The Information Content of PIPE Offerings

Transcription

1 The Information Content of PIPE Offerings Steven Freund* Kose John** Gopala Vasudevan*** November 2006 The authors are *Assistant Professor, College of Management, University of Massachusetts, Lowell, **Charles William Gerstenberg Professor Of Banking and Finance, Stern School of Business, New York University, and ***Associate Professor of Finance and Accounting, Charleston College of Business, University of Massachusetts, Dartmouth. Vasudevan acknowledges the receipt of a Summer Research Grant from the Charlton College of Business.

2 The Information Content of PIPE Offerings Abstract We examine the stock price reaction and long run stock price performance for a large sample of issues known as Private Investment in Public Equity (PIPE). We find that equity and preferred PIPE issues convey positive information about firm value, while convertible debt PIPE offerings have an insignificant stock price reaction around the announcement day. Our cross sectional regressions relating announcement period returns to firm characteristics show the stock price reaction is more positive for larger issues and for firms that have poor operating performance at the time of the issue. PIPE issues, which that have downside protection, price protection, or both types of protection, have a lower announcement stock price reaction. Issuers of equity and preferred PIPEs continue to outperform their matching portfolios for up to 9 months following the issue. Issuers of convertible debt with floating or reset features underperform their matching portfolios for up to 36 months following the issue. 1

3 The Information Content of PIPEs Offerings U.S firms have traditionally raised capital either through the public markets or through private placements. Recently, firms have been raising capital through a new source of financing, private investments in public equity (PIPEs). PIPEs are hybrids between a private placement and public offerings in the secondary markets; they are marketed as if they are private placements but the investor receives securities that are traded immediately in the public markets. The PIPEs market has experienced dramatic growth during the periods, with companies raising more than $60 billion through PIPEs issue. In this study we examine the information released by firms issuing securities in the PIPEs markets. There are several reasons this is an important research topic. First, this is an important part of the market for firms issuing securities, particularly for medium and small firms raising funds. The PIPE market comprised 25% of the equity private placement market in This increased to 41% in 2002 (Dresner and Kurt, 2003). Yet we know little about the information conveyed by these issues, the characteristics of the issuers, and the relation between the information and firm characteristics. Second, the PIPE market has introduced several innovations not commonly seen in traditional corporate security offerings. A distinctive feature of some PIPEs is that they provide the investor price protection to investors. A convertible floating PIPEs issuer for example has a conversion ratio that decreases with the stock price of the issuer. We focus on the information conveyed at the time of issuance, based on two characteristics: the seniority of the security and whether it provides any other protection to the new investors. We classify issues that are senior securities, such as convertible debt, as an 2

4 offering that provides downside protection". We classify issues that include a reset or floating rate feature as one that provides "price protection". Hence, common stock and convertible preferred provides neither downside nor price protection; convertible debt with a fixed conversion price provides downside protection but no price protection; common stock with reset and convertible preferred with a floating or reset conversion price provides price protection but no downside protection; convertible debt with a floating or reset conversion price provides both downside and price protection. We examine the information released by these securities through the stock price reaction to the security announcements. We also examine the relation between the announcement period returns to firm characteristics through cross-sectional regressions. We further analyze the longterm returns to PIPE issuers by following the stock price performance for the three year period following the issuance. These results are consistent with the underreaction hypothesis formalized in Daniel, Hirshleifer, and Subramanyam (1998). We find that the information conveyed by the offering PIPEs is quite different from the information conveyed by public offering of corporate securities. In contrast to the finding that in the public market the information conveyed by issuing informationally sensitive securities such as equity is negative, we find that issuing additional informationally sensitive securities leads to higher announcement period returns in the PIPE markets. We find that the stock price reaction is positive for announcements of common stock and preferred stock issues with no reset. It is not different from zero for security issues that are less sensitive to firm value such as convertible debt. We further find that the stock price reaction to the security issue is lower when the security offers protections to the new investors. Our cross-sectional regressions relating announcement period returns to firm characteristics show the stock price reaction is higher for 3

5 larger issues and for firms that have poor operating performance at the time of the issue. This indicates that the market expects these PIPE investors to create information on firm quality and would invest in information sensitive securities such as equity only when the information produced is very positive. The stock price reaction is also inversely related to firm size and is lower for firms that offer either downside protection, price protection or both types of protection. Our evidence is broadly consistent with the empirical predictions of Fulghieri and Lukin (2001). We also examine the long run stock price performance of the issuing firms. We find significant differences of performance based on the protections offered to new investors. Issuers of PIPEs that offer no downside or price protection to new investors outperform their matched counterparts during the nine month period following the issue; issuers that offer PIPEs with downside protection only perform at the same level as the matched firms; issuers that offer PIPEs with price protection only underperform their matched firms for 18 months; issuers of PIPEs with both price and downside protection underperform their match firms for 36 months. The rest of the paper is organized as follows: Section 2 has the background to the PIPE markets and related papers; Section 3 has the data and method; Section 4 has the results; Section 5 concludes. II. Background We first provide the institutional background for PIPEs, followed by a survey of relater literature. A. Institutional background The equity private placement market consists of three types of issues; Rule 144A transactions, NonRule 144A transactions, and registered direct transactions. During the

6 periods, $91.8 billion of the securities issued were Rule 144A transactions, $36.9 billion were nonrule 144A PIPEs and $2.7 billion were registered direct offerings (Dresner and Kim, 2003). Rule 144A transactions are private placements that involve equity or equity linked securities sold only to a Qualified Institutional Buyers (QIB) with the understanding that these buyers may resell the securities to another QIB. The essential definition of a QIB is an institutional investor that owns and invests in the aggregate at least $100 million in securities of issuers not affiliated with that buyer. NonRule 144 A transactions which are considered PIPEs are private placements that involve equity or equity linked securities executed in compliance with certain exemptions provided for under the Securities Act of The issuers of securities in these transactions must be accredited investors. An accredited investor is any individual or institution deemed capable of understanding the financial risks associated with the issue of restricted securities. Deal structures associated with nonrule 144 A PIPEs include common and preferred equity, convertible preferred equity, convertible debt, and warrants. Registered direct transactions, which are also considered PIPEs, involve the issue of equity and equity linked securities to accredited investors. However, in contrast to nonrule 144A PIPEs, the securities issued in a registered direct transaction have already been registered for sale to investors through a primary registration statement filed with and declared effective by the SEC. PIPEs have received considerable attention in both the academic and popular press because of their floating or reset feature. Consider an investor who invests $1000 in a convertible preferred having the reset feature and trading at $10 per share. The investor can convert and receive 100 shares if the share price is $10 and 125 shares when the share price goes down to $8. 5

7 This essentially protects the investor against any loss in value as long as the company continues to be solvent. Academics view this feature as helping to mitigate the costs of financial distress and the adverse selection problems that investors face when investing in equity like securities (Brennan, 1985). However, the popular press refers to the convertible debt and convertible preferred stock with a reset or floating feature as 'death spiral convertibles' because issuers of these securities have experienced large declines in their stock prices subsequent to their issuance. These declines are partly attributed to some of the investors in these securities, especially hedge funds, shorting the stock before conversion. This enables them to convert at lower prices and receive more shares that they can use to close out their short positions. These actions have prompted the SEC to take action against several hedge funds for manipulating the stock of these PIPE issuers. 1 B. Related Papers There is a large body of theoretical and empirical work that has examined the information conveyed by issuing securities in the public markets. Most of these papers focus on the adverse selection faced by outside investors who buy these newly issued securities. Myers and Majluf (1984) show that the best choice for a firm would be to use a security such as riskless debt that is insensitive to firm value and then progressively issue more information sensitive securities such as risky debt, preferred stock and finally equity. This issue has been further modeled in Narayanan (1988) and Nachman and Noe (1994). Both these papers show the optimal choice of high quality firms would be to issue securities such as debt that are less sensitive to firm value. This "pecking order theory" of the security issuance 1 See SEC widens probe into 'death-spiral' schemes, The Financial Times, March 9,

8 decision receives empirical support in Myers and Sundar (1999) and Lemmon and Zender (2002). Fulghieri and Lukin (2001) relax the assumption of fixed information asymmetry and model a scenario where outside investors can produce costly information about the firm and therefore become more informed about firm value. In their model, firms of higher quality would encourage outside investors to produce information by issuing a more information sensitive security such as equity or preferred stock. Hence their model predicts the pecking order theory would be reversed when the costs of becoming informed are sufficiently low and the pecking order for security issues would hold when the costs of information production are large. Several empirical studies have examined the information conveyed by private placements of equity and debt (Hertzel and Smith, 1993, and Wruck, 1989). They find that in contrast to the negative information conveyed by public equity issues the market reacts positively to private placements of equity. Hertzel et al examine the long run stock price performance of firms that place equity privately. They find that similar to issuers of public equity, these firms tend to have poor stock price during the three year period following the private placement. In the first study on PIPE issuers, Hillion and Vermalen (2004) examine the short-term and long-term performance of issuers of reset and floating convertible issuers. They find that the issuers of these securities have poor operating and stock price performance at issue. They also find that approximately 85% of their sample firms have negative returns in the year following issuance and investors lose on average 34% of their investment. In related papers, Chaplinsky and Haushalter (2005), Gomes and Phillips (2004), and Brophy, Ouimet, and Sialm (2006) examine PIPEs. Chaplinsky and Haushalter (2005) examine the reason for investing in PIPEs in spite of their poor stock price and operating performance 7

9 following the issue. They find that after including the discounted price that PIPE investors receive for their investments, these investors do substantially better than the existing shareholders. Gomes and Phillips (2005) examine the public versus private market security choice of firms. They show that conditional on the public security choice, firms with higher information asymmetry will prefer to issue debt securities that are less sensitive to firm value. They also document a reversal of the pecking order in the private markets. Firms with higher degree of information asymmetry are more likely to issue equity. Brophy et al. (2006) examine the role of hedge funds in the PIPE markets. They find that companies that get financing from hedge funds do worse than companies that are financed by other investors. They also find that issuers where investors have significant repricing rights tend to do worse. Our study differs from the above mentioned studies in several ways. Our primary objective is to examine the information conveyed by PIPEs and its relationship to security characteristics and the level of information asymmetry. We focus on the information sensitivity of the security issued based on two characteristics, the seniority of the security and whether it offers any other protection to new investors. We classify issues that are senior securities, such as convertible debt, as an offering that provides downside protection". We classify issues that include a reset or floating rate feature as one that provides "price protection". We examine the information conveyed by these securities through the stock price reaction to the security announcements and we also relate the announcement period returns to firm characteristics through cross-sectional regressions. We further analyze the long-term returns to PIPE issuers by following the stock price performance for the three year period 8

10 following the issuance. Our results show the market does not fully incorporate all the information at the announcement of these security issues. Loughran and Ritter (1997) and Spiess and Affleck-Graves provide evidence on the long run stock price performance of public issuers of equity. They find that these issuers have poor long run stock price performance following the issue. These findings are in clear violation of market efficiency. Other behavioral theories hypothesize that investors can either underreact to the issue announcement or be overoptimistic at issue announcement. Daniel, Hirshleifer, and Subramanyam (1998) formalize the underreaction hypothesis. In their model investors are overconfident and react only partially to news announcements. They show that subsequent abnormal performance can continue in the direction of the announcement period returns. Loughran and Ritter (1997) hypothesize that investors can place more emphasis on recent performance stock price and operating performance instead of putting more weight on long-term averages. Consistent with this view, they find that firms tend to issue equity at times when the stock price and operating performance are at its highest. Issuers suffer declines in both operating and stock price performance following the issue. In this part of our study we examine whether the long run stock price performance of the PIPE issuers are consistent with a market efficiency hypothesis, an underreaction hypothesis, or an overoptimism hypothesis. III. Data and Method: We first describe the data and then the method. Our data set is for the period A. Data We obtain our data on PIPE issuers during the period from Sagient Research's PlacementTracker Service. The data has information on the registration date, security issued, the 9

11 offer discount (for equity), the placement agent and the investor identity. We include in our sample those issuers that meet the following criteria: 1. The issuer is a publicly traded firm. 2. Only the first transaction is included per firm year. 3. Stock return data are available in the Center for Research in Security Prices (CRSP) database with sufficient returns to estimate the market model. 4. Firm data is available in Compustat. Our screening of the Sagient Database for the above steps yields 2906 firms. Table I reports the distribution of the 2906 issuers by calendar year. We see a large increase in the number of issues over the period. The largest number of issues, 672, takes place in The lowest number, 74, is in The most frequently issued security is common stock with 1174 issues representing approximately 40% of the sample. The smallest number of issues, 40, is common stock with reset. The next largest number of issues, 619, is convertible debt with a fixed conversion price (fixed) and represents 21.3% of the sample. There are 421 issuers of convertible preferred fixed. There are 424 issuers of convertible preferred floating or reset and 228 issuers of convertible debt floating or reset. The data also indicate the number of firms issuing securities with floating or reset features have gone down substantially in recent periods because of the bad publicity associated with these securities. Table II reports summary sample statistics for the overall sample and by issuer type. All values shown are for the year prior to the issue (year 1). In the year before the issue, the mean (median) book value of assets of all the issuers is $ ($40) million. This is substantially smaller than the mean $1449 million book value of assets for public equity issuers reported in Gomes and Philips (2005). There are large differences in firm size between PIPE issuers. The 10

12 largest firms with a mean book value of assets of $ are issuers of convertible debt without reset or floating features. The next largest group of firms with a mean book value of assets of $ million issue convertible preferred without reset or floating features. The third largest group of issuers with a mean book value of assets of $ million issue common stock without any reset features. Issuers of equity with reset, convertible debt with reset and convertible floating with reset are smaller with mean book value of assets of $ million, $80.79 million and $51.92 million respectively. The issuing firms have a pretax operating cash flow of $51.18 million. Convertible debt issuers have a mean operating cash flow of $ million in year -1. Issuers of common stock with or without reset features, convertible preferred securities and convertible with floating or reset all have negative pretax operating cash flows in the year of the issue. Most studies on public equity issues show that issues have high operating cash flows in the year of the issue. The mean offer size for all PIPE issues is $63.73 million. The largest issue size is $ for convertible debt issuers and the smallest offer size of $9.68 million is for convertible preferred floating or reset issuers. The mean offer size to market capitalization for the PIPE issuers is The smallest offer size to capitalization ratio, 11%, is for common stock reset issuers and the largest offer size to capitalization ratio, 30%, is for convertible debt fixed issuers. Our sample of common stock issues are sold to investors with an average discount of 6.75% and the sample of common stock with reset issues have a discount of 22.02%. A large number of these issuers have negative pretax operating cash flows. One measure of their ability to survive is the cash depletion rate, calculated as the absolute value of cash flow from operations divided by cash and short-term securities. We calculate this measure only for firms 11

13 with negative operating cash flows. The inverse of this ratio gives the number of years the firm can survive using the existing cash and marketable securities. The overall value of the cash depletion rate is showing the firms cannot survive even for a year with the existing cash. The lowest value, 2.86, is for common stock issuers and the highest value, is for issuers of convertible preferred fixed. The average number of years that the issuer has been listed on an exchange is 7.74 years; the oldest firms with a mean age of years are convertible debt fixed issuers. The youngest issuers with an average age of 6.51 years are convertible preferred fixed issuers. We calculate the Tobin's q as the (book value of assets-book value of equity+ market value of equity)/book Value of assets. The average Tobin's q of the issuing firm is 4.17 years. This indicates that these are firms with relatively valuable growth opportunities. Issuers of convertible debt fixed and convertible preferred fixed have Tobin's q ratios of 3.86 and 3.01 respectively indicating that they are relatively more mature firms. B. Method We examine the stock price reaction to the issue announcement, using the standard eventstudy method of Brown and Warner (1985) to calculate the daily excess returns. Average daily abnormal returns are computed in a two-step procedure, using stock price data from CRSP. We report results using both the CRSP equal-weighted and value- weighted indices as market proxies. First, we estimate the parameters of a single-factor market model for each firm. We use the returns from day 255 to day 46 to estimate each firm s alpha and beta coefficients. Second, we estimate the excess return by subtracting a firm s expected daily return from its 12

14 actual return. Cumulative abnormal returns are calculated by summing the abnormal returns over the period from day 1 to day +1, where day 0 represents announcement of the issue. We further examine the relationship between the stock price reaction to the issue and firm variables, using cross-sectional regressions. We regress cumulative abnormal returns against EBITDA divided by the book value of assets for the issuer, firm size, offer size, cash depletion rate and three dummy variables to measure the sensitivity of the security issued to firm value. The three dummies are: 1. PP (Price Protection): A dummy variable set equal to one if the security issued offers price protection to the investor by having a floating or reset feature. We set the PP dummy to 1 for issues of equity with reset or convertible preferred with floating or reset. 2. DP (Downside Protection): A dummy variable set equal to one when the security issued has a debt like feature and does not offer price protection to investors. Issues of convertible debt fixed would have this dummy set equal to DP_PP dummy (Price and Downside Protection): A dummy variable set equal to 1 if the security issued offers both price and downside protection to investors. Hence an issue of convertible debt floating or reset would have a value of 1 because it offers both price protection (through the floating or reset feature) and downside protection (through the debt feature) to investors. These three dummy are mutually exclusive. This enables us to avoid multicollinearity in the cross-sectional regressions relating announcement period returns to firm variables. When DP 13

15 or PP take on a value of one, this indicates that the security offers either downside or price protection to investors. When the security issued offers both downside and price protection to investors the DP_PP dummy is set equal to 1 and both DP and PP dummies are set equal to zero. We use EBITDA divided by the book value of asset as a proxy for the operating performance of the issuer at issue. When there is asymmetric information and outside investors are not able to produce information regarding the firm because information production is costly, we can expect investors to use the current operating performance as a predictor of future operating performance. This would imply a positive coefficient for this variable. However, if the cost of information production is not sufficiently large, outside investors will be able to expend resources and get information about the future cash flows. Other less informed investors will revise their estimates of the issuer after observing the actions taken by the more informed. These revisions would be larger for firms that currently have lower or more negative operating performance. Therefore the information production hypothesis would predict a negative coefficient for the EBITDA/Book Value of Assets. The second variable we use is firm size proxied by the book value of assets. We can expect the information asymmetry between insiders and outside investors to be lower for larger firms since they attract more analysts and larger amounts of public information is available. Hence we can expect a negative coefficient for the size variable. The third variable we use is the offer size scaled by the market value of equity. The adverse selection hypothesis would predict a negative coefficient for this variable (Myers and Majluf, 1984). The negative relationship between the offer size and announcement period returns is confirmed by Mikkelson and Partch (1986) and Masulis and Korwar (1986). However if outside investors are able to produce costly information on firm value, we can expect that they 14

16 will invest more when the firm is undervalued. Rational outsiders would revise their valuation upward as the informed outsiders take larger positions. The information production hypothesis would predict a positive coefficient for this variable. The next variable we use is the cash depletion rate of the firm, calculated as the absolute value of cash flow from operations divided by cash and short-term securities. This variable is set equal to zero for firms that have a positive cash flow from operations. Firms with higher values have less cash available to continue their operations and will need larger infusions of cash. The adverse selection theories would predict a negative coefficient for this variable and the information production theories would predict a positive coefficient for this variable. The last firm variable that we use is the age of the firm given by the number of years since it has been listed on the exchange. We can expect older firms to be larger, attract more analysts, and produce more publicly available information. The incremental information conveyed by these issues would be lower and hence we can expect a negative coefficient for this variable for the information production hypothesis. The last three variables are the mutually exclusive dummy variables: PP for price protection; DP for downside protection; DP_PP for both downside and price protection. Issuers of securities with price protection such as equity reset and preferred reset are issuing a security that is less sensitive to firm value. The adverse selection theories would predict a positive coefficient for this variable while theories on information production predict a negative coefficient for this variable. Issuers of securities with downside protection only such as convertible debt are issuing a security that is less sensitive to firm value when compared to a more sensitive security such as equity. Therefore, the adverse selection theories would predict a positive coefficient for this 15

17 variable. The information production theories would predict a negative coefficient for this variable. The last dummy variable is DP_PP set equal to 1 when the security issued offers both downside and price protection to investors. The adverse selection theories would predict a positive coefficient for firms that issue securities that have downside protection. The information production theories would predict a positive coefficient for this variable because outside investors are investing in a security that offers protection against any losses in firm value. Since these securities are the least sensitive to firm value, the information production theory would expect DP_PP to be negative and larger in magnitude than both DP and PP. We examine both the announcement-period returns and long-run post-issue holdingperiod returns. We estimate announcement-period excess returns by cumulating daily marketmodel prediction errors for a ten day (Day 5 to +5) and a three day (Day 1 to +1) Periods. We have the issue dates for the sample of issuers. These firms are not allowed to announce these issues before the placement of the securities and hence, we can use the dates given by Placement Tracker. C. Measuring Long Run Stock Returns Fama (1998) find the buy and hold abnormal return method for measuring long run stock returns problematic because it does not account for potential cross-sectional dependence in returns. To address this possibility, we estimate abnormal returns using the calendar-time portfolio approach used by Mitchell and Stafford (2000). A portfolio of equally weighted firms is formed for each month in calendar time. Sample firms are included in the portfolio if the sample firm event month for issuing the PIPE is in the prior m months. We report the following regression for m = 9, 18, 27, and 36 months: 16

18 R R = α + β( R R ) + ssmb + hhml + (1) pt ft mt ft t ε πτ R pt is the return on the sample portfolio, R ft is the one-month T-Bill rate, R mt is return on all NYSE, Amex, and Nasdaq firms, SMB is the return on a portfolio of small firms minus the return on a portfolio of large firms, HML is the return on a portfolio of high book-to-market ratio firms minus the return on a portfolio of low book-to-market ratio firms. These are the Fama-French (1993) factors. If the model adequately describes then the expected value of the intercept, α, which describes the monthly abnormal return, is zero under the null hypothesis of no abnormal performance. As shown by Fama and French, the three factor model is unable to completely describe the cross section of expected returns particularly for small, low book to market stocks that comprise a large part of the sample. IV. Results First, we report the overall issuers announcement period returns for different intervals around the issue announcement and the results based on issuer type. Second, we provide the results of regressing the announcement period returns on firm and security characteristics. Finally, we provide the long run stock price performance of the issuers. A. Announcement Period Returns Three-day abnormal stock returns to the issuing firms are reported in Table III, Panel A. For our sample of 2906 issuers, the overall announcement period return is positive and significant for all intervals. This implies that the market perceives PIPE issuer as increasing firm value. For the day -1 to +1 interval the mean (median) return is 2.24% (0.24%)and significant at the 1% level. We find that mean (median) return for the -5 to +1 window is 2.64% (0.34%) and 17

19 the mean (median) return for the -10 to +2 window is 3.64% (0.41%) and significant at the 1% level. Panel B reports the results for the issuers of common stock with and without price protection. We report the results only for the -1 to +1 interval. We find that the mean (median) stock returns are positive, 3.76% (0.82%) and significant for firms that issue equity without any reset or floating feature. The mean (median) stock price reaction is not significantly different from zero for firms that issue equity with a floating or reset feature. The difference between the two groups is not significant at conventional levels. Panel C reports the results for convertible preferred issuers. We find the stock price reaction is highest with a mean (median) return of 4.62% (1.33%) for issuers of convertible preferred stock without price protection. The mean (median) return for issuers of convertible preferred stock with price protection is not significantly different from zero. The two groups are significantly different from each other at the 1% level. Panel D reports the results for issuers of convertible debt fixed and convertible debt floating or reset securities. We find that both the mean and median announcement period returns are not significantly different from zero for both issuers. Panel E reports the results for the securities based classified by price and downside protection. We find the announcement period returns are positive and significant, with mean (median) returns of 3.98% (0.95%) for the 1595 issuers that offer securities with no downside or price protection. The mean (median) announcement period returns are zero (negative) for the 1311 issuers that offer both price and downside protection. The difference between the two groups is significant at the 1% level. 18

20 Our univariate tests show the information conveyed by issuing securities that are sensitive to firm value is positive and significantly different from the information conveyed by offering securities that are less sensitive to firm value. However our tests do not control for other firm characteristics that can vary across our sample of issuers. Hence in Table IV we report the results of cross-sectional regressions where we regress the announcement period returns for different intervals against security characteristics such as price protection, downside protection and other firm characteristics. In model 1, we regress the announcement period returns against dummy variables for price protection, downside protection, and both downside protection. The other variables we use are the EBITDA/Book Value of Assets, Offer Size/Market Value of Equity, Cash Depletion Rate and the age of the firm. In model 2 we replace EBITDA/Book Value of Assets with firm size, proxied by the log of the Book Value of Assets. We do not include both EBITDA/Book Value of Assets and firm size in the same regressions because we find that they are positively correlated; larger firms tend to have higher EBITDA/Book Value of Assets ratios. The first column of Table IV reports the results with the -1 to +1 returns as the dependent variable. The coefficient of EBITDA/Book Value of Assets is negative and significantly different from zero at the 10% level. This shows the announcement period returns are more positive for PIPE issuers that have lower EBITDA/Book Value of Assets indicating the market perceives that outside investors who invest in these PIPEs produce positive information about the issuer. The value of this information is higher for firms that are not currently doing well. The coefficient of Offer Size/Market Value of Equity is positive and significantly different from zero at the 1% level. This shows the market reaction is more positive for larger issues. The results are in sharp contrast to the evidence on equity issues in the public markets which find a negative 19

21 between offer size and the announcement period returns (Asquith and Mullins, 1986). The coefficients of burn rate and the age of the firm are not significantly different from zero. The coefficients of our three dummy variables, PP (the dummy set equal to 1 if the security offers price protection), DP (the dummy set equal to 1 if the security offers downside protection), DP_PP (the dummy set equal to 1 if the security offers both price protection and downside protection) are all negative and significantly different from zero at the 1% level. The coefficient for the price protection dummy indicates the announcement period returns would be 3% lower than a PIPE issuers when the firm offers equity securities that offer price protection alone. The R-squared of the regression is 0.03 and the F-value is 11.61, significant at the 1% level. The second column of the regression reports the results of model 2. The coefficient of Firm Size is negative and significant at the 1% level. This indicates the announcement period returns are lower for larger firms. This can possibly happen because there is more public information about larger firms and the incremental value of the information conveyed by issuing these securities is lower. The coefficients of the other variables are similar to model 1 and therefore we do not report them here. Columns 3 and 4 report the results of regressions with the days -5 to +1 as the dependent variable. Columns 5 and 6 report the results of regressions with the announcement period returns from days -10 to +2 as the dependent variable. In all regressions we find the coefficients of price protection, downside protection and both downside and price protection all negative and significant at the 1% level. Overall, we find the announcement period returns are positive and significant for issuers of securities that are more sensitive to firm value. We also find the announcement period returns are higher for smaller firms, larger issue sizes and firms that have lower EBITDA/Book Value of 20

22 Assets at issue. Our results are broadly consistent with the predictions of the information production theories of security issuance. Outsiders who invest in the PIPE issues generate information about firm value. The market revises their estimates when the outsiders invest in these new issues. B. Long Run Stock Price Performance In the previous section we analyzed the market reaction to the announcement of PIPE issues. In this section we analyze the long run stock price performance of these issuers. Loughran and Ritter (1997) and Spiess and Affleck-Graves document that public issuers of equity have poor long run stock price performance following the issue. Daniel, Hirshleifer and Subrahmanyam (1998) formalize the underreaction hypothesis. They show that subsequent abnormal performance can continue in the direction of the announcement period returns. In this part of our study we examine whether the long run stock price performance of the PIPE issuers are consistent with market efficiency hypothesis, the underreaction hypothesis, or the overoptimism hypothesis. Panel A of Table V reports the results of issuers that offer no price or downside protection. These are firms that issue straight equity or preferred stock with no floating or reset features for the nine months, eighteen months, twenty seven months and 36-month periods following the issue. The intercept in each regression measures the risk-adjusted abnormal performance of the PIPE issuers. The first column reports the results for the nine month period following the issue. The intercept is positive, 1.36 and significant at the 10% level indicating the issuing firms have a monthly excess return of 1.36% during the nine month period following the issue. The adjusted R-square of the regression is 78.82%. The coefficients of the regressions for 18 months, 27 months and 36-month intervals are not significant. 21

23 Panel B reports the results for issuers that offer downside protection. These are convertible debt issuers without any floating or reset features. The intercept of the regressions is not significant for any time periods showing that these issuers perform at the same level as their matching portfolios. Panel C reports the results for issuers that offer price protection. These include equity and preferred stock with reset or floating features. The first column reports the results for the 9 month period following the issue. The intercept is negative, , and significant at the 1% level showing that these issuers under perform their matched portfolios by 3.19% per month during the 9-month period following the issues. We find the intercept is less negative, , and, significant for the 18-month period also. The intercept is negative but not significant for the 27-month and 36-month periods following the issue. These results offer some support for the underreaction hypothesis of Daniel, Hirshleifer and Subramanyam (1998). The market only partially incorporates the information conveyed by issuing securities that offer price protection. Panel D reports the results for issuers that offer both price and downside protection to investors. These are convertible issues with a reset or floating feature. The first column reports the results for the 9-month period following issuance. The intercept is negative, , and significant at the 1% level suggesting that these issuers under perform the market by 2.48% per month during this period. We further find the intercept is negative and significant for the 18- month, 27-month and 36-month periods following the issuance. These results suggest the market only partly incorporates the information conveyed by these issuers. Hillion and Vermalen, 2004 find similar results for their sample of convertible debt issuers. Overall our results in this section offer some support for the underreaction hypothesis and are in violation of market efficiency. We find the positive announcement period returns to equity 22

24 and preferred issuers with no reset or floating features continue to persist for up to 9 months following the issue. Similarly the negative announcement period returns to issuers of less sensitive securities such as convertible reset or floating and equity reset continues to persist for up to 3 years following the issue. Although we find that issuers of convertible reset or convertible floating have poor stock price performance, the new investors in their PIPEs might still make positive abnormal returns because they typically receive warrants and other securities as part of their investment. Chaplinski and Haushalter (2005) value these securities and show that these investors are able to get positive returns inspite of the poor long run stock price performance of these issuers. V. Summary and Conclusions We examine the information conveyed by PIPE issues. A PIPE is essentially a hybrid between a public offering in the secondary markets and a private placement. The PIPEs are marketed as if they are private placements but the investor receives securities that can be traded immediately in the public markets. We examine the stock price reaction to the announcements of PIPE issues and the long run stock price performance for up to 36 months following the issue. In contrast to the finding that in the public markets the information conveyed by issuing equity is negative, we find that issuing equity PIPEs results in a positive announcement period returns. We further find that among each type of securities, the stock price reaction to the security issue is negatively related to the protections offered to the new investors. Our crosssectional regressions relating announcement period returns to firm characteristics show that the stock price reaction is more positive for larger issues and for firms that have poor operating 23

25 performance at issue. Overall, our evidence on the stock price reaction to PIPE issues is broadly consistent with the empirical predictions of Fulghieri and Lukin (2001). We also examine the long run stock price performance of the issuing firms. Our evidence is broadly consistent with the underreaction hypothesis formalized in Daniel, Hirshleifer and Subramanyam (1998). Issuers of securities that offer no downside or price protection to new investors such as equity or preferred stock with no reset or floating features outperform their matched counterparts during the nine month period following the issue. Issuers of securities that offer both price and downside protection to new investors such as convertible debt floating or reset under perform their matching portfolios for up to 36 months following the issue. References Allerhand, J., New Floorless Convertible Securities Generate Debate and Litigation, The New York Law Journal September 3, Asquith, P. and D. Mullins, 1986, Equity Issues and Offering Dilution, Journal of Financial Economics 15, Barber, B.M. and J.D. Lyon, 1997, Detecting Long-Run Abnormal Stock Returns: The Empirical Power and Specification of Test Statistics, Journal of Financial Economics 43, Brennan, M.,1985, Costless Financing Under Asymmetric Information, WorkingPaper, UCLA. Brown, S.J. and J.B. Warner, 1985, Using Daily Stock Returns: The Case of Event Studies, Journal of Financial Economics 14 (1), Brophy, D., P.G. Ouimet, and C. Sialm, 2006, PIPE Dreams? The Performance of Companies Issuing Equity Privately, Working Paper, University of Michigan. Chaplinski, S. and D. Haushalter, 2005, Financing Under Extreme Uncertainty: Evidence from PIPEs, Working Paper, University of Virginia. Cramer, James, A Deal With The Devil, Part I The Street.com, September 24,

26 Cramer, James, A Deal With The Devil, Part II, The Street.com, September 24, Daniel, K, D. Hirshleifer, and A. Subramanyam, 1998, Investor Psychology and Security Market Under and Overreactions, Journal of Finance 53, Dresner, S. and E.K. Kim, 2003, PIPEs: A Guide to Private Investments in Public Equity, Bloomberg Press. Fama, Eugene, 1998, Market Efficiency, Long Term Returns, and Behavioral Finance, Journal of Financial Economics 49, Fama, E. and K.R. French, 1993, Common Risk Factors in the Returns on Stocks and Bonds, Journal of Financial Economics 33, Fulghieri, P and D. Lukin, 2001, Information Production, Dilution Costs and Optimal Security Design, Journal of Financial Economics 61, Gomes, A and G. Phillips, 2004, Why do Public Firms Issue Private and Public Equity, Convertibles and Debt? Working Paper, University of Pennsylvania. Hertzel, M. and R.L. Smith, 1993, Market Discounts and Shareholder Gains for Placing Equity Privately, Journal of Finance 48, Hertzel, M., M. Lemmon, J. Linck, and L. Rees, 2002, Long Run Performance following Private Placement of Equity Journal of Finance, 27, Hillion P. and T. Vermalen, 2004, Death Spiral Convertibles, Journal of Financial Economics 71 (2), Kothari, S.P. and J.B. Warner, 1997, Measuring Long Horizon Security Performance, Journal of Financial Economics 43, Krishnamurthy, S., P. Spindt, and V. Subramaniam, and T. Woidtke, 2005, Does Investor Identity Matter inequity Issues? Evidence from Private Placements, Journal of Financial Intermediation 14, Lemmon, M. and J. Zender, 2002, Debt Capacity and Tests of Capital Structure Theories, Working Paper, University of Utah. Loughran, T. and J.R. Ritter, 1997, The Operating Performance of Firms conducting Seasoned Equity Offerings, Journal of Finance 52, Masulis, R.W. and A.N. Korwar, 1986, Seasoned Equity Offerings: An Empirical Investigation, Journal of Financial Economics 15,

27 Nachman, D. and T. Noe, 1994, Optimal Design of Securities under Asymmetric Information, Review of Financial Studies 7, Narayanan, M.P. 1988, Debt versus Equity under Asymmetric Information, Journal of Financial and Quantitative Analysis 23, Spiess, D.K. and J. A. Graves, 1995, Underperformance in Long-Run Stock Returns Following Seasoned Equity Offerings, Journal of Financial Economics 38, Sunder, L. and S.C. Myers, 1999, Testing Static Tradeoff Against Pecking Order Models of Capital Structure, Journal of Financial Economics 51 (2), Mitchell, M and E. Stafford, 2000, Managerial Decisions and Long Run Stock Price Performance, Journal of Business 73, Wruck, K., 1989, Equity Concentration and Firm Value: Evidence from Private Equity Financing, Journal of Financial Economics 23,

28 Table I. Distribution of PIPE Issues by Calendar Year and Type The distribution of 2906 PIPE issues by industrial firms by year and type of issue during the period From the original data of 5668 observations taken from the Placement Tracker company database, the following table includes issues of firms which have data available in both CRSP and Compustat, and where the type of PIPE issue is clearly identifiable. Year All PIPE Issues Common Stock Common Stock with Reset Convertible Preferred Fixed Convertible Preferred Floating or Reset Convertible Debt Fixed Convertible Debt Floating or Reset Total

29 Table II. Mean and Median Statistics for PIPE Issuers Mean and median values of characteristics of 2906 PIPE issues by industrial firms during the period The sample is taken from the Placement Tracker database. Except for offer size, values are for the year prior to the issue. Pretax operating cash flow is defined as net sales, minus cost of goods sold, minus selling and administrative expenses, but before deducting interest, depreciation, and amortization expenses. Cash flow to book value is pretax operating cash flow divided by the book value of total assets and measures the sample firm s raw performance. Capitalization is market price of shares times shares outstanding. Offer size is the dollar value of the issue. Issue discount is calculated as (1 (purchase price of issued equity /market price of present equity))x100. This measure is the percentage that the issue is below the market price of existing stock and is calculated only for equity offerings. If the cash flow from operations (compustat data308) is negative, cash depletion rate is the absolute value of cash flow from operations divided by cash and short term investments (compustat data01). If cash flow from operations is positive, cash depletion rate is set to zero. Age is the number of years firm is listed on an exchange before the issue. Variable Book Value ($MM) Mean (Median) Observations All PIPE Issues ( 40.58) 2906 Common Stock (30.16) 1174 Common Stock with Reset (27.81) 40 Convertible Preferred Fixed (67.72) 421 Convertible Preferred Floating or Reset (22.75) 424 Convertible Debt Fixed (244.25) 619 Convertible Debt Floating or Reset (28.48) 228 Pretax Operating Cash Flow in ($MM) ( -4.36) (-5.88) (-5.27) (-4.82) (-4.64) (8.03) (-4.43) 228 Cash Flow to Book (-0.17) (-0.26) (-0.27) (-0.13) (-0.27) (-0.04) (-0.18) 228 Capitalization ($MM) (84.88) (80.54) (73.28) (58.38) (49.99) (615.91) (62.92) 228 Offer Size ($MM) (8.00) (7.00) (5.00) (10.00) (5.00) (115.00) (4.50) 228 Offer Size to Capitalization 0.18 (0.11) (0.10) (0.06) (0.16) (0.10) (0.12) (0.09) 228 Issue Discount (% below market price) 6.75% (10.14%) % (9.94%) 40 Cash Depletion Rate (0.63) (0.69) (1.03) (0.64) (1.12) (0.00) (1.09) 226 Age 7.74 (5.00) (5.00) (5.00) (5.00) (5.00) (6.00) (5.00) 228 Tobin s q 4.17 (2.46) (2.88) (2.88) (1.82) (2.96) (2.12) (