Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 IPO Value Cécile Carpentier and Jean-Marc Suret In this paper, we exploit a particular Canadian setting where firms enter a main stock exchange using two different paths. The first one is the classical IPO, and the second one is the move from a new or junior stock exchange to the main board. This graduation does not required the firm to provide a prospectus. Both groups of firms should be relatively similar in terms of expectations for growth and liquidity. We use this particular context to study in what extent the IPOs are fairly priced by the investment bankers. After a propensity score matching, we compare the value of comparable firms that list following each path. We show that the IPO provides a higher value, consistent with the hypothesis that the prospectus and the process surrounding the IPO have significant value in reducing the information asymmetry and the risk perceived by the market. The difference in values is economically significant, and cannot be explained by the expected differences in liquidity in the secondary market, by the distribution of IPOs in hot and cold issue markets neither by the investor’s recognition hypothesis. Our findings have implications along several dimensions. For the entrepreneurs, this indicates that the IPO process on the main board has value, despite the underpricing generally observed on the market. For researchers, our results suggest that the price set during the first days of trading could be irrational, because it is set higher than an issue price that is itself higher than the value based on fundamentals. For regulators, at a time of the debate surrounding proportional disclosure and the prospectus (Pritchard 2012), we evidence that the IPO process, including the prospectus, seems to have significant value effect. 1. Introduction A large body of research is devoted to initial public offering (IPO) pricing and underpricing (Boutron et al., 2007; Dissanaike and Amel-Zadeh, 2007; Ljungqvist, 2007; Pukthuanthong-Le, 2008). However, the main question of the fairness of the valuation at IPO time remains unsolved. Worldwide, the initial return is large, albeit a function of the regulatory framework and of the market cycle. It is generally assumed that IPOs are sold below their fair value, for several reasons associated with investment bankers who undervalue the IPOs to create a positive abnormal return during the first transaction ____________ Prof. Cécile Carpentier and Prof. Jean-Marc Suret, Laval University, Québec, Canada, CIRANO Fellows and associate members of the European Center for Corporate Control Studies. Cécile Carpentier: Laval University, Faculty of Administrative Science, School of Accountancy, 6244 Pavilion Palasis-Prince, G1V0A6, Québec, Canada, Phone: (418) 656 2131 # 6385, Fax: (418) 656 7746. Email: Cecile.carpentier@fsa.ulaval.ca, Jean-marc.suret@fsa.ulaval.ca, __________ The authors gratefully acknowledge financial support from the Social Sciences and Humanities Research Council of Canada (SSHRC). 1 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 days (Ljungqvist, 2007). Other studies conclude that IPOs are overvalued (Purnanandam and Swaminathan, 2004) but Zheng (2007) contradicts this conclusion. The overvaluation of IPOs could explain why, on average, the long-run performance following IPOs is generally considered abnormally poor (Baker et al., 2006). Determining the extent to which IPOs are fairly priced is important. If the pricing is fair, then the abnormal returns generally observed during the first trading days on the secondary market should be associated with some kind of irrational exuberance. If it is not, then a credible explanation is needed for why the issuers left so much money on the table. However, assessing the fairness of IPO valuation is a complicated challenge, because it requires determining a firm’s intrinsic value in a particular context. At the IPO time, firms enter a public market and their expected growth and liquidity are not easily comparable to that of firms already listed on the market, or to that of private companies. Finding a benchmark for valuation is thus very complicated (Derrien, 2005; Pukthuanthong-Le, 2008). In this paper, we exploit a particular setting where firms enter a main stock exchange using two different paths. The first one is the classical IPO, and the second one is graduation, i.e. the move from a new or junior stock exchange to the main board. Both groups should be relatively similar in terms of expectations of growth and liquidity. The importance of the prospectus and information therein is a timely research topic in the current context: the equity crowdfunding phenomenon implies that investors should be allowed to buy shares outside the protection of the IPO process; backdoor listings are becoming more frequent; and the willingness of several countries to stimulate the development of junior stock markets will increase the number of firms that list on main boards using graduation. We examine the Canadian market, where a junior stock exchange, the TSX Venture Exchange (TSXV), has become the main provider of new listings for the main board, the TSX Exchange (TSX). We construct a sample of IPOs and graduations, and control for potential endogeneity problems using a propensity score matching method. Using the valuation model suggested by Aggarwal et al. (2009), we study the difference in valuation between IPOs and comparable graduate firms. We show that IPOs are priced higher than similar graduate firms, controlling for variables generally considered value relevant and using a propensity score matched sample. The coefficient of the IPO dummy variable (0.60) indicates an economically significant effect. This difference in values can be linked neither to the differences in expected liquidity on the after-market, nor to a market timing effect. We do not observe a significant influence of the amount raised at the listing time, as suggested by the investor recognition hypothesis. Our results are consistent with the proposition that the prospectus contains more relevant information than the financial statements, which could decrease the risk perceived by investment bankers and investors. Our results support the hypothesis that the growth opportunities associated with IPO issuers are greater than those of graduates, and the hypothesis that investment bankers partially 2 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 consider the results of their marketing efforts in IPO pricing. Overall, our results indicate that the IPO issue price is higher than the price of similar firms graduating to the same exchange during the same period. In Section 2, we review previous studies and outline institutional settings. The conceptual framework and hypotheses are presented in Section 3. Empirical method and data are then discussed, followed by the results. Section 6 concludes the paper. 2. Previous Studies and Institutional Settings 2.1 The conundrum of IPO value A large part of the IPO literature is devoted to explaining underpricing (Ljungqvist, 2007). According to this perspective, IPOs are generally priced below their fair value and underpricing (or initial return) is explained by the price change toward its intrinsic value on the secondary market. Underpricing is often used as a measure of the value effect of the involvement of intermediaries (Chen and Mohan, 2002; Chahine and Filatotchev, 2011), regulatory framework (Hopp and Dreher, 2013; Shi et al., 2013) or informational content of prospectus and accounting data (Boulton et al., 2011; Wyatt, 2013).1 Several explanations for underpricing have been proposed, including the willingness of investment bankers to protect themselves against legal liability resulting from false information in the registration statement against lawsuits, to create an interest in the stock, or to keep uninformed investors involved in the IPO market. The underpricing proposition is confirmed by the observation, based on interviews, that underwriters consciously underprice IPOs, by applying a deliberate discount to estimated free cash flows (Deloof et al., 2009). The use of the initial return as a metric for a real departure from fair value rests on the idea that a stock is fairly priced if it is priced at the same level as that of the first trading day’s closing price (Daily et al., 2003 p.275). However, as underlined by Chemmanur (2010 p.472), for “underpricing to be a meaningful measure (…) one has to make the crucial (and rather strong) assumption that the closing price of a firm’s stock on the first day of secondary market trading is equal to the intrinsic value of that stock.” There is ample evidence that the long-run underperformance following IPOs is generally weak (Jenkinson and Ljungqvist, 2001; Eckbo and Norli, 2005; Gregory et al., 2010). This seems to contradict the hypothesis that the price observed on the secondary market reflects the fair stock value. Price support actions by underwriters, window-dressing by entrepreneurs, irrational factors, and restrictions on short selling in a context of large heterogeneity of expectations can explain the high prices observed on the secondary market (Jenkinson and Ljungqvist, 2001 p.149). According to Cook et al. (2006), the presumption that early aftermarket prices are unbiased estimates of an IPO’s share value is inconsistent with recent findings. 1 Daily et al. (2003) provide a meta-analysis of this stream of research. 3 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 Miller (1977) suggests that the market value of equity reflects the valuation of the most optimistic investors when investors subject to short-sale constraints have heterogeneous prior beliefs. In this setting, stocks will be priced at a premium over their fundamental value. Firms going public are characterized by greater heterogeneity in investor beliefs and tighter short-sale constraints (Chemmanur and Krishnan, 2012). This induces possible overpricing on the secondary market that decreases when new information reduces heterogeneity (L'Her and Suret, 1996). We are left with two prices that presumably reflect the intrinsic value of the firm. Indeed, Derrien (2005) suggests that underwriters price IPOs at a range between the intrinsic value and the price noise traders are ready to pay on the secondary market. IPO stocks are consequently overpriced with respect to their intrinsic value. 2.2 IPO Valuation To solve the controversy surrounding IPO pricing, a small number of researchers have attempted to use valuation models to determine to what extent the price set by investment bankers could be considered fair. However, IPO offer pricing remains relatively unexplored (Daily et al., 2005). A review of this research field is proposed by Pukthuanthong-Le (2008). The results are contradictory. The median IPO was significantly overvalued at the offer price relative to valuations based on industry peer price multiples, according to Purnanandam and Swaminathan (2004). However, Zheng (2007) contradicts this conclusion after correcting two concerns of Purnanandam and Swaminathan’s (2004) study. The first problem is related to the omission of expected growth. The second refers to a scaling problem, associated with the fact that IPO firms will rapidly increase in size. Employing the comparable firm approach and a regression valuation model, Pukthuanthong-Le and Varaiya (2007) show that the median IPO offer value is significantly greater than estimated intrinsic value. Using a sample of 1,655 IPOs, Aggarwal et al. (2009) observe that firms with more negative earnings have higher valuations than firms with less negative earnings, and firms with more positive earnings have higher valuations than firms with less positive earnings. They suggest that negative earnings are a proxy for growth opportunities, which are a significant component of IPO firm value. Daily et al. (2005) examine the relation between several prospectus-based variables (including governance, profitability, CEO-retained equity, age and venture capital backing) and the offer price and underpricing. They find no evidence that these variables are related to either IPO offer price spread or IPO offer price. However, Mousa and Reed (2013) observe a significant association between slack resources, several pieces of prospectus information and high tech IPO value. Previous research on fairness of IPO pricing and relevance of prospectus information produced contradictory results, probably because “evaluating the pricing accuracy of IPO stocks is a hard task because of the difficulty of finding satisfactory comparable seasoned stocks” (Derrien, 2005 p.488). IPOs are likely to have more growth opportunities than seasoned firms, but present particular risk factors. In this paper, we exploit the particular situation that prevails in Canada, where firms can access the main market via two different paths. This allows us to compare an IPO’s value with that of 4 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 similar firms entering the market with potentially similar growth opportunities and risk factors. 2.3 Institutional settings There are two main stock exchanges in Canada: The TSX and the TSXV. From an international standpoint both markets apply low listing requirements. To list on Tier 2 of the TSXV, a firm must meet the following conditions: stock price over CAN$0.15 and post-IPO net tangible assets and market capitalization higher than CAN$500,000. The TSXV has no requirements relative to issuer profitability; it simply stipulates sufficient working capital for 12 months of operations. The minimum market capitalization required for IPOs on the NASDAQ is US$45 million in 2012. In Australia, where listing requirements can be considered permissive, the general admission criteria of the ASX, according to the asset test, require at least AU$2 million in net tangible assets or AU$10 million in market capitalization. As a direct consequence of the low listing requirements, more than 95% of Canadian IPOs from 1986 to 2006 could be considered micro caps according to the SEC criteria. Feng and Yue (2013 p.1943) report that a newly TSXVlisted company is approximately 10 percent of the size of a US penny-stock IPO. Most Canadian IPO issuers are not profitable, and a significant proportion of them report no revenues (Carpentier et al., 2010). The ultimate goal of the TSXV is to graduate its best performers to the main exchange, the TSX (Carpentier et al., 2010). In Canada, from 1992 to 2011, there were 951 graduations, and 562 IPOs on the TSX. The graduation process is easy and its costs are modest. According to the exchange, the issuer information that is on file with the TSXV can be provided directly to the TSX, reduced listing fees apply, and the exchange supports a pre-file meeting. The firms have to provide listing documents, which can be public continuous disclosure documents or prospectuses, including audited financial statements, most recent quarterly statements and the TSX listing application.2 Upon approval of the listing application and documents by the TSX listing committee, trading begins on the main board.3 These graduations provide a benchmark to study IPO valuation. Using graduates as a benchmark for IPO valuation presents the following advantages: both groups of firms list on the same market during the same period, and are likely to be at the same stage of development. Moreover, as evidenced by Carpentier et al. (2010), and unlike IPOs in several studies, graduates do not exhibit any abnormal long-run returns. This indicates that they are on average fairly priced by the market. However, a significant difference exists between the two listing modes. Firms that use an IPO issue new shares to public investors and increase the number of their shareholders, their float and market value significantly. In most cases, graduates do not issue new equity. 2 See: Graduating To Toronto Stock Exchange, available online at: http://www.tmx.com/en/pdf/graduation_info.pdf 3 As an example, Amaya graduates on 1 October 2013. The only filing on the SEDAR electronic system relevant to this operation is a Material Change Notice that indicates that the firm had received final approval from the TSX to graduate from the TSXV and list its securities on the TSX. 5 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 Approximately one-third of graduates close a private placement within one year of listing on the main board. The median amount collected by this subsample of graduates is less than half of the median IPO gross proceeds. These differences should be considered in our model. 3. Conceptual Framework and Hypotheses We consider that IPOs should be priced higher than comparable graduates for several reasons, namely marketing effect, informative content of the IPO process and growth opportunity effect. However, these positive effects on value could be mitigated by the reduction of information asymmetry caused by the previous trading of the graduates’ stocks on the junior market before their listing on the main board. We first test the general hypothesis that there is a difference in value. We then study the extent to which the potential difference can be linked to the liquidity effect, timing effect or market recognition hypothesis. 3.1 The positive value of the IPO process Marketing effect IPOs receive more media coverage than graduations. This is partly explained by the fact that firms that graduate do not sell stock to the public, and by the regulatory IPO setting. IPOs involve investment bankers, whereas graduations do not. Investment bankers play a significant role in IPOs, attributed to a certification or market power hypothesis. Chemmanur and Krishnan (2012, p.570) propose that underwriters are “motivated to obtain high valuations for the equity of firms they take public by making use of their long-term relationships with various participants in the IPO market. These relationships enable them to attract greater participation by these market players (…) leading to higher valuations for the equity of these firms both in the IPO and in the secondary market.” Their empirical tests confirm this hypothesis for prestigious investment bankers. Similarly, Cook et al. (2006) suggest that investment bankers have an “incentive to promote an IPO to induce sentiment investors into the market for it.” Their results show that the promotional efforts of investment bankers influence IPO valuation. Investment bankers can include the potential effect of their marketing efforts in the issue price. Because the investment bankers are not involved in graduations, their market power can explain why the offer prices of IPOs are higher than the prices of graduates. The information content of prospectuses A second perspective states that the prospectus provides a unique set of information not available in financial statements. Wyatt (2013) provides evidence of the valuation relevance of the “use of proceeds” disclosure in the IPO prospectus. Using a sample from 34 countries, Shi et al. (2013) find a significant negative association between IPO underpricing and disclosure regulation, and conclude that more extensive disclosure requirements reduce information asymmetry in IPO markets. Using content analysis, 6 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 Hanley and Hoberg (2009) show that the greater the informative content in the prospectus, the better the pricing accuracy. Carpentier et al. (2012) compare IPOs and a backdoor listing method that implies smoother oversight by securities commissions and a shorter process based on private placements. They find that the choice of the listing method and regulation strictness significantly influences the value of new listings. The set of information available in the prospectus is larger than that provided in financial statements. Moreover, it is partly related to expected growth and future cash flows, whereas the information collected on the junior market is mainly related to past results. Accordingly, we consider that the prospectus and the regulatory process surrounding the IPO, which includes the prospectus visa by securities commissions, is likely to give higher value to IPO firms than to similar firms entering the market by graduation. Signal for expected growth By making an IPO, a firm signals that it has growth opportunities to finance. Conversely, graduation does not involve a similar increase in equity, even if private placements can occur around the listing time. Indeed, Carpentier et al. (2010) observe that firms listed on the TSXV graduate after a strong increase in stock value, which could reflect the exploitation of growth opportunities. During the three-year pre-graduation period, Canadian graduate firms realized a significant equally weighted abnormal return (BHAR) of 206% (p.415). However, Carpentier et al. (2010) do not observe abnormal positive returns following graduation. This indicates that most of the growth of graduate firms occurs before their listing on the main board. If the growth of the graduates is in the past while the growth of the IPO is in the future, then higher values should be observed for IPOs. An opposite perspective: the reduction in information asymmetry When a firm is already listed, uncertainty should be lower because their shares have already been traded for a few years, which implies that their managers’ behavior is known by the market. According to Derrien and Kecskes (2007 p.447), “To reduce valuation uncertainty, a natural solution for firms wishing to raise equity is to proceed in two stages, listing and letting develop a public market in the firm’s existing shares in the first stage, and selling new shares to the public in the second stage. The more active the market that develops in the firm’s existing shares, the greater the reduction of valuation uncertainty, and, in turn, the less the underpricing required when the firm sells new shares.” Following this reasoning, one can expect that graduates will be valued higher than IPOs, because the information asymmetry has been reduced by trades, prices and disclosure on the junior market. However, as suggested by Derrien and Kecskes, trades occurring before the IPO are likely to reduce underpricing and not to provide higher valuation. Derrien and Kecskes show that underpricing is 10% to 30% lower for firms using this two-step process in the UK. Accordingly, we consider that the effect of marketing, information and expected growth on the issue price will dominate that of reduction of information asymmetry. Hence the following hypothesis: H1: With all things being equal, the value of an IPO firm should be higher than the value of a firm that graduates from the junior market to the main board. 7 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 3.2 Liquidity effect An IPO implies selling stock to a large number of investors. Such an operation might create a liquid market, which positively influences underpricing and survival, as evidenced by Ellul and Pagano (2006, p. 381). This is not the case for graduation. When the graduate issues new stocks after listing on the main board, it uses private placements that cannot increase liquidity. Vismara et al. (2012) discuss the reasons why many of the European second markets have failed. They show that firms in these markets raise only a few million euros and rarely develop liquid trading. Investment bankers can incorporate the expected positive effect of the liquidity in the issue price. This can explain why Ellul and Pagano observe a negative relation between expected liquidity and underpricing. Thus we expect that liquidity on the aftermarket positively influences issue prices, and explains the difference between IPO and graduation values. However, as underlined by Ellul and Pagano (2006), investors do not know precisely how liquid the aftermarket will be, and so we should take expected liquidity into account, as follows: H2: With all things being equal, the value of an IPO firm should be higher than the value of a firm that graduates from the junior market to the main board, when expected liquidity is accounted for. Timing effect It is well known that cycles in the IPO market influence the type of issues, the volume of issues, underpricing and the fate of IPOs (He, 2007). Hot and cold issue periods are likely to influence the issue price. Helwege and Liang (2004) evidence that IPOs in hot periods raise more money than IPOs in other periods, and that IPOs in hot periods have higher market-to-book ratios. He (2007) shows that IPOs in hot periods result in higher share prices. Managers can choose the IPO time to benefit from market sentiment and get higher values. It is thus important to determine to what extent the hot-cold dichotomy explains the differences in valuation. H3: With all things being equal, the value of an IPO firm should be higher than the value of a firm that graduates from the junior market to the main board, when market sentiment is accounted for. Investor recognition effect The investor recognition hypothesis proposes that raising more financing leads to greater investor recognition, i.e., greater investor awareness of the firm (Kecskes, 2008). Indeed, Kecskes observes that firms collecting larger amounts of money at IPO are worth more than those collecting smaller amounts, and shows that raising more financing leads to greater investor recognition and thereby increases IPO firm value. Greater investor recognition also increases firm value for seasoned equity offerings (Autore and Kovacs, 2014). Proxies for investor recognition are better underwriting services, greater liquidity, and increase in analyst following. A difference in equity value between IPO firms and graduates can be a consequence of greater investor recognition 8 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 of IPOs than of graduations. Coverage by financial analysts is a classical estimator of investor recognition, but a very small proportion of IPOs and graduates are actually followed by analysts, because of their small size. Following Kecskes (2008), who posits that investor recognition is caused by the raising of more financing, we use issue size as a proxy. Accordingly, we test the hypothesis that the differences between IPOs and graduations can be explained by differences in the relative size of the issues surrounding the listing. H4: The difference in value between IPO firms and graduates is explained by the differences in the amounts of financing raised around the listing time. 4. Empirical Method and Data 4.1 Empirical model In this paper, we consider neither the market price for IPOs nor underpricing. We focus on the price set by the investment banker, although we use the graduation’s market price as a benchmark comparison for graduation.4 First, our main hypothesis is that IPO provides higher value than graduation. Using issue prices that are on average lower than market prices, we biased the results against our hypothesis. Second, we are interested in the price and amount that the firm can obtain through an IPO. Third, we consider that the effect of the demand for the stock is partially incorporated in the issue price by the investment banker, as suggested by Derrien (2005). Analyzing valuation at IPO or graduation is difficult because of the numerous firms reporting no sales, negative earnings or negative shareholders’ equity. In these cases, it is impossible to use Q ratios (King and Segal, 2009) or multiples (Purnanandam and Swaminathan, 2004) to estimate value. In this study, we use the model proposed by Aggarwal et al. (2009), where the dependent variable is the total offer value defined as the offer price multiplied by the post-IPO shares outstanding. This empirical model allows detection of cases of negative earnings, a situation that prevails in Canada, for both IPOs and graduations. Following Hand (2003) and Aggarwal et al. (2009), we use the Log transformation for each continuous variable as follows: L(W) = loge (1+W) when W ≥ 0 in $ million; and L(W) = -loge (1-W) when W < 0 in $ million. We consider that the value at the listing time on the main market is explained by the three main valuation drivers: revenues, book value of equity and earnings. We use earnings before extraordinary items following Aggarwal et al. (2009). Stock prices are 4 Analysis of the IPO market price and underpricing could enrich this paper by providing a path for analyzing the sources and consequences of reduction in information asymmetry. However, this implies consideration of another stream of conceptual models and introduction of a complementary set of variables. Overall, the reduction of the heterogeneity provided by listing on the venture market is likely to be commensurate with the length of the listing and with the trading volume. Moreover, underpricing cannot be estimated for graduation. This dimension of the research has been left for further studies. 9 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 partially determined by growth opportunities; this is particularly true in the situation of new ventures at the IPO stage. Classical indicators of growth opportunities like Q ratios or Book-to-Market ratio cannot be introduced in a model explaining price. We follow Aggarwal et al. (2009), who include a dummy high-tech as a proxy for growth opportunities. Prestigious auditors are associated with high market valuations (Firth and Liau-Tan, 1998; Aggarwal et al., 2009). In the IPO case, the choice of a prestigious investment banker has generally been considered a positive signal (Demers and Joos, 2007), characterized by lower underpricing and better long-run performance, associated with a decrease in information asymmetry (Carter and Manaster, 1990). Chemmanur and Krishnan (2012) evidence that market power explains the higher valuation observed for prestigious investment banker-backed IPOs. However, similar to a venture capitalbacked dummy, this variable is not available for graduations and should be omitted from the model. Because graduations do not involve changes in ownership, we do not consider ownership retention by pre-IPO shareholders, as in Aggarwal et al. (2009). The model takes the following form: L(OV)i = α1+ α2 L(INC)i + α3 L(BV)i + α4 L(Sales)i + α5 DHTi + α6 DIPOi + α7PAUDITi + ei (1) where L(W) stands for the L transformation and OV means Overall Value and is measured as follows: for IPOs, OV = Offer price x number of shares outstanding immediately after the IPO (in $ million). This value is equivalent to market value based on the issue price. For graduation, OV is the market capitalization at the date of listing on the main market, because there is no real offering in this case. Private placements that generally occur after the listing are not included in this value.5 Inc: Net Income before extraordinary items at fiscal year-end closed before IPO or graduation BV: book value of equity at fiscal year-end closed before IPO or graduation Sales: Revenues reported at the fiscal year-end closed before IPO or graduation DHT: dummy variable set to 1 if the industry of the issuer is a high-tech company, and 0 otherwise. DIPO: dummy variable set to 1 if the company makes an IPO and 0 otherwise (if the company graduates). Paudit is a dummy variable set to 1 if the auditor before the IPO or graduation is prestigious (‘‘Big 5’’ or ‘‘Big 4’’), and 0 otherwise. 5 For IPOs, the overall value is equivalent to the offer value defined by Aggarwal et al. (2009). 10 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 Our main explanatory variable is DIPO. To test our second and third hypotheses, we also include the following explanatory variables: Age: Age of the company, in years, at the IPO or graduation L(exp. liq.): Log of the expected rotation (trade amount/market value of a size-sector matched IPO or graduation 12 months after the matched IPO) Dhot: dummy variable that has a value of 1 if the corresponding month is in the upper (lower) third of the moving average distribution of IPOs. Dcold: dummy variable that has a value of 1 if the corresponding month is in the upper (lower) third of the moving average distribution of IPOs. GPTA: Ratio of gross proceeds to total assets. 4.2 Self-selection control Firms that list through an IPO can differ from those that graduate from a junior market. For example, firms backed by private equity investors or venture capital funds are less likely to use the junior market to finance their development because they can afford the IPO directly. Conversely, firms rejected by venture capitalists or whose managers are more concerned by keeping control are more likely to use the junior market and then to graduate. Large firms can access the main board directly, and should not initially list on the venture exchange. This situation requires a control for this self-selection effect. The limitations of the classical two-step procedure developed by Heckman (1979) and the instrumental approach in general are now well established (Larcker and Rusticus, 2010; Lawrence et al., 2011; Lennox et al., 2012; Peel and Makepeace, 2012). In the particular case of the choice of listing mode, the exclusion condition required for the Heckman correction cannot be easily satisfied, because this condition entails finding a variable that influences the listing mode choice and has no influence on the value. We consequently use a propensity score matching approach following Lawrence et al. (2011) and Lennox et al. (2012). This method matches observations based on the probability of making an IPO and generates two samples with similar characteristics. Propensity scores are calculated using the following logistic model: DIPO = α1+ α2 Lasset + α3ROAi + α4 agei + ei (2) where Lasset is the log transformation of pre-listing total assets. ROA is the ratio of net income before extraordinary items to total assets, using prelisting amounts. Age is the age of the company, in years, at the IPO or graduation. 4.3 Data source and variable measurement We collected the lists of IPOs from FPInfomart.ca from 1997 to 2011. We excluded privatization of state-owned companies, demutualizations, creation of income trusts and 11 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 Capital Pool Company Program IPOs.6 Because it is difficult to model value for natural resources, oil and gas, financial and real estate companies, we exclude companies operating in these industries. We got a final sample of 186 IPOs. We obtained prospectuses from SEDAR (the Canadian equivalent of EDGAR). IPOs’ accounting data come from prospectuses. The TSXV provided a list of 724 graduated companies from 1997 to 2011. We excluded 59 graduations that resulted from the acquisition of a TSXV company by a TSX company. As for IPOs, we excluded natural resources, oil and gas, financial and real estate companies. We got a final sample of 218 graduates. Unlike an IPO, a company can graduate to the TSX without raising any gross proceeds, but can issue stock privately after its graduation. We hand-collected the new placements from FPinfomart.ca for the year following graduation. We summed the equity issued during the first year after graduation to get an estimate of the gross proceeds to compare with that of IPOs. We took the median Canadian issue price of these issues following graduation to put the IPO issue price in perspective. Graduates’ accounting data come from Mergent on line, and we verified and supplemented these data with SEDAR. Several companies make a private or public placement several months or weeks before their graduation, but after the fiscal year-end. These new placements were tracked in SEDAR or the FPinfomart.ca database of new issues, press releases and financial statements following graduation. We added the gross proceeds to the total assets and shareholders’ equity of the companies. Several companies (14) graduated just after or shortly (within six months) after a merger or a reverse takeover. For these companies we also hand-collected the accounting data of the entrant (not public) companies by analyzing the management information circular sent to the shareholders of the public companies before the reverse takeover. We then summed the accounting data of the public company and the entrant company to figure out the financial position of the merged entity before graduation. Market data comes from Datastream. For graduation, OV is the market capitalization at the graduation date, measured by the unadjusted price multiplied by the number of shares outstanding. For IPOs, OV is the issue price (from FPinfomart.ca) multiplied by the number of shares outstanding. We got the missing data and verified OV using the financial statements after the IPO or graduations from SEDAR. Based on SIC codes, we defined two dummy variables, associated with High Tech, and Other (DHT, Dother), respectively. AGE is the number of years since the incorporation of the firm (or since the founding date if the firm reincorporates prior to listing). Size is the post-IPO or post-graduation shareholders’ equity. We characterized Hot and Cold 6 The Capital Pool Company (CPC) program has been implemented in Canada to ease the creation of shells, ultimately used in reverse merger listings by operating companies (Carpentier and Suret, 2006). Their IPOs result in the listing of a non-operating company; we consequently excluded CPC IPOs from our sample. 12 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 issue market periods following Helwege and Liang (2004), by estimating the 3-month centered moving average of the IPO number. DHOT (DCOLD) has a value of one if the corresponding month is in the upper (lower) third of the moving average distribution. We follow the methodology developed by Ellul (2006) in a similar context, to develop an ex ante estimator of liquidity, based on a matched-firm approach. We assume that to forecast an IPO’s (graduate’s) future liquidity, investors use the value observed for a previous IPO (graduate) of comparable size, belonging to the same sector and occurring shortly before the listing observed. We estimate the trading volume and the rotation twelve months after this matched IPO (graduation) and use this estimation as a proxy for the expected liquidity of our sample. We consider that the expected rotation is an acceptable proxy for the real unobservable ex-post rotation. 5. Results 5.1 Characteristics of the sample Table 1 presents the main characteristics of the 186 Canadian IPOs and 218 graduations from 1997 to 2011.7 Before listing, IPO firms are approximately 30% larger than graduates, with median total assets of CAN$31.92 million and CAN$24.72 million respectively. Revenues are twice as large for IPO firms than for graduates (CAN$22.70 vs CAN$10.02). Net earnings are very small, but higher for IPOs. As expected, larger firms list directly on the main board, but both Canadian IPOs and graduates could be considered small firms.8 In both groups, the proportion of firms reporting losses is approximately 45%, and the proportion of firms reporting no revenues is about 9%. In Canada, IPOs and graduations are mainly done by small firms, according to international standards. The proportions of firms in the High Technology sectors are in the same range: 60% for IPOs and 55% for graduates. IPO firms are slightly older than graduates. An important difference is in the level of shareholders’ equity. The median shareholders’ equity is CAN$9.75 million for IPOs versus CAN$16.65 million for graduates. Graduates have already done an IPO on the venture exchange, and have raised cash at that time. They also exhibited a steady increase in stock price during the three years before their graduation, which has increased their capacity to raise more money through private placements or seasoned equity offerings. This implies that before the IPO, graduates were less leveraged than IPO firms. 7 Almost all distributions exhibit a large right skewness as indicated by the high value reported as a maximum. For this reason, we discuss and test the median of the distributions. 8 According to the European Commission’s definition, the category of micro, small and medium-sized enterprises consists of enterprises that have revenues lower than 50 million euros or an annual balance sheet total not exceeding 43 million euros, about CAN$70 million (European Commission, 2005). Statistics Canada defines small businesses as firms with fewer than 500 employees and less than CAN$50 million in annual revenues. 13 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 Each IPO firm raised money; the median gross proceeds are CAN$42.16 million. Only 76 of the 218 graduates raised money within a year of listing on the main market. The median amount of these private placements is CAN$14.90 million. On average, graduates raised CAN$7.70 million after their listing on the main board. However, the median expected liquidity is higher for graduates (0.32) than for IPOs (0.23). After listing, IPO firms are larger than graduates. The median overall value reaches CAN$138.45 million for IPOs but only CAN$66.43 for graduates. The median postlisting net assets is CAN$56.41 million for IPOs and CAN$21.81 million for graduates. Such differences imply that we should use a control method for endogeneity. Insert Table 1 here 5.2 Propensity Score matching Propensity scores are calculated using the logistic model presented in equation 2. The maximum caliper distance is set to 10%. According to the propensity score matching method, the pairs of listing firms are formed based on firms whose propensity scores differ by at most the caliper width. Reducing the caliper width increases the likelihood that the distribution of observed baseline covariates will be the same in both groups, but is detrimental to the number of observations. There are no clear rules to determine which variables to include in the model, but only variables influencing the choice decision and the outcome variable simultaneously should be included (Caliendo and Kopeinig, 2008). We follow the recommendation of Heckman et al. (1998) to begin with a parsimonious specification of the model, and iteratively add variables to the specification. A new variable is kept if it is statistically significant at conventional levels. We present the results obtained with the parsimonious model and with a larger caliper to keep the largest possible number of observations in the sample. In Table 4, we analyze the robustness of our results to the choice of caliper width. We also present the results obtained with the complete propensity score matching model, which greatly reduces the number of observations. The overall conclusion is that our results are robust to caliper distance and specification of the propensity score matching model. In Table 2, we present the statistical tests of the differences between IPO firms and graduates for the whole and propensity score-matched samples. To test the differences between medians of the distributions of continuous variables, we use the nonparametric Kruskal-Wallis test. For dummies, we use the Pearson chi-square test. In this table, the caliper width is set to 10%. For the whole sample, significant differences exist between the size of IPO firms and graduates. IPO firms have revenues and total assets significantly higher than graduates, but graduates exhibit higher shareholders’ equity. The median ages also differ significantly between both groups. The propensity score-matching reduces the difference in the size variable, but it remains significant. We do not observe significant differences between the two groups regarding profitability, proportion of firms without revenues or earnings, and sectoral distribution. Gross proceeds differ significantly before and after the match. A similar proportion of prestigious auditors are involved in both groups. The distribution of the two types of 14 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 listing across the hot and cold issue market period differs. The gross proceeds difference and the hot and cold dichotomy will be specifically tested in the next steps of the analysis. IPOs firms are larger before the listing than the graduates and raise significantly higher amounts of cash than the graduates do. Their post-listing size is consequently significantly higher than that of the graduates. We can consider OV as an estimate of the listing market value of the firms, and Size as an estimate of the book value of equity. The market-to-book ratio based on the medians of the distributions is close to 3 for both groups of firms, similar to that reported in the US just before the IPOs (Chahine et al., 2012 p.185). Both IPO firms and graduates are valued as if they had significant growth opportunities. From 1995 to 2005, the average market-to-book ratio of Canadian firms fluctuates between 1 and 2 (King and Segal, 2008). Insert Table 2 here 5.3 First hypothesis and robustness Our first hypothesis states that with all things being equal, the value of an IPO firm should be higher than that of a graduate. Table 3 presents the correlation matrix. Table 4 reports the results of Model 1, which is similar to the model proposed by Aggarwal et al. (2009). The dependent variable is the log transformation of the overall value (estimated using the issue price in the case of IPOs). Accounting variables are measured before the issue or graduation. In column 1, we use the full sample. In column 2, we apply the model to the propensity score matched sample, with a relatively large caliper width of 10% and the parsimonious matching model. In columns 3 and 4, we reduce the caliper distance (and the sample) to check the robustness of our results to the matching process parameters. The rightmost column presents the results when the complete matching model is used. This model greatly reduces the number of observations. The coefficient of the DIPO dummy variable is positive and significant, and the influence of the propensity matching parameters is slight. The main result appears robust along this dimension. The coefficient ranges from 0.52 to 0.78. A coefficient of 0.60 indicates that the average value advantage for IPOs relative to graduation is 60% of the log of the overall value (82% of the overall value), when income, book value, sales and the high tech sector are considered in the model. This is economically significant and consistent with the hypothesis that IPOs are, on average, overpriced based on the issue price. The effect of the listing mode remains even if we use a complete propensity score matching model, which reduces the total number of observations in the sample to 68 (Column 5). Sales and shareholders’ equity are positively associated with overall value, as expected. However, the sign of the LINC variable is negative. This is not a totally unexpected result. Several studies summarized by Joos and Zhdanov (2008) report an “anomalous” relation between negative earnings and equity value (Collins et al., 1999; Core et al., 2003; Joos and Plesko, 2005). Jan and Ou (2011) report that the market, on 15 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 average, prices negative-book-value firms higher than positive-book-value firms. Aggarwal et al. (2009) find that IPO firms with more negative earnings have higher valuations than do firms with less negative earnings, and suggest that negative earnings are a proxy for growth opportunities. This explanation is consistent with Jane and Ou’s suggestion that R&D accumulated over time plays an important role in the market valuation of firms reporting negative earnings and negative book value. Our sample is composed mostly of high tech firms, and more than 45% of the sample report losses. This probably explains the observed relation. As a robustness check, we omit LINC in the first model reported in Table 5, but this change does not affect the main results of the model. Insert tables 3 and 4 here 5.4 Complementary hypotheses Table 5 reports the tests of the three complementary hypotheses proposed to explore possible explanations for the differences observed between the value of IPOs and graduates. Our second hypothesis states that with all things being equal, the value of an IPO firm should be higher than that of a graduate, when expected liquidity is accounted for. We introduce expected liquidity as a new explanatory variable in column 2 of Table 5. Note that expected liquidity is higher for graduates than for IPO firms, although the difference is non-significant. This can be traced to the fact that graduates have already developed a pool of interested individual shareholders while the participation of institutional investors in the largest IPOs reduces float and liquidity. As anticipated, expected liquidity is positively associated with overall value. However, the size of the coefficient and the significance level of the DIPO dummy are not affected by this modification of the model. This indicates that Hypothesis 2 is valid; the value differential attributed to IPOs cannot be explained by a liquidity effect. Our third hypothesis states that the value of an IPO firm should be higher than that of a graduate, when liquidity and market sentiment are accounted for. This hypothesis deserves attention because the frequency of IPOs during hot issue markets is significantly higher than that of graduations during the same period (Table 2). The model incorporating the hot and cold dummy variables is presented in column 3. Valuations are negatively associated with hot periods and positively linked with cold periods. The coefficient of the dummy related to the listing mode (DIPO) is not changed. The higher IPO value is not attributable to the exploitation of market sentiment. Note that the effects of hot and cold markets on prices observed are not consistent with previous evidence on IPO market cycles. However, the literature indicates that hot markets are associated with high volumes of IPOs and large underpricing (Helwege and Liang, 2004), but large underpricing is not inconsistent with lower issue prices. He (2007) shows that IPOs in hot periods result in higher share prices on the secondary market, and considers that this secondary market price does reflect the firm’s 16 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 fundamentals. Because underpricing is higher during hot issue markets, issue price is likely to be lower. Indeed, He (2007) explains that in hot periods, the information produced by investment bankers improves the quality of IPO firms; this allows ex ante low quality firms to go public. Hence the lower issue price observed during hot issue markets. The fourth hypothesis states that the difference in value between IPO firms and graduates is explained by the differences in the size of the financing raised at listing time, according to the investor recognition hypothesis. We add the gross proceeds expressed as a proportion of pre-listing total assets (GPTA) to the model (Table 5, Column 4). As expected, the coefficient is positive: value is positively influenced by the amount of money raised at the listing date, as suggested by Kecskes (2008). Investor recognition is not the sole explanation for this result. Because their fees are based on a proportion of the gross proceeds, investment bankers have a greater incentive to price larger IPOs higher than smaller IPOs. Large issues can also be more attractive for institutional investors, which can lead to higher value at the end of the book building process. Nevertheless, the positive association between gross proceeds and value does not change the level of significance of the DIPO variable. IPOs remain better priced than comparable firms. Insert table 5 here 6. Conclusion In this paper, we contribute to the literature on IPO pricing by comparing the equity value of a group of small IPO firms with graduates that list on the same market during the same period. The benchmark firms share several characteristics likely to induce relatively high value. First, they have been listed for several years on a junior exchange and the prices and trades observed on this market should have reduced the information asymmetry and the risk perceived by investors. Second, the graduates exhibit, on average, a steady performance on the junior markets, allowing them to fulfill the listing requirements of the main board. Third, they have already implemented substantial changes in their organization, ownership structure and relations with the capital market that characterize the IPO firms (Derrien, 2010). Our research design compares the market prices of graduates with the issue price of IPOs, which biased our results against our hypothesis. Despite the previous elements, we observe that IPOs are priced higher by investment bankers than comparable firms are. The result is robust to several specifications of the propensity score matching models. The difference we observe is economically significant, and cannot be explained by the expected differences in liquidity in the aftermarket, the distribution of IPOs in hot and cold issue markets, or by the investor recognition hypothesis. If graduates are fairly priced by the market, our results indicate that investment bankers overvalue the IPO. This result is in line with some previous 17 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 conclusions (Purnanandam and Swaminathan, 2004; Derrien, 2005) and is not inconsistent with the observation of underpricing. As proposed by Derrien (2005, 2010), the underwriter sets the prices higher than the fundamental value per share but lower than the price sentiment investors are willing to pay. This implies both overpricing relative to fundamentals and abnormal initial returns. Our findings have implications along several dimensions. For entrepreneurs, the results indicate that the IPO process on the main board has value, despite the underpricing generally observed on the market. For researchers, our results suggest that the price set during the first days of trading could be irrational because it is set higher than an issue price that is also higher than the value based on fundamentals. For regulators, amidst the debate surrounding proportional disclosure and the prospectus (Pritchard, 2012), we evidence that the IPO process, including the prospectus, seems to have a significant value effect. References Aggarwal RK, Bhagat S and Rangan SP (2009) The Impact of Fundamentals on IPO Valuation. Financial Management 38(2): 253-284 Autore DM and Kovacs T (2014) Investor Recognition and Seasoned Equity Offers. Journal of Corporate Finance 25(0): 216-233 Baker MP, Ruback RS and Wurgler JA (2006). Behavioral Corporate Finance: A Survey. Handbook of Corporate Finance: Empirical Corporate Finance. B. Espen Eckbo (ed.): Elsevier/North-Holland. Boulton TJ, Smart SB and Zutter CJ (2011) Earnings Quality and International IPO Underpricing. Accounting Review 86(2): 483-505 Boutron E, Gajewski J, Gresse C and Labégorre F (2007) Are IPOs Still a Puzzle? A Survey of the Empirical Evidence from Europe. Finance 28(2): 5-41 Caliendo M and Kopeinig S (2008) Some Practical Guidance for the Implementation of Propensity Score Matching Journal of Economic Surveys 22(1): 31-72 Carpentier C, Cumming DJ and Suret J-M (2012) The Value of Capital Market Regulation and Certification: IPOs Versus Reverse Mergers. Journal of Empirical Legal Studies 9(1): 56-91. Carpentier C, L'Her J-F and Suret J-M (2010) Stock Exchange Markets for New Ventures. Journal of Business Venturing 25 (4): 403-422 Carpentier C and Suret J-M (2006) Bypassing the Financial Growth Cycle: Evidence from Capital Pool Companies. Journal of Business Venturing 21(1): 45-73 Carter R and Manaster S (1990) Initial Public Offerings and Underwriter Reputation. The Journal of Finance 45(4): 1045-1067 Chahine S, Arthurs JD, Filatotchev I and Hoskisson RE (2012) The Effects of Venture Capital Syndicate Diversity on Earnings Management and Performance of IPOs in the US and UK: An Institutional Perspective. Journal of Corporate Finance 18(1): 179-192 18 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 Chahine S and Filatotchev I (2011) The Effects of Corporate Governance and Audit and Non-audit Fees on IPO Value. The British Accounting Review 43(3): 155-172 Chemmanur TJ (2010) Venture Capital, Private Equity, IPOs, and Banking: An Introduction and Agenda for Future Research. Journal of Economics and Business 62(6): 471-476 Chemmanur TJ and Krishnan K (2012) Heterogeneous Beliefs, IPO Valuation, and the Economic Role of the Underwriter in IPOs. Financial Management 41(4): 769811 Chen CR and Mohan NJ (2002) Underwriter Spread, Underwriter Reputation, and IPO Underpricing: A Simultaneous Equation Analysis. Journal of Business Finance & Accounting 29(3-4): 521-540 Collins DW, Pincus M and Xie H (1999) Equity Valuation and Negative Earnings: The Role of Book Value of Equity. The Accounting Review 74(1): 29-61 Cook D, Kieschnick R and Ness RV (2006) On the Marketing of IPOs. Journal of Financial Economics 82(1): 35-61 Core JE, Guay WR and Buskirk AV (2003) Market valuations in the New Economy: an investigation of what has changed. Journal of Accounting and Economics 34(13): 43-67 Daily CM, Certo ST and Dalton DR (2005) Investment Bankers and IPO Pricing: Does Prospectus Information Matter? Journal of Business Venturing 20(1): 93-111 Daily CM, Certo ST, Dalton DR and Roengpitya R (2003) IPO Underpricing: A MetaAnalysis and Research Synthesis. Entrepreneurship Theory and Practice 27(3): 271-295 Deloof M, Maeseneire WD and Inghelbrecht K (2009) How Do Investment Banks Value Initial Public Offerings (IPOs)? Journal of Business Finance & Accounting 36(12): 130-160 Demers EA and Joos P (2007) IPO Failure Risk. Journal of Accounting Research 45(2): 333-371 Derrien F (2005) IPO Pricing in Hot Market Conditions: Who Leaves Money on the Table? The Journal of Finance 60(1): 487-521 Derrien F (2010). Initial Public Offerings. Behavioral Finance. H. Kent Baker and John R. Nofsinger (eds): John Wiley & Sons, Inc.: 475-490. Derrien F and Kecskes A (2007) The Initial Public Offerings of Listed Firms. The Journal of Finance 62(1): 447-479 Dissanaike G and Amel-Zadeh A (2007) Discussion of Venture Capitalists, Business Angels, and Performance of Entrepreneurial IPOs in the UK and France. Journal of Business Finance & Accounting 34(3-4): 529-540 Eckbo BE and Norli O (2005) Liquidity Risk, Leverage and Long-run IPO returns. Journal of Corporate Finance 11(1/2): 1-35 Ellul A and Pagano M (2006) IPO Underpricing and After-Market Liquidity. Review of Financial Studies 19(2): 381-421 European Commission (2005). The New SME Definition: User Guide and Model Declaration. Enterprise and Industry Publication, Available at http://ec.europa.eu/enterprise/policies/sme/files/sme_definition/sme_user_guide_ en.pdf 19 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 Feng C and Yue L (2013) Voluntary Adoption of More Stringent Governance Policy on Audit Committees: Theory and Empirical Evidence. Accounting Review 88(6): 1939-1969 Firth M and Liau-Tan CK (1998) Auditor Quality, Signalling and the Valuation of Initial Public Offerings. Journal of Business Finance & Accounting 25(1/2): 145-165 Gregory A, Guermat C and Al-Shawawreh F (2010) UK IPOs: Long Run Returns, Behavioural Timing and Pseudo Timing. Journal of Business Finance & Accounting 37(5-6): 612-647 Hand JRM (2003). Profits, Losses and the Non-Linear Pricing of Internet Stocks. Intangible Assets: Values, Measures and Risks. J.R.M.Hand and B. Lev Eds. New York, NY: : Oxford University Press. Hanley KW and Hoberg G (2009) The Information Content of IPO Prospectuses Review of Financial Studies 23(7): 2821-2864 He P (2007) A Theory of IPO Waves. Review of Financial Studies 20(4): 983-1020 Heckman J, Ichimura H, Smith J and Todd P (1998). Characterizing Selection Bias Using Experimental Data. NBER Working Paper No. 6699 Heckman JJ (1979) Sample Selection Bias as a Specification Error. Econometrica 47(1): 153-162 Helwege J and Liang N (2004) Initial Public Offerings in Hot and Cold Markets. Journal of Financial and Quantitative Analysis 39(3): 541-569 Hopp C and Dreher A (2013) Do Differences in Institutional and Legal Environment Explain Cross-country Variations in IPO Underpricing? Applied Economics 45(4): 435-454 Jan C-L and Ou JA (2011) Negative-Book-Value Firms and Their Valuation. Accounting Horizons 26(1): 91-110 Jenkinson T and Ljungqvist A (2001). Going Public: The theory and Evidence on How Companies Raise Equity Finance, Second edition. Oxford : Oxford University Press. Joos P and Plesko GA (2005) Valuing Loss Firms. The Accounting Review 80(3): 847870 Joos P and Zhdanov A (2008) Earnings and Equity Valuation in the Biotech Industry: Theory and Evidence. Financial Management 37(3): 431-460 Kecskes A (2008). Why Are Firms That Raise More Financing Worth More? University of Toronto, Available at :http://people.ucalgary.ca/~nfa99/2008/papers/50.pdf King M and Segal D (2008) Market Segmentation and Equity Valuation: Comparing Canada and the United States. Journal of International Financial Markets, Institutions and Money 18(3): 245-258 King MR and Segal D (2009) The Long-Term Effects of Cross-Listing, Investor Recognition, and Ownership Structure on Valuation. Review of Financial Studies 22(6): 2393-2421. L'Her J-F and Suret J-M (1996) Consensus, Dispersion and Security prices. Contemporary Accounting Research 13(1): 209-228 Larcker DF and Rusticus TO (2010) On the Use of Instrumental Variables in Accounting Research. Journal of Accounting and Economics 49(3): 186-205 20 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 Lawrence A, Minutti-Meza M and Ping Z (2011) Can Big 4 versus Non-Big 4 Differences in Audit-Quality Proxies Be Attributed to Client Characteristics? The Accounting Review 86(1): 259-286 Lennox CS, Francis JR and Wang Z (2012) Selection Models in Accounting Research. The Accounting Review 87(2): 589-616 Ljungqvist A (2007). IPO Underpricing: A Survey. Handbook in Corporate Finance : Empirical Corporate Finance, Volume 1. B. Espen Eckbo ed: North-Holland: 375422. Miller EM (1977) Risk, Uncertainty, and Divergence of Opinions. The Journal of Finance 32(4): 1151-1168 Mousa F-T and Reed R (2013) The Impact of Slack Resources on High-Tech IPOs. Entrepreneurship Theory and Practice 37(5): 1123-1147 Peel MJ and Makepeace GH (2012) Differential Audit Quality, Propensity Score Matching and Rosenbaum Bounds for Confounding Variables. Journal of Business Finance & Accounting 39(5-6): 606-648 Pritchard AC (2012) Facebook, the JOBS Act, and Abolishing IPOs. Regulation 35(3): 12-7 Pukthuanthong-Le K (2008). IPO Valuation. Handbook of Finance: John Wiley & Sons, Inc. Pukthuanthong-Le K and Varaiya N (2007) IPO Pricing, Block Sales, and Long-Term Performance. Financial Review 42(3): 319-348 Purnanandam AK and Swaminathan B (2004) Are IPOs Really Underpriced? Review of Financial Studies 17(3): 811-848 Shi C, Pukthuanthong K and Walker T (2013) Does Disclosure Regulation Work? Evidence from International IPO Markets. Contemporary Accounting Research 30(1): 356-387 Vismara S, Paleari S and Ritter JR (2012) Europe's Second Markets for Small Companies. European Financial Management 18(3): 352-388 Wyatt A (2013) Is There Useful Information in the ‘Use of Proceeds’ Disclosures in IPO Prospectuses? Accounting and Finance Forthcoming (2013). Zheng SX (2007) Are IPOs really overpriced? Journal of Empirical Finance 14(3): 287309 21 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 Table 1 Descriptive statistics of TSX IPOs and graduation towards the TSX between 1997 and 2011. n IPO Before listing Total assets, in M$ Net earnings (NE), in M$ Sales, in M$ Shareholders' equity, in M$ Age, in years DHT Listing Gross proceeds, in M$ Issue price, in $ Exp. Liq. Paudit Dhot Dcold Post listing OV, in M$ Size, in M$ Graduation Before listing Total assets, in M$ Net earnings (NE), in M$ Sales, in M$ Shareholders' equity, in M$ Age, in years DHT Listing Gross proceeds, in M$ Issue price, in $ Exp. Liq. Paudit Dhot Dcold Post listing OV, in M$ Size, in M$ mean median std min max 186 186 186 143.99 0.002 138.02 31.92 0.18 22.70 292.42 0.42 13.89 -40.99 264.33 0.00 1417.51 40.13 1148.02 186 186 186 34.15 17.53 0.60 9.75 9.63 77.83 -35.10 21.12 0.69 390.13 100.88 186 186 186 186 186 186 64.47 10.02 0.63 0.76 0.53 0.12 42.16 9.50 0.23 66.02 5.99 1.06 2.48 0.70 0.02 348.00 26.00 5.28 186 186 239.05 102.54 138.45 56.41 314.28 153.34 15.00 1.68 1752.03 953.95 218 218 218 42.48 -1.12 20.33 24.72 0.08 10.02 48.11 4.85 5.20 -24.36 26.54 0.00 303.20 9.48 97.95 218 218 218 27.01 10.93 0.55 16.65 7.88 33.30 10.38 2.84 1.90 230.43 54.66 76 76 218 218 218 218 22.09 4.72 0.48 0.70 0.35 0.20 14.90 3.00 0.32 23.58 5.40 0.49 2.06 0.40 0.02 143.75 22.50 2.17 218 218 107.90 35.40 66.43 21.81 120.86 43.06 5.16 3.91 647.49 280.38 % NE<0 % sales=0 45.16 8.60 46.33 9.17 The sample excludes natural resources, oil and gas, financial and real estate companies. Accounting variables are measured before the issue or graduation. Net earnings (Sales) means net income before extraordinary items (Revenues). Age is the age of the firm, in years, at the IPO or graduation, High Tech is a dummy variable set to 1 if the industry of the issuer is High Tech. Exp. Liq. is the expected turnover estimated by the trade amount/market value of an IPO (graduation)-size-sector matched IPO (graduation) 12 months after the matched IPO (graduation). Paudit is a dummy variable set to 1 if the auditor before the IPO or graduation is prestigious, and 0 22 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 otherwise. Dhot (Dcold) has a value of 1 if the corresponding month is in the upper (lower) third of the moving average distribution of IPOs. OV means Offer Value. For IPOs, OV is the offer price multiplied by the number of shares outstanding immediately after the IPO; for graduation, OV is the market capitalization just after graduation on the main market. Size is the post-IPO or postgraduation net assets. Table 2 Descriptive statistics and test of the difference in means between the IPO and graduation groups, for the full sample and the propensity score matched sample Full sample Propensity score matched sample All Obs. IPO non IPO Diff All Obs. IPO non IPO Diff. Median Median Median Median Median Median Median Median p value p value Before listing Totasset, M$ 27.54 31.92 24.72 0.03 ** 27.00 27.76 25.02 0.42 SE, M$ 14.08 9.75 16.65 0.00 *** 13.45 7.58 17.77 0.00 *** Sales, M$ 15.96 22.70 10.02 0.00 *** 15.85 20.03 9.75 0.00 *** Norev, mean 0.09 0.09 0.09 0.84 0.09 0.09 0.08 0.69 EPS<0, mean 0.46 0.45 0.46 0.81 0.48 0.48 0.48 1.00 ROA 0.01 0.01 0.00 0.80 0.00 0.00 0.00 0.96 Age, years 8.47 9.63 7.88 0.00 *** 8.22 8.36 8.11 0.63 DHT, mean 0.57 0.60 0.55 0.29 0.64 0.67 0.62 0.35 Dother, mean 0.43 0.40 0.45 0.29 0.36 0.33 0.38 0.35 Listing GP, M$ 31.67 42.16 14.90 0.00 *** 30.10 40.02 15.00 0.00 *** Exp. liq. 0.30 0.23 0.32 0.33 0.31 0.23 0.34 0.24 Paudit, mean 0.73 0.76 0.70 0.21 0.73 0.75 0.72 0.53 Dhot, mean 0.44 0.53 0.35 0.00 *** 0.46 0.56 0.36 0.00 *** Dcold, mean 0.16 0.12 0.20 0.02 ** 0.15 0.10 0.19 0.02 ** Post listing OV, M$ 90.59 138.45 66.43 0.00 *** 93.84 136.78 67.32 0.00 *** Size, M$ 32.03 56.41 21.81 0.00 *** 32.03 52.18 22.95 0.00 *** Number of obs. 404 186 218 324 162 162 ***, **, * indicates significance at the 0.01, 0.05 and 0.1 levels respectively. GP means gross proceeds. Accounting variables are measured before the issue or graduation. Totasset (SE, sales) is total assets (shareholders’ equity, revenues). Norev is a dummy variable equal to 1 if the firm has no revenues, and 0 otherwise. EPS<0 is a dummy variable equal to 1 if the firm reported a net loss before extraordinary items. ROA is the net income before extraordinary items divided by total assets. Age is the age of the firm, in years, at the IPO or graduation. DHT (Dother) is a dummy variable set to 1 if the industry of the issuer is High Tech (Other) and 0 otherwise. Exp. liq. is the expected turnover estimated by the trade amount/market value of an IPO (graduation)-size-sector matched IPO (graduation) 12 months after the matched IPO (graduation). Paudit is a dummy variable set to 1 if the auditor before 23 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 the IPO or graduation is prestigious, and 0 otherwise. Size is the post-IPO or post-graduation net asset. Dhot (Dcold) has a value of 1 if the corresponding month is in the upper (lower) third of the moving average distribution of IPOs. Diff median p value means p value of non-parametric KruskalWallis test of difference between the IPO and non-IPO groups. For dummies, we report the mean rather than the median, and diff median p value means p value of a Pearson chi-square test of difference between the IPO and non-IPO group. OV means Offer Value. For IPOs, OV is the offer price multiplied by the number of shares outstanding immediately after the IPO; for graduation, OV is the market capitalization just after graduation on the main market. Size is the post-IPO or post-graduation net assets. Table 3 Pearson Correlation Coefficients, propensity score matched sample (324 observations) LOV LINC LOV 1.00000 LINC -0.02604 1.00000 LBV Lsales DHT DIPO Paudit Lasset age DHOT DCOLD 0.6405 LBV Lsales DHT DIPO paudit Lasset Age DHOT 0.09485 0.27188 0.0883 <.0001 0.20973 0.50108 0.0001 <.0001 0.07731 1.00000 0.09090 0.1650 -0.30374 0.10942 -0.33548 0.1024 <.0001 <.0001 0.30776 0.02649 0.0491 0.34543 0.23170 0.05150 <.0001 0.6347 <.0001 0.3554 0.04379 <.0001 -0.08151 0.05987 -0.02861 0.04143 0.03482 0.4321 0.1432 0.2826 0.6079 0.4574 0.5323 0.46758 0.16327 0.37031 0.65061 -0.23382 0.06873 -0.05480 <.0001 0.0032 <.0001 <.0001 <.0001 0.2173 0.3255 0.07585 0.25172 0.00614 0.30158 -0.04031 0.11363 0.06686 0.19881 0.1732 <.0001 0.9123 0.01858 <.0001 0.4697 0.0410 0.2301 0.0003 0.02054 -0.02601 0.19827 -0.01761 -0.08222 0.16980 1.00000 0.7390 -0.01662 0.01278 0.7126 0.6409 0.0003 0.7522 0.1397 0.14498 -0.14941 -0.13146 -0.08661 0.14206 0.7657 0.0090 0.0071 0.0179 0.1197 0.0105 0.17966 0.8187 -0.07954 0.03370 -0.09206 0.08417 0.08356 -0.06111 0.0012 0.1532 0.0981 0.1306 0.1334 0.2727 -0.11452 0.13679 0.0394 DCOLD 0.12181 0.0284 Exp. liq. 1.00000 0.0137 0.5455 1.00000 1.00000 1.00000 1.00000 1.00000 0.0022 0.07182 -0.37773 1.0000 0.03587 0.1973 <.0001 0.08373 0.00257 -0.0446 0.5200 0.1326 0.4230 0.9632 LOV is the log transformation of the offer value (in M$). Accounting variables are measured before the issue or graduation. Linc is the log transformation of net earnings before extraordinary items (in M$). LBV (Lasset) is the log transformation of shareholders’ equity (total assets) in M$. Lsales is the log transformation of revenues (in 24 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 M$). DHT is a dummy variable set to 1 if the industry of the company is High Tech and 0 otherwise. DIPO is a dummy variable set to 1 if the company issues an IPO and 0 otherwise. Paudit is a dummy variable set to 1 if the financial statements of the company are audited by a prestigious auditor and 0 otherwise. Age is the age of the firm, in years, at the IPO or graduation. Dhot (Dcold) has a value of 1 if the corresponding month is in the upper (lower) third of the moving average distribution of IPOs. Exp. liq. is the expected turnover estimated by the trade amount/market value of an IPO (graduation)-size-sector matched IPO (graduation) 12 months after the matched IPO (graduation). 25 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 Table 4 Ordinary least square regression of the offer value on the fundamentals Sample Whole Matched Matched Matched Matched Sample sample sample sample sample Propensity score model model 1 model 1 model 1 model 2 maximum caliper distance 10% 3% 0.05% 5% Intercept 3.16322 3.24243 3.19696 3.16236 2.98599 student's t 20.39 *** 16.84 *** 15.70 *** 11.00 *** 7.80 *** LINC -0.11344 0.11500 0.12930 0.13814 0.10826 student's t -3.53 *** -3.10 *** -3.25 *** -2.64 *** -1.19 LBV 0.16957 0.14317 0.15958 0.17562 0.19666 student's t 6.08 *** 4.61 *** 4.57 *** 3.84 *** 2.20 ** Lsales 0.12904 0.15283 0.12942 0.15879 0.14945 student's t 4.02 *** 4.04 *** 3.13 *** 2.90 *** 1.20 DHT 0.23452 0.27239 0.27988 0.28237 0.45414 student's t 2.34 ** 2.33 ** 2.19 ** 1.66 * 1.78 * DIPO 0.75381 0.69283 0.68410 0.52190 0.78436 student's t 7.34 *** 6.00 *** 5.47 *** 3.13 *** 3.25 *** Paudit 0.19509 0.07905 0.14157 0.10667 0.13209 student's t 1.91 * 0.67 1.12 0.64 0.56 Number 404 324 264 134 68 Adjusted R2 0.2356 *** 0.1835 *** 0.1930 *** 0.1691 *** 0.2400 *** Student’s t are significant for an accepted error risk of 10 percent *, 5 percent **, and 1 percent ***. The dependent variable is the log transformation of the offer value (in M$). Accounting variables are measured before the issue or graduation. Linc is the log transformation of net earnings before extraordinary items (in M$). LBV is the log transformation of shareholders’ equity (in M$). Lsales is the log transformation of sales (in M$). DHT is a dummy variable set to 1 if the industry of the company is High Tech and 0 otherwise. DIPO is a dummy variable set to 1 if the company issues an IPO and 0 otherwise. Paudit is a dummy variable set to 1 if the financial statements of the company are audited by a prestigious auditor and 0 otherwise. Propensity scores are calculated using the following logistic models. Model1: DIPO=F(Lasset, ROA, Age), where Lasset is the log transformation of total assets. Model2: DIPO=F(Lasset, ROA, Age, Lsales), where Lasset (Lsales) is the log transformation of total assets (revenues). 26 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 Table 5 Ordinary least square regression of the offer value on the fundamentals – Robustness check Matched Matched Matched Matched Parameter Sample sample sample sample Intercept 3.40656 3.11725 3.15026 3.04009 student's t 18.16 *** 16.11 *** 15.51 *** 14.84 *** LINC -0.11133 -0.08691 -0.10121 student's t -3.04 *** -2.39 ** -2.79 *** LBV 0.11964 0.14111 0.14770 0.16731 student's t 3.92 *** 4.61 *** 4.92 *** 5.48 *** Lsales 0.10244 0.16179 0.13417 0.16687 student's t 2.96 *** 4.33 *** 3.59 *** 4.30 *** DHT 0.32591 0.25738 0.28161 0.29723 student's t 2.78 *** 2.23 ** 2.47 ** 2.63 *** DIPO 0.68822 0.65451 0.76214 0.66615 student's t 5.88 *** 5.72 *** 6.61 *** 5.58 *** Paudit 0.10129 0.10612 0.12813 0.09526 student's t 0.85 0.92 1.12 0.84 Expected liquidity 0.20838 0.20928 0.16596 student's t 3.27 *** 3.35 *** 2.60 *** Dhot -0.23900 -0.24158 student's t -2.15 ** -2.20 ** Dcold 0.37324 0.34474 student's t 2.35 ** 2.19 ** GPTA 0.04463 student's t 2.76 *** Number 324 324 324 324 Adjusted R2 0.1614 *** 0.2077 *** 0.2407 *** 0.2565 *** Student’s t are significant for an accepted error risk of 10 percent *, 5 percent **, and 1 percent ***. The models are estimated using a sample of 404 observations (whole sample) and 324 observations (matched sample) of IPOs and graduations. The dependent variable is the log transformation of the offer value (in M$). Accounting variables are measured before the issue or graduation. Linc is the log transformation of net earnings before extraordinary items (in M$). LBV is the log transformation of shareholders’ equity (in M$). Lsales is the log transformation of sales (in M$). DHT is a dummy variable set to 1 if the industry of the company is High Tech and 0 otherwise. DIPO is a dummy variable set to 1 if the company issues an IPO and 0 otherwise. Paudit is a dummy variable set to 1 if the financial statements of the company are audited by a prestigious auditor and 0 otherwise. Dhot (Dcold) has a value of one if the corresponding month is in the upper (lower) third of the moving average distribution of IPOs. Expected liquidity is the expected turnover estimated by the trade amount/market value of an ipo(graduation)-size-sector matched IPO (graduation) 12 months after the matched IPO (graduation). GPTA means gross proceeds divided by total assets. Propensity scores are calculated using the following logistic model: 27 Proceedings of Paris Economics, Finance and Business Conference 13 - 15 April 2015, Crowne Plaza Hotel Republique, Paris, France ISBN: 978-1-922069-73-3 DIPO= F (Lasset, ROA, age), where Lasset is the log transformation of total assets. The maximum caliper distance is set to 10%. 28