The Internet Boom in a Corporate Finance Retrospective

The Internet Boom in a Corporate Finance Retrospective 1 Ulrich Hege 2 HEC School of Management Sébastien Michenaud 3 HEC School of Management 1 st version: April 2004 2 nd version: September 2004 Abstract This paper reviews the important transformations, some of them short-lived and some lasting, in the financing of innovative companies that emerged with the advent of the internet and the “internet bubble” between 1998-2000, and tries to interpret them from the perspective of modern corporate finance theory. The internet had a strong impact on the investment flows and the operations of the venture capital industry in the US and in Europe. We argue that the superstar valuations of the bubble years, the value drop in the internet shakeout and the volatile markets in the period thereafter are best understood from the vantage point of network economics, and we focus in particular on the network effects that explain the recent success of well-known internet companies, and the failure of many others. We review innovative valuation techniques methods that emerged during that time to justify exorbitant stock prices. A centerpiece of our interpretation are the specific agency conflicts that emerged in the financing of internet start-ups, and the mechanisms that evolved in response to them. In addition, the paper briefly discusses the pay incentives and capital structure of internet startups, and explores finally the particularities of the IPO wave during the boom, that was characterized by the large number of companies going public without near-term profitability. Classification JEL: G31, G32, G35, L86. 1 We are grateful to Hervé Tanguy, Nicolas Curien and an anonymous referee for many helpful comments. 2 Dept. of Finance and Economics, 1 rue de la Libération, 78351 Jouy-en-Josas Cedex, France. Email hege@hec.fr. Also affiliated to CEPR, ECGI, GREGHEC, Europlace Institute of Finance and CentER. 3 Dept. of Finance and Economics, 1 rue de la Libération, 78351 Jouy-en-Josas Cedex, France. Email michenauds@hec.fr.- 2 - Introduction Soon after the World Wide Web became popular in the second half of the 1990s, many expected the internet to lead to a major technological revolution that would fundamentally transform consumer behavior and the mode of competition among firms. The ubiquitous term of the “New Economy” epitomized the widely accepted idea that new internet-based companies and business models had the potential to supplant existing firms and industries, and that they would give rise to a period of strong economic growth. These beliefs about the internet, and more specifically, the exuberant expectations about growth rates of the new sectors and the potential prize in a winner-takes-all competition, fed a wave of broad-based economic optimism that nourished, in the period 1998-2000, a major speculative bubble, the “internet bubble”. In New York, the NASDAQ, the major high-tech stock index, more than tripled in value between October 1998 and March 2000. The backlash was equally dramatic, with the NASDAQ index losing more than 75% in the following two years, and perhaps even excessive, considering that it subsequently increased by more than 80% from October 2002 to April 2004. With the NASDAQ index in April 2004 still accounting for only 40% of its peak value, it firmly looks today as if March 2000 marked indeed the peak of a speculative bubble. In the bubble period, there was a widespread belief among venture capital investors and financial markets that the economics of internet-based networks would convey formidable market power to successful internet start-ups. “Old economy” firms could hardly challenge that and this situation would enable to create hitherto unknown levels of growth. Venture capital funds aggressively injected capital into a wide array of internet-related start-ups, and a flurry of internet-related initial public offerings occurred with an unprecedented level of initial returns. In this article, we review some of the key developments of this remarkable period and offer an interpretation of the events from the viewpoint of corporate finance theory - with the benefit of hindsight, as we freely admit. Our goal is to elucidate the link between the new financial phenomena and the internet, namely how the advent of the World Wide Web could give rise to such a transformative change of financial markets. More specifically, we focus on six aspects that we think are of peculiar importance: (i) the impact of the internet on the role and the efficiency of the venture capital industry in Europe and the United States; (ii) the beliefs of financial markets’about internet competition and internet business models; (iii) valuation techniques applicable to these companies; (iv) agency conflicts between investors, in particular venture capital firms, and internet entrepreneurs, as well as agency conflicts within venture funds, between fund providers and fund managers; (v) specific features in incentive compensation and financial policies of internet firms and finally (vi) the initial public offerings of these start-ups. 1. Internet and Venture Capital The arrival of the internet and of internet start-ups had dramatic effects on equity markets everywhere. In the United States, the largest and most developed market for venture capital, venture capital investments have seen a very strong and persistent growth throughout the 1990s and a virtual explosion during the two years between 1998 and 2000, with a fivefold increase of investments in these two years alone. This abrupt development was largely driven by the internet: in 2000, more than 80% of all venture capital investments have been made in internet-related companies, for a total of more than $56 billion (PricewaterhouseCoopers (2001)). By contrast, less than a quarter of all venture capital investments went to internetrelated firms prior to 1997. A rapid decline both in venture capital funding in general and in - 3 - internet investment in particular has taken place since, even though the share of internetrelated investments in venture capital disbursements is, with more than a third, perhaps surprisingly high. While other sources of funds may be available to internet start ups, like traditional bank loans, venture capital is widely believed to be more efficient relative to other financing modes when it comes to funding low collateral, high risk, high return projects (see Ueda (2004) for a convincing theory model on this topic). This interpretation is consistent with the strong growth since the 1980s, when bank lending to small firms was stagnant or even falling during this period (Ueda (2002)). Figure 1. Venture Capital Funding in the United States and Share of Internet Start-Ups 55% 44% 80% 79% 53% 39% 34% 20% 68% 0 20 40 60 80 100 120 1995 1996 1997 1998 1999 2000 2001 2002 2003 Total VC funding (billion USD) Internet (as a percentage of total) 7.6 11.5 14.8 21.2 54.3 105.8 40.7 21.2 18.4 Source: PricewaterhouseCoopers-Ventureconomics In Europe, a rather similar development took place, but it is worthwhile to review the most important differences. First, Europe is a relative newcomer to venture capital funding, as the numbers published annually by the European Venture Capital Association (EVCA (2002)) show. In fact, the investment figures given by EVCA have traditionally been dominated by investments into more mature companies, like leveraged buy-outs or corporate restructurings, operations known as private equity. The arrival of the internet led for the first time to a massive investment flow into start-up companies that jump-started venture capital funding in the American sense of the term. In 1999, venture capital investments in start-up companies in high-tech industries amounted to € 8.9 billion. While this was only about a quarter of American investments during the same year, it indicates a dramatic catch-up compared with earlier years. In fact, venture capital funding for start-up companies increased by a staggering 166% in 1999 alone. The share of funds provided to high-tech start-ups, as a total of all venture capital and private equity investments, increased over the same period from 23% to 35%. This rapid evolution was largely driven by the advent of the internet. In fact, as recent empirical work argues, the internet revolution was the true starting point of a genuine venture capital industry in Europe dedicated to the financing of high-tech start-ups (Hege et al. (2003)).- 4 - To understand the dramatic swings in venture capital commitments on both sides of the Atlantic, it is useful to put them into the historical context of the venture capital industry. The strong growth of venture funding in the US, and the appetite for similar investments elsewhere, have largely been fuelled by the astounding rates of returns of venture capital investments. While estimates on the long-run rates of returns vary, a typical estimate available prior to the take-off of the internet bubble was that by Ibbotson Associates, who estimated the annual average return for American venture capital at 45% for the 1960 to 1995 period. Moreover, a common argument was that venture capital returns appeared to be almost uncorrelated with the market return of the equity market, which – according to the CAPM asset pricing model – should mean that its expected return is close to the riskfree return, i.e., in the case of the American market, close to a 4% (nominal) return in the long run. 4 Compared to the United States, rates of return have traditionally been lower in Europe. To give an example prior to the internet bubble, NVP, the Dutch venture capital association, calculated a mean rate of return for its members of 15% for the 1989-1998 period, and of only 6% for start-up financing activities (NVP 2000). Just prior to the year 2000, a clear improvement in venture capital returns occurred, with the mean return of its member firms increasing from 11.9% in 1998 to 14.5% in 1999, and a particular sharp increase in early stage investments, according to EVCA. The reluctance of European investors to fund early stage activities, as well as their sudden interest in internet start-ups during the bubble must be seen against the backdrop of such numbers. A lot of money was piling into venture capital funds during a rather short period of time, and apparently rather indiscriminately; this is perhaps less surprising if one tries to adopt the vantage point of an investor in 1998. Another element that helps to understand the dramatic increase in available funds is that venture capital has always been a highly cyclical industry. The cyclicality not only appears in fund commitments, but also in the industry sectors where investments occur. Ultimately, the cyclicality in venture capital investments and returns is driven by what is known as the market conditions for venture “exits”, the sale of venture capital investments, typically in the form of an initial public offering (IPO) or a trade sale (i.e. selling it to another firm, usually an industry incumbent). Venture capital depends very much on the exit market conditions since the bulk of venture capital returns are generated by a minority of projects. Typically, among the successful ventures, the best performers are sold in an IPO (see, for example, Cochrane (2004)). IPOs have known highly cyclical market conditions for many decades, with hot issue markets of intensive IPO activity and high first-day returns alternating with cold issue markets of low IPO activity and paltry or negative initial returns. The internet bubble marked the last hot issue period from 1998 until summer 2000. The second important exit road for venture capital investments, acquisitions or trade sales, as they are known in the industry, are also intimately linked to the stock market. The current stock market conditions play a crucial role for the parameters that determine the valuation of any non-listed firm in an acquisition: for example, a peer group of comparable but listed firms is used to extract the multiples that determine the price range for the acquisition. In addition, the forecasts of future cash flows and their growth rates are certainly influenced by stock market conditions. 4 The stated numbers on returns on correlations are not uncontested, see e.g. Cochrane (2004) for different results. As far as the correlation with the market index is concerned, a major difficulty is that rates of returns are calculated from accounting figures and that venture capital firms tend to delay write-downs. Returns calculated from accounting figures probably tend to underestimate the return correlation.- 5 - The arrival of the internet and the concomitant explosion in investments had two clear consequences on the venture capital industry: i) The strong increase in funds was followed by the market entry of new venture capital investors and the creation of new venture capital funds. Most of the new money inflow came from institutional investors, like pension funds and university endowments, that previously had a much smaller commitment in venture funding. The capital inflow meant that, for a few years, “money was chasing deals” and that the balance of power between fund providers and entrepreneurs shifted in favor of the latter. ii) This large money inflow resulted in an increased scarcity of experienced venture capitalists. Being a successful venture capitalist requires substantial experience, and venture capital partners have typically garnered extensive knowledge as fund managers, investment bankers or former entrepreneurs before joining a venture fund. Venture capitalists learn their trade by costly trial and error rather than in business schools. This idea has been summarized bluntly in the saying that the training of a venture capitalist would be as expensive as crashing an airplane. In an interesting theory model, Michelacci and Suarez (2004) have highlighted the effect of a shortage of experienced venture capitalists by considering a venture capital sector that employs two complementary resources, capital and venture capitalists. While competitive financial markets guarantee an elastic supply of capital, experienced venture capitalists are in very inelastic supply because of their lengthy training process. Thus, scarce venture capitalists are the factor ultimately limiting innovation and economic growth. At the same time, the practices of the venture capital business were transformed by i) the advent of new players in the field and ii) shortened business cycles. i) Concerning the type of players in the venture capital industry, there were both seemingly permanent and rather short-lived transformations. On the one hand, well-established large companies set up venture funds of their own, called corporate venture capital funds, with the double objective to better follow and control the new modes of innovation and to participate in the expected boon of returns. While many of the internet-based corporate funds faltered as well, the trend of companies to spin off parts of their research funding into separate corporate venture funds clearly did not stop altogether with the bursting of the bubble. On the other hand, much of what was hailed as new practice was in fact the short-lived byproduct of the bubble, like venture capital incubators and venture catalysts. Instead of putting funds into clearly defined start-up firms, venture investors created research laboratories, called incubators, where individual researchers where assigned to projects and frequently reassembled to new teams in the hope of ultimately hatching successful start-up firms. Venture catalysts were essentially holding companies, like Softbank, CMGI, and Internet Capital Group. They combined investments in newly started venture firms with holdings in stock market-listed internet firms. Those firms were aiming for a public listing themselves, unlike venture funds that hitherto had chosen the form of closed-end funds. Incubators and venture catalysts quickly fell in oblivion after the internet bubble burst. ii) An accelerated venture capital cycle and premature exits. The prototypical venture-funded start-up firm normally goes through a succession of stages that characterize different phases in its development, and each stage is normally associated with one or several financing rounds. In the first stage, called seed financing, typically, the goal is to identify the potential of the initial innovative idea through the establishment of a business plan. The next stage is typically early-stage financing, followed by development and expansion stages (Gompers and Lerner (1999)) and, finally, the exit of the venture capitalists. Traditionally, the entire cycle- 6 - used to last perhaps two to four years on average. Already starting since 1995, and even more so during the internet bubble years, a clear acceleration in the venture capital cycle could be observed; the data in Cochrane (2004), for example, document this trend very clearly in the US venture capital sector. This acceleration was closely linked to strategies of premature exits by venture capitalists, for which the tendency to take internet companies public long before they were even close to reaching breakeven became a powerful symbol. Such a development was unconceivable even a few years earlier. 2. Internet, network competition, and superstar valuation Looking back at the financial markets during the internet bubble years, one may wonder what kind of rationale at the time was strong enough to persuade so many people, and to drive and sustain such a frenzy in start-up funding, initial public offerings and stock prices. While fully acknowledging that this is a difficult and inevitably subjective exercise, in this section we will attempt an interpretation, based on the economics of network competition. We hypothesize that market participants, or at least an influential segment of them, entertained a belief that winner-takes-all competition would become the norm across all internet-based industries. This belief, so our argument, may have played a substantial role in the creation of the speculative bubble specific to internet start ups and, later, when contrarian evidence could no longer be ignored, in its precipitous downfall. Network industries are characterized by the fact that a consumer’s value of the goods and services depend on his use of the associated network, namely both on the state of the network and on the user’s integration within the network. In addition, an agent’s use of the network affects the state of the network for other users by creating network externalities. This will be the case for example because she increases the network’s value through an increase in availability, quality and the diversity of the network’s offerings, thereby creating positive externalities, or because she contributes to the network’s congestion, which creates negative externalities. These effects are not specific to the internet and have long been a prominent feature of traditional network industries like e.g. electricity distribution, telecommunication and the railroad industry. The computer industry itself was hailed as the perfect example of a network industry long before the advent of the World Wide Web. In this market, the end users’ requirements for compatibility between different systems gave rise to a strong demand for standardization in hardware as well as in software. This, in turn, gave a decisive advantage to the market leaders who could impose their technologies as de facto standards. For internet-related industries, network effects were expected to result in the emergence of players with strong market positions. These effects were created and sustained through a demand for compatibility, as well as a club e ect related to both the size of the user base and its usage, along with the quantity and the quality of content that was generated. An early example that powerfully foreshadowed the demand for compatibility and network effects in internet-based activities was the browser market. In 1994, when Netscape was launched, the browser market was considered as a natural monopoly market akin to the market for PC operating systems where Microsoft’s Windows was then considered a near-monopoly product. In 1995, when Netscape went public on NASDAQ, financial markets seemed to be fully swayed by this logic, as the surge in the stock price following the IPO and the huge level of media attention attest. Microsoft’s subsequent aggressive strategy to contest Netscape’s initial market leadership, which consisted in providing Internet Explorer, its competing product, for free and bundling it with its Windows operating system, was viewed as further evidence to the power of network economics. Microsoft’s strategy consisted essentially in increasing the installed base in order to marginalize Netscape, made possible by Microsoft’s grip on the- 7 - operating systems market and a policy of frequent upgrades. With its determined response, Microsoft made it clear that it viewed competition in the browser market as network competition, and its subsequent success in marginalizing Netscape seemed to demonstrate that an established market leader in the internet era could not easily be challenged. During the internet bubble – which in this article we date to the period from summer 1998 to spring/summer 2000 – there was a widespread and indiscriminate belief in the network economics of any internet-related business. In some cases, the belief in the network effects was fully justified; eBay’s success is probably the best illustration of such a case. Valuations were largely overstated in most other cases. This was probably due to the lack of understanding of the future winning business models, a lack that in turn was owed to the large uncertainty with regard to customer behavior and innovation. eBay is the perfect illustration of a successful internet marketplace that has clearly benefited from a first mover advantage and strong network effects. eBay was the first player to launch an auction web site in the US in 1995 and it quickly established a dominant position on its domestic market. In the following years it expanded internationally through acquisitions of, or partnership with established local online auction companies and other similar companies. So far, it has managed to sustain its domestic leadership, and to a remarkable extent, to put it across other regions of the world. Likewise, Yahoo! was launched early in 1994. It clearly benefited from a similar first-mover advantage and emerged as an unchallenged portal leader in 1998-2000 – at least this was the belief that financial markets expressed with Yahoo’s market valuation at the time. However, Google, started only in 1999, was able to capture a large share of the search engine business through its innovative technology and the build-up of a large market share in the internet advertising market. It looks like a curiosity today that Google was initially not among the first players in the search engine market. Microsoft and Yahoo! were major competitors and claimed millions of users. The success of Google shows that dominant market positions in the internet era were often rather contestable. In the case of the market for web searches, one could argue that switching costs for a single end-user are low relative to other market segments such as internet auctions. In the market for internet auctions, a coordinated move among most or at least some of the agents linked to the network would be required to justify the decision by a single customer to switch to a competitor. Such a coordination problem is typical of a public good problem and naturally leads to the increased sustainability of a player’s natural monopoly position. In addition, the value added by Google’s technology had a huge impact on usage and justified uncoordinated switching from leading portals (Yahoo! and Microsoft’s MSN) to the new entrant. As of early 2004, competition in this market seems to focus much more on new technological developments rather than customer acquisition. On the other hand, the economics of network effects were probably much less important for ecommerce start-ups. Nevertheless, during the internet bubble, financial markets clearly assumed market leadership in these markets to be a lasting advantage and valued it accordingly, notably for B2C and B2B internet start-ups. In B2C, it now seems evident that market leadership was not a decisive factor for success, let alone survival, as the number of now-defunct e-commerce web sites suggests (famous examples include Pets.com and boo.com). There are, however, examples showing a role of network economics in internet retailing. The case of amazon.com is probably the most illustrative one. Closer examination of amazon’s success story suggests that a dominant market position was not sufficient; its success is probably as much rooted in the ability to build up massive capacities for logistics that guaranteed smooth and fast execution on an unrivalled range of books than in assembling- 8 - a larger customer base than its rivals. The ability to build the initial business on a profitable market segment (books, and later music and media) also enabled amazon.com to eventually turn a profit. As a result, amazon.com quickly established itself as the leading virtual bookstore and built on its initial leadership to extend activities in a large array of unrelated products such as clothing and consumer electronics. The fate of B2B marketplaces was even more sobering than that of B2C companies: practically none of them proved capable of delivering a viable business model, contrary to the initial expectations. In 1999, the widespread belief was that marketplaces would match suppliers with purchasers, that they would exchange information on their products and prices and do business online. In this emerging industry, network effects were expected to be high, and large exchange platforms promised to deliver big savings for their customers on their procurement, savings of which the platform were supposed to capture their share. It turned out, however, that suppliers were very reluctant to post their product catalogues online since this would have meant to circulate sensitive information like prices and rebates in the public domain. At the same time, purchasers did not join massively these marketplaces, considering that better prices were not all that mattered. Basically, it turned out that the B2B marketplace business model had a fundamental flaw in misjudging the incentives of both sides to participate. Interestingly enough, the financial markets erred as much as the media in their expectation of the B2B business model. By 2002 the large majority of players in this segment of the internet economy had disappeared. Most of the companies that survived the B2B shakeout changed their focus, for example by providing web-based software for private marketplaces. During the internet bubble from 1998 to 2000, the prevalent belief in financial markets was that network effects would play a predominant strategic role for most internet start-ups. Network effects would provide the leader with a very profitable position that would be difficult to challenge for any of its followers. As a result, any start-up that launched a new venture on the internet had, first and foremost, to build a dominant market share on its target market. Because of the network effects described above, competition among start-ups was perceived as a winner-takes-all competition (Noe and Parker (2000)) akin to a patent race where the lion’s share of the profits are earned by the leader. This type of competition is very similar to that of entertainment and sports superstars who capture most of the rent available in their field, at the expense of second-tier competitors (Rosen (1981)). In the internet economy, this meant superstar valuations of the market leaders by the financial markets. Players with a strong market share position and/or brand position in any internet field earned a substantial market premium relative to their competitors. With hindsight, equity markets were showing the right intuition when they identified network effects as the potential value drivers in the internet economy: successful internet business models were exclusively based on them. It even happened in a very small number of cases that the market initially underestimated the value premium that an internet-based company could sustain, as the example of eBay’s subsequent success suggests. In fact, in April 2004, eBay’s stock price was more than 60% higher than at the height of the bubble in March 2000. But this, of course, is a rare exception. The problem was that in the 1998-2000 period, market operators had a poor understanding of internet economics, and were not (yet) capable to discern the network-based economic models that would eventually be able to generate a profit. Instead, and certainly helped by fascination about all the possibilities that the internet seemed to open, the value of network effects was generally overrated and indiscriminately applied to every business models alike, if only it showed some relation with the internet. The near-total wipeout of companies that were once considered stars and uncontested market- 9 - leaders in the most promising segments of the “new economy”, such as Ariba (-98%) in the B2B market or AOL (now Time Warner) in the ISP market (-80%), is a stark reminder of this episode. Whatever the financial markets’ mistakes as they became obvious soon after, during the bubble years start-ups were all too willing to adapt their strategy to the stock market context. In order to command a high valuation and ensure survival, what mattered in these days was to build market leadership and grow market share. Obviously, in a winner-takes-all competition, speed is a key success factor. However, unlike a patent race for a new prescription drug say, the leading position in an internet-based market is not determined by a single approval decision by a drug administration agency or patent office, but by the independent decisions of a large number of customers. Their aggregate decision to adopt the service will convey a decisive advantage to an internet company which will generally be highly valuable only if the customers’ switching costs are so high that customers can effectively not be acquired or poached by competitors later on. Differences in the character of network effects could also explain why companies fared vastly differently when adopting seemingly identical strategies with regard to the choice between the possibilities to grow organically or acquiring competing firms. eBay is a fine example of a rather successful acquisition strategy meant to rapidly capture and subsequently consolidate leadership positions in a slate of foreign markets. An intriguing lesson is that it proved impossible for eBay to arrive in a dominant position whenever the company shunned its usual acquisition strategy of taking over an already established local internet auctioneer and instead tried to build its operations from scratch. This was the case in Japan: Yahoo! was able to hold on to its leading market position against a massive attempt by eBay to create a foothold, and eBay subsequently withdrew from the Japanese market. On the other hand, in countries where leadership was attained, the economics of internet-based networks and the associated club e ects played in full in eBay’s favor. Width and depth of the offering attracted more buyers who, in turn, increased the quality and quantity of the goods offered by sellers. At the end of 2003, eBay boasted 95 million users worldwide and 45,000 categories of merchandise sold through its trading platform. More importantly, its business turned out to be very profitable, with operating margins in the range of 30% in 2002 and 2003. In contrast to eBay’s successful acquisition strategy, amazon.com acquired a large number of companies in 1999 and 2000 that apparently were overvalued, and whose rationale for amazon.com appeared to be rather doubtful after the internet shakeout. Tellingly, most of these acquisitions were made when amazon.com’s stock price was at its peak and were paid for in stock or a combination of stock and cash. This suggests that opportunistic market timing may have played as much a role as sound business strategy. Taken together, the different tales of eBay and of amazon.com illustrate the nuanced character of network economics that led to almost opposite strategic effects in different market segments. So while the eventual success of most of the internet survivors was indeed ultimately grounded in network effects, only a fraction of the internet business models that looked promising initially turned out to be sustainable. Financial markets went through a cycle of boom, bust and finally selective rehabilitation of internet stocks between 1998 and 2004, during which the valuation of internet stocks abruptly changed. We suggest viewing this cycle as a collective learning process of market operators grappling to understand and value the- 10 - internet phenomenon 5 . In this interpretation, we distinguish three different periods in stock markets: In the period 1998-2000, we observed extremely high valuations for a very large number of internet start-ups. Two effects can explain these optimistic valuations: on the one hand, there was a strong belief in the natural monopoly value effect, as discussed above, and on the other hand, the financial markets expected high growth. In those times of overoptimism, the former effect was not challenged. Most of the investors believed that the winner-takes-all competition was a game where, on average, players would earn more profits than under traditional competition in spite of the risks associated. In the period 2000-2001, a drastic correction occurred that hit all dot.com and internet-based stocks alike. This adjustment expressed the newfound skepticism that the hoped-for effects that had driven those superstar valuations would not materialize. Growth expectations were sharply revised downwards, and in addition the market accepted the idea that competition would be fiercer than initially anticipated, and that dominant market positions would be harder to sustain. The first effect affected all internet firms more or less with the same magnitude, while the second effect naturally had a stronger impact on e-commerce, telecom, network operators, and ISPs, industries where the stock price drop was even more severe. Finally, starting in late 2002, a rehabilitation occurred for a selective group of surviving internet stocks. In the 2002-2004 period, internet companies that were clear market leaders and managed to convincingly assert their leadership position and, above all, started to generate profits, outperformed the general market. This group includes Amazon.com and Yahoo! stocks. While both companies showed in spring 2004 market capitalizations still far below their peak valuations in 2000/2001, they command high valuation relative to their current financial results and fared much better than NASDAQ and the Dow Jones internet Services index in the period 2003 to early 2004. Interestingly, eBay commands a far higher valuation in 2004 than at its previous peak during the internet bubble. Our interpretation of the speculative internet bubble is based on the idea that market participants overrated the benefits from network effects in internet-based industries. A similar idea is at the heart of a theory model by Scheinkman and Xiong (2003) who show that heterogeneous beliefs among investors, in the form of overconfidence about a private signal, and the presence of short sales constraints alone suffice to generate a speculative bubble. 3. The valuation of internet start-ups During the internet speculative bubble, issues related to the valuation of internet start-ups and appropriate valuation techniques created an intensive debate among finance professionals, which inevitably spread to include academics as well. In this debate, opinions were voiced purporting that traditional valuation techniques would no longer be applicable for internet start-ups, typically based on the argument that (i) these methods would require forecasts of future expected cash flows, while a vast majority of the start-ups posted huge losses and did not even expect to break-even in the medium term; (ii) internet start-ups were deprived of operational and financial histories on which forecasts could credibly be built on. With hindsight however, this debate was clearly driven by an 5 Our implicit theoretical reference is the semi-strong market efficiency hypothesis.- 11 - exaggerated belief in the rationality of stock markets: markets were considered to be infallible in their pricing of internet stocks, market participants were desperately searching for the right valuation models that would justify what later turned out to be a bubble. A short review of ths valuation debate is nonetheless interesting. To put things straight, the valuation of a company rests on the general principle that a firm should be valued on the basis of its future expected cash flows to its investors, primarily equity- and debtholders. There is no valid reason to challenge this general principle. Only the methodology how to forecast future expected cash flows or how to proxy for them can be the subject of legitimate discussion. The two traditional techniques against which objections were raised were, first, the discounted cash flow (DCF) technique that heavily relies on projections of future cash flows that the firm is expected to generate. Second, the technique of earnings multiples that relies on recent historical earnings of the firm or sometimes on consensus forecasts by analysts on these earnings. The DCF technique, in essence, applies the Net Present Value concept to the valuation of a firm, by discounting expected future cash flows at an appropriate rate of return that reflects the underlying non-diversifiable investor risk. The usefulness of the approach in practice depends on the ability to correctly estimate future cash flows. Generally, either these future cash flows are derived from past historical financial data, when the firm is sufficiently mature, or growth rates are built into a model of future cash flows. Given the uncertainty surrounding most of the internet start-ups and the lack of historical data to conduct the prerequisite financial analysis, the practical application of the DCF technique was difficult. Another set of arguments against traditional valuation techniques was based on the idea that internet start-ups are totally different from traditional firms and should therefore be valued with different tools. Their key assets were human capital and brand, not tangible assets. Their value no longer seemed to depend on their ability to generate cash flows from existing assets and existing, well-identified markets, but rather on their ability to capture growth opportunities that had not yet materialized. This is precisely where real options theory appeared to provide significant insights. The surge in interest in real options method during the internet boom was largely driven by the desire to justify astronomical stock prices (see Schwartz and Moon (2000) for an example). However, adding real options techniques to the portfolio of valuation methods is certainly an interesting effort that should be dissociated from such misguided uses. Real options theory, which is based on the tools of financial option models, has become popular in the 1990s (Dixit and Pindyck (1994)) but its use has largely been confined to the valuation of technical natural resources projects (oil, mining). Real options theory has been useful at pointing out the shortcomings in the practical use of traditional valuation techniques, which are implicitly based on the notion of a single, perfect forecast of the firm’s future. DCF valuation results will lead to a systematic downwards bias in the valuation of projects that feature with both a high-level uncertainty and managerial flexibility. However, it is difficult to make use of the option valuation technique, since there are no methods to identify the most valuable real options in a systematic way. Even in a mining firm, which is a single-project firm, the analysis will be made difficult because numerous options are present. In the case of a technology start-up, which typically is a multi-project firm, the valuation of the implicit real options will be even more difficult. Not only is the number of available options multiplied in a multi-project firm, but real options are also more likely to create synergies so that their values interact and create compound options, i.e. options on options. A truly satisfying valuation of real options will be almost intractable analytically.- 12 - These practical difficulties help to explain why, since the internet bubble has burst, the interest in using real options techniques for the valuation of start-ups has largely faded away. Copeland et al. (2000) probably expressed a general sentiment with their dictum that real option values can reasonably be approximated through a traditional decision-tree technique,keeping in mind that the implementation of even that simplified technique is not trivial. For the time being, real options valuation are likely to remain a marginal practice among professional bankers and investors, and the internet boom will have produced little or no lasting change in the practice of valuation in the end. 4. Agency conflicts and the financing of internet start-ups The peculiar competitive environment of the internet boom created new challenges that fundamentally altered the relationship between start-ups and their financial backers. In this section, we investigate the nature of this change and the reaction of venture capitalists, entrepreneurs and investors (VCs). Specifically, we explore the hypothesis that magnified agency problems constituted an essential part of these challenges. Agency costs should not be thought of as a constant tax on business operations, they are likely to get amplified in times of crisis, which is, in our view, what happened during the internet bubble. We will discuss two distinct fault lines of agency conflicts. The first one concerns conflicts between the venture capital firm and the entrepreneur, the second one those between fund providers and fund managers, or between the general partners that run venture capital firms and their limited partners. In any start-up, the entrepreneur or founding team uses funds provided by outsiders, the investors. This raises an agency problem, which is in principle no different from the agency conflicts in large companies and multinationals that are well known since Berle and Means (1932). But the magnitude of the agency conflicts between insiders (the founding team) and outside financiers depends on the ability of the latter to implement appropriate safeguarding mechanisms and to exert appropriate level of monitoring 6 . In large and mature companies, addressing these issues is a matter of establishing a set of governance rules that create an efficient balance of power between managers, and their need for control, and shareholders, who need to make sure funds are not diverted. In smaller businesses and start-ups, financiers will have to exercise monitoring in order to keep potential agency conflicts under control. Innovative activities are in general associated with a high level of uncertainty and contractual incompleteness. For example it might be difficult or almost impossible to define unambiguously what determines success or failure of the project, or to define relevant benchmarks against which the managers’ performance can be evaluated. The likelihood of conflicts between the founding team and the investors is larger than in more mature activities, where unforeseen contingencies can be more easily addressed, consistent with evidence that financial constraints have always been particularly elevated in high-tech industries (Himmelberg and Petersen (1994)). The venture capital industry has long resorted to a wide array of mechanisms to protect itself against such agency conflicts. Those mechanisms typically involve the presence of venture capitalists on the board of directors. They also 6 Too much monitoring may discourage the managers’ investment in firm-specific effort while too little monitoring may induce managers to run the firm to maximise their own private benefits at the expense of shareholders (see Aghion and Tirole (1997)).- 13 - provide the VC firm with the possibility to take effective control of the start-up at any moment, should this be optimal (see Hellmann (1998) and Sahlman (1990) for more details). During the internet boom, these well-known agency problems became further aggravated for the following reasons : (a) The winner-takes-all competition in internet-related industries meant that a massive and prompt flow of funding was required. Most of the expenses were difficult to control through appropriate contracting, e.g. marketing expenses. As a result, it was hard to reign in any tendency to overspend. The founding team was naturally inclined to overinvest, a tendency that is in fact fuelled by overly optimistic views on the expected returns on investment and by a bias towards risk-taking. Both effects are classical agency costs. The former is a typical behavioral bias among entrepreneurs (see Landier and Thesmar (2003) for a theoretical and empirical study), while the latter is rational as long as the entrepreneurs are equityholders protected by limited liability. Winner-takes-all competition can dramatically reinforce this effect: If all competitors behave the same way, this will result in ever higher marketing expenditures all around, and each competitor will be convinced that a high stake in this race is indispensable to marginalize the competition and win the prize. (b) Tangible assets and tangible investments only accounted for a small share of total assets and total investments in most e-commerce start-ups. Typically, the costs related to hardware servers and to the web site’s design and coding – the only expenditure that could be easily verified by an outsider or a court – are limited. On the other hand, a large part of the initial investments and expenses are for items such as marketing expenses, customer base acquisition and retention costs (see Copeland et al. (2000) and Hand (2000)), which are not only intangible but also difficult to benchmark. To illustrate this point, eBay’s marketing and product development expenses have consistently accounted for around 70% of all operating expenses throughout the 1996 to 2003 period 7 . Like in the luxury goods industry, goods and services purchases on the internet were believed to strongly depend on the “cash burn rate” to promote products and brands. This was best exemplified by boo.com (e-commerce) and World Online (ISP), who both disappeared through bankruptcy or acquisition. (c) The relation of bargaining power was turned upside down. Start-up entrepreneurs suddenly were in a strong position because of the massive funds inflow into the industry. Moreover, the start-ups could secure a disproportionate share of rights and cash flows because the bargaining power in the hands employees was considerably increased at the expense of shareholders. In large parts, this was just a consequence of a tight labor market that put qualified and scarce staff in a strong bargaining position. In addition, in internet start-ups, like in other firms whose business model and value is predominantly based on skill and creativity, the most valuable assets were easily its human capital resources, i.e. its employees (Zingales (2000)). As a result, a star employee who threatened to leave the company could extract a large rent from the firm and its investors. Let us briefly discuss how the two probably most important agency conflicts in venture financing were magnified in internet start-ups during the boom years. 7 Source: eBay’s SEC 10K filings. Total operating expenses were calculated excluding amortization of acquired intangibles and merger related costs.- 14 - (1) Bias towards independence. Among the various available exit strategies, venture capitalists generally prefer the most profitable one, with a slight bias towards IPOs, as they are associated with a positive reputation effect. The situation is different for the founding team who systematically prefers the exit strategy that will uphold the start-up’s independence. Indeed, there are private benefits associated with the managing of a company, and the founding team will always prefer to preserve its managing team status. This means that they will tend to avoid a trade sale and favor either an IPO or a management buy-out, and adopt product market strategies that commit the venture to remain independent (see Schwienbacher (2003) for a theoretical model). During the internet boom, this conflict became explosive, as the expectation of large fortunes in the wake of going public, coupled with the general tendency towards premature stock market listings, raised the stakes for entrepreneurs and made them even less neutral in respect to the choice of the exit route. (2) Bias towards continuation. When the founding team is faced with the risk of failure, liquidation of the operations should be considered. However, it is inclined to, ex-ante, favor continuation at any costs and avoid liquidation. This could be (i) because of the entrepreneurs’ excessive optimism, (ii) because of the private benefits associated with the management of a company, or (iii) because the founding team is an equityholder with a wealth constraint and is not asked to contribute to the next stage financing. This effect arises because, as equityholders, the company founders are protected by limited liability against the downside while it can take advantage of the upside. Cornelli and Yosha (2003) show that this bias towards continuation induces “window dressing” strategies by entrepreneurs, i.e. short term results manipulation, so as to favor investors’ next stage financing. Internet start-ups were particularly exposed to the possibility of such short-term manipulations. They had discretionary power over short-term expenses, as much of their expenditures accrued for items like marketing, and they could represent business activity through non-pecuniary measures of output such as unique web site visitors, etc.. The manipulations in revenues by Lernout&Hauspie, a former market leader in voice recognition, or by Comroad, a supplier of navigation systems that invented 96% of its sales, were extreme cases illustrating the potential for abuse. The amplification of agency conflicts between venture capitalists and entrepreneurs had dramatic consequences. On the entrepreneurial side, an internet start-up without new financing from the venture capitalist can survive for an extended period of time by adjusting its cash burn rate, for example by cutting its marketing spending. Bias towards continuation is probably the main reason for the start-ups exhausting all of their funds when possible. As a consequence, although bankruptcy filing began at the beginning of summer 2000, the internet shakeout was a gradual and almost smooth process. On the financier side, one should have expected an increase in contractual provisions and other arrangements safeguarding against the moral hazard risk. The second major fault line of agency conflicts during the internet boom was the one between the ultimate providers of the invested funds, i.e. institutional investors, and the fund managers, i.e. general partners of venture capital funds. First, general partners were all too willing to cooperate with entrepreneurs in nurturing unrealistic ambitions on scale and scope. For example, during the internet bubble, most startups, including the smallest ones, expressed the strategic objective of operating on a European or even a global scale. Similarly, proven management rules, like those insisting on a clear corporate strategy that focuses on core activities and key markets, were ignored, or, at least,- 15 - seemed not to be applicable to internet start-ups. Like amazon.com, a number of multi-market e-commerce sites appeared. Again, a strong belief in the network economics of the internet provided a rationalization for both types of costly and overreaching ambitions. But crucially, general partners in venture funds were probably willing to go along because their incentives were not aligned with those of fund providers. The structure of their compensation contracts, with typically a large (20%) retainer on profits and no penalty in case of losses, meant that they participated in the upside potential, but barely had to suffer the consequences of investment failure. This may not be a big issue in normal times, but faced with the high level of risk that characterized the internet boom, they were willing to take a gamble. Second, why did fund managers accept all this money in spite of what turned out to be a paucity of “good” internet projects? Certainly, as discussed in the previous section, fund managers and investors were overoptimistic and considerably overestimated the quality of their internet projects. But importantly, given the compensation structure of a general partner in a venture fund, who typically received a percentage of managed funds plus a share in the ROI, it was obviously in their interest to accept as many funds as they could. With hindsight, obviously their incentives were no longer appropriate given the market conditions in this particular period. They clearly contributed to agency issues similar to the Free Cash Flow problem described by Jensen (1986): general partners were mostly unwilling to return excess funds to their shareholders even when they could not identify enough sound projects. A compensation structure geared more heavily towards realized returns and including actual penalties for underperformance, with a cut in management fees, would have been more appropriate, together with provisions ensuring the swift return of “excess” funds. The institutional investors’ failure to implement adequate incentives, even though they must have been aware that the unprecedented level of capital inflows created additional risks, was perhaps itself sign of excessive optimism, in this case of the investors themselves as opposed to the entrepreneurs and general partners 8 . The general partners’ inadequate compensation structures probably continued to produce devastating consequences after the internet shakeout unraveled. With most of their investments recorded in the books as ongoing projects, general managers largely had a free hand to window-dress the results of their funds. They could use this discretion to “gamble for resurrection” of distressed portfolio companies by keeping them alive. This probably meant often that good money was thrown after bad in order to mask problems with individual portfolio companies. 5. Incentives and financial policies of internet start-ups As far as the capital structure and the financial policies of internet start-ups are concerned, two salient differences compared with more traditional firms emerge: (i) the wide-spread use of stock-options (ii) the surprising absence of debt financing. (i) Incentives. Internet companies have quickly become champions in offering performancerelated incentives not only to their executives, but also to a wide array of employees. In fact, internet firms, like other high-tech firms, had drastically “democratized” the use of 8 Another argument would be that in the short term, it was rational to direct funds towards internet start-ups to take advantage of the bubble. This explanation, however, cannot address the following question: why did the fund providers not insist on better compensation contracts for fund managers to make sure they would not lose too much with such a strategy later on?- 16 - performance-related pay and stock options. The use of equity-linked compensation was viewed as essential to attract and maintain talent at a time when not only executives with experience in high-tech start-ups, but also programmers, web designers and other staff members had become a scarce human resource. Managers and key employees would often be granted compensation packages that overwhelmingly consisted of stock options and other performance-sensitive securities. Stock as a means of providing incentives was also offered to other groups of “stakeholders”, like suppliers and customers. Incidentally, as far as suppliers were concerned, stock was also a convenient means of payment. To understand the explanations given at the time for this startling development, let us start from the hypothesis that the nexus of contracts holding together a typical internet start-up vastly differs from that of a more traditional firm. In a widely cited survey, Zingales (2000) coined the paradigm of the “new firm” to mark this difference which he summarizes in three points: first, internet start-ups are characterized by much more volatile boundaries than traditional firms; second, the contractual relationships that bind together company and employees are much looser and more temporary in nature, as is evident in the development of markets for temporary specialized workers. Finally, whereas in a traditional firm, the assets are first physical assets and only then intangible assets, for the new firm the overwhelming part of its valuable assets is the human capital of its employees, their knowledge and creativity. In the new firm, the asset value of human capital assets is the “inalienable” property of the employees who can easily walk away. As a consequence, the bargaining power between companies and employees had shifted in favor of the latter, and companies appeared to have no choice but to offer generous incentives in order to create the employee loyalty and motivation that was seen as the essence of their valuable assets. (ii) All-equity financing and absence of dividends. The large majority of internet start-ups during the bubble years relied exclusively on equity financing, and tried to avoid the issuance of debt by all means. 9 During the internet bubble, only a handful of internet companies had issued debt, for example amazon.com, but there are few other examples. The reliance on equity is in stark contrast with the classical financing pattern of start-up firms, where entrepreneurs, prior to the emergence of venture capital, used bank debt as the primary source of financing and switched to equity issuance only after reaching the limit of what they considered a sustainable debt level. Three reasons might explain the exclusive reliance on equity financing during the boom years. First, as we had already seen, internet start-ups placed a high value on financial flexibility. A company without debt obligation has an additional debt capacity to draw on if it wants to finance expansion. Moreover, financial flexibility included also the more defensive objective of avoiding fixed commitments in form of debt repayments that could be in harm’s way in case of adverse contingencies. In fact, the brutal stock market reversal in 2000-2001, as well as the ensuing difficulty for many companies to secure funding and their survival can be viewed as an ex post justification of this instinctive prudence. Second, according to the prominent capital structure theory of Jensen and Meckling (1976), the agency costs of debt, i.e. the incentives to take on risks even if they are value-destroying, is an increasing function of a company’s leverage. Following up on our earlier argument that internet start-ups were already operating in an environment of increased agency problems and 9 Moreover, their debt is rated as junk bonds or contains convertibility features and therefore, is itself a security class with a strong equity-like risk component.- 17 - de facto large risks (both upside and downside), the absence of debt financing ultimately was in the interest of shareholders as well. Finally, according to the more prosaic view expressed in Baker and Wurgler (2002), which has gained prominence in corporate finance since the bursting of the internet bubble, the reliance on all-equity financing expressed merely an effect of shrewd “market timing”. That is, managers tend to issue equity when they perceive equity as overvalued, and they rely on debt issuance when they view the stock price as low relative to its fundamental value. What other episode in recent financial history would be a stronger example than the internet bubble for a clear case of managers perceiving their companies as overvalued, and opportunistically taking advantage of this situation by exclusively relying on equity issuance! Finally, even though this discussion is not fully specific to internet start-ups as such, we discuss the absence of dividend payouts during the internet bubble as one of the more salient financial phenomena during this period. In fact, dividend payouts continuously decreased during the 1980s and 1990s, where many companies, in high-tech and also in traditional sectors, abandoned them in favor of accumulated earnings and share buybacks . . Fama and French (2001) observe that the proportion of listed US companies with dividend payments diminished from 2/3 in the late 1970s to 21% in 1998. What can financial theory tell us to make sense of such strong movements in the payout policy? Traditional theories on payout policy, based on “clientele” effects or tax advantages, seem hardly relevant. It is hard to see how tax clientele would have changed so much that they could explain a seismic shift in dividend payments, and if anything, the tax legislation has become more favorable to dividends than it was twenty years ago. By contrast, two other explanations seem relevant, and they are closely related with the prerogatives of internet companies. First, the payout policies of firms have been influenced by the extensive use of equity-based compensation awards and particularly of stock options. As a consequence, managers and employees had a direct interest in the stock price of their company, which determined the value of their stock option grants. This implied that they clearly preferred stock repurchases over dividend payouts to distribute income to shareholders, and their preference was amplified by the net dilution caused by stock options exercises. Indeed, when employees exercise their stock options, the company has to issue new stock, so that, to avoid any net dilution, many companies launched substantial buyback programs. In fact, in 2000-2001, for the first time, the value of share buybacks of US companies exceeded the value of total dividend payments. About half of the share buybacks were made with the latter goal in mind, i.e. in order to neutralize the dilution effect of newly issued shares in conjunction with employee stock options. Second, profitable high-tech companies also avoided dividend payouts in order to build up cash reserves. This was viewed as strategically important in order to maintain maximum financial flexibility, in view of future investments like mergers and acquisitions. The extraordinary investors’ patience and tolerance regarding the non-payment of dividends is perhaps best understood as the expression of an optimistic belief that the build-up of internal funds reflected future “external” growth opportunities. 6. Initial Public Offerings- 18 - We analyze in a bit more detail how the internet bubble changed the character of IPOs, with the acceleration of the venture capital cycle and initial public offerings prior to breakeven. The internet bubble presented a classic hot issue market for IPOs as Figure 2 shows. The brisk pace of initial public offerings in the years 1998-2000 shows all the typical signs that were previously seen in preceding hot issue markets (Ritter and Welch (2003)): strong volatility associated with unprecedented levels of underpricing i.e. initial returns of the companies taken public surpassed everything seen in earlier hot issue markets (Loughran and Ritter (2004), Demers and Lewellen (2003)), and finally a rather abrupt ending of the issue activity. The massive wave of internet-based stock market listings had also been a key to the initial success of new stock markets that had been created in Europe in the late 1990s to emulate the success of the NASDAQ market in the US. The Nouveau Marché in Paris, Techmark in London, and similar markets in Amsterdam, Brussels and Milan were created to provide an outlet for such stocks. The German Neuer Markt became the most notorious poster-child of this short-lived success. It attracted almost 300 new listings in two years, only to be closed in 2003, after a drop in value of 97% of its leading NMAX index, many scandals and delistings. To understand why public listings have become so popular with internet start-ups, let us begin with the theoretical explanations for the decision to go public. The IPO marks in fact the exit of venture capital providers from the newly listed firm: even if they do not sell all of their stockholdings in the IPO immediately, venture capitalists will normally try to unwind their remaining holdings rather rapidly. Therefore, it also points out the transition from a mode of relationship financing to a form of arm’s length financing. Chemmanur and Fulghieri (1999) and Pagano and Roell (1998), among others, argue that the advantage of relationship financing resides in valuable monitoring services provided by the venture capitalists. The benefit of a public listing is seen in the continuous production of information that the stock market offers through analysts' coverage and the efficient aggregation of information in the stock price. As soon as this informational role, together with the gains in transparency and stock liquidity conveyed by a market listing, outweighs the initial monitoring role, a firm will go public. So according to this view, for internet companies the need for such a continuous stock market update arises early on in their life cycle. Another explanation that seems particularly relevant to internet start-ups emphasizes the advertising effect of an IPO. In this view, an important audience of an IPO is not the financial market, but the product market where the newly listed firm sells its products. Stoughton et al. (2000) propose a model where start-up firms use the IPO as a device to credibly signal to their customers that their products are of high quality, whereas low-quality producers cannot adopt a similar strategy as it would be too costly. One can stipulate that the announcement effect linking IPO and free marketing with an audience of current or potential customers is even more pertinent for internet start-ups: indeed, what a better kind of publicity for a dotcom company than that of its stock market debut with its large media exposure. All the more so if the name of the newly listed firm happens to be identical to the URL of its company website and if the IPO is accompanied by an investors’ frenzy that lead to a first-day return of more than 100%. Some recent empirical studies supply evidence in support of this idea. Blass and Yafeh (2002) show that, among Israeli firms undertaking an IPO, those with a large client base abroad show a clear preference for a listing on NASDAQ while those with a domestic customer base opt for a listing in Tel Aviv. Demers and Lewellen (2003) show that the web traffic of internet companies significantly increases in the month following their IPO, and that the increase is all the more important as the IPO had substantial media exposure and generated a high first-day return.- 19 - In a more behavioral vein, the flurry of accelerated IPO activity can be seen as a rational response of company insiders to a market frenzy that generated a huge demand for internet stocks at hugely inflated prices. According to Baker and Wurgler’s (2002) idea, insiders opportunistically take advantage of market misperceptions, and they tend to issue shares when they are overpriced. Notice that the volume of primary shares issued in IPOs, which, in the US alone, reached a volume of around $65 billion in each 1999 and 2000, about twice the average level of the 1990s, is only one side of the medal. In addition, newly listed companies issued large volumes of equity that were distributed to employees, former shareholders of acquired companies, and other stakeholders. Numerous internet IPOs have chosen to accelerate the decision to go public. Until the late 1990s, it was conceived wisdom that companies could only go public if they had established a solid record of profitability. Compared with this traditional tenet, internet firms have massively opted for early IPO dates. Only 21% of the companies listed in the US in 1999 have been profitable at the moment of their IPO – compared with 68% in the 1983-1998 period – and the IPOs of unprofitable companies generated a first-day return of 71.5% in those years compared to a return of 42.6% for profitable companies (Loughran and Ritter (2004)).- 20 - Figure 2. Total number of IPOs, Share of Internet IPOs and First-day returns 22% 17% 14% 22% 72% 56% 14% 9% 12% 0 100 200 300 400 500 600 700 800 1995 1996 1997 1998 1999 2000 2001 2002 2003 uN b m r e f o PI s O 0% 10% 20% 30% 40% 50% 60% 70% 80% First-day return Total number of IPOs Internet IPOs (as a percentage of total) % First-day return 419 3% 601 3% 416 4% 227 15% 185 60% 200 42% 64 16% 62 8% 57 8% Source : Ritter (2004) and authors’ calculations from data provided by Jay Ritter It is also interesting to relate this hot issue market and its characteristics with the formation and the burst of the speculative bubble as we interpreted it in section 2. Indeed, a common IPO practice that was prevalent during this period was to lock up around 80% of the shares offered at IPO by insiders and other equity holders. As many of the internet stocks were recent initial public offerings, the number of tradeable shares was thus limited, typically for a period of 6 months. After this lock-up period, insiders and other equity holders were no longer constrained by their inability to trade their shares, and this was known to all market participants in advance. Hong, Scheinkman and Xiong (2004) discuss a theoretical model with such features. In their model, the market’s limited risk absorption capacity, created by an artificially small free float of the newly issued equity, explains that prices could be pushed well above fundamental values even though the expiration of lock-up periods was anticipated by market participants. This model offers a rationale both for the very high underpricing observed during the boom period, when free floats were routinely limited to small fractions of the market capitalization, and for the dramatic correction in asset prices afterwards, when the asset float of internet stocks often increased dramatically. We will briefly mention another implication of the premature IPO wave that is directly linked to the likely agency conflicts in internet start-ups that we mentioned earlier. Recall that we had postulated that in internet start-ups, with their high cash burn rates and low average value of tangible assets, managers could easily manipulate accounting returns. In this context, the transition to a less flexible financing mode may become interesting as an incentive tool. Take the example of an ailing internet start-up that has run out of cash and wants to raise a a new round of financing. Its current financiers might be tempted to throw good money after bad, in order to rescue their past investments. At the same time, new investors are probably not inclined to provide funding, given the value of the securities already issued. Then, without an IPO, the company would normally be granted yet another round of cash injection. After an IPO on the other hand, , the presence of many dispersed shareholders can act as a powerful commitment device that prevents new financing from coming forward. In other words,- 21 - relationship financing is ex post efficient, but arm’s length financing can be in certain cases ex ante efficient. This will be the case if the incentives for the entrepreneurial team to finish the project successfully without burning more cash are sufficiently improved by the new, rigid financial contracts. In this view, a positive aspect of the massive wave of premature IPOs for internet start-ups was that it imposed a hard, non-renegotiable budget constraint, which prevented further cash injections once the company is publicly listed. 10 This hypothesis might appear far-flung, but it should be remembered that numerous internet start-ups had little prospects of ever becoming profitable, especially towards the end of the 1998-2000 bubble. Those still held by venture capital funds stood a larger chance to secure additional, and likely wasteful, funding than those that had been taken public before. Conclusion This overview of the “internet bubble” and its aftermaths started out with its unique impact on venture funding. It argued that the European venture capital, in the sense of early-stage financing of innovative companies, was only truly started with the advent of the internet, and the massive flows of funds that came with it. It remains yet to be seen whether early-stage financing in Europe will survive the demise of so many venture-backed internet start-ups. We attributed the large flow of funds to internet companies, the subsequent sharp correction and selective rehabilitation to a learning process about the true nature of internet-based network economics in which winner-takes-all competition was overestimated. These beliefs drove the quest for alternative valuation techniques, based on internet-specific performance metrics and real options techniques, which were largely a bubble-driven, temporary phenomenon. We argued that investors faced heightened agency conflicts, due to the need for flexibility and the perceived imperatives of a winner-takes-all competition in network industries. Institutional investors faced agency costs as well in their relationship with fund managers, in the form of excessive fund allocations and delayed value corrections. Finally, internet start-ups have been at the forefront of a premature IPO market, where for the first time companies were massively listed on the stock market while showing large operating losses. We suggested that premature IPOs might have acted as a commitment device against further cash burn of failing internet companies. 10 Dewatripont-Maskin (1995), among others, present a related formal model in a different context.- 22 - References Baker, Malcolm and Jeffrey Wurgler. 2002. "Market Timing and Capital Structure" J. Finance, 57:1, pp. 1-32. Berle, Adolph A. and Gardiner C. Means. 1932. The Modern Corporation and Private Property. New York: Macmillan. Black, Bernard S. and Ronald J. Gilson. 1998. "Venture Capital and the Structure of Capital Markets : Banks versus Stock Markets" J. Finan. Econ., 47:3, pp. 243-77. Blass, Asher and Yishay Yafeh. 2001. "Vagabond Shoes Longing to Stray: Why Foreign Firms List in the United States" J. Banking Finance, 25:3, pp. 555-72. Bolton, Patrick and David Scharfstein. 1996. "Optimal Debt Structure and the Number of Creditors" J. Polit. Economy, 104:1, pp. 1-25. Chemmanur, Thomas J. and Paolo Fulghieri. 1999. "A Theory of the Going Public Decision" Rev. Finan. Stud., 12:2, p. 249-79. Cochrane, John H. 2004. "The Risk and Return of Venture Capital", Mimeo, University of Chicago, Graduate School of Business. Cooper, Michael J.; Orlin Dimitrov and P. Raghavendra Rau. 2001. "A Rose.Com By Any Other Name" J. Finance, 56, pp. 2371-88. Copeland, Tom; Tim Koller and Jack Murrin. 2000. Valuation. Measuring and Managing the Value of Companies. New York: Wiley, 3rd edition. Cornelli, Francesca and Oved Yosha. 2003. "Stage Financing and the Role of Convertible Securities" Rev. Econ. Stud., 70:1, pp. 1-32. Demers, Elizabeth and Katharina Lewellen. 2003. "The Marketing Role of IPOs: Evidence from Internet Stocks", J. Finan. Econ., 68:3, pp.413-37. Dewatripont, Mathias and Eric Maskin. 1995. "Credit and Efficiency in Centralized and Decentralized Economies" Rev. Econ. Stud., 62:4, pp. 541-65. Dixit, Avinash K. and Robert S. Pyndick R. 1994. Investment Under Uncertainty. Princeton: Princeton University Press. EVCA. 2003. European Venture Capital Association. 2003 Yearbook. Zaventem, Belgium. Fama, Eugene F. and Kenneth French. 2001. "Disappearing Dividends: Changing Firm Characteristics Or Lower Propensity To Pay?" J. Finan. Econ., 60:1, pp. 3-43. Gompers, Paul and Josh Lerner. 1999. The Venture Capital Cycle. Cambridge: MIT Press.- 23 - Hand, John R.M. 2000. "Profits, Losses and the Non-Linear Pricing of Internet Stocks" Mimeo, University of North Carolina, Chapel Hill. Hege, Ulrich; Frédéric Palomino and Armin Schwienbacher. 2003. "Determinants of Venture Capital Performance: Europe and United States" Mimeo, HEC Paris, http://campus.hec.fr/profs/hege. Hong Harrison; José Scheinkman and Wei Xiong. 2004. "Asset Float and Speculative Bubbles" Mimeo, Princeton University. Hellmann, Thomas. 1998. "The Allocation of Control Rights in Venture Capital Contracts" RAND J. Econ., 29:1, pp. 57-76. Himmelberg, Charles P. and Bruce C. Petersen. 1994. "R&D and Internal Finance: A Panel Study of Small Firms in High Tech Industries" Rev. Econ. Statist., 76:1, pp. 38-51. Jensen, Michael C. 1986. "Agency Costs of Free Cash Flow" Amer. Econ. Rev., 76, pp. 323- 29. Landier, Augustin and David Thesmar. 2003. "Financial Contracting with Optimistic Entrepreneurs: Theory and Evidence" Mimeo, University of Chicago. Loughran, Tim and Jay Ritter. 2004 "Why has IPO Underpricing Changed Over Time?" Mimeo, University of Florida. Michelacci, Claudio and Javier Suarez. 2004. "Business Creation and the Stock Market", Rev. Econ. Stud., 71:2, pp. 459-81. Noe, Thomas H. and Geoffrey Parker. 2000 "Winner Take All: Competition, Strategy, and the Structure of Returns in the Internet Economy" Mimeo, Tulane University. NVP. 2000. Nederlandse Vereniging van Participatiemaatschappijen. "Rendementsmeting van Nederlandse Venture Capital Activeiten. Rapport NVP-onderzoek over de Periode 1989- 1998", Mimeo, La Haye. Pagano, Marco and Ailsa Roell. 1998. "The Choice of Stock Ownership Structure: Agency Costs, Monitoring and the Decision to Go Public" Quart. J. Econ., 113:1, pp. 187-225. PricewaterhouseCoopers [2004] Money Tree US Report. Q4 2003 Results. Available at http://www.pwcmoneytree.com. Ritter, Jay R. 2004. "Some Factoids about the 2003 IPO Market" Mimeo, University of Florida, http://bear.cba.ufl.edu/ritter. Ritter, Jay R. and Ivo Welch. 2003 "A Review of IPO Activity, Pricing and Allocations" J. Finance, 57:4, pp. 1795-1828. Rosen, Sherwin. 1981."The Economics of Superstars" Amer. Econ. Rev., 71:5, pp. 845-58. Sahlman, William A. 1990. "The Structure and Governance of Venture Capital Organizations" J. Finan. Econ., 27:2, pp. 879-893.- 24 - Scheinkman, José and Wei Xiong. 2003. "Overconfidence and Speculative Bubbles" J. Polit. Economy, 111:6, pp. 1183-1219. Schwartz, Eduardo S. and Mark Moon. 2000 "Rational Pricing of Internet Companies" Financial Analysts Journal, pp. 62-75. Schwienbacher, Armin. 2003. "Venture Capital Investment, Product Innovation, and Exit Decision" Mimeo, University of Amsterdam. Stoughton, Neal; Kit Pong Wong and Josef Zechner. 2001 "IPOs and Product Quality", J. Bus., 74:3, pp. 375-408. Ueda, Masako. 2002. "Banks versus Venture Capital " CEPR Discussion Paper 3411. Ueda, Masako. 2004. "Banks versus Venture Capital: Project Evaluation, Screening, and Expropriation" J. Finance, 59:2, pp. 601-21. Zingales, Luigi. 2000. "In Search of New Foundations" J. Finance, 55:4, p. 1623-54.