The effect of trade secret legal protection on venture capital investments: Evidence from the inevitable disclosure doctrine

Journal of Business Venturing 31 (2016) 524–541

Contents lists available at ScienceDirect

Journal of Business Venturing

The effect of trade secret legal protection on venture capital investments: Evidence from the inevitable disclosure doctrine Francesco Castellaneta b,⁎, Raffaele Conti b, Francisco M. Veloso b, Carlos A. Kemeny a,b a b

Carnegie Mellon University - Department of Engineering and Public Policy, 5000 Forbes Ave., Baker Hall 129, Pittsburgh, PA 15213, United States CATÓLICA-LISBON School of Business & Economics, Palma de Cima, 1649-023 Lisboa, Portugal

a r t i c l e

i n f o

Article history: Received 19 December 2015 Received in revised form 15 July 2016 Accepted 18 July 2016 Available online xxxx JEL classification: L26 O34 O43

a b s t r a c t This study investigates how the inevitable disclosure doctrine, a form of trade secret legal protection, affects venture capital (VC) investment. Using a data set of VC deals realized in the United States from 1980 to 2012, we find that a rule in favor of inevitable disclosure increases the amount of VC investment. We address mechanisms that can explain these findings by assessing how the inevitable disclosure doctrine (a) displays a different impact on VC investments according to the characteristics of the state and the industry where the start-ups operate and (b) affects the performance of VC-backed firms. We also discuss managerial and policy implications of our findings. © 2016 Elsevier Inc. All rights reserved.

Keywords: Intellectual property rights protection Trade secrets Inevitable disclosure doctrine Venture capital Legal environment

1. Executive summary We discuss and empirically test how the Inevitable Disclosure Doctrine (IDD) – which is a particularly strong form of trade secret legal protection available in some US states – affects Venture capital (VC) investments. In states where the IDD is embraced through a court precedent, a firm might in fact obtain a court injunction to prohibit a departing employee from walking out with valuable trade secrets and joining a competitor or founding a rival firm. Therefore, we argue that a judicial precedent clearly in favor of the IDD possibly enhances VC investments, as it reassures VC investors about the possibility to keep the VC-backed firms' key trade secrets and employees. We test our argument in the US venture capital industry from 1980 to 2012. We take advantage of the fact that, in the considered time period, some US states have embraced the inevitable disclosure doctrine whereas other states have clearly rejected it or still have an unclear position. Results show that a precedent in favor of the IDD exerts a positive effect on VC attraction, and particularly so in industries where patents are less used – and so trade secrets are likely to be more important – and in states where alternative legal means to block employee mobility – such as non-compete agreements – are less enforceable. Overall, this paper contributes to a better understanding of the relationship between the institutional environment and VC investments in particular, and, more in general, on entrepreneurship. Previous studies have found that institutional factors that limit

⁎ Corresponding author. E-mail address: [email protected] (F. Castellaneta).

http://dx.doi.org/10.1016/j.jbusvent.2016.07.004 0883-9026/© 2016 Elsevier Inc. All rights reserved.

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

525

employee mobility might be detrimental for the formation of new firms; however, their effect might be more nuanced than it is often assumed when considering other measures of the entrepreneurial ecosystem performance. Our findings suggest in fact that states constraining employee mobility might display less new companies, but, at the same time, more of those companies might become attractive to VC investors. 2. Introduction VC plays a critical role in fostering entrepreneurship, innovation, and, ultimately, economic growth (Samila and Sorenson, 2011a). Given its importance, scholars and policy makers have been trying to better understand the factors that may condition the development of VC in a given geographic area, be it a country or a region. Prior work—mainly at the country level—has considered several factors possibly enhancing VC investment, including efficient financial markets (Jeng and Wells, 2000), a favorable taxation regime (Keuschnigg and Nielsen, 2003), and the extent to which the legal environment protects and enforces property rights in general (Cumming et al., 2010) and—given that the assets of VC-backed firms are mainly intangible (Croce et al., 2013)—intellectual property rights (IPR) in particular. With respect to the role of IPR in VC, particular attention has been paid to the patent system (e.g., Hsu and Ziedonis, 2013; Mann and Sager, 2007), an important dimension of the IPR environment. Despite their role, patents are only part of IPR. A further dimension of IPR policy and practice is represented by trade secrets, defined as any information that derives independent economic value, actual or potential, from not being generally known.1 Given the breadth of knowledge potentially covered by the term, Halligan (2008: 3) argues that “the vast bulk of intangible assets are trade secret assets,” whereas Risch (2007: 656) notices that trade secrets are “the most important and most heavily litigated intellectual property rights.” Therefore, it appears significant to extend this line of inquiry into how legal protection associated with trade secrets might influence the presence and role of VC in a geographic area. To help fill this gap, we focus on the IDD whose rule determines whether the owner of a trade secret can (if the rule is in favor of the IDD) or cannot (if the rule is against the IDD) obtain a court injunction to prohibit a departing employee from working for a competitor or founding a rival firm, on the grounds that she could inevitably disclose trade secrets (Lowry, 1988). The IDD can thus allow a company not only to protect its extant trade secrets but also to avoid the loss of valuable human capital to a competing firm. In the United States, where the IDD has predominantly been developed, the extent to which a state jurisdiction embraces it varies. Through its court precedents, a state may adopt (a) a rule clearly in favor of inevitable disclosure, (b) a rule clearly against it, or (c) no clear rule. These scenarios could condition the decisions of venture capitalists (VCs) to invest in a state. In particular, we argue that a state embracing a rule clearly in favor of the IDD should attract VC investments more than any other possible scenario (against, or no clear rule), for reasons related to the likelihood of obtaining a court injunction to restrict employee mobility. Given the VCs' ability to select high-quality companies, for VC-backed firms the possibility of avoiding any knowledge leakages due to outbound employee mobility is likely to be more substantive than the possibility of receiving knowledge inflows from inbound mobility (Chemmanur et al., 2011; Sorensen, 2007).2 Therefore, VC investors would generally prefer to invest in a legal environment where key employees of an invested firm cannot easily leave the company to pursue opportunities in competition with the former employer (Baron et al., 2001). In this respect, a precedent in favor of inevitable disclosure increases the likelihood that a VC-backed firm obtains a court injunction against a former employee hired by a competitor (Png and Samila, 2015). Furthermore, given that a state court will tend to make decisions consistent with the precedent in any similar case at hand, a case clearly in favor of inevitable disclosure also enhances the predictability of this court injunction. Higher predictability is desired not only by risk-averse VC investors, who prefer a more stable institutional environment (Malesky and Samphantharak, 2008), but also by risk-neutral investors, who might otherwise prefer to wait and see how the regulatory environment evolves before making investments (Bloom et al., 2007). To empirically assess whether and how rulings on inevitable disclosure stimulate investment from VCs, we exploit longitudinal variation in inevitable disclosure rules in U.S. states, as determined by court precedents (Kahnke et al., 2008; Klasa et al., 2014; Malsberger, 2011; Milgrim and Bensen, 2015; Png and Samila, 2015; Quinto and Singer, 2009; Wiesner, 2012). We find that a rule in favor of inevitable disclosure significantly increases the amount of VC available for local start-ups compared with where there is no rule or where the rule is against inevitable disclosure. 3. Theory: Trade secret legal protection, inevitable disclosure doctrine, and venture capital It is well established that the extent to which formal institutions protect and enforce property rights might play a crucial role for VC attraction, as most investees' assets are in fact intangible (Cumming et al., 2010; Croce et al., 2013). In this respect, trade 1 More precisely, the U.S. Uniform Trade Secrets Act, §1.4, defines a trade secret to mean “information, including a formula, pattern, compilation, program device, method, technique, or process, that: (i) derives independent economic value, actual or potential, from not being generally known to, and not being readily ascertainable by proper means by, other persons who can obtain economic value from its disclosure or use, and (ii) is the subject of efforts that are reasonable under the circumstances to maintain its secrecy.” 2 The papers of Sorensen (2007) and Chemmanur et al. (2011) refer to the United States, where most value created by a VC-backed start-up was already present when VCs began investing. However, an important caveat is that this finding does not hold in Europe (e.g., Croce et al., 2013), neither for traditional VC (Cumming et al., 2014) nor for other types of VCs, such as corporate VCs (Colombo and Murtinu, 2016), governmental VCs (Grilli and Murtinu, 2015, 2014), or bank-affiliated VCs (Cumming and Murtinu, 2016).

526

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

secret legal protection is generally considered to be broader – and so potentially more valuable – than protection granted by other IPR, including patents, copyrights, and trademarks (Besen and Raskind, 1991). First, whereas patent and copyright protection require subject matter to be novel, trade secret protection requires only that the subject matter derive some commercial value from not being known (Kitch, 1980). Second, patents, copyrights, and trademarks protect only explicit knowledge—that is, knowledge already articulated and stored in certain media. Trade secrets, in contrast, protect any proprietary knowledge that is not known to others and that can provide a competitive advantage, be that explicit or tacit knowledge.3 Finally, unlike patent protection, trade secret protection is exempt from having an expiration date and may live as long as the knowledge is kept secret. A critical element associated with the role of trade secrets, as in all dimensions of IPR, is enforceability. Although several aspects contribute to the enforceability of trade secrets (David, 1993), the extent to which they are in fact protected may critically depend on whether a jurisdiction embraces the IDD. This doctrine determines whether the owner of a trade secret can (if the rule is in favor of the IDD) or cannot (if the rule is against the IDD) obtain a court injunction to prohibit a departing employee from working for a competitor or founding a rival firm, on the grounds that he or she could inevitably disclose trade secrets (Lowry, 1988). In the United States, where the IDD has been predominantly developed, the rule is established through court precedents. This means that once the rule of law is established for the first time by a court for a particular case, it is thereafter referred to when similar cases are decided, effectively binding the future—which is the essence of the stare decisis principle (Hart, 2012; Horwitz, 1977; Landes and Posner, 1976).4 The modern form of the IDD was defined in 1995 with the Seventh Circuit's decision in PepsiCo, Inc. v. Redmond. Despite other decisions having embraced the IDD as a further reason to enforce a non-compete agreement5 or to grant a limited injunction,6 PepsiCo, Inc. v. Redmond was the first court decision to issue a very broad injunction, which prohibited de facto a departing employee – in that case a high-level manager at PepsiCo – who had not signed a non-compete agreement from working for a competitor, on the grounds that he or she would inevitably disclose trade secrets (Mulcahy and Tassin, 2003).7 By hampering the mobility of those key employees who have access to a firm's most valuable trade secrets, the IDD might exert an ambiguous effect on start-ups' profitability—and so on VCs' incentive to invest in a start-up. On one side, in new firms, performance outcomes substantially depend on a few critical employees, such as founders, who possess the relevant knowledge for nurturing the business and so ensure its future profitability (Campbell and Ganco, 2012; Colombo and Grilli, 2005). On the other side, growing a business involves hiring people who possess the relevant technical or managerial skills (Hellmann and Puri, 2002), especially from established companies. For the average start-up, it is debatable whether the possibility of reducing mobility-related knowledge leakages to competitors is more important than the possibility of benefiting from inbound knowledge inflows through recruiting. However, we argue that for VC-backed firms, the former possibility is more important than the latter, because most of the VC-backed firms' value is already present when the VCs begin investing. For instance, Chemmanur et al. (2011) find that, compared with firms that do not obtain VC financing, firms financed by VCs are already 7% more productive before receiving VC funding, whereas after receiving the funding they gain only five points more in productivity. Similarly, Sorensen (2007) finds that about two-thirds of the value created by VCs derives from selection—from VCs picking those start-ups that are ex-ante more valuable—and only one-third from value addition—from VCs nurturing ex-post the businesses they have chosen. Given that VC-backed firms possibly represent the most valuable new businesses among the overall start-up population, and that most of this value is already present when VCs start investing, VC investors will make every effort to ensure the permanence of the VC-backed companies' key employees (including founders).8 In fact, at investment time, VCs often establish contractual clauses that make it difficult for key employees to depart from the financed company, such as vesting clauses – that is, a legal arrangement in which a key employee's shares are originally held by the company and awarded over a multiyear period conditional on the employee not leaving the company (Gompers and Lerner, 2001) – and non-compete clauses – which prohibit key employees from joining a competitor for a specified period (Kaplan and Strömberg, 2003).9 Although those clauses might be effective for reducing outbound mobility, entrepreneurs or key employees may seek extra compensation in exchange for signing them, which would be reflected naturally in lower VC-backed firm profits and returns to VC investments. In contrast, once a 3 For instance, trade secret protection may encompass not only chemical formulae and customer lists, both of which can be stored and represent explicit knowledge, but also “negative know-how” obtained through previously attempted but failed techniques or procedures (Graves, 2006), which is essentially tacit knowledge. 4 Although it is possible that individual courts will deviate from precedent, the process of appeals to higher courts within each state lowers the probability that any disregard for precedent is systematic (Landes and Posner, 1976). 5 Eastman Kodak Co. v. Powers Film Prod., 189 A.D. 556 (N.Y.A.D. 1919). 6 B. F. Goodrich Co. v. Wohlgemuth, 117 Ohio App. 493, 192 N.E.2d 99 (1963). 7 Furthermore, another important characteristic of PepsiCo, Inc. v. Redmond is that the trade secrets involved in the case were not of a technical nature, as was generally true in earlier cases considering inevitable disclosure. 8 Even anecdotal evidence confirms how VC investors place a great importance on keeping the key employees of a start-up they invest in. For instance, Fred Wilson—the co-founder of Union Square Ventures, a New York City–based venture capital firm with investments in Web 2.0 companies such as Twitter, Tumblr, Foursquare, Zynga, Kickstarter, and 10gen—says that “it's not a good thing when one of your key employees, particularly one that knows your entire customer base or your entire code base, or worse knows your entire employee population and who is a superstar and who is not, leaves to join a direct competitor. It's in the company's interest to stop that from happening. So I am in favor of non-competes for senior members of a management team and key employees” (http://avc.com/2007/12/ thinking-about-4/). To get a sense of the importance of IDD for venture capitalists, we also performed a short survey collecting responses by a convenience sample of 28 US VCs. The survey was composed of two questions in order to enhance the likelihood of response. In the first question, we asked VCs how important it is (on a 1 to 5 scale) for them to make sure that key employees stay with the VC-backed firm. In the second question, we asked whether, when making VC investments, they take into account the possibility of relying on the IDD. The results indicate that 19 VCs (67.9%) rated as 5 (the maximum) the importance of keeping key employees; 8 VCs (28.6%) rated this importance as 4, and one VC (3.6%) rated it as 3. At the same time, 16 VCs (57.1%) take into account the existence of the IDD when making investments, whereas 12 (42.9%) do not. 9 As reported by Kaplan and Strömberg (2003: 12), “non-compete clauses are used in approximately 70% of the financings.”

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

527

jurisdiction has embraced the IDD, VC-backed firms could limit employee mobility through a court injunction even in the absence of any specific contractual clause. The extent to which U.S. state jurisdictions embrace the IDD varies substantially according to court precedents. In particular, whereas some state courts have ruled clearly in favor of or against inevitable disclosure, others have reached unclear decisions—that is, ambiguous decisions that do not clarify exactly the IDD's scope of application—or have yet to decide on the doctrine. This generates three possible contexts for a given state: (a) a rule clearly in favor of inevitable disclosure, (b) a rule clearly against inevitable disclosure, (c) no clear rule. With respect to the baseline case of having no clear rule, a rule in favor of inevitable disclosure should attract VC investors for two reasons related, respectively, to the expected value of obtaining a court injunction against a former employee hired by a competitor and to the predictability of that injunction. First, in the presence of a clear precedent embracing the IDD, the likelihood of a court injunction that blocks departing key employees increases (Png and Samila, 2015). This result would naturally attract VC investors, whose returns depend on the ability of the firms they invest in to retain their best employees. In fact, when a rule in favor of inevitable disclosure is adopted in a state, local firms experience an increase in value (Qiu and Wang, 2015), which is precisely a consequence of their greater ability to reduce unintended outbound mobility, and thus both to keep their most valuable knowledge inside the organizational boundaries (Gao and Ma, 2016; Png and Samila, 2015) and to decrease the intensity of local competition (Klasa et al., 2014).10 Second, with a clearly favorable precedent, any court will most likely apply the IDD to any similar case at hand in the same direction as previous cases. Higher predictability of court decisions in inevitable disclosure cases, compared with a scenario where there is no rule at all, should attract both risk-averse and risk-neutral investors. Higher predictability clearly benefits risk-averse investors, who naturally prefer to invest in more predictable institutional environments (Malesky and Samphantharak, 2008). And it may also benefit risk-neutral investors, who might otherwise prefer to wait and see how the regulatory environment evolves before making investments.11 In real option terms, uncertainty about the legal environment makes the “option to wait” more valuable (Bittlingmayer, 2000; Bloom et al., 2007; Dixit and Pindyck, 1994). It is not apparent whether, unlike with a rule in favor of inevitable disclosure, a rule against is superior to the baseline scenario of no clear rule. For instance, a clear rule against inevitable disclosure increases the predictability of a court decision and, as a result, should attract VC investment. However, it also decreases the possibility of retaining employees, as state courts would most likely allow employees to leave for other opportunities, which should translate into a decrease in VC investment. Hence, whether a rule against the IDD increases VC investment more than not having any rule is an empirical matter. In fact, it depends on whether the benefits due to the reduction in regulatory uncertainty outweigh the costs related to the possible loss of the investee firm's key employees. Table 1 summarizes the potential effects of having a clear rule in favor of or against inevitable disclosure, compared with the baseline scenario of having no rule, on the expected value of obtaining a court injunction against an employee deciding to leave for joining a competitor and the predictability of that injunction. Compared with the baseline no-rule scenario, a rule in favor of the IDD—different from no rule or a rule against inevitable disclosure—increases both the expected value of obtaining an injunction and the predictability of such injunction. Hence, VCs should be attracted more to jurisdictions where there exists a rule clearly in favor of inevitable disclosure than to jurisdictions with rules against or no rules. Accordingly, we formulate the following baseline hypothesis: A rule in favor of inevitable disclosure increases the amount of venture capital investment available for local start-ups more than a rule clearly against inevitable disclosure or the absence of a clear rule. In the next sections, we first verify the validity of this hypothesis. Then, we try to uncover the mechanisms driving the findings. In particular, we explore how inevitable disclosure conditions the overall level of VC investments, as well as the number of start-ups. We also analyze how alternative institutions that VCs might leverage to reduce the possibility of key employees leaving the invested firm with valuable knowledge condition the effect of inevitable disclosure. Finally, we look at how the performance of a VC-backed firm is influenced by the state where it is located has embraced the IDD. 4. Data, variables and methodology 4.1. Data Our empirical analysis relies on a balanced panel of the 50 U.S. states and the District of Columbia from 1980 to 2012. Because variation in inevitable disclosure occurs at the state level, this is the most appropriate level of analysis. The sample period begins in 1980 because the Thomson Reuters' VentureXpert data set, which was used to gather data on VC investment, has limited 10 The increase in value—and in the corresponding VC investment—following the adoption of the IDD should be particularly salient for start-ups that mainly rely on trade secrets (vis-à-vis patents) to protect their knowledge base. For example, patents are typically barely used in the software industry (Hall, Helmers, Rogers, and Sena, 2014), where instead companies rely on trade secrets to protect their most important ideas—the classical example being the trade secret protecting Google's search engine algorithm. 11 For instance, when there is no clear precedent on the IDD—such that it is not certain whether and to what extent courts might embrace the doctrine in future cases—VC investors may turn to non-compete clauses to prevent employee departure. Yet, money associated with drafting and negotiating these terms with key employees might be wasted if, at some point in the future, a court precedent in favor of inevitable disclosure is established.

528

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

Table 1 Anticipated effect of inevitable disclosure court rulings. Court ruling on inevitable disclosure

With respect to having no rule Expected value of obtaining a court injunction

Predictability of a court injunction

Favorable rule Against rule

Increase Decrease

Increase Increase

coverage of investments realized in the 1970s and is thus fully reliable only beginning with the 1980s (Gompers and Lerner, 1999). The sample period ends in 2012 because this is the last year for which we could have observations for all our variables. From the VentureXpert database, we collected information about the amount of equity invested per deal, the investment date, and the location of both investee and investor.12 There were 83,595 deals completed within the relevant period, from which we excluded 152 because investee location information was unknown. State-level rulings of inevitable disclosure were gathered from a variety of sources (Kahnke et al., 2008; Klasa et al., 2014; Malsberger, 2011; Milgrim and Bensen, 2015; Png and Samila, 2015; Quinto and Singer, 2009; Wiesner, 2012).13 Moreover, we collected data on the number of new firms from the Business Dynamics Statistics database. Finally, we gathered state gross domestic product (GDP) and population data from the U.S. Department of Commerce's Bureau of Economic Analysis (BEA). Overall, we constructed a balanced panel data set of 1683 state-year observations. 4.2. Variables 4.2.1. Dependent variables 4.2.1.1. Venture capital investment per start-up. Similar to Samila and Sorenson (2011a), we defined VC investment as the equity investment associated with any VC deal at different stages of financing: seed, early, later, or in balanced stages. Deal equity was aggregated into state-year observations (by investee headquarters location and investment year). Then we divided the overall amount of VC investment in a certain state and year by the number of local start-ups, that is, the number of new firms created in that state and year. Compared to the mere amount of VC, the amount of VC investments per start-up has the advantage of being a measure net not only of any change in the entrepreneurship rate possibly determined by precedents in favor of or against the IDD but also of any scale effect. At any rate, as a robustness check, or whenever data on the number of start-ups are not available (e.g., at the industry level), we also consider the sheer amount of VC as a dependent variable. 4.2.2. Independent variables 4.2.2.1. Inevitable disclosure rule. To measure the inevitable disclosure rule, we (a) established the criteria for defining when a decision is in favor of the IDD, against it, or unclear; (b) identified the landmark decisions about inevitable disclosure in each state; and (c) classified each decision according to the aforementioned criteria. Throughout this process, we relied on the pro bono help of a leading legal scholar in the field of U.S. intellectual property law.14 In more detail, we first defined the IDD consistent with PepsiCo, Inc. v. Redmond as a legal doctrine through which, even without a covenant not to compete, an employer can enjoin a former employee from working for a competitor or from founding a rival firm by demonstrating that the employee's new job duties will inevitably lead to trade secret misappropriation (Kahnke et al., 2008). Based on this definition, we created three dummy variables measuring the rule on inevitable disclosure in each U.S. state, as established by court precedents, in any given year from 1980 to 2012. In particular, the favorable dummy equals 1 if the state courts have clearly embraced a rule in favor of inevitable disclosure. The against dummy equals 1 if the state courts have clearly embraced a rule against inevitable disclosure. The no clear rule dummy equals 1 if (a) the state courts have not ruled on inevitable disclosure, or (b) they have ruled but without clarifying whether the doctrine can be applied to block an employee who has not signed a covenant not to compete, or (c) they have adopted a position not in line with a former precedent—as in the case of New York, whose state courts embraced the doctrine in 1997 but became more cautious and ambiguous after 2003. Details on the criteria we used to identify and codify the most important rulings on inevitable disclosure in any state are discussed in Appendix B. The resulting measures of inevitable disclosure rule are presented in Table 2. A more detailed list of all the rulings we took into account for measuring state rule on inevitable disclosure is presented in Appendix Table B1. Overall, our identification strategy relies on the assumption that case-specific court decisions are exogenous and thus not driven by a willingness to attract VC investments or by the presence of an already active VC community. However, to verify the exogeneity assumption, we also check whether a ruling in favor of or against inevitable disclosure is related to the state's political 12

Data were collected from the VentureXpert database on November 3, 2014. See Appendix Table B1. The legal scholar providing his help is a full professor of law at a U.S. university and an attorney. He is listed as among the “Best Lawyers in America” in intellectual property law and in information technology law, as well as in Who's Who in America, Who's Who in Law, the International Who's Who of Internet Lawyers, the International Who's Who of Business Lawyers, and the Who's Who of Intellectual Property Lawyers, and he is listed in Martindale and Hubble as AV, Preeminent. This legal scholar (a) validated the criteria for defining when a decision is in favor of the IDD, against it, or unclear; (b) checked our original list of landmark decisions on inevitable disclosure and suggested additions to or deletions from that list; and (c) reviewed each decision to corroborate or change our initial classification. 13 14

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

529

Table 2 State rule on inevitable disclosure. Sources: (Kahnke et al., 2008; Klasa et al., 2014; Malsberger, 2011; Milgrim and Bensen, 2015; Png and Samila, 2015; Quinto and Singer, 2009; Wiesner, 2012). State

Year

Rule

California Delaware Florida Illinois Iowa Louisiana Maryland Massachusetts Minnesota New Jersey New York New York North Carolina Pennsylvania Utah Virginia Washington

1944 2006 2001 1995 2002 1967 2004 1995 1992 1980 1997 2003 1976 2010 1998 1999 1997

Against Favorable Against Favorable Favorable Against Against Against Against Against Favorable No clear rule Against Favorable Favorable Against Favorable

orientation—which might be more or less prone to enact pro-VC policies (e.g., Pe'Er and Gottschalg, 2011)—the presence of lobbying activities in the state and the past amount of VC investment in the state (see Appendix Table A5). The list of all variables and their measures is provided in Appendix Table A1. 4.3. Methodology We evaluated the 33-year panel using a state fixed-effects model. Thus, our methodology resembles a typical difference-indifference (diff-in-diff) strategy, through which we compare, for instance, whether states that adopted a rule in favor of inevitable disclosure experience a change in VC investments compared to the other states. VC investment per start-up is a continuous, nonnegative, and skewed variable. Therefore, we use a Poisson quasi-maximum likelihood estimator (QMLE). As pointed out by Cameron and Trivedi (1998, p. 104), “for a right-skewed continuous nonnegative dependent variable, there are reasons for using the Poisson QMLE rather than the more customary OLS in a log-linear model.” First, a Poisson QMLE, compared to a log-linear OLS, has the fundamental advantage of handling zero values of the dependent variable without the need to add an arbitrary constant value that might bias the estimates (Cameron and Trivedi, 1998). Second, even assuming that all observations are positive, “in the presence of heteroskedasticity, estimates obtained using log-linearized models are severely biased” (Silva and Tenreyro, 2006, p. 641). Hence, in the main specification we estimated the following model using a Poisson QMLE:   VC investment per start−upi;t ¼ f β0 þ β1 FAVORABLEi;t−1 þ β2 AGAINST i;t−1 þ β3 STATEGDPi;t−1 þ γ i þ ci þ εit

ð1Þ

where i indexes the state and t indexes the year; VC investment per start-upi,t is the amount of money invested in a certain state and year in VC deals, divided by the overall number of start-ups (i.e., new firms) in that state and year; FAVORABLEi,t − 1 and AGAINSTi,t − 1 are the inevitable disclosure rule dummy variables equal to 1 if, respectively, a rule clearly in favor of or against inevitable disclosure was already enacted in the state, and 0 otherwise. We assume that there exists a one-year lag from the time a court decision is made to when it actually has an effect on VC investments. STATEGDPi,t − 1 is the state GDP control variable, γi represents the series of year fixed effects, ci represents state fixed effects, and εit is the error term. Regarding the error term, to account for the presence of serial correlation and to avoid inconsistent standard errors, we clustered observations at the state level—specifically, the state where companies that receive VC investment are located (Bertrand et al., 2004). According to the baseline hypothesis, we expect β1 to be significantly positive and greater than β2, which means that a favorable rule on inevitable disclosure leads to more VC investment per start-up, when compared with the other possible scenarios (i.e., a rule against or no rule). 5. Results Appendix Table A2 shows the distribution of VC investments across states. It emerges that, as expected, equity investment is unevenly distributed across states: California and Massachusetts display the highest VC investments per start-up, whereas Alaska and Wyoming report almost no VC investment per start-up. In Table 3 we present the summary statistics for the variables used in the empirical analysis. (Their pairwise correlations are presented in Appendix Table A3.) Overall, the treated group of the seven U.S. states deciding in favor of the IDD corresponds to 231 state-year observations, of which 161 are before the decision and 70—about 7% of the overall observations—are after the decision.

530

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

Table 3 Descriptive statistics of the main variables.

VC investment per start-up Overall VC investment Number of start-ups Favorable rule Against rule No decision rule State GDP Lobbying Stock option tax rate UTSA enactment Non-compete enforceability Internal Rate of Return Number of IPOs Number of IPOs divided by VC-backed firms

N

Mean

Std. Dev.

Min

Max

1683 1683 1683 1683 1683 1683 1683 1683 1650 1683 1683 338 1683 1497

0.0246359 408.1336 9536.681 0.0415924 0.1188354 0.8395722 2.070292 38.51753 0.0150395 0.6506239 0.3575956 0.3562416 4.246583 0.1048002

0.0664815 1991.312 11,188.19 0.1997152 0.3236909 0.3671114 2.649651 42.94863 0.0086403 0.4769147 0.1532628 1.623785 14.06906 0.1560795

0 0 569 0 0 0 0.1054506 0 0 0 0 −1 0 0

1.23891 52,086.15 74,879 1 1 1 20.08805 337 0.0358871 1 0.75 20 178 1

Figs. 1 and 2 provide preliminary support for the prediction that a decision in favor of inevitable disclosure increases VC investment per start-up. Before a decision in favor of inevitable disclosure, treated states (states embracing the IDD) and control states (only states with no clear rule on the IDD in Fig. 1, or also states against the IDD in Fig. 2) experience similar patterns of VC investment per start-up—consistent with the parallel-path assumption we need for estimating the effect of a decision in favor of the IDD in a diff-in-diff framework. Yet, after the decision, treated states experience a sharp increase in VC investment per start-up, which is particularly evident a few years after the decision but also persists over time. Interestingly, this increase in treated states seems not to correspond to a decrease in the control states; in other words, VCs seem not to move their investments from control to treated states; rather, they just invest more in the latter. This evidence is consistent with the stable unit treatment value assumption (SUTVA), requiring—in order to properly estimate the treatment effect in any experimental or quasi-experimental design—that the outcome of one unit should be unaffected by the treatment assignment to other units.15 The previous non-parametric evidence should be complemented with the findings of a multivariate Poisson regression estimating the effect of a decision in favor of the IDD on the amount of VC per start-up, as shown in Table 4. In model 1, the favorable dummy is not only positive and significant (β1 = 0.237, p b 0.05) but also greater than the against coefficient dummy, which is instead not significantly different from zero. Thus, consistent with our baseline hypothesis, a favorable rule increases VC investment per start-up by about 27% more than a rule against inevitable disclosure or the absence of a clear rule.16 We obtain similar estimates if we use, rather than a Poisson model, an OLS regression (Table 4, column 6). The positive impact of a rule in favor of inevitable disclosure also holds when we control for state industrial composition—as proxied by the share of establishments in different SIC 1-digit industries—(column 2); the importance of lobbying activity in the state—as measured by the number of political organizations as in Sobel (2008)—(column 3)17; changes in taxation on employee stock options—as measured by the individual income tax rate (Bickley, 2012)—(column 4); and the enactment of the Uniform Trade Secrets Act (UTSA)— (column 5) (Castellaneta et al., 2016).18 Our results substantially hold even when we restrict the time period to either three, four, or five years before and after a decision in favor of the IDD (Appendix Table A4). Our main findings also hold when considering the ratio between overall VC investments and the number of innovative start-ups—as proxied by the number of organizations patenting for the first time (results available upon request).19 In Table 5, we disentangle the effect of inevitable disclosure on the numerator (overall VC investments) and the denominator (number of start-ups) of our main dependent variable (VC investments for start-ups). It turns out that a decision in favor of inevitable disclosure does not enhance the number of start-ups (columns 3 and 4) – and neither the number of innovative start-ups as measured by the number of new patent assignees (results available upon request). Rather, when compared to a decision against inevitable disclosure—which allows employees to move and found their own companies—a favorable decision might even be detrimental for the number of new firms (column 3).20 However, it substantially increases the incentive of VCs to invest in companies located in the state, such that the overall amount of VC investments increases (columns 1 and 2). In particular, a decision in favor of the IDD increases the number of VC-backed firms (column 5) rather than the average VC investment per VC-backed firm (column 6). In other words, it seems that a decision in favor of inevitable disclosure benefits a group of existing

15 However, even assuming that this assumption does not hold in our context, we still believe our result would be interesting and would have relevant policy implications. To be sure, if the SUTVA were not true, we could no longer interpret the estimated coefficient of a favorable decision as the increase in VC investments following a decision in favor of inevitable disclosure. Yet, that coefficient would still be a valid measure of the gap that a decision in favor of ID would create between the states embracing ID and the others. 16 This implies that each start-up in a state might potentially receive about 6898 dollars more of VC after a precedent in favor of inevitable disclosure is established. 17 Our measure of lobbying captures aggregate lobbying effects, not necessarily the specific lobbying by the VC industry. Therefore, our aggregate measure cannot fully rule out the possibility that VC lobbying might determine changes in the IDD rule. 18 To control for the enactment of the UTSA, we include a variable equal to 1 after the year of enactment in a certain state and 0 otherwise (Castellaneta et al., 2016;Png and Samila, 2015). 19 The mere number of start-ups might in fact be an imprecise measure of the truly entrepreneurial start-ups that VCs invest in (Henrekson and Sanandaji, 2014). 20 Since this finding is not robust to using different model specifications (i.e., Poisson versus OLS), it should be interpreted with caution.

531

.02

.04

.06

.08

.1

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

-5

0

5

10

Time from treatment Treated states

Control states

.02

.04

.06

.08

.1

.12

Fig. 1. The impact of inevitable disclosure on VC investment per start-up (control group includes states with no clear rule) Notes. Investment per start-up is in millions of 2009 USD.

-5

0

5

10

Time from treatment Treated states

Control states

Fig. 2. The impact of inevitable disclosure on VC investment per start-up (control group includes state with against or no clear rule) Notes. Investment per start-up is in millions of 2009 USD.

companies that without the IDD would have not received any VC investments. These companies might be start-ups that rely extensively on trade secrets to protect their most valuable knowledge, for instance because they operate in industries that do not rely on patents. When we analyze the impact of inevitable disclosure at the state-industry (SIC 2-digit)-year level, we find that the positive effect of a favorable decision on VC investments is weaker in industries where firms rely more on patents to protect their knowledge (Table 6, column 1)—where the industry reliance on patents is measured by the industry average number of patents per firm.21 The effect of the IDD should also depend on alternative institutions that VCs can leverage to reduce the possibility of key employee leaving the invested firm with valuable knowledge. For instance, VCs often ask key employees of the VC-backed firm to sign a formal non-compete agreement; however, the enforceability of such agreements varies substantially across states (Bird and Knopf, 2014; Garmaise, 2011).22 Hence, we expect that a decision in favor of inevitable disclosure should particularly benefit VCs in states where non-competes are barely enforceable. This is exactly what we find in Table 6, column 2.

21 This is an industry time invariant measure. In particular, to compute the average number of patents per firm in any industry (SIC 2-digits) in the overall time period, we assigned each patent to a (public) parent company, based on Bessen's (2009) concordance file. 22 The strength of enforceability is available only until 2004; in particular, Garmaise's (2011) enforceability index covers the period 1992–2004, whereas Bird and Knopf's (2014) index covers the period 1976–1994.

532

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

Table 4 The impact of inevitable disclosure on VC investment per start-up. Variable

(1) VC investment per start-up

(2) VC investment per start-up

(3) VC investment per start-up

(4) VC investment per start-up

(5) VC investment per start-up

(6) VC investment per start-up (log)

Favorable rule

0.237⁎⁎ (0.094) 0.025 (0.093)

0.293⁎⁎⁎ (0.113) −0.022 (0.110) −1.110 (18.926) 9.045 (7.184) −3.640 (6.478) 50.288⁎⁎

0.242⁎⁎⁎ (0.093) 0.021 (0.090)

0.236⁎⁎⁎ (0.089) 0.019 (0.084)

0.239⁎⁎ (0.093) 0.017 (0.099)

0.410⁎⁎⁎ (0.110) 0.342 (0.224)

−0.040 (0.095) 0.030 (0.018) Yes Yes 1683 −91.516

0.136⁎⁎⁎ (0.023) Yes Yes 1683

Against rule Share of firms in SIC 0 Share of firms in SIC 1 Share of firms in SIC 2–3 Share of firms in SIC 4

(22.242) 2.877 (5.819) −8.775 (6.914) 6.455 (6.017)

Share of firms in SIC 5 Share of firms in SIC 6 Share of firms in SIC 7–8

−0.001⁎⁎ (0.001)

Lobbying

−21.415⁎⁎⁎ (7.443)

Stock option tax rate UTSA enactment State GDP State fixed effects Year fixed effects Observations Log likelihood R-squared

0.035⁎ (0.020) Yes Yes 1683 −91.340

0.030⁎ (0.018) Yes Yes 1683 −91.517

0.040⁎⁎ (0.018) Yes Yes 1683 −91.509

0.041⁎⁎ (0.018) Yes Yes 1650 −87.245

0.646

Notes. Poisson regression results in columns (1)–(5). OLS regression results in column (6). Robust standard errors clustered by state in parentheses. The dependent variable is the VC investment per start-up. The “stock option tax rate” variable is not available for the District of Columbia. The unit of observation is the state-year, and the data cover the 50 U.S. states (and the District of Columbia) from 1980 to 2012. ⁎ p b 0.1. ⁎⁎ p b 0.05. ⁎⁎⁎ p b 0.01.

Table 5 The impact of inevitable disclosure on overall VC investment, the number of start-ups, VC-backed firms, average VC investment. Variable

(1) Overall VC investment

(2) Overall VC investment (log)

(3) Number of start-ups

(4) Number of start-ups (log)

(5) Number of venturebacked firms

(6) Average VC investment per venture-backed firm

Favorable rule

0.180⁎⁎⁎ (0.066) −0.061 (0.122) 0.051⁎⁎⁎ (0.012) Yes Yes 1683 −28,303.856

0.527⁎⁎⁎ (0.194) 0.182 (0.143) 0.099⁎⁎ (0.038) Yes Yes 1683

0.030 (0.045) 0.114⁎⁎

0.100⁎ (0.088) 0.063 (0.044) 0.022⁎⁎ (0.010) Yes Yes 1683

0.239⁎⁎⁎ (0.079) −0.029 (0.086) 0.030⁎⁎⁎ (0.009) Yes Yes 1683 −4954.627

0.084 (0.139) −0.010 (0.061) 0.018 (0.011) Yes Yes 1497 −3904.225

Against rule State GDP State fixed effects Year fixed effects Observations Log likelihood R-squared

0.588

(0.053) 0.012⁎⁎ (0.006) Yes Yes 1683 −81,501.552

0.437

Notes. Poisson regression results in columns (1), (3), (5), (6). OLS regression results in columns (2) and (4). Robust standard errors clustered by state in parentheses. In model (1) and (2), the dependent variable is the overall amount of VC investment. In model (3) and (4), the dependent variable is the number of start-ups. In model (5), the dependent variable is the number of venture backed firms. In model (6), the dependent variable is the average VC investment per venture-backed firm. The unit of observation is the state-year, and the data cover the 50 U.S. states (and the District of Columbia) from 1980 to 2012. ⁎ p b 0.1. ⁎⁎ p b 0.05. ⁎⁎⁎ p b 0.01.

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

533

Table 6 The impact of inevitable disclosure on: VC investment according to industry patent reliance and state non-compete enforceability (models 1–2); VC-backed firm performance (models 3–5). Variable

(1) Overall VC investment

(2) VC investment per start-up

(3) Number of IPOs

(4) Number of IPOs divided by VC-backed firms

(5) Internal Rate of Return

Favorable rule

0.293⁎⁎ (0.121) −0.091 (0.134) 0.112⁎⁎⁎ (0.020) −0.054⁎⁎ (0.022)

0.993⁎⁎⁎ (0.174) −0.076 (0.086)

0.400⁎⁎⁎ (0.149) −0.141 (0.089)

0.549⁎⁎ (0.227) 0.131⁎

0.390⁎⁎ (0.187) −0.157 (0.270)

−1.002 (0.993) −1.998⁎⁎⁎ (0.537) −0.011 (0.027) Yes Yes 1275 −55.820

0.038⁎⁎ (0.015) Yes Yes 1584 −1933.116

0.050⁎⁎ (0.022) Yes Yes 1497 −348.848

Against rule Industry patent reliance Favorable rule ∗ Industry patent reliance Non-compete enforceability Favorable rule ∗ Non-compete enforceability State GDP State fixed effects Year fixed effects Observations Log likelihood R-squared

0.036⁎⁎ (0.015) Yes Yes 12,383 −796,497.923

(0.217)

0.034 (0.035) Yes Yes 338 0.136

Notes. Poisson regression results in models (1)–(3); fractional logit regression results in column in model (4); OLS regression in model (5). Robust standard errors clustered by state in parentheses. In model (1), the dependent variable is overall VC investment in any SIC-2 digit. In model (2), the dependent variable is VC investment per start-up. In model (3) the dependent variable is the number of IPOs; in model (4) is the share of VC-backed companies eventually exiting through an IPO; in model (5) is the average IRR of VC-backed firms in the state. The unit of observation is the industry-state-year in model 1 and state-year in model 2, 3, 4, and 5. “Industry patent reliance” is a time invariant measure. The data cover the 50 U.S. states (and the District of Columbia) from: 1980 to 2012 in model 1, 3, 4, and 5; 1980 to 2004 in model 2—data on non-compete enforceability are missing after 2004. ⁎ p b 0.1. ⁎⁎ p b 0.05. ⁎⁎⁎ p b 0.01.

However, VCs investors might also rely on informal enforcement mechanisms based on reputation and trust, which usually develop over time (e.g., Greif, 1993). If so, the positive effect of a decision in favor of inevitable disclosure on VC investments should be more salient for first-round than for later-round investments. In the latter, compared to the former, the risk of startup founders and key employees leaving the firm is lower, since in the eyes of actual and future investors they may have developed some level of reputation, which they are unwilling to harm with opportunistic behaviors. In fact, we find the positive effect of a decision in favor of the IDD to be particularly evident for first-round investments—(results available upon request). Similarly, we find that the positive effect of a decision in favor of the IDD is particularly evident for early-stage vis-à-vis late-stage investments, which might (also) be because the importance of avoiding key employees' departure is greater for less developed start-ups (e.g., Colombo and Grilli, 2005)— (results available upon request). Finally, our prediction that inevitable disclosure increases VC investments is essentially based on the idea that keeping key employees is crucial for VC-backed firms' performance—and thus for VC investors' returns. Hence, we should observe that in states embracing the IDD the performance of VC-backed firms increases. Based on previous research, we measure the performance of VC-backed firms in terms of both the likelihood of an exit through an IPO (Lerner, 1994) and—for a limited subset of stateyear observations for which we have private data on VC investments—the average investment rate of returns (IRR) of exited investments (Mason and Harrison, 2002).23 Table 6 shows that a decision in favor of inevitable disclosure raises not only the number and the share of state VC-backed companies exiting through an IPO (columns 3 and 4) but also the mean IRR obtained by VCs when investing in the firms located in these states (column 5), consistent with our expectations.

6. Robustness checks The validity of our empirical results relies on the assumption that decisions in favor of or against inevitable disclosure are exogenous, conditional on the control variables included in the regressions. As these decisions represent court rulings about specific cases and are not aimed at enacting general policies for attracting VC, we believe this assumption might be reasonable. At any rate, we verified empirically whether states' economic and political conditions might have actually influenced the judicial rulings on inevitable disclosure cases. Appendix Table A5 reports the results of both probit and linear probability models where the dependent variables 23 We base this analysis on a sample of 1729 VC investments realized by 60 VC firms in 37 U.S. states between 1980 and 2009. The data were assembled by collecting fundraising prospectuses—documents usually referred to as private placement memoranda (PPMs)—from various investment firms. PPMs contain the performance—measured as IRR—of prior VC investments that a VC firm has made. Data regarding the state of incorporation of the start-ups were collected from VentureXpert.

534

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

are dummies equal to 1 when a state court precedent is in favor (columns 1 and 3) or against (columns 2 and 4) inevitable disclosure, and the covariates are VC investment per start-up to check for possible reverse causality, state GDP to account for economic conditions, the presence of lobbying activity in the state—measured, as in Sobel (2008), considering the presence of political establishments in SIC code 8650—and a dummy variable indicating whether the state displays a Red (i.e., Republican) political orientation, as proxied by the last presidential election results (Pe'Er and Gottschalg, 2011). The coefficients of these variables reveal that they have no significant impact on the enactment of judicial reforms. These findings certainly do not completely rule out the possibility that a decision in favor of or against the IDD is driven by some interest groups (e.g., Grossman and Helpman, 2002; Peltzman, 1976), but—at least— they alleviate any concern about the endogeneity of the inevitable disclosure rulings. A further challenge to the diff-in-diff approach is that differential changes between the states in favor of inevitable disclosure and the other states may be determined by pre-ruling differences in the time trend of dependent variables. Figs. 1 and 2 provide preliminary support to the idea that treated and control states display similar trends before treatment. However, to better tackle this issue, we constructed a dynamic diff-in-diff model employing a set of dummies that measure the distance in years from a ruling in favor of inevitable disclosure. Results, reported in Appendix Table A6, show that the coefficients prior to change in favorable rulings are small and statistically insignificant. Moreover, all post-change coefficients but one are positive and significant at conventional levels and are highest in the two years immediately following the decision. This is consistent with the evidence of Figs. 1 and 2, and supports the idea that the main factor driving the increase in VC investment is exactly the inevitable disclosure ruling rather than other unobserved changes. Another potential concern is that our findings might be driven by just one state where a court has established a precedent in favor of the IDD. To address this concern, we perform a series of leave-one-out analyses. In more detail, we analyze the impact of a favorable rule on the amount of VC investments leaving out of the sample one of the states in favor of inevitable disclosure: the results (available upon requests) do not substantially change. To further rule that our results might be driven by specific characteristics of the states' ruling on inevitable disclosure—with respect to the not-ruling states—we re-run our analysis on just the subsample of states that ruled in favor of or against inevitable disclosure and find that the results remain robust—(results available upon request). We also run tests to evaluate how robust the models are to different codifications of the ruling in New York, which embraced the doctrine in 1997 but then seemingly switched its position on inevitable disclosure in 2003. As a result of this change, scholars disagreed about whether New York should be considered to be in favor of or against inevitable disclosure (e.g., Malsberger, 2011; Milgrim and Bensen, 2015; Quinto and Singer, 2009): for this reason, in the main analysis we considered New York as a state in favor of inevitable disclosure from 1997 to 2003, and with no clear rule after 2003. However, as a robustness check, we re-run our analysis classifying New York either as a state always in favor of inevitable disclosure or as a state in favor from 1997 to 2003 and against after 2003. Results (available upon request) show that our results are robust to the use of different codifications. 7. Discussion and conclusion Most of the literature on IPR and VC has focused on patents (Hsu and Ziedonis, 2013; Mann and Sager, 2007). However, far less is known about the impact of trade secret protection on VC. In this paper, we focus specifically on the IDD, a strong form of trade secret protection. By exploiting a longitudinal variation in inevitable disclosure rule in the United States, we show that a rule in favor of inevitable disclosure increases VC investment per start-up in a state by about 27%. This result is driven by an increase in the number of companies receiving VC financing, mainly in industries where trade secret protection (vis-à-vis patent protection) is more important and in states where non-competes are less enforceable. We believe this research contributes to existing literature in several ways. First, it contributes to the literature on the impact of institutions on investment (Lerner and Tåg, 2013; Pe'Er and Gottschalg, 2011; Taussig and Delios, 2015). In particular, it suggests that IPR protection not only reduces the risk of mobility-related knowledge leakages (Hsu and Ziedonis, 2013; Mann and Sager, 2007) but also increases the predictability of court decisions on IPR cases—which resonates with the Coasian argument that an institutional environment providing a clear definition of property rights always leads to the socially efficient outcome regardless of the initial allocation of those rights (Coase, 1960). Second, this work extends the literature on the role of trade secrets as an important IPR protection mechanism that affects entrepreneurial ecosystems. Previous literature has already shown that trade secret protection increases firm profits and stimulates clustering (Fosfuri and Rønde, 2004), encourages R&D investment (Png, 2015), and decreases labor mobility (Png and Samila, 2015). In this study, we find that the protection of trade secrets through the adoption of the IDD plays a role in attracting VC investment, reinforcing the idea that a form of trade secret protection, namely that of inevitable disclosure, may play an important role in the entrepreneurial environment of a geographic area and, ultimately, in its economic growth. Third, our research contributes to the literature on the effect of employer-friendly labor regulations on innovation and entrepreneurship. Some studies propose that the enforceability of non-compete agreements affect entrepreneurship negatively (Gilson, 1999; Stuart and Sorenson, 2003; Samila and Sorenson, 2011b). However, other studies argue that non-compete enforceability might have a positive impact on the overall entrepreneurial ecosystem. On one side, non-competition agreements might induce established corporations to be more entrepreneurial and invest in riskier R&D projects (Conti, 2014); on the other side, non-competes might exert a “screening effect” on the formation of new firms, which, when the enforceability of non-competes increase, start and stay larger, are founded by higherearners and are more likely to survive (Starr et al., forthcoming.). In this regard, we show that the IDD—which, similar to noncompetes, might severely limit employee mobility—increases the level of VC investment per start-up, which is seen as an important instrument to support entrepreneurial growth firms. Taken together, these findings suggest that legal means limiting employee mobility (such as inevitable disclosure and non-competes) may have a more nuanced impact than it is often assumed on the development of the overall economy.

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

535

As with any empirical study, this work has limitations. First, a shortcoming of our study is that we were unable to sort out whether the positive relationship between a favorable rule on inevitable disclosure and VC investments was due mainly to an increase in the probability of obtaining a court injunction limiting employee mobility or to a decrease in the predictability of such an injunction. Future research, such as a survey of VC investors, should seek to disentangle these two mechanisms and their effects on VC investment. Second, our study has considered the U.S. VC industry only. Hence, our results might not extend to other contexts where, based on previous studies, VCs tend to have a lower ability to select high-quality companies (e.g., Colombo and Murtinu, 2016; Croce et al., 2013; Cumming and Murtinu, 2016; Cumming et al., 2014; Grilli and Murtinu, 2014). Future research should therefore verify whether our results hold when considering the effect of a stronger trade secret protection on VC investments in other countries. Finally, we did not consider how the efficiency of the judicial system—whose positive impact on entrepreneurship has already been established (Chemin, 2009)—influences the relationship between a decision in favor of inevitable disclosure and VC investments. In this respect, future research might analyze whether the effect of a decision in favor of the IDD on VC investment is stronger in those states where the judicial system is more efficient and effective in enforcing precedents. Despite these limitations, our study might have important managerial and policy-making implications. With a better understanding of how to attract VC, entrepreneurs might pursue pertinent strategies for reducing the risk of losing valuable knowledge due to employee departure and, in so doing, receive more funding. For instance, entrepreneurs who live in countries and/or regions where formal institutions such as the IDD or non-competes are not effective should seek to leverage social ties to decrease the possibility of employee departure. From a policy-maker perspective, we show whether and to what extent the enactment of laws protecting trade secrets by regulating employee mobility might affect the entrepreneurial ecosystem. In this respect, future studies should attempt to capture the effect of inevitable disclosure on the overall economic performance of a region. Assessing the social desirability of inevitable disclosure is an important topic we leave for future research. Appendix A

Table A1 Operationalization of variables. Variable

Operationalization

VC investment per start-up

The amount of VC equity invested in each state (by location of investee company), in millions of 2009 USD, divided by the number of new firms in a state. Sources: Thomson Reuters' VentureXpert; Business Dynamics Statistics. The amount of VC equity invested in each state (by location of investee company), in millions of 2009 USD. Source: Thomson Reuters' VentureXpert. The number of new firms in a state. Source: Business Dynamics Statistics. Dummy equal to 1 if a state has ruled in favor of inevitable disclosure. Sources: (Kahnke et al., 2008; Klasa et al., 2014; Malsberger, 2011; Milgrim and Bensen, 2015; Png and Samila, 2015; Quinto and Singer, 2009; Wiesner, 2012). Dummy equal to 1 if a state has ruled against inevitable disclosure. Sources: (Kahnke et al., 2008; Klasa et al., 2014; Malsberger, 2011; Milgrim and Bensen, 2015; Png and Samila, 2015; Quinto and Singer, 2009; Wiesner, 2012). Dummy equal to 1 if a state (a) has not ruled on inevitable disclosure, (b) has ruled but without clarifying its position, (c) has changed its position over time. Sources: (Kahnke et al., 2008; Klasa et al., 2014; Malsberger, 2011; Milgrim and Bensen, 2015; Png and Samila, 2015; Quinto and Singer, 2009; Wiesner, 2012). Share of establishment in any industry defined at the SIC 1-digit level. Source: County Business Patterns Number of political establishments in SIC 8650. Source: County Business Patterns State revenues by individual income tax divided by state GDP. Source: Census Bureau Dummy equal to 1 if a state has enacted the UTSA. Source: (Castellaneta et al., 2016). Annual state gross domestic product, in hundred billions of 2009 USD. Source: BEA. Non-compete agreement enforceability index (on a scale from 0 to 1), where 0 equals no enforcement and 1 equals highest possible level of enforcement. Sources: (Bird and Knopf, 2014; Garmaise, 2011). Dummy equal to 1 if a state has voted for a Republican presidential candidate in the last presidential election. Source: www.uselectionatalas.org The average Investment Rate of Returns (IRR) of exited investments. Source: Proprietary dataset Number of investments exited through an IPO. Source: Thomson Reuters' VentureXpert. Number of investments exited through an IPO divided by the number of VC-backed firms. Source: Thomson Reuters' VentureXpert.

VC investment Number of Start-ups Favorable rule

Against rule

No clear rule

Share of firms in any SIC 1-digit industry Lobbying Stock option tax rate UTSA enactment State GDP Non-compete enforceability index

Presidential election (red) Internal Rate of Return Number of IPOs Number of IPOs divided by VC-backed firms

536

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

Table A2 Yearly average VC equity investment (in millions of 2009 USD) per start-up, by state. State

1980–1989

1990–1999

2000–2012

State

1980–1989

1990–1999

2000–2012

Alabama Alaska Arizona Arkansas California Colorado Connecticut Delaware District of Columbia Florida Georgia Hawaii Idaho Illinois Indiana Iowa Kansas Kentucky Louisiana Maine Maryland Massachusetts Michigan Minnesota Mississippi Missouri

0.002 0.000 0.006 0.000 0.032 0.016 0.016 0.001 0.010 0.002 0.007 0.000 0.001 0.004 0.002 0.001 0.001 0.001 0.001 0.004 0.007 0.050 0.003 0.008 0.000 0.001

0.007 0.000 0.016 0.002 0.110 0.051 0.054 0.005 0.074 0.014 0.025 0.003 0.003 0.022 0.006 0.003 0.003 0.005 0.008 0.006 0.030 0.152 0.007 0.027 0.010 0.018

0.011 0.000 0.026 0.002 0.269 0.093 0.079 0.020 0.111 0.017 0.042 0.017 0.005 0.041 0.015 0.006 0.015 0.011 0.004 0.012 0.093 0.414 0.014 0.047 0.003 0.016

Montana Nebraska Nevada New Hampshire New Jersey New Mexico New York North Carolina North Dakota Ohio Oklahoma Oregon Pennsylvania Rhode Island South Carolina South Dakota Tennessee Texas Utah Vermont Virginia Washington West Virginia Wisconsin Wyoming

0.001 0.000 0.001 0.008 0.009 0.002 0.004 0.002 0.002 0.004 0.002 0.013 0.005 0.007 0.002 0.000 0.007 0.009 0.006 0.002 0.005 0.008 0.001 0.002 0.000

0.001 0.006 0.004 0.031 0.025 0.003 0.025 0.019 0.002 0.012 0.006 0.018 0.030 0.009 0.011 0.000 0.018 0.027 0.021 0.004 0.039 0.047 0.001 0.006 0.000

0.004 0.012 0.006 0.091 0.070 0.015 0.055 0.042 0.003 0.023 0.006 0.033 0.059 0.041 0.012 0.002 0.019 0.058 0.054 0.023 0.067 0.086 0.005 0.013 0.001

Table A3 Correlations.

1. VC investment per start-up 2. Overall VC investment 3. Number of start-ups 4. Favorable rule 5. Against rule 6. No decision rule 7. State GDP 8. Lobbying 9. Stock option tax rate 10. UTSA enactment 11. Non-compete enforceability 12. Internal Rate of Return 13. Number of IPOs 14. Number of IPOs divided by VC-backed firms

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

1.00 0.66 0.23 0.09 0.31 −0.32 0.35 0.32 0.18 0.02 −0.03 0.05 0.38 −0.08

1.00 0.49 0.04 0.31 −0.30 0.60 0.46 0.11 0.02 −0.19 0.00 0.62 −0.04

1.00 0.05 0.38 −0.36 0.90 0.75 0.01 −0.15 −0.11 −0.03 0.67 0.11

1.00 −0.08 −0.48 0.12 0.17 −0.02 0.12 0.11 0.03 −0.00 −0.07

1.00 −0.84 0.40 0.29 0.14 0.04 −0.12 0.04 0.39 −0.01

1.00 −0.42 −0.34 −0.12 −0.10 0.05 −0.06 −0.34 0.05

1.00 0.86 0.10 −0.06 −0.14 −0.01 0.61 −0.02

1.00 0.08 −0.01 −0.00 −0.01 0.50 −0.05

1.00 0.15 −0.03 0.09 0.11 −0.09

1.00 −0.05 −0.01 −0.05 −0.24

1.00 0.00 −0.24 0.02

1.00 0.00 −0.03

1.00 0.15

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

537

Table A4 The impact of inevitable disclosure on VC investment per start-up three, four and five years before and after a decision in favor of the IDD. Variable

(1) VC investment per start-up (plus/minus 3 years)

(2) VC investment per start-up (plus/minus 4 years)

(3) VC investment per start-up (plus/minus 5 years)

Favorable rule

0.252⁎ (0.140) −0.001 (0.167) 0.038 (0.031) Yes Yes 950 −70.806

0.266⁎⁎ (0.136) 0.014⁎⁎⁎

0.246⁎ (0.130) −0.021 (0.138) 0.040 (0.029) Yes Yes 1050 −74.057

Against rule State GDP State fixed effects Year fixed effects Observations Log likelihood

(0.148) 0.040 (0.030) Yes Yes 1000 −72.422

Notes. Poisson regression results in columns (1)–(3). Robust standard errors clustered by state in parentheses. The dependent variable is the amount of VC investment per start-up. Only states for which there are at least three years of observations after a decision in favor of the IDD are kept. The unit of observation is the state-year, and the data cover the 50 U.S. states (and the District of Columbia) from 1980 to 2012. ⁎ p b 0.1. ⁎⁎ p b 0.05. ⁎⁎⁎ p b 0.01.

Table A5 Determinants of rulings on inevitable disclosure: probit and OLS regressions. Variable

(1) Favorable rule probit

(2) Against rule probit

(3) Favorable rule OLS

(4) Against rule OLS

Presidential election (red)

−0.579 (0.450) −8.994 (8.581) 0.042 (0.111) 0.006 (0.005) Yes Yes 242 −24.691

−0.222 (0.403) 0.076 (2.322) 0.062 (0.135) −0.001 (0.008) Yes Yes 230 −23.565

−0.006 (0.005) −0.014 (0.016) −0.001 (0.003) 0.000 (0.000) Yes Yes 1559

0.003 (0.005) −0.010 (0.039) 0.002 (0.002) 0.000 (0.000) Yes Yes 1467

0.031

0.023

VC investment per start-up State GDP Lobbying organizations State fixed effects Year fixed effects Observations Log likelihood R-squared

Notes. Probit regression in models (1) and (2). OLS regression results in columns (3) and (4). Robust standard errors clustered by state in parentheses. In all models the dependent variable is equal to 1 in the year when a precedent against inevitable disclosure is established and 0 otherwise. The unit of observation is the state-year, and the data cover the 50 U.S. states (and the District of Columbia) from 1980 to 2012. State-year observations after a change is implemented are discarded.

538

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541 Table A6 Effects of inevitable disclosure on VC investment prior to and after rulings. Variable

(1) VC investment per start-up

Favorable rule (=t − 4)

0.047 (0.215) −0.025⁎⁎ (0.223) −0.004⁎⁎⁎

Favorable rule (=t − 3) Favorable rule (=t − 2)

Favorable rule (=t + 1)

(0.184) 0.053 (0.160) −0.057 (0.124) 0.164⁎

Favorable rule (=t + 2)

(0.091) 0.249⁎

Favorable rule (=t − 1) Favorable rule (=t)

Favorable rule (=t + 3) Favorable rule (N = t + 4) Against rule State GDP State fixed effects Year fixed effects Number of states Log likelihood

(0.134) 0.086 (0.134) 0.199⁎ (0.116) 0.024 (0.095) 0.028 (0.019) Yes Yes 51 −91.522

Notes. Poisson regression results in column (1). Robust standard errors in parentheses. Disturbances are clustered by state. Coefficients of “Favorable rule” are relative to the period five years or earlier before a precedent in favor of the IDD. In model (1), the dependent variable is VC investment per start-up. The unit of observation is the state-year, and the data cover the 50 U.S. states (and the District of Columbia) from 1980 to 2012. ⁎ p b 0.1. ⁎⁎ p b 0.05. ⁎⁎⁎ p b 0.01.

Appendix B. Criteria to codify court decisions B.1. Decisions in favor We defined a favorable decision on inevitable disclosure to be a decision that recognized the applicability of the doctrine for preventing an employee from moving to a rival without requiring an accompanying non-compete agreement. For example, the prominent Illinois case PepsiCo, Inc. v. Redmond is a clear instance of a favorable decision on inevitable disclosure.24 William Redmond Jr. sought to leave PepsiCo for Quaker, a competitor, but was prohibited from doing so on the basis that his new employment would “inevitably lead him to rely on the plaintiff’s [PepsiCo’s] trade secrets.” B.2. Decision against We defined an against decision as a decision that clearly rejected the possibility of restricting employee mobility to a competitor based on the IDD, where there is also the absence of a non-compete agreement and actual misappropriation. For instance, in the Louisiana case Standard Brands, Inc. v. Walter T. Zumpe et al., Standards Brands sought to enjoin a former employee from working for a competitor in the coffee and tea business.25 The court decided that “absent disclosure or imminent threat of disclosure, injunction should not be granted.” B.3. Unclear decision We define an unclear decision—the equivalent of having no rule—as a decision that may acknowledge the existence of the doctrine but that does not clarify its scope and conditions of applicability. An example of unclear decision is the Connecticut case Branson Ultrasonics Corp. v. Stratman.26 The court accepted the IDD only as a reinforcement of a non-compete agreement but did not clarify whether the IDD would have been applied in the absence of a non-compete.

24 25 26

PepsiCo, Inc. v. Redmond, 54 F.3d 1262, 1272 (7th Cir. 1995). Standard Brands, Inc. v. Zumpe et al., 264 F. Supp. 254 (E.D. La. 1967). Branson Ultrasonics Corp. v. Stratman, 921 F. Supp. 909 (D. Conn. 1996).

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

539

In general, once a state precedent has been set, state courts have been consistently applied that precedent; the only exception is New York, where state courts initially adopted the doctrine in 1997 but rejected it in 2003. Hence, we consider New York to be a state in favor of inevitable disclosure from 1997 to 2003, and as having no clear position thereafter. Table B1 Precedents in favor and against inevitable disclosure. Sources: (Kahnke et al., 2008; Klasa et al., 2014; Malsberger, 2011; Milgrim and Bensen, 2015; Png and Samila, 2015; Quinto and Singer, 2009; Wiesner, 2012). State

Year

Case

Decision

California

1944 Continental Car-Na-Var Corp. v. Moseley, 24 Cal. 2d 104, 107, 148 P.2d 9, 11 (1944)

Against

Delaware

2006

Favorable

Florida

2001

Illinois

1995

Iowa

2002

Louisiana

1967

Maryland

2004

Massachusetts

1995

Minnesota

1992

New Jersey

1980

New York

1997

New York

2003

North Carolina

1976

Pennsylvania

2010

Utah

1998

Virginia

1999

Washington

1997

“The mere fact that the defendant knew the formulae for plaintiff's products when he left plaintiff's employ … was not sufficient evidence from which an inference could be drawn that he was using or intended to use such formulae on behalf of defendant.” W.L. Gore & Associates, Inc. v. The court, relying on many decisions from other states, determined that even in the Huey-Shen Wu, et al. C.A. No. absence of a current covenant not to compete, the court had the authority to “limit 263-N (Del. Ch. 2006) a defendant working in a particular field if his doing so poses a substantial risk of the inevitable disclosure of trade secrets.” Del Monte Fresh Produce Co. v. “A court should not allow a plaintiff to use inevitable disclosure as an after-the-fact Dole Food Co., 148 F. Supp. 2d non-compete agreement to enjoin an employee from working for the employer of 1326 (S.D. Fla. 2001). his or her choice.” PepsiCo, Inc. v. Redmond, 54 F.3d “We affirm the district court's order enjoining Redmond from assuming his 1262, 1272 (7th Cir. 1995). responsibilities at Quaker through May, 1995, and preventing him forever from disclosing PCNA trade secrets and confidential information.” Barilla America, Inc. v. Wright, No. “The Court will also craft the injunction broadly, enjoining Wright from taking 4-02-CV-90267, 2002 WL. any position in the pasta industry, so as to prevent any incentive, financial or 31165069 (S.D. Iowa Jul. 5, 2002). otherwise, to disclose trade secret information.” Standard Brands, Inc. v. Zumpe The court stated, “while it does not appear here that the disclosure of et al., 264 F. Supp. 254 (E.D. La. confidential information by [the defendant] will inevitably result from his 1967). employment by [a competitor], even if this were the consequence, no remedy could be afforded.” The court cited Louisiana's statutory prohibition on non-compete agreements and strong public policy of free labor. “The chief ill in the covenant not to compete imposed by the inevitable LeJeune v. Coin Acceptors, Inc., 849 A.2d 451, 471 (Md. 2004). disclosure doctrine is in its after-the-fact nature: The covenant is imposed after the employment contract is made and therefore alters the employment relationship without the employee's consent.” Campbell Soup Co. v. Giles 47 F.3d The court states that an injunction blocking the employee cannot be granted, 467, 472 (1st Cir. 1995). since “the public interest tilted in Giles' [the defendant] favor, especially given the absence of a non-competition agreement.” International Business Machine “In the absence of a covenant not to compete or a finding of actual or an intent Corp. v. Seagate Technology Inc. to disclose trade secrets, employees ‘may pursue their chosen field of endeavor in direct competition’ with their prior employer.” 941 F. Supp. 98 (D. Minn. 1992). Continental Group, Inc. v. Amoco “Risk of harm if information is inadvertently disclosed, however, is not Chem. Corp., 614 F.2d 351, 359 (3d sufficient to satisfy the standard for granting a preliminary injunction.” Cir. 1980). DoubleClick, Inc. v. Henderson, No. “Defendants are enjoined, for a period of six months from the date of this 116914/97, 1997 N.Y. Misc. Lexis opinion, from launching any company, or taking employment with any 577 (Sup. Ct. N.Y. Co. Nov. 7, 1997). company, which competes with DoubleClick.” Marietta Corp. v. Fairhurst, 301 The court states that the inevitable disclosure doctrine should be applied in A.D.2d 734 (N.Y. App. Div. 2003). “rare instances,” suggesting a decision in favor of inevitable disclosure. However, the court also specifies that “the doctrine of inevitable disclosure is disfavored […] absent evidence of actual misappropriation by an employee.” Travenol Labs., Inc. v. Turner, 228 “North Carolina courts have never enjoined an employee from working for a S.E.2d 478, 483 (N.C. Ct. App. 1976). competitor merely to prevent disclosure of confidential information. We approve and affirm only that part of the preliminary injunction which enjoins the defendant Turner from revealing, and the defendant Cutter from seeking to obtain any confidential information concerning the modification of the Westphalia centrifuge by plaintiff Travenol.” Bimbo Bakeries USA Inc. v. “We are satisfied that there is a substantial likelihood that Defendant will not be Botticella, No. 10-cv-00! 4 (E.D. able to perform his duties at Hostess and will not perform those duties without Penn. Feb. 9, 2010). disclosing, whether intentionally or inadvertently, Bimbo’s trade secrets.” Novell, Inc. v. Timpanogos “I have found that it is inevitable that defendants will traffic upon Novell’s Research Group, Inc., 46 U.S.P.Q.2d trade secrets and confidential technical information unless they are restrained 1197 (Utah Dist. Ct. 1998). from being in the same business Novell is in.” Government Technology Services, The court stated that Virginia does not recognize the inevitable disclosure doctrine Inc. v. Intellisys Technology Corp., 51 because the “mere knowledge of a trade secret is insufficient to support an Va. Cir. 55 (Va. Cir. Ct. Oct. 20, 1999). injunction order.” Solutec Corp, Inc. v. Agnew, 1997 The courts recognize that in some situations a broad injunction for a limited period is the only effective remedy to a threatened trade secrets violation, because “[a] WL 794496, 8 (Wash. Ct. App.). narrow injunction which would allow [Mr. Agnew and Mr. Ingle] to make competitive apple waxes would be extremely difficult to police, and would be unduly burdensome and expensive, in order to determine whether [Solutec's] trade secrets were used.”

State rule after the decision

Against

Favorable

Favorable

Against

Against

Against

Against

Against

Favorable

Not clear

Against

Favorable

Favorable

Against

Favorable

540

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

References Baron, J.N., Hannan, M.T., Burton, M.D., 2001. Labor pains: change in organizational models and employee turnover in young, high-tech firms. Am. J. Sociol. 106, 960–1012. Bertrand, M., Duflo, E., Mullainathan, S., 2004. How much should we trust differences-in-differences estimates? Q. J. Econ. 119, 249–275. Besen, S., Raskind, L., 1991. An introduction to the law and economics of intellectual property. J. Econ. Perspect. 5, 3–27. Bickley, J.M., 2012. Employee Stock Options: Tax Treatment and Tax Issues. Bird, R.C., Knopf, J.D., 2014. The Impact of Local Knowledge on Banking. Bittlingmayer, G., 2000. Regulatory uncertainty and investment: evidence from antitrust enforcement. Cato J. 20, 295–325. Bloom, N., Bond, S., Van Reenen, J., 2007. Uncertainty and investment dynamics. Rev. Econ. Stud. 74, 391–415. Cameron, C.A., Trivedi, P.K., 1998. Regression Analysis of Count Data. Cambridge University Press, Cambridge, UK. Campbell, B., Ganco, M., 2012. Who leaves, where to, and why worry? Employee mobility, entrepreneurship and effects on source firm performance. Strateg. Manag. J. 33, 65–87. Castellaneta, F., Conti, R., Kacperczyk, A.“.O.”., 2016. Money secrets: how does trade secret legal protection affect firm market value? Evidence from the Uniform Trade Secret Act. Strateg. Manag. J. Chemin, M., 2009. The impact of the judiciary on entrepreneurship: evaluation of Pakistan's “Access to Justice Programme.”. J. Public Econ. 93, 114–125. Chemmanur, T.J., Krishnan, K., Nandy, D.K., 2011. How does venture capital financing improve efficiency in private firms? A look beneath the surface. Rev. Financ. Stud. 24, 4037–4090. Coase, R.H., 1960. The problem of social cost. J. Law Econ. 16. Colombo, M.G., Grilli, L., 2005. Founders' human capital and the growth of new technology-based firms: a competence-based view. Res. Policy 34, 795–816. Colombo, M.G., Murtinu, S., 2016. Venture capital investments in Europe and portfolio firms' economic performance: independent versus corporate investors. J. Econ. Manag. Strateg. Conti, R., 2014. Do non-competition agreements lead firms to pursue risky R&D projects? Strateg. Manag. J. 35, 1230–1248. Croce, A., Martí, J., Murtinu, S., 2013. The impact of ventre capital on the productivity growth of European entrepreneurial firms:“screening”or “value added”effect? J. Bus. Ventur. 28, 489–510. Cumming, D., Murtinu, S., 2016. On the Efficiency of Bank-affiliated Venture Capital. Cumming, D., Schmidt, D., Walz, U., 2010. Legality and venture capital governance around the world. J. Bus. Ventur. 25, 54–72. Cumming, D.J., Grilli, L., Murtinu, S., 2014. Governmental and independent venture capital investments in Europe: a firm-level performance analysis. J. Corp. Financ. David, P.A., 1993. Intellectual property institutions and the panda's thumb: patents, copyrights, and trade secrets in economic theory and history. Global Dimensions of Intellectual Property Rights in Science and Technology. National Research Council, Washington D.C., pp. 19–64. Dixit, A., Pindyck, R., 1994. Investment Under Uncertainty. Princeton University Press, Princeton, NJ. Fosfuri, A., Rønde, T., 2004. High-tech clusters, technology spillovers, and trade secret laws. Int. J. Ind. Organ. 22, 45–65. Gao, H., Ma, Y., 2016. Human capital driven acquisition: evidence from the inevitable disclosure doctrine (Available at SSRN 2713600). Garmaise, M.J., 2011. Ties that truly bind: noncompetition agreements, executive compensation, and firm investment. J. Law Econ. Org. 27, 376–425. Gilson, R.J., 1999. The legal infrastructure of high technology industrial districts: Silicon Valley, Route 128, and covenants not to compete. N. Y. Univ. Law Rev. 74, 575–629. Gompers, P., Lerner, J., 2001. The venture capital revolution. J. Econ. Perspect. 15, 145–168. Gompers, P.A., Lerner, J., 1999. The Venture Capital Cycle. MIT Press, Cambridge, MA, MA and London. Graves, C., 2006. Law of negative knowledge: a critique. Tex. Intell. Prop. Law J. 15, 387–416. Greif, A., 1993. Contract enforceability and economic institutions in early trade: the Maghribi traders' coalition. Am. Econ. Rev. 83, 525–548. Grilli, L., Murtinu, S., 2014. Government, venture capital and the growth of European high-tech entrepreneurial firms. Res. Policy 43, 1523–1543. Grilli, L., Murtinu, S., 2015. New technology-based firms in Europe: market penetration, public venture capital, and timing of investment. Ind. Corp. Chang. 24, 1109–1148. Grossman, G., Helpman, E., 2002. Special Interest Politics. MIT Press, Cambridge, MA. Halligan, R., 2008. Protection of US trade secret assets: critical amendments to the economic Espionage Act of 1996. J. Marshall Rev. Intell. Prop. L. 7, 656–681. Hart, H.L.A., 2012. The Concept of Law. Oxford University Press, Oxford. Hellmann, T., Puri, M., 2002. Venture capital and the professionalization of start-up firms: empirical evidence. J. Finance 57, 169–197. Henrekson, M., Sanandaji, T., 2014. Small business activity does not measure entrepreneurship. Proc. Natl. Acad. Sci. U. S. A. 111, 1760–1765. Horwitz, M.J., 1977. The Transformation of American Law, 1780–1860. Harvard University Press, Cambridge, MA. Hsu, D., Ziedonis, R., 2013. Resources as dual sources of advantage: implications for valuing entrepreneurial-firm patents. Strateg. Manag. J. 34, 761–781. Jeng, L.A., Wells, P.C., 2000. The determinants of venture capital funding: evidence across countries. J. Corp. Financ. 6, 241–289. Kahnke, R.E.R., Bundy, K.K.L., Liebman, K.A.K., 2008. Doctrine of Inevitable Disclosure. Faegre Benson LLP. Kaplan, S., Strömberg, P., 2003. Financial contracting theory meets the real world: an empirical analysis of venture capital contracts. Rev. Econ. Stud. 70, 281–315. Keuschnigg, C., Nielsen, S., 2003. Tax policy, venture capital, and entrepreneurship. J. Public Econ. 87, 175–203. Kitch, E., 1980. The law and economics of rights in valuable information. J. Leg. Stud. 9, 683–724. Klasa, S., Ortiz-Molina, H., Serfling, M.A., Srinivasan, S., 2014. Protection of Trade Secrets and Capital Structure Decisions. Landes, W., Posner, R., 1976. Legal precedent: a theoretical and empirical analysis. J. Law Econ. 19, 249–307. Lerner, J., 1994. Venture capitalists and the decision to go public. J. Financ. Econ. 35, 293–316. Lerner, J., Tåg, J., 2013. Institutions and venture capital. Ind. Corp. Chang. 22, 153–182. Lowry, S., 1988. Inevitable disclosure trade secret disputes: dissolutions of concurrent property interests. Stanford Law Rev. 40, 519–544. Malesky, E.J., Samphantharak, K., 2008. Predictable corruption and firm investment: evidence from a natural experiment and survey of Cambodian entrepreneurs. Q. J. Polit. Sci. 3, 227–267. Malsberger, B., 2011. Trade Secrets: A State-by-state Survey. fourth ed. Bloomberg BNA, Arlington, VA. Mann, R.J., Sager, T.W., 2007. Patents, venture capital, and software start-ups. Res. Policy 36, 193–208. Mason, C.M., Harrison, R.T., 2002. Is it worth it? The rates of return from informal venture capital investments. J. Bus. Ventur. 17, 211–236. Milgrim, R.M., Bensen, E.E., 2015. In: Milgram, RM (Ed.), Use of agreements to protect trade secrets in the employment relationship. In Milgrim on Trade Secrets Vol. 2. Bender, New York, p. 2. Mulcahy, J., Tassin, J., 2003. Is PepsiCo the Choice of the Next Generation: The Inevitable Disclosure Doctrine and Its Place in New York Jurisprudence. 21. Hofstra Lab. Emp. LJ, pp. 2–59. Pe'Er, A., Gottschalg, O., 2011. Red and blue: the relationship between the institutional context and the performance of leveraged buyout investments. Strateg. Manag. J. 32, 1356–1367. Peltzman, S., 1976. Toward a more general theory of regulation. J. Law Econ. 19, 245–248. Png, I., 2015. Law and innovation: evidence from state trade secrets laws. Rev. Econ. Stat. Png, I., Samila, S., 2015. Trade Secrets Law and Mobility: Evidence from Inevitable Disclosure. Qiu, B., Wang, T., 2015. Does Knowledge Protection Benefit Shareholders? Evidence from stock market reaction and firm investment in knowledge assets (econfin.massey.ac.nz.). Quinto, D., Singer, S., 2009. Trade Secrets. Oxford University Press, New York. Risch, M., 2007. Why Do We Have Trade Secrets? Marq. Intell. Prop. Law Rev. 11, 1–76. Samila, S., Sorenson, O., 2011a. Venture capital, entrepreneurship, and economic growth. Rev. Econ. Stat. 93, 338–349.

F. Castellaneta et al. / Journal of Business Venturing 31 (2016) 524–541

541

Samila, S., Sorenson, O., 2011b. Noncompete covenants: incentives to innovate or impediments to growth. Manag. Sci. 57, 425–438. Silva, J.M.C.S., Tenreyro, S., 2006. The log of gravity. Rev. Econ. Stat. 88, 641–658. Sobel, R.S., 2008. Testing Baumol: institutional quality and the productivity of entrepreneurship. J. Bus. Ventur. 23, 641–655. Sorensen, M., 2007. How smart is smart money? A two-sided matching model of venture capital. J. Financ. 62, 2725–2762. Starr, E. P., Balasubramanian, N., Sakakibara, M., n.d. Screening Spinouts? How noncompete enforceability affects the creation, growth, and survival of new firms. Manag. Sci. (forthcoming). Stuart, T., Sorenson, O., 2003. The geography of opportunity: spatial heterogeneity in founding rates and the performance of biotechnology firms. Res. Policy 32, 229–253. Taussig, M., Delios, A., 2015. Unbundling the effects of institutions on firm resources: the contingent value of being local in emerging economy private equity. Strateg. Manag. J. 36, 1845–1865. Wiesner, R.M., 2012. A state-by-state analysis of inevitable disclosure: a need for uniformity and a workable standard. Marq. Intell. Prop. Law Rev. 16, 211–230.