عملکرد شبکه ای بانک های سرمایه گذاری : شواهد حاصل از سرمایه گذاری های خصوصی در حقوق صاحبان سهام عمومی

We examine investment banks' networking function in capital markets, using a sample of Private Investments in Public Equity (PIPEs). We argue that investment banks develop relationships with investors through repeat dealings, and that investment banks' networks of relationship investors form the basis of their networking function. We find that investment banks, especially those with larger investor networks, help issuers attract investors. Correspondingly, an issuer that desires more investors is more likely to hire an investment bank than place the shares directly. We also find that issuers pay higher fees to hire investment banks with larger investor networks. Our empirical findings suggest that the networking function of investment banks is important in securities offerings.

Investment banking is a relationship-based rather than transaction-based business. Investment banks accumulate relationship-specific assets with securities issuers through repeat dealings (James, 1992), and develop relationships with investors through repeat dealings in securities offerings, brokerage services, and analyst research coverage. The resulting investor networks are useful for investment banks in lowering the costs of searching for potential investors. Such investor networks also help investment banks win trust from investors and induce investors to produce and truthfully reveal information (e.g., Benveniste and Spindt, 1989, Benveniste and Wilhelm, 1990, Cornelli and Goldreich, 2001, Sherman and Titman, 2002, Sherman, 2000 and Sherman, 2005). We refer to the use of investor networks by investment banks to certify, market, and distribute securities to investors as their networking function. In many financial markets, financial institutions and investors form widespread networks. Hochberg et al. (2007) show that better networked venture capital (VC) firms demonstrate better investment performance, and that start-up companies supported by better networked VCs are more likely to survive. For securities offerings, issuing firms often cite distributional ability as one of the most important factors in selecting managing underwriters (Corwin and Schultz, 2005). Despite the anecdotal evidence, it is challenging to empirically document the importance of investment banks' networking function because investment banks are not required to disclose and are highly protective of order book and allocation information in public securities offerings.3 In this paper, we shed light on the importance of investment bank–investor relationships for the success of new securities issues by studying a sample of 2096 Private Investments in Public Equity (PIPEs) that occurred during 2000–2005. PIPEs are ideal for this analysis for two reasons. First, PIPE issuers can choose to place securities either directly with investors or indirectly with the help of one or more investment banks. Second, allocation information for PIPEs is disclosed. Using data on prior deals, we are able to construct a measure of investor network for each investment bank in our sample: the number of investors related to the investment bank in prior deals (also referred to as the investment bank's relationship investors). This measure helps capture the investment bank's networking ability. With this measure, as well as the fact that not all PIPEs in our sample use the service of investment banks, we are able to answer three questions. First, are investment banks with larger investor networks able to attract more investors to subsequent deals? Second, do issuing firms recognize the importance of networking ability when deciding whether to employ an investment bank? Third, do issuing firms pay for investment banks' networking abilities? We first provide evidence on the importance of investment banks' networking function and on the source of their networking abilities by examining the determination of the number of investors in a PIPE deal. We find that investment banks help attract investors to a deal and that those with stronger networking abilities (larger investor networks) can attract more investors, especially new investors without a prior relationship with the issuer. Specifically, the number of investors in a PIPE deal increases by 41.9% if the deal is intermediated instead of being directly placed, other things being equal. Conditional on intermediation, a one standard deviation increase from the sample median in networking ability of the deal's placement agent(s) further increases the number of participating investors by 14.1% (throughout the paper, we use the terms “investment bank” and “placement agent” interchangeably). These findings suggest that investment banks play an important networking role and they help issuing firms achieve more dispersed ownership structures. To further understand how investment banks help attract investors, we also examine the determinants of investor participation in a deal from the investor's perspective. We find that an investor is more likely to participate in a PIPE deal if it has a prior relationship with the placement agent(s). While the unconditional likelihood of investor participation in a PIPE deal is 2.5%, this probability increases to 8.8% if an investor has a prior relationship with the placement agent(s). This finding demonstrates that prior relationships with investors help improve an investment bank's networking ability. Second, we provide new insight into issuing firms' selective use of investment banking service. Several papers have compared security offerings with and without investment banks. Smith (1977) raises the question of why underwriters are employed in the vast majority of public offerings, despite lower issuing costs for right offerings (see Eckbo et al. (2007) for an excellent survey of seasoned equity offerings). Similarly, Scholes and Wolfson (1989) question why many eligible shareholders do not participate in discount dividend-reinvestment and stock-purchase plans. For mergers and acquisitions, Servaes and Zenner (1996) find that transaction costs, contracting costs, and information asymmetries are related to the acquirer's decision of whether to hire an investment bank. We build on these studies and examine whether issuers recognize the networking role of investment banks in their decisions of whether to hire an investment bank. We use issue and firm characteristics as proxies for issuing firms' different needs for access to investors. We find that smaller issues and issues motivated by strategic alliances are less likely to be intermediated. The fact that investment banks are selectively used by issuers based on their need for access to investors suggests that an important reason for hiring investment banks is their access to networks of investors. The selective use of investment banks also provides new insight into broad issues such as the way issuers choose to raise capital and its potential association with their desired ownership structures. Finally, we examine the relation between the fees paid by issuers and their placement agents' networking abilities. We find that issuers pay higher fees to investment banks with larger investor networks. More specifically, when an issuing firm moves from a placement agent with the median number of relationship investors of 42 to one with the mean of 67, the fees increase by 12 basis points, or $28,000. Our results thus suggest that issuing firms do pay for investment banks' networking function. Our major contribution is threefold. First, to the best of our knowledge, we are the first to emphasize the networking function of investment banks and use their investor networks to explicitly measure their networking abilities. Second, we provide the first direct evidence on how investment banks' networking abilities can affect investor participation and how issuing firms can selectively use investment banks to achieve their desired ownership structures when they raise capital. The direct evidence on the networking function of investment banks adds to the vast literature on the role of investment banks in securities offerings for information production, certification, and marketing (e.g., Benveniste and Spindt, 1989, Benveniste and Wilhelm, 1990, Carter and Manaster, 1990, Corwin and Schultz, 2005, Gao and Ritter, 2008 and Huang and Zhang, 2008). Third, the existing literature focuses on investment bank compensation for equity issuances in the public market (e.g., Altinkiliç and Hansen, 2000 and Chen and Ritter, 2000). We add to the literature by providing evidence on how investment banks are compensated in the PIPE market, and more importantly, whether investment banks are compensated for their networking abilities. The rest of the paper is organized as follows. Section 2 presents the hypotheses. Section 3 describes the data and reports the summary statistics. Section 4 analyzes the determination of the number of investors in a PIPE deal and an investor's participation decision. Section 5 analyzes the issuer's decision of whether to hire placement agents and the determination of agent compensation. Section 6 concludes the paper.

Investment banks develop relationships with investors through their repeat dealings in brokerage services, analyst research coverage, and securities offerings. The resulting networks of investors assist investment banks in performing a networking function inwhich the investment banks certify, market, and distribute securities to investors. Because the PIPE market differs from the public market in several aspects, it provides us with a unique opportunity to shed new light on investment banks' networking function in capital markets. We fi nd that investment banks, especially those with larger investor networks, help issuers attract investors. We also fi nd that investors are more likely to participate in an issue if they have an existing relationship with the hired investment bank(s). These fi ndings suggest that investment banks' investor relationships contribute to their networking function, and that investment banks playan important role in helping issuing fi rms achieve less concentrated ownership structures. We also fi nd that issuing fi rms that desire more investors are more likely to hire investment banks than place the shares directly, suggesting that they do recognize investment banks' networking role. Finally, we fi nd that PIPE issuers pay higher fees when they hire investment banks with larger investor networks. The PIPE market has become an increasingly signi fi cant source of external fi nancing for public fi rms. Sagient Research Systems, a leadingproviderof information on PIPEs,reports total PIPE issuances of less than $2 billion in 1995 andabout $141 billionin 2007 (excluding structured equity lines and Canadian-domiciled issuers). Our paper complements other recent papers in understanding this increasingly important market (see Brophy et al. (2006) , Chaplinsky and Haushalter (2006) , Huson et al. (2006) , Meidan (2006) , and Dai (2007) , among others). In this paper, we focus on the bene fi ts of investment banks' networking function. We note that repeat dealings between investment banks and investors could also cultivate agency problems. For example, investment banks could allocate underpriced PIPEs to their favored investor clients (e.g., certain hedge funds) in order to attract brokerage commissions, just as they do with IPOs ( Loughran and Ritter, 2004; Reuter, 2006 ). We leave this important issue to future research.