Analysts Coverage and Long-term Performance of Initial Public Offerings

Transcription

1 Analysts Coverage and Long-term Performance of Initial Public Offerings Vijay Jog, Carleton University Bruce J. McConomy, Wilfrid Laurier University * May 2003 * Corresponding author's address: School of Business & Economics, Wilfrid Laurier University, Waterloo, Ontario, Canada, N2L 3C5. bmcconom@wlu.ca Phone (519) x2890; fax (519) Acknowledgements: We thank the Canadian Academic Accounting Association for financial support. We greatly appreciate access to analyst forecast data provided by I/B/E/S, a service of Thomson Financial, as part of its broad academic program to encourage earnings expectations research. Lastly, we acknowledge the excellent research assistance provided by Qian Hu.

2 Analysts Coverage and Long-term Performance of Initial Public Offerings ABSTRACT It is well known that, on average, initial public offerings (IPOs) significantly underperform the market over the long-run. Various explanations for this underperformance (including the heterogeneous expectations of investors) have been offered. In this paper, we focus on the evolution of financial analyst coverage and the effect of the initiation of analyst coverage on the long-term performance of Initial Public Offerings (IPOs). We find that IPOs followed by IBES financial analysts significantly outperform those that do not receive any coverage. We also examine factors associated with financial analysts initiating coverage subsequent to companies initial public offerings. Our results indicate that the original analyst coverage decision is affected by: the extent of earnings history provided in the prospectus; underwriter prestige; firm size; whether management included an earnings forecast in the IPO prospectus and whether the post-ipo earnings indicate good news. 2

3 Analysts Coverage and Long-term Performance of Initial Public Offerings 1.0 Introduction This paper examines the evolution of financial analyst coverage and the effect of the initiation of analyst coverage on the long-term performance of Initial Public Offerings. We first assess the factors associated with financial analysts decisions to initiate coverage on recent IPO firms. We then assess the extent to which analyst following serves as an important signal of firm quality to investors in the Canadian market. Our research takes advantage of the Canadian IPO environment, where analysts do not cover many IPOs in the first few years after the IPO, in contrast to the U.S. environment where financial analysts cover the vast majority of IPOs immediately after the quiet period. This allows us to investigate factors that influence analysts coverage decisions, and to assess whether there is a difference between the stock market performance of IPOs that are followed by analysts in the post-issue IPO period, as compared to those that are not. We first examine factors associated with financial analysts initiating coverage on Canadian public companies. Our results indicate that the original analyst coverage decision is affected by: the extent of earnings history provided in the prospectus; underwriter prestige; firm size; whether management included an earnings forecast in the IPO prospectus; whether the forecast was audited and whether the earnings performance could be considered as good news. Secondly, our assessment of long run performance indicates that IPOs covered by IBES financial analysts significantly outperform a typical IPO. 3

4 The paper makes the following contributions to the literature. It is the first study to systematically document the evolution of analysts coverage following an IPO. It also provides evidence of the impact of analyst following on the performance of Canadian IPOs. Ongoing analysis will also allow us to discriminate between the effects of earnings estimates provided by analysts from the lead-underwriting firm, versus those who had no part in the IPO issuance Prior Research 2.1 Analyst Following Prior research on analyst following has tended to focus on large U.S. firms and the interplay of analyst following and institutional ownership (e.g., O Brien and Bhushan 1990). In an early study, Bhushan (1989) notes that analyst following can serve as a proxy for information gathering about the firm. He finds that analyst following is related to ownership structure, institutional ownership percentage and the percentage of firms shares held by insiders. He also finds that analysts' coverage is positively related to firm size, return variability and negatively related to the number of lines of business of the firm. O Brien and Bhushan (1990) examine analyst following and institutional ownership for a sample of U.S. public companies from 38 industries for the period. Because analyst following and institutional ownership are interrelated they use simultaneous equations to evaluate the joint behaviour of the two decisions. They find that analysts prefer to follow firms from regulated industries, industries with increasing numbers of firms and tend to avoid firms with higher return volatility and competition from pre-existing analyst following. 1 Subsequent testing will also investigate the influence of the number of analysts following the firm on returns and the effects of a lack of coverage ( neglected firms ) on post-ipo stock market performance. 4

5 In research focusing on the interrelationship of analyst following and IPOs, Rajan and Servaes (1997) found that higher underpricing led to increased analyst following for their sample of U.S. IPOs from More recently, research has focused on the interrelationship between analyst coverage and the underwriting relationship. Lin, McNichols and O Brien (2003) examine a sample of IPOs and seasoned equity offerings (SEOs) and find that in the U.S. lead analysts cover on average 74% of the companies whose offerings they underwrite within one year of the offering. Similarly, they find that investors have access to less unaffiliated research about IPOs, because analysts affiliated with offerings issue recommendations sooner following an offering and in greater numbers than unaffiliated analysts (ibid. p. 3). More specifically, Lin et al. (2003) find that for U.S. IPOs, 75% are covered by affiliated (lead or co-underwriter) analysts within 12 months of the IPO, and 48.3% are covered by unaffiliated analysts. Twenty-two percent of their sample of IPOs had no coverage during the first year after the IPO. Analyst coverage can be an important factor in the post-issue success of an IPO. For example Bradley, Jordan and Ritter (2003) find that not only is coverage initiated immediately after the 25 day quiet period for most U.S. IPO firms, the coverage almost always starts with a buy or strong buy rating. Consistent with this finding, Bradley et al. note that firms who have coverage initiated after the quiet period experience a significant, positive market-adjusted return of 4.1% in the five day period surrounding the end of the quiet period (-2, +2 day window). Also they note that coverage initiated immediately after the quiet period increased over their period of study, reaching 95% of their sample of IPOs by

6 2.2 Post-Issue/Long Run Return Performance An interesting observation about IPOs is their generally poor post-issue return performance. For example, Ritter and Welch (2002) document that, based on a sample of U.S. IPO issuers during the period, the three-year average degree of market-adjusted return was negative 23.4 percent over the post-issue period. Similarly, Page and Reyneke (1997) assess a four-year postissue period on the Johannesburg Stock Exchange and document an underperformance of 18.4 percent per annum. They also discuss significant underperformance internationally based on of the London Stock Exchange, Helsinki Stock Exchange and Latin America. In Canada, Jog and McConomy (2003) report median underperformance of negative 25.1 percent for IPOS in the 24- month post-issue period. Further, they find that IPOs with at least five years of operating results at the time of the IPO, and firms audited by BIG6 auditors, perform significantly better in the post-issue period than other IPOs, on average. 2 Existing explanations for long-term underperformance are more speculative than based on sound theoretical models. For example, Ritter and Welch (2002, p. 1821) focus on two potential explanations for the long-run underperformance of IPOs. First, is the notion that investors have heterogeneous expectations about the valuation of an IPO. The most optimistic investors tend to buy the IPO, and, as the variance of opinion decreases, the marginal investor s valuation will converge towards the mean valuation. The result being that, on average, prices will fall. This is also consistent with drops in share price at the end of the lock-up period (see Bradley et al. 2003). Second, is the contention that in general more IPOs follow successful IPOs, such that the last large group of IPOs would underperform and be a relatively large fraction of the pool of IPOs. However, although this may help to explain poor performance of a portfolio of equally 6

7 weighted IPOs, it does not explain underperformance when each period is weighted equally (e.g., see Loughran, Ritter and Rydqvist (1994)). Finally, Ritter and Welch (2002) discuss several other potential explanations for poor post-issue performance that point towards overconfidence by both entrepreneurs and investors. This latter explanation is consistent with Loughran and Ritter (1995) who suggest that investors seem to be systematically misestimating the probability of finding a big winner. According to them, it is the triumph of hope over experience (p. 47). Each of these explanations relies on the notion that investors are either unwilling or incapable of properly evaluating future prospects of firms. In the context of this paper, it is possible that analyst following may improve the firms information environment and allow investors to better assess the prospects of IPOs. 3.0 SAMPLE DESCRIPTION AND METHODOLOGY 3.1 Sample Description The sample consists of Toronto Stock Exchange (TSE) IPOs that satisfied the following criteria: 1. The firm applied for initial listing on the TSE between January 1983 and December 1994 (based on listings from the TSE Review) The firm issued equity shares that were not previously publicly traded (i.e., IPOs) and whose fiscal year end was prior to or on December 31, Limited life investment funds, limited partnerships, mining firms and firms issuing only preferred shares were excluded. 4 2 For a review of related results in the Canadian context see Jog (1997). 3 Since we are interested in long run performance of IPOs, we have stopped our sample period as of Prior research suggests that the vast majority of mining firms choose not to include an earnings forecast in their prospectus, in part because valuation of such firms is based on the extent of mineral deposits rather than short-run 7

8 4. IPOs that issued units (e.g., common shares plus warrants) were excluded, as the individual components of units are generally not separately priced in IPOs. 5. Firms that included a projection in their prospectus were excluded, because projections may include hypotheses or assumptions that are not necessarily the most probable in management s judgment (CICA, 1989, p. 2). A total of 225 firms met these criteria and had the necessary pricing data available to complete our testing Table 1 about here Table 1 provides a description of the sample considered in this paper. The sample consists of 225 firms; ninety-eight of them received analysts coverage during the course of the 24 month post- IPO issue period. Table 1 - Panel A indicates that typically firms with higher gross proceeds receive coverage in the post-issue period. This is not surprising as higher gross proceeds also indirectly indicate larger firm size, and size is also associated with lower return volatility. The conjecture that, on average, smaller firms are subjected to a higher degree of post-ipo return volatility is also supported by the correlation matrix in Table 3 (correlation between LNGP and VOL is negative and significant at per Table 3 Table 3 is discussed in greater detail below). There is little difference across the two groups with respect to the degree of retained ownership or mean underpricing. Table 1 - Panel B indicates that analysts seem to not initiate coverage for IPOs in the finance industry, but for other industries there is no major difference in terms of coverage initiation. earnings prospects. Therefore mining firms were excluded from the analysis (consistent with Brown et al., 2000 and Jog and McConomy, 2003). 5 Actually a total of 258 firms met the criteria listed above, however we are still in the process of collecting the necessary data to complete our regression analyses for 24 firms. 8

9 3.2 Methodology Methodology Factors related to the initiation of Analyst Coverage We measure the financial analyst(s) decision to follow IPO firms as follows financial analyst following (FAF) equals 1 for firms covered by at least one financial analyst during the course of the 24-month post-issue period and 0 otherwise. The following logit model is used to assess factors affecting financial analyst coverage: Log {FAF i /(1-FAF i )} = α o + α 1 FCST i + α 2 LNAL i + α 3 GN i + α 4 RS i + α 5 VOL i + α 6 EH i + Where, for firm i, α 7 PRES i + α 8 ECP i + α 9 UP i + α 10 LNGP i + ε i (1) FCST i = 1 for firms that include a forecast in their prospectus, and 0 otherwise; LNAL i = absolute value of the natural log of the Leland and Pyle (1977) retained ownership variable; 6 GN i = 1 if actual earnings in the first period ending after the IPO are greater than prior period earnings (i.e., good news based on a random walk model), and 0 otherwise; RS i = 1 if management s IPO is from the period where IPO forecasts have audit assurance (i.e., after the regime shift in 1989), and 0 if reviewed; VOL i = standard deviation of daily stock price returns (volatility) for the sixty days after the IPO; EH i = 1 if earnings history is provided in the IPO for at least 5 periods, and 0 otherwise; PRES i = 1 for IPOs with a prestigious underwriter, and 0 otherwise; 7 ECP i = 1 for IPOs with an executive compensation plan (bonus or stock options), and 0 otherwise; UP i = underpricing is the difference between the closing price on the first day of trading and the initial offering price, expressed as a percentage of the initial offering price; LNGP i = the natural log of gross proceeds from the IPO issue; and ε i = the error term. 6 More specifically, following Downes and Heinkel (1982) and Clarkson et al. (1992), α = (N - Np - Ns)/N where N is the total number of shares outstanding after the initial offer, Np is the number of primary shares in the initial offer and Ns is the number of secondary shares offered by the entrepreneur for resale to the public. LNALi is equivalent to the absolute value of the retained ownership proxy per Downes and Heinkel (1982) and Clarkson et al. (1992), such that LNALi = αi + ln (1-αi). We utilize the absolute value of their measure in order to allow a more intuitive interpretation (i.e., higher retained ownership is represented by a higher LNAL) consistent with related valuation and post-issue performance research (e.g., Feltham et al. [1991]; Li and McConomy [2003]). 7 Underwriter prestige is based primarily on investment dealer rankings published in the Financial Post 500, with the top 10 investment dealers each year considered prestigious (McConomy, 1998; Espenlaub et al., 2001). 9

10 We expect that analysts are more likely to cover firms where management has already released an earnings forecast (FCST=1), as analysts tend to provide more accurate forecasts of earnings after the release of related management earnings forecasts (Hassell and Jennings 1986). Consistent with Bhushan (1989) we expect that analysts are more likely to follow firms with high-retained ownership (LNAL). McNichols and O Brien (1997) note that analysts are more likely to cover firms that they view favourably. As a proxy for this favourable view we utilize good news, where GN=1 for IPOs that have good news in terms of earnings in the first year following their IPO based on a random walk expected earnings model. McConomy (1998) notes that audited management earnings forecasts in IPO prospectuses are less biased. Therefore, if this increases the likelihood that they will be relied on by financial analysts, we would expect the RS variable to be positive. This is also consistent with the empirical regularity that financial analyst coverage has tended to increase over time (Rajan and Servaes [1997] and Bradley et al. [2003]). Prior research into financial analyst following suggests that coverage is related to return variability (Bhushan [1989]; O Brien and Bhushan [1990]). Similarly, in the context of management earnings forecasts, it has been argued that an increase in the ex ante variability in reported earnings may lead to increased use of earnings forecasts (Hughes 1986), or it may have a negative impact on the disclosure of earnings estimates (Waymire 1985). Following Li and McConomy (2003) we use VOL as our proxy for returns volatility and uncertainty about earnings. Also, to the extent that longer earnings history may increase the likelihood that an earnings forecast is released, we expect EH to be positive (see Mak 1996). As discussed above, the lead underwriter is expected to be more likely to follow the IPO in the post-issue period (Lin 10

11 et al. 2003). Our preliminary proxy for this is PRES, such that IPO firms taken public by a more prestigious underwriter are also more likely to have analyst coverage in the post-issue period. Similarly, the existence of an executive compensation plan, such as a bonus or stock option plan, is generally often contingent on being able to provide a hard estimate of expected earnings (Li and McConomy 2003), such that whether a firm has an ECP may influence financial analysts decisions to follow the IPO in the post-issue period. Rajan and Servaes (1997) note that higher underpricing (UP) is expected to lead to increased analyst following. Finally LNGP provides a proxy for the ex ante uncertainty of IPOs, consistent with Kim, Krinsky and Lee (1993) and Beatty and Ritter (1986, 219), it also captures the empirical regularity that smaller IPOs tend to be more speculative (ibid.) Methodology Post-Issue long run Performance Typically, the post-issue return performance of a group of stocks is analyzed by investigating the returns earned by an investor whose investment strategy is to invest in each IPO as it lists on the stock exchange. However, since this performance may be affected simply by the overall performance of the stock market, it is also necessary to analyze it on a relative basis by comparing it with a widely based stock market index. In this paper, we measure the post-issue return performance (LTP i ) of an IPO (or a specific group of IPOs) as (R ipo, T - R m, T ) where R ipo,t is the average total return (where a 50% return is expressed as 0.5) measured from the first day of the month after the IPO was listed, until the earlier of the delisting date or two years for the IPOs comprising that group; and R m,t is the average of the TSE300 index over the same 8 As a robustness check we reran the results presented below using LNTA (the natural log of total assets as of the end of the last fiscal period before the IPO) rather than LNGP. The results for LNTA are consistent with those presented for LNGP. 11

12 interval. 9 The results shown in the paper are based on the cumulative abnormal return (relative to the TSE 300) at the end of twenty-four months. We perform multivariate analysis of the post-issue performance of IPOs during the 24-month period following the offering date, using the following regression: LTP i = 1 FAF24 i + 2 SPEC i + 3 LNAL i + 4 ECP i + 5 EH i + 6 BIG6 i + 7 LNGP + ζ i (2) Where, in addition to the variables defined above: FAF24 i = 1 for an IPO firm where an IBES financial analyst initiates coverage during the 24 month post-ipo period, and 0 otherwise; SPEC i = 1 for speculative offerings, and 0 otherwise; 10 and BIG6 i = 1 if a BIG6 (or BIG8) accounting firm is associated with the offering, and 0 otherwise. If the analyst coverage decision provides an important indication to investors regarding the suitability of the companies for long-term investing purposes, FAF24 is expected to be positive. The SPEC variable reflects the fact that the OSC requires prospectuses of speculative IPOs to be specifically labeled as speculative (usually on the first page of the prospectus), which is expected to reflect increased ex ante uncertainty. We also control for the extent of retained ownership (LNAL), the existence of an executive compensation plan (ECP) and the extent of operating history in our analysis of post-issue performance consistent with Jog and McConomy 9 Lyon et al. (1999) have questioned the exclusive use of such a benchmark in comparing relative performance. However, our use of a benchmark such as the TSE300 is consistent with many other studies of IPO performance (e.g., Ritter, 1991; Lyon et al., 1999). 10 Beatty and Ritter (1986) use a count of the number of uses of proceeds as the SEC requires more speculative issues to provide relatively more detailed enumerations of the uses of proceeds. In the absence of a similar 12

13 (2003). Existence of an executive compensation plan indicates a higher degree of information asymmetry, because such plans are usually contingent on one or both of future net income and share price performance (Scott 2000, p. 317). Therefore, on average, a manager of an IPO with an ECP plan would have better private information about earnings than one without such a plan. We control for the extent of operating history of the firm using the EH dummy variable, which equals 1 for firms with at least five periods of prior operating history (this disclosure is non-voluntary, and tends to decrease ex ante uncertainty). Auditor quality has been used in prior research to signal firm value (Titman and Trueman [1986]), and has also been found to have a significantly positive influence on post-issue performance (Jog and McConomy 2003). Finally we control for firm size using the natural log of gross proceeds from the IPO, which prior research suggests is another proxy for ex ante uncertainty (Kim, Krinsky and Lee [1993], Beatty and Ritter [1986]); ζ i is the error term. 4.0 Results 4.1 Results Financial Analyst Coverage Table 2 variables relate to the specific properties of the sample firms, and provides univariate results regarding the differences between IBES forecast firms and those that do not receive coverage by IBES analysts in the 24 month post-issue period. It is clear that IPOs receiving analysts coverage display different characteristics than those that receive no coverage. First, firms receiving coverage are also more likely to provide a forecast in their own prospectuses (FCST significant at 0.05). 11 Of the firms receiving analyst coverage, the vast majority had requirement regarding use of proceeds in Canada, the OSC requirement provides a direct indication of speculative issues for our sample prospectuses, which similarly serves to reduce the ex ante uncertainty of investors. 11 In an effort to ensure that management is able to convey all relevant information to all investors, in 1982 the Ontario Securities Commission (OSC), which acts as a counterpart to the Securities Exchange Commission in the 13

14 good news in terms of better earnings performance in their first post-ipo period (GN significant at 0.01). Similarly, analyst covered firms include (in terms of a relatively higher percentage within the group) more firms in the audit assurance ( ) time period (RS significant at 0.01), firms with a longer earnings history (EH significant at 0.01), those with high-prestige underwriters (PRES significant at 0.001) and those that have a performance based executive compensation plan (ECP significant at 0.05). Results from tables 1 and 2 provide an indication that the initiation of analysts coverage is not a random event; the firm descriptors, as well as specific firm characteristics, might influence the coverage Table 2 about here Table 3 provides a correlation matrix for the independent variables used in the multivariate analyses per Tables 4 and 5. The correlation matrix gives no indication of multicollinearity concerns, and only two of the bivariate correlations are greater than 0.40 (the correlations of LNGP (the natural log of gross proceeds from the IPO issue) with returns volatility (VOL) and underwriter prestige (PRES)) Table 3 about here Table 4 shows the results of the LOGIT estimates as per equation 1, and confirms the conjectures based on the univariate results across the individual variables from Table 2. The LOGIT models provides an overall percentage of correct predictions of 69.3 percent, and all variable coefficients show the predicted signs. The most dominant variables are: RS U.S. introduced a policy allowing firms to voluntarily include management earnings forecasts in IPO prospectuses filed with the Commission. Firms making an initial public offering must file a prospectus with the OSC prior to obtaining a listing on the Toronto Stock Exchange. The Toronto Stock Exchange is Canada s premier exchange, which accounts for more than eighty percent of Canadian market capitalization. 14

15 (management forecasts from the audit assurance period ( ) consistent with greater reliance by financial analysts on prospectuses from this period, and also consistent with the general trend to increased analyst coverage over time (significant at 0.001); the size of gross proceeds indicting analysts preference for larger firms (LNGP significant at 0.01); good news meaning that the firm actually performed better than a random walk prediction (GN significant at 0.02); inclusion of a forecast by the firm reinforcing management s ability and willingness to forecast its firms earnings (FCST significant at 0.03); presence of a prestigious underwriter (PRES significant at 0.07); and firms with a longer earnings history (EH marginally significant at 0.10). 12 The ownership percent, volatility of returns in the post-ipo period, the degree of underpricing 13 and the existence of executive compensation plan do not seem to influence the initiation of analysts coverage Table 4 about here The overall results indicate a systematic choice made by analysts to restrict their coverage to a select sub sample of firms that are relatively large, to assure marketability, and to those whose underlying quality has been demonstrated whether by management (by providing forecasts and performing well), or by market intermediaries (e.g., the choice of prestigious underwriters). Firms that do not display such characteristics seem to have a much lower chance of receiving analysts coverage. 12 If large size and high liquidity are the characteristics of institutional investor holding stocks, the same characteristics should be found in analyst following IPOs. Existing empirical evidence also indicates that infrequently traded stocks lack analysts following (Easley, Kiefer, O Hara, and Paperman (1996)).Our results confirm this hypothesis. 13 Note that there is no statistically significant difference between the two groups in the relative degree of underpricing and that the average degree of underpricing in this sample is lower than what is typically reported in 15

16 4.2 Results - Post-Issue long run Return Performance It is possible that analysts coverage does not add any more information to the market place in terms of post-issue performance. So interesting questions to investigate are: whether firms that receive analysts coverage perform differently than those who do not receive coverage; and whether or not the analysts coverage is incrementally important (in addition to firm descriptors and firm characteristics). Table 5 shows the corresponding results Table 5 about here As can be seen, the three variables that have the most explanatory value in explaining the longrun performance of IPOs are whether: firms receive the IBES coverage, have a long earnings history and utilize a BIG 6 accounting firm. All three variables are statistically significant at These results show that investors may wish to consider firms that receive analysts coverage in the post-ipo period when making their investment decisions, as analyst coverage is one of the most significant influencing variables Overall Conclusions and Areas of Further Research The intent of this paper is two fold: to investigate whether there is any systematic pattern in the initiation of analysts coverage and whether this initiation of coverage is related to long-run performance. Our results show that analysts coverage is generally restricted to a subset of firms that are large, are managed by those willing to provide future guidance, and have an independent quality check by a prestigious underwriter. Our results also show that analysts coverage is the U.S. Thus, our results do not support Rajan and Servaes (1997) who found that higher underpricing led to increased analyst following for their sample of U.S. IPOs from We consider these post-issue performance results as being preliminary. Ongoing testing is aimed at establishing the relative performance of two portfolios of recent IPOs in the post-issue period - to determine if those followed by financial analysts outperform those neglected by financial analysts, after controlling for other related variables. 16

17 statistically and positively related to firms showing good earnings performance in the post-issue period, and that it tended to increase over time (i.e, more coverage in the 1990s than the 1980s). We also show that analyst coverage provides a key indication of the long-run stock market performance of IPOs. Analyst coverage is significant in addition to established factors from prior research, such as the extent of earnings history provided in the IPO prospectus, and the use of a prestigious underwriter. There are at least two areas where this paper will be extended. First, it will be interesting to see what comes first: the analysts coverage or the better performance. More specifically, the results presented do not show the causality between long run performance and analysts coverage, except to say that these are related. Second, it would be interesting to know whether the breadth of coverage (number of analysts) influences (or is influenced by) long-run performance. We are in the process of investigating these lines of inquiry. 17

18 REFERENCES Beatty, R. and J. Ritter (1986), Investment Banking, Reputation, and the Underpricing of Initial Public Offerings, Journal of Financial Economics 15, pp Bhushan, R. (1989), Firm Characteristics and Analyst Following, Journal of Accounting and Economics 11, pp Bradley, D., B. Jordan and J. Ritter (2003), The Quiet Period Goes Out with a Bang, Journal of Finance 58, pp Brown, P., A. Clarke, J. How and K. Lim (2000), The Accuracy of Management Dividend Forecasts in Australia, Pacific-Basin Finance Journal 8, pp Canadian Institute of Chartered Accountants (CICA), (1989), Examination of A Financial Forecast Or Projection Included In A Prospectus Or Other Public Offering Document, CICA Handbook - Auditing and Related Services Guideline, September, Clarkson, P. A. Dontoh, G. Richardson and S. Sefcik (1992), The Voluntary Inclusion of Earnings Forecasts In IPO Prospectuses, Contemporary Accounting Research 8, pp Downes, D. and R. Heinkel (1982), Signalling and the Valuation of Unseasoned New Issues, Journal of Finance 37, pp Easley, D., Kiefer, N., O Hara, M., and Paperman, J., (1996), Liquidity, Information, and Infrequently Traded Stocks, Journal of Finance 51, pp Espenlaub, S., M. Goergen and A. Khurshed (2001), IPO Lock-in Agreements in the UK, Journal of Business Finance & Accounting 28, pp Feltham, G., J. Hughes, and D. Simunic Empirical Assessment of the Impact of Auditor Quality on the Valuation of New Issues. Journal of Accounting and Economics, Hassell J. and R. Jennings, Relative Forecast Accuracy and the Timing of Earnings Forecast Announcements The Accounting Review 61, pp Hughes, P. (1986), Signalling by Direct Disclosure under Asymmetric Information, Journal of Accounting and Economics 8, pp Jog, V. (1997), The Climate for Canadian Initial Public Offerings, In: Halpern, P., (Ed) Financing Growth In Canada, University of Calgary Press, pp Jog, V. and B. McConomy, "Voluntary Disclosure of Management Earnings Forecasts in IPOs," Journal of Business, Finance and Accounting 30 (1) & (2), 2003, pp Kim, J., I. Krinsky and J. Lee (1993), Motives for Going Public and Underpricing: New Findings from Korea, Journal of Business Finance & Accounting 20, pp Leland, H. and D. Pyle. (1977), Information Asymmetries, Financial Structure, and Financial Intermediation, Journal of Finance 32, pp

19 Li, Y. and B. McConomy (2003), Simultaneous Signaling in IPOs via Management Earnings Forecasts and Retained Ownership: An Empirical Analysis of the Substitution Effect, forthcoming, Journal of Accounting, Auditing and Finance. Lin, H., M. McNichols and P. O Brien (2003), Analyst Impartiality and Investment Banking Relationships, Working Paper, Stanford Graduate School of Business. Loughran, T. and J. Ritter (1995), The New Issues Puzzle, Journal of Finance 50, pp Loughran, T., J. Ritter and K. Rydqvist (1994), Initial Public Offerings: International Insights, Pacific- Basin Finance Journal 2, pp Lyon, J., B. Barber and C. Tsai (1999), Improved Methods for Tests of Long-Run Abnormal Stock Performance, Journal of Finance 54, pp Mak, Y Forecast Disclosure by Initial Public Offering Firms in a Low-litigation Environment. Journal of Accounting and Public Policy 15, McConomy, B. (1998), Bias and Accuracy of Management Earnings Forecasts: An Evaluation of the Impact of Auditing, Contemporary Accounting Research 15, pp McNichols, M. and P. O Brien (1997), Self-Selection and Analyst Coverage, Journal of Accounting Research 35 (Supplement), pp O Brien, P. and R. Bhushan (1990), Analyst Following and Institutional Ownership, Journal of Accounting Research 28 (Supplement), pp Page, M. and I. Reyneke (1997), The Timing and Subsequent Performance of Initial Public Offerings (IPOs) on the Johannesburg Stock Exchange, Journal of Business Finance & Accounting 24, pp Rajan, R. and H. Servaes. (1997), Analyst Following of Initial Public Offerings, Journal of Finance 52, pp Ritter, J. (1991), Long Term Performance of Initial Public Offerings, Journal of Finance 46, pp Ritter, J. and I. Welch (2002), A Review of IOP Activity, Pricing and Allocations, Journal of Finance 57, pp Scott, W. (2000), Financial Accounting Theory Second Edition. Prentice Hall Canada Inc. Scarborough, Ontario. Titman, S. and B. Trueman (1986), Information Quality and the Valuation of New Issues, Journal of Accounting and Economics 8, pp Toronto Stock Exchange Review. Toronto Stock Exchange, Toronto, Canada, Waymire, G Earnings Volatility and Voluntary Management Forecast Disclosure. Journal of Accounting Research 23 (1):

20 White, H. (1980), A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity, Econometrica 48,

21 TABLE 1 Firm Descriptors - Full Sample (225 firms: 98 IBES forecasters and 127 non-forecast firms) Panel A: Firm descriptors Mean Median Std. Error of Mean IBES Forecast firms Retained ownership, LNAL i 50.0% 46.8% 2.9% Returns volatility, VOL i 2.5% 2.4% 0.1% Underpricing, UP i 6.3% 1.0% 1.7% Gross Proceeds, GP i (millions) Non-forecast firms Retained ownership, LNAL i 51.1% ns 43.5% 2.7% Returns volatility, VOL i 2.9% 2.7% 0.1% Underpricing, UP i 9.3% ns 1.6% 2.4% Gross Proceeds, GP i (millions) 32.7 *** Panel B: Distribution of firms by industry Forecast Non-forecast Total firms firms Industry HighTech Manufacturing Transport/Comm Retail/Distribution Oil & gas Finance Services Other Total where, for firm i LNAL i = absolute value of the natural log of the Leland and Pyle (1977) retained ownership variable; VOL i = standard deviation of daily stock price returns (volatility) for the sixty days after the IPO; UP i = underpricing is the difference between the closing price on the first day of trading and the initial offering price, expressed as a percentage of the initial offering price; LNGP i = the natural log of gross proceeds from the IPO issue;, ***, ns = Differences in panel A significant at 0.001, 0.10 and not significant respectively (two-tailed). The reported significance levels are based on a t-test for the difference between IBES forecast and non-forecast firms. 21

22 TABLE 2 Firm Characteristics - Full Sample (225 firms: : 98 IBES forecasters and 127 non-forecast firms) Firm characteristic Forecast firms Non-forecasters variables Number Percentage Number Percentage Forecast, FCST i 72/ / ** Good news, GN i 83/ / * Regime shift, RS i 67/ / Earnings history, EH i 77/ / * Underwriter prestige, PRES i 90/ / Exec. Comp. Plan, ECP i 45/ / ** where, for firm i: FCST i = 1 for firms that include a forecast in their prospectus, and 0 otherwise; GN i = 1 if actual earnings in the first period ending after the IPO are greater than prior period earnings (i.e., good news based on a random walk model), and 0 otherwise; RS i = 1 if management s IPO forecast has audit assurance (i.e., after the regime shift), and 0 if reviewed; EH i = 1 if earnings history is provided in the IPO for at least 5 periods, and 0 otherwise; PRES i = 1 for IPOs with a prestigious underwriter, and 0 otherwise; ECP i = 1 for IPOs with an executive compensation plan (bonus or stock options), and 0 otherwise;,*, **, ***, ns = difference significant at 0.001, 0.01, 0.05 and 0.10 and not significant respectively (one-tailed, except for LNTA which is two-tailed). The reported significance levels are based on a test of the difference between two proportions. 22

23 23

24 TABLE 3 Correlation Matrix for Multivariate Analyses per Tables 4 and 5 (225 firms) FCST LNAL GN RS VOL EH PRES ECP UP LNGP FAF SPEC BIG6 FCST *** ** ** * LNAL ** * ** GN 0.12*** * 0.20 * *** 0.20 * RS *** *** 0.15 ** *** VOL * * ** EH * ** * 0.17 * PRES *** ** ECP *** *** ** UP * *** LNGP * * * FAF 0.14** * * 0.17 * ** SPEC -0.17** * * BIG *** *** * *** 0.17 * * Where, for firm i, FCST i = 1 for firms that include a forecast in their prospectus, and 0 otherwise; LNAL i = absolute value of the natural log of the Leland and Pyle (1977) retained ownership variable; GN i = 1 if actual earnings in the first period ending after the IPO are greater than prior period earnings (i.e., good news based on a random walk model), and 0 otherwise; RS i = 1 if management s IPO forecast has audit assurance (i.e., after the regime shift), and 0 if reviewed; VOL i = standard deviation of daily stock price returns (volatility) for the sixty days after the IPO; EH i = 1 if earnings history is provided in the IPO for at least 5 periods, and 0 otherwise; PRES i = 1 for IPOs with a prestigious underwriter, and 0 otherwise; ECP i = 1 for IPOs with an executive compensation plan (bonus or stock options), and 0 otherwise; UP i = underpricing is the difference between the closing price on the first day of trading and the initial offering price, expressed as a percentage of the initial offering price; 24

25 LNGP i = the natural log of gross proceeds from the IPO issue FAF24 i = 1 for an IPO firm where an IBES financial analyst initiates coverage during the 24 month post-ipo period, and 0 otherwise; SPEC i = 1 for speculative offerings, and 0 otherwise; and BIG6 i = 1 if a BIG6 (or BIG8) accounting firm is associated with the offering, and 0 otherwise.,*, **, *** = correlation significant at 0.001, 0.01, 0.05 and 0.10 respectively (with Pearson correlations showing to the left of the diagonal and Spearman correlations to the right of the diagonal. 25

26 TABLE 4 LOGIT estimation of the IBES coverage decision (225 firms) Variable Pred. Sign Estimated Coefficient Asymp. t-ratio (sig) a FCST i (.03) LNAL i ns GN i (.02) RS i (.001) VOL i +/ ns EH i (.10) PRES i (.07) ECP i ns UP i ns LNGP i (.01) Constant +/ Likelihood ratio test (10 d.f.) Cragg-Uhler R 2.30; Percentage of correct predictions 69.3% Where, for firm i, FCST i = 1 for firms that include a forecast in their prospectus, and 0 otherwise; LNAL i = absolute value of the natural log of the Leland and Pyle (1977) retained ownership variable; GN i = 1 if actual earnings in the first period ending after the IPO are greater than prior period earnings (i.e., good news based on a random walk model), and 0 otherwise; RS i = 1 if management s IPO forecast has audit assurance (i.e., after the regime shift), and 0 if reviewed; VOL i = standard deviation of daily stock price returns (volatility) for the sixty days after the IPO; EH i = 1 if earnings history is provided in the IPO for at least 5 periods, and 0 otherwise; PRES i = 1 for IPOs with a prestigious underwriter, and 0 otherwise; ECP i = 1 for IPOs with an executive compensation plan (bonus or stock options), and 0 otherwise; UP i = underpricing is the difference between the closing price on the first day of trading and the initial offering price, expressed as a percentage of the initial offering price; LNGP i = the natural log of gross proceeds from the IPO issue

27 TABLE 5 Regression Results for Post-Issue Performance (225 firms) Pred. White s Variable Sign Coefficient t-value Signif. Panel A: Valuation based on retained ownership, forecast, auditor quality and underwriter prestige signals FAF24 i SPEC +/ ns ECP ns EH BIG LNGP ns Intercept +/ ns Adjusted R Specifically, we provide results for the regression: LTP i = 1 FAF24 i + 2 SPEC i + 3 LNAL i + 4 ECP i + 5 EH i + 6 BIG6 i + 7 LNGP + ζ i Where: LTP i = market adjusted post issue performance for an IPO, measured from the first day of the month after the IPO was listed FAF24 i = 1 for an IPO firm where an IBES financial analyst initiates coverage during the 24 month post- IPO period, and 0 otherwise; SPEC i = 1 for speculative offerings, and 0 otherwise; EH i = 1 if earnings history is provided in the IPO for at least 5 periods, and 0 otherwise; LNAL i = absolute value of the natural log of the Leland and Pyle (1977) retained ownership variable; ECP i = 1 for IPOs with an executive compensation plan (bonus or stock options), and 0 otherwise; EH i = 1 if earnings history is provided in the IPO for at least 5 periods, and 0 otherwise; BIG6 i = 1 if a BIG6 (or BIG8) accounting firm is associated with the offering, and 0 otherwise; and LNGP i = the natural log of gross proceeds from the IPO issue All significance tests in panel B are based on White's Heteroscedasticity-consistent covariance matrix (White 1980). 27