant in both subperiods. We also use the wealth relatives using the NASDAQ Composite index as the benchmark as our dependent variable. Results are shown in Table 5. We find that most of our earlier findings remain the same. We also use three-year CARs as our dependent variable (not reported). We find similar results. We conclude that there is mixed support for the divergence of opinion hypothesis. Three-year wealth relatives are inversely related to first-day returns for the companies that went public during 1999–2000. But three-year wealth relatives are only weakly inversely related to high-low spreads and not related to bid-ask spreads and volume ratios. In the next subsection we investigate daily abnormal returns surrounding lockup expiration.
346
Table 5. Determinants of Long-Term Returns Using NASDAQ as a Benchmark. Full Period 1996–2000 Divergence of opinion variables First-day return
Pre-Bubble Period 1996–1998
Bubble Period 1999–2000
Full Period 1996–2000
Pre-Bubble Period 1996–1998
−0.158 (−0.47)
−0.685 (−3.67)***
−0.365 (−1.80)* −1.751 (−1.90)* −1.210 (−0.90) 0.033 (1.03)
0.065 (0.17) −1.191 (−0.58) −5.971 (−0.78) 0.064 (0.73)
−0.420 (−1.82)* −1.379 (−1.31) −1.318 (−0.95) 0.021 (0.60)
−0.186 (−2.60)*** −0.066 (−0.26) 0.135 (1.49) −0.431 (−2.57)** −0.273 (−2.14)** −0.291 (−0.65) 0.724 (1.08)
−0.061 (−0.44) −0.400 (−0.99) −0.070 (−0.39) −0.588 (−1.55) −0.103 (−0.41) 0.043 (0.07) 1.015 (0.97)
−0.166 (−1.85)* 0.008 (0.02) 0.178 (1.64) −0.345 (−1.86)* −0.346 (−2.30)** −0.365 (−0.55) −0.037 (−0.04)
−0.193 (−2.25)** −0.133 (−0.44) 0.118 (1.19) −0.330 (−1.83)* −0.313 (−2.27)** −0.309 (−0.58) 0.832 (1.08)
−0.092 (−0.45) −0.122 (−0.24) −0.088 (−0.39) −0.303 (−0.61) −0.390 (−1.23) 0.495 (0.63) 0.867 (0.57)
−0.154 (−1.52) −0.328 (−0.81) 0.141 (1.24) −0.290 (−1.51) −0.346 (−2.20)** −0.775 (−1.05) 0.011 (0.01)
Bid-ask spread Volume ratio
Log (market-to-book ratio) Log (1 + age) EBITDA < 0 dummy Internet and technology dummy Dilution factor Participation ratio
GIANCARLO GIUDICI AND PETER ROOSENBOOM
−0.575 (−3.36)***
High-low spread
Firm and offer characteristics Log (market value)
Bubble Period 1999–2000
VC backing dummy Intercept R2 adjusted F-statistic Observations
1.630 (1.93)* 0.191 (1.41) 0.833 (0.85) 15.44% 11.12*** 555
−1.552 (−1.35) 0.530 (2.05) 0.466 (0.28) 3.24% 1.55 164
2.477 (2.08)** 0.093 (0.57) 0.408 (0.34)
2.936 (2.82)*** 0.105 (0.74) 1.152 (1.06)
17.51% 9.28*** 391
17.38% 8.51*** 465
−1.764 (−0.73) 0.313 (1.04) 0.567 (0.28) 1.78% 0.78 107
3.572 (3.09)*** 0.109 (0.64) 1.059164 (0.76) 18.00% 7.03*** 358
Note: Table shows the OLS regression results using the log of three-year wealth ratio (NASDAQ composite index) as the dependent variable. See Table 3 for variable definitions. White (1980) heteroscedastic-consistent t-statistics are within parentheses. ∗ Significant at the 10% level. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
The Long-Term Performance of Initial Public Offerings
Certification variables Underwriter market share
347
348
GIANCARLO GIUDICI AND PETER ROOSENBOOM
5.2. Daily Abnormal Returns Before and After Lock-Up Expiration Most companies in our sample have mandatory lock-up agreements ranging from 6 months (Neuer Markt, NASDAQ Europe) to 12 months (Nouveau March´e, Nuovo Mercato and EuroNM Belgium) following the IPO date. Insiders are prohibited from selling any of their shares during this period. Ofek and Richardson (2003) argue that the lock-up agreement can be thought of as a stringent form of short sale restriction. The expiration of the lock-up agreement loosens this short-sale constraint and allows insiders to start trading in their firm’s shares. It is likely that these insiders have private and more realistic information about the firm’s prospects than (overoptimistic) outside investors that have been determining stock prices before lock-up expiration. We hypothesize this private information gradually gets incorporated into stock prices resulting in a stock price decline after lock-up expiration. We investigate daily abnormal returns surrounding the earliest unlock date. This is the first time that (part of the) pre-IPO owners are allowed to sell their shares and private information about the future prospects of the company can get incorporated into share prices.7 We are able to determine the earliest unlock date for 547 sample firms. In eight cases we could not determine the exact unlock date because it is conditioned on the company being profitable in the future or the information is missing from the prospectus. In our sample, the average (median) lock-up agreement expires 13.4 (12.1) months after the IPO date. Table 6 shows that there is an average cumulative abnormal return of −0.2% during day −1 to day 0, where day 0 is the earliest unlock date. A total of 54% of our sample firms experience negative abnormal returns during this interval. However, the stock price reaction is not significantly different from zero. We find that the average five-day cumulative abnormal return during days [−4,0] is significantly negative at −1.4% with 62% of sample firms having negative returns. The average cumulative abnormal return during days [−10,0] equals −2.1%. This is consistent with U.S. studies that report a significantly negative stock price reactions surrounding lock-up expiration (Field & Hanka, 2001; Brav & Gompers, 2003). Table 6 also shows that the cumulative abnormal returns for the various time intervals are more negative for those companies that went public during 1999–2000. In fact, the cumulative abnormal returns during [−4,0] and [−10,0] are only significant for companies that went public during the stock market bubble. Table 6 also shows cumulative abnormal returns using the NASDAQ Composite index as the benchmark. Results are similar. In order to examine the long-run effects of lock-up expiration we measure returns during a period of 100 trading days before and 100 trading days after the earliest lock-up expiration. Each company has to have at least 50 trading
349
The Long-Term Performance of Initial Public Offerings
Table 6. Abnormal Returns Around Lockup Expiration. Cumulative Abnormal Returns (Local Market Index) (%)
t-Stat
Cumulative Abnormal Returns (NASDAQ Index) (%)
t-Stat
−0.23 −1.40 −2.06 −0.25
−0.70 −2.36** −2.31** −0.30
−0.14 −1.30 −1.90 −0.47
−0.40 −2.12** −2.17** −0.56
Pre-bubble period 1996–1998 Days −1 to 0 0.16 Days −4 to 0 1.12 Days −10 to 0 1.83 Days 1 to 10 −0.63
0.32 1.16 1.28 −0.45
0.13 0.35 0.26 −1.34
0.24 0.36 0.18 −0.94
Bubble period 1999–2000 Days −1 to 0 Days −4 to 0 Days −10 to 0 Days 1 to 10
−0.94 −3.30*** −3.29*** −0.09
−0.25 −1.96 −2.77 −0.12
−0.57 −2.57** −2.56** −0.12
Full period 1996–2000 Days −1 to 0 Days −4 to 0 Days −10 to 0 Days 1 to 10
−0.38 −2.42 −3.63 −0.10
Note: Table shows abnormal returns around the earliest lockup expiration date at which at least one of the pre-IPO owners is allowed to sell (part of) his shares. Day 0 is the earliest expiration day of the lockup agreement. The abnormal returns are computed as the difference between the return on the stock and the return on the local market index or the NASDAQ composite index. The dataset includes 547 initial public offerings during the full period of 1996–2000, 157 initial public offerings during the pre-bubble period of 1996–1998 and 390 initial public offerings during the bubble period of 1999–2000. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
days during the pre-lock period to be included in the sample. We exclude the 21 days surrounding lock-up expiration during days [−10,+10] that we investigated earlier. Following Ofek and Richardson (2003), we argue that the 100 days trading interval after the earliest unlock date is a long enough period to allow more “pessimistic” investors to sell their shares. The comparison to the period before lock-up expiration controls for the inability to sell shares. Table 7 reports the results. We find that there is a highly significant difference between the average daily abnormal return in the period before and after lockup expiration. Before lock-up expiration the average daily abnormal return is positive at 0.07%, whereas after lock-up expiration the average daily abnormal return is negative at −0.06%. The difference of −0.13% (or −12.2% during the 100 trading day interval) is highly significant. This suggests that the long-run underperformance sets in after lock-up expiration. Table 7 also shows that this pattern applies to
350
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Table 7. Comparison of Daily Abnormal Returns Before and After the Expiration of the Lockup Agreement. Daily Abnormal Returns (Local Market Index) (%)
Daily Abnormal Returns (NASDAQ Index) (%)
Observations
Full period 1996–2000 Post-lockup expiration Pre-lockup expiration Difference t-stat (difference = 0)
−0.06 0.07 −0.13 −3.72***
−0.04 0.07 −0.11 −3.03***
54,700 54,415
Pre-bubble period 1996–1998 Post-lockup expiration Pre-lockup expiration Difference t-stat (difference = 0)
−0.01 0.16 −0.17 −2.90***
−0.04 0.11 −0.15 −2.34**
15,700 15,591
Bubble period 1999–2000 Post-lockup expiration Pre-lockup expiration Difference t-stat (difference = 0)
−0.09 0.03 −0.12 −2.66***
−0.04 0.06 −0.10 −2.17**
39,000 38,824
Note: Table compares average daily abnormal returns between a period of 100 trading days before and after expiration of the lockup agreement. These averages exclude days −10 to +10 around lockup expiration (day 0). At lockup expiration at least one of the pre-IPO owners is allowed to sell (part of) his shares. The dataset includes 547 initial public offerings during the full period of 1996–2000, 157 initial public offerings during the pre-bubble period of 1996–1998 and 390 initial public offerings during the bubble period of 1999–2000. Each company has to have at least 50 trading days during the pre-lock period to be included in the sample. The abnormal returns are computed as the difference between the return on the stock and the return on the local market index or the NASDAQ composite index. ∗∗ Significant at the 5% level. ∗∗∗ Significant at the 1% level.
both the firms that went public during 1996–1998 and the companies that went public during 1999–2000. In addition, the abnormal returns using the NASDAQ Composite index as the benchmark display a similar pattern. We interpret this finding as consistent with the divergence of opinion hypothesis of Miller (1977).
6. CONCLUSIONS In this chapter we investigate the determinants of the long-run performance of IPOs on Europe’s new stock markets. We report that the average company that
The Long-Term Performance of Initial Public Offerings
351
went public on these markets has been a very poor long-term investment. Investors would be left with an average of only 68 cents (72 cents) compared to one euro invested in the local market index (NASDAQ Composite index). We test the divergence of opinion hypothesis of Miller (1977) as one possible explanation for why the average company performs so poorly. We find that the stock price performance during a three-year window is inversely related to first-day returns. This is consistent with the divergence of opinion hypothesis. This hypothesis states that overoptimistic investors initially set market prices above fundamental values (resulting in high first-day returns) and that prices gradually decline to fundamental values over time as more pessimistic investors enter the market. However, the other three proxies for divergence of opinion (high-low spread, bid-ask spread and volume ratio on the first trading day) are not significantly associated with long-run stock price performance. We therefore conclude that there is mixed support for the divergence of opinion hypothesis. In addition, we find that Internet and technology companies that went public during the bubble period of 1999–2000 performed very poorly. We also report that IPOs completed in 1999–2000 that were underwritten by reputable investment banks performed significantly better than other IPOs. Next, we investigate daily abnormal returns surrounding lockup expiration. Lock-ups prevent insiders from selling their shares in the period immediately following the IPO. The expiration of the lock-up agreement enables insiders to sell their shares. In our sample, the average (median) lock-up agreement expires 13.4 (12.1) months after the IPO date. It is likely that these insiders have private and more realistic information than the (overoptimistic) investors that have been determining market prices in the period before lock-up expiration. The divergence of opinion hypothesis predicts that this private information gradually gets incorporated into stock prices resulting in a stock price decline after lock-up expiration. Accordingly, we find that the average daily abnormal return during 100 trading days after lock-up expiration is significantly lower than the average daily abnormal return during 100 trading days before lock-up expiration. Overall, our results provide some support for the divergence of opinion hypothesis of Miller (1977). However, it is difficult to directly measure the divergence of opinion among investors. It may be impossible to distinguish between measures of uncertainty and divergence of opinion (Houge et al., 2001). Our results should therefore be interpreted with care.
NOTES 1. The most notorious example is that of Comroad, a “traffic-navigation technology” company that went public in 1999. In April 2002 it was revealed that nearly all of the
352
GIANCARLO GIUDICI AND PETER ROOSENBOOM
company’s $94 million in reported revenue for 2001 was fictitious. Comroad has been delisted and Comroad’s CEO is facing criminal charges (Wall Street Journal, 1 October 2002). 2. High-tech companies are active in SIC codes 3571, 3572, 3575, 3577, 3578 (computer hardware), 3661, 3663, 3669 (communications equipment), 3674 (semiconductors), 3812 (navigation equipment), 3823, 3825, 3826, 3827, 3829 (measuring and controlling devices), 3841, 3845 (medical instruments), 4812, 4813 (telephone equipment), 4899 (communications services) and 7370, 7371, 7372, 7373, 7374, 7375, 7378 and 7379 (software). We collect SIC codes from COMPUSTAT Global Vantage and Worldscope Disclosure. We identify European Internet firms using the list provided by Knauff et al. (2003). They provide a list of 138 European Internet IPOs based on membership of the Bloomberg European Internet Index and talks with investment bankers. 3. Note that the percentage market share of the lead manager is based on all IPOs on the local stock market. For example, we calculate the percentage market share of underwriters in Germany including all IPOs on the new market segment (Neuer Markt) and main market segments (Amtlicher Handel and Geregelter Markt) of the Frankfurt Stock Exchange. 4. We do not investigate long-term operating performance because several Neuer Markt companies have published false annual and quarterly accounting data (D’Arcy & Grabensberger, 2003). 5. The average holding period is 35.2 months. Firms are delisted because of takeovers (13 firms) and financial distress (22 firms). Although many of the Neuer Markt companies in our sample became insolvent during the 36-month holding period, it was relatively difficult for these companies to be expelled from the market. It became easier to delist companies as from October 2001. Neuer Markt companies with a stock price less than d1 and a market capitalization of less than d20 million for a period of 30 days were put on a surveillance list. If there was no change in this situation for a further period of 90 days these companies were delisted. 6. Lyon et al. (1999) and Brav (2000) also show that long-horizon event studies are associated with statistical difficulties. 7. For example, existing shareholders of ABIT AG agreed not sell their shares for a period of 6 months from the IPO date. In addition, they agreed (with the exception of the venture capitalist 3i plc) not to sell shares for a further period of 6 months without consent of the lead manager. In this example, we take the earliest unlock date to be 6 months after the IPO date (the earliest time at which one of the pre-IPO shareholders (3i plc) is allowed to sell its shares).
ACKNOWLEDGMENTS Giancarlo Giudici acknowledges funding from Cofinanziamento MIUR. Peter Roosenboom acknowledges funding from ERIM. We thank Janice Tjon Sien Kie for research assistance and seminar participants at Erasmus University Rotterdam for valuable comments and suggestions. All errors are our own.
The Long-Term Performance of Initial Public Offerings
353
REFERENCES Aggarwal, R. K., Krigman, L., & Womack, K. L. (2002). Strategic IPO underpricing, information momentum, and lockup expiration selling. Journal of Financial Economics, 66, 105–137. Barber, B. M., & Lyon, J. D. (1997). Detecting long-run abnormal stock returns: The empirical power and specification of test statistics. Journal of Financial Economics, 43, 341–372. Brav, A. (2000). Inference in long-horizon event studies: A bayesian approach with application to initial public offerings. Journal of Finance, 55, 1979–2016. Brav, A., & Gompers, P. (1997). Myth or reality? The long-run underperformance of initial public offerings: Evidence from venture and non-venture-capital-backed companies. Journal of Finance, 52, 1791–1822. Brav, A., & Gompers, P. (2003). The role of lockups in initial public offerings. Review of Financial Studies, 16, 1–29. Carter, R. B., Dark, F. H., & Singh, A. K. (1998). Underwriter reputation, initial returns and the long run performance of IPO stocks. Journal of Finance, 53, 285–311. D’Arcy, A., & Grabensberger, S. (2003). The quality of Neuer Markt quarterly reports – An empirical investigation. International Journal of Accounting, 38, 329–346. Dorn, D. (2003). Does sentiment drive the retail demand for IPOs? Working Paper, Drexel University. Espenlaub, S., Gregory, A., & Tonks, I. (2000). Re-assessing the long-term underperformance of UK initial public offerings. European Financial Management, 6, 319–342. Field, L. C., & Hanka, G. (2001). The expiration of IPO share lockups. Journal of Finance, 56, 471–500. Gezcy, G. C., Musto, D. K., & Reed, A. V. (2002). Stocks are special too: An analysis of the equity lending market. Journal of Financial Economics, 66, 241–269. Houge, T., Loughran, T., & Yan, X. (2001). Divergence of opinion, uncertainty, and the quality of initial public offerings. Financial Management, 30, 5–23. Jain, B. A., & Kini, O. (1994). The post-issue performance of IPO firms. Journal of Finance, 49, 1699–1726. Jaggia, S., & Thosar, S. (2004). The medium-term aftermarket in high-tech IPOs: Patterns and implications. Journal of Banking & Finance, 28, 931–950. Knauff, P., van der Goot, T., & Roosenboom, P. (2003). Is accounting information relevant to valuing European Internet IPOs? In: A. Ginsberg & I. Hasan (Eds), New Venture Investment: Causes and Consequences (pp. 257–280). Amsterdam: North Holland. Levis, M. (1993). The long-run performance of initial public offerings: The UK experience 1980–1988. Financial Management, 22, 28–41. Ljungqvist, A. P. (1997). Pricing initial public offerings: Further evidence from Germany. European Economic Review, 41, 1309–1320. Ljungqvist, A. P., & Wilhelm, W. J. (2002). IPO allocations: Discriminatory or discretionary? Journal of Financial Economics, 65, 167–201. Ljungqvist, A. P., & Wilhelm, W. J. (2003). IPO pricing in the dot-com bubble. Journal of Finance, 58, 723–752. Logue, D. E., Rogalski, R. J., Seward, J. K., & Foster-Johnson, L. (2002). What is special about the roles of underwriter reputation and market activities in initial public offerings? Journal of Business, 75, 213–243. Loughran, T., & Marietta-Westberg, J. (2001). When any news is bad news: Extreme events following an initial public offering. Working Paper, University of Notre Dame and Michigan State University. Loughran, T., & Ritter, J. R. (1995). The new issues puzzle. Journal of Finance, 50, 23–51.
354
GIANCARLO GIUDICI AND PETER ROOSENBOOM
Lyon, J. D., Barber, B. M., & Tsai, C-L. (1999). Improved methods for tests of long-run abnormal stock returns. Journal of Finance, 54, 165–201. Megginson, W. L., & Weiss, K. A. (1991). Venture capitalist certification in initial public offerings. Journal of Finance, 46, 879–903. Miller, E. M. (1977). Risk, uncertainty, and divergence of opinion. Journal of Finance, 32, 1151–1168. Mikkelson, W. H., Partch, M. M., & Shah, K. (1997). Ownership and operating performance of companies that go public. Journal of Financial Economics, 44, 281–307. Ofek, E., & Richardson, M. (2003). DotCom mania: The rise and fall of Internet stock prices. Journal of Finance, 58, 1113–1137. Rajan, R., & Servaes, H. (1997). Analyst following of initial public offerings. Journal of Finance, 52, 507–529. Ritter, J. R. (1991). The long-run performance of initial public offerings. Journal of Finance, 46, 3–27. Stehle, R., Ehrhardt, O., & Przyborowsky, R. (2000). Long-run stock performance of German initial public offerings and seasoned equity issues. European Financial Management, 6, 173–196. Teoh, S., Welch, I., & Wong, T. J. (1998). Earnings management and the long-run performance of initial public offerings. Journal of Finance, 53, 1935–1974. White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica, 48, 721–746.
