Financial contracts in PIPE offerings: the role of expert placement agents.
Private Investments in Public Equity (PIPEs) have emerged in the last decade as an important source of financing, in particular for poorly performing companies (Brophy, Ouimet, and Sialm, 2009). From 1999 to 2012, there were about 15,000 US PIPE offerings that combined, raised over $500 billion. (1) Yet, when compared to many other types of equity issuances that have been studied extensively by researchers, we know relatively little about the inner workings of PIPEs.
A key feature of PIPEs is the use of complicated contract terms that allocate contingent cash flow and control rights between investors and issuers (Hillion and Vermaelen, 2004; Chaplinsky and Haushalter, 2010). One view regarding PIPE contract terms is that sophisticated investors include them to exploit the urgent financing needs of struggling companies. According to this view, the typical PIPE contract design is harmful to the issuer as it gives the investor too many protections and too few restrictions. An alternative view, emphasized by financial contracting theory, is that this contract design is the value-enhancing solution to the severe frictions (moral hazards, adverse selection, etc.) that characterize the PIPE market. According to this view, state contingent cash flow and control rights are able to lower transaction costs below what can be achieved by only adjusting the offer price.
In this paper, we undertake a comprehensive investigation of 14 key contract terms in PIPE offerings. We document the frequency of each term and describe how it protects the investor at the expense of the issuer. We explain how investor-friendly terms can minimize financing frictions thereby allowing investors to offer more issuer-friendly pricing. Building on this insight, we analyze how the inclusion of investor-friendly terms in PIPE contracts relates to the expertise of the placement agent who advises the issuer during the offering process. Our main finding is that issuers with high-ranked agents agree to more investor-friendly terms.
One explanation as to why the contract design and the agent's expertise are related is that there could be endogenous matching in the PIPE market. Our analysis demonstrates that this matching does exist. We find that high-ranked agents generally match with larger and higher quality issuers, perhaps because these issuers wish to certify to PIPE investors that they have less pronounced investment risks (Dai, Jo, and Schatzberg, 2010). Certification from expert agents lowers the risks that investors face, thereby decreasing the need for investor-friendly terms that protect against such risks. However, as demonstrated by Chaplinsky and Haushalter (2010) and our own analysis, lower risks and other favorable issuer characteristics are associated with PIPE contracts with fewer investor-friendly terms. As a result, the matching explanation predicts a correlation between agent expertise and investor-friendly terms that is precisely the opposite from what we observe in the data. Thus, even though there is endogenous matching and certification in the PIPE market, the pattern of matching cannot explain why issuers with high-ranked agents agree to more investor-friendly terms.
To reconcile our findings, we build on Tirole's (2009) argument that a contracting party may find it difficult to compute the payoff consequences of complex state contingent contract terms. This difficulty is particularly acute in situations such as a PIPE offering where financial contracts include a myriad of complicated cash flow and control right terms. PIPE investors, hedge funds, private equity funds, and venture capital funds, are well versed in financial contracting, so they can compute the payoff consequences of negotiable terms with relatively high accuracy. In contrast, PIPE issuers have limited experience with negotiating complicated financial contracts, and their urgent need for capital gives them little time to decipher the exact payoff consequences of the terms. Consequently, issuers may form biased assessments of each negotiable term, by which they either "optimistically" underestimate or "pessimistically" overestimate how much expected surplus the term will transfer to investors. Although PIPE issuers may be susceptible to both types of biases, their poor preissue performance makes them more prone to overestimating the consequences of investor-friendly terms that would occur if postissue performances are weak.
Issuers' biases can affect how PIPE contracts are structured and negotiated. Concretely, the investor would have to refrain from including any investor-friendly term for which the issuer, due to her overestimation bias, demands a pricing compensation that is above the term's correct payoff consequences. Certain excluded terms could increase the overall surplus of PIPE offerings by signaling preinvestment issuer quality (i.e., overcoming adverse selection) or providing issuers with favorable postinvestment incentives (i.e., overcoming moral hazards). Consequently, both the contract design and the postissue surplus can be vastly different between an uninformed issuer who is susceptible to biases and an informed issuer who understands the payoff consequences of the terms.
We expand on this insight and confirm that placement agents play a critical role in the PIPE market by helping their issuer-clients understand the payoff consequences of the contract terms. Placement agents take a leading role in PIPE offerings. They spend considerable time briefing their clients on the meaning and the importance of the various terms. They also actively participate in contract negotiations. There is substantial heterogeneity of agent expertise on contracts in the PIPE market. High-ranking agents are more capable than their low-ranking counterparts of helping their issuer-clients better understand the payoff consequences of the contract terms. One reason for this difference is that expert agents have more experience in negotiating the types of terms that are typical of PIPE contracts. Our empirical evidence indicates that agents with greater market share do provide better service during contract negotiations. Our empirical evidence is derived from a sample of 2,414 intermediated US PIPE offerings that were completed from 1999 to 2012. We aggregate 14 functionally distinct contingent cash flow rights into an Investor-Friendly Index (IFI). As in Chaplinsky and Haushalter (2010), we treat warrants as a contract term rather than as a pricing adjustment. (2) This approach is consistent with the contract-theoretical argument that contingent cash flow rights are used as a contractual solution to financing problems rather than simply as a price adjustment (even though their inclusion also affects how much investors pay for a security).
Our IFI captures how many investor-friendly terms are included in PIPE contracts and how many issuer-friendly terms are excluded. (3) In our baseline tests, we calculate IFI by adding the individual contract term dummies together. This aggregation method, which is similar to the method used by Gompers, Ishii, and Metrick (2003) to aggregate corporate governance provisions, has the advantage of being simple and transparent. However, one obvious problem with such an equal-weighted addition is that it implicitly assumes that all contract terms have equal complexity and importance. To address this problem, we conduct several robustness tests in which we calculate IFI in different ways, including specifications where we do not treat warrants as a contract term. We find qualitatively very similar results from our baseline and robustness tests, signifying that our main findings are not simply due to the aggregation method we use.
Our first result is that issuers advised by expert agents have higher IFI scores than companies advised by low-ranking, "nonexpert" counterparts. We validate this finding in multivariate tests where we control for several company and investor characteristics, as well as the choice between a traditional versus a structured PIPE offering (Brophy et al., 2009).
Next, we examine how agents influence the relation between the use of investor-friendly terms and the offering price. More precisely, we determine how the IFI varies with the pricing discount, defined as the difference between the PIPE offer price and the traded common equity price. We analyze a "raw" pricing discount that does not account for the pricing effects of warrants and other contingent cash flow rights (otherwise we would include these terms on both sides of the regression). We find that companies that agree to more investor-friendly terms are compensated with lower PIPE discounts, demonstrating a trade-off between terms and price. Moreover, we confirm that this trade-off varies with the expertise of the placement agent. Issuers advised by expert agents receive discounted compensation for investor-friendly terms that is about three times greater than that received by issuers advised by low-ranking advisors. This result indicates that expert agents give their issuer-clients advice that allows them to extract higher pricing compensation in exchange for investor-friendly terms.
Finally, we analyze the long-run stock returns to PIPE issuers to further determine whether expert agents provide good or bad advice. We find that stock returns at 12,24, and 36 months after a PIPE offering date are significantly greater for companies with expert agents than for companies with low-ranking agents. This finding runs against the expectation that expert advisors recommend contract designs that allow investors to expropriate issuers. As additional direct evidence that investor-friendly terms in PIPE contracts are generally not linked to such expropriation, we find that long-run stock returns are significantly greater for companies that have contracts with more investor-friendly terms. These findings suggest that expert agents offer advice that is beneficial to their issuer-clients.
In summary, our empirical results indicate that the expertise of issuers' placement agents relates to the structuring and the performance of PIPE offerings. These patterns are consistent with the agent role that we propose. Agents reduce imbalances in contracting parties' assessments of contracts. An issuer advised by an expert agent comes to share the investor's understanding of the consequences of the negotiable terms. The parties thereby agree on how large a reduction in the pricing discount is appropriate for each investor-friendly term. As a result, such a negotiated contract will include all investor-friendly terms that mitigate moral hazards, adverse selection, and other financing problems. This explains our findings that expert agents are associated with more investor-friendly terms. The greater overall surplus (from mitigated problems) associated with such a contract design explains our findings that expert agents are associated with higher long-run returns. Moreover, expert agents prevent investors from undercompensating issuers for terms that have greater payoff consequences than anticipated by biased issuers. This confirms our findings that expert agents are associated with a greater reduction in pricing discounts in exchange for the negotiated terms.
From a conceptual perspective, the information that placement agents provide to PIPE issuers about contract terms could have several possible benefits. First, the involvement of expert agents causes investor-friendly terms to be included if, and only if, these contingent cash flow rights increase the overall surplus. This allows the negotiated contract to address the severe financing problems that challenge PIPE issuers. In addition, the involvement of expert agents can mitigate the various forms of underinvestment that arise from issuers' lack of understanding of the consequences of the contract terms. One form of underinvestment occurs when an uninformed issuer demands overly steep compensation for terms that the investor sets as a necessary financing condition. Another form of underinvestment occurs when an uninformed issuer refuses to conduct a PIPE to avoid expropriation at the hands of an informed investor. Moreover, the involvement of expert agents can simply lower "haggling costs" during contract negotiations, allowing offerings to be completed more swiftly. This benefit is important as most PIPE issuers urgently need capital.
Our paper contributes to the large body of literature that investigates the role of placement agents and financial advisors in equity issuances. Unlike existing research on initial public offerings (IPOs), seasoned equity offerings (SEOs), debt issuances, and private placements that focus on the association between an agent's rank and offer price (Carter and Manaster, 1990; Beatty and Welch, 1996; Datta, Iskandar-Datta, and Patel, 1997; Livingston and Miller, 2000; Cooney et al., 2001; Logue et al., 2002; Ljungqvist and Wilhelm, 2002; Loughran and Ritter, 2004; Fang, 2005), we present novel evidence pertaining to the association between agent rank and contract design, as well as the associated effect on pricing and returns.
It is important to emphasize that the placement agent role we discuss is conceptually distinct from the certification role that has been examined extensively in the IPO and SEO literature. Certification reduces the imbalance in contracting parties' assessments of the object of contracting (e.g., asset valuation, growth opportunities, and risks). The role that we propose reduces the imbalance in the parties' assessments of the contracts themselves.
We also contribute to the growing literature on PIPEs (Hillion and Vermaelen, 2004; Dai, 2007; Brophy et al., 2009; Chaplinsky and Haushalter 2010; Chen, Dai, and Schatzberg, 2010). By relating the inclusion of contract terms to the agent's rank, our analysis extends the work of Dai et al. (2010), who determine that high-ranking PIPE agents are associated with lower pricing discounts. We validate their result and further confirm that this difference can be explained, in part, by high-ranking agents providing their issuer-clients with information about the consequences of the contract terms. We also extend their analysis by relating long-run stock returns to the ranking of PIPE agents and the inclusion of contract terms, respectively.
Finally, we contribute to the literature that examines the real-world structure of sophisticated financial contracts. Our analysis suggests that PIPE contracts often include many of the same protections that are found in contracts between venture capitalists and private entrepreneurial companies including special dividend rights, antidilution rights, first refusal rights, and redemption rights (Kaplan and Stromberg, 2003). The prevalence of such investor-friendly terms in PIPE contracts lends support to the argument of Chaplinsky and Haushalter (2010) that PIPE offerings are surrounded with particularly large information and agency problems. We find, similar to the studies of venture capital contracts, that PIPE contracts include more investor-friendly terms when these financing problems are more pronounced (Gompers, 1998; Kaplan and Stromberg, 2004; Bengtsson, 2011; Bengtsson and Sensoy, 2011).
The remainder of the paper is organized as follows. Section I describes the institutional setting in which PIPE agents operate, and contrasts the role of PIPE agents with that of IPO underwriters. Section II introduces our data and describes how we identify high-ranking placement agents. Section III discusses the terms included in the PIPE contracts that we study, and explains how we code them. In Section IV, we describe our empirical strategy. In Section V, we present the empirical results, while Section VI provides our conclusions.
I. Institutional Background on PIPE Agents
Unlike the relatively concentrated market for IPO underwriting, the market for advice concerning PIPEs is quite fragmented. (4) There are a large number of placement agents (368 in our sample) who compete against each other to provide advice for those companies that seek to issue a PIPE. The group of agents includes large investment banks and specialized corporate finance advisory firms, as well as "boutique" advisors who specialize in PIPEs and other types of private placements. Not surprisingly, a large degree of fragmentation means that many aspects of the relationship between issuers and agents are nonstandardized. Concretely, there is large variation in what is included in the legal contracts that issuers and agents sign, how agents are being compensated, and what work tasks agents fulfill.
The relationship between the issuer and the agent begins with informal discussions that quickly evolve into a formalized relationship. The milestone for this formalization is when the parties sign a Private Placement Agreement (sometimes also called an Engagement Letter), which is a legal document specifying the responsibilities of each party. Under this agreement, the issuer authorizes the agent to be the exclusive advisor for the offering, except in deals where more than one advisor is engaged. The issuer also promises to give the agent access to the company's detailed financial reports, management team, auditor, legal advisor, and consultants. This allows the agent to prepare the material that is necessary for pitching the deal to potential investors. To ensure that this information does not spread, the agent agrees to owe duty of confidence to the issuer.
The agent takes a leading role in initiating, progressing, and closing the PIPE deal. The agent controls the interactions between the issuer, the investor, the law firm, auditor, and other parties. This role differs from what advisors do in, for example, private equity, where the lead investor is more in charge of the transaction.
The first task of the agent is to conduct its business and financial due diligence on the issuer. In some instances, the agent draws on this information to prepare a formal Private Placement Memorandum that will be circulated to interested investors. In other cases, the agent only prepares an informal "teaser" where the main items regarding the company are summarized. More importantly, these documents do not reveal the identity of the issuer. This anonymity is a safeguard against any investor trying to exploit the information about the PIPE offering in order to trade in the issuer's stock.
The second task for the agent is to identify a group of investors interested in the deal. Here, the agent faces a delicate trade-off. If they were to solicit the deal to too few investors, then the offering could fail. If, by contrast, they were to solicit the deal to too many investors, then the SEC may view the PIPE as being a general solicitation of the sale of securities, which is illegal.
The third task, often viewed as the hardest one for the agent, is to "bring the deal over the wall," which is industry jargon for the point in time when the investor expresses their serious interest in the PIPE offering. Once the deal is over the wall, the agent will reveal the identity of the issuer and give the investor more detailed financial and operational information. Before this disclosure, the agent must secure a promise that the investor will not further disclose the information and refrain from trading the issuer's shares. These promises could be made orally or in a Non-Disclosure Agreement and a Non-Trading Agreement.
The fourth task of the agent is to advise the issuer on how to price the PIPE offering and which contract terms to use. Our empirical evidence demonstrates that agents fulfill this task in a way that benefits the issuer. The agent provides advice to the issuer both prior and during the negotiation of the deal structure that takes place with the investors. One part of such advice is to explain the meaning of typical PIPE contract terms. Since many of the terms are esoteric in the sense that they are used primarily in PIPE offerings, the issuer may not understand what they give up by agreeing to a particular term. Another part of this advice is to demonstrate the payoff consequences of each negotiable term. Calculating these payoff consequences can be very complicated. Each term has consequences that depend upon which state-of-the-world is realized, the investor's choice to use the right implied by the term, and which other terms are included in the contract. In practice, even expert agents may not be able to perfectly calculate payoff implications, but their experience with the contract terms means that they are better able to do so than their issuer-clients. Finally, the agent assists with closing, by collecting comfort letters from the auditors and legal opinions from the advising law firms.
The role of the placement agents in PIPE offerings is in some ways similar to the role of underwriters in IPOs. Both types of advisors interact with issuers, investors, lawyers, and auditors. Both types of advisors receive compensation of similar magnitudes, which is paid upon consummation of the deal. Both types of advisors are also informational intermediaries, who help issuers find interested investors and compile information that these investors want.
However, PIPE agents play an intermediary role that is not shared with IPO underwriters. Expert advisors can provide information about the payoff consequences of negotiable contract terms. There are three reasons why this role is unique. First, PIPE contracts include more complex terms than IPO contracts, which are fairly standardized. In addition, PIPE issuers, who are often struggling companies, may have less time to digest the details of the offering. An IPO is a big event in a company's history, and considerable time is spent on understanding its implications. Moreover, the market for advice is more fragmented in the PIPE market. The IPO market is concentrated to a small number of underwriters who arguably all understand how a typical IPO is structured. In contrast, there are many inexpert advisors in the PIPE market who lack this understanding. This suggests that there is more variation in the PIPE market with respect to how much information about contract terms expert advisors can give their clients about payoff consequences.
Another key difference is that PIPE advisors, in a typical deal, do not risk any capital in underwriting and after-market support. Thus, unlike IPO underwriters, PIPE advisors do not buy any shares during or after the offering. One implication of this difference is that PIPE agents can operate on a much smaller scale than IPO underwriters, who must have large pools of capital (in practice usually large investment banks). This may explain why the market for PIPE advice is so fragmented. Another implication is that PIPE advisors subsume much less of the investment risk. As a result, PIPE advisors may provide less certification of their clients than IPO underwriters do. Other advisory roles, including the one we emphasize in this paper, to provide information about contract terms, are important in the PIPE market.
We obtain data on 14,881 US PIPEs that closed from 1999 to 2012 from Sagient Research, Inc.'s Placementtracker database. (5) Following Chaplinsky and Haushalter (2010), we exclude structured equity lines (1,309 observations) and common stock reset PIPEs (77 observations). We also exclude PIPEs that have no disclosed agent information (2,288 observations). We further exclude issuers not covered by Compustat and Center for Research in Security Prices (CRSP) as we need data concerning company characteristics and stock returns for our analysis. This restriction eliminates 6,986 observations, signifying that many PIPE issuers are small and thinly traded companies. Of the remaining 4,221 observations with complete data, we exclude 1,807 "direct" PIPEs, which are offerings without a placement agent, from our main sample. We focus our attention on "intermediated" PIPEs, which are offerings with agents, because of the potential concern that the choice of employing an agent may be correlated with discounts or contract terms. In robustness tests, we examine both direct and intermediated PIPEs, and obtain qualitatively similar results. Our final sample covers 2,414 unique intermediated PIPE offerings, 1,181 unique issuers, and 368 unique agents. Panels A and B of Table I describe the year and industry distribution of the sample.
B. PIPE Offering Characteristics
In Panel C, we summarize the variables that capture the characteristics of the PIPE offering. The average offer size is $51.3 million, with the median at $ 11.1 million. For each PIPE, we calculate the pricing discount, which captures the percentage by which the PIPE price is below (or, in rare cases, above) the traded equity price. (6) For common stock PIPEs, we calculate discounts as the percentage difference between closing price one day before the closing date and the offer price. For PIPEs with fixed-price convertibles, discounts are measured as the percentage difference between the closing price one day before the closing date and the conversion price. For PIPEs with floating rate convertibles, we calculate discounts as the difference between the closing price one day before the closing date and the specified floor price. (7) The mean and median discounts are 9.5% and 8.6%, respectively.
About 90% of our sample represents traditional PIPEs, in which investors hold common stock or fixed-price convertibles. The remaining 10% are structured PIPEs where investors hold securities with repricing rights, such as floating price convertibles or convertible resets. We control for PIPE type in our empirical tests. More than half of all PIPE investors are hedge funds. The average investor has issued 6.6 within-sample PIPEs previously.
C. PIPE Issuers Characteristics
Panel D of Table I examines the characteristics of our sample PIPE issuers, including market capitalization prior to closing date, Analyst Coverage, CAR (-12, -1), Debt/Assets, EV (enterprise value)/Assets, R&D (research & development)/Assets, Intangible/Assets, Cash Burn, and the Altman's Z-score. Detailed definitions of these variables are provided in Appendix B. Consistent with earlier findings from the PIPE literature, our data demonstrate that PIPE issuers typically exhibit weak operating performance and display other characteristics consistent with a high degree of information asymmetry and agency cost. These issuers are small, with mean market capitalization measured on the day prior to the PIPE transaction of $271 million and a median of $91 million. Not surprisingly, given their size, they are not often followed by many (or any) analysts, with a mean analyst coverage equal to 2.2 and median of 1.0. Consistent with Brophy et al. (2009) and Chaplinsky and Haushalter (2010), we note a positive average cumulative abnormal return (CAR) for PIPE issuers before the offering. Furthermore, these firms often exhibit high EV/Assets ratios, with a mean of 4.1 and a median of 2.0, and R&D/Assets ratios, with a mean of 29.2% and a median of 12.0%. About 33% of the issuers are financially distressed or with a Z-score below 1.8.
That the typical PIPE issuer is a small, badly performing company means that the role of the placement agents that we propose, providing information about contract terms, is relevant in practice. A struggling issuer that is desperately in need of new capital to survive is forced to focus on improving his weak operating performance and, as such, cannot afford to devote much attention to understanding the intricacies or the terms of PIPE contracts.
III. PIPE Contract Terms
We now discuss the types of contract terms that are commonly used in the U.S. PIPE offerings. Our goal is to introduce the contract terms that are the focal point of our tests, present their sample frequencies, and discuss how we choose to aggregate them for our empirical analysis. We further describe how these terms translate into contingent cash flow rights between issuers and investors. Such allocation leads to a transfer of surplus between the contracting parties, and potentially affects the overall surpluses associated with PIPE offerings. Finally, we illustrate our contention that contractual arrangements characteristic of PIPE offerings are so complicated that a boundedly rational issuer may find it prohibitively difficult to compute the precise consequences of each term.
Table II describes the 14 contract terms that we later empirically analyze. For ease of presentation, we group these terms into three categories. The first category, investor protections (Terms 1 to 7), lists terms that attach various protections to PIPE investors' stocks. The terms in this category are favorable to investors at the expense of issuers. The second category, trading restrictions (Terms 8 to 11), lists terms that determine how investors can trade underlying stocks after the offerings. The third category, issuer rights (Terms 12 to 14), contains terms that grant to issuers the right to force investors to take certain actions. The terms in these latter categories favor issuers at the expense of investors.
A. Investor Protections
1. Dividend, Interest, and Warrants
Dividends, interest, and warrants are key deal features in PIPE offerings. About 26% of our sample contracts include provisions that entitle investors to fixed payments at prespecified dividend or interest rates. These dividends are sometimes cumulative, which means that an investor does not receive regular dividends, but rather a lump sum at a later time. About half of our sample contracts include warrants that allow investors to purchase, in the future, predetermined numbers of certain securities at specified prices.
Since warrants have cash flow consequences only in certain states of the world, their inclusion introduces a state-contingent payoff curve. Almost all contract theoretical work takes as its point of departure the proposition that a contingency-based payoff curve is not used to adjust pricing, but rather to provide an optimal solution to a variety of contracting frictions (e.g., agency or information problems). (8) Instead of including warrants, which are relatively complicated to price, the PIPE contracting parties can adjust the negotiated pricing discount. This adjustment is easier to carry out and has more obvious payoff consequences than including warrants. In other words, while there are many contract theoretical reasons why warrants are included in a financial contract, simply lowering the transaction price is unlikely to be the primary reason.
Chaplinsky and Haushalter (2010) analyze the use of warrants in PIPE offerings and find that consistent with contract theory, this term is more common when the issuer presents higher risks to the investor. Our data further support the notion that warrants are used as a contract term rather than a simple price adjustment. We find that the inclusion of warrants is positively correlated with the inclusion of other investor-friendly PIPE terms. This correlation should be negative if warrants were only used as pricing compensation for harsher terms (as more terms imply lower pricing; i.e., no use of warrants).
A similar line of reasoning may motivate our choice of treating interest and dividends as contract terms rather than as pricing factors. As derived in several theoretical models, the promise of regular payments to investors can remove suboptimal overinvestment decisions and induce managerial effort. Thus, these deal features are in place to overcome agency and information problems, rather than to provide investors with greater expected cash flows.
2. Investor Registration Rights
The key feature of PIPE offerings is that firms can close these offerings before filing registration statements with the SEC making PIPE offerings time efficient. However, this solution means that investors must assume the risk of illiquidity as they are not allowed to resell the acquired securities before the Registration Statement becomes effective. To mitigate this risk, PIPE contracts can force issuers to file Registration Statements within a short time period after offer closings. This protection is included in about 65% of the PIPEs in our sample. In some cases, this contract places a cap on the amount of capital that an issuer can draw down before the Registration Statement takes effect. Some contracts include penalty terms if registration fails, such as cancellation of the financing.
3. Antidilution Protection
Antidilution provisions protect PIPE investors against future financing at lower valuations than those of the current offerings. In its harshest form, antidilution prohibits an issuer from issuing or selling any equity securities (or securities convertible into equity) during a certain period after a PIPE offering. A typical period is 90 trading days following the effectiveness of the Registration Statement. This contract could also prohibit an issuer from issuing or selling these securities at a price below what the PIPE investor paid or below a specified benchmark price.
In a less severe form, antidilution terms protect investors from future price decreases by reducing offer prices (or, alternatively, conversion prices) to equal the lowest prices paid for any equity securities in future financing. In this case, an investor may also have the right to receive cash or additional common shares without additional consideration. (9) About 43% of our sample provides investors with some form of antidilution protection.
4. Right of First Refusal and Investor Call Options
Investor call options and investor right of first refusal give investors the right to purchase additional shares of a company's securities during a certain period in the future. Like warrants and antidilution, these contract terms are in place to protect investors against future dilution from price decreases or equity offerings at below market prices. About 20% of our sample includes a right of first refusal. Investor call options are included in 5% of our contracts.
5. Redemption Rights
Investor optional redemption is sometimes used to strengthen the liquidation rights of an investor's investment. This protection gives an investor the right to demand that a firm redeem the investors' claim upon a change in control. This conversion is typically priced at face value or at a certain percentage above face value (often higher than 100%; occasionally higher than 200%) in addition to the value of any accrued unpaid interest. About 12% of our sample includes an investor redemption option. Redemption rights matter as PIPE contracts often do not specify any contracted payments on which the issuer can default. Thus, redemption rights may offer an investor the only available means by which to force an issuer to repay an investment.
B. Trading Restrictions
Many PIPE offerings include provisions restricting how an investor can trade the underlying stock for a certain period after the offer closing. These restrictions, which are much less prevalent than investor protections we have previously discussed, favor issuers at the expense of investors.
The most common trading restriction prohibits investors from engaging in short transactions, hedging a company's common stock, or taking a position that is in excess of the value of shares owned (i.e., an offsetting long position) prior to the effectiveness of the Registration Statement. Sometimes, a contract also requires that an investor not engage in shorting or hedging for a longer period than the SEC's requirement, sometimes as long as the purchased PIPE security remains outstanding. About 8% of our sample explicitly forbids short selling before a certain date and 3% of the sample does not allow investors to hedge a company's common stock in excess of the value of shares owned before a certain date.
An additional trading restriction applies a lock-up period to a PIPE transaction. This provision prohibits an investor from selling any shares of an issuer's common stock purchased or received through the exercise of warrants for a duration typically lasting a few months following the closing. We find lock-up provisions in 3% of our sample.
Finally, in very rare cases (0.1 % of our sample), PIPE contracts prohibit investors from affecting any sales to the public of a company's shares for a certain number of days after the Registration Statement takes effect. This restriction is useful if a company plans a public offering (i.e., an SEO) shortly after the closing of the PIPE offering as it avoids price pressure from the investors' resale of shares to the public.
C. Issuer Rights
1. Company-Forced Conversion
PIPE contracts sometimes include company-forced conversion options in which shares held by investors will automatically convert into common stock under certain conditions, typically related to the issuers' stock performance during a given period following a PIPE offering. For instance, investors may have to convert their shares if stock price or weighted average stock prices exceed certain benchmark numbers. In an alternative formulation, issuers may have to convert their shares if daily trading volumes exceed certain levels for a specified number of consecutive trading days. In some extreme cases, such as a company's taking a 10,000-to-l reverse stock split, investors will also be forced to convert. About 14% of the contracts in our sample include company-forced conversion options.
The effect of these company-forced conversion provisions is to require investors to give up their contractual protections when companies attain a desired level of performance. In particular, if a company performs well, an investor will retain only the same rights as common shareholders, but if the same company performs poorly, an investor will retain superior cash flow rights. The usefulness of such a provision has been demonstrated in the extant theoretical work (see Bengtsson and Sensoy, 2011, for a discussion).
2. Company Put Option and Optional Redemption
About 12% of our sample PIPEs include company optional redemption provisions that provide issuers with the right to force PIPE investors to exercise redemption rights after a certain date or due to certain events. About 3% of the PIPEs in our sample include company put options where a company has the right to force an investor to purchase additional shares at a specified price. The effect of put options and optional redemption is that an investor would receive less favorable cash flow rights if a company were to achieve strong performance.
D. Overall Structure of PIPE Contracts
We construct five general observations concerning the contract terms in PIPE offerings. First, there is substantial variation in how the terms are included in PIPE contracts. (10) In addition, the overall structure of PIPE contracts is such that investors obtain superior rights if a company performs poorly. As company performance improves, an investor must relinquish these superior rights. Moreover, many contract terms appear to be designed to overcome agency and information problems. This is not surprising given that these problems are particularly severe for the types of companies that issue PIPEs. Further, many of the trading restrictions directly address incentives associated with hedge funds to undertake actions that can adversely affect a PIPE offering company (Hillion and Vermaelen, 2004). Finally, we note that trading restrictions and issuer rights increase with investor protections. (11) Investors request more contractual protection when they are restricted in trading/shorting and must give issuers more rights.
Given the complexity and intricacy of PIPE contracts, it is plausible that issuers, unless they are advised by an expert placement agent, face several hurdles when trying to understand the consequences of each negotiable term. First, a given term is described using intricate legal language that must be "translated" into a contingent cash flow right. In addition, the contingent nature of the cash flow right makes it necessary to compute the term's financial consequences in a range of future states of the world, and to estimate the probability distribution of these states. Furthermore, the consequences of the term cannot be analyzed in isolation as its consequences depend upon how the other terms are presented in the contract. For example, the value of an investor redemption right is lower if an investor also has an antidilution right (which allows share repricing following weak stock performance). Finally, the term's contractual consequences may be very different from its actual consequences due to renegotiations and hold-ups that occur when a contract is being enforced.
IV. Empirical Strategy
Our goal is to demonstrate that the expertise of the issuer's placement agent can affect the inclusion of investor-friendly terms in PIPE contracts. Our identification follows three steps. First, we aggregate the 14 individual contract terms to examine the aggregate investor-friendliness of the contract. In addition, we develop proxies to determine how well placement agents help their issuer-clients understand the payoff consequences of negotiable terms. Moreover, we ensure that endogenous matching between issuers and agents or agent certification do not bias our results toward finding the empirical relationships that we present.
B. Investor-Friendly Index (IFI)
The first step of our identification strategy is to aggregate the 14 PIPE contract terms to an index. We create an "Investor-Friendly Index" (IFI) that adds all contract terms favorable to investors (i.e., investor protections) and deducts all terms favorable to issuers (i.e., trading restrictions and issuer rights). We add seven in order to ensure that all contracts have a positive IFI (in total, we have seven possible deductions). In our baseline tests, we use simple addition where all terms have equal weights. As reported in Panel C of Table I, the mean of the IFI is 8.6, with a standard deviation of 1.7. The contract with the most investor-friendly contract terms has an IFI of 14, and the contract with the fewest has an IFI of five.
In our baseline tests, we calculate the IFI in a way that builds on the implicit assumption that an included issuer-friendly term is the same as an excluded investor-friendly term. To illustrate this principle, about one in ten PIPE contracts in our sample includes a forced conversion term giving an issuer the right to force a conversion of the investor's stock if certain conditions are met. The inclusion of an issuer-friendly term is assumed to equate to the exclusion of a hypothetical, inversely defined, investor-friendly term that does not give the issuer this conversion right. This assumption may appear strange at first glance, but it fits well with our reasoning pertaining to the placement agent's role in a PIPE offering. Uninformed issuers find it equally hard to understand the consequences of an inversely defined investor-friendly term and those of an actual issuer-friendly term, so these could be analyzed symmetrically.
Our method of calculating IFI also builds on the assumption that all contract terms are equally valuable and complex for all types of PIPE offerings. In practice, the payoff consequences of investor protection, trading restrictions, and issuer rights vary considerably across PIPE offerings, depending upon company and investor characteristics, as well as on market-wide conditions. The problem is that it is prohibitively difficult to come up with a model for estimating the relative importance of PIPE terms. Such a model would be very complex and based on many arbitrary assumptions to derive payoff consequences. Moreover, the model would also have to incorporate how difficult it would be for a boundedly rational PIPE issuer to estimate the payoff consequences. In light of these difficulties, our baseline method uses the simplest aggregation method, addition, which has the advantage of being the most transparent one. Our method reflects the spirit of Gompers et al. (2003), who, like us, use simple addition to aggregate corporate governance provisions.
In our robustness tests, we relax the above-discussed assumptions and recalculate IFI in different ways. To ensure that our results are not driven by one individual contract term, we exclude each term once and calculate IFI by adding the other 13 terms together (i.e., we recalculate IFI in 14 different ways). We also run tests based on each individual term, and on simulated IFIs based on randomly assigned weights on each of the 14 contract terms. We find that our main results are qualitatively similar in the baseline and these robustness tests. This reinforces our conclusion that our findings are not driven by a certain aggregation method.
C. Expertise of the Placement Agent
The second step of our identification strategy is to capture how well placement agents are able to credibly and accurately convey information about contract terms to their issuer-clients. Our sample covers 368 unique placement agents whose experience with PIPEs and other types of equity offerings varies. We base our identification on the assumption that placement agents with more experience are, on average, better able to convey information about contract terms. This assumption can be motivated by agents' acquiring this ability through experience (i.e., "learning-by-doing") or by experienced agents who are frequently chosen by issuers due to their superior ability.
We create two dummy variables that classify a PIPE agent as high ranking or low ranking. The first dummy, High Ranking Agent (C-M Ranking), is based on the Carter-Manaster (1990) rankings. These rankings were introduced by Carter and Manaster (1990) and extended by Carter, Dark, and Singh (1998) and Loughran and Ritter (2004). Underwriters with the lowest reputation are given a ranking of zero while the most reputable agents have a ranking of nine. The dummy High Ranking Agent (C-M Ranking) takes a value of one if the agent has a high (above seven) Carter-Manaster (1990) ranking, and zero otherwise. (12) As reported in Panel C of Table 1,19.7% of the PIPEs in our sample have a high-ranking agent based on the Carter-Manaster (1990) ranking.
A potential problem with the Carter-Manaster (1990) ranking is while it captures an agent's broader reputation for and experience with equity offerings, it does not necessarily report an agent's specific expertise in the PIPE market. Therefore, as a robustness check, we design a second reputation measure based on the agent's market share in the PIPE market. The second dummy, High Ranking Agent (Market Share), is determined by the following two steps. First, we calculate agent market share by comparing an agent's PIPE volume (in dollars) in the three preceding years to the total volume of intermediated PIPE offerings during the same period. (13) To give the reputation measure a sense of stability over time, we count the number of times the agent was on the top 15 lists over our sample period. The 10 agents that appear on the annual top 15 lists most often are designated as reputable agents. Appendix A lists the top 10 agents based on market share. Most of them have a Carter-Manaster (1990) ranking of eight or nine. As reported in Panel C of Table 1,15.0% of the PIPEs in our sample have a high ranking agent based on market share.
D. Matching and Certification
The third step in our identification strategy is to carefully consider how agent-issuer matching may bias our empirical results pertaining to contract term determinants. In an ideal experiment, we would eliminate this bias by identifying an exogenous instrument for how issuers and agents match. In practice, however, it is prohibitively difficult to find such an instrument. Any observable company characteristic that explains how issuers choose their placement agents may also explain how investor-friendly terms are included in the PIPE contract. This identification problem is not unique to our study, but is a general issue in studies of underwriters and placement agents.
We tackle this concern in two ways. First, our empirical tests include a large battery of issuer (and investor) controls in our main specifications. We find that our results are statistically significant after including these controls. This rules out the influence of matching based on observable issuer characteristics. Second, we infer that the process of unobserved matching is likely as there is positive assortive matching between issuers and agents. Researchers have observed there is positive assortive matching between IPO and SEO candidates and their underwriters (see Fernando, Gatchev, and Spindt, 2005, for evidence and a literature review). Positive assortive matching can have a variety of rationales. Better companies may want to employ high-ranking advisors in order to signal their superior quality (Titman and Trueman, 1986), or high-ranking advisors may prefer to advise better companies in order to maintain their strong reputations. Similar economics are likely at play in the PIPE market.
Panel A of Table III presents evidence that there is positive assortive matching in the PIPE market based on observable issuer characteristics. We report summary statistics separately for the 475 PIPE offerings associated with high-ranking agents based on the dummy High Ranking Agent (C-M Ranking) and the remaining 1,939 offerings. We use this sample split to examine how PIPE issuers match up with their placement agents. We find that issuers with high-ranking agents are significantly larger than their counterparts in terms of market capitalization prior to the offering. In particular, the mean capitalization is $466 million (median is $145 million) for issuers with high-ranking agents, while it is only $223 million (median is $80 million) for issuers with other agents. We also find that PIPE issuers advised by high-ranking agents exhibit better analyst coverage and greater financial leverage as measured by Debt/Asset. Furthermore, high-ranking agents appear to advise issuers with lower EV/Assets ratios, R&D/Assets ratios, and Intangible/Assets. This pattern, which is broadly consistent with expert agents matching with better company issuers, mimics the findings of Dai et al. (2010). (14)
With positive assortive matching, one possible role of placement agents is to certify the quality of their issuer-clients to investors. Agent certification may result as high-ranking agents have private information regarding the quality of the issuer through their due diligence, which is valuable as many PIPE issuers are small and struggling companies with limited coverage from analysts and the media. When a high-ranking agent signals (or directly transmits) private information about the high quality of the issuer to the investor, this reduces the investor's concern about agency and asymmetric information.
The problem with the matching and certification explanations is that they predict exactly the opposite empirical relationship between agent expertise and investor-friendly terms (and pricing) from what we observe in our data. As demonstrated in numerous theoretical models, an investor responds to less severe financing problems associated with higher quality issuers by lowering their demand for investor-friendly terms. This lower demand suggests that issuers with high-ranking agents are able to negotiate PIPE contracts with fewer terms, but receive less pricing compensation for the included terms. Thus, the matching and certification are difficult to reconcile with our findings. From an identification standpoint, this is an important conclusion. If matching and certification were to affect the inclusion of investor-friendly terms, then our results would be biased against finding the results we document. In other words, our findings might be even stronger if we were able to control for how agents and issuers match with each other.
E. Matching Based on Issuer Experience
We also examine how issuers and placement agents match up based on an issuer's experience with PIPE offerings. We note that 40% of the PIPE offerings in our sample are completed by first time issuers.
As presented in Panel B of Table III, we find that the probability that an issuer matches up with a high-ranking agent lies with how many PIPEs the issuer has negotiated previously. This finding adds additional evidence against the matching and certification explanations as companies that repeatedly issue PIPEs have more pronounced agency problems (Floras and Sapp, 2012). Moreover, this finding supports our explanation that the role of expert agents is to help their issuer-clients better understand the payoff consequences of negotiable terms. Repeated issuers have, by definition, negotiated PIPE contracts before, so they have a better understanding of the complex terms than first time issuer do. Thus, repeated issuers benefit less from matching with a high-ranking agent, consistent with our results.
V. Empirical Results
A. Investor-Friendly Contract Terms
First, we analyze how an agent's rank relates to the IFI. Since the dependent variable in the Poisson regressions must be nonnegative, we add a base value of seven to our IFI index (as there are seven terms that are unfavorable to investors). In unreported tests, we also run negative binomial regressions and obtain very similar results. Robust standard errors, which account for clustering at the issuer level, are reported in parentheses. Our results are robust to clustering on the agent, the issuer industry, or the offering year.
Our main variable is High Ranking Agent, which takes a value of one if the issuer is associated with a high-ranking agent and zero otherwise. In Panel A, we use Carter-Manaster (1990) rankings to identify high-ranking agents. For robustness, we also use the market share to identify high-ranking agents, with the results presented in Panel B. As control variables in all of the regression models, we include various measures that capture the issuers' level of information asymmetry and agency costs, as well as their historical performance. These variables include Ln(MV), Ln(Analyst), CAR (-12,-1), Debt/Assets, EV/Assets, RD/Assets, Intangible/Assets, Cash Burn Dummy, and Distress Dummy. The detailed definitions of these variables are provided in Appendix B. In Models (1) and (7), we use the full sample and include a dummy, Traditional PIPE, which takes a value of one if common stock or fixed price convertibles are offered, and zero otherwise, to control for PIPE type. One potential concern is that the correlation between High Ranking Agent and the IFI does not apply to all types of PIPEs. To rule this out, in Models (2) and (8), we restrict the sample to traditional PIPEs, while in Models (3) and (9), we restrict the sample to common stock PIPEs. In all of the above specifications, we also include investor-type fixed effects, industry fixed effects, and year fixed effects.
Our sample period includes the 2008 financial crisis period. To explore whether this period (which exhibits extreme liquidity constraints) affects PIPE contracting, we repeat our analysis in Models (4) to (6) and Models (10) to (12) by including a Crisis Dummy that is equal to one if the PIPE deal was closed in 2008 and 2009, and zero otherwise.
In Table IV, we find that high-ranking agents are significantly and positively associated with the IFI in all specifications. Examining the coefficients on the control variables, we confirm that the IFI is significantly higher for issuers with lower market capitalization, less analyst coverage, and higher EV/Assets ratios. Also, distressed firms are associated with higher IFI. These findings support the argument that investors request and receive more contractual protection when issuers encounter greater moral hazards, adverse selection, or other financing problems. We also determine that structured PIPEs, which are used by more distressed issuers (Brophy et al., 2009), also include more investor-friendly terms, further supporting this argument. Moreover, we find that more investor-friendly terms are included in PIPE contracts during the crisis period. This is consistent with the notion that investors demand more protection and have greater negotiation power in a tough environment for financing. In unreported robustness checks, we repeat the above-mentioned analysis for PIPEs invested by hedge funds separately and find consistent results.
The positive correlation between agent ranking and IFI is also robust to expanding the sample to include 1,807 direct PIPEs offerings where the issuer does not have a placement agent. We define direct PIPEs as a separate omitted category and find that these offerings have fewer investor-friendly terms than PIPEs with a low-ranking agent. This result suggests that inexpert agents can also provide some help to their issuer-clients pertaining to contract terms.
A potential concern is that high-quality issuers may signal their quality by utilizing high-ranking agents and agreeing to more investor-friendly terms. Our findings mentioned above that high-quality firms are often associated with higher IFI indicate that this is unlikely the case. To have a clearer understanding of the relation between firm quality and IFI without the influence of an agent, we conduct robustness checks using the sample of direct PIPEs. Consistent with the agency story, we find higher quality firms (larger, more analyst coverage, and less distressed) are associated with lower IFI.
Another potential concern is that the correlation between High Ranking Agent and the IFI is specific to the way we aggregate individual contract terms. To address this concern, we relax the above discussed assumptions and recalculate IFI in different ways. We construct IFI by omitting any of the 14 contract terms and reexamine the relationship between IFI and high-ranking agents. We find consistently positive coefficients on High Ranking Agent, all significant at least at the 5% confidence level. We also calculate a simulated IFI based on randomly assigned weights to the 14 contract terms. Across 10,000 simulated regressions, we find that the correlation between high-ranking agent and IFI is always positive (and significant in the majority of all specifications). This test rules out the possibility that the results are due to the equal-weighting scheme that we employ in our baseline tests.
An additional concern is that in our design of IFI, if warrant is attached, we add one. One could argue that the nature of warrant is much more complicated as it varies in terms of coverage ratio, term, and exercise price discount. In order to determine whether our simplification biases our results, in unreported regressions, we examine warrant coverage, which is measured as the ratio of proceeds from exercising warrants over the proceeds from the current offering, the term of the warrants, and the discount of the warrants, which is measured as the percentage difference between the warrant exercise price and the closing price. We find that agent reputation does affect warrant coverage, but this result is not significant. We note that more reputable agents are associated with significantly longer warrant terms and higher warrant discounts. These results support our inference based on IFI (Table IV) that more reputable agents are associated with more investor-friendly contractual protection.
Finally, we explore the possibility that expert agents focus on the more complex deals, while inexpert agents focus on simpler deals with fewer specific contract terms. If this were the case, the higher value of IFI might merely reflect the complexity of the PIPE contract and not its degree of investor-friendliness. To rule out this possibility, we
construct a contract complexity index that adds all contract terms regardless as to whether they are investor-friendly or issuer-friendly (i.e., we add issuer rights terms and trading restrictions to investor rights terms). We then relate this new index to the agent's expertise, but find no significant correlation.
B. Agent Rank, Investor-Friendly Contract Terms, and Pricing Discounts
1. Pricing-Term Trade-Offs
We now mm to the pricing effects of investor-friendly terms in PIPE contracts and the role of agents in this process. We analyze a "raw" pricing discount that is defined as the percentage difference between the closing price one day prior to the closing date and the offer price. By using this definition, we do not account for the pricing effects of warrants and other contingent cash flow rights. This approach is consistent with the contract theoretical argument that contingent cash flow rights are used as a contract solution to financing problems rather than simply as a price adjustment. Accordingly, our goal is not to study how expert agents relate to the total pricing of PIPEs (which would account for the pricing effects of all state contingent cash flow rights), but to investigate how expert agents relate to the trade-off between contract terms and the raw pricing dimension of PIPEs. In other words, we would like to know whether issuers with an expert agent are able to receive higher pricing in the form of raw discounts as compensation for agreeing to the lower pricing implied by the included state contingent cash flow rights. Our evidence suggests that this is the case.
Table V presents the results of the ordinary least square (OLS) regressions in which the PIPE pricing discount is the dependent variable. In all of the regression models, we include IFI and the same control variables as in Table IV. In models other than (1), (4), and (7), we also include High Ranking Agent. In Panel A, we use the Carter-Manaster (1990) ranking to identify high-ranking agents. For robustness, we also use historical PIPE market shares to identify high-ranking agents and present the results in Panel B. We begin with the full sample of both structured PIPEs and traditional PIPEs. Then, we repeat the analysis for traditional PIPEs and common stock PIPEs. The robust standard errors, which account for clustering at the company level, are reported in parentheses.
In Model (1), we find that the IFI is significantly and negatively associated with the pricing discount, which is evidence of a trade-off between more investor-friendly contract terms and a higher offer price. The coefficient of the IFI is about -0.01 suggesting that each added investor-friendly term is associated with an approximately 1% reduction in the pricing discount. For a PIPE transaction of median size ($10 million in our sample), this association corresponds to an increase of about $100,000 in net proceeds, a cash amount that is not trivial for a small struggling company that issues PIPEs. When we restrict the sample to traditional PIPEs (Model (4)) and common stock PIPEs (Model (7)), similarly, we find a negative coefficient. However, they are not statistically significant.
2. Agent Expertise and the Pricing-Term Trade-Off
In Models (2), (5), and (8) of Table V, we further include our proxies for agent expertise and find that this is negatively related to the pricing discount, but insignificant. When we repeat this analysis in Panel B using market share to define agent reputation, similarly, we confirm a negative, but insignificant coefficient on agent ranking for the sample of all PIPEs and traditional PIPEs. Interestingly, we find a significant and negative relationship between agent expertise and discounts for common stock PIPEs. In Models (3), (6), and (9), we include an interaction variable between High Ranking Agent and IFI. We find a significantly negative coefficient on this interaction variable no matter which sample we use. We find similar results in Panel B. The significance of the coefficient on High Ranking Agent (as shown in Model (14)) disappears when the interaction term is included. Thus, when compared with issuers advised by low-ranking agents, issuers advised by high-ranking agents receive greater compensation in the form of lower PIPE discounts when they agree to more investor-friendly terms. In untabulated tests, we split the sample into issuers with high- and low-ranking agents. We find that the trade-off between terms (i.e., IFI) and pricing (i.e., discount) is about three to seven times more pronounced for PIPEs involving high-ranking agents than it is for those involving low-ranking agents.
To summarize, our analysis of the pricing discounts demonstrates that issuers who agree to more investor-friendly terms are compensated by a lower pricing discount. This suggests that a contract with many investor-friendly terms may not be bad for the issuer as long as there is sufficient pricing compensation. We find that the pricing compensation is higher for issuers with high-ranked agents, which is consistent with our argument that expert agents help their clients understand, and thereby better negotiate, contracts with investor-friendly terms. Overall, this contract design may be beneficial for the issuer as investor-friendly terms have the potential to mitigate agency and information problems. Next, we investigate whether we can find evidence of these benefits by analyzing long-run stock returns after PIPE offerings.
C. PIPE Issuers' Long-Run Stock Performance
Although the extant literature has established that PIPE issuers experience significantly negative long-run performance subsequent to a PIPE'S issuance (Brophy et al., 2009; Chaplinsky and Haushalter, 2010), this empirical pattern is yet to be explained. We contribute to this debate by investigating how PIPE issuers' long-run performance relates to contract design and agent expertise. These tests are interesting in their own right, but also shed light on whether the inclusion of investor-friendly terms increases the overall surplus or merely transfers wealth between the contractual parties.
We examine returns for a horizon up to three years. We analyze a relatively long time horizon rather than event window returns as it is plausible to expect the gains from contract designs to materialize slowly. (15) The sample includes 2,163 PIPEs for which we have sufficient data to construct CARs for a period of up to 36 months. We begin with a univariate comparison of the long-run stock performance of PIPE issuers using high-ranking agents and those using low-ranking agents. We measure long-run stock performance by calculating equal-weighted market-adjusted CARs at 12, 24, and 36 months following PIPE issuance. (16) We also provide CARs based on the alphas from time-series regressions of PIPE issuers' monthly excess returns on the Fama-French Four-Factor model. These results are presented in Panel A of Table VI. We find that PIPEs associated with high-ranking agents exhibit significantly less negative CARs than their counterparts do.
In Panel B of Table VI, we exhibit a similar comparison of long-run stock performance on the part of PIPE issuers with high (above median) IFI scores and those with low (below median) IFI scores. We find that a high IFI is associated with less negative long-run performance than those with a low IFI. The differences are significant at the 1% confidence level across all time windows.
Our cross-sectional regressions of PIPE issuers' long-run stock performance on agent reputation and the IFI confirm the above findings from the univariate test. The regression results are presented in Table VII. The dependent variables are the equal-weighted market-adjusted CARs at 12,24, and 36 months (depending on the regression model) following PIPE issuance, respectively. The key independent variables of interest are High Ranking Agent and IFI. Agent reputation is determined by the Carter-Manaster (1990) rankings in Panel A and based on previous market share in Panel B. In each panel, we begin with the full sample of both structured PIPEs and traditional PIPEs. Then, we repeat the analysis for traditional PIPEs and common stock PIPEs. In addition, we control for PIPE type, industry fixed effects, various issuer characteristics, and PIPE transaction costs (discounts). We further include the investor-type fixed effect in all specifications, as previous studies indicate that investor identity is an important indicator of PIPE issuers' long-run stock performance. For instance, Dai (2007) finds that venture capitalist invested PIPEs perform better than their counterparts. Brophy et al. (2009) confirm that PIPEs associated with hedge funds exhibit particularly poor long-run stock performance.
We observe a significantly positive correlation between the IFI and PIPE issuers' long-run stock performance across all specifications. The inclusion of an additional investor-friendly term increases CAR (1, 12) by 4.3% to 5.3%, increases CAR (1, 24) by 5.8% to 9.5%, and increases CAR (1, 36) by 10.6% to 16.9%. Thus, the effect of the IFI is both statistically and economically significant. We further demonstrate that PIPE issuers associated with high-ranking agents perform significantly better than those with low-ranking agents when all PIPEs are included in the regressions. When we restrict the sample to traditional PIPEs, we find that High Ranking Agent is significant for all three event windows when we use the market share based ranking, but significant only for 36-month CARs following issuance when we use the Carter-Manaster (1990) rankings. When the sample is restricted to common stock PIPEs, we do not find agent ranking to significantly affect issuers' long-run stock performance.
This paper presents detailed evidence regarding the design of financial contracts used in PIPE offerings. We document how PIPE contracts may include many possible combinations of terms that allocate contingent cash flow rights. These terms are in place to address agency and information problems, which are particularly severe financing frictions for companies that typically issue PIPEs. We relate the inclusion of these terms, and their associated effects on pricing and returns, to the expertise of the issuer's placement agent.
Our findings can be summarized as follows. First, issuers advised by high-ranking agents provide investors with more contractual protection than those advised by low-ranking agents. In addition, high-ranking agents allow issuers to extract more compensation in exchange for investor-friendly contract terms. Moreover, both agent ranking and the IFI are associated with stronger (i.e., less negative) long-run stock performance following PIPE offerings.
Our findings are consistent with an explanation that builds on the idea that it is sometimes difficult to understand the payoff consequences of complex contract terms. We argue that the presence of high-ranking agents in PIPE offerings allows the contracting parties to agree on the consequences of investor-friendly terms. PIPE investors are familiar with complicated contract design and can correctly understand the consequences of their typically esoteric terms. PIPE issuers, who are often small, distressed companies, suffer from bounded rationality with respect to their ability to decipher and evaluate contract terms. In such a contracting environment, placement agents play an important role by bridging the contract knowledge gap between these two parties' knowledge about contracts. This agent role allows contracts to include more contingent cash flow contingencies, features that are viewed as optimal in many contract theoretical models. Our findings indicate that an agent's ability or incentive to fill this function varies with their experience and reputation. High-ranked agents provide higher quality services as they can more credibly and accurately convey valuable information to their clients.
Appendix A: List of Expert Placement Agents
This table lists the 10 reputable PIPE placement agents based on market share.
Agent Name Nr of Sum of Gross C&M PIPES Proceeds ($M) Ranking Goldman, Sachs & Co. 17 23,092 9 JP Morgan Chase & Co. 37 8,175 9 Credit Suisse Securities 27 5,007 9 (USA) LLC UBS Investment Bank 34 3,801 8 Citigroup Global 17 3,773 9 Markets, Inc. Rodman & Renshaw, LLC 183 2,912 2 Lehman Brothers, Inc. 48 2,670 8 Morgan Stanley 10 2,314 9 Banc of America Securities LLC 50 1,613 9 Bear, Steams & Co. Inc. 12 809 9
Appendix B: Definitions of Key Variables
Variables Definitions Investor- Index that captures the aggregate Friendly Index inclusion of investor-friendly (i.e., IFI) contract terms. IFI is calculated by adding each investor protection term, deducting each investor trading restriction, deducting each issuer right, and adding seven (to make nonnegative). High Ranking A dummy variable that is equal to Agent (C&M one if the placement agent belongs Ranking) to the group of the most reputable agents, based on its Carter and Manaster (1990) rankings, and zero otherwise. The cutoff for high-ranking agents is a C&M ranking of seven. Our findings are robust to other cutoffs. Discount For common stock PIPEs, the percentage difference between closing price one day before the closing date and the offer price. For PIPEs with fixed price convertibles, the percentage difference between the closing price one day before the closing date and the conversion price. For PIPEs with floating rate convertibles (i.e., structured PIPEs), the difference between the closing price one day before the closing date and the specified floor price. High-Ranking First, we calculate agent market Agent (Market share by comparing an agent's PIPE Share) volume (in dollars) in the three preceding years to the total volume of intermediated PIPE offerings during the same period. Then, we count the number of times the agent was on the top 15 lists over our sample period. The 10 agents that appear on the annual top 15 lists most often are designated as reputable agents. Offer Size The amount of gross proceeds (i.e., capital raised in the PIPE offering). Traditional A dummy variable that is equal to PIPE one if the PIPE offers common stocks and fixed price convertibles to investors, and zero otherwise. Structured A dummy variable that is equal to PIPEs one if the PIPE includes re-pricing rights, such as floating price convertibles or convertible resets, and zero otherwise. Investor The within sample number of PIPE Experience transactions that an investor has previously participated in. Nr of PIPEs The within sample number of PIPE Issued Before transactions that an issuer has previously completed. Market Value The market capitalization in ($M) millions of the PIPE issuer one day prior to the closing date. Analyst The maximum number of analysts Coverage following the PIPE issuer over the 12 months prior to the PIPE. CAR (-12,-1) The equal-weighted market adjusted cumulative abnormal returns (CARs) 12 months prior to the PIPE. CAR (1,12) The equal-weighted market adjusted CARs from one month to 12 months after the PIPE. CAR (1,24) The equal-weighted market adjusted CARs from one month to 24 months after the PIPE. CAR (1,36) The equal-weighted market adjusted CARs from one month to 36 months after the PIPE. Debt/Assets The ratio of long-term debt to total assets. Both numbers are from the financial statement of the nearest fiscal year prior to the PIPE. R&D/Assets The ratio of the R&D expenses to total assets. Both numbers are from the financial statement of the nearest fiscal year prior to the PIPE. EV/Assets The ratio of enterprise value, which is the sum of market capitalization and debt minus cash to total assets. Both numbers are from the financial statement of the nearest fiscal year prior to the PIPE. Intangible/Assets The ratio of intangible assets to total assets. Both numbers are from the financial statement of the nearest fiscal year prior to the PIPE. Cash Burn This measure is computed as cash flow from operations/cash and cash equivalents. For issuers with positive cash flow, this value is set equal to zero. More negative values indicate less time before running out of cash. BURN dummy This dummy is set equal to one if cash burn rate is above the sample median and to zero if it is below the median. Z-Score This measure is computed according to Altman's original formula: Z = 1.2T1 + 1.4T2 + 3.3T3 + 0.6T4 + .999T5, where T1 = Working Capital/Total Assets T2 = Retained Eamings/Total Assets T3 = Earnings Before Interest and Taxes/Total Assets T4 = Market Value of Equity /Total Liabilities T5 = Sales/Total Assets Distress dummy This dummy is set equal to one if the z-score is below 1.8 and to zero otherwise. Crisis dummy This dummy is set to equal to one if the PIPE deal was closed in 2008 and 2009, the financial crisis period, and zero otherwise.
Beatty, R.P. and I. Welch, 1996, "Issue Expenses and Legal Liability in Initial Public Offerings," Journal of Law and Economics 39, 545-602.
Bengtsson, O., 2011, "Covenants in Venture Capital Contracts," Management Science 57, 1926-1943.
Bengtsson, O. and B. Sensoy, 2011, "Investor Abilities and Financial Contracting: Evidence from Venture Capital," Journal of Financial Intermediation 20, 477-502.
Brophy, D.J., P.P. Ouimet, and C. Sialm, 2009, "Hedge Funds as Investors of Last Resort," Review of Financial Studies 22, 541-574.
Byoun, S. and W.T. Moore, 2003, "Stock vs. Stock-Warrant Units: Evidence from Seasoned Offerings," Journal of Corporate Finance 9, 575-590.
Carter, R.B., F.H. Dark, and A.K. Singh, 1998, "Underwriter Reputation, Initial Returns, and Long Run Performance of IPO Stocks," Journal of Finance 53,285-311.
Carter, R.B. and S. Manaster, 1990, "Initial Public Offerings and Underwriter Reputation," Journal of Finance 45, 1045-1068.
Chaplinsky, S. and D. Haushalter, 2010, "Financing Under Extreme Uncertainty: Contract Terms and Returns to Private Investments in Public Equity," Review of Financial Studies 23, 2789-2820.
Chemmanur, T.J. and P. Fulghieri, 1997, "Why Include Warrants in New Equity Issues? A Theory of Unit IPOs," Journal of Financial and Quantitative Analysis 32, 1-24.
Chen, H.C., N. Dai, and J.D. Schatzberg, 2010, "The Choice of Equity Selling Mechanisms: PIPEs versus SEOsJournal of Corporate Finance 16, 104-119.
Cooney, J.W., A.K. Singh, R.B. Carter, and F.H. Dark, 2001, "IPO Initial Returns and Underwriter Reputation: Has the Inverse Relationship Flipped in the 1990s?" University of Kentucky, Case Western Reserve, and Iowa State University, Working paper.
Dai, N., 2007, "Does Investor Identity Matter? An Empirical Examination of Investments by Venture Capital Funds and Hedge Funds in PIPEs," Journal of Corporate Finance 13, 538-563.
Dai, N., H. Jo, and J.D. Schatzberg, 2010, "The Quality and Price of Investment Banks' Service: Evidence from the PIPE Market," Financial Management 39, 585-612.
Datta, S., M. Iskandar-Datta, and A. Patel, 1997, "The Pricing of Initial Public Offers of Corporate Straight Debt," Journal of Finance 52, 379-396.
Fang, L.H., 2005, "Investment Bank Reputation and the Price and Quality of Underwriting Services," Journal of Finance 60, 2729-2761.
Fernando, C.S., V.A. Gatchev, and P.A. Spindt, 2005, "Wanna Dance? How Firms and Underwriters Choose Each Other," Journal of Finance 60, 2437-2469.
Floras, I. and T. Sapp, 2012, "Why Do So Many Firms Issue Private Equity Repeatedly? On the Information Content of Repeated PIPE Offerings," Journal of Banking and Finance 36, 3469-3481.
Gamer, J.L. and B.B. Marshall, 2005, "Unit IPOs: What the Warrant Characteristics Reveal About the Issuing Firm," Journal of Business 78, 1837-1858.
Gompers, P.A., 1998, "An Examination of Convertible Securities in Venture Capital Investments," Harvard University, Working paper. Gompers, P. A., J. Ishii, and A. Metrick, 2003, "Corporate Governance and Equity Prices," Quarterly Journal of Economics 118, 107-155.
Hillion, P. and T. Vermaelen, 2004, "Death Spiral Convertibles," Journal of Financial Economics 71, 381-415.
How, J.C.Y. and J.S. Howe. 2001, "Warrants in Initial Public Offerings: Empirical Evidence," Journal of Business 74, 433-457.
Jain, B.A., 1994, "The Underpricing of "Unit" Initial Public Offerings," Quarterly Review of Economics and Finance 34, 309-327.
Kaplan, S. and P. Stromberg, 2003, "Financial Contracting Meets the Real World: An Empirical Analysis of Venture Capital Contracts," Review of Economic Studies 70, 281-316.
Kaplan, S. and P. Stromberg, 2004, "Characteristics, Contracts, and Actions: Evidence from Venture Capitalist Analyses," Journal of Finance 59, 2177-2210.
Livingston, M. and R.E. Miller, 2000, "Investment Bank Reputation and the Underwriting of Nonconvertible Debt,". Financial Management 29, 21-34.
Ljungqvist, A.P. and W.J. Wilhelm, Jr., 2002, "IPO Allocations: Discriminatory or Discretionary?" Journal of Financial Economics 65, 167-201.
Logue, D.E., R.J. Rogalski, J.K. Seward, and L. Foster-Johnson, 2002, "What Is Special About the Roles of Underwriter Reputation and Market Activities in Initial Public Offerings?" Journal of Business 75, 213-243.
Loughran, T. and J.R. Ritter, 2004, "Why Has IPO Underpricing Changed over Time?" Financial Management 33, 6-37.
McDonald, J.G. and A.K. Fisher, 1972, "New-Issue Stock Price Behavior," Journal of Finance 27, 97-102.
Megginson, W.L. and K.A. Weiss, 1991, "Venture Capitalist Certification in Initial Public Offerings," Journal of Finance 46, 879-904.
Schultz, P., 1993, "Unit Initial Public Offerings," Journal of Financial Economics 34, 199-229.
Simon, C.J., 1990, "The Role of Reputation in the Market for Initial Public Offerings," University of Chicago, Working paper.
Tirole, J., 2009, "Cognition and Incomplete Contracts," American Economic Review 99, 265-294.
Titman, S. and B. Trueman, 1986, "Information Quality and the Valuation of New Issues," Journal of Accounting and Economics 8, 159-172.
(1) PIPE statistics are reported by Sagient Research (www.sagientresearch.com/pt).
(2) Our data support this viewpoint. The use of warrants is positively correlated with the use of other contract terms. If warrants were used as a pricing adjustment, then one would expect a negative correlation.
(3) For convenience, we discuss PIPE terms as being only investor-friendly (which are the most prevalent in practice). This practice also captures issuer-friendly terms as including an issuer-friendly term is functionally the same as excluding an investor-friendly term that is inversely defined. In our coding scheme, we deduct issuer-friendly terms from the IFI.
(4) This section is, to a large extent, based on information from the book "The Issuer's Guide to PIPEs: New Markets, Deal Structures, and Global Opportunities for Private Investments in Public Equity," as well as from memos prepared by law firms. The information about the structure of private placement agreements comes from a sample of Securities and Exchange Commission (SEC) filings that we have analyzed. Finally, we have interviewed Brett Goetschius, who is the chief executive officer (CEO) of MarketNexus Media and an expert on PIPEs.
(5) As pointed out in Chaplinsky and Haushalter (2010), relative to the private placements available in the Security Data Corporation's New Issues database, the Placementtracker database lists more PIPE offerings and provides more detailed coverage of contract terms.
(6) Unlike the all-in-net discounts calculated in Chaplinsky and Haushalter (2010), in this paper, we treat the option feature of PIPE securities (such as warrants) as contractual terms and study how they are correlated with the discounts.
(7) The purchase price of a floating convertible PIPE is conditional upon the trading prices of the PIPE issuer's stocks during a specified period, typically 10 to 30 days prior to conversion. The floor price is the lowest purchase price or conversion price if the stock performance of the issuer deteriorates badly. Thus, a discount based on the floor price for such a PIPE represents the maximum discount the investor can receive.
(8) Extant work on warrants describes how this deal feature can reduce informational frictions in IPOs and SEOs (Schultz, 1993; Jain, 1994; Chemmanur and Fulghieri, 1997; How and Howe, 2001; Byoun and Moore, 2003; Gamer and Marshall, 2005).
(9) PIPE contracts use the lull ratchet antidilution provision, which gives existing investors the right to receive additional shares at the price of a new financing round at a lower valuation. This is in contrast to venture capital contracts in which full ratchet is a relatively uncommon version of the antidilution provisions.
(10) For instance, the number of terms varies from 0-11, with a standard deviation of 2.2. Furthermore, IFI ranges from 5-14, with a standard deviation of 1.7. We also analyze the variation in the number of terms and IFI by agent and find no evidence that agents simply utilize boiler plate contracts in different PIPE deals.
(11) The correlation coefficient between investor protection terms and trading restriction term is 0.144 (with p-value of 0.000). The correlation coefficient between investor protection and issuer right is 0.063 (with a /)-value of 0.000).
(12) Our Carter-Manaster (1990) ranking is obtained from Jay Ritter's Website. He provides the Carter-Manaster ranking for IPO underwriters in several subperiods. We search for the PIPE placement agent's name in the Carter-Manaster (1990) ranking list during the period when a specific PIPE was issued.
(13) Market share has been used frequently in the literature as an empirical proxy for reputation. See, for example, McDonald and Fisher (1972), Simon (1990), Megginson and Weiss (1991), Beatty and Welch (1996), Fang (2005), and Dai et al. (2010).
(14) In an unreported probit regression on matching, we find results consistent with the univariate analysis.
(15) We also test for short-run announcement returns, but find no significant results. This suggests that the market is not aware of the gains associated with an expert agent and PIPE contracts with higher IFI.
(16) We calculate value-weighted market-adjusted CARs as well and find similar results. As the majority of PIPE issuers are small firms, we view equal-weighted returns as a more appropriate metric.
Ola Bengtsson and Na Dai *
We appreciate the valuable comments from Raghu Rau (Editor) and an anonymous referee. The paper also benefits from comments from Heitor Almeida, Alex Borisov, Murillo Campello, seminar participants at the University of Illinois at Urbana-Champaign, SUNY at Albany, the 2011 Mid-Western Finance Association Annual Meeting, and the 2011 FMA Annual Meeting. Na Dai owes thanks to the Center for Institutional Investment Management (CIIM) at University at Albany for financial support. All remaining errors are our own.
* Ola Bengtsson is a Professor of Finance in the Department of Economics at the Lund University of Economics and Management in Lund, Sweden. Na Dai is an Associate Professor of Finance in the School of Business at the State University of New York at Albany in Albany, NY.
Table I. Summary Statistics The sample is 2,414 intermediated PIPEs (i.e., offerings where the issuer employs a placement agent) that closed from 1999 to 2012. The data are retrieved from Sagient Research, Inc.'s Placementtracker. See Appendix B for variable definitions. In Panel D, the variables reflect the financial statement of the nearest fiscal year prior to the offering. Panel A. Year of Offering Year of No. of Per- Year of Offering Offerings centage Offering 1999 120 (5%) 2006 2000 210 (9%) 2007 2001 190 (8%) 2008 2002 159 (7%) 2009 2003 253 (10%) 2010 2004 219 (9%) 2011 2005 170 (7%) 2012 Panel B. Industry Group of Issuer Pharma- Biotech- Healthcare Internet ceuticals nology Products 359 323 207 167 (15%) (13%) (9%) (7%) Panel C. Characteristics of Offering Mean Median 25% Offer Size ($M) 51.3 11.1 5.0 Discount 9.5% 8.6% -0.3% Investor-Friendly Index (i.e., IFI) 8.6 9.0 7.0 Investor Experience 6.6 1.0 0.0 Traditional PIPE 90.3% Common Stock PIPE 62.3% Structured PIPE 9.7% Hedge Funds Investor 53.0% High-Ranking Agent (C&M 19.7% Ranking) High Ranking Agent (Market 14.8% Share) Panel D. Characteristics of Issuers Mean Median 25% Market Value ($M) 271.0 90.9 42.8 Analyst Coverage 2.2 1.0 0.0 CAR (-12,-1) 19.7% 2.4% -58.5% Debt/Assets 14.6% 3.5% 0.0% EV/Assets 4.1 2.0 0.9 R&D/Assets 29.2% 12.0% 0.0% Intangible/Assets 11.6% 1.7% 0.0% Cash Burn -6.6 -0.6 -1.7 Z-Score -0.5 0.8 -4.7 Panel A. Year of Offering Year of No. of Per- Offering Offerings centage 1999 162 (7%) 2000 155 (6%) 2001 126 (5%) 2002 196 (8%) 2003 203 (8%) 2004 123 (5%) 2005 128 (5%) Panel B. Industry Group of Issuer Pharma- Telecommunications Software Others ceuticals 359 163 139 1056 (15%) (7%) (6%) (44%) Panel C. Characteristics of Offering 75% Std. Dev. Offer Size ($M) 26.1 320.0 Discount 17.5% 16.6% Investor-Friendly Index (i.e., IFI) 10.0 1.7 Investor Experience 7.0 15.7 Traditional PIPE Common Stock PIPE Structured PIPE Hedge Funds Investor High-Ranking Agent (C&M Ranking) High Ranking Agent (Market Share) Panel D. Characteristics of Issuers 75% Std. Dev. Market Value ($M) 198.0 905.5 Analyst Coverage 3.0 3.6 CAR (-12,-1) 82.9% 141.4% Debt/Assets 19.5% 26.3% EV/Assets 4.4 8.6 R&D/Assets 38.0% 54.1% Intangible/Assets 17.6% 17.8% Cash Burn 0.0 178.5 Z-Score 4.2 28.7 Table II. Overview of PIPE Contract Terms The sample consists of 2,414 intermediated PIPEs. See Table I for the sample overview. Panel A lists contract terms that provide the investor with protection, Panel B lists terms that place limits on investor trading, and Panel C lists terms that grant the issuer rights. Panel A. Investor Protections (Favorable to Investor, Added to Investor-Friendly Index, IFI) ID Term Definition Frequency 1 Interest rate or dividend Issuers pay periodic 26.4% interest or dividends to investors at the specified rate. They can be paid with cash, shares, or the same security as issued to investors. 2 Warrants Investors are granted a 54.5% certain number of warrants with specified exercise price and expiration dates. 3 Registration right Investors request that a 64.7% company file a registration statement covering the resale of common stocks (underlying the issued securities) no later than a certain number of days after the closing and make it effective within a certain time window. 4 Antidilution The antidilution 42.6% provision protects investors against future financing at a lower valuation than the valuation of the current (protected) offering. In extreme cases, a company is not allowed to issue or sell any equity securities or securities convertible into equity during a certain period after closing. 5 First refusal right This provision provides 20.1% investors with the right to purchase additional shares of a company's security under specified terms during a certain period before issuers sell shares to third parties. 6 Investor call option Investors have the right 5.3% to purchase additional shares with specified terms prior to the expiration date of an option. 7 Redemption right The redemption right 12.2% provides investors with the right to demand that firms redeem investors' claims upon the occurrence of certain events, such as a change of control, typically at face value or a certain percentage of face value plus accrued and unpaid interest. Panel B. Trading Restrictions (Unfavorable to Investor, Deducted from IFI) 8 No shorting/hedging This provision asks 8.0% investors not to engage in any short transactions or hedging of a company's common stock prior to the effectiveness of the Registration Statement. 9 Offsetting long position This provision asks 2.7% investors not to engage in any short transactions or hedging of a company's common stock in excess of the amount of shares owned (an offsetting long position) prior to the effectiveness of the Registration Statement. 10 Public offering If a company is planning 0.1% a public offering shortly following a PIPE issuance, the company will ask investors not to affect any sales to the public of shares of the company for a certain period of days following the effectiveness of the Registration Statement to avoid price pressure from investors' resale of shares to the public. 11 Lock up With this provision, 2.9% investors may not sell any shares of a company's common stock purchased or received through the exercise of warrants for the duration of a few months following the closing. Panel C. Issuer Rights (Unfavorable to Investor, Deducted from IFI) 12 Company-forced conversion Securities held by PIPE 13.5% investors will automatically convert or be forced to convert into common stock under certain conditions. These conditions often relate to company stock performance. For instance, when the stock price or the weighted average stock price during a period exceeds a certain benchmark or the daily trading volume exceeds a certain level for some consecutive trading days. 13 Company put option A company put option 2.9% gives a company the right to request that PIPE investors purchase additional securities at a specified price in the future. 14 Company optional This provision gives a 11.6% redemption company the right to force PIPE investors to exercise redemption rights after a certain date or upon the occurrence of certain events. Table III. Matching between PIPE Issuers and Placement Agents The sample consists of 2,414 intermediated PIPEs. See Table I for the sample overview. Panel A lists the sample means for key company characteristics for issuers with high-ranking agents and low-ranking agents, respectively. Panel B tabulates the number of PIPEs issued previously for high-ranking agents and low-ranking agents, respectively. Agent reputation is based on the C&M ranking. Panel A. Company Characteristics (Mean) for Issuers with High- and Low-Ranking Agents High-Ranking Low-Ranking Difference Agents Agents Market Value ($M) 466.2 223.2 243.0 *** Analyst Coverage 4.0 1.7 2.3 *** CAR (-12,-1) 1.6% 24.1% -22.5% *** Debt/Assets 16.8% 14.1% 2.7% ** EV/Assets 2.3 4.6 -2.3 *** R&D/Assets 22.1% 31.0% -8.9% *** Intangible/Assets 8.7% 12.3% -3.6% *** Cash Burn -1.3 -7.9 6.6 Z-score -1.8 -1.2 -0.7 Number of Obs. 475 1,939 Panel B. Number of PIPEs Issued Before By Issuer with High- and Low-Ranking Agents No. of PIPEs Issued High-Ranking Low-Ranking Before Agents Agents 0 232 (49%) 686 (35%) 1 84 (18%) 346 (18%) 2 46 (10%) 241 (12%) 3 39 (8%) 181 (9%) 4 19 (4%) 111 (6%) 5 15 (3%) 91 (5%) >5 40 (8%) 283 (15%) *** Significant at the 0.01 level. ** Significant at the 0.05 level. Table IV. Relation between the Investor-Friendly Index (IFI) and High-Ranking Agents The sample consists of 2,414 intermediated PIPEs. See Table I for the sample overview. These are Poisson regressions. The dependent variable is the IFI, which is the sum of the investor protection provisions (+), the trading restriction provisions (-), and the issuer right provisions (-) plus seven (to make the dependent variable nonnegative). Agent ranking is based on the Carter and Manaster (1990) ranking in Panel A, and based on market share in Panel B. The definitions of all of the other variables are provided in Appendix B. Standard errors are clustered by company and reported in the parentheses. All Traditional Common PIPES PIPEs Stock PIPEs Specification (1) (2) (3) Panel A. Carter and Manaster Rankings (1990) High-Ranking Agent 0.0273 *** 0.0237 ** 0.0172 (0.0099) (0.0099) (0.0111) Crisis Dummy Ln(Investor Experience) 0.0098 *** 0.0077 *** 0.0060 ** (0.0026) (0.0025) (0.0028) Ln(MV) -0.0199 *** -0.0189 *** -0.0209 *** (0.0040) (0.0040) (0.0049) Ln(Analyst) -0.0115 ** -0.0058 -0.0064 (0.0055) (0.0055) (0.0062) CAR (-12,-1) -0.0034 -0.0011 -0.0055 (0.0029) (0.0033) (0.0039) Debt/Assets 0.0017 -0.0065 -0.0169 (0.0155) (0.0127) (0.0128) EV/Assets 0.0011 *** 0.0012 ** 0.0012 ** (0.0003) (0.0005) (0.0006) RD/Assets -0.0003 0.0007 -0.0046 (0.0066) (0.0070) (0.0101) Intangible/Assets 0.0381 0.0291 -0.0065 (0.0244) (0.0258) (0.0305) Burn Dummy 0.0001 0.0037 -0.0011 (0.0077) (0.0079) (0.0091) Distress Dummy 0.0240 *** 0.0221 ** 0.0257 ** (0.0087) (0.0090) (0.0107) Traditional PIPE -0.0936 *** (0.0141) Common Stock PIPE -0.0805 *** (0.0090) Intercept 2.1717 *** 2.1634 *** 2.1092 *** (0.0305) (0.0292) (0.0350) Investor-Type Fixed Effect Yes Yes Yes Industry Fixed Effect Yes Yes Yes Year Fixed Effect Yes Yes Yes Observations 2,414 2,180 1,505 Pseudo [R.sup.2] (%) 6.73 6.90 7.22 All Traditional Common PIPES PIPEs Stock PIPEs Specification (4) (5) (6) Panel A. Carter and Manaster Rankings (1990) High-Ranking Agent 0.0392 *** 0.0375 *** 0.0259 ** (0.0113) (0.0112) (0.0128) Crisis Dummy 0.0758 *** 0.0736 *** 0.0607 *** (0.0141) (0.0140) (0.0162) Ln(Investor Experience) 0.0233 *** 0.0230 *** 0.0246 *** (0.0032) (0.0031) (0.0037) Ln(MV) -0.0269 *** -0.0240 *** -0.0239 ** (0.0047) (0.0047) (0.0061) Ln(Analyst) -0.0132 ** -0.0092 -0.0128 * (0.0062) (0.0063) (0.0073) CAR (-12,-1) -0.0075 ** -0.0019 -0.0062 (0.0035) (0.0040) (0.0046) Debt/Assets -0.0027 -0.0059 -0.0260 (0.0164) (0.0144) (0.0173) EV/Assets 0.0015 * 0.0005 0.0008 (0.0008) (0.0008) (0.0008) RD/Assets -0.0008 0.0041 0.0000 (0.0100) (0.0111) (0.0140) Intangible/Assets 0.1005 *** 0.0895 *** 0.0592 * (0.0290) (0.0304) (0.0353) Burn Dummy -0.0068 -0.0023 -0.0114 (0.0093) (0.0096) (0.0108) Distress Dummy 0.0346 *** 0.0327 *** 0.0390 *** (0.0106) (0.0107) (0.0122) Traditional PIPE -0.0349 ** (0.0172) Common Stock PIPE -0.0639 *** (0.0107) Intercept 2.1340 *** 2.1356 *** 2.0778 *** (0.0361) (0.0339) (0.0422) Investor-Type Fixed Effect Yes Yes Yes Industry Fixed Effect Yes Yes Yes Year Fixed Effect No No No Observations 2,414 2,180 1,505 Pseudo [R.sup.2] (%) All Traditional Common PIPES PIPEs Stock PIPEs Specification (7) (8) (9) Panel B. Carter and Manaster Rankings (1990) High-Ranking Agent 0.0463 *** 0.0440 *** 0.0394 *** (0.0099) (0.0099) (0.0105) Crisis Dummy Ln(Investor Experience) 0.0088 *** 0.0067 *** 0.0055 ** (0.0026) (0.0025) (0.0028) Ln(MV) -0.0220 *** -0.0211 *** -0.0223 *** (0.0040) (0.0040) (0.0049) Ln(Analyst) -0.0093 * -0.0039 -0.0056 (0.0055) (0.0055) (0.0062) CAR (-12,-1) -0.0035 -0.0010 -0.0055 (0.0030) (0.0034) (0.0039) Debt/Assets -0.0071 -0.0150 -0.0194 (0.0138) (0.0119) (0.0133) EV/Assets 0.0012 *** 0.0012 ** 0.0012 ** (0.0004) (0.0005) (0.0006) RD/Assets -0.0047 -0.0033 -0.0083 (0.0059) (0.0063) (0.0088) Intangible/Assets 0.0328 0.0242 -0.0103 (0.0242) (0.0257) (0.0303) Burn Dummy 0.0029 0.0067 0.0002 (0.0076) (0.0079) (0.0090) Distress Dummy 0.0256 *** 0.0233 *** 0.0267 ** (0.0087) (0.0090) (0.0106) Traditional PIPE -0.0942 *** (0.0142) Common Stock PIPE -0.0789 *** (0.0090) Intercept 2.1910 *** 2.1810 *** 2.1206 *** (0.0301) (0.0291) (0.0346) Investor-Type Fixed Effect Yes Yes Yes Industry Fixed Effect Yes Yes Yes Year Fixed Effect Yes Yes Yes Observations 2,414 2,180 1,505 Pseudo [R.sup.2] (%) 6.76 6.93 7.25 All Traditional Common PIPES PIPEs Stock PIPEs Specification (10) (11) (12) Panel B. Carter and Manaster Rankings (1990) High-Ranking Agent 0.0512 *** 0.0447 *** 0.0353 ** (0.0126) (0.0130) (0.0149) Crisis Dummy 0.0740 *** 0.0692 *** 0.0697 *** (0.0141) (0.0146) (0.0158) Ln(Investor Experience) 0.0231 *** 0.0229 *** 0.0247 *** (0.0032) (0.0031) (0.0037) Ln(MV) -0.0255 *** -0.0224 *** -0.0228 *** (0.0048) (0.0049) (0.0061) Ln(Analyst) -0.0130 ** -0.0098 -0.0136 * (0.0062) (0.0062) (0.0073) CAR (-12,-1) -0.0080 ** -0.0020 -0.0061 (0.0035) (0.0040) (0.0046) Debt/Assets -0.0095 -0.0112 -0.0279 (0.0154) (0.0143) (0.0178) EV/Assets 0.0013 0.0001 0.0007 (0.0008) (0.0008) (0.0008) RD/Assets -0.0057 -0.0001 -0.0042 (0.0087) (0.0100) (0.0126) Intangible/Assets 0.0944 *** 0.0839 *** 0.0576 (0.0292) (0.0306) (0.0353) Burn Dummy -0.0013 0.0033 -0.0073 (0.0092) (0.0095) (0.0106) Distress Dummy 0.0386 *** 0.0363 *** 0.0430 *** (0.0106) (0.0107) (0.0121) Traditional PIPE -0.0351 ** (0.0169) Common Stock PIPE -0.0627 *** (0.0108) Intercept 2.1414 *** 2.1414 *** 2.0777 *** (0.0360) (0.0339) (0.0421) Investor-Type Fixed Effect Yes Yes Yes Industry Fixed Effect Yes Yes Yes Year Fixed Effect No No No Observations 2,414 2,180 1,505 Pseudo [R.sup.2] (%) 1.49 1.59 1.68 *** Significant at the 0.01 level. ** Significant at the 0.05 level. * Significant at the 0.10 level. Table V. Relation between Discounts (PIPE Pricing), Investor-Friendly Index (IFI), and High-Ranking Agents The sample consists of 2,414 intermediated PIPEs. See Table I for the sample overview. These are OLS regressions. The dependent variable is the discount, which for common stock PIPEs is the percentage difference between the closing price one day prior to the closing date and the offer price, for PIPEs with fixed price convertibles it is the percentage difference between the closing price one day prior to the closing date and the conversion price, and for PIPEs with floating rate convertibles (i.e., structured PIPEs), it is the difference between the closing price one day prior to the closing date and the specified floor price. The IFI is the sum of the investor protection provisions (+), trading restriction provisions (-), and issuer right provisions (-) plus seven (to make the dependent variable nonnegative). Agent ranking is based on the Carter and Manaster (1990) rankings in Panel A, and based on market share in Panel B. The definitions of all of the other variables are provided in Appendix B. All specifications also include an intercept. Standard errors are clustered by company and reported in the parentheses. Common All Traditional Stock PIPES PIPEs PIPEs Specifications (1) (2) (3) Panel A. Carter and Manaster (1990) Rankings IFI -0.0101 *** -0.0100 *** -0.0076 ** (0.0034) (0.0034) (0.0034) High-Ranking Agent -0.0156 0.1192 * (0.0113) (0.0634) IFI x High-Ranking -0.0157 * Agent (0.0072) Ln(Investor -0.0077 -0.0082 -0.0081 Experience) (0.0053) (0.0053) (0.0052) Ln(MV) -0.0066 -0.0058 -0.0058 (0.0044) (0.0043) (0.0043) Ln(Analyst) -0.0138 * -0.0132 * -0.0139 * (0.0054) (0.0055) (0.0054) CAR (-12,-1) 0.0072 * 0.0071 ** 0.0070 ** (0.0031) (0.0031) (0.0031) Debt/Assets -0.0135 -0.0141 -0.0166 (0.0136) (0.0137) (0.0140) EV/Assets 0.0009 * 0.0009 * 0.0008 ** (0.0004) (0.0004) (0.0004) RD/Assets 0.0152 0.0147 0.0140 (0.0118) (0.0120) (0.0120) Intangible/ -0.0214 -0.0241 -0.0253 Assets (0.0222) (0.0223) (0.0222) Burn Dummy 0.0261 *** 0.0270 *** 0.0265 *** (0.0080) (0.0080) (0.0080) Distress Dummy -0.0123 -0.0121 -0.0110 (0.0096) (0.0096) (0.0095) Traditional PIPE -0.0901 *** -0.0905 *** -0.0879 *** (0.0214) (0.0214) (0.0213) Common Stock PIPE Intercept 0.3504 *** 0.3553 *** 0.3325 *** (0.0702) (0.0700) (0.0688) Investor-Type Yes Yes Yes Fixed Effect Industry Yes Yes Yes Fixed Effect Year Fixed Effect Yes Yes Yes Observations 2,414 2,414 2,414 Adjusted [R.sup.2] (%) 10.18 10.25 10.51 All Traditional Stock PIPES PIPEs PIPEs Specifications (4) (5) (6) Panel A. Carter and Manaster (1990) Rankings IFI -0.0050 -0.0049 -0.0029 (0.0033) (0.0033) (0.0033) High-Ranking Agent -0.0090 0.1075 * (0.0108) (0.0523) IFI x High-Ranking -0.0136 * Agent (0.0060) Ln(Investor -0.0087 -0.0089 -0.0089 Experience) (0.0055) (0.0055) (0.0054) Ln(MV) -0.0050 -0.0046 -0.0045 (0.0042) (0.0042) (0.0041) Ln(Analyst) -0.0186 ** -0.0183 ** -0.0188 ** (0.0054) (0.0055) (0.0055) CAR (-12,-1) 0.0048 * 0.0048 0.0046 (0.0029) (0.0029) (0.0029) Debt/Assets -0.0119 -0.0122 -0.0148 (0.0117) (0.0118) (0.0118) EV/Assets 0.0009 0.0008 0.0008 (0.0005) (0.0005) (0.0005) RD/Assets 0.0149 0.0146 0.0140 (0.0121) (0.0122) (0.0123) Intangible/ 0.0003 -0.0014 -0.0030 Assets (0.0227) (0.0228) (0.0227) Burn Dummy 0.0325 *** 0.0330 *** 0.0327 *** (0.0077) (0.0077) (0.0077) Distress Dummy -0.0128 -0.0127 -0.0117 (0.0099) (0.0099) (0.0099) Traditional PIPE Common Stock PIPE 0.0610 *** 0.0609 *** 0.0599 *** (0.0096) (0.0096) (0.0096) Intercept 0.1630 ** 0.1659 *** 0.1497 ** (0.0634) (0.0632) (0.0617) Investor-Type Yes Yes Yes Fixed Effect Industry Yes Yes Yes Fixed Effect Year Fixed Effect Yes Yes Yes Observations 2,180 2,180 2,180 Adjusted [R.sup.2] (%) 12.56 12.57 12.78 Common All Traditional Stock PIPES PIPEs PIPEs Specifications (7) (8) (9) Panel A. Carter and Manaster (1990) Rankings IFI -0.0018 -0.0018 0.0001 (0.0041) (0.0041) (0.0041) High-Ranking Agent 0.0045 0.1217 * (0.0113) (0.0567) IFI x High-Ranking -0.0141 * Agent (0.0069) Ln(Investor -0.0057 -0.0055 -0.0056 Experience) (0.0044) (0.0045) (0.0045) Ln(MV) -0.0017 -0.0020 -0.0019 (0.0043) (0.0043) (0.0042) Ln(Analyst) -0.0108 * -0.0109 * -0.0114 * (0.0051) (0.0052) (0.0051) CAR (-12,-1) 0.0025 0.0025 0.0024 (0.0031) (0.0031) (0.0031) Debt/Assets -0.0076 -0.0077 -0.0090 (0.0130) (0.0129) (0.0130) EV/Assets 0.0012 * 0.0012 * 0.0012 * (0.0006) (0.0006) (0.0006) RD/Assets 0.0013 0.0015 0.0007 (0.0084) (0.0084) (0.0083) Intangible/ -0.0029 -0.0022 -0.0030 Assets (0.0244) (0.0245) (0.0244) Burn Dummy 0.0302 *** 0.0300 *** 0.0297 *** (0.0075) (0.0076) (0.0076) Distress Dummy -0.0075 -0.0076 -0.0068 (0.0097) (0.0097) (0.0096) Traditional PIPE Common Stock PIPE Intercept 0.1231 ** 0.1217 ** 0.1070 * (0.0592) (0.0596) (0.0585) Investor-Type Yes Yes Yes Fixed Effect Industry Yes Yes Yes Fixed Effect Year Fixed Effect Yes Yes Yes Observations 1,505 1,505 1,505 Adjusted [R.sup.2] (%) 13.60 13.55 13.80 Common All Traditional Stock PIPES PIPEs PIPEs Specifications (10) (11) (12) Panel B. Market Share-Based Rankings IFI -0.0096 *** -0.0090 *** -0.0045 (0.0035) (0.0035) (0.0033) High-Ranking Agent -0.0172 0.0065 -0.0146 (0.0109) (0.0165) (0.0110) IFI x High-Ranking -0.0029 * Agent (0.0015) Ln(Investor -0.0076 -0.0074 -0.0085 Experience) (0.0053) (0.0053) (0.0055) Ln(MV) -0.0054 -0.0053 -0.0039 (0.0045) (0.0045) (0.0044) Ln(Analyst) -0.0141 ** -0.0141 ** -0.0189 *** (0.0055) (0.0055) (0.0055) CAR (-12,-1) 0.0072 ** 0.0070 ** 0.0047 (0.0031) (0.0031) (0.0029) Debt/Assets -0.0108 -0.0111 -0.0096 (0.0137) (0.0136) (0.0119) EV/Assets 0.0009 ** 0.0008 ** 0.0008 (0.0004) (0.0004) (0.0005) RD/Assets 0.0163 0.0164 0.0158 (0.0117) (0.0116) (0.0120) Intangible/ -0.0212 -0.0205 0.0004 Assets (0.0221) (0.0221) (0.0227) Burn Dummy 0.0256 *** 0.0252 *** 0.0319 *** (0.0080) (0.0079) (0.0076) Distress Dummy -0.0126 -0.0130 -0.0130 (0.0096) (0.0096) (0.0099) Traditional PIPE -0.0894 *** -0.0877 *** (0.0215) (0.0215) Common Stock PIPE 0.0606 *** (0.0096) Intercept 0.3402 *** 0.3334 *** 0.1551 ** (0.0710) (0.0712) (0.0640) Investor-Type Yes Yes Yes Fixed Effect Industry Yes Yes Yes Fixed Effect Year Fixed Effect Yes Yes Yes Observations 2,414 2,414 2,180 Adjusted [R.sup.2] (%) 10.26 10.36 12.62 Common All Traditional Stock PIPES PIPEs PIPEs Specifications (13) (14) (15) Panel B. Market Share-Based Rankings IFI -0.0037 -0.0010 -0.0000 (0.0034) (0.0041) (0.0041) High-Ranking Agent 0.0123 -0.0305 *** -0.0073 (0.0161) (0.0098) (0.0151) IFI x High-Ranking -0.0033 ** -0.0030 * Agent (0.0015) (0.0016) Ln(Investor -0.0083 -0.0057 -0.0055 Experience) (0.0055) (0.0044) (0.0044) Ln(MV) -0.0038 0.0001 0.0001 (0.0044) (0.0044) (0.0044) Ln(Analyst) -0.0190 *** -0.0109 ** -0.0108 ** (0.0055) (0.0052) (0.0052) CAR (-12,-1) 0.0045 0.0025 0.0022 (0.0029) (0.0031) (0.0031) Debt/Assets -0.0098 -0.0054 -0.0064 (0.0119) (0.0126) (0.0125) EV/Assets 0.0008 0.0012 ** 0.0012 ** (0.0006) (0.0006) (0.0006) RD/Assets 0.0160 0.0036 0.0038 (0.0120) (0.0085) (0.0084) Intangible/ 0.0014 -0.0021 -0.0015 Assets (0.0228) (0.0244) (0.0244) Burn Dummy 0.0314 *** 0.0295 *** 0.0291 *** (0.0076) (0.0075) (0.0074) Distress Dummy -0.0136 -0.0076 -0.0079 (0.0098) (0.0096) (0.0095) Traditional PIPE Common Stock PIPE 0.0617 *** (0.0097) Intercept 0.1476 ** 0.1100 * 0.1037 * (0.0643) (0.0592) (0.0592) Investor-Type Yes Yes Yes Fixed Effect Industry Yes Yes Yes Fixed Effect Year Fixed Effect Yes Yes Yes Observations 2,180 1,505 1,505 Adjusted [R.sup.2] (%) 12.81 14.07 14.27 *** Significant at the 0.01 level. ** Significant at the 0.05 level. * Significant at the 0.10 level. Table VI. Relation between Long-Run Stock Returns, High-Ranking Agents, and Investor-Friendly Index (IFI) The sample consists of 2,163 intermediated PIPEs (PIPEs in 2011 and 2012 are excluded for the long-run performance analysis). See Table I for the sample overview. This table reports PIPE issuers' long-run stock returns as market- adjusted equal-weighted cumulative abnormal returns (CARs) and Fama-French four-factor model adjusted CARs over 12 months, 24 months, and 36 months following PIPE issuance, respectively. In Panel A, the issuers' returns are reported for categories with high-and low-ranking agents, respectively. In Panel B, the issuers' returns are provided for groups with above median IFI and below median IFI, respectively. Equal-Weighted Market Adjusted CAR (1,12) CAR (1,24) CAR(1,36) Panel A. CARs by Agent Reputation High-Ranking Agents -5.84% -1.79% 2.43% Low-Ranking Agents -17.20% -33.29% -46.75% Difference 11.36% 31.50% 49.18% p-Value on Differences 0.062 * 0.000 *** 0.000 *** Panel B. CARs by IFI IFI above Median -3.42% -9.98% -13.06% IFI below Median -25.81% ^13.69% -60.55% Difference 22.39% 33.71% 47.49% p-Value on Differences 0.000 *** 0.000 *** 0.000 *** Fama-French Four-Factor CAR (1,12) CAR (1,24) CAR(1,36) Panel A. CARs by Agent Reputation High-Ranking Agents 1.56% 2.67% 9.50% Low-Ranking Agents -13.50% -29.41% -41.12% Difference 15.06% 32.08% 50.62% p-Value on Differences 0.000 *** 0.000 *** 0.000 *** Panel B. CARs by IFI IFI above Median -0.73% -9.45% -16.41% IFI below Median -19.94% -36.59% -46.32% Difference 19.21% 27.14% 29.91% p-Value on Differences 0.000 *** 0.000 *** 0.000 *** *** Significant at the 0.01 level. ** Significant at the 0.05 level. * Significant at the 0.10 level. Table VII. Relation between Long-Run Stock Returns, High-Ranking Agents, and Investor-Friendly Index (IFI) The sample consists of 2,163 intermediated PIPEs (PIPEs in 2011 and 2012 are excluded for the long-run performance analysis). See Table I for the sample overview. These are OLS regressions. The dependent variable is the PIPE issuers' long-run stock returns as market-adjusted equal-weighted cumulative abnormal returns (CARs) over 12 months, 24 months, and 36 months following PIPE issuance, respectively. The IFI is the sum of the investor protection provisions (+), trading restriction provisions (-), and issuer right provisions (-) plus seven (to make the dependent variable non-negative). Agent ranking is based on Carter and Manaster (1990) ranking in Panel A, and based on market share in Panel B. The definitions of all of the other variables are provided in Appendix B. All specifications also include an intercept. Standard errors are clustered by company and reported in the parentheses. Dependent Variable All PIPES CAR (1,12) CAR (1,24) CAR (1, 36) Specifications (1) (2) (3) Panel A. Carter and Manaster (1990) Rankings IFI 0.0511 *** 0.0768 *** 0.1205 *** (0.0194) (0.0269) (0.0362) High-Ranking Agent 0.1166 * 0.2505 ** 0.3989 *** (0.0680) (0.1050) (0.1399) Discounts -0.0966 -0.1969 -0.1797 (0.1915) (0.2569) (0.3164) Ln(MV) -0.1523 *** -0.2291 *** -0.3051 *** (0.0354) (0.0471) (0.0600) Ln(Analyst) 0.0928 ** 0.1713 *** 0.2698 *** (0.0375) (0.0543) (0.0720) CAR (-12,-1) 0.1097 *** 0.1850 *** 0.2668 *** (0.0288) (0.0365) (0.0484) Debt/Assets 0.5658 0.7908 ** 0.8541 ** (0.3593) (0.3769) (0.3800) EV/Assets -0.0027 -0.0092 -0.0122 (0.0048) (0.0088) (0.0128) RD/Assets 0.2529 ** 0.3184 ** 0.3724 ** (0.1201) (0.1565) (0.1735) Intangible/Assets 0.1883 0.1487 0.1520 (0.2514) (0.3105) (0.3876) Burn Dummy -0.0488 0.0076 0.0170 (0.0660) (0.0865) (0.1068) Distress Dummy -0.0243 0.0806 0.1675 (0.0663) (0.1000) (0.1341) Traditional PIPE 0.1992 * 0.2477 0.2905 (0.1031) (0.1550) (0.1919) Common Stock PIPE Intercept -0.3048 -0.4545 -0.8001 * (0.2605) (0.3665) (0.4576) Investor Fixed Effect Yes Yes Yes Industry Fixed Effect Yes Yes Yes Observations 2,163 2,163 2,163 Adjusted [R.sup.2] (%) 9.88 11.80 12.60 Dependent Variable Traditional PIPEs CAR (1,12) CAR (1,24) CAR (1,36) Specifications (4) (5) (6) Panel A. Carter and Manaster (1990) Rankings IFI 0.0463 ** 0.0606 ** 0.1071 *** (0.0209) (0.0284) (0.0380) High-Ranking Agent 0.0868 0.1769 0.2993 ** (0.0705) (0.1076) (0.1439) Discounts -0.0814 0.0586 0.0610 (0.2108) (0.2716) (0.3398) Ln(MV) -0.1520 *** -0.2240 *** -0.3027 *** (0.0396) (0.0520) (0.0664) Ln(Analyst) 0.0994 ** 0.1674 *** 0.2567 *** (0.0399) (0.0571) (0.0750) CAR (-12,-1) 0.1105 *** 0.2000 *** 0.3095 *** (0.0332) (0.0426) (0.0562) Debt/Assets 0.5564 0.7611 * 0.8211 ** (0.3724) (0.3908) (0.3908) EV/Assets -0.0029 -0.0156 -0.0192 (0.0078) (0.0132) (0.0192) RD/Assets 0.2869 ** 0.3554 ** 0.3729 ** (0.1228) (0.1607) (0.1767) Intangible/Assets 0.2364 0.1875 0.1848 (0.2697) (0.3327) (0.4083) Burn Dummy -0.0670 0.0034 -0.0008 (0.0738) (0.0942) (0.1145) Distress Dummy -0.0433 0.0773 0.1761 (0.0675) (0.1013) (0.1333) Traditional PIPE Common Stock PIPE -0.0316 -0.0966 -0.1067 (0.0669) (0.0869) (0.1087) Intercept -0.0016 0.0518 -0.2206 (0.2530) (0.3476) (0.4312) Investor Fixed Effect Yes Yes Yes Industry Fixed Effect Yes Yes Yes Observations 1,937 1,937 1,937 Adjusted [R.sup.2] (%) 9.83 12.03 13.29 Dependent Variable Common Stock PIPEs CAR (1,12) CAR (1,24) CAR (1,36) Specifications (7) (8) (9) Panel A. Carter and Manaster (1990) Rankings IFI 0.0527 ** 0.0952 *** 0.1692 *** (0.0259) (0.0343) (0.0486) High-Ranking Agent 0.0432 0.1330 0.2931 * (0.0684) (0.1157) (0.1521) Discounts -0.0457 0.0514 0.0400 (0.2797) (0.3451) (0.4455) Ln(MV) -0.1308 *** -0.2488 *** -0.3252 *** (0.0377) (0.0557) (0.0747) Ln(Analyst) 0.1415 *** 0.2201 *** 0.2730 *** (0.0468) (0.0654) (0.0895) CAR (-12,-1) 0.1284 *** 0.2331 *** 0.3657 *** (0.0367) (0.0538) (0.0737) Debt/Assets 0.1012 0.2986 * 0.4006 * (0.1185) (0.1729) (0.2220) EV/Assets -0.0001 -0.0081 -0.0090 (0.0087) (0.0139) (0.0198) RD/Assets 0.0876 0.1461 0.1180 (0.0853) (0.1565) (0.1869) Intangible/Assets 0.0273 0.0581 0.2023 (0.2683) (0.3494) (0.4731) Burn Dummy -0.0389 0.0358 0.0267 (0.0681) (0.0969) (0.1287) Distress Dummy -0.0269 0.0831 0.1880 (0.0749) (0.1182) (0.1573) Traditional PIPE Common Stock PIPE Intercept -0.0747 -0.1320 -0.6658 (0.2697) (0.4008) (0.5237) Investor Fixed Effect Yes Yes Industry Fixed Effect Yes Yes Yes Observations 1,333 1,333 1,333 Adjusted [R.sup.2] (%) 7.69 11.26 13.04 Dependent Variable All PIPES CAR (1,12) CAR (1,24) CAR (1, 36) Specifications (1) (2) (3) Panel B. Market Share-Based Rankings IFI 0.0475 ** 0.0749 *** 0.1206 ** (0.0192) (0.0269) (0.0364) High-Ranking Agent 0.2341 * 0.2562 * 0.2723 * (0.0966) (0.1219) (0.1562) Discounts -0.0864 -0.1982 -0.1945 (0.1901) (0.2589) (0.3211) Ln(MV) -0.1653 ** -0.2379 *** -0.3086 ** (0.0391) (0.0503) (0.0634) Ln(Analyst) 0.1040 ** 0.1925 *** 0.3019 *** (0.0371) (0.0544) (0.0722) CAR (-12,-1) 0.1104 *** 0.1852 *** 0.2664 *** (0.0284) (0.0364) (0.0487) Debt/Assets 0.5240 0.7448 * 0.8050 * (0.3334) (0.3524) (0.3580) EV/Assets -0.0023 -0.0092 -0.0127 (0.0047) (0.0089) (0.0130) RD/Assets 0.2457 ** 0.3112 ** 0.3654 ** (0.1162) (0.1536) (0.1730) Intangible/Assets 0.1773 0.1198 0.1029 (0.2456) (0.3059) (0.3841) Burn Dummy -0.0362 0.0292 0.0482 (0.0637) (0.0855) (0.1065) Distress Dummy -0.0172 0.0931 0.1859 (0.0667) (0.1008) (0.1354) Traditional PIPE 0.1903 * 0.2384 0.2813 (0.1025) (0.1553) (0.1927) Common Stock PIPE Intercept -0.2038 -0.3455 -0.6859 (0.2601) (0.3706) (0.4667) Investor Fixed Effect Yes Yes Yes Industry Fixed Effect Yes Yes Yes Observations 2,163 2,163 2,163 Adjusted [R.sup.2] (%) 10.26 11.78 12.31 Dependent Variable Traditional PIPEs CAR (1,12) CAR (1,24) CAR (1, 36) Specifications (1) (2) (3) Panel B. Market Share-Based Rankings IFI 0.0425 * 0.0577 * 0.1060 *** (0.0207) (0.0283) (0.0379) High-Ranking Agent 0.2313 * 0.2639 * 0.2756 * (0.0985) (0.1238) (0.1571) Discounts -0.0626 0.0743 0.0687 (0.2101) (0.2732) (0.3442) Ln(MV) -0.1666 *** -0.2377 *** -0.3127 *** (0.0435) (0.0555) (0.0701) Ln(Analyst) 0.1093 *** 0.1842 ** 0.2823 *** (0.0396) (0.0574) (0.0753) CAR (-12,-1) 0.1121 *** 0.2020 *** 0.3117 *** (0.0328) (0.0424) (0.0562) Debt/Assets 0.5152 0.7140 * 0.7720 * (0.3469) (0.3658) (0.3681) EV/Assets -0.0025 -0.0155 -0.0198 (0.0078) (0.0132) (0.0193) RD/Assets 0.2792 ** 0.3468 ** 0.3641 ** (0.1188) (0.1576) (0.1753) Intangible/Assets 0.2315 0.1707 0.1509 (0.2638) (0.3268) (0.4024) Burn Dummy -0.0554 0.0218 0.0260 (0.0711) (0.0925) (0.1135) Distress Dummy -0.0375 0.0869 0.1906 (0.0677) (0.1019) (0.1343) Traditional PIPE Common Stock PIPE -0.0270 -0.0927 -0.1047 (0.0646) (0.0858) (0.1087) Intercept 0.0861 0.1543 -0.1099 (0.2569) (0.3527) (0.4392) Investor Fixed Effect Yes Yes Yes Industry Fixed Effect Yes Yes Yes Observations 1,937 1,937 1,937 Adjusted [R.sup.2] (%) 10.31 12.21 13.22 Dependent Variable Common Stock PIPEs CAR (1,12) CAR (1,24) CAR (1, 36) Specifications (1) (2) (3) Panel B. Market Share-Based Rankings IFI 0.0514 * 0.0945 ** 0.1694 *** (0.0257) (0.0344) (0.0487) High-Ranking Agent 0.1230 0.1463 0.2102 (0.0872) (0.1354) (0.1853) Discounts -0.0170 0.0807 0.0762 (0.2808) (0.3457) (0.4472) Ln(MV) -0.1370 *** -0.2527 *** -0.3263 *** (0.0388) (0.0578) (0.0775) Ln(Analyst) 0.1454 *** 0.2295 *** 0.2925 *** (0.0468) (0.0660) (0.0893) CAR (-12,-1) 0.1289 *** 0.2340 *** 0.3672 *** (0.0366) (0.0537) (0.0736) Debt/Assets 0.0930 0.2908 * 0.3919 * (0.1169) (0.1716) (0.2206) EV/Assets -0.0001 -0.0084 -0.0099 (0.0086) (0.0139) (0.0199) RD/Assets 0.0901 0.1504 0.1260 (0.0858) (0.1577) (0.1899) Intangible/Assets 0.0303 0.0537 0.1860 (0.2668) (0.3478) (0.4707) Burn Dummy -0.0374 0.0407 0.0377 (0.0677) (0.0966) (0.1283) Distress Dummy -0.0245 0.0913 0.2065 (0.0748) (0.1181) (0.1584) Traditional PIPE Common Stock PIPE Intercept -0.0398 -0.0921 -0.6105 (0.2724) (0.4104) (0.5360) Investor Fixed Effect Yes Yes Yes Industry Fixed Effect Yes Yes Yes Observations 1,333 1,333 1,333 Adjusted [R.sup.2] (%) 7.83 11.26 12.88 *** Significant at the 0.01 level. ** Significant at the 0.05 level. * Significant at the 0.10 level.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Private Investment in Public Equity|
|Author:||Bengtsson, Ola; Dai, Na|
|Date:||Dec 22, 2014|
|Previous Article:||Sales of secondary shares in SEOs: a comparison across top managers, other insiders, and outsiders.|
|Next Article:||The option and decision to repurchase stock.|