新闻传播学相关参考资料合集：媒体广告策略

MARKET SENTIMENT, MEDIA HYPE AND THE UNDERPRICING OF INITIAL PUBLIC OFFERINGS: THE CASE OF AUSTRALIAN TECHNOLOGY IPOS. BEAUTY HO, MAYA TAHER, ROBERT LEE AND NEIL FARGHER SCHOOL OF ACCOUNTING THE UNIVERSITY OF NEW SOUTH WALES AUGUST 10, 2001 Acknowledgements The authors thank participants in the Current Developments in Financial Accounting Research Seminar at the University of New South Wales, Elizabeth Carson, and Mike Wilkins for their helpful comments. We gratefully acknowledge the provision of data and assistance from William Huang of the Securities Industry Research Centre of the Asia-Pacific (“SIRCA”), and Atika Lenz from JPMorgan. We also acknowledge the research assistance of Wendy Chan. Any errors or omissions remain the responsibility of the authors alone. All data is publicly available. 1 MARKET SENTIMENT, MEDIA HYPE AND THE UNDERPRICING OF INITIAL PUBLIC OFFERINGS: THE CASE OF AUSTRALIAN TECHNOLOGY IPOS. Abstract This study examines the relationship between the extent of initial public offering (IPO) underpricing and market sentiment surrounding technology issues listing on the Australian Stock Exchange (ASX) during 1999 and 2000. We consider ‘hype’ surrounding these issues as reflected in the media and as reflected in the market’s sentiment towards recent offerings by similar firms. We also consider the relationship between technology firms’ need for follow-on offerings due to ‘cash burn’ and the level of underpricing. Finally, we examine the information content of management and accountant going concern warnings as a signalling mechanism to reduce ex ante uncertainty regarding the relative risk of IPO candidates. Our preliminary results indicate that the extent of underpricing is systematically related to variables measuring the market sentiment surrounding the listing of an IPO. Specifically, underpricing is higher following high underpricing in similar recent issues. There is some evidence of higher underpricing for firms with higher media interest and in the period of the hot IPO market prior to April 2000. We find that firms that experience a greater rate of cash burn also experience greater underpricing consistent with the conjecture that such firms are more likely to need additional financing shortly after they go public. The association between cash burn and underpricing is however reliant on inclusion of a few issues with very high underpricing. We also find evidence consistent with warnings in the prospectus regarding going concern issues providing a valuable signal to mitigate investors’ ex ante uncertainty about the value of an offering, thereby reducing the subsequent level of underpricing achieved by that firm. 2 1. INTRODUCTION We examine the economic factors determining the level of underpricing in recent initial public offerings (“IPOs”) by Australian technology firms during 1999 and 2000. The paper contributes to the existing literature by extending the findings of DuCharme et al. (2001) that link unprecedented levels of underpricing achieved by technology firms with measures of stock hype and a firm’s need for follow-on financing. Secondly, we examine the information content of management and accountant warnings regarding going concern issues as a signalling mechanism to mitigate ex ante uncertainty for IPO candidates (Willenborg and McKeown 2000) during a period where regulatory changes facilitated the listing of companies with high levels of inherent risk. The volatile market conditions experienced by so called ‘new economy’ firms in recent years have, on average, culminated in very high levels of underpricing of IPOs. KPMG reports in their 1999 capital markets survey that Australian technology stocks achieved an average level of first day underpricing of 57 per cent to issue prices. Underpricing varied within the converging business sector from a positive return of 236 per cent for LibertyOne to a decline of 45 per cent for Spike Networks. Whether the high levels of valuation for technology stock offerings is regarded as rationally reflecting very high growth options, or is an example of a ‘hot issue market’ (e.g. Ritter 1984) that reflects hypaethral (excessive, open to the sky) expectations for future growth for these firms, the period examined provides a suitable period in which to examine the association between market sentiment and IPO underpricing. In such markets a ‘hype’ develops surrounding the market for new issues. We view ‘hype’ as a measure of the excitement surrounding a pending issue. We use the term ‘hype’ to include excitement generated by: (1) the print and electronic media in relation to an IPO candidate’s listing, and (2) the hype derived from the equity market’s sentiment towards public offerings by technology firms as reflected in recent offerings.1 DuCharme et al.(2001) find that the extent of underpricing is systematically related to greater levels of news exposure for the IPO candidate in a seven-day period prior to the IPO for a sample of US-based internet companies. This paper examines the degree to which DuCharme et al.’s (2001) results for a narrow set of internet stocks may be generalised for a broader sample of Australian firms operating in the so called ‘convergent business sector’.2 Further, we expand the notion of hype as it was operationalised by DuCharme et al. (2001) to include both hype 1 While we measure several sources of ‘hype’ we readily acknowledge that all the characteristics of ‘hype’ are clustered in a ‘hot IPO’ market and can not be regarded as independent factors. 3 generated by the print and electronic media, and hype derived from the equity market’s sentiment towards public offerings by technology firms. Previous research by Willenborg and McKeown (2000) suggests the role of going-concern opinions in reducing information asymmetry, and therefore ex ante uncertainty, for IPO candidates. We extend this research on the information content of audit qualifications issued pre- IPO to an Australian setting where recent amendments to the ASX Listing Rules have tempered the emphasis on profitability as a prerequisite to raising public equity. We also introduce a broader measure of ‘going concern warnings’ beyond an audit qualification designed to incorporate circumstances where investigating accountants may formally approve an entity’s accounts while implicitly questioning the ability of that firm to operate as a going-concern. We examine initial public offerings by technology companies on the ASX during 1999 and 2000. This two-year period includes a period of high growth expectations for technology stocks and a period of diminished expectations following the dramatic reduction in technology share prices that occurred in April 2000. By restricting our analysis to technology industry IPOs we examine variation within a relatively homogeneous sample of firms, but we expect cross- sectional variation in underpricing and market hype surrounding these issues across the period examined. Our preliminary results indicate that the extent of underpricing is systematically related to variables measuring the hype surrounding the listing of an IPO candidate on the ASX. We find that the market sentiment as reflected in the underpricing of recent, comparable IPO candidates is systematically related to the underpricing performance of the current technology offerings. Excluding a few IPOs from the sample, based upon statistical criteria, there is greater underpricing of technology IPO candidates prior to the technology market correction in April 2000. Stock hype as measured by the print and electronic media coverage of IPO candidates in the period preceding their listing on the ASX is only marginally significant in explaining subsequent underpricing achieved upon listing for the full sample. The media coverage is associated with underpricing when extreme observations are excluded based upon statistical criteria. There is evidence that some Australian technology firms that experienced high underpricing also had a greater rate of cash burn, consistent with the conjecture that such firms are more likely to need additional financing shortly after they go public. Finally, we find that ‘going concern warnings’ issued by management or the investigating accountant for an IPO candidate are 2 The ‘convergent business sector’ generally refers to those firms operating in the sectors of 4 valuable in mitigating investors’ ex ante uncertainty about the ‘true’ value of an offering, thereby reducing the subsequent level of underpricing achieved by that firm. This paper is structured as follows. Section 2 discusses background literature with a focus on the theoretical models that have been developed in a US setting to explain underpricing. There is also consideration given to the direction that underpricing research has taken in an Australian setting. Section 3 is concerned with the development of hypotheses. Sample selection, empirical design and the definition and measurement of explanatory variables used to test the hypotheses is considered in Section 4. Section 5 presents the empirical findings. Concluding remarks are offered in Section 6. 2. REVIEW OF PRIOR LITERATURE 2.1 General Background An IPO is the first effort by private firms to raise capital in the public equity market. Many empirical studies have documented that IPOs are typically underpriced, that is, an investor who purchases new issues at the offering price can, on average, make relatively large returns 3. Loughran and Ritter (2000) report that for the period 1990 to 1998, IPO candidates left over US$27 billion of potential IPO proceeds ‘on the table’ because of underpricing. For the same period, the first day returns of IPOs averaged approximately 15 per cent, indicating there is a systematic downward bias in the offer price compared with the price in the secondary trading market. However, recent years have witnessed unprecedented levels of underpricing driven largely by the spectacular stock exchange debuts achieved by firms with Internet-focused business models. Ritter (2001) reports that US$65 billion was left on the table from all IPOs during 1999 and 2000 alone. Further, DuCharme et al. (2001) document that the mean (median) underpricing of 238 Internet IPOs listed in the US during the period 1988 through 1999 was a staggering 113.8 per cent (45.6 per cent). Various theoretical models have been developed to explain underpricing as an equilibrium phenomenon in the IPO market. These include models based on the institutional framework hypothesis (Chalk and Peavy 1986, Finn and Higham 1988, and Taylor and Walter 1990), the litigation hypothesis (Tinic 1988), and the information asymmetry hypothesis, which includes consideration of reputation effects for auditors and underwriters of new issues (Titman and Trueman 1986, Beatty 1989, and Balvers, McDonald and Miller 1988) and the signalling telecommunications, information technology, electronics, multimedia, the Internet, or biotechnology. 3 Ibbotson and Ritter (1993), How (1994) and Loughran et al. (1994) summarise international evidence of IPO underpricing, as well as potential determinants thereof. 5 hypothesis (Grinblatt and Hwang 1989, Welch 1989, Allen and Faulhaber 1989 and How and Low 1993). The phenomenon of underpricing has not received as much attention in an Australian setting by researchers when compared to the significant literature that has emerged in the US. Lee, Taylor and Walter (1996a) find that Australian IPO underpricing varies in a manner consistent with the model of Rock (1986), and the extension of this by Beatty and Ritter (1986). In contrast to Lee, Taylor and Walter (1996a), How (2000) documents that the degree of underpricing is not systematically related to long run returns for mining companies. The underpricing literature in Australia has also examined the robustness of the predictions generated by the signalling model of Datar, Feltham and Hughes (1991). To this end, Lee, Stokes, Taylor and Walter (1999) report a significant positive relation between ex ante proxies for an IPO candidate’s firm-specific risk and the selection of a high quality auditor as an indication to the market of the underlying quality of the IPO candidate. 2.2 Underpricing and Stock Hype Lang and Lundholm (2000) suggest that increasing disclosure can be used to ‘hype the stock’ and thereby reduce the firms’ cost of equity. Lang and Lundholm (2000) examined the disclosure practices and associated stock market responses of 41 small companies in the US around the time of seasoned equity offerings. The authors report significant results in support of their hypothesis that a higher level of disclosure during the period leading up to a stock-offering announcement is associated with higher stock returns. On announcement date, however, Lang and Lundholm found that the market penalises firms that achieved higher levels of disclosure by altering their previous disclosure patterns and they interpret this as evidence that firms ‘hyped’ their stock4. 4 However, the full penalty is not imposed on the announcement date for those firms that actively inflated their stock prices without an economic basis. The stock prices of these companies continue to decline for a period up to 390 days after the equity offering is announced. 6 DuCharme et al. (2001) find that the extent of underpricing for US-based Internet companies is systematically related to greater levels of news exposure for the IPO candidate in a seven-day period prior to the IPO. Bhartov, Mohanram and Seethamraju (2001) find that, even for non-Internet firms, earnings were significant for valuation prior to 1999 but lost their significance as the market turned a blind eye toward financial statement information in 1999. Further, Loughran and Ritter (2000) found that the higher the expected demand for stock prior to listing, the more likely it is that issuers will accept a lower offer price because of the expected appreciation in share price once the stock is listed. When coupled with the unprecedented first day returns achieved in recent years by firms with Internet-focused business models, the findings of Loughran and Ritter (2000) and DuCharme et al. (2001) suggest that stock hype may potentially be playing a significant role in explaining the underpricing in technology IPOs. In this paper we argue that the concept of hype is related to the phenomenon of ‘hot IPO’ markets. It has been well documented that there are pronounced cycles in the number of new issues per month and in the average initial returns per month (e.g. Ibbotson and Jaffe 1975, Ritter 1984, Ibbotson, Ritter and Sinclair 1988, Lowery and Schwert 2001). We use the word ‘hype’ to capture information effects reflecting market sentiment, whether rational or irrational, in the pre- IPO period.5 We examine the extent to which this sentiment influences underpricing during a ‘hot IPO’ period. Another aspect of media coverage and underpricing is the recent evidence by Demers and Lewellen (2001) that underpricing is associated with post-IPO advertising benefits. That is, the underpricing leads to increased exposure and increased website traffic. While we focus on the period prior to the IPO, Demers and Lewellen suggest that any observed underpricing in the finance market can also provide future benefits from increased exposure in the post-IPO period. 2.3 CLERP, Increased Going Concern Risk and Underpricing Substantial reforms to the ASX Listing Rules and the Corporations Law through the Corporations Law Economic Reform Act (“CLERP Act”) were implemented on September 1, 1999 and have significantly improved the regulatory environment for small companies seeking to raise capital from the public equity market in Australia. The prerequisite of profitability that had traditionally prevented loss-making or marginally profitable operations from accessing the ASX 5 Lower and Schwert (2001) find that the cycles in initial returns and high volumes in IPO’s are predominantly driven by information learned during the registration period and argue that this is more consistent with sequential learning. Lower and Schwert measure ‘information’ content with reference to returns to recent similar IPO firms and the market in general, with the potential that these returns reflect a bubble rather than explain information about future prospects. In this study we consider ‘hype’ to include all potential ‘information’ from both news sources and recent market returns. We do not attempt to distinguish between rational or irrational expectations of high future growth. 7 Official Listing has been replaced by a broader series of tests designed to evaluate an entity’s ability to operate as a going concern. An analysis of ASX listings during the 1999 and 2000 hot IPO market reveals that the regulatory amendments did, in fact, facilitate improved access to public equity capital than what was available for SMEs under the earlier listing framework. A total of 69 technology firms secured an official listing on the ASX during 1999. However, 42 firms from this sub-sample (60.9 per cent) secured their listing in the four months following September 1, 1999 when the amended ASX Listing Rules came into effect. A further 87 technology firms were listed on the ASX during 2000. With profitability no longer a disqualifying factor for the achievement of a listing, we anticipate that the average risk profile and level of ex ante uncertainty associated with technology firm IPO candidates since September 1, 1999 is particularly high. This provides a valuable setting in which to examine the effectiveness with which the investigating accountants, auditors and management are able to communicate the inherent risk associated with investing in a technology offering to potential investors. One mechanism for increasing the credibility of management statements regarding performance is the independent accountant’s report contained in the prospectus documents. Willenborg and McKeown (2000) find that for a sample of small U.S. offerings, offering firms with pre-issue going-concern opinions suffer less first-day underpricing than similar securities without going-concern opinions. The Australian Corporations Law effectively prevents companies with a going concern qualification from going public. We therefore consider a broader definition of warnings contained in the prospectus that indicate some need for an investor to specifically consider going concern risk for the IPO candidate firm. 3. DEVELOPMENT OF HYPOTHESES 3.1 Underpricing and Stock Hype One of the objectives of our research is to investigate whether market sentiment surrounding the listing of IPO candidates is a possible explanation of the extent of observed underpricing. As noted above, previous research suggests hype as a measure of the market sentiment and excitement surrounding a pending issue. The interest in an IPO is fostered by investment bankers, managers and potential investors. Hype surrounding an IPO candidate in the period immediately prior to listing assists in fostering a market for the firm’s shares. We consider two types of measures of hype that reflect the level of interest surrounding a pending issue. We consider hype as reflected by exposure in the media and hype as reflected in the market for similar issues. The exposure in the print and electronic media generates awareness among potential investors, which results in an increase in demand for the newly listed stock. This provides 8 relatively more price ramp up on the actual day of offer, which, in turn, induces more underpricing. High underpricing in recent similar offerings can indicate that during the period between the registration of the offer and the initial trading of the issue, investors’ perceptions of the prospects for the firm have changed. If these improved perceptions are maintained then high underpricing will result. Such underpricing is consistent with rational revelation of “information about industry prospects” (Lowry and Schwert 2001), or because the underpricing reflects excess demand for these type of stocks (Loughran and Ritter 2000). Our hypothesis that stock hype created by both the media and underlying market sentiment fuels momentum trading in the stock of an IPO candidate on listing date has received support in the prior literature. DuCharme et al. (2001) find that the extent of underpricing for US-based Internet companies is systematically related to greater levels of news exposure for the IPO candidate in a seven-day period prior to the IPO. Further, Loughran and Ritter (2000) found that the higher the expected demand for stock prior to listing, the more likely it is that issuers will accept a lower offer price because of the expected appreciation in share price once the stock is listed. When coupled with the unprecedented first day returns achieved in recent years by firms with Internet-focused business models, the findings of Loughran and Ritter (2000) and DuCharme et al. (2001) suggest that stock hype may potentially be playing a significant role in explaining the underpricing in technology IPOs. This study extends the findings of DuCharme et al. (2001) in an Australian context where domestic corporate law imposes a seven-day media black-out period pre-listing which prevents systematic attempts by issuers to enhance publicity about a prospective offering by publishing media releases. There are possible real economic actions, other than the IPO, that would reasonably be expected to affect firm value. Therefore we have chosen a 10-day window preceding the ‘media black-out’ period to reduce the likelihood of picking up corporate announcements of economic substance. We expand the notion of hype as it was operationalised by DuCharme et al. (2001) to include hype derived from the Australian public equity market’s sentiment towards public offerings by technology firms. This dimension is motivated by Ritter (2001) who suggests that the secondary market performance of comparable firms is likely to influence investor reaction to the present offering. To summarise, we posit that, at least in the period examined (1999 and 2000), the level of stock hype is positively associated with stock returns. We examine the first day of trade for initial public offerings by technology firms and predict that: 9 H1: The extent of first day underpricing of Australian technology IPOs is positively correlated with the degree of stock hype about an IPO candidate prior to listing. We measure hype along two dimensions. First, we focus on the reflection of hype in the media as measured by the number of mentions a particular IPO candidate receives in the press and electronic media. Secondly, we measure the underlying sentiment of the market as reflected in recent underpricing of similar IPOs. 3.2 Underpricing and the Rate of Cash Burn DuCharme et al. (2001) observe that the establishment of a systematic relationship between stock hype and the underpricing of technology offerings is valuable in its own right, however it does not explain why an issuing firm would leave potential IPO proceeds on the table. To this end, we examine a second hypothesis, consistent with DuCharme et al. (2001), whereby firms anticipating the need for considerable additional financing in the future to fund growth opportunities have an incentive to tolerate IPO underpricing. The decision to go public typically reflects the need for additional equity capital to finance expenditures associated with future growth. Business models in the technology sector during the period we examine were frequently only sustainable to the extent that entrepreneurs were able to return to the capital markets for a follow-on offering as IPO proceeds were consumed (or ‘burnt’). This suggests that entrepreneurs, anticipating the need for considerable additional financing in the future to fund growth opportunities, have an incentive to tolerate greater underpricing so as to leave investors with a sound impression of the firm at the time of going public. We hypothesise that the desire to return to the capital markets, proxied by the rate at which the firm burns through its IPO proceeds on operating activities, is systematically related to the extent of underpricing. This suggests the second hypothesis to be examined. H2: The extent of underpricing of Australian technology IPOs is positively correlated with the rate at which an IPO candidate is expected to burn cash. Explanations of the cross-sectional variation in stock prices for technology firms were prevalent in the literature during the bull market of the late 1990s (e.g. Cohen 1999, Hand 2000a, Hand 2000b). Demers and Lev (2001) find that an Internet firm’s rate of cash burn was a value driver that differed in its pervasiveness as the sector matured and shifted focus towards sustainable business models and bottom (as opposed to top) line financial statement analysis. In 1999, capital markets encouraged aggressive spending behaviour by technology firms to develop the 10 infrastructure required to secure a position of market leadership and to acquire a customer base. However, this perspective was largely reversed in late 2000 as investors realised that ongoing infusions of cash could not underpin the development of a company that lacked a sustainable business model. To the extent that our sample includes post-April 2000 IPO’s the power of our tests to find a relation between underpricing and cash burn will be reduced. 3.3 Underpricing and the Information Content of Going Concern Warnings This hypothesis is motivated by recent research in the US undertaken by Willenborg and McKeown (2000) on the information content of audit qualifications issued pre-IPO. The Australian IPO market provides a different setting in which to examine the extent to which information asymmetries between issuer and investor are mitigated by the signalling effect of audit qualifications, and the extent to which this signalling is incorporated into the relative underpricing of a particular issue. In Rock’s (1986) model, underpricing is required to compensate uninformed investors for the winner’s curse problem. Beatty and Ritter (1986) argue that the level of underpricing is related to the ex ante uncertainty of the aftermarket value of the IPO. IPOs with greater ex ante uncertainty are more difficult to value. To compensate investors for the greater uncertainty, higher risk IPOs have higher initial returns. Therefore, IPOs with higher ex ante uncertainty are more underpriced than those with lower ex ante uncertainty. Empirical support for this monotonic relationship can be found in Ritter (1984), Beatty and Ritter (1986), Wolfe and Cooperman (1990), How, Izan and Monroe (1995), and Lee, Taylor and Walter (1996a). Willenborg and McKeown (2000) extend the model by predicting that a going-concern opinion issued for an IPO candidate has the ability to affect the level of uninformed investors’ ex ante uncertainty about the ‘true’ value of an offering and hence the anticipated level of underpricing. This argument draws on the information content of the audit qualification as a proxy for the volatility or skewness of future returns in the secondary market, rather than as a predictor of investment failure. Consistent with their hypothesis, Willenborg and McKeown (2000) find that securities with going-concern opinions suffer less first-day underpricing than similar securities without going-concern opinions. Notwithstanding the robustness of Rock’s (1986) model of underpricing and the significant results documented by Willenborg and McKeown (2000) in a US setting concentrating on the “small deal” (IPO proceeds up to US$10 million) segment of the IPO market, the direction of our final hypothesis is not obvious. It appears equally reasonable to contend that the issuance of a qualified audit opinion for an IPO candidate is a signal of a poor quality investment opportunity. In this respect, investors would demand a greater risk premium to subscribe to the issue which in 11 turn induces a greater level of underpricing when the firm lists in the secondary market. While this enhanced risk premium may not be fully reflected in a firm’s stock price on the first day of trading, it is reasonable to assume that a proportion of the cost of the premium is borne by the entrepreneur (in the form of underpricing) who, in effect, ‘leaves money on the table’ in order to allow the underwriter to sell the issue. Investigation of audit qualification using Australian IPO firms offers a different perspective to the US market setting since IPO auditors in Australia can not formally ‘qualify’ their attestations as investigating accountants. This is an important regulatory dimension and the extant literature has not explored how the accounting profession reconciles this apparent ‘constraint’ on signaling power with the onus imposed by professional guidelines. In the Australian regulatory environment it is necessary to consider a broader measure of audit qualification designed to incorporate circumstances where investigating accountants, or management, can formally approve an entity’s accounts while implicitly questioning the ability of that firm to operate as a going- concern. In the Australian setting there is an “investigating accountant’s report” contained in the prospectus. Further, the accounting firms can influence management in the preparation of the prospectus to adequately disclose risks of concern to the accounting firm thereby mitigating the need for comment by the accountant or auditor. We refer to this broader measure as the issuance of a going-concern warning. An opinion of this nature can be contained in a number of locations within the company’s prospectus including: the ‘Risk Factors’ section prepared by company directors; or the report issued by the Investigating Accountant. Our hypothesis extends research on the information content of pre-IPO audit qualifications to the Australian context where recent amendments to the ASX Listing Rules have enhanced the average risk profile of a firm raising public equity through an IPO. Consistent with Rock’s (1986) model of underpricing and the results documented by Willenborg and McKeown (2000) in a US setting, we hypothesise that the inclusion of a going concern warning by management or the investigating accountant will reduce the extent of underpricing. H3: The extent of underpricing for Australian technology IPOs is negatively associated with the inclusion in the prospectus of a ‘going-concern’ warning by either the management or the independent accounting firm retained by the IPO candidate. 12 4. Sample Selection and Methodology 4.1 Sample Selection Using the Connect 4 database of company prospectuses, we identify a population of 403 firms that filed for official quotation on the ASX by lodging a prospectus with the Australian Securities and Investment Commission (ASIC) between January 1, 1999 and December 31, 2000. The Connect 4 database classifies firms into year groups on the basis of prospectus filing dates. As a result, our sample excludes those firms that filed for quotation before January 1, 1999 but did not, in fact, list on the ASX until after December 31, 1998. A total of 165 firms filed for quotation in 1999 and a further 238 firms filed for quotation in 2000. We eliminate 27 firms that filed for quotation but did not list on the ASX until after December 31, 2000 because of the need to calculate the standard deviation of post-IPO returns using intra-day return data for all IPO candidates for 180 days after listing. This restriction also ensured that every firm in the final sample had been operational as a public company for at least six months. We also eliminate 36 firms that filed for quotation during the two-year window but later cancelled their IPO by withdrawing their prospectus, as well as 57 firms that lodged a prospectus in relation to a rights issue as part of a seasoned equity offering. [INSERT TABLE 1 HERE] Our study focuses on the phenomenon of underpricing in the context of Australian technology firms. To this end, we limit our sample of IPO candidates to those Australian-based firms that may be classified into technology segments on the basis of the industry definitions developed by the Wall Street Research Net (WSRN). The technology industry segments identified by WSRN are: Search/portals; content/communities; e-tailers; financial services; e-commerce enablers; security; performance software; hardware; high technology manufacturing; Internet services; advertising/marketing; consultants/designers; speed/bandwidth; ISPs/access providers; and biotechnology (including biopharmaceutical). These data screens culminate in a final sample size of 156 Australian technology firms. 4.2 Methodology Model of Underpricing We estimate the following regression model to test our explanations of underpricing for Australian technology firms: 13 Underpricingi = α + β1MEDIA i + β2TECHCRASH i + β3MKTSENTIMENT i + β4BURN i + β5QGC i + β6INSTOWNER i + β7CONTROL i + β8RETAINED i + β9SIGMA_180 i + β10PROSPRISK i + β11AGE i + β12ACCTREPUTE i + β13AUDREPUTE i + β14UWREPUTE i + β15DELAY i + β16LN_E(PROCEEDS) i + β17LN_TA i + β18UNIT i + β191/PRICE i + εi (1) Definition and measurement of the specific variables are discussed below. We first consider three alternative measures of the dependent variable, the level of underpricing. We then consider, in turn, variable measurement for the three hypotheses of interest, and other variables expected to impact the underpricing for Australian technology firms. Definition and measurement of the specific variables are summarised in Table 2. [INSERT TABLE 2 HERE] Underpricing Consistent with the extant Australian literature, we use three distinct measures of the underpricing phenomenon. First, an unadjusted measure of underpricing (Lee et al., 1996a), RawRtn, is calculated as the closing sale price on the first day of listing divided by the subscription price per share, minus unity. Secondly, a continuously compounded measure of underpricing (How et al., 1995), LnRtn, is calculated as the natural logarithm of the closing price of the first day of listing divided by the subscription price per share. This measure reduces the influence of outlying observations and increases the normality of the distribution of returns. Finally, a market index adjusted measure of underpricing (Lee et al., 1996a), MktAdjRtn, is calculated as per the unadjusted underpricing measure (RawRtn) less the New All Ordinaries value on the listing date divided by the New All Ordinaries value on the prospectus registration date, minus unity. These underpricing metrics are calculated using aftermarket data obtained from the Securities Industry Research Centre of the Asia-Pacific (“SIRCA”) for each firm included in the sample. A market index adjusted measure of underpricing is of particular relevance in an Australian capital markets setting where issuance procedures prevent the issuing firm from changing the offer price and (or) the quantity of shares issued pursuant to the prospectus lodged with ASIC. An adjustment to the offer price and (or) quantity may only be made by lodging a supplementary prospectus with ASIC. This represents an important difference from the prevailing US 14 environment, where subscription prices are often not determined until (non-binding) offers have been received from potential subscribers (Hanley, 1993). Stock Hype Drawing from prior discussions, we expect β1 to be positive if stock hype (MEDIA) is associated with underpricing. However, Hypothesis 1 is not solely motivated by a belief that media-induced hype influences the initial pricing of offerings in the secondary market. A further dimension to the hype story incorporates the underlying sentiment of the market per se. To this end, we include an indicator variable for the period before the tech-crash (TECHCRASH)6 and a measure of recent underpricing on previous IPOs (MKTSENTIMENT). We predict that β2 an and β3 will be positive since those firms listing prior to the ‘bursting of the dot com bubble’ are likely to have greater euphoria associated with their listing. We operationalise media-induced hype (MEDIA) by reviewing John Fairfax-owned print publications 7, The Industry Standard (an Australian venture capital and technology magazine and web site8), and AustraliaInternet.com (an Australian web site that publishes technology-specific content and Internet industry commentary9) for articles relating to those firms included in our sample. We select the John Fairfax publications archives because it includes daily newspapers (such as The Sydney Morning Herald and The Australian Financial Review) as well as periodicals (such as Shares magazine, Business Review Weekly and Personal Investor). We select a 10-day window prior to the seven-day period immediately preceding the quotation of the firm on the ASX. This observation window is chosen because Australian corporate law imposes a seven-day ‘media black-out’ period pre-listing, which prevents systematic attempts by issuers to enhance publicity about a prospective offering by publishing media releases. There are possible real economic actions, other than the IPO, that would reasonably be expected to affect firm value. Therefore we have chosen a 10-day window preceding the ‘media black-out’ period to reduce the likelihood of picking up corporate announcements of economic substance. The number of mentions in the media for each IPO range from zero to 35 stories. Because it is unlikely that the information content of news varies directly with the each news story, to reduce measurement error we measured the news intensity on a five point scale based upon the frequency of stories. A value of 1 was assigned for MEDIA to those firms with less than 6 ‘mentions’ in the 6 Refer Demers and Lev (2001) and Keating, Lys and Magee (2001) for discussion on the downturn in technology stocks during this period. Because the decline in the secondary market for technology stocks in April, 2000 is known ex poste the effect of this change is considered further in the sensitivity analysis. 7 http://www.newsstore.f2.com.au 8 http://www.thestandard.com.au 9 http://www.australia.internet.com 15 press; a value of 2 was assigned to firms with between 6 and 10 ‘mentions’; a value of 3 was assigned to firms with between 11 and 15 ‘mentions’; a value of 4 was assigned to firms with between 16 and 20 ‘mentions; and, finally, any firm receiving in excess of 20 ‘mentions’ in the Australian print and electronic media was assigned a scaled value of 5.10 We use two variables, TECHCRASH and MKTSENTIMENT, to reflect the underlying sentiment of the Australian capital markets during the period January 1999 through December 2000. The first variable, TECHCRASH, is operationalised as a time period indicative variable that measures 1 if an IPO candidate listed on the ASX prior to April 17, 2000 and 0 if a firm listed after this date. April 17, 2000 is the date generally recognised by Australian financial market participants as coinciding with the ‘bursting of the dot come bubble’ and a material adjustment in market conditions and perceptions of the sustainability of numerous new economy business models. The TECHCRASH variable therefore incorporates a market anomaly dimension that global equity markets were experiencing prior to April 2000 into our model. The second variable, MKTSENTIMENT is operationalised as the mean level of unadjusted underpricing (RawRtn) achieved by the previous four firms in our sample that listed on the ASX immediately prior to the IPO candidate.11 This variable is motivated by Ritter (2001) who suggests that the secondary market performance of comparable firms that listed prior to the IPO candidate is likely to influence investor reaction to the present offering. Rate of Cash Burn Under Hypothesis 2, we predict that β4 will be negative since the greater (more negative) the rate at which an IPO candidate spends its proceeds from the public offering, the greater the extent of underpricing. 12 Ideally we would like an ex ante measure of expected cash burn. The disclosure requirements for prospectuses of IPO candidates in Australia do not require an issuer to publish forward-looking statements.13 In order to obtain cash burn observations for each firm in our sample we elected to measure burn rates for IPO candidates from an ex post perspective. We define a firm’s burn rate (BURN) as the actual net operating cash flow less interest received on 10 We also measured media intensity using quintiles based upon the frequency of stories. The results are discussed in the sensitivity analysis. 11 To avoid losing any observations from a relatively small sample the second through fourth IPOs have the average for the previous available IPO’s. The first IPO includes an estimate of 49%, based upon a prior NASDAQ listing and the inferences reported are not influenced by reasonable alternative estimates. 12 Only 23 firms (about 15% of our sample) have positive operating cash flows (after deducting interest received on invested funds). Hence, the interpretation more negative the BURN, the greater the underpricing. 13 Notwithstanding this, forward-looking financial statements can be included in prospectuses of Australian companies (Cheung et al., 1999). Only 61 firms (39 per cent) in our sample disclosed directors’ forecasts as to expected cash usage rates for the period subsequent to listing. Further the disclosures are inconsistent across firms both in relation to forecast period and the components of the financial statements that are forecast. 16 invested funds, scaled by the amount of IPO proceeds, for the 12 months subsequent to listing on the ASX. Interest received is deducted from the measure of cash burn in order to eliminate distortions that may arise, for example, when a firm raises a significant amount of capital via an IPO and subsequently invests that money in the short-term money market. In this scenario, the firm may be able to achieve positive net operating cash flow because the interest received from the money market offsets expenditure incurred to operate the firm. We sourced net operating cash flow information for each firm in our sample from InvestorWeb Ltd. This information is compiled on a quarterly basis for a large sample of Australian technology and mining firms. In instances where we did not have four quarters of cash flow information for a firm subsequent to listing, we extrapolated the firm’s net operating cash flow to obtain an annualised figure. Information Content of Management or Accountant Warnings Under Hypothesis 3, we expect β5 to be negative if the prospectus included a going concern warning by the investigating accountant or management (QGC) that reduces informed investors’ ex ante uncertainty about the ‘true’ value of an offering and therefore decreases the level of underpricing. The measure is designed to incorporate circumstances where investigating accountants formally approve an entity’s accounts while implicitly questioning the ability of that firm to operate as a going-concern in the absence of mitigating factors such as ongoing access to finance. We refer to this broader measure as a ‘going concern warning’. We would expect that the placement of such a warning would be the subject of negotiation between management and the investigating accountant. An opinion of this nature can be contained in a number of locations within the company’s prospectus. Review of prospectus documents revealed that the warnings were primarily included in the ‘Risk Factors’ section prepared by company directors or the report issued by the Investigating Accountant (which incorporates an examination of the assumptions underlying forward-looking statements). Examples of these warnings are included as Appendix A. Table 3 provides a summary of the types of warnings and the section of the prospectus where they were found. As can be seen from Table 3 the most frequent type of going concern warning (14.7%) was a broad mention of the need for further capital to be raised in the longer term if the corporate strategy is to be executed. A similar number of warnings pointed to the specific need for financing to secure short-term capital adequacy (14.1%). Such warnings were typically included by management under the title “risk factors” with only 8.3% included in the investigating accountant’s report. Despite the common perception that IPO risk sections include 17 ‘boiler plate’ warnings, relatively few warnings were found given the unproven profitability of these companies. Consistent with the Corporations Law no prospectus was found to contain an audit qualification. 14 [INSERT TABLE 3 HERE] Institutional Ownership and Retained Ownership of IPO Candidate As far as the predicted signs on the control variables are concerned, we predict β6 to be negative if underpricing is perceived as a resolution to an agency phenomenon. That is, issuing firms consent to leave money on the table in order to reward institutional investors for the establishment of mechanisms that facilitate monitoring activities. Stoughton and Zechner (1998) find results suggesting that the value of a firm’s IPO is determined by the ownership structure resulting from the offering mechanism. It is theorised that the reconciliation of this agency problem culminates in a positive relationship between underpricing and strategic rationing in favour of institutional investors. Consistent with this prediction, Koh and Walters (1989) find for their sample of IPOs in Singapore, a negative correlation between underpricing and the proportion of the issue allocated to small investors. Supportive evidence is also reported in Hanley and Wilhelm (1995) for U.S. IPOs. We therefore expect a similar relation to hold for Australian IPOs. We predict that β7 and β8 will be positive if the percentages of voting stock retained by pre- IPO shareholders and controlling equity interests are associated with underpricing. Leland and Pyle (1977) posits that, under conditions of information asymmetry between the issuer and the investor, insiders’ fractional ownership holding of the equity of the firm post-IPO serves as a credible signal of firm value in the secondary market. This model has been examined in an Australian context with mixed results.15 Recent research in the US by Habib and Ljungqvist (2000) and Loughran and Ritter (2001) suggests that the higher the proportion of shares retained by pre-IPO shareholders in the secondary market, the higher the level of underpricing that can be accepted by the issuing firm. We measure the ownership interest retained by pre-IPO shareholders in an IPO candidate at the ‘share allotment’ date. This information was sourced from company prospectuses and cross- checked for consistency with company allotment filings with the ASX. This is to control for the 14 One might raise the issue of how a grossly unprofitable company receives no audit qualification when the prospectus states that the ability to operate as a going concern is dependent upon the ongoing support from controlling entities or on the success of the current IPO or further capital raisings. Discussions with auditors suggest that the potential flow of funds from existing or future equity holders is a mitigating factor in avoiding the necessity to qualify these companies’ financial statements during 1999 and 2000. 15 Lee, Taylor and Walter (1996a) find evidence that industrial IPOs in Australia that have higher retained ownership are significantly more underpriced. How (2000) found a negative but not significant relation between retained ownership and underpricing. 18 effect of secondary trading on the ownership structure of the IPO and post-IPO variability in the number of ordinary shares outstanding. The finance industry measures institutional ownership by the ownership interest held by the top 20 shareholders in a firm. In order to control for circumstances where pre-IPO shareholders retain a material interest in the firm at the ‘share allotment’ date we separate the institutional ownership variable into two components. CONTROL measures the proportion of outstanding shares held by the top three shareholders in an IPO candidate. In contrast, INSTOWNER measures the ownership interest held by the top 20 shareholders. This variable is adjusted to eliminate the CONTROL component. This information is obtained from company allotment filings with the ASX that were sourced from JPMorgan. In each of our empirical models, CONTROL is operationalised as an indicative variable that takes the value of 1 if the proportion of fully paid ordinary shares outstanding of an IPO candidate held by the top three shareholders exceeds 50 per cent, and a value of 0 otherwise. Underwriter and Auditor Reputation Auditor and investigating accountant are included as indicator variables with a value of one if the firm is a Big Five firm and zero otherwise. While these two accountant reputation variables are highly correlated the collinearity does not impact the other coefficients in the model so we include both accountant reputation measures in the results reported. For Australian IPOs How et al. (1995) predict that high quality auditors and investigating accountants should be associated with smaller underpricing. Hence, we predict that β12 and β13 will be negative. By agreeing to underwrite the issue at the set offer price lends credibility to the IPO. Based upon prior literature, we expect that high reputation underwriters underprice less (Beatty and Ritter 1986). Because underwriter reputation may be specific to the technology sector we use the number of IPOs underwritten as the measure of underwriter opinion rather than a more general reputation measure. Size and Risk We draw from Rock’s (1986) model of underpricing and Beatty and Ritter’s (1986) seminal study to argue that β9 and β10 will be positive. This prediction is motivated by a belief that the standard deviation of returns (SIGMA_180) and the number of risks disclosed in the prospectus (PROSPRISK) are appropriate proxies for a firm’s ex ante level of uncertainty and, in combination with firm age (AGE), these variables measure (with error) the degree of risk associated with an IPO candidate independent of the quasi going concern opinion. We predict that β11 will be negative on the basis that less established (or younger) firms tend to have less publicly 19 available information available about them and therefore higher ex ante uncertainty and underpricing (Lee et al., 1996a). Risk is measured by the standard deviation of returns in the first 180 days subsequent to listing (SIGMA_180). We further include the number of risk factors listed in the prospectus (PROSPRISK) and the AGE of the firm as possible further measures of risk. Lee et al. (1996a) use the number of days between prospectus registration and exchange listing as a measure of the level of informed demand. Those issues which close (and therefore list) most rapidly are expected to have the highest level of informed demand. We expect these issues to display relatively larger underpricing and therefore predict that β15 will be positive. Issue size and firm size have frequently been used to proxy investors’ ex ante uncertainty (Beatty and Ritter, 1986). Banz (1981) reports evidence of a ‘size’ anomaly where smaller firms, on average, earn higher risk-adjusted returns than larger firms. He argues that differential information across firms of different sizes, possibly due to informed investors concentrating on larger firms, may be an explanation for this ‘size effect’. We expect larger issuing firms to have more information available about them and therefore lower ex ante uncertainty and underpricing. Empirical support for a negative relationship between underpricing and size can be found in Wolfe and Cooperman (1990) and How et al. (1995). We proxy size by the offer size (LN_E(PROCEEDS)), measured as the natural logarithm of the product of the offer price and the number of shares offered to the public and firm employees, and the natural logarithm of the IPO candidate’s total assets (LN_TA). Drawing on this extant literature, we predict that β16 and β17 will be negative. Unit IPOs combine the issue of shares with warrants or options, which give the holder the right to purchase further shares at a stipulated price within a specified period of time. Lee et al. (2000) report that such arrangements, by setting in advance the price at which subsequent equity will be raised, motivates managers to provide sufficient information to ensure that warrants are exercised if the ‘project’ proves to be viable. We therefore expect that β18 will be negative. Finally, to consider small share-price effects, we take the reciprocal of the IPO offer price per share or per unit. Consistent with Willenborg and McKeown (2000), we predict that β19 will be positive. 20 5. DATA AND RESULTS 5.1 Descriptive Statistics Table 4 reports the summary statistics for our sample of Australian technology firms. We report a mean level of unadjusted underpricing, RawRtn, of 49.73%. This high level of underpricing is consistent with evidence presented from a study of US Internet firms, which documents mean underpricing of a staggering 114% (Ducharme et al, 2001). Previous IPO studies of Australian industrial firms report comparatively moderate levels of unadjusted underpricing (see for example Lee et al., 1996a who document a mean level of underpricing of 16%). The high initial return observed for technology offerings relative to industrial firms is consistent with the notion that technology firms are subject to greater information asymmetry and ex ante uncertainty, on account of their comparatively smaller size16 and relatively ‘unproven’ business models, and thus tend to be underpriced by a greater amount (Lowry and Schwert, 2001). [INSERT TABLE 4 HERE] The median one day return (RawRtn) is 20.5% with standard deviation of 95.61% and minimum and maximum levels of underpricing of –71% and 725% respectively. These statistics indicate the presence of outlying observations within the sample. In a relatively small sample the observations with extremely high underpricing pose a problem of observations with extreme influence. On the other hand, we do not consider it reasonable to exclude highly underpriced firms from an analysis of the influence of hype on underpricing. In order to mitigate the concern regarding the impact of these outliers on the reported results, we report results using the natural log of returns as a second measure of underpricing that reduces the impact of the outliers on the residuals. We also report results for sub-samples that exclude influential observations or exclude extreme underpricing. The mean number of mentions that Australian technology firms received in the print and electronic media during a ten day period prior to listing is 4.6 (scaled to 1.46 when a five point scale is used for media intensity in subsequent analysis). This result is lower than the level of media exposure reported in DuCharme et al. (2001) where Internet firms received, on average, 23 mentions from newspapers and 11 mentions from electronic media. This difference is likely to be attributed to the lower number of news outlets in Australia, and possibly the influence of the pre- IPO ‘media black-out’ period that is operative in Australia immediately prior to the issue. The ten 21 day news period in our study is measured prior to the seven-day pre-IPO media black-out required in Australia. On average, firms in our sample ‘burn’ 34% of their IPO proceeds in undertaking operating activities in the 12 months subsequent to listing. That the mean level of cash generation for technology firms is negative confirms our expectation that sustained negative net operating cash flows are likely to provide a greater incentive for management to raise a follow-on offering to finance ongoing expenditure until the underlying business becomes cash flow positive. Pre-IPO shareholders retain a 48.64% equity interest in the firms in our sample post IPO (RETAINED). This figure is consistent with a retained equity interest of 54.6% reported by Lee et al. (1996a) for Australian industrial firms. The median operating history of Australian technology firms is less than three years. On average, Australia technology firms raised $11.6 Million via IPOs, which is significantly lower than the US$65 Million raised by US Internet firms (DuCharme et al, 2001). The mean total assets of $14.7 million is also significantly lower than the total assets of Australian industrial firms reported in Lee et al (1996a). These figures indicate that the sample firms in our study have less established operations than firms with a comparable operating history from non-technology sectors. This result may be attributed to reforms of ASX Listing Rules, which enabled non-profit making firms to access the public capital markets. The surge in the number of firms listing after September 1, 1999 illustrates that it is likely that the opportunity to go public was increasingly seized by smaller technology firms. These firms clearly perceived raising public equity as a cheaper and more viable source of additional financing. Finally, 60.4% of technology firms have Big 5 member firms as their auditors and 65% retained a Big 5 firm as the Investigating Accountants. The Big 5 have a significantly higher presence in US IPOs with over 93% of US Internet firms electing to retain Big 5 auditors (DuCharme et al, 2001). The Pearson correlation coefficients for selected variables are reported in Table 5. As would be expected, a number of variables are highly correlated. Consistent with our first hypothesis, MEDIA is positively correlated with the level of underpricing indicating that the greater the number of ‘mentions’ an IPO candidate receives in the print and electronic media prior to its listing, the greater the extent of underpricing. The first day return is also positively correlated with the TECHCRASH indicative variable indicating that prior to the ‘bursting of the 16 Lee et al., 1996 report a mean level of total assets of $40.4 Million for a sample of 266 Australian industrial firms. In contrast, we document a mean level of total assets of $14.7 Million. This ‘size effect’ dimension of the underpricing phenomenon is examined in How et al. (1995). 22 dot com bubble’ in April 2000, Australian technology IPOs were more highly underpriced relative to firms that listed on the ASX after the technology sector correction. MKTSENTIMENT is also highly positively correlated with RawRtn at a 1% level of significance. This result suggests that those firms with greater underpricing typically listed on the ASX during a period when the capital markets had reacted favourably to comparable offerings from other firms in the technology sector. The correlation matrix also illustrates that the greater the firm size, the greater the number of mentions (MEDIA) received by that firm in the period preceding its ASX debut. As anticipated, LN_E(PROCEEDS) and LN_TA are highly positively correlated at a 1% level of significance suggesting that larger firms have the ability (and the expected need) to raise larger tranches of equity financing from the public markets. Due to the high correlation between some of the independent variables, there is a need to consider the analysis within a multivariate framework. [INSERT TABLE 5 HERE] 5.2 Multivariate Analysis Table 6 reports the results from three multivariate regressions, which differ in terms of the measurement of the dependent variable. Consistent with the extant Australian literature, we report results using three measures of the first day underpricing, an unadjusted measure of underpricing (RawRtn), log returns (LnRtn), and finally, a market index adjusted measure of underpricing, (MktAdjRtn). [INSERT TABLE 6 HERE] All three specifications explain a substantial portion of the cross-sectional variation in underpricing. The empirical model that uses an unadjusted, raw return as the measure of underpricing shows an adjusted R2 of 33%. Where the first day return is measured as a log return, the empirical model shows an adjusted R2 of 38%. Finally, in Regression 3 where a market index adjusted measure of underpricing is used, we show an adjusted R2 of 34%. A number of results for the control variables are worthy of comment. Consistent with How et al. (1995), we find evidence that firms with larger total assets (LN_TA) are significantly less underpriced. We also document support, at the 5% level of significance, that those firms that raise public equity at a lower subscription price per share achieve a greater amount of underpricing upon listing on the ASX. Some ownership, size and reputation variables are highly correlated reducing the statistical significance of test statistics on some control variables due to 23 the collinearity. We retain all the control variables in the model to provide a more conservative test of the variables of interest. We find a negative and significant relationship between the variance of returns (SIGMA_180) and the level of underpricing. This result is consistent across all three specifications and suggests that those firms with a greater level of risk (as proxied by the ex poste standard deviation of stock returns for the 180 days subsequent to listing) have reduced levels of underpricing. To the extent that SIGMA_180 is a valid proxy for the ex ante level of uncertainty associated with a particular offering, we therefore fail to get results in this instance that are consistent with Rock (1986) and Beatty and Ritter (1986). This may be attributed to the period of our study, the capital markets appear to be ‘punishing’ inherently risky offerings with lower underpricing motivated by a fundamental belief that the potential for loss from this risky setting far exceeds the upside potential from investing in a firm with a significant amount of ex ante uncertainty. Extant literature suggests that reputable auditors and underwriters tend to lessen the amount of underpricing achieved by an IPO candidate since they are construed as a providing a signal of the quality of information to potential investors (How et al., 2000). In contrast, we report that for our sample of technology issues there is a positive and significant relationship, across all three specifications, between the reputation of the underwriter retained by the issuer and the resultant level of underpricing. This suggests that, for our specific sample and time period examined, rather than the reputation of the underwriter reducing uncertainty, it could be that higher reputation underwriters gain the IPOs with the higher upside potential to the investor, and have higher underpricing. Our initial results show some weak support for a significant relationship, between variables measuring the hype surrounding the listing of an IPO candidate on the ASX during this period and the resultant level of underpricing. As expected, the coefficient of the MKTSENTIMENT variable is positive and significant at the one percent level for all three specifications. The mean level of underpricing on recent IPO’s is a strong indicator of issue underpricing. The coefficient on the TECHCRASH indicator variable is marginally significant in explaining the underpricing of Australian technology firms in two of the three specifications and strongly significant in the third specification. The level of media coverage is only somewhat weakly associated with underpricing in the full sample. The coefficient on the MEDIA variable is positive but only significant at the marginal level of ten percent using a one tailed test. As discussed in the sensitivity analysis below, this result is influenced by a few issues with very high underpricing but relatively low media coverage. 24 All three specifications of the empirical model for underpricing indicate that the greater the rate at which an IPO candidate is expected to ‘burn’ cash, the greater the level of underpricing achieved by that firm upon its listing on the ASX (BURN). This result provides support for Hypothesis 2 and suggests that firms anticipating the need for considerable additional financing in the future to fund growth opportunities would have an incentive to tolerate greater underpricing so as to leave investors with a sound impression of the firm at the time of going public. Consistent with Hypothesis 3, our results show a negative and significant relationship, between a ‘going-concern warning’ in relation to an IPO candidate (QGC) and the subsequent level of underpricing achieved by that firm upon listing. Our result indicates that the information content of a going concern warning issued for an IPO candidate has the ability to affect the level of informed investors’ ex ante uncertainty about the ‘true’ value of an offering. This result is consistent with Rock’s (1986) model of underpricing where the level of underpricing is related to the ex ante uncertainty of the aftermarket value of the IPO. The result is also consistent with the results documented by Willenborg and McKeown (2000) regarding the informativeness of pre- IPO going concern qualifications. Further, this is an important finding in an Australia setting where recent amendments to the ASX Listing Rules have enhanced the average risk profile of a firm raising public equity through an IPO. Due to the potential impact of observations with extreme underpricing in this sample we replicate the analysis excluding potentially influential observations. We exclude observations where the studentized residuals are greater than two. This reduces the sample size from 156 to 149 offerings. The results for this analysis are reported as Table 7. [INSERT TABLE 7 HERE] The inferences reported above are somewhat stronger with the exclusion of potentially influential observations. Despite the reduction in sample size, the coefficient on the MEDIA variable is positive and significant at the five percent level when outlying observations are removed. Further investigation suggests that at least three of the extreme underpricing IPOs had significant holdings related to a media owner. These IPOs had less coverage relative to the amount of underpricing. Removing these observations increases the significance of the media variable in explaining underpricing. The coefficient on the variable capturing the period prior to April, 2000 is also positive and significant at the five percent level when outlying observations are removed. The period prior to April, 2000 had a level of underpricing significantly in excess of more typical underpricing in the technology sector as proxied by offerings after April, 2000. All three specifications indicate that underpricing was higher for IPO candidates prior to the ‘tech wreck’ of April 2000 (TECHCRASH). Consistent with the inferences in the unrestricted sample, 25 underpricing performance of Australian technology offerings is positively associated with the hype generated from the secondary market’s treatment of comparable IPO candidates (MKTSENTIMENT). After exclusion of potentially influential observations the results for all three proxies for hype are consistent with the signs predicted by Hypothesis 1. The inferences which relate to Hypothesis 2 and Hypothesis 3 are robust to the exclusion of potentially influential outliers. As predicted, underpricing is higher for firms expected to need future financing (BURN, Hypothesis 2) and lower where some form of going concern warning was included in the prospectus (QGC, Hypothesis 3). 5.3 Further Sensitivity Analysis Due to the impact of extreme observations discussed above we also replicated the analysis excluding IPOs with first day returns in excess of 200%. This reduces the sample size from 156 to 148 firms. Despite the reduction in sample size, the coefficient on the MEDIA variable is positive and significant at the five percent level when extremely underpriced observations are removed. Further investigation suggests that at least three of the extreme underpricing IPOs had significant holdings related to a media owner. These IPOs had less coverage relative to the amount of underpricing. Removing these observations increases the significance of the media variable in explaining underpricing. The support for hypothesis two is sensitive to the removal of outliers. That is, the coefficient on cashburn (BURN) is not significantly different from zero after removing the IPOs with first day returns in excess of 200%. The association between underpricing and the cash burn rate is contingent upon the sample examined, with the results influenced by inclusion of a few IPOs with very high underpricing. The other inferences are not influenced by the exclusion of the extremely highly underpriced offerings. Demers and Lev (2001) provide evidence of a structural change in the valuation of internet stocks around April, 2000. This is of most concern with respect to testing the hypothesis measuring going concern warnings. If the warnings were found to only be significant after the ‘tech wreck’ the interpretation might suggest that these warnings were only issued after investors had lost heavily and that the variable acted as a further proxy for the pre- and post- crash differences in underpricing. We therefore partition our sample into IPOs before and after the ‘tech wreck’. We re-estimate the underpricing model removing the TECHCRASH variable. The results are reported in Table 8. The coefficient on the going concern warning is coefficient is significant in the pre-crash period. The result suggests that going concern warnings have impact on investors perceptions during the ‘hot IPO market’ period. The lack of significance in the post- 26 crash period could potentially be due to the nature of the companies brought to market after April 2000, with more risky companies not proceeding with floats during this post-crash period. [INSERT TABLE 8 HERE] Table 8 also provides some insight into the other effects. The coefficients on MEDIA and market sentiment remains significant at the ten percent level in both the pre-crash and post-crash period. The coefficient on cash burn is significant at an alpha level of less than one percent only in the post-crash period. The lack of significance of variables in this table must be however be interpreted with respect to the relatively small sample sizes of 84 IPOs pre-crash and 65 post- crash. 17 As noted above the number of mentions in the media for each IPO range from zero to 35 stories. It is unlikely that the information content of news varies directly with the each news story. To reduce measurement error we measured the news intensity on a five-point scale based upon the frequency of stories, as defined above. We also measured media intensity using quintiles of news mentions. When measured this way the MEDIA variable is significant at the five percent level in two of the three models reported in table 6, but is not significant using log returns (LNRTN, p=0.1204) or when influential outliers are excluded (Table 7, p=0.1030 using market adjusted returns). Further investigation suggests that when the media mentions are merely divided into quintiles by number of stories the bottom four quintiles include one to six stories. The top quintile includes firms with between 7 and 35 news stories. This scaling does not adequately differentiate the highly newsworthy IPO’s and hence reduces the power of our tests and the statistical significance of the MEDIA variable in this relatively small sample. 6. CONCLUSIONS This study examines the economic factors determining the level of underpricing in recent initial public offerings by Australian technology firms. The market conditions for initial public offerings by so called ‘new economy’ firms in recent years have, on average, culminated in significant levels of underpricing. For our sample of 156 Australian technology firms that listed on the ASX between January 1999 and December 2000 we report that substantial sums of money were left ‘on the table’ on account of severe IPO underpricing. The mean level of unadjusted underpricing achieved by firms in our sample upon listing is 49.7%. Our study extends the findings of DuCharme et al. (2001). We have expanded the notion of hype as it was operationalised by DuCharme et al. (2001) to include both hype generated by the 17 We find no significant differences between the coefficients in each period using F tests to compare coefficients in a combined model allowing coefficients to vary between the pre-crash and post-crash period. 27 print and electronic media in relation to an IPO candidate’s listing on the ASX, and hype derived from the public equity market’s sentiment towards offerings by technology firms. Further, we extend the research of Willenborg and McKeown (2000) who find that audit qualifications provide a mechanism to mitigate ex ante uncertainty regarding IPO candidates. We examine a broader measure of going concern warnings designed to incorporate circumstances where Investigating Accountants may formally approve an entity’s accounts while implicitly questioning the ability of that firm to operate as a going-concern. This is an important dimension in an Australian setting in view of the substantial reforms to the ASX listing rules that have meant that the prerequisite of profitability that had traditionally prevented loss-making or marginally profitable operations from accessing the ASX Official Listing has been replaced by a broader series of tests designed to evaluate an entity’s ability to operate as a going concern. It has therefore become increasingly important to examine the effectiveness with which the investigating accountant and the firm’s directors are able to communicate the inherent risks associated with investing in a technology offering to potential investors. Our results indicate that the extent of underpricing is systematically related to variables measuring the hype surrounding the listing of an IPO candidate on the ASX. Underpricing is positively correlated with underpricing by comparable recent IPOs. There is some evidence that stock hype, measured by the print and electronic media coverage of IPO candidates in the period preceding their listing on the ASX, is associated with subsequent underpricing. There is also some evidence that underpricing is higher during the hot listing market for technology IPO candidates that existed prior to the technology market correction in April 2000. Australian technology firms that experience a greater rate of cash burn also experience greater underpricing consistent with the conjecture that such firms are more likely to need additional financing shortly after they go public. This result is however contingent upon the sample examined, with the results influenced by inclusion of a few IPOs with very high underpricing. Finally, the results indicate that the information content of a going concern warning issued for an IPO candidate is valuable in mitigating investors’ ex ante uncertainty about the ‘true’ value of an offering, thereby reducing the subsequent level of underpricing achieved by that firm. Our analysis has concentrated on explaining first day underpricing of technology firms. It is likely that many of the technology IPOs had poor performance in the longer term. This suggests the need for research on the long run performance of the offerings allowed by the ASX listing on the basis of commitments rather than past profits or the existence of assets. There is also a need for examination of auditor responses to the high risk faced by these businesses, particularly where 28 the prospectuses included warnings of issues regarding the ability to operate as a going concern. Another area that warrants future research is the possible effect of prestige investors on the level of underpricing. Those firms recording the highest three levels of unadjusted underpricing on listing date in our sample are each owned, in part, by investment vehicles affiliated with a high profile investor. Future research could examine the influence of a ‘prestige investor effect’ on IPO candidate underpricing in relatively thin markets such as Australia. 29 REFERENCES Balvers, R.J., B. McDonald and R.E. Miller. 1988. Underpricing of new issues and the choice of auditor as a signal of investment banker reputation. The Accounting Review 43: 605-622. Barry, C.B. 1989. Initial Public Offering Underpricing: The Issuer’s View – A Comment. Journal of Finance 44: 1099-1103. Barzel, Y., M.A. Habib and D.B. Johnsen. 2000. IPO syndicates, private foreknowledge, and the economics of excess search. George Mason University School of Law – Law and Economics Working Paper Series: Working Paper No. 00-29. Beatty, R.P. and J.R. Ritter. 1986. Investment banking, reputation, and underpricing of initial public offerings. Journal of Financial Economics 15: 213-232. Benveniste, L.M. and P.A. Spindt. 1989. How investment bankers determine the offer price and allocation of new issues. Journal of Financial Economics 24: 343-361. Bhartov, E., P. Mohanram and C. Seethamraju. 2001. Valuation of internet stocks - an IPO perspective. forthcoming Journal of Accounting Research April, 2001. Booth, J.R. and L. Chua. 1996. Ownership dispersion, costly information, and IPO underpricing. Journal of Financial Economics 41: 291-310. Botosan, C.A. 2000. Evidence that greater disclosure lowers the cost of equity capital. Journal of Applied Corporate Finance. 12 (4): 60-69. Brennan, M.J. and J. Franks. 1997. Underpricing, ownership and control in the initial public offerings of equity securities in the U.K. Journal of Financial Economics 45: 391-413. Carter, R. and S. Manaster. 1990. Initial public offerings and underwriter reputation. Journal of Finance 45: 1045-1067. Chen, H-S. and J.R. Ritter. 2000. The seven percent solution. Journal of Finance 55: 1105-1131. Clarkson, P.M. and Merkley, J. 1994. Ex ante uncertainty and the underpricing of initial public offerings: Further Canadian evidence. Canadian Journal of Administrative Sciences 11: 54- 66. Cohen, J. 1999. Internet Investing White Paper. Wit Capital (http://www.witcapital.com). Day, G.S. and P.J. Schoemaker. 2000. Wharton on managing emerging technologies. Published by John Wiley & Sons, Inc. Demers, E. and B. Lev. 2001. A rude awakening: Internet shakeout in 2000. University of Rochester Working Paper. January 2001. 30 Demers, E. and K. Lewellen. 2001.The interaction between product and financing markets: evidence on the marketing role of IPOs. University of Rochester Working Paper. May, 2001. DuCharme, L.L., S. Rajgopal and S.E. Sefcik. 2001. Lowballing for “Pop”: The case of Internet IPO underpricing. University of Washington Working Paper. January 2001. Finn, F. J. and R. Higham. 1988. The performance of unseasoned new equity issues-cum-stock exchange listings in Australia. Journal of Banking and Finance 12: 1-19. Hand, J.R.M. 2000. Profits, losses, and the non-linear pricing of Internet stocks. University of North Carolina Working Paper. November 2000. Hand, J.R.M. 2000. The role of accounting fundamentals, web traffic, and supply and demand in the pricing of US Internet stocks. University of North Carolina Working Paper. February 2000. Hanley, K.W. 1993. The underpricing of initial public offerings and the partial adjustment phenomenon. Journal of Financial Economics 34: 231-250. Hanley, K.W. and Willhelm, W.M. 1995. Evidence on the strategic allocation of initial public offerings. Journal of Financial Economics 37: 239-257. How, J.C.Y. 2000. Initial and long run performance of mining IPOs in Australia. Australian Journal of Management. 25: 95. How, J.C.Y., Izan, H.Y. and Monroe, G.S. 1995. Differential information and the underpricing of initial public offerings: Australian evidence. Accounting and Finance. 87-104. Ibbotson, R. G., J. R. Ritter and J. L. Sinclair. 1988. Initial public offerings. Journal of Applied Corporate Finance 1: 37-45. Ibbotson, R. G. and Jaffe, J.F. 1975. ‘Hot issue’ markets. The Journal of Finance 30: 1027-1042. Keating, E. K., T. Z. Lys and R. P. Magee. 2001. The Internet Downturn: Finding internet valuation factors in Spring 2000. Working Paper. Northwestern University. Koh, F. and Walter, T. 1989. A direct test of Rock’s model of the pricing of unseasoned issues. Journal of Financial Economics 23: 251-272. KPMG Corporate Finance’s Annual Survey of Australian Capital Markets, June 1999. Krigman, L., W. H. Shaw and K.L. Womack. 1999. The persistence of IPO mispricing and the predictive power of flipping. Journal of Finance: 1105. Lang, M. H. and R. J. Lundholm. 2000. Voluntary disclosure and equity offerings: reducing information asymmetry and hyping the stock. Contemporary Accounting Research (Winter): 623-663. 31 Lee, P., Stokes, D., Taylor, S. and Walter, T. 1999. The association between audit quality, accounting disclosures and firm-specific risk: Evidence from the Australian IPO market. University of Sydney Working Paper. Lee, P.J., Taylor, S.L. and Walter, T.S. 1996a. Australian IPO pricing in the short and long run. Journal of Banking and Finance 20: 1189-1210. Lee, P.J., Taylor, S.L. and Walter, T.S. 1996b. Expected and realised returns for Singaporean IPOs: Initial and long run analysis. Pacific-Basin Finance Journal. 4: 153-180. Lerner, J. 1994. Venture capitalists and the decision to go public. Journal of Financial Economics 35: 293-316. Loughran, T. and J.R. Ritter. 2000. Why don’t issuers get upset about leaving money on the table in IPOs? University of Notre Dame and University of Florida Working Paper. Loughran, T. and J.R. Ritter. 1995. The new issues puzzle. Journal of Finance 50: 23-49. Lowry, M. and G.W. Schwert. 2001. IPO Market cycles: Bubbles or Sequential Learning?. University of Rochester – The Bradley Policy Research Centre Financial Research and Policy: Working Paper No. FR 00-21. Lowry, M. and G.W. Schwert. 2001. Biases in the IPO pricing process. University of Rochester – The Bradley Policy Research Centre Financial Research and Policy: Working Paper No. FR 01-02. Mello, A.S. and J.E. Parsons. 1998. Going public and the ownership structure of the firm. Journal of Financial Economics 49: 79-109. Ritter, J.R. 1984. The ‘hot issue’ market of 1980. Journal of Business 57: 215-240. Ritter, J.R. 1991. The long-run performance of initial public offerings. Journal of Finance 46: 3- 27. Ritter, J.R. 2001. Flipping made easy: The in-n-out IPO. Red Herring. May 4, 2001. Rock, K. 1986. Why new issues are underpriced. Journal of Financial Economics 15: 187-212. Stoughton, N.M. and J. Zechner. 1998. IPO-mechanisms, monitoring and ownership structure. Journal of Financial Economics 49: 45-77. Willenborg, M. 1999. Empirical analysis of the economic demand for auditing in the initial public offerings market. Journal of Accounting Research. 37 (1): 225-238. Willenborg, M. and J. McKeown. 2000. Going-concern initial public offerings. University of Connecticut and Pennsylvania State University Working Paper. December, 2000. Wolfe, G. A. and E. S. Cooperman 1990. A reassessment of the excess return phenomenon for initial public offerings of common stock, Journal of Applied Business Research 6: 40-50. 32 APPENDIX A: Examples of Going Concern Warnings Example 1: eisa Limited Prospectus lodged 16/7/1999 Investigating accountant: KPMG Investigating Accountant’s Report 13/7/1999 Appendix 4 “The ability of the company to continue as a going concern is currently dependent upon the ongoing support of its shareholders and a major creditor, Edge. There is a formal commitment from Edge to provide ongoing financial support for the lesser of: 12 months from 17 June 1999; and the period from 17 June 1999 to the commencement of trading eisa’s shares on the ASX.” Example 2:Wavenet Ltd Prospectus lodged 18/2/2000 Investigating accountant: PricewaterhouseCoopers Securities Ltd Investigating Accountant’s report 11/2/2000. In the body of the Independent Accountant’s Report: “As set out in Note (a) of Appendix D, the continuing viability of the Company's wholly owned subsidiary Wavenet Technology Pty Ltd to continue as a going concern and to realise its assets and extinguish its liabilities in the normal course of business and at amounts stated in the Unaudited Consolidated Proforma Balance Sheet is dependent upon the Company being successful in raising additional funds from its shareholders or other sources sufficient to enable it to advance funds to Wavenet Technology Pty Ltd in order for its obligations to be met.” In an appendix to the report (Appendix D): Significant Accounting Policies “(a) Continuation as a going concern Wavenet Technology Pty Ltd has experienced operating losses during the years ended 30 June 1998 and 30 June 1999 and the three months ended 30 September 1999. The continuing viability of Wavenet Technology Pty Ltd and its ability to continue as a going concern and to meet its debts as and when they fall due is subject to significant uncertainty and is dependent upon the Company being successful in raising funds from this Prospectus or from other sources sufficient to enable it to advance funds to Wavenet Technology Pty Ltd in order for its obligations to be met.” 33 Table 1: Selection of Sample of Australian Technology Firms Listed During 1999 and 2000 Sample 1999 2000 Total No % No % No % Prospectuses (sourced from Connect 4 database) 1 165 40.9 238 59.1 403 100.0 LESS: Firms that registered prospectus with ASIC but did not list on ASX pre-December 31, 2000 10 37.0 17 63.0 27 100.0 Non-technology firms 2 47 37.0 80 63.0 127 100.0 Withdrawn issues 0 0.0 36 100.0 36 100.0 Rights Issues 33 57.9 24 42.1 57 100.0 Total firms eliminated through data screens 90 36.4 157 63.6 247 100.0 Final sample – Australian technology firms 75 48.1 81 51.9 156 100.0 Notes: 1. Those firms that filed for official quotation on the ASX during the two-year window, January 1, 1999 through December 31, 2000, by lodging prospectus with ASIC. 2. Classified using Wall Street Research Net Decision Framework 34 Table 2: Predicted Relationship Between the Extent of Underpricing and Explanatory Variables Explanatory Variables Represented By Measured As PredictedSign Dependent Variables Unadjusted Measure of Underpricing RawRtn Closing sale price on the first day of listing divided by the subscription per share, minus unity. n/a Continuously Compounded Measure of Underpricing LnRtn The natural logarithm of the closing price of the first day of listing divided by the subscription price per share. n/a Market Index Adjusted Measure of Underpricing MktAdjRtn The unadjusted underpricing (RRtn) less the New All Ordinaries value on the listing date divided by the New All Ordinaries value on the prospectus registration date, minus unity. n/a Independent Variables Stock Hype MEDIA Number of ‘mentions’ sourced from John Fairfax-owned print publications (http://newsstore.f2.com.au), The Industry Standard, and http://www.australia.internet.com. ‘Mentions’ were then scaled to measure media intensity on a five point scale. + TECHCRASH Indicative variable: TECHCRASH = 1 if the firm listed on ASX pre-April 17, 2000 (the date nominated by Australian capital markets as ‘bursting of the dot com bubble’, 0 if the firm listed after this date. + MKTSENTIMENT The mean level of unadjusted underpricing (RRtn) achieved by the four firms that listed on the ASX immediately prior to the IPO candidate. + Actual Rate of Cash Burn BURN Actual net operating cash flow less interest received on invested funds, scaled by the expected amount of IPO proceeds, for the 12 months subsequent to listing on the ASX. 1 - Quasi Going Concern Risk Deduced from Investigating Accountant’s Report and/or Prospectus Risk Factors QGC Indicative variable: QGC = 1 if going concern risk identified in Risk Factors section of prospectus filed with ASIC and/or in the Investigating Accountant’s Report, 0 if no specific identification of the going concern risk. - Control Variables Institutional Ownership Interest at IPO INSTOWNER Proportion of fully paid ordinary shares outstanding of IPO candidate held by shareholders ranked 4 to 20. Shares outstanding was calculated at the ‘allotment’ date for the offering – typically 4 days prior to quotation on ASX + Controlling Shareholder Interest at IPO CONTROL Indicative variable: CONTROL = 1 if proportion of fully paid ordinary shares outstanding of IPO candidate held by top 3 shareholders exceeds 50%, 0 otherwise. Shares outstanding was calculated at the ‘allotment’ date for the offering – typically 4 days prior to quotation on ASX + Retained Ownership Interest at IPO RETAINED Proportion of fully paid ordinary shares retained by pre- IPO shareholders. + Degree of Risk Associated with IPO Candidate SIGMA_180 The standard deviation of the first 180 days’ returns, including the initial return. + PROSPRISK The number of risk factors identified by company directors in the prospectus filed with ASIC. The number of factors was then used to generate a five-point scale. + AGE Length of prior operating history – number of years from incorporation to listing on ASX. - Reputation of Investigating Accountant ACCTREPUTE Indicative variable: ACCTREPUTE = 1 if investigating/independent accountant is a ‘Big 5’ firm, 0 otherwise. - 35 Explanatory Variables Represented By Measured As PredictedSign Reputation of Auditor AUDREPUTE Indicative variable: AUDREPUTE = 1 if auditor (distinct from investigating/independent accountant) of IPO candidate is a ‘Big 5’ firm, 0 otherwise. - Reputation of Underwriter UWREPUTE The number of sample IPOs underwritten by a given underwriter (excluding this IPO). - Time to Listing DELAY The number of days from registration of prospectus with ASIC to the date of listing on ASX. - Issue Size LN_E(PROCEEDS) The natural logarithm of the product of offer price and the number of shares offered to the public (including portion of offer reserved for employees). - Firm Size LN_TA The natural logarithm of the IPO candidate’s total assets as recognised in pro forma financial statements immediately prior to listing on ASX. - Unit IPO UNIT Indicative variable: UNIT = 1 if issue is a unit IPO (for example, options or warrants issued to investor in combination with an allotment of shares), 0 otherwise. - Issue Price 1/PRICE The reciprocal of the IPO offer price per share (or per unit) + 36 Table 3: Sources of Going Concern Warnings for Sample of 156 Australian Technology Initial Public Offerings Type of Going Concern Warning1 Number % Audit qualifications 0 0 Going concern warnings in both the Risk Factors and Investigating Accountant’s Report 7 4.5% Going concern warnings in the Investigating Accountant’s Report only. 6 3.8% Going concern warnings in the Risk Factors Section of Company Prospectus only - Narrowly defined. 22 14.1% Going concern warnings in the Risk Factors Section of Company Prospectus only - Broadly defined. 23 14.7% Subtotal 58 37.2% No going concern risk noted 98 62.8% Total 156 100% Notes: 1. The categorization of warnings is as follows: Type of Going Concern Warning Characteristics / Examples Theoretical Extreme: Audit Qualification AUS 708: Going Concern § Inability to pay debts as and when they fall due § Inability to continue in operation without the necessity or intention to liquidate within 12 months of conclusion of the current reporting period Going concern warnings in the Investigating Accountant’s Report § Short term capital inadequacy (insufficient working capital or operating cash flow) § Dependent upon the ongoing support from controlling entities § Dependent upon the success of the current IPO or the success of further capital raising § Inability to secure additional capital or successfully list on the ASX will prevent management from executing business plan § Going concern warnings in Risk Factors Section of Company Prospectus (Narrowly defined) § Dependent upon the success of the current IPO or the success of further capital raising § Firm will be materially and adversely affected if further capital cannot be obtained on favourable terms § Corporate strategy cannot be executed effectively or business plan objectives have to be changed if further capital cannot be obtained on favourable terms within a 12-month time frame. Going concern warnings in Risk Factors Section of Company Prospectus (Broadly defined) § Corporate strategy cannot be executed effectively or business plan objectives have to be changed if further capital cannot be obtained over the ‘longer term’. 37 Table 4: Descriptive Statistics (n = 156) Median Mean Std. Deviation Minimum Maximum Dependent Variables - First day returns Raw Returns (RawRtn) 0.2050 0.4974 0.9561 -0.7100 7.2500 Log Returns (LnRtn) 0.1865 0.2751 0.4804 -1.2379 2.1102 Market Adjusted Returns (MktAdjRtn) 0.1878 0.4804 0.9549 -0.6342 7.1813 Independent variables Media Mentions 3.0000 4.5833 5.2205 0.0000 35.0000 MEDIA 1.0000 1.4615 0.9115 1.0000 5.0000 TECHCRASH 1.0000 0.5769 0.4956 0.0000 1.0000 MKTSENTIMENT 0.3800 0.4998 0.6376 -0.3375 2.9850 BURN -0.2797 -0.3387 0.7338 -8.2141 0.8162 QGC 0.0000 0.3718 0.4848 0.0000 1.0000 Other sample characteristics INSTOWNER 0.2119 0.2189 0.1100 0.0209 0.5414 CONTROL 0.0000 0.4872 0.5014 0.0000 1.0000 RETAINED 0.4917 0.4864 0.2036 0.0899 0.9223 SIGMA_180 0.0540 0.0565 0.0195 0.0216 0.1244 PROSPRISK 2.0000 2.6218 1.1379 1.0000 5.0000 AGE (years) 2.7849 5.0946 5.5751 0.0986 31.2356 ACCTREPUTE 1.0000 0.6510 0.4783 0.0000 1.0000 AUDREPUTE 1.0000 0.6040 0.4907 0.0000 1.0000 UWREPUTE 3.0000 3.1090 2.7884 0.0000 10.0000 DELAY (days) 45.5000 51.0513 26.9113 14.0000 278.0000 E(PROCEEDS) ($ Millions) 9.2160 11.6000 51.7807 1.5000 396.0000 Market Capitalisation ($ Millions) 32.1589 81.7884 161.7378 2.5200 1268.2500 Total Assets (TA) ($ Millions) 12.8256 14.7000 79.6429 1.1171 876.4150 UNIT 0.0000 0.1544 0.3625 0.0000 1.0000 1/PRICE 2.0000 2.0421 1.3458 0.2273 5.0000 Notes: Sample is 156 Australian technology initial public offerings of ordinary equity made between January 1999 and December 2000. Refer to Table 2 for definitions of variables. Table 5: Pearson Correlation Coefficients for Selected Variables from Australian Technology Firms RRtn MEDIA TECHCRASH MKTSENTIMENT INSTOWNER RETAINED LN_E(PROCEEDS) LN_TA PROSPRISK RawRtn 1.0000 - - - - - - - - MEDIA 0.1488 1.0000 - - - - - - - (0.0637)* TECHCRASH 0.3053 0.1780 1.0000 - - - - - - (0.0001)*** (0.0262)** MKTSENTIMENT 0.4307 0.1540 0.4156 1.0000 - - - - - (0.0000)*** (0.0549)* (0.0000)*** INSTOWNER -0.0594 -0.0646 -0.0436 -0.0023 1.0000 - - - - (0.4613) (0.4228) (0.5887) (0.9774) RETAINED 0.1305 0.1310 -0.1600 -0.0586 -0.5823 1.0000 - - - (0.1045) (0.1030) (0.0461)** (0.4677) (0.0000)*** LN_E(PROCEEDS) 0.0043 0.4606 0.1032 0.0935 -0.0303 0.1565 1.0000 - - (0.9573) (0.0000)*** (0.2000) (0.2457) (0.7073) (0.0510)* LN_TA -0.0575 0.3564 -0.0164 0.0823 0.0026 0.1847 0.8173 1.0000 - (0.4761) (0.0000)*** (0.8390) (0.3070) (0.9746) (0.0210)** (0.0000)*** PROSPRISK 0.1044 0.2565 -0.0110 0.0237 0.2293 0.0144 0.0958 0.1056 1.0000 (0.1948) (0.0012)*** (0.8916) (0.7687) (0.0040)*** (0.8587) (0.2342) (0.1897) Notes: Sample is 156 Australian technology initial public offerings of ordinary equity made between January 1999 and December 2000. Refer to Table 2 for definitions of variables. Numbers in parentheses are p-values. *** Statistic associated with this p-value is significantly different from zero at a 1 per cent level ** Statistic associated with this p-value is significantly different from zero at a 5 per cent level * Statistic associated with this p-value is significantly different from zero at a 10 per cent level Table 6: Regressions examining the Association between Media, Market sentiment and Cash burn with First Day Underpricing for Australian Technology Firms Underpricing a = α + β1MEDIA + β2TECHCRASH + β3MKTSENTIMENT + β4BURN + β5QGC + β6INSTOWNER + β7CONTROL + β8RETAINED + β9SIGMA_180 + β10PROSPRISK + β11AGE + β12ACCTREPUTE + β13AUDREPUTE + β14UWREPUTE + β15DELAY + β16LN_ E(PROCEEDS) + β17LN_TA + β18UNIT + β191/PRICE Predicted Sign RawRtn a (n = 156) LnRtn a (n = 156) MktAdjRtn a (n = 156) Intercept No prediction 1.0679 (0.4938) 1.8214 (0.0170) 0.9770 (0.5294) MEDIA + 0.1245 (0.0697)δ 0.0580 (0.0775)δ 0.1199 (0.0765)δ TECHCRASH + 0.2307 (0.0651)δ 0.2308 (0.0010)** 0.2308 (0.0641)δ MKTSENTIMENT + 0.6313 (0.0000)** 0.3034 (0.0000)** 0.6333 (0.0000)** BURN - -0.2896 (0.0012)** -0.1320 (0.0020)** -0.2925 (0.0010)** QGC - -0.2340 (0.0483)* -0.1031 (0.0651)δ -0.2407 (0.0431)* INSTOWNER + 0.6696 (0.2076) 0.3327 (0.2016) 0.6311 (0.2202) CONTROL + 0.2454 (0.1717) 0.0953 (0.2235) 0.2205 (0.1961) RETAINED + 0.4417 (0.2446) 0.2913 (0.1732) 0.4982 (0.2166) SIGMA_180 + -6.5577 (0.0452) -3.6454 (0.0261) -6.3503 (0.0497) PROSPRISK + 0.0462 (0.2388) 0.0076 (0.4045) 0.0446 (0.2453) AGE - 0.0134 (0.1329) 0.0080 (0.0890) 0.0137 (0.1299) ACCTREPUTE - -0.1658 (0.2005) -0.0269 (0.3892) -0.1800 (0.1798) AUDREPUTE - 0.2526 (0.0923) 0.0976 (0.1445) 0.2621 (0.0833) UWREPUTE - 0.0595 (0.0083) 0.0202 (0.0454) 0.0579 (0.0097) DELAY - -0.0018 (0.2417) -0.0017 (0.0788)δ -0.0020 (0.2137) LN_E(PROCEEDS) - 0.1172 (0.1707) -0.0024 (0.4842) 0.1207 (0.1625) LN_TA - -0.2322 (0.0202)* -0.1253 (0.0113)* -0.2305 (0.0204)* UNIT - -0.1021 (0.3060) -0.0678 (0.2434) -0.1165 (0.2806) 1/PRICE + 0.1426 (0.0111)* 0.0311 (0.1493) 0.1442 (0.0101)* Adjusted R2 33.20% 38.00% 33.66% F-statistic 5.06** 6.00** 5.14** Notes: a. Sample is 156 Australian technology initial public offerings of ordinary equity made between January 1999 and December 2000. Consistent with the extant Australian literature on IPO underpricing, three measures of the dependent variable are used: (1) an unadjusted measure of underpricing (Lee et al., 1996), RawRtn, is calculated as the closing sale price on the first day of listing divided by the subscription price per share, minus unity; (2) a continuously compounded measure of underpricing (How et al., 1995), LnRtn, is the natural logarithm of the closing price of the first day of listing divided by the subscription price per share; and (3) a market index adjusted measure of underpricing (Lee et al., 1996), MktAdjRtn, is calculated as per the unadjusted underpricing measure (RawRtn) less the New All Ordinaries value on the listing date divided by the New All Ordinaries value on the prospectus registration date, minus unity. Refer to Table 2 for definition of variables. Numbers in parentheses are p-values. ** T-statistic associated with this p-value is significantly different from zero at the 1 per cent level (one-tailed t test). * T-statistic associated with this p-value is significantly different from zero at the 5 per cent level (one-tailed t test). δ T-statistic associated with this p-value is significantly different from zero at the 10 per cent level (one-tailed t test). Table 7: Sensitivity Analysis - Regressions Examining First Day Underpricing for a sub-sample of 149 Australian Technology Firms excluding potentially influential observations a Predicted Sign RawRtn b (n = 149) LnRtn b (n = 149) MktAdjRtn b (n = 149) Intercept No prediction 2.7367 (0.0050) 2.3114 (0.0007) 2.6387 (0.0069) MEDIA + 0.0964 (0.0334)* 0.0680 (0.0308)* 0.0915 (0.0413)* TECHCRASH + 0.2565 (0.0033)** 0.2347 (0.0002)** 0.2579 (0.0034)** MKTSENTIMENT + 0.4354 (0.0001)** 0.2819 (0.0001)** 0.4324 (0.0001)** BURN - -0.2962 (0.0001)** -0.1282 (0.0007)** -0.2995 (0.0001)** QGC - -0.1539 (0.0395)* -0.0924 (0.0626)δ -0.1629 (0.0315)* INSTOWNER + -0.4085 (0.2144) -0.0023 (0.4974) -0.4436 (0.1958) CONTROL + -0.0458 (0.3898) 0.0324 (0.3867) -0.0744 (0.3245) RETAINED + 0.4095 (0.3908) 0.2746 (0.1555) 0.4749 (0.1138) SIGMA_180 + -5.7314 (0.1487) -3.952 (0.0086) -5.502 (0.0111) PROSPRISK + 0.0326 (0.2058) 0.0099 (0.0359) 0.0307 (0.1108) AGE - 0.0208 (0.0061) 0.0140 (0.0074) 0.0209 (0.0059) ACCTREPUTE - -0.0299 (0.4038) 0.0054 (0.4748) -0.0425 (0.3649) AUDREPUTE - 0.0667 (0.2849) 0.0554 (0.2475) 0.0750 (0.2619) UWREPUTE - 0.0403 (0.0055) 0.0207 (0.0286) 0.0383 (0.0079) DELAY - -0.0004 (0.3990) -0.0014 (0.0893)δ -0.0006 (0.3403) LN_E(PROCEEDS) - -0.0466 (0.2798) -0.0797 (0.0748)δ -0.0427 (0.2968) LN_TA - -0.1372 (0.0289)* -0.0748 (0.0667)δ -0.1354 (0.0309)* UNIT - -0.0617 (0.3077) -0.0555 (0.2568) -0.0747 (0.2722) 1/PRICE + 0.0368 (0.1826) 0.0235 (0.2013) 0.0383 (0.1736) Adjusted R2 43.03% 40.40% 42.94% F-statistic 6.88** 6.28** 6.86** ** T-statistic associated with this p-value is significantly different from zero at the 1 per cent level (one- tailed t test). * T-statistic associated with this p-value is significantly different from zero at the 5 per cent level (one- tailed t test). δ T-statistic associated with this p-value is significantly different fro m zero at the 10 per cent level (one-tailed t test). Notes: a. Potentially influential observations, where the studentized residual is in excess of two, are removed from this analysis. b. All other definitions are per the previous table. c. Numbers in parentheses are p-values. Table 8: Regressions of IPO Underpricing for Australian Technology Firms Pre and Post April 2000 Market Adjusted Return Day 1 = α + β1MEDIA + β3MKTSENTIMENT + β4BURN + β5QGC + β6INSTOWNER + β7CONTROL + β8RETAINED + β9SIGMA_180 + β10PROSPRISK + β11AGE + β12ACCTREPUTE + β13AUDREPUTE + β14UWREPUTE + β15DELAY + β16LN_ E(PROCEEDS) + β17LN_TA + β18UNIT + β191/PRICE Predicted Sign Pre Crash (n=84) Post Crash (n=65) Intercept No prediction 3.068 (0.0354) 1.808 (0.1406) MEDIA + 0.098 (0.0866)δ 0.166 (0.0414) * MKTSENTIMENT + 0.481 (0.0001)** 0.149 (0.0979)δ BURN - -0.235 (0.1004) -0.288 (0.0001) ** QGC - -0.333 (0.0273)* 0.010 (0.4617) INSTOWNER + -1.140 (0.1025) 1.342 (0.0195)* CONTROL + -0.111 (0.3400) 0.398 (0.0412)* RETAINED + 0.220 (0.3685) 0.532 (0.1453) SIGMA_180 + -3.895 (0.1566) -3.717 (0.9505) PROSPRISK + -0.033 (0.2914) -0.009 (0.4316) AGE - 0.013 (0.1549) 0.020 (0.0356) ACCTREPUTE - -0.066 (0.3715) -0.024 (0.4338) AUDREPUTE - 0.122 (0.2590) 0.043 (0.3814) UWREPUTE - 0.027 (0.1349) 0.042 (0.0201) DELAY - 0.007 (0.0645) -0.003 (0.0165)* LN_E(PROCEEDS) - -0.0242 (0.4285) -0.090 (0.1525) LN_TA - -0.170 (0.077)δ -0.079 (0.1553) UNIT - -0.087 (0.3428) -0.143 (0.1451) 1/PRICE + 0.028 (0.3308) 0.065 (0.1111) Adjusted R2 24.72% 60.39% F-statistic 2.51** 6.42** ** T-statistic associated with this p-value is significantly different from zero at the 1 per cent level (one-tailed t test). * T-statistic associated with this p-value is significantly different from zero at the 5 per cent level (one-tailed t test). δ T-statistic associated with this p-value is significantly different from zero at the 10 per cent level (one-tailed t test).