Venture Capital Firms and the Retention of Founder-Ceos Timothy G

Strategic Entrepreneurship Journal Strat. Entrepreneurship J., 3: 199–217 (2009) Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/sej.71

DANCE WITH THE ONE THAT BROUGHT YOU? VENTURE CAPITAL FIRMS AND THE RETENTION OF FOUNDER-CEOS TIMOTHY G. POLLOCK,1* BRET R. FUND,2 and TED BAKER3 1Smeal School of Business, The Pennsylvania State University, University Park, Pennsylvania, U.S.A. 2Leeds School of Business, University of Colorado-Boulder, Boulder, Colorado, U.S.A. 3College of Management, North Carolina State University, Raleigh, North Carolina, U.S.A.

We consider how a venture capital fi rm’s perceived uncertainty in new and uncertain industry environments affects its decisions to retain founder-CEOs at companies they take public. We further consider how the human capital of the founder-CEO, the overall experience of the venture capital fi rm (VCF), and the VCF’s specifi c experience with the new industry moderate the relationship between industry-based uncertainty and founder-CEO retention. We explore these issues in the context of 340 venture capital fi rm investments in Internet sector start-ups that went public from 1995 to 2000. We fi nd evidence that industry-based uncertainty decreases the likelihood of founder-CEO retention, that founder-CEO human capital and VCF Internet- sector experience decreases the effects of these uncertainties on founder-CEO retention for business-to-business (B2B) fi rms, but increases them for business-to consumer (B2C) fi rms, and that VCF age further decreases the likelihood that the founder-CEOs of B2C fi rms will be retained. Copyright © 2009 Strategic Management Society.

INTRODUCTION founded and grown their new companies (Boeker and Karichalil, 2002; Flamholtz and Randle, 2000; Two key assumptions underlying research on upper Nelson, 2003; Wasserman, 2003). An equally key echelons and corporate governance are that who assumption that has received far less scholarly atten- leads a company matters greatly, and that changes tion is that owners somehow know whether and how in leadership can have signifi cant consequences for to use their ability to replace fi rms’ leaders in ways a fi rm’s strategy and performance (e.g., Bebchuk that will enhance fi rm performance. However, the and Fried, 2004; Hambrick and Mason, 1984; Shen general assumption that owners are skilled in their and Cannella, 2002; Westphal and Fredrickson, performance of governance activities has been chal- 2001). In the context of entrepreneurial start-ups, lenged (see Bebchuk and Fried [2004] for a review these leaders are generally the individuals who have of this literature), and research on the effects of founder-CEO replacement on fi rm performance has been mixed (e.g., Beckman, Burton and O’Reilly, Keywords: Initial public offerings; founder; venture capitalist; corporate governance; cognition; symbolic action; decision 2007; Certo, et al., 2001; Daily and Dalton, 1992; making He, 2008; Jayaraman, et al., 2000; Nelson, 2003; *Correspondence to: Timothy G. Pollock, Smeal School of Willard, Krueger, and Feeser, 1992). Business, The Pennsylvania State University, 417 Business Building, University Park, PA 16802, U.S.A. Further, little research has considered what E-mail: [email protected] happens to owners’ attempts to shape fi rm

200 T. G. Pollock, B. R. Fund, and T. Baker leadership during upheavals in the organizational in an emerging and fast-changing industry. In addi- landscape around the creation of major new indus- tion, we consider how founder-CEOs’ human capital tries and the promulgation of new technologies. The and different types of VCF experience shape the novelty and uncertainty of such developments creates extent to which industry-based uncertainty infl u- new ambiguities, confl icting interests, different ences founder-CEO retention decisions. means of success, and alternative ways for the con- We test our hypotheses using a sample of 340 fl icting interests and preferences to play out (Santos VCF investments in Internet fi rms that went public and Eisenhardt, 2009; Sine, Mitsuhashi, and Kirsch, from 1995 to 2000. VCFs play a critical role in 2006), all of which can challenge existing under- shaping young fi rms’ leadership and governance standings of who is appropriate to lead a fi rm. structures as they attempt to deal with major issues We attempt to address this issue by studying how of uncertainty and incentive alignment between owners’ perceptions of industry-based uncertainty sources and users of risk capital, especially in prom- infl uence decisions to retain or replace founder- ising new industries (Sahlman, 1988; Sapienza and CEOs in companies preparing to conduct initial Gupta, 1994). The infusion of venture capital into a public offerings (IPOs). Prior research has consid- young fi rm is a fundamental step in the process of ered the important role that governance and execu- separating ownership from control (Berle and Means, tive recruiting plays in managing uncertainties as 1932; Wasserman, 2003), and the leadership deci- fi rms prepare for their IPOs (Beatty and Zajac, 1994; sions made at this early stage in an organization’s Certo, et al., 2003; Chen, Hambrick, and Pollock, life can have signifi cant, ongoing consequences for 2008). In this study, we consider how one particular the fi rm (Baron, Hannan, and Burton, 1999; Fischer set of active and powerful owners—venture capital and Pollock, 2004; He, 2008). fi rms (VCFs)—infl uenced the leadership of start-ups We contribute to the entrepreneurship literature they invested in during the Internet bubble of the by exploring the important role that VCFs’ perceived mid- to late-1990s (Hendershott, 2004; Sine, et al., uncertainty plays in shaping the founder-CEO reten- 2006). Both academics and practitioners have noted tion decision in an emerging industry. Our study that replacing founder-CEOs with outside execu- contributes to the governance literature by showing tives is one of the most signifi cant ways that venture how differing experiences infl uence owners’ percep- capital fi rms shape their portfolio companies (Boeker tions of uncertainty, thereby shaping fundamental and Karichalil, 2002; Flamholtz and Randle, 2000; governance decisions. Levensohn, 2006; Nelson, 2003; Wasserman, 2003). Indeed, as one experienced venture capitalist recently wrote, ‘Executing that task [managing the founder- THEORY AND HYPOTHESES CEO transition] is arguably the single most important value a VC board will add in the company’s The benefi ts and costs of founder-CEO retention and replacement lifetime’ (Levensohn, 2006: 5). And while prior research has explored how factors such as stock Deciding whether or not to retain the founding CEOs ownership, founder characteristics, and life cycle of promising fi rms in their investment portfolios is factors (such as hitting milestones) infl uence this a key strategic decision made by VCFs, second in decision (e.g., Boeker and Karichalil, 2002; Jain and importance only to their decisions about which Tabak, 2008; Wasserman, 2003), questions regard- investments to make (Hellmann and Puri, 2002; ing the role played by VCFs’ perceptions of the Kaplan and Stromberg, 2004; Wasserman, 2003). industry uncertainties they face, and how these per- This decision can have substantial consequences for ceptions infl uence whether a founder-CEO is retained the structure of the fi rm (Baron, Burton, and Hannan, or replaced, have been left largely unaddressed. 1999) and its ability to retain key employees (Baron, We draw on cognitive theories of perception and Hannan, and Burton, 2001; Beckman, et al., 2007), decision making (Carpenter, Pollock, and Leary, acquire fi nancial resources (Boeker and Karichalil, 2003; Cyert and March, 1963; Haunschild and 2002; Certo, et al., 2001; Nelson, 2003) and even Miner, 1997; Sitkin and Pablo, 1992) and theories survive (Fischer and Pollock, 2004; He, 2008). Evi- of symbolic action (DiMaggio and Powell, 1983; dence regarding the performance benefi ts of founder- Meyer and Rowan, 1977; Pfeffer and Salancik, CEO retention remains quite mixed. Some studies 1978) to examine how perceived uncertainty infl u- fi nd that replacing a founder-CEO enhances perfor- ences whether VCFs retain or replace founder-CEOs mance outcomes (Beckman et al., 2007; Certo et al.,

2001; Hambrick and Crozier, 1985), while others substantially less relationship confl ict among board fi nd no signifi cant performance differences (Daily members. and Dalton, 1992; Willard, et al., 1992), and yet Mixed effects are also found when studying the others fi nd benefi ts to founder-CEO retention relationship between founder-CEO presence and (Fischer and Pollock, 2004; Forbes, Korsgaard, and initial fi nancial market reactions at IPO. Nelson Sapienza, Forthcoming; He, 2008; Jayaraman et al., (2003) found that fi rms going public with founder- 2000; Kor, 2003; Nelson, 2003). CEOs receive a higher price premium over the book The relationship between founder-CEO retention value of the company, suggesting investors may or replacement and fi rm performance is complicated. place a higher value on the intangible benefi ts asso- The presence of a founder-CEO who retains sub- ciated with the founder’s presence. In contrast, other stantial ownership power and, thus, greater control scholars (Certo et al., 2001) found that founder-CEO over the practices and strategic direction of the presence results in a greater run-up in stock price on company increases the probability the fi rm will the fi rst day of trading—that is, greater underpric- survive during the fi ve years following its IPO ing—which may signal that there is greater uncer- (Fischer and Pollock 2004). Similarly, recent tainty about the fi rm’s future prospects among research has shown that founder-CEO replacements investors when the founder-CEO continues to lead resulting in changes to the organizational blueprint the fi rm. (i.e., the norms and culture of the organization) Given the complicated relationship between toward more bureaucratic forms of organizing founder-CEO presence and various organizational (Baron et al., 2001) and fi lling key jobs with succes- outcomes, it is perhaps unsurprising that research sors who do not fi t the local defi nitions of jobs estab- fi nds substantial variance across deals in whether lished by their original holders (Burton and Beckman, founder-CEOs are retained. A recent survey of VCFs 2007), increase employee turnover rates, exacerbat- suggests founder-CEOs are replaced about 67 ing the loss of key scientifi c and other human capital. percent of the time (Adams, 2005); academic studies This loss of talent and the organizational disruptions typically put the number between 40 and 60 percent1 it engenders can threaten fi rm performance and sur- (Boeker and Karichalil, 2002; Certo et al., 2001; vival (Aldrich, 1999; Baron et al., 2001). In another Fischer and Pollock, 2004; Jain and Tabak, 2008; intriguing study, Beckman, Burton, and O’Reilly Nelson, 2003; Wasserman, 2003). Although in prac- (2007) discovered that while founder exit from a tice founder-CEOs are retained almost as often as company increased the probability the fi rm would they are replaced, when going into deals VCFs fre- eventually go public (VCFs’ generally preferred quently adopt as their starting position the institu- form of exit) by 25 percent, top management team tionalized assumption that founding executives may (TMT) exits had an even greater negative effect need to be replaced by more professional managers (34%) on the likelihood of going public. This fi nding, with different kinds of experience (e.g., Flamholz combined with the insight from Baron and col- and Randle, 2000; Rubenson and Gupta, 1996; leagues’ earlier study (Baron et al., 2001) that Wasserman, 2003). Illustrating this pattern, founder-CEO replacement increases TMT turnover, Wasserman (2003: 154–5) quotes a partner of one suggests that while the direct effects of founder exit venture capital fi rm he studied as noting that, ‘Our increase the likelihood of an IPO, the indirect effects, default assumption when we fi rst look at a company via their infl uence on TMT turnover, reduce the like- is that the founder-CEO can’t lead this company lihood of an IPO. going forward.’ When pushed for an explanation of Finally, in exploring the effects of founder-CEO how the VC ‘had arrived at this rule of thumb,’ the presence on board dynamics and confl ict, Forbes and partner ‘admitted that it was a widely held belief colleagues (Forbes, et al., Forthcoming) found that, within the venture capital industry and one that he contrary to their expectations, when a fi rm was going had not questioned (Wasserman, 2003: 167).’ Simi- through a down round (i.e., accepting VC invest- larly, another VC said that one of his criteria when ment at a lower valuation than in the previous round) deciding whether to invest in a fi rm is whether or the presence of a founder-CEO led to about the same amount of relationship confl ict among directors as a nonfounder CEO. However, during normal fi nanc- 1 Whereas the Adams (2005) survey focuses only on VC-backed start-ups, some of the other studies include mixed samples of ing rounds, where the valuation increases over the VC and non-VC backed companies. Founder-CEOs may be prior round, the presence of a founder-CEO led to less likely to be replaced in non-VC backed fi rms.

Copyright © 2009 Strategic Management Society Strat. Entrepreneurship J., 3: 199–217 (2009) DOI: 10.1002/sej 202 T. G. Pollock, B. R. Fund, and T. Baker not there is likely to be a fi ght if he decides to replace their investment portfolios, VCFs face high levels the founder as CEO. He asks founders, ‘Are you the and multiple sources of uncertainty—including guy (or woman) to take this company all the way?’ technological, product market, fi nancial market, and If they insist that they are, he becomes less inclined regulatory issues—over which they have little or no to invest in the company. Thus, what drives the deci- control. Further, as we’ve already noted, VCFs fre- sion to ultimately retain or replace a founder-CEO quently enter a relationship with substantial uncer- becomes an important question.2 tainty about whether the founder will be the right The fi rst and most obvious answer would be poor individual to lead the company all the way to and performance. However, this assumption is problem- through an IPO. Perceptions of uncertainty about a atic. Given the early stage of both market and fi rm’s leadership, however, are a potential source of company development at which many VCF invest- perceived uncertainty4 that VCFs can address. ments are made, CEO performance may sometimes Prior research suggests organizations frequently be diffi cult to judge, especially during periods draw on their existing repertoires of routines when between major milestones. More importantly, in responding to new and uncertain situations (Baker empirically testing the assumption that founder and Nelson, 2005; March and Olsen, 1976; Weick, replacements occur in response to poor performance, Sutcliffe, and Obstfeld, 1999), even if the action or Wasserman (2003) found just the opposite: founder- routines were developed to deal with a different CEOs were more likely to be replaced when the environmental context or set of issues. This is company was performing well. Wasserman’s study because existing routines provide a useful starting showed that founders who led their fi rms to success- place from which organizations can begin enacting fully develop their initial product, complete more their environments (Weick, 1995), as they allow the rounds of fi nancing, and successfully raise larger organization to engage in concrete behaviors over amounts of fi nancing were more likely to be replaced. which it has some control. Research has also dem- In Wasserman’s interpretation, VCFs do not see the onstrated that in uncertain conditions organizations founder-CEO’s skills in dealing with the fi rm’s criti- are more likely to take what they believe are safe cal contingencies at one stage as evidence that the courses of action (Haunschild and Miner, 1997) that founder-CEO’s skills will continue to be a good are accepted as a reasonable and legitimate means match for fi rm needs in the future. He also suggested for responding to uncertainty within their industry that VCFs make these judgments proactively, making (Cyert and March, 1963; DiMaggio and Powell, ‘CEO changes before a mismatch would cause prob- 1983; Haunschild, 1994). Retaining a founder-CEO lems’ (Wasserman, 2003: 165). Thus, poor prior is a potentially uncertainty-increasing gamble that performance does not appear to be the primary driver the current leadership will be able to continue to of the founder retention or replacement decision, as meet future demands, while replacing a founder- is often the case in large public corporations (see CEO with a proven professional whose leadership Finkelstein, et al., 2009, for a recent review). capabilities are better known reduces a specifi c Instead, we argue that VCFs’ perceived uncer- source of leadership uncertainty (Flamholtz and tainty about future performance may strongly infl u- Randle, 2000; Rubenson and Gupta, 1996). ence whether the founder-CEO is retained.3 Across The notion that founding entrepreneurs are not necessarily the right individuals to take a company through all stages of its life cycle not only has 2 In addition to the benefi ts of founder-CEO retention discussed earlier, prior research has suggested that founders who retain high levels of ownership increase their likelihood of maintain- ing the CEO position (Boeker and Karichalil, 2002; Wasser- man, 2003). We acknowledge that there are additional factors or not these perceptions are accurate), and through their effects such as CEO ownership power that infl uence the founder-CEO on a VCF’s perceived uncertainty, infl uence the likelihood that retention outcome, and attempt to control for these additional a founder-CEO will be retained. Our empirical analysis tests infl uences in our analysis. the relationships between these indicators and the likelihood of 3 We differentiate in this study between sources of uncertainty founder-CEO retention. and perceptions of uncertainty. This distinction is important, 4 It is important to recognize that we are not claiming this is the because we are making no theoretical claims about whether or only, or even the primary, source of uncertainty. Rather, we are not particular actions VCFs may take—or indicators VCFs may arguing that it is the potential source of uncertainty over which rely on in their decision making—affect actual levels of uncer- the VCF is likely to have the most direct control and, thus, is tainty. Rather, our theoretical and analytical approach is to the source of uncertainty VCFs will be most likely to attempt explore how different indicators may affect the extent to which to infl uence if they feel the need to act, regardless of its role in a VCF perceives more or less uncertainty (regardless of whether the overall level of uncertainty perceived.

Copyright © 2009 Strategic Management Society Strat. Entrepreneurship J., 3: 199–217 (2009) DOI: 10.1002/sej Founder-CEO Retention 203 currency among venture capitalists (Wasserman, going public at an unprecedented rate and at a 2003), it is also widely perceived outside the VC younger age than had previously been typical. Many industry as a legitimate belief (e.g., Flamholtz and fi rms were going public with untried business models Randle, 2000; Hambrick and Crozier, 1985) and has and limited revenue streams. In addition, although been used to justify the actions of other institutional the Internet has existed since 1969, the technologies actors, such as investment banks (Certo et al., 2001). underlying the development of commercializable Replacing the founder-CEO is a concrete action Internet-related business applications were still in VCFs can take that is widely perceived as legitimate, their infancy in the mid- to late-1990s, and it was is associated with uncertainty reduction, and is seen not clear how much sustained demand there was as an important signal of good and active gover- going to be for the products and services being nance by other stakeholders (Shen and Cannella, offered (Santos and Eisenhardt, 2009; Sine et al., 2002). Thus, replacing the founder-CEO may not 2006). As a consequence, the purpose and nature only yield instrumental benefi ts because it replaces of the fi rms themselves changed rapidly (Rindova an unproven leader with an individual who has expe- and Kotha, 2001), dot com founder-CEOs were rience dealing with the myriad sources of uncer- popularly characterized5 as young and unproven tainty the fi rm will face as it moves forward, it can (Kuemmerle, 2002), investment analysts were creat- also yield symbolic benefi ts (Meyer and Rowan, ing new metrics to justify sky-high fi rm valuations 1977; Pfeffer and Salancik, 1978). Meyer and Rowan that could not be justifi ed using traditional indica- (1977: 355) note that, ‘activity has ritual signifi - tors, and management gurus were publishing books cance . . . it maintains appearances and validates an with titles like Blown to Bits, arguing that the skills organization’ even if its substantive effects on actual and capabilities needed to run old economy bricks organizational effi ciency are unclear. Replacing the and mortar companies had been rendered moot. CEO is often employed as a ritual means of moving Facing these considerable uncertainties as unprec- an organization in a new strategic direction or dis- edented amounts of investment dollars poured into tinguishing between past and future organizational their coffers, VCFs that invested in Internet compa- events (Boeker, 1992; Finkelstein et al., 2009; nies were under considerable pressure to make good Wiesenfeld, Wurthmann, and Hambrick, 2008). leadership decisions in order to infl uence the perfor- Further, professionalization can play an important mance of their portfolios. However, it is also impor- symbolic as well as substantive role because it is an tant to note that not all subsegments of the emerging activity that absorbs perceived uncertainty (Meyer Internet industry were likely to be equally unfamiliar and Rowan, 1977). Therefore, we expect that high to VCFs. Indeed, consideration of the different types levels of perceived uncertainty—whatever its of fi rms identifi ed as ‘Internet companies’ reveals source—will make founder-CEO retention less signifi cant differences in their similarity to non- likely (Boeker and Karichalil, 2002) and that lower Internet technology fi rms. Infrastructure companies VCF perceptions of uncertainty increase the likeli- such as UUNet, Cobalt Networks, Netzero, and hood the founder-CEO will be retained. AboveNet Communications had business models and assets that were the most similar to other non- Internet related technology companies. Business-to- Founder-CEO retention decisions in a new business (B2B) fi rms such as AdForce, Doubleclick, industry context Ariba, and Marimba often pursued business models An important source of uncontrollable uncertainty is that were similar to traditional companies selling the nature and stage of development of the start-up to other businesses, but introduced products and fi rm’s industry. Although VCFs specialize in making services (e.g., those related to Internet advertising) investments in highly promising but highly uncer- that were quite novel. Finally, Business to consumer tain growth industries, even for these fi rms the early (B2C) fi rms such as Amazon, eBay, WebVan, days of Internet commercialization were extraordi- Peapod, Pets.com, and others put many traditional narily novel (Aldrich and Baker, 2001). While all IPOs are uncertain (Pollock and Rindova, 2003; Pollock, Rindova and Maggitti, 2008; Ritter and Welch, 2002), the level of uncertainty surrounding 5 Although the assumption that most dot coms were founded by individuals in their twenties has become a popular myth, in our an IPO was magnifi ed for Internet sector fi rms, espe- sample the average founder-CEO was 41 (S.D. 7.5) at the time cially during the mid- to late-1990s when fi rms were of the IPO, and ages ranged from 24 to 65.

Copyright © 2009 Strategic Management Society Strat. Entrepreneurship J., 3: 199–217 (2009) DOI: 10.1002/sej 204 T. G. Pollock, B. R. Fund, and T. Baker consumer sales processes on the Web, but also intro- experience—or human capital—perceived as neces- duced some of the most novel and untried business sary to successfully run a fast-growing enterprise, models seen during this period. VCFs may be more confi dent in the CEO’s ability Thus, the level of industry-based uncertainty a to deal effectively with industry uncertainty. This VCF perceived was likely to vary based on the reduces their perceptions of the risks associated with industry subsegment in which a portfolio fi rm com- industry uncertainty, thus reducing the effects of peted. Whether B2B or B2C industries were per- industry subsegment on founder retention decisions. ceived as more uncertain is largely an empirical Thus, we hypothesize: question that we will explore in our analysis; Hypothesis 2: The founder-CEO’s human capital however, as a group, B2B and B2C companies intro- will reduce the negative effects of Internet indus- duced new and untried products, strategies, and try subsegment uncertainty on founder-CEO technologies that likely made them less familiar and retention. more uncertain than infrastructure companies. We, thus, expect that VCFs were less likely to retain a Prior research suggests that simply having prior fi rm’s founder-CEO in B2B and B2C companies experience with an activity or situation can reduce than in infrastructure companies and hypothesize: associated perceptions of uncertainty (Carpenter et al., 2003; Sitkin and Pablo, 1992). Both the VCF’s Hypothesis 1: The more uncertainty associated experience with Internet companies generally, and with an IPO fi rm’s Internet industry subsegment, the VCF’s overall experience with emerging indus- the less likely a VCF will be to retain the tries are types of VCF experience that may reduce founder-CEO. its perceived uncertainty about the B2B and B2C industry subsegments and, therefore, increase the likelihood that a founder-CEO will be retained. Moderators of industry uncertainty Research suggests that VCFs learn important Hypothesis 1 represents our baseline argument. industry lessons by paying attention to their experi- However, we also anticipate that different kinds of ences with portfolio fi rms (De Clercq and Sapienza, experience can infl uence the extent to which indus- 2005). Thus, a VCF’s prior experience in the Internet try uncertainty affects VCFs’ overall perceptions of industry could provide it with the opportunity to uncertainty and decreases the likelihood a founder- develop the skills and deal access necessary to locate, CEO will be retained. We consider three types of choose, and manage investments in this challenging experience that may reduce a VCF’s perceived context, thereby affecting its perceptions of industry uncertainty and increase the likelihood of founder- uncertainty. Consistent with these expectations, in a CEO retention: (1) the prior experience of the recent study of VC experience on portfolio fi rms’ founder-CEO; (2) the VCF’s experience with strategic decision making in the medical device indus- Internet companies; and (3) the VCF’s general try, Carpenter and his colleagues (2003: 806) experience with emerging industries. suggested that, ‘perceptions can be driven by the As noted previously, a major concern leading to expectation that particular experiences will lead to an the replacement of founder-CEOs is that they do not actual decrease in the risks being faced.’ They found have the requisite skills and experience to run a that if a venture capitalist sitting on the fi rm’s board company as it grows and develops (Flamholtz and did not have prior international experience, the fi rm Randle, 2000; Rubenson and Gupta, 1996). However, was less likely to pursue an internationalization strat- individual entrepreneurs vary signifi cantly in the egy; but if the VC had prior experience the effect was amount of knowledge and experience they bring reversed, and the fi rm became more likely to pursue to their new venture. Many entrepreneurs have this strategy. They argued this was because the VC’s substantial experience working in new start-ups prior experience and exposure to international issues (Wasserman, 2003) and have served as CEOs and markets reduced his/her perceived uncertainties (Dimov and Shepherd, 2005; Jain and Tabak, 2008; of the risks associated with these actions. Following Wasserman, 2003). In addition, prior research has this same logic, we expect that VCFs with greater shown that possession of an MBA degree is also experience guiding Internet fi rms to IPOs will per- viewed positively by venture capitalists and other ceive lower levels of uncertainty than will VCFs with investors (Dimov and Shepherd, 2005). Thus, to the less experience in successfully developing Internet extent that a founder-CEO has the training and companies.

Similar arguments can be made about a VCF’s eliminated observations where founder-CEO replace- experience with emerging industries, more gener- ments occurred more than 60 days before the close ally. Over the last 40 years, VCFs have participated of the fi rst investment round in which the VCF par- in the genesis and development of other new and ticipated. We used a 60-day buffer because it is unfamiliar industries, including semiconductors, possible that replacing the founder-CEO could have personal computing, and biotechnology. Each of been a condition for the VCF’s investment.6 Our these industries offered substantial uncertainties fi nal sample yielded 340 VCF/IPO observations, regarding the nature of the technology, its market refl ecting the participation of 114 venture capital potential, and the business models necessary to fi rms in 193 IPOs. The number of IPOs a VCF par- realize the new technology’s value. Since older ticipated in ranged from two to 23. Except where VCFs are more likely to have been involved at the otherwise noted, the data used to create the variables beginning of a variety of emerging industries and, described below were obtained from the IPO fi rms’ therefore, have had more experience and opportuni- offering prospectuses. ties to develop routines for dealing with the types of uncertainties they are likely to face in the current Dependent variable context (Nelson and Winter, 1982), they may per- ceive less uncertainty in dealing with Internet com- Founder-CEO retention. This was a dummy vari- panies than newer venture capital fi rms that have not able coded 1 if a company’s CEO at the time of its had similar experiences. Recent research suggests IPO was a founder and 0 otherwise, even if the that a large number of new venture capital fi rms founder continued to hold another executive position were founded during the 1990s as investment capital within the company and/or serve on the board of poured into the industry and the IPO market heated directors. We cannot directly measure whether up (Lee and Wahal, 2004). Thus, to the extent that founder-CEO replacements were initiated by VCFs. a fi rm is older and has developed the experience and However both the practitioner (e.g., Gordon, 2002; routines necessary to invest in and develop fi rms in Levensohn, 2006) and academic literatures (e.g., emerging industries, the less likely perceived indus- Boeker and Karichalil, 2002) suggest that founder- try uncertainty is to result in replacing the founder- CEO replacements at highly promising ventures are CEO. Therefore, we hypothesize: generally instigated by VCFs and outside board members. Hypothesis 3: The experience of the VCF with taking IPO fi rms public will reduce the negative Independent variables effects of Internet industry subsegment uncertainty on founder-CEO retention. Subindustry categories. Following Pollock and Gulati (2007) we initially created dummy variables Hypothesis 4: The age of the VCF will reduce the for each of the following Internet subindustry negative effects of Internet industry subsegment classifi cations: uncertainty on founder-CEO retention. 1. e-Infrastructure: Data, voice and video carriers, Web hosts, hardware suppliers DATA AND METHODS 2. e-Services/solutions: Consultants, software/ applications, back offi ce services We collected data on 389 U.S. Internet-related fi rms 3. e-Advertising/marketing/media: Internet that went public between 1995 and 2000. Our unit advertising and research of analysis was the VCF/IPO combination. Of our 4. e-Retail: Consumer products and services original sample, 320 companies received venture 5. e-Finance: Banks, brokers, and credit fi nancing. We eliminated fi rms where no founder companies data were available—because all founders had left 6. e-New media: Advertising/subscription- the organization prior to registering for the IPO— supported communities and all observations where VCFs participated in only one IPO (since VCFs needed to participate in at least two IPOs for us to be able to calculate 6 We also experimented with a 30-day buffer. The results were our lagged and cumulative variables). We also substantively the same.

7. e-Internet service providers: Toll-supported research to operationalize multidimensional con- access providers structs (Boyd, Gove, and Hitt, 2005). 8. Infomediaries: Firms that act as liaisons Number of prior Internet-industry deals. This between buyers and sellers and do not carry measure equaled the total number of Internet IPOs a signifi cant amount of inventory; we further in which the VCF had participated prior to the restrict the defi nition to include only fi rms that current IPO. have a consumer on at least one side of the VCF age. Consistent with prior research (e.g., deal (i.e., business-to-consumer and consumer- Gompers, 1996; Lee and Wahal, 2004), this measure to-consumer middlemen) was calculated as the difference between the year the 9. Business-to-business: Firms involved in busi- VCF raised its fi rst fund and the year of the IPO. ness-to-business e-commerce Because the effects of VCF age are likely to be nonlinear (e.g., the difference between a one-year- We then collapsed these dummy variables into three old VCF and a six-year-old VCF is likely to be broad categories: (1) B2B companies (subindustry greater than between a 35-year-old VCF and a categories 2, 3, and 9) that had businesses on both 40-year-old VCF), and VC fi rms in our sample sides of business transactions; (2) B2C companies range from zero to 54 years in age at the time of the (subindustry categories 4, 5, 6, and 8) that had con- IPOs, we transformed this measure into its natural sumers on at least one side of the transaction; and logarithm.7 Data on the year each VCF raised its (3) infrastructure companies (subindustry categories fi rst fund were obtained from the VentureXpert 1 and 7) that provided Internet access and other database. infrastructure products and services. Using these broader classifi cations captures the major differ- Control Variables ences within the industry and avoids problems with increased collinearity and certain industry categories Perceived market uncertainty. We operationalized perfectly predicting the outcome. We created dummy VCFs’ perceived market uncertainty as the absolute variables, coded 1 if a fi rm was in the B2B or B2C value of the amount of underpricing a VCF experi- subindustry segment, respectively, and 0 otherwise. enced on its prior Internet-sector IPO. Underpricing, We treated infrastructure companies as the excluded defi ned as the percentage change in stock price category. between the initial price set for an IPO fi rm’s shares and the closing price at the end of the fi rst day of trading, is one of the most studied phenomena within Moderators the extensive fi nance literature on IPOs (Benveniste Founder-CEO human capital index. This measure and Spindt, 1989; Loughran and Ritter, 2004; Ritter was an index comprised of three dimensions of a and Welch, 2002; Rock, 1986). Although average founder-CEO’s human capital: whether or not the levels of underpricing have varied during different founder-CEO had previous experience in a start-up, eras (Loughran and Ritter, 2004), there is generally whether or not the founder-CEO had previous expe- a signifi cant jump in price the day a company goes rience as a CEO, and whether or not the founder- public. The larger this jump, the more the fi rm is said CEO possessed an MBA. Thus, the index ranged in to have been underpriced relative to its market value from zero to three. In developing the index we value. also considered founder-CEO age, but unlike the Although a variety of factors have been identifi ed other three measures founder-CEO age did not have over the last 30 years that infl uence underpricing a signifi cant main or moderating effect when used (see Loughran and Ritter, 2004 and Ritter and Welch, in the models individually, so we excluded it from 2002 for reviews), a common assumption underly- our analysis. Employing an index of these three ing all of these studies is that underpricing refl ects dimensions was a parsimonious way to test our investor uncertainty about a fi rm’s likely future per- second hypothesis. Further, combining these three formance. These uncertainties can be driven by lack indicators into single index avoided collinearity of information about specifi c aspects of the company problems associated with the high correlation between prior CEO and prior start-up experience. Our approach was also consistent with recent calls 7 Because there were zero values, a 1 was added to all observa- for greater use of indices in strategy and governance tions before transforming the measure.

Copyright © 2009 Strategic Management Society Strat. Entrepreneurship J., 3: 199–217 (2009) DOI: 10.1002/sej Founder-CEO Retention 207 and also by more general concerns about the indus- (the amount of underpricing the stock experiences) try in which the fi rm competes and the general from this measure since it is captured as a separate market conditions it faces (Gulati and Higgins, independent variable. The data used to calculate this 2003). On average, high levels of underpricing rela- measure were collected from the Center for Research tive to the norms of the period refl ects greater dis- on Securities Pricing (CRSP) database. sensus and higher levels of perceived uncertainty VCF ownership power. We measured VCF own- about the value and future potential of the compa- ership power as the pre-IPO percentage of the fi rm’s ny’s stock; the less the perceived uncertainty, the equity owned by the focal VCF.8 Overall, VCF own- greater the consensus regarding the value of the ership in a fi rm may be concentrated in the hands of company and the lower the underpricing the IPO one or two VCFs, but is often spread across several experiences. different VCFs (Lerner, 1995). Prior research We are unable to measure VCFs’ perceived suggests that ownership concentration may affect uncertainly directly. However, given the rapidly governance in important ways, and it is often argued changing conditions during the late-1990s, we that more concentrated ownership adds to a VCF’s expect that VCFs would treat the level of underpric- power (Boeker and Karichalil, 2002; Flamholtz and ing experienced on their most recent Internet IPO Randle, 2000). as an important source of timely and relevant infor- Time since last Internet IPO. The time that elapsed mation about the market uncertainties present in between a VCF’s prior Internet IPO and the current this new industry. Recent experience with under- IPO can infl uence its ability to infer meaning from pricing, therefore, serves as a useful proxy for levels its experience and make behavioral adjustments that of perceived market uncertainty that a VCF faces. can infl uence the current IPO (Hayward, 2002). This This proxy is likely to have been particularly salient measure was operationalized as the number of days given the many sources of uncertainty that VCFs between the previous Internet IPO in which the VCF faced during the Internet bubble. Indeed, underpric- participated and the current IPO. IPO dates were ing averaged approximately 7.4 percent during the obtained from the SDC New Issues database. 1980s and 12 percent during the early-1990s, but Total number of VCFs invested in the IPO. This averaged 65 percent in 1999 and 2000. During these variable equaled the total number of VCFs that two years, underpricing varied widely, demonstrat- owned equity in an IPO fi rm at the time it went ing the substantial levels of uncertainty during this public. More than one VCF is usually involved in period. Although the vast majority of IPOs experi- the funding of IPO fi rms (the average in our sample ence positive changes in market value on the fi rst was 3.35 and ranged from one to seven). The number day of trading, some IPOs experience declines in of VCFs involved in an IPO can affect the infl uence price, which also refl ects dissensus and higher levels of any one VCF on the founder-CEO retention of perceived uncertainty. Thus, we used the decision. absolute value of underpricing (scaled as a decimal California dummy. This location variable was percentage) to capture the general amount of included to account for the fact that Internet fi rms uncertainty. located in California may have been geographically VCF’s prior Internet-sector investment perfor- closer to many of the VCFs that specialize in tech- mance. This measure was operationalized as the nology companies and, thus, may have been subject average six-month post-IPO performance of all to greater supervision and VCF involvement in the Internet-sector IPOs in which the VCF participated company’s affairs. prior to the current IPO. We chose the six-month Bubble year dummy. The Internet market bubble time period because stock ownership lockup provi- hit its peak in 1999 and 2000 and burst in the latter sions that restrict a VCF’s sale of its stock typically half of 2000. Thus, we coded this measure equal to expire after six months of trading (Field and Hanka, 1 if a company went public in 1999 or 2000 and 0 2001; Levin, 2001). Thus, the six-month time frame otherwise to control for a variety of factors that may be of particular relevance to venture capitalists. This measure was calculated as the percentage change in stock price between the price at the end of the fi rst day of trading and the price of the stock 8 We also considered the focal VCF’s ownership as a percentage of the total portion of the fi rm’s equity owned by all VCFs at the close of the 180th day of trading. We removed and the number of board seats a VCF controlled, but these the change in stock price on the fi rst day of trading measures were not signifi cant in any of our models.

Copyright © 2009 Strategic Management Society Strat. Entrepreneurship J., 3: 199–217 (2009) DOI: 10.1002/sej 208 T. G. Pollock, B. R. Fund, and T. Baker could be associated with the year in which the fi nancing after 20009, and were still private at the end company went public. of 2000. We combined these companies with the 193 CEO ownership. This variable equaled the per- IPO fi rms included in our sample on which Ventur- centage of a company’s outstanding shares owned eXpert also had data. We collected the following vari- by the CEO prior to the IPO. This variable was ables for each fi rm: (1) the number of rounds of included to control for the CEO’s ownership power. venture fi nancing received; (2) the total venture Prior research suggests that the greater the retained funding raised; (3) the number of VCFs invested in ownership, the less likely a founder is to be replaced the company; (4) its subindustry category within the as CEO (Boeker and Karichalil, 2002). Internet sector; and (5) the year founded, which was IPO fi rm age. Firm age at IPO was calculated as used to create two founding dummy variables (one the months since the fi rm’s incorporation date. Older indicated the fi rm was founded before 1995 and the fi rms may have gone through more rounds of pre- other indicated the fi rm was founded between 1995 IPO fi nancing, putting the founder-CEO at greater and 1997). Firms founded after 1997 were treated as risk of being replaced (Wasserman, 2003). Because the excluded category. These variables were used in fi rms ranged in age from three to 217 months in our a probit regression to predict whether or not a company sample, this variable was log transformed to reduce went public during our period of study. The overall the effect of extreme values on the analysis. model was highly predictive, with all three VCF- Offer value. This measure equaled the total number related variables and two of the industry dummies of shares offered during the IPO multiplied by the signifi cantly predicting the likelihood a fi rm would offering price. The size of the offering can send go public. This regression was then used to create signals to the market about the relative quality and the selectivity instrument that was included in our stability of the offering and is frequently used as a logistic regressions predicting founder-CEO retention control by fi nance scholars conducting IPO-related (Higgins and Gulati, 2003). research (Ibbotson and Ritter, 1995). Because the offer values in our sample ranged from $14.4 million Method of analysis to $375 million, this variable was log transformed Our hypotheses were tested using logistic regres- to reduce the effect of extreme values. sions. Because the same VCFs were included in the Board size. This measure equaled the number of sample multiple times, the observations are not inde- individuals listed in the offering prospectus who are pendent. In order to correct for this, we calculated identifi ed as members of the board of directors. This standard errors using the robust and cluster com- measure was used to control for the fact that the mands in Stata 9.0. These commands calculated larger the board, the more diffi cult it is for the CEO Huber–White–Sandwich estimates for the standard or any one director to dominate its decision making errors and correct for nonindependence of observa- (Pearce and Zahra, 1992). tions within VC clusters.10 Prone to IPO. In addition to funding IPO fi rms, VCFs also funded fi rms that did not go public during our period of study. To the extent these companies RESULTS did not go public for reasons associated with the retention or replacement of the founder-CEO, this Table 1 presents descriptive statistics and correla- could result in an over- or under-representation tions for all the variables used in our analysis. To of founder-CEO retentions among venture-backed companies that ultimately went public, creating a potential source of bias in our sample. In order to 9 Because we are controlling for the probability that a fi rm went address this issue, we followed previous research public during our period of study, which ended in 2000, it is important to make sure the data only refl ects VCF fi nancing (e.g., Higgins and Gulati, 2003; Stuart, Hoang, and activities during the period a fi rm is at risk in our analysis. Hybels, 1999) and created an instrumental variable Since VentureXpert only reports this information cumulatively that captured the likelihood a company would go (i.e., as of the date the data were collected), we had to exclude fi rms whose data refl ected fi nancing activities that occurred public between 1995 and 2000. after 2000, as we were unable to separate pre- and post-2000 First, we identifi ed 222 companies listed in the activities. VentureXpert database that were in the Internet sector, 10 In analyses not reported here we also clustered based on company, rather than VC fi rm, since some IPO fi rms are also received at least one round of venture capital fi nanc- represented in our sample multiple times. The results were the ing from 1995 to 1999, did not receive any venture same as those presented here.

ease interpretation, the means and standard devia- 0.70

− tions reported in Table 1 were calculated using untransformed variables.

0.06 1.00 Table 2 presents the results of logistic regressions − predicting retention of the founder-CEO. In logistic 0.04 0.04 0.02 1.00 regression, the unstandardized regression coefﬁ - − − cients represent the logarithm of the odds of the focal event occurring (i.e., the odds ratio, which 0.01 − equals the probability of the event occurring divided by the probability the event will not occur). To ease 0.02 0.01 0.07 0.02 − − interpretation in Table 2, we exponentiated the coef- ﬁ cients and present the odds ratios. Thus, the values 0.08 0.05 0.00 0.05 0.05 0.01 0.31 1.00 − − − in our regression table represent the multiplicative impact of the predictor on the probability of the focal 0.04 0.03 0.01 1.00 event—in our models, retention of the founder- − − − CEO—occurring divided by the probability the 0.01 0.04 0.04 0.09 0.11 1.00 0.03 1.00 event will not occur. An odds ratio above 1 indicates − − − that a one-unit increase in the independent variable increases the odds of retention, while an odds ratio 0.23 0.11 0.09 0.03 0.06 0.00 0.09 0.24 0.33 1.00 e appropriate. − − − − below 1 indicates that a one-unit increase in the independent variable reduces the odds of retention. 0.11 0.01 0.20 0.19 0.06 0.08 0.01 0.15 0.18 − − − − − In order to reduce nonessential collinearity in our models (Cohen et al., 2003), the variables used in 0.09 0.06 0.12 0.02 0.05 0.33 0.10 0.16 0.26 0.03 0.03 0.03 0.03 0.25 1.00 0.25 0.15 0.23 1.00 0.03 1.00 − − − − − − − our interactions were mean centered. Collinearity diagnostics show that while the overall condition 0.05 0.07 0.08 0.08 0.06 0.07 0.35 0.00 0.29 0.01 0.03 1.00 index is relatively high (approximately 23–24) for − − − − − − − − our models, it is still well below the recommended 0.02 0.15 0.08 0.110.05 1.00 0.06 0.12 cutoff level of 30 (Belsley, Kuhn, and Welsch, 1980) − − − − − − in all but one instance, which is when the interactions between founder-CEO human capital and the 0.06 0.02 0.05 0.13 0.03 0.03 0.05 1.00 0.27 0.04 − − − − − − B2B and B2C dummy variables are included in the model at the same time (condition index = 35). 0.01.

0.06 0.21 0.05 0.01 0.08 0.00 0.10 0.09 < − − − − − − Therefore, we present each of our interactions testing the hypotheses separately.11 Model 1 in Table 2 0.12 0.01 0.09 0.08 0.07 0.12 0.01 0.01 0.10 0.18 0.07 0.04 0.00 0.13 0.15 0.03 0.09 0.26 0.21 0.20 0.31 0.29 1.00 0.03 0.07 0.05 cant at p − − − − − − − − − includes the control variables. Model 2 adds the

main effects of B2B and B2C and tests Hypothesis 0.04 0.12 0.01 0.21 0.14 0.10 0.06 0.08 0.03 0.02 0.02 0.13 0.01 0.14 0.01 0.13 1. Models 3 and 4 add the interactions testing − − − − − − − − − 0.12 signiﬁ

Hypothesis 2, Models 5 and 6 add the interactions > testing Hypothesis 3, and Models 7 and 8 add the interactions testing Hypothesis 4. Model 1 reveals that a number of our control variables have signifi cant relationships with the like- 0.05, correlations lihood of CEO retention and that the overall model < fi t is very good. In particular, IPO fi rm age, the number of VCs invested in the company, the per- cant at p centage of stock owned by the founder-CEO, and 0.09 signifi

> 11 The pattern of interactions for prior deals and VCF age are rm age 2 51.09 32.21 0.00 1.00

the same when the respective interactions between these vari- g. prior post-IPO perf. 12 165.12 170.64 ble 1. Descriptive statistics and correlations riable ID Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 tal VCFs 6 3.39 1.49 0.07 0.06 0.01ables 0.05 and the B2B and B2C dummies are included in the me since last IPO 7 194.10 277.77 B2C 18 0.28 0.45 B2B 17 0.56 0.50 0.01 0.22 Founder-CEO HC indexFounder-CEO 16 1.82 0.91 0.57 VCF age 15 16.46 9.99 0.00 0.10 Number of prior deals 14 4.14 3.94 Abs val. prior underpricing 13 0.94 1.00 Av All means and standard deviations are based on untransformed measures; correlations log-transformed measures, wher VCF stock ownership 11 15.25 9.94 CEO stock ownership 10 11.41 11.93 0.47 Selection instrument 9 0.78 0.23 Ta Va retentionFounder-CEO IPO ﬁ 1 0.60 0.49 1.00 Correlations Offer valueOffer 3 82.04 54.45 0.00 Bubble years dummy dummyCA To 4 0.90 0.30 5 0.03 0.48 0.18 0.23 0.50 1.00 Ti Board size 8 6.74 1.57 models together.

Table 2. Predicting founder-CEO retention

Variable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8

Ln ﬁ rm age 2.584* 3.183** 3.701* 3.709* 3.724* 3.740* 3.184** 3.315* (1.088) (1.427) (1.954) (2.041) (1.950) (1.944) (1.409) (1.569) Ln offer value 2.139† 1.317 1.427 1.628 1.523 1.702 1.404 1.569 (0.931) (0.652) (0.684) (0.802) (0.794) (1.004) (0.658) (0.743) Bubble years dummy 0.292† 0.428 0.272* 0.385 0.411 0.357 0.451 0.475 (0.214) (0.290) (0.178) (0.266) (0.272) (0.244) (0.298) (0.320) CA dummy 0.304* 0.372† 0.311* 0.327† 0.327† 0.353† 0.406† 0.391† (0.176) (0.213) (0.181) (0.200) (0.188) (0.194) (0.222) (0.215) Total VCFs 1.303† 1.313† 1.379† 1.426* 1.268 1.272 1.366* 1.390* (0.190) (0.216) (0.228) (0.246) (0.212) (0.213) (0.207) (0.215) Time since last IPO 0.998* 0.998* 0.998* 0.998* 0.998* 0.998* 0.998* 0.998** (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) Board size 0.950 1.078 1.026 1.026 1.090 1.087 1.051 1.063 (0.133) (0.135) (0.127) (0.134) (0.151) (0.146) (0.136) (0.135) Selection instrument 14.308* 9.970* 9.451* 8.904* 13.866** 13.918** 9.357* 9.909* (15.502) (9.451) (9.234) (9.494) (13.406) (13.346) (8.990) (9.076) CEO stock ownership 1.715** 1.885** 1.914** 1.912** 1.949** 1.930** 1.912** 1.929** (0.168) (0.231) (0.246) (0.244) (0.270) (0.261) (0.246) (0.244) VCF stock ownership 1.021 1.035* 1.035† 1.037† 1.033† 1.035* 1.035* 1.030† (0.018) (0.018) (0.019) (0.019) (0.018) (0.017) (0.018) (0.017) Avg. prior post-IPO perf. 0.999 0.998 0.999 0.999 0.998 0.999 0.998 0.999 (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) (0.001) Abs. val. prior underpricing 0.634† 0.685 0.650 0.635† 0.741 0.748 0.681 0.649† (0.161) (0.170) (0.174) (0.168) (0.189) (0.195) (0.166) (0.153) Number of prior deals 0.891** 0.879** 0.884** 0.879* 0.793** 0.946 0.882* 0.884* (0.038) (0.044) (0.042) (0.046) (0.060) (0.045) (0.043) (0.047) Ln VCF age 1.548 1.665 1.777 1.863† 1.670 1.617 0.853 2.406** (0.566) (0.635) (0.655) (0.687) (0.655) (0.613) (0.589) (0.792) Founder-CEO HC index 7.232** 8.619** 5.093** 18.922** 9.248** 8.839** 9.524** 9.973** (2.275) (3.099) (2.213) (8.566) (3.473) (3.255) (3.711) (4.132) B2B 0.115** 0.087** 0.101** 0.094** 0.119** 0.098** 0.116** (0.052) (0.042) (0.050) (0.049) (0.055) (0.056) (0.054) B2C 0.194** 0.227** 0.335* 0.195** 0.255* 0.199** 0.261* (0.108) (0.119) (0.162) (0.114) (0.164) (0.110) (0.137) B2B × founder-CEO HC 3.276* index (1.985) B2C × founder-CEO HC 0.160** index (0.092) B2B × prior deals 1.218* (0.105) B2C × prior deals 0.815† (0.096) B2B × VCF age 3.056 (2.268) B2C × VCF age 0.170† (0.162) Log-likelihood −84.119 −78.838 −77.162 −74.756 −76.925 −77.136 −77.369 −76.381 Wald chi-square 83.75 85.66 76.073 75.91 87.64 87.24 80.39 77.88 Pseudo-R2 0.6330 0.6561 0.6634 0.6739 0.6644 0.6635 0.6625 0.6668 Observations 340 340 340 340 340 340 340 340

Effects reported as odds ratios. Robust standard errors in parentheses. †signifi cant at 10%; *signifi cant at 5%; **signifi cant at 1%.

Copyright © 2009 Strategic Management Society Strat. Entrepreneurship J., 3: 199–217 (2009) DOI: 10.1002/sej Founder-CEO Retention 211 the founder-CEO’s human capital are all signifi cant that when this interaction is included in the model, and increase the likelihood the founder-CEO will be the main effect odds ratio of B2C remained signifi - retained. In addition, once the industry measures are cant and was less than 1 and the interaction odds added to the model, VCF percentage stock owner- ratio was signifi cant; however, it was also less than ship also becomes signifi cant and has a positive 1 (OR = 0.16). Interpretation of the interaction using effect on founder-CEO retention. On the other hand, lincom suggests that a one-unit increase in the going public in 1999 or 2000, being located in Cali- founder-CEO human capital index resulted in an fornia, the amount of time since the VCF last took odds ratio of 0.053 (p < 0.001); that is, the odds ratio an Internet company public, and the number of prior was approximately cut in half by a one-unit increase Internet-company IPOs the VCF has done are all in founder-CEO human capital index. signifi cant and decrease the likelihood a founder- Hypothesis 3 suggested that the number of prior CEO will be retained. Internet IPO fi rms a VCF had taken public would Hypothesis 1 suggested that greater levels of reduce the effects of Internet industry subsegment industry uncertainty would decrease the likelihood a uncertainty on the likelihood of founder-CEO reten- founder-CEO would be retained. This hypothesis tion. The results in Models 5 and 6 provide mixed was supported. Both B2B and B2C had signifi cant support for this hypothesis. Consistent with the effects with odds ratios less than 1 in Models 2 to 8. hypothesis, the B2B × prior Internet IPOs interaction Model 2 indicates that the odds ratio for being in the odds ratio was signifi cant and resulted in an odds B2B industry subsegment was 0.115, and the odds ratio of 0.115 (p < 0.001), suggesting that each addi- ratio for the B2C industry subsegment was 0.194, tional prior IPO reduced the negative effect of B2B both in comparison to fi rms in the infrastructure by approximately 21 percent. However, the B2C × segment. We employed the lincom command in prior Internet IPOs interaction had a signifi cant odds STATA to determine whether the difference between ratio less than 1, and the comparison odds ratio was being in the B2B and B2C industry subsegments was 0.208. Thus, a one-unit change in the number of signifi cant; results of this analysis showed no statis- prior deals increased the negative effect of B2C on tically signifi cant difference between the effects of retention by approximately 19 percent. Therefore, B2B and B2C. Hypothesis 3 was supported for B2B fi rms, but not Hypothesis 2 suggested that higher levels of for B2C fi rms. founder-CEO human capital would reduce the nega- Hypothesis 4 suggested that VCF age would tive effects of industry uncertainty on the likelihood reduce the effects of industry uncertainty on the of retention. This hypothesis was partially supported. likelihood of founder-CEO retention. This hypothe- When the B2B × founder-CEO human capital index sis was not supported. The B2B × VCF age interac- interaction was added in Model 3 the main effect tion odds ratio was greater than 1, but it was not odds ratio remained signifi cant and less than 1, while signifi cant. The B2C × VCF age interaction odds the interaction term odds ratio was signifi cant and ratio was signifi cant, but it was less than 1. Because greater than 1. Interpretation of the interaction odds the moderator is a continuous measure, in order to ratio indicates that a one-unit increase in the founder- interpret the interaction we had to assign a value to CEO human capital index reduced the negative VC age. Using the mean value for VC age in our effect of B2B on the odds of retention by 3.276 sample, the resulting comparative odds ratio was times. To interpret the effect of CEO characteristics 0.002 (p < 0.02), suggesting an average-aged VC on the perceived uncertainty associated with B2B, fi rm further reduced the odds ratio to nearly 0. we employed the lincom command in STATA to calculate the odds ratio for B2B when the founder- CEO human capital index was equal to 1 versus DISCUSSION when it was equal to 0. The resulting odds ratio (0.287) was signifi cant at p < 0.07. Thus, although In this study, we explored whether and how owners’ the overall effects of B2B on the likelihood of reten- perceptions of uncertainty in a novel industry context tion was still negative, the odds ratio was roughly infl uenced their decisions to retain or replace tripled by a one-unit increase in the CEO index. A founder-CEOs. We developed and tested theoretical different pattern of results was obtained, however, arguments that—in the context of a highly uncertain when the founder-CEO human capital index was and rapidly changing new industry—venture capital interacted with the B2C indicator. Model 4 shows fi rms’ decisions to retain founder-CEOs would be

Copyright © 2009 Strategic Management Society Strat. Entrepreneurship J., 3: 199–217 (2009) DOI: 10.1002/sej 212 T. G. Pollock, B. R. Fund, and T. Baker infl uenced by their perceived uncertainty about challenge to the common assumption that owners the industry subsegment and that the effects of this somehow know whether and how to effectively use perceived uncertainty would be moderated by the their ability to replace fi rms’ leaders and suggests founder-CEO’s experience and education and by the that—especially in the face of high levels of VCF’s relevant prior experience in the Internet perceived novelty—important governance decisions industry and in prior emerging industries. may be characterized fundamentally as attempts to Our results showed that VCFs adapted their deci- reduce perceived uncertainty. sions based on perceived industry sector uncertainty. Further, our study contributes to the literature sug- Founder-CEOs in Internet fi rms providing infra- gesting that venture capital fi rms may make key structure products and services were more likely to decisions for symbolic reasons (e.g., Gompers, 1996; keep their jobs than were founder-CEOs in the more Lee and Wahal, 2004). For example, scholars have uncertain B2B and B2C sectors. We did not theorize suggested that newer VCFs will grandstand by specifi c differences between B2B and B2C, but the taking companies public earlier and with more results of our moderating hypotheses showed dis- underpricing to signal their capabilities to current tinctly different patterns. For B2B, our fi rst two and potential investors in their venture funds, thus moderating hypotheses were supported. Founder- enhancing their ability to raise larger venture funds CEO and prior VCF experience successfully taking in the future. Our fi ndings are consistent with the Internet companies public both reduced the likeli- idea that VCFs may also replace founder-CEOs hood that industry uncertainty would lead to founder- for at least partially symbolic reasons, since such CEO replacement. The VCF’s prior experience in replacements can be a means of managing investors’ other industries had no signifi cant effect. For the perceptions of uncertainty even if they have no sub- B2C sector, all three moderators increased the likeli- stantive impact on the fi rm’s performance. Thus, our hood that the founder-CEO would be replaced, theory and fi ndings may begin to provide an expla- which is the opposite of what we hypothesized. nation why the literature on the effects of founder- Below, we speculate about possible explanations for CEO replacement on fi rm performance has found the distinctive patterns of fi ndings between the B2B such mixed fi ndings. Future research should con- and B2C sectors. We also discuss an interesting tinue to explore the symbolic, as well as substantive, pattern we had not hypothesized but noted for one motivations that may underlie VCFs’ governance, of our control variables. investment, and other activities, and how the factors motivating a founder-CEO retention or replacement decision relate to subsequent fi rm performance. Theoretical contributions Our study also contributes to the literature on The results support our most general argument that corporate governance more generally. Our primary uncertainty plays an important role in shaping what contribution here is to begin developing and testing may be the most important governance decision theory that does not assume governance skills are facing young fi rms. Our fi ndings suggest that VCFs automatic, general, or stable. For the important and display a level of subtlety and nuance in their gov- perhaps increasingly prevalent condition of environ- ernance decisions, fi tting founder retention and mental uncertainty and change, our results demon- replacement to levels of subsector uncertainty, their strate how owners are likely to adjust their governance own relevant industry experience, and to the charac- decisions based on perceptions of environmental teristics of the founder-CEO. But they also suggest uncertainty, and also suggest that prior experience that in the face of high levels of perceived uncer- can alter perceived uncertainty and shape owners’ tainty, important governance decisions may be a interpretations in ways that lead to different gover- little bit blunter, and aimed simply at reducing per- nance decisions. Thus, in unstable and highly uncer- ceived uncertainty through whatever tools may be tain environments, we should not expect to observe available. Further, at very high levels of perceived the rapid emergence of a consistent pattern of uncertainty or with particular sorts of uncertainty— strategic governance actions. Instead, idiosyncratic as we suggest below may have characterized the patterns are likely to emerge, creating the possibility B2C sector—VCFs may have been still trying to for substantial industry-level trial and error learning fi gure out how to govern their portfolio fi rms rather (Miner and Haunschild, 1995). than simply or confi dently applying approaches We are intrigued by our fi ndings of support that had worked before. This provides an additional for most of our moderating hypotheses for fi rms

Copyright © 2009 Strategic Management Society Strat. Entrepreneurship J., 3: 199–217 (2009) DOI: 10.1002/sej Founder-CEO Retention 213 providing products and services to other businesses individual VCF ownership control variable is also and the contradiction to most of our moderating an indirect measure of how much a VCF has chosen hypotheses for fi rms providing products and services to invest in a particular fi rm and, therefore, may to consumers. We had noted earlier that the levels indirectly indicate how much they liked the manage- and sources of uncertainty might differ between ment team, including the founder-CEO. If the VCF these sectors, but we did not hypothesize and are is betting heavily on the jockey, then it stands to unable to fully explain the observed differences reason it may use its power to keep this individual between them. Given the pattern of results, we spec- in place. ulate that the notion everything is different or blown This interpretation is consistent with a folk theory to bits that was rampant in popular characterizations related to us by one of the venture capitalists we of the early Internet industry as a whole may have interviewed informally. He claimed that when a had more relevance for VCFs when investing in B2C VCF has a high ownership stake there is a ten- fi rms than when investing in other sectors. While dency—in the VC’s words—to ‘go native’ and infrastructure and B2B business models were often identify more heavily with the founder. Similarly, as directly analogous to those of more traditional fi rms, a general partner from another VCF put it, ‘Once B2C business models were often characterized by you’ve really gotten into a fi rm (fi nancially) . . . you diffi culty answering fundamental questions, such as may have fallen in love with the CEO.’ Perhaps how to turn eyeballs (the number of people esti- when VCFs have made large investments in fi rms mated to see a Web page) into cash fl ow, and other they believe to be endowed with good management, fundamental issues about how to monetize Web site they tend to exhibit a degree of consistency and activity. We speculate that in the face of very novel commitment (Cialdini, 2004), using their power to and uncertain B2C business models, VCFs may have encourage retention. Beyond such speculation and discounted the ability of traditional indictors of VC folk theories, our fi ndings regarding VCF own- human capital to predict success, and that the more ership and power point to the usefulness of further they knew or learned about the uncertainty of this research exploring the effects of the relationship sector, the less confi dent they became in the wisdom between structural sources of power and owners’ of retaining founder-CEOs. As these interpretations perceptions on governance. are purely speculative, future research should continue to explore this issue. Practical implications Finally, we were struck by the consistent pattern of results for our control variable measuring the We believe that our study also has signifi cant impli- VCF’s pre-IPO percentage ownership of the fi rm. cations for practitioners. Our overall models were We included this measure to control for differences highly predictive of the founder-CEO retention deci- among VCFs in the ability to assert preferences sion; thus, we believe we have captured many of the regarding whether to retain a founder-CEO. Although key factors that infl uence this decision. Some of we had not developed a hypothesis in this regard, these, such as founder-CEO stock ownership, are we expected that this proxy for VCF power might fairly obvious. But others, such as our theoretical be associated with a greater likelihood of replace- variables, offer an interesting and nuanced set of ment, as powerful owners took action to make factors entrepreneurs should consider when pursu- governance changes. Instead, our results suggest ing venture capital investments. Our fi ndings show that this proxy for individual VCF power is as- that VCFs’ industry familiarity can have a signi- sociated with a greater likelihood of founder-CEO fi cant effect on whether or not founders will be retention. retained as CEOs, and that their own personal quali- To the extent that future research supports our fi cations and experience—as well as the age and somewhat counterintuitive fi nding that greater VCF experience of the venture capital fi rm—can dramati- power can increase the likelihood of founder-CEO cally reduce or enhance these effects. To the extent retention, it may point to the importance of the rela- entrepreneurs have the option to choose among dif- tionship between power and an owner’s prior invest- ferent venture capital fi rms, they may want to con- ment decisions. The perceived quality of the sider these characteristics as additional criteria when management team is often described as an important deciding whose money to accept. Further, the driver of VCF investment decisions (Zacharakis and fi ndings for VCF ownership control suggest that Meyer, 1998; Zacharakis and Shepherd, 2005). Our (once he/she has decided to take on venture capital

Copyright © 2009 Strategic Management Society Strat. Entrepreneurship J., 3: 199–217 (2009) DOI: 10.1002/sej 214 T. G. Pollock, B. R. Fund, and T. Baker at all) if an entrepreneur wants to increase the into once he/she steps down as CEO. There may be likelihood he/she will maintain control of his/her signifi cant differences associated with whether the company, higher levels of ownership by individual founder is retained in another operational role, VCFs may improve the odds. remains involved in the company in a nonopera- tional role—such as director or chairman of the board—or ceases involvement with the company Future research directions altogether. Future research should explore how Like all studies, this one has limitations that provide different factors infl uence the specifi c forms that opportunities for further research. One set of future founder-CEO replacements take and the conse- research directions arises from the fact that we do quences they have for the fi rm. not directly ascertain VCFs’ perceptions of uncer- A fi nal set of opportunities arises from the fact tainty, but rather infer them based on more indirect that we cannot determine with certainty whether external indicators we expect will infl uence their a VCF initiated replacement or the founder-CEO perceptions. Although our measures may be some- stepped aside voluntarily. While we recognize that what crude proxies, our results suggest the assump- founders may and do voluntarily step down as tion we make appears to be reasonable. Nonetheless, CEOs, there are many reasons to expect that this future research conducted in real time as VCFs make does not occur in the majority of cases. Both the these decisions is necessary to further validate our empirical and practitioner literatures indicate that assumption and to continue exploring the important founders generally prefer to stay at the helms of the issue of how VCF perceptions infl uence founder- companies they started, and that it often takes a CEO retention and replacement. board vote to remove them (e.g., Boeker and A second future research direction results from Karichalil, 2002; Gordon, 2002; Levensohn, 2006: the fact that we are unable to capture a VCF’s prior Wasserman, 2003). Indeed, one of the most signifi - experience in other industry contexts. However, it cant indicators that a founder is likely to remain may be reasonable to assume that a VCF’s initial CEO of his/her company—in our study and others— behaviors will tend to refl ect their prior experience. is the percentage of stock ownership he/she retains, Further, the extent to which there are other, unmea- which gives him/her some power to resist replace- sured factors infl uencing the founder-CEO retention ment. If he/she were more likely than not to step decision serves to make ours a conservative test of aside voluntarily, then this would not be the case, our hypotheses. A related limitation is that in situa- and we would also be unlikely to observe the VCF- tions where there are multiple VCFs invested in a based effects we fi nd in our study. Further, the company, we have no way of knowing for sure extent to which there are instances in our sample which VCF was most responsible for replacing or where the founder-CEOs step aside voluntarily adds retaining a founder-CEO. Although a VCF does not error variance to our models, making ours a conser- have to be the primary decision maker for its experi- vative test of our arguments. Future research, using ence on a deal to infl uence its subsequent actions, more fi ne-grained approaches (such as surveys or our VCF ownership measure, as well as our control qualitative methods) should continue to explore for the number of VCFs invested in the company, this issue. address many of the potential variations in VCF infl uence. Nonetheless, future research should continue to conduct fi ner-grained analyses on the dynamics of VCF syndicates and how they shape ACKNOWLEDGEMENTS key strategic governance decisions. A second related limitation is that the infrastructure industry segment We would like to thank Associate Editor Harry Sapienza that we use as our baseline for comparison may still and two anonymous SEJ reviewers for their many helpful generate relatively high levels of uncertainty, com- suggestions and comments. We would also like to thank Don Hambrick, Dave Harrison, Anne Miner, Yuri pared to other industries. However, this would serve Mishina, Teresa Nelson, Dean Shepherd, and the partici- only to reduce variance between our industry sub- pants in the Organizations Research Group at Penn State segments, again making this a conservative test of University, the RH Smith Entrepreneurship Conference our hypotheses. at the University of Maryland, and the London Business A third opportunity arises from the fact that we School Entrepreneurship Conference for their helpful do not consider the specifi c role a founder moves comments on earlier drafts of this study.

REFERENCES Boyd BK, Gove S, Hitt MA. 2005. Construct measurement in strategic management research: illusion or reality? Adams DD. 2005. A synopsis from shining a light on the Strategic Management Journal 26(3): 239–257. critical role of boards: A survey of boards of private Burton MD, Beckman CM. 2007. Leaving a legacy: posi- equity-backed companies. Available at: http://www. tion imprints and successor turnover in young fi rms. capitalval.com [accessed 21 July 2009]. American Sociological Review 72: 239–266. Aldrich HE. 1999. Organizations Evolving. Sage: London. Carpenter MA, Pollock TG, Leary MM. 2003. Governance, Aldrich HE, Baker T. 2001. Learning and legitimacy: entre- the experience of principals and agents, and global stra- preneurial responses to constraints on the emergence of tegic intent: testing a model of reasoned risk taking. new populations. In The Entrepreneurship Dynamic: Strategic Management Journal 24(9): 803–820. Origins of Entrepreneurship and its Role in Industry Certo ST, Covin JG, Daily CM, Dalton DR. 2001. Wealth Creation and Evolution, Schoonhoven CB, Romanelli and the effects of founder management among IPO-stage E (eds.). Stanford University Press: Stanford, CA; new ventures. Strategic Management Journal 22(6–7): 207–235. 641–658. Baker T, Nelson RE. 2005. Creating something from Certo ST, Daily CM, Cannella AA Jr., Dalton DR. 2003. nothing: resource construction through entrepreneurial Giving money to get money: how CEO stock options bricolage. Administrative Science Quarterly 50: and CEO equity enhance IPO valuations. Academy of 329–366. Management Journal 46(5): 643–653. Baron JN, Burton MD, Hannan MT. 1999. Engineering Chen G, Hambrick DC, Pollock TG. 2008. Puttin’ on the bureaucracy: the genesis of formal policies, positions ritz: pre-IPO enlistment of prestigious affi liates as dead- and structures in high technology fi rms. Journal of Law, line-induced remediation. Academy of Management Economics and Organization 15: 1–41. Journal 51(5): 954–975. Baron JN, Hannan MT, Burton MD. 1999. Building the iron Cialdini RB. 2004. The science of persuasion. Scientifi c cage: determinants of managerial intensity in the early American, Mind, Special Edition; 70–77. years of organizations. American Sociological Review Cohen J, Cohen P, West PW, Aiken LS. 2003. Applied 64: 527–547. Multiple Regression/Correlation Analysis for the Behav- Baron JN, Hannan MT, Burton MD. 2001. Labor pains: ioral Sciences (3rd edn). Lawrence Erlbaum Associates: change in organizational models and employee turnover Mahwah, NJ. in young, high-tech fi rms. American Journal of Sociology Cyert RM, March JG. 1963. A Behavioral Theory of the 106(4): 960–1012. Firm. Prentice-Hall: Englewood-Cliffs, NJ. Beatty RP, Zajac EJ. 1994. Managerial incentives, monitor- Daily CM, Dalton DR. 1992. Financial performance of ing, and risk bearing: a study of executive compensation, founder-managed versus professionally managed corpora- ownership, and board structure in initial public offerings. tions. Journal of Small Business Management 30: 25–34. Administrative Science Quarterly 39: 313–335. De Clercq D, Sapienza HJ. 2005. When do venture capital Bebchuk L, Fried J. 2004. Pay Without Performance: The fi rms learn from their portfolio companies? Entrepre- Unfulfi lled Promise of Executive Compensation. Harvard neurship Theory & Practice 29(4): 517–535. University Press: Cambridge, MA. DiMaggio P, Powell W. 1983. The iron cage revisited: Beckman CM, Burton MD, O’Reilly C. 2007. Early teams: institutional isomorphism and collective rationality in the impact of team demography on VC fi nancing and organizational fi elds. American Sociological Review 48: going public. Journal of Business Venturing 22: 147–160. 147–173. Dimov D, Shepherd DA. 2005. Human capital theory and Belsley DA, Kuhn E, Welsch RE. 1980. Regression Diag- venture capital fi rms: exploring ‘home runs’ and ‘strike nostics: Identifying Infl uential Observations and Sources outs.’ Journal of Business Venturing 20: 1–21. of Collinearity. John Wiley and Sons: New York. Field LC, Hanka G. 2001. The expiration of IPO share Benveniste LM, Spindt PA. 1989. How investment bankers lockups. Journal of Finance 56(2): 471–500. determine the offer price and allocation of new issues. Finkelstein S, Hambrick DC, Cannella AA Jr. 2009. Stra- Journal of Financial Economics 24: 343–361. tegic Leadership: Theory and Research on Executives, Berle AA, Means GC. 1932. The Modern Corporation Top Management Teams, and Boards. Oxford University and Private Property. The Macmillan Company: New Press: New York. York. Fischer HM, Pollock TG. 2004. Effects of social capital and Boeker W. 1992. Power and executive dismissal: scape- power on surviving transformational change: the case of goating at the top. Administrative Science Quarterly initial public offerings. Academy of Management Journal 27: 538–547. 47: 463–481. Boeker W, Karichalil R. 2002. Entrepreneurial transitions: Flamholtz EG, Randle Y. 2000. Growing Pains: Transition- factors infl uencing founder departure. Academy of Man- ing from Entrepreneurship to a Professionally Managed agement Journal 45(3): 818–826. Firm (3rd edn). Jossey-Bass: San Francisco, CA.

Forbes DP, Korsgaard MA, Sapienza HS. Financing deci- Kuemmerle W. 2002. A Test for the Fainthearted. Harvard sions as a source of confl ict in venture boards. Journal Business Review 80(5): 122–127. of Business Venturing. Forthcoming. Lee PM, Wahal S. 2004. Grandstanding, certifi cation and Gompers PA. 1996. Grandstanding in the venture capital underpricing of venture backed IPOs. Journal of Finan- industry. Journal of Financial Economics 42: 133–156. cial Economics 73: 375–407. Gordon SR. 2002. Easing out the founder: managing a suc- Lerner J. 1995. Venture capital and the oversight of private cessful founder/CEO transition. White paper, Spencer fi rms. Journal of Finance 50(1): 301–318. Stuart, New York. Levensohn PN. 2006. Rights of passage: managing CEO Gulati R, Higgins MC. 2003. Which ties matter when? The transition in venture-backed technology companies. contingent effects of interorganizational partnerships on White paper, Levensohn Venture Partners, San Fran- IPO success. Strategic Management Journal 24(2): cisco, CA. 127–144. Levin JS. 2001. Structuring Venture Capital, Private Equity Hambrick DC, Crozier LM. 1985. Stumblers and stars in and Entrepreneurial Transactions. Panel Publishers: the management of rapid growth. Journal of Business Frederick, MD. Venturing 1: 31–45. Loughran T, Ritter JR. 2004. Why has underpricing changed Hambrick DC, Mason PA. 1984. Upper echelons: the orga- over time? Financial Management 33(3): 5–37. nization as a refl ection of its top managers. Academy of March JG, Olsen JP. 1976. Ambiguity and Choice in Orga- Management Review 9: 193–206. nizations. Universitetsforlaget: Bergen, Norway. Haunschild PR. 1994. How much is that company worth— Meyer JW, Rowan B. 1977. Institutionalized organizations: interorganizational relationships, uncertainty, and formal structure as myth and ceremony. American acquisition premiums. Administrative Science Quarterly Journal of Sociology 83(2): 340–363. 39(3): 391–411. Miner AS, Haunschild PR. 1995. Population-level learning. Haunschild PR, Miner AS. 1997. Modes of interorganiza- In Research in Organizational Behavior, Cummings tional imitation: the effects of outcome salience and uncer- LL, Staw BM (eds). JAI Press: Stamford, CT; 115– tainty. Administrative Science Quarterly 42(3): 472–500. 166. Hayward MLA. 2002. When do fi rms learn from their Nelson RR, Winter SG. 1982. An Evolutionary Theory acquisition experience? Evidence from 1990–1995. of Economic Change. Harvard University Press: Strategic Management Journal 23(1): 21–39. Cambridge, MA. He L. 2008. Do founders matter? A study of executive Nelson T. 2003. The persistence of founder infl uence: man- compensation, governance structure and fi rm perfor- agement, ownership, and performance effects at initial mance. Journal of Business Venturing 23: 257–279. public offerings. Strategic Management Journal 24(8): Hellmann T, Puri M. 2002. Venture capital and the profes- 707–724. sionalization of start-up fi rms: empirical evidence. Pearce JA, Zahra SA. 1992. Board composition from a Journal of Finance 57(1): 169–197. strategic contingency perspective. Journal of Manage- Hendershott RJ. 2004. Net value: wealth creation (and ment Studies 29: 411–438. destruction) during the internet boom. Journal of Pfeffer J, Salancik GR. 1978. The External Control of Corporate Finance 10(2): 281–299. Organizations: A Resource Dependence Perspective. Higgins MC, Gulati R. 2003. Getting off to a good start: Harper & Row: New York. the effects of upper echelon affi liations on underwriter Pollock TG, Gulati R. 2007. Standing out from the crowd: prestige. Organization Science 14: 244–263. the availability enhancing effects of IPO-related signals Ibbotson RG, Ritter JR. 1995. Initial public offerings. In on alliance formation by entrepreneurial fi rms. Strategic Handbooks in Operations Research and Management Organization 5(4): 339–372. Science, Jarrow R, Maksimovic V, Ziemba W (eds). Pollock TG, Rindova VP. 2003. Media legitimation effects Elsevier Sciences B.V.: London; 993–1016. in the market for initial public offerings. Academy of Jain BA, Tabak F. 2008. Factors infl uencing the choice Management Journal 46: 631–642. between founder versus non-founder CEOs for IPO fi rms. Pollock TG, Rindova VP, Maggitti PG. 2008. Market Journal of Business Venturing 23: 21–45. watch: Information and availability cascades among the Jayaraman N, Khorana A, Nelling E, Covin J. 2000. CEO media and investors in the U.S. IPO market. Academy of founder status and fi rm fi nancial performance. Strategic Management Journal 51: 335–358. Management Journal 21(12): 1215–1224. Rajan R, Servaes H. 1997. Analyst following of initial Kaplan SN, Stromberg P. 2004. Characteristics, contracts, public offerings. Journal of Finance 52: 507–529. and actions: evidence from venture capitalist analyses. Rindova VP, Kotha S. 2001. Continuous ‘morphing:’ com- Journal of Finance 59(5): 2177–2210. peting through dynamic capabilities, form, and function. Kor Y. 2003. Experience-based top management team Academy of Management Journal 44: 1263–1280. competence and sustained growth. Organization Science Ritter JR, Welch I. 2002. A review of IPO activity, pricing 14: 707–719. and allocations. Journal of Finance 57: 1795–1828.

Rock K. 1986. Why new issues are underpriced. Journal of Wasserman N. 2003. Founder CEO succession and the Financial Economics 15: 187–212. paradox of entrepreneurial success. Organization Science Rubenson GC, Gupta AK. 1996. The initial succession: a 14(2): 149–172. contingency model of founder tenure. Entrepreneurship Weick KE. 1995. Sensemaking in Organizations. Sage Theory & Practice 21: 21–35. Publications: Thousand Oaks, CA. Sahlman WA. 1988. Aspects of fi nancial contracting in Weick KE, Sutcliffe KM, Obstfeld D. 1999. Organizing for venture capital. Journal of Applied Corporate Finance high reliability: processes of collective mindfulness. In 1(2): 23–36. Research in Organizational Behavior, Sutton RI, Staw Santos FM, Eisenhardt KM. 2009. Constructing markets BM (eds). JAI Press: Stamford, CT; 81–123. and shaping boundaries: entrepreneurial agency in Westphal JD, Fredrickson JW. 2001. Who directs strategic nascent fi elds. Academy of Management Journal 52: change: director experience, the selection of new CEOs, 643–671. and change in corporate strategy. Strategic Management Sapienza HJ, Gupta AK. 1994. Impact of agency risks and Journal 22(12): 1113–1138. task uncertainty on venture capitalist CEO interaction. Wiesenfeld BM, Wurthmann K, Hambrick DC. 2008. Academy of Management Journal 37(6): 1618–1632. The stigmatization and devaluation of elites associated Shen W, Cannella AA Jr. 2002. Revisiting the performance with corporate failures: a process model. Academy of consequences of CEO succession: the impacts of succes- Management Review 33(1): 231–251. sor type, post-succession senior executive turnover, and Willard GE, Krueger DA, Feeser HR. 1992. In order to departing CEO tenure. Academy of Management Journal grow, must the founder go: a comparison of growth 45: 717–733. between founder and non-founder managed high-growth Sine WD, Mitsuhashi H, Kirsch DA. 2006. Revisiting manufacturing fi rms. Journal of Business Venturing 7: Burns and Stalker: formal structure and new venture 181–194. performance in emerging economic sectors. Academy of Zacharakis AL, Meyer GD. 1998. A lack of insight: do Management Journal 49: 121–132. venture capitalists really understand their own decision Sitkin S, Pablo A. 1992. Reconceptualizing the determi- process? Journal of Business Venturing 13(1): 57– nants of risk behavior. Academy of Management Review 76. 17: 9–38. Zacharakis AL, Shepherd DA. 2005. A non-additive deci- Stuart TE, Hoang H, Hybels RC. 1999. Interorganizational sion aid for venture capitalists’ investment decisions. endorsements and the performance of entrepreneurial European Journal of Operational Research 162: ventures. Administrative Science Quarterly 44: 315–349. 673–689.