The tortoise, the hare, and the hybrid: effects of prior ...maryannfeldman.web.unc.edu/files/2019/07/The-tortoise-the-hare-an… · The tortoise, the hare, and the hybrid: effects

The tortoise, the hare, and the hybrid: effects of

prior employment on the development of an

entrepreneurial ecosystem

Mary Donegan,1 Allison Forbes,2 Paige Clayton,3 Alyse Polly,4

Maryann Feldman,5 and Nichola Lowe6

1Department of Urban and Community Studies, University of Connecticut, 10 Prospect Street, Hartford CT

06103. e-mail: [email protected], 2Center for Regional Economic Competitiveness, PO Box 100127,

Arlington VA 22210. e-mail: [email protected], 3Department of Public Policy, University of North Carolina,

Abernethy Hall, CB 3435, Chapel Hill NC 27599, Chapel Hill. e-mail: [email protected], 4CREATE at

Kenan Institute of Private Enterprise, University of North Carolina, Chapel Hill, Kenan Center Suite 403, CB

3440, Chapel Hill NC 27599. e-mail: [email protected], 5Kenan Institute and Department of Public Policy,

University of North Carolina, Chapel Hill, Kenan Center Suite 403, CB 3440, Chapel Hill NC 27599. e-mar-

[email protected] and 6Department of City and Regional Planning, University of North Carolina,

Chapel Hill, New East Building, CB 3140, Chapel Hill NC 27599. e-mail: [email protected]

Abstract

Prior employment imprints nascent entrepreneurs with logics for organizing startups. Within a regional

ecosystem, entrepreneurs with different employment backgrounds pursue alternative entrepreneurial

pathways, each generating distinct, though complementary, regional impacts. By analyzing diverse

pre-entrepreneurial employment experiences, no one pathway leads to superior firm performance; prior

industry experience generates strong early performance that tapers off, while prior academic experience

engenders slow, steady, long-lasting growth. Our approach is well-suited for theorizing ecosystem devel-

opment and generating policy recommendations in support of ecosystem diversity.

JEL classification: J21, L65, R11, O18, E24

1. Introduction

Entrepreneurs are key players in the establishment and growth of local ecosystems. Rather than being passive, entre-

preneurs actively engage in building networks, resources and institutions that affect dynamics within their regional

economy and, with it, the contours of the surrounding entrepreneurial ecosystem. However, entrepreneurs are highly

differentiated—not simply because they have idiosyncratic personalities or distinct skill sets, but because entrepre-

neurial decisions and aspirations are also shaped by existing industry norms. We hypothesize that entrepreneurs

draw inspiration and insights from their prior local work experience. Despite popular perceptions, new ventures rare-

ly start from scratch, but instead are formed by founders who previously worked at established firms and universities.

This suggests the need to look more closely at work experiences that precede the formation of an entrepreneurial

firm. A fuller understanding of local professional experiences prior to starting a new firm provides insights into the

VC The Author(s) 2019. Published by Oxford University Press on behalf of Associazione ICC. All rights reserved.

Industrial and Corporate Change, 2019, 1–22

doi: 10.1093/icc/dtz037

Original article

Dow

nloaded from https://academ

ic.oup.com/icc/advance-article-abstract/doi/10.1093/icc/dtz037/5526892 by U

niversity of North C

arolina at Chapel H

ill user on 18 July 2019

https://academic.oup.com/

dynamics of the regional entrepreneurial landscape, and subsequent outcomes in terms of the performance of

regions.

Prior employment imprints individuals with organizational logics and expectations that shape the ways they or-

ganize new firms (Baron et al., 2001; Burton et al., 2002). This organizational imprinting also affects how new firms

interact with their external environments—a relationship which appears to have strong localized effects (Marquis,

2003). The types of funding and mentoring opportunities that new entrepreneurs rely on is affected by their prior

experiences (Sørensen and Phillips, 2011). Furthermore, prior employment may influence entrepreneurs’ perception

of how to best connect with organizations within the regional economy, including which types of intermediary insti-

tutional supports to utilize and engage (see Clayton et al., 2018 for a review of these supports).

An entrepreneur’s prior career, including her immediately prior employment affiliation, is not just a precursor to

ecosystem engagement. We argue it defines the characteristics of the regional ecosystem, providing know-how, organ-

izational insights and cultural awareness to strengthen the talent pipeline and inform entrepreneurial development and

decision-making. Thus, within a single regional ecosystem, entrepreneurs with different backgrounds may pursue alter-

native and different pathways, each potentially generating distinct, though complementary, economic impacts.

This article simultaneously examines a range of local pre-entrepreneurial employment experiences within one

regional ecosystem in order to capture their differential contributions to the ecosystem. Our study region is the

13-county Research Triangle region of North Carolina, where we draw on a unique dataset that captures entrepre-

neurial development of the life sciences industry. Incorporating founder career histories, we are able to draw out the

mix of prior work experiences that founders bring to their new firms. This data allows us to extend previous analysis

in two important ways. First, we create an analytical framework for integrating insights from organizational manage-

ment and entrepreneurial studies, recognizing that ecosystem boundaries are not limited to entrepreneurial firms and

their obvious support institutions. Rather, elements of the entrepreneurial ecosystem include large regional employ-

ers, which offer micro-spaces for aspiring entrepreneurs to test out and incubate new ideas and foster valuable rela-

tionships and resources that can support eventual entrepreneurial development and success. Second, we conduct a

deep dive into one regional economy in order to capture and compare the varying influences of multiple types of prior

work experience on key entrepreneurial milestones and longer-term performance. In this sense, we also align insights

from organizational and regional sciences, recognizing that regions are comprised of a diverse ecology of entrepre-

neurial experiences, each potentially adding different, though complementary, economic strengths. By studying one

region, we are able to draw out intra-regional heterogeneity that is often suppressed with cross-regional comparison

(Kenny and von Burg, 1999). Our approach is therefore well suited for theorizing ecosystem development and for

generating policy recommendations in support of ecosystem diversity.

Our empirical results demonstrate that there is considerable career heterogeneity among entrepreneurial founders,

even within a single industry in a regional economy. We also find similar variation in how those prior work and career

experiences translate into entrepreneurial development and performance. Entrepreneurial firms whose founders were

once affiliated with big pharma companies do not generate substantial jobs in their first year of operation. It is not until

year three of their existence that they have significantly higher positive employment, and this trend of strong employ-

ment continues from that point forward. In contrast, firms started by a founder that gained prior work experience as

an employee at a local entrepreneurial firm consistently start with a comparatively larger numbers of employees—but

their fast start does not necessarily ensure they maintain a long-term lead. Beyond employment, we find divergent

patterns in other areas of firm performance. For example, firms founded by individuals with a hybrid of both big

pharma and entrepreneurial experiences produce a larger number of patents, whereas firms founded by academics are

more successful at raising public funding. These findings are consistent overall with those of Decker et al. (2014),

which examines the complexities of startup dynamics at the national scale and finds that despite the considerable

contribution of a small group of startups and high-growth firms to job creation, most of these firms fail, and most of

those that do survive remain small. The average age at firm closure for firms in our sample is 6.7 years, while the

average maximum employment of our sample in North Carolina over the time period studied is 16.2 employees.

The different patterns of prior work history, and their influence on growth, invoke the classic parable of the tor-

toise and the hare. Some firms, particularly those created by big pharma founders or academic founders, may be lik-

ened to tortoises, whose growth begins small but steadily surpasses that of other firms. In contrast, other firms, such

as second-generation entrepreneurial firms, act more like hares—with a notable growth spurt coming out of the start-

ing gate that then tapers over time, and they are eventually surpassed by others in size. These differences point to

specific strengths and weaknesses of firms that may be inherent, depending on the prior experience of the founders.

2 M. Donegan et al.

Dow




arolina at Chapel H


Deleted Text: ; Baron, Hannan, and Burton, 2001

Deleted Text: paper

Deleted Text: -

Deleted Text: -

For example, while firms founded by academics may benefit from the academics’ ability to raise public funding, such

firms may also be stifled by academics’ difficulty balancing university obligations. Finally, some firms exhibit features

of both the tortoise and the hare, with less discernable patterns of initial size and growth. If regional development is

to have maximum effect, recognizing these differences in how firms are created and grow is important. This article is

an effort to begin to uncover these differences.

The paper is organized as follows. We start with a review of existing literature to make the case for situating the

employment and career experiences that precede entrepreneurship within an ecosystem framework. This is followed

by a description of our database, including our methods for capturing detailed career histories of entrepreneurial

founders, and information on firm characteristics, performance and outcomes across the seven types of prior experi-

ences. The next sections include empirical models of a variety of economic outcomes to demonstrate variation among

different prior work experiences, and a discussion that includes illustrative firms from our dataset. This article con-

cludes with reflections on next steps and policy implications.

2. Entrepreneurs in their ecosystem context

Entrepreneurial ecosystems have been defined in a number of ways, though most definitions recognize some kind of

interaction between actors, institutions, and organizations within a given geographic region or place. Mason and

Brown (2014), for example, define the entrepreneurial ecosystem as “a set of interconnected entrepreneurial actors,

institutions, entrepreneurial organizations and entrepreneurial processes which formally and informally coalesce to

connect, mediate and govern the performance within the entrepreneurial environment” (p. 5). In a related definition,

Stam and Spigel (2018) call the entrepreneurial ecosystem “a set of interdependent actors and factors coordinated in

such a way that they enable productive entrepreneurship within a particular territory” (p. 407), one that remains

highly spatially concentrated despite being increasingly facilitated by digital infrastructure (Autio et al., 2018).

Ecosystems provide a conceptual umbrella for studying a community of entrepreneurs and the surrounding sup-

ports that enable new high-growth ventures to form, survive, and expand. But as Spigel and Harrison (2018) note,

studies of entrepreneurial ecosystems often generate a laundry list, which obscures causal factors. This also risks the

creation of formulaic policy recommendations that fail to recognize differences in place-specific dynamics (Feldman

and Lowe, 2018). Spigel and Harrison (2018) argue instead for greater focus on the underlying mechanisms that en-

able entrepreneurial firms to benefit from, and further strengthen, their surrounding environment. This reflects their

push for a more fluid definition of entrepreneurial ecosystems as “ongoing processes through which entrepreneurs ac-

quire resources, knowledge and support, increasing their competitive advantage and ability to scale up” (p. 158)—in

other words, a framework that treats ecosystems as dynamic, contextually dependent phenomena.

A process perspective supports cross-regional comparisons, recognizing that each region will have a unique mix

of resources and supports and thus produce different entrepreneurial opportunities and outcomes, as well as chal-

lenges. Additionally, it fosters deeper appreciation of variation within a single regional ecosystem, capturing differen-

tial levels of resource availability by local entrepreneurs but also differences in how entrepreneurial actors perceive

and engage those supports. In this respect, entrepreneurs are not passive actors that inherit a fixed regional ecosys-

tem. In the process of building their firm, entrepreneurs encounter a range of regional resources and supports and

must choose which to pursue and engage. Additionally, entrepreneurs may be in a position to build collective resour-

ces that promote their interests, engaging in consensus building and sense-making processes to further the develop-

ment of their technology (Feldman and Lowe, 2008; Lowe and Feldman, 2017). As this suggests, the actions of

entrepreneurs are endogenous to the development of ecosystem resources (Feldman et al., 2005).

But entrepreneurial actions also stem from prior experience, particularly career and employment experiences that

pre-date the formation of the firm. These previous experiences can affect essential decisions related to launching a

new firm or pursuing an entrepreneurial opportunity. Prior employment can also influence entrepreneurs’ perception

of how to participate in local ecosystems, specifically dictating their involvement with intermediary institutional sup-

ports. In this regard, it can affect the types of funding options that entrepreneurs pursue, which in turn can reflect, or

influence, decisions around innovation, firm expansion (including hiring employees), or eventual industry exit

(Sørensen and Phillips, 2011). Thus, within one ecosystem, entrepreneurs may pursue alternative entrepreneurial

pathways, and these co-existing pathways may in turn affect ecosystem growth.

While underexplored in relation to ecosystem development, there is a strong tradition in organizational manage-

ment of connecting prior work experience and entrepreneurial development. Existing scholarship in this area tends

The tortoise, the hare and the hybrid 3

Dow




arolina at Chapel H


Deleted Text: paper

Deleted Text: paper

Deleted Text: E

Deleted Text: C

to concentrate on the entrepreneurial impact of two dominant career types: prior employment either at large corpora-

tions or prominent academic institutions. These two types of previous work experiences are typically studied in isola-

tion, considered independent of one another (for an exception, see Wennberg et al., 2011). Yet within a regional

ecosystem, both experiences can coexist and can determine the orientation and vibrancy of the resulting entrepre-

neurship. Entrepreneurs previously employed in a large corporate setting often create businesses in closely related

fields to that of their former employer (Dencker et al., 2009; Roberts et al., 2011; Campbell et al., 2012). In some

cases, they might even pursue ideas and technologies that their former employers did not fully develop, either because

the ideas would have cannibalized the parent’s existing products, or because the market was estimated to be too small

(Klepper, 2001; Baker et al., 2003; Klepper and Sleeper, 2005). Academic entrepreneurs—including faculty members,

affiliated researchers and post-doctoral fellows—often generate new product and process ideas through laboratory

and applied research. In-depth analysis of academic entrepreneurs recognizes the supporting role that diverse univer-

sity resources, including law and business schools, play in developing networking, mentoring and funding, ultimately

shaping the structure those companies take (Wright et al., 2004; Donegan and Lowe, 2017). Yet academic entrepre-

neurs remain tied to their primary academic job, potentially leading to difficulties balancing that career with entre-

preneurial work—difficulties that may ultimately dampen the firm’s long-term performance (Franklin et al., 2001).

The literature also recognizes additional constraints that academic entrepreneurs face—including the need to con-

front potentially hostile campus commercialization cultures while also navigating campus TTOs and related offices

(Siegel, 2006; Siegel et al., 2007; O’Gorman et al., 2008).

A third, though comparatively less analyzed, pathway includes prior employment at other entrepreneurial firms.

While the literature has long devoted attention to serial or habitual entrepreneurs who have founded multiple firms

either sequentially or in tandem (see e.g., Westhead and Wright, 1998; Ucbasaran et al., 2010), far less attention has

been paid to how employment at other entrepreneurial firms shapes emerging entrepreneurs and their firms. Some

studies, such as Astebro and Thompson (2011), argue that workers with inherently stronger entrepreneurial ambi-

tions and abilities self-select into entrepreneurial workplaces, where they remain until they have gained the experi-

ence they need to form their own firms (see also Elfenbein et al., 2010). Other studies instead suggest that the

experience of work at an entrepreneurial firm encourages workers to found a firm, provides mentoring opportunities

and offers hands-on learning (Dobrev and Barnett, 2005). The nature of work at entrepreneurial firms requires that

workers rotate through a variety of tasks—a practice helped along by these firms’ comparatively relaxed organiza-

tional hierarchies (Elfenbein et al., 2010). Prior work experience at entrepreneurial firms allows workers to develop

this awareness and sensibility over time within a particular social and organizational setting, which in turn can mo-

tivate and inform their subsequent decision to launch their own venture (Aldrich and Kim, 2007).

Rounding out the corporate and academic pathways to entrepreneurship are firms whose founding teams com-

bine different backgrounds. Hybrid teams, and individuals within teams, can draw from, and combine, both experi-

ences. Teams can also include non-local actors that introduce perspectives from outside the local region, adding to

the mix of founders that remain local (Dahl and Sorenson, 2012).

Heterogeneity within founding teams has been a focus of scholars for some time, including sources of variation

that might stem from diverse employment backgrounds and work histories. Early research by Ensley et al. (1998)

found team heterogeneity to be a source of entrepreneurial disadvantage that negatively impacted firm performance.

Recent research has pushed this line of inquiry further. In one study, Ucbasaran et al. (2003) found the more diverse

a team, in terms of their background experience, the greater the likelihood of churning, resulting in higher rates of

team exit. These findings suggest the possibility that intra-team differences could have a destabilizing, or disruptive,

effect. But other studies have challenged claims that there are limits to team diversity. Vanaelst et al. (2006) looked

at teams that brought together a range of prior experiences, but discovered this variety had little effect on the overall

approach to doing business. Related to this, Chowdhury (2005) found that challenges associated with team diversity

can be overcome when diverse teams combine complementary skill sets and share a commitment to team effective-

ness. This suggests that diversity, per se, is not the problem. Rather, there are limits to cooperation and knowledge

sharing that can be overcome through team building.

Management scholars typically examine each respective entrepreneurial pathway independent of any others, thus

missing a textured understanding of heterogeneity across co-existing pathway types. This suggests an opportunity to

study concurrent and, at times, intersecting pathways to entrepreneurship. Situating such an analysis within a single

study region also means expanding our current understanding of entrepreneurial ecosystems. It allows us to make the

case that ecosystem connections and processes do not just originate when the firm is officially formed or first

4 M. Donegan et al.

Dow




arolina at Chapel H


Deleted Text: Campbell, Ganco, Franco, and Agarwal, 2012;

Deleted Text: Baker, Miner, and Eesley, 2003;

Deleted Text: <xref ref-type=

Deleted Text: O'Gorman, Byrne, and Pandya, 2008;

Deleted Text: for example

conceptualized. They can also be initiated during earlier, formative career experiences. By making room for these

multiple pathways into entrepreneurship, we are in a better position to consider whether, and how, these varying

routes contribute differently to overall ecosystem performance and also reflect on the broader policy implications for

intra-regional diversity. With that in mind, we turn next to a description of our study region in North Carolina.

3. Empirical setting and life sciences industry

The choice to focus on North Carolina’s Research Triangle region’s life sciences cluster reflects both the industry’s

long tenure in the region and its status in policy circles. The region’s life sciences cluster is one of the largest in the

country, anchored by the region’s three R1-ranked research universities, a long history of pharmaceutical branch

plants locating in the region, and a growing number of entrepreneurial startups. Yet, unlike Cambridge, MA, and the

San Francisco Bay Area—the country’s leading life sciences regions—the Research Triangle region was not an obvi-

ous candidate to develop a life sciences industry. While major pharmaceutical firms have had a local presence since

the 1960s, the region was initially a hub for satellite branch plants in the mid-1990s, with few corporate spawns

emerging (Luger and Goldstein, 1991; Markusen, 1996). Despite heavy involvement of the region’s research univer-

sities in life sciences research, the universities lagged in spinning out life science companies, when compared to their

peers (Bercovitz et al., 2001; Lowe and Feldman, 2008; Donegan, 2019). Two of the region’s research universities

(Duke University and the University of North Carolina at Chapel Hill) have medical schools, while the third (North

Carolina State University) has a veterinary school. North Carolina State University and the University of North

Carolina at Chapel Hill share a joint department of biomedical engineering. Furthermore, all three universities have

business schools with students and faculty involved in local entrepreneurial activity. The result is that the region lacks

a single, dominant pathway to life sciences’ entrepreneurship, instead nurturing an entrepreneurial environment

reflecting diverse career experiences in high-profile branch plants (like GlaxoSmithKline), the region’s research uni-

versities, the steadily growing number of entrepreneurial firms, or some combination thereof.

Reflecting these diverse roots, the region’s life sciences industry is comprised of a diversity of firms. The majority

of these firms focus on what can be thought of as “traditional” pharmaceutical, medical device, or life sciences activ-

ities, such as early- to mid-stage drug development, drug commercialization, medical device development, diagnostic

development, and the manufacturing of these products. We also include firms that focus on less traditional, but critic-

al, supportive functions—including the development of related software systems (e.g., drug management, virtual

health sciences laboratories), specialized law and consulting. The classification also includes a wide array of contract

research organizations (CROs), which have a strong presence in the region. This conception of the region’s industry

is thus more holistic than if it were based on industrial codes (e.g., Standard Industrial Classification or North

American Industry Classification System codes), but we believe it to be a more accurate representation of the region’s

entrepreneurial life sciences ecosystem.

The diverse work histories that life science firm founders have in the region make it an ideal place for exploring rela-

tionships between employment backgrounds and firm performance. Yet, we also believe that the selection of life scien-

ces remains relevant from a policymaking perspective, since the life sciences continue to be a prime target for

policymakers searching for sources of future regional economic growth (Feldman and Francis, 2004; Donegan, 2019).

By focusing on a second-tier region that was not an obvious candidate for the development of a life sciences industry,

our findings offer greater applicability to other regions seeking to develop industry concentrations (Mayer, 2011).

4. Data and methods

This article focuses on entrepreneurial life science firms, a sub-set of firms within the PLatform for Advancing

Community Economies (PLACE): Research Triangle database. The PLACE: Research Triangle database documents

the genesis of the Research Triangle region’s entrepreneurial ecosystem, drawing information on firms and their

founders from over 30 underlying data sources (Feldman and Lowe, 2015). We have identified the near universe of

942 entrepreneurial life science firms founded in the 13-county region between 1960 and 2016.1 We limit our sample

1 Our data triangulation methods, as described in Feldman and Lowe (2015), and our review of our firm list by local indus-

try experts allow us to have confidence in our ability to claim the near universe of entrepreneurial firms in this 13-county

region and broad industry.


Dow




arolina at Chapel H


Deleted Text: S

Deleted Text: L

Deleted Text: S

Deleted Text: I

Deleted Text: M

Deleted Text: paper

to 488 firms that were founded between 1990 and 2012, and for which we have both founder employment history

and firm employment data.2

To categorize prior work experience—our independent variables of interest—we draw on the detailed founder data con-

tained in the PLACE: Research Triangle database. The database includes records of each founder’s places of employment

prior to founding a firm, as well as the start and end dates for each employment episode. This allows us to assess whether

each founder experienced any of the three prior regional employment types that are core to the bioscience industry: (1) an aca-

demic experience at one of the region’s three large research universities (Duke University, North Carolina State University,

and the University of North Carolina Chapel Hill); (2) “big pharma” work at one of the region’s prominent pharmaceutical

corporations (GlaxoSmithKline,3 Pfizer, Dupont, Eli Lilly, and Ciba-Geigy)4; or, (3) a regional entrepreneurial experience at

another of the region’s entrepreneurial life science firms.5 To account for founders with more than one type of work experi-

ence before founding a firm, we also create a series of hybrid experiences that span multiple prior career experiences.

The 488 firms in our sample are linked to 924 founders.6 As described in Table 1, of those 924 founders, 191

have academic experience (or work experience at the three large academic employers in the region), 112 have entre-

preneurial experience (work experience at other entrepreneurial life science firms in the region), and 93 have big

Table 1. Public funding, private funding and patent publication (firm-level outcomes) by associated experience

Experience Total number

of founders

Percentage of associated firms that:

Received public fundingb Received private funding Published a patent

Academic experience 191 54.3% 22.1% 29.3%

Big pharma experience 93 24.7% 23.3% 27.4%

Entrepreneurial experience 112 33.3% 19.8% 22.9%

No experience 343 26.5% 17.0% 22.1%

Any hybrid experiencea 185 38.6% 22.1% 22.1%

Academic-big pharma 18 50.0% 5.6% 11.1%

Academic-big pharma-entrepreneurial 20 40.0% 30.0% 30.0%

Big pharma-entrepreneurial 74 28.6% 25.4% 20.6%

Academic-entrepreneurial 73 46.6% 25.9% 31.0%

Total and average 924 38.0% 21.1% 24.3%

Source: PLACE: Research Triangle database.aAny Hybrid totals the number of founders that have any type of hybrid experience.bIncludes Small Business Innovation Research (SBIR) and Small Business Technology Transfer (STTR) funding from the Small Business Administration, and state

funding.

2 The underlying employment data from the PLACE: Research Triangle database is the National Establishment Time Series

(NETS) database (Walls & Associates, 2013). Neumark et al (2011) suggest that 3-year intervals reduce much of the em-

ployment stickiness found in the NETS data. We drew our sample from PLACE: Research Triangle database in

December of 2016. As of that date the underlying data on firm survival and employment reflected the NETS 2013 data,

which includes data from 1990 to 2012. This may result in an undercount of 2009–2012 firms (Kolko & Neumark, 2007),

but we do not believe the undercount is enough to bias study findings (see also Neumark et al., 2011).3 GlaxoSmithKline also includes work histories at Burroughs Wellcome and Glaxo Wellcome.4 Following Burton et al. (2002), we chose these employers for the many new firms that were created by former

employees.5 Though some founders also have prior entrepreneurial experience at firms in different industries and/or regions, we

focus in this article only on prior experience at entrepreneurial firms in our sample so that we can directly analyze the

influence of prior same industry and region experience on entrepreneurial outcomes. For an investigation of a broader

range of entrepreneurial experiences, see Clayton et al. (2018).6 A founder is counted at the firm level, such that if a single individual founded two firms, that founder is included twice in

the count of 924 founders. This is necessary to account for temporal changes in experience served, since over time the

same founder can add experience (e.g., move from an academic experience type with one firm to a hybrid academic-

entrepreneurial experience type with a subsequent firm).

6 M. Donegan et al.

Dow




arolina at Chapel H


Deleted Text: , Wall, and Zhang

Deleted Text: three

Deleted Text: -

Deleted Text: , Wall, & Zhang

Deleted Text: , Sorensen, and Beckman

Deleted Text: paper

Deleted Text: , Feldman, & Lowe

pharma experience (work experience at the most prominent pharmaceutical corporations in the region). An addition-

al 185 firms have had some form of a hybrid work experience: 74 have experience in both big pharma corporations

and entrepreneurial firms in the region (big pharma-entrepreneurial), 73 have both local academic and local entrepre-

neurial experience (academic-entrepreneurial), and 18 have experience in both academic and big pharma settings in

the region (academic-big pharma). Finally, 20 founders have all three types of experience (academic-big pharma-

entrepreneurial) in the region. Our interest is in the development of the local entrepreneurial ecosystem: the remain-

ing 343 founders (37% of 924) have no prior employment with employers that are core to the region’s life science in-

dustry (none). Founders with no local work experience, either in our three categories or at other local employers, are

present at 27 firms (6%) in our sample. These seven work experience categories are mutually exclusive and cumula-

tively exhaustive.

We include each of the prior experiences in our empirical models. For the 269 firms founded by teams, the mean

founding team size is 2.63 founders. 219 firms are started by solo entrepreneurs. We code prior experience as a pro-

portion of total firm founders. For example, for a firm with four founders—one of whom had academic experience,

one of whom had big pharma experience, and two of whom had both big pharma and entrepreneurial experience—

the firm would have the following work experience proportion values: 0.25 for academic experience, 0.25 for big

pharma experience, and 0.50 for the hybrid big pharma-entrepreneurial experience. In our estimation, the omitted

category is founders with no local experience in the core background sectors. Our interviews reveal that founders

who move to the region are a diverse group, representing primarily team members with some prior expertise that is

significantly valuable for them to incur the costs of relocating. Our interest is in the endogenous development of the

regional ecosystem; thus, we compare founders with local experience against founders with any type of non-local

experience.

In order to understand whether different local work experiences are associated with variations in firm-level eco-

nomic performance, we consider a variety of outcomes. The primary firm-level outcomes in which we are interested

include patenting, firm survival, likelihood of experiencing a merger or acquisition, and employment. Patents indicate

a firm’s productive focus on research, development and commercialization, and are used as a measure of entrepre-

neurial firm success in research-intensive or technology-oriented industries. With the understanding that a firm’s abil-

ity to develop technology to patent, survive and generate employment depends on its ability to raise funding, we also

assess the association between experience types and the levels of private and public funding received.

Table 1 presents data on accessing funding and patenting. Private funding includes all venture capital, crowd-

sourced funding, and angel funding the firm raised over the course of its life. Public funding includes federal and state

funding. We derive two variables from this data: first, dummy variables that reflect whether a firm raised any funding

or not, and second, a continuous variable that reflects the amount of funding raised. Private funding is measured in

millions of dollars, and public funding is measured in thousands of dollars. As with funding, we include a dummy

variable to reflect whether the firm published any patents in its lifetime. We also include a count variable reflecting

the total number of patents the firm published over the course of its life. We find considerable variation among our

firms, associated with different founder work experiences. Firms with founders that have an academic experience—

either just academic, academic combined with big pharma, academic combined with entrepreneurial, or academic

combined with both big pharma and entrepreneurial experience—have a higher association with receiving public

funding. There is not as much variation in receipt of private funding. Firms with founders who have academic-big

pharma—entrepreneurial experience have the highest rate of receipt of private funding. Regarding patents, firms

with the academic-entrepreneurial experience combination—either alone or with big pharma added—are granted the

most patents. All data on funding and patents is drawn from the PLACE: Research Triangle database.

Firm survival contributes to the regional economy. A firm is considered “born” in the first year it reports employ-

ment, and is considered “dead” in the first year it does not report employment in the state. Table 2 reports how ex-

perience type is associated with time-to-closure. We do not observe a closure for any firm reporting employment in

2012; therefore, our data is right-censored. There are 122 firms (25%) in our unbalanced panel that exited the region

prior to 2012. Firms whose founders had entrepreneurial experience, on average, survive longest. Survival data is

drawn from the PLACE: Research Triangle database.

Another way to look at differential survival outcomes is to examine the type of firm exits. When a firm experien-

ces a major event, such as a merger or acquisition, its location might change, or the firm might cease to exist under

the same name. These major events have implications for measures like survival, as well as employment, since the dis-

ruption to the firm’s identity might mean it becomes a completely different type of employer, or employment may no


Dow




arolina at Chapel H


Deleted Text: -

Deleted Text: percent

longer be formally associated with the company. Also, while employment may go to zero or a firm may officially exit

(the market and the dataset), these events provide wealth to the entrepreneurs and the region. Here, the key event is

merger or acquisition, recorded as a binary variable in each year (0¼never merged or acquired, 1¼merger or acqui-

sition). Again, the model recognizes that some firms may experience a merger or acquisition, but we will not observe

the event in our data because our data is right-censored, unnaturally cut off after the year 2012. We observe merger

or acquisition events for 81 of the 488 firms in our dataset. Firms with big pharma experience are most likely to

merge with, or be acquired by, another company. Of course, the gold standard for an entrepreneurial company is an

initial public offering (IPO), and eight firms in our data experienced an IPO. Firms with big pharma experience are

most likely to experience an IPO. The level of IPOs by firms associated with an academic-big pharma-entrepreneurial

experience is high as well—there are very few of these firms (20), but they are successful in this regard.

Employment is a measure of a firm’s growth and its role in the region’s labor market, and indicates the extent to

which companies—regardless of their ultimate survival—gainfully employ a region’s labor force and contribute to

the region’s success. We measure firm employment as a count of employees. There are no 0 counts of employment,

but the data is highly skewed toward 1 and over-dispersed above 50; for example, there are 72 observations of one

Table 3. Initial and maximum employment levels, and employment growth, by associated experience


of founders

Average max

employment

per associated firm

Average initial

employment

per associated firm

Average growth from Year 1

Year 3 Year 6 Year 9

Academic experience 191 14.36 5.09 29.5% 126.7% 270.5%

Big pharma experience 93 20.32 5.40 64.8% 221.4% 176.1%

Entrepreneurial experience 112 14.67 5.72 74.0% 189.1% 174.6%

No experience 343 19.54 6.74 34.2% 125.3% 196.4%

Any hybrid experiencea 185 8.61 4.16 25.0% 98.1% 168.5%

Academic-big pharma 18 7.75 5.10 5.6% 161.4% 0%**

Academic-big pharma-entrepreneurial 20 12.54 4.29 5.1% 85.3% 256.5%

Big pharma-entrepreneurial 74 7.79 3.79 4.4% 121.0% 107.6%

Academic-entrepreneurial 73 10.19 4.30 7.5% 103.8% 174.5%

Total and average 924 13.4 5.1 28.1% 141.8% 193.7%

Source: PLACE: Research Triangle database.aAny Hybrid totals the number of founders that have any type of hybrid experience.bAll firms of this category type had the same number of employees from year one through to year 9. Hence, 0% growth.

Table 2. Merger or acquisition, moves out of state, and IPO events; age at closure and establishment date


of founders

Percentage of associated

firms with exit events:

Average age of

firms at closure

Average firm

estab. date

Merged or

acquired

Moved out

of state

IPO

Academic experience 191 18.6% 5.7% 2.1% 6.6 2003

Big pharma experience 93 30.1% 4.1% 5.5% 6.3 2002

Entrepreneurial experience 112 24.0% 1.0% 1.0% 8.1 2003

No experience 343 19.4% 3.2% 0.8% 7.0 2001

Any hybrid experiencea 185 16.7% 0.0% 0.0% 6.0 2002

Academic-big pharma 18 20.0% 0.0% 0.0% 3.3 2002

Academic-big pharma-entrepreneurial 20 25.4% 3.2% 3.2% 7.6 2003

Big pharma-entrepreneurial 74 10.3% 1.7% 1.7% 5.7 2006

Academic-entrepreneurial 73 18.6% 2.1% 2.1% 5.9 2004

Total and average 924 20.8% 2.6% 2.1% 6.7 2003

Source: PLACE: Research Triangle database.aAny Hybrid totals the number of founders that have any type of hybrid experience.

8 M. Donegan et al.

Dow




arolina at Chapel H


Deleted Text:

employee in year three. Table 3 examines initial and maximum firm-level counts of employment, and growth rates

from year one to years three, six, and nine. The data suggests that no single type of firm performs consistently better

across all time periods.

Our models include a series of control variables. First, number of founders may be related to the number of

employees in each firm. The number of founders may also reflect the potential interactions between various founder

networks and available resources, increasing the likelihood of gaining access to funding, patenting and survival. We

therefore include founding team size as a control variable in all of our models. Firm development milestones are

accomplished over time, suggesting that the total number of years a firm has been in existence may positively influ-

ence certain outcomes. We therefore include a variable for age. Within the life sciences industry, entrepreneurial firms

vary by technology focus. The Research Triangle life science industry is highly specialized in human therapeutics. We

include a human therapeutics dummy in our funding and patenting models, since we expect that firms focusing on

this technology will patent more and raise more funding. We do not include this dummy in our models of survival

and employment, as we see no theoretical link between a firm’s concentration in human therapeutics and its ultimate

employment growth and survival.

Tables 4 and 5 report summary statistics and correlations between variables. The correlations between variables

range from �0.22 to 0.35, with the exception of three hybrid experience types that have correlations with each other

of up to 0.68. We add these carefully into our models, observing any changes in other variables’ significance levels

and coefficients, and measure overall model fit. There are no other variables with correlations higher than 0.35 that

are used together in our regression models. Multicollinearity does not appear to be a problem.

4.1 Models

We use a Cox proportional hazards model to evaluate whether independent variables are associated with firm sur-

vival and firm merger or acquisition events. The Cox proportional hazard model is semi-parametric, with no strong

Table 4. Variable summary statistics, for variables used in regression models

Variable Mean Std. Dev. Min Max N

Age 8.789 5.051 0 26 488

Number of founders 1.893 1.028 1 6 488

Human therapeutics focus 0.322 0.468 0 1 488

Academic experience 0.191 0.335 0 1 488

Big pharma experience 0.104 0.273 0 1 488

Entrepreneurial experience 0.131 0.296 0 1 488

Any hybrid experiencea 0.190 0.336 0 1 488

Academic-big pharma 0.023 0.131 0 1 488

Academic-big pharma-entrepreneurial 0.019 0.104 0 1 488

Big pharma-entrepreneurial 0.084 0.242 0 1 488

Academic-entrepreneurial 0.064 0.196 0 1 488

Patents (binary) 0.205 0.404 0 1 488

Patents granted (among patents recipients) 10.990 17.799 1 87 100

Total public funding (in $1000s) 436.119 1646.046 0 19257.07 488

Total private funding (in $millions) 7.153 28.760 0 221.68 488

Public Funding (binary) 0.305 0.461 0 1 488

Private Funding (binary) 0.158 0.365 0 1 488

Firms with observed closures 0.176 0.381 0 1 488

Firms with observed merger or acquisition 0.195 0.396 0 1 488

Employment in year 1 5.477 7.673 1 66 488




Employment max 16.203 38.133 1 466 488

Source: PLACE: Research Triangle database.aAny Hybrid totals the number of founders that have any type of hybrid experience.


Dow




arolina at Chapel H


Deleted Text: -

Tab

le5.V

ari

ab

leco

rre

lati

on

ma

trix

,fo

rv

ari

ab

les

use

din

reg

ress

ion

mo

de

ls

12

34

56

78

910

11

12

13

14

15

16

17

18

1.A

ge

1.0

00

2.N

um

ber

of

founder

s�

0.1

36

1.0

00

3.H

um

an

ther

apeu

tics

0.2

18

0.0

22

1.0

00

4.A

cadem

icex

per

ience

�0.1

65

0.0

52�

0.0

38

1.0

00

5.B

igpharm

aex

per

ience

0.1

74

0.0

06

0.1

24�

0.3

32

1.0

00

6.E

ntr

epre

neu

rialex

per

ience

�0.1

45�

0.0

27

0.0

26�

0.1

59�

0.1

07

1.0

00

7.H

ybri

d-

Aca

dem

icPharm

a0.0

48�

0.2

16

0.1

73�

0.1

27�

0.0

86�

0.0

71

1.0

00

8.H

ybri

d-

Aca

dem

icPharm

a�

0.0

75

0.1

02

0.1

73�

0.1

27

0.0

09

0.2

30�

0.0

31

1.0

00

9.H

ybri

d-

Entr

epre

neu

rialPharm

a0.0

84

0.0

75

0.1

96�

0.1

83�

0.0

30�

0.0

92�

0.0

68�

0.0

68

1.0

00

10.H

ybri

d-

Entr

epre

neu

rialA

cadem

ic�

0.1

05

0.0

45�

0.0

58�

0.2

21�

0.1

59

0.1

93�

0.0

57

0.1

21

0.1

91

1.0

00

11.H

ybri

d—

any

a�

0.0

07�

0.0

47

0.1

98�

0.3

17�

0.1

56

0.0

71�

0.5

28

0.2

19

0.5

35

0.6

61

1.0

00

12.Pate

nts

gra

nte

d(o

fth

ose

rece

ived

pate

nts

)

0.2

69

0.2

98

0.1

57�

0.0

84

0.1

06�

0.1

46�

0.0

34�

0.0

21

0.1

59�

0.0

31

0.0

34

1.0

00

13.T

ota

lpublic

fundin

g($

1000s)

�0.1

56

0.0

17

0.0

52

0.0

30

0.0

28�

0.1

09�

0.2

89�

0.0

66�

0.0

71

0.0

01

0.1

26

0.0

24

1.0

00

14.T

ota

lpri

vate

fundin

g($

millions)

0.1

29

0.2

82

0.2

97�

0.0

29

0.1

21

0.0

80�

0.0

82�

0.0

60

0.1

98�

0.0

12

0.0

27

0.3

22�

0.1

42

1.0

00

15.E

mplo

ym

ent

inyea

r1

�0.0

40

0.1

68

0.0

65�

0.0

85�

0.0

17�

0.1

56�

0.0

69

0.0

55�

0.0

86

0.0

30�

0.0

54

0.2

72�

0.1

32

0.2

02

1.0

00

16.E

mplo

ym

ent

inyea

r3

0.0

71

0.0

62

0.1

35

0.1

13�

0.0

79�

0.1

69�

0.0

60

0.0

35�

0.1

16�

0.0

01�

0.0

85

0.2

79�

0.1

24

0.1

30

0.8

62

1.0

00

17.E

mplo

ym

ent

inyea

r6

0.0

02

0.2

46

0.1

29

0.0

12�

0.1

04�

0.1

32�

0.0

41�

0.0

36

0.0

14�

0.0

68�

0.0

65

0.5

12�

0.0

86

0.0

35

0.3

85

0.3

91

1.0

00

18.E

mplo

ym

ent

inyea

r9

0.0

67

0.2

24

0.2

11�

0.0

13�

0.0

94�

0.1

17�

0.0

11�

0.0

49

0.0

32�

0.0

71�

0.0

43

0.4

97�

0.0

98

0.1

47

0.3

16

0.3

52

0.9

60

1.0

00

aA

ny

Hybri

dto

tals

the

num

ber

of

founder

sth

at

have

any

type

of

hybri

dex

per

ience

.

10 M. Donegan et al.

Dow




arolina at Chapel H


assumptions about the distribution of the data (or the underlying hazard/survival functions), and it allows us to in-

clude time-varying variables. To test that variables do not violate the assumption of proportional hazards in this spe-

cific model (the assumption that the baseline hazard rate does not change over time—that is, it is the same in year

one as in year ten), we test equality of survival (over time) for each variable, and the model as a whole using the

survival time (sts) test commands in Stata—log rank tests and graphs of Kaplan-Meier survival estimates and Nelson-

Aalen cumulative hazard rates—and the post-estimation command estat phtest to test for violations. While the base-

line hazard rates for some variables do change over time, the model as a whole does not violate the proportionality

assumption.

We model both forms of funding—private and public—using a binary outcome, for whether the firm received

funding or not. We do this because the numbers of observations with any of these outcomes are quite small: 77 firms

receive private funding, and 147 firms receive public funding (state or federal funding). We use a probit model to esti-

mate associations with private funding, but use a logit model to estimate associations with any public funding.

We rely on two separate models to evaluate the association between our independent variables and firm-level

measures of patenting. We first evaluate the relationships between our independent variables and whether firms have

published any patents (e.g., any patents is 1, no patents is 0). As with funding, few firms—just 100 of 488—publish

patents. We use a probit model to estimate associations with any patenting. We present the findings from this model,

but also use this model to generate a score for each observation on the probability of producing any patents. Next,

and looking only at firms that have published at least one patent, we include this probability score as a control meas-

ure in estimating the relationship between the independent variables and the number of patents. The total number of

patents is a count variable for which we use a Poisson distribution to model.

In our employment models we use four cross-sectional measures of employment: at one, three, six and nine years

old. We use a negative binomial distribution to estimate the relationship between the independent variables and em-

ployment outcomes. A negative binomial distribution is an alternative to Poisson, appropriate for count data that is

highly skewed and dispersed.

5. Results

Table 6 provides results for public funding. Model 1 is the base case with control variables for age; number of found-

ers, a binary measure of whether a firm’s technology focus is in human therapeutics or not; and amount of private

funding. Two control variables are statistically significant; age is negatively related across all models, and human

therapeutics is positive across all models. The number of founders and private funding are not significant, which

reflects either that private funding is a later-stage resource than public funding, or that different groups of firms re-

ceive public funding. Models 2, 3 and 4 separately incorporate the three core experience variables: academic experi-

ence (model 2), big pharma experience (model 3), and entrepreneurial experience (model 4). Model 5 simultaneously

incorporates all three core experience variables. Model 6 includes only the catch-all variable for hybrid experience,

and Model 7 includes the three core experience variables and the catch-all variable for hybrid experience. Model 8

includes the three core experience variables, and variables for the four categories of hybrid experience.

There is a positive, statistically significant relationship between firms with higher proportions of founders with aca-

demic experience and the receipt of public funding, and this relationship is consistent across all model permutations.

This result is not surprising, given that this key early-stage influx of funding comes largely from the federal govern-

ment, and acquisition depends on the firm’s technology focus and ability to complete the application process.

Universities may provide this support through their technology transfer offices. On the other hand, firms with higher

proportions of founders with either big pharma or entrepreneurial experience are less likely to have received public

funding. Yet, this statistically significant negative relationship does not hold for firms with higher proportions of

founders with hybrid experiences. Firms with higher proportions of founders who have served these hybrid experiences

are not statistically significant different from founders with no such experience in terms of getting funding. In contrast,

firms with higher proportions of founders with hybrid big pharma-entrepreneurial experience, and no academic experi-

ence, were far less likely to have received public funding than firms with none of these founder experiences.

Table 7 shows results for receiving private funding. Model 1 is the base case with control variables for age; num-

ber of founders, a binary measure of whether a firm’s technology focus is in human therapeutics or not; and amount

of public funding. Age is negative and statistically significant across all models. The relationships between the control

variables for human therapeutics and number of founders are positive and statistically significant, though the


Dow




arolina at Chapel H


relationships are weak and only appear in later models. There is also a positive and statistically significant relation-

ship between a firm’s receipt of public funding and the likelihood of receiving private funding, consistent across all

models. Work experience variables are added in the same manner as for public funding. Firms with higher propor-

tions of founders with academic experience and firms with higher proportions of founders with entrepreneurial ex-

perience are associated with a decreased likelihood of raising private financing, significant at the 0.01 level and

consistent across all models. Firms with higher proportions of three hybrid experience types—academic-big pharma,

big pharma-entrepreneurial, and academic-entrepreneurial—are negatively and statistically significantly associated

with lower levels of private funding, with the significance at a lower threshold than that associated with non-hybrid

experience.

Tables 8 and 9 present results from our patent models, the first of which is a probit model of whether a firm pub-

lishes at least one patent versus none at all (Table 8). Variables are added in an identical fashion to all earlier regres-

sion tables. The second patent model is a Poisson model with a count dependent variable modeling how many

patents each firm receives (Table 9). Variables for age and the number of founders are negative and statistically sig-

nificant across all the models in Table 8, while both forms of funding are positive and significant. All core experien-

ces are negatively related to publishing patents, and statistically significant in combined models. Firms with hybrid

big pharma-entrepreneurial experience and hybrid academic-big pharma experience are less likely to publish patents,

though the result for the latter is only weakly significant.

Table 9 also includes a control variable for the probability of having any patents at all, a measure (probability

score) generated by the probit model. This control variable, along with control variables age and number of founders,

Table 6. Logit regression analysis: association of experience with binary variable for receipt of public financing

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

Age �0.09*** �0.10*** �0.08*** �0.08*** �0.08*** �0.09*** �0.08*** �0.08***

(0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02)

Number of founders �0.00 �0.07 0.00 0.02 �0.03 0.03 0.01 0.00

(0.07) (0.07) (0.07) (0.07) (0.08) (0.07) (0.08) (0.08)

Human therapeutics 0.51** 0.49** 0.56*** 0.49** 0.52** 0.57*** 0.59*** 0.63***

(0.20) (0.21) (0.21) (0.21) (0.21) (0.21) (0.22) (0.22)

Private funding ($millions) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

Academic experience 1.14*** 0.93*** 0.82*** 0.79***

(0.29) (0.29) (0.30) (0.30)

Big pharma experience �1.01** �0.87** �0.98** �0.98**

(0.42) (0.42) (0.43) (0.43)

Entrepreneurial experience �0.81** �0.72* �0.81** �0.82**

(0.36) (0.37) (0.37) (0.37)

Hybrid experience [4 types] �0.50* �0.48

(0.29) (0.31)

Academic-pharma experience 0.70

(0.74)

Academic-pharma-ent

experience

�0.71

(1.05)

Entrepreneurial-pharma

experience

�1.24***

(0.48)

Entrepreneurial-academic

experience

0.04

(0.49)

Log Likelihood �308.1 �299.9 �304.7 �305.3 �296.0 �306.6 �294.7 �291.2

Wald Chi Squared 51.24 62.84 55.19 54.63 67.41 53.73 69.49 73.76

Observations 488 488 488 488 488 488 488 488

Note: Values are presented as coefficients (standard errors).


Dow




arolina at Chapel H


Deleted Text: s.

is both positive and significant across all the models. Controls for human therapeutics and public funding are sporad-

ically statistically significant. Among those firms publishing patents, firms with higher proportions of founders with

academic or entrepreneurial work experiences are less likely to publish higher numbers of patents. In contrast, firms

with higher proportions of founders with hybrid big pharma-entrepreneurial experience are more likely to publish

higher numbers of patents.

Table 10 includes results from our survival models. Model 1 is the base case, with control variables for the first

year the firm is in business (in place of age); the number of founders; and the amount of private funding. We omit age

because the model considers the number of years a firm is in the dataset, a measure that is highly correlated with firm

age, and instead include a variable for the first year in business to control for a cohort effect. Founding year is nega-

tive and statistically significant, though weakly, across all models. The number of founders is positive and statistically

significant across all models. Regarding the experience variables, the only variable with a statistically significant ef-

fect on the likelihood of a firm closure event is the variable representing academic experience: firms with higher pro-

portions of academic founders are more likely to survive, a finding that is consistent with the literature (Degroof and

Roberts, 2004; Rothaermel et al., 2007).

Table 11 includes results from our time-to-merger-or-acquisition models. Model 1 is the base case with control

variables for the first year the firm is in business (in place of age); the number of founders; and funding. The first year

Table 7. Probit regression analysis: association of experience with binary variable for receipt of any private financing

(binary)


Age �0.11*** �0.10*** �0.11*** �0.10*** �0.10*** �0.11*** �0.10*** �0.10***

(0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01)

Number of founders �0.02 0.00 �0.02 0.00 0.03 0.02 0.10* 0.10*

(0.05) (0.05) (0.05) (0.05) (0.05) (0.05) (0.06) (0.06)

Human Therapeutics 0.14 0.14 0.13 0.13 0.14 0.20 0.24 0.27*

(0.14) (0.14) (0.14) (0.14) (0.14) (0.14) (0.15) (0.15)

Total public funding

($thousands)

0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00***

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

Academic experience �0.34 �0.43** �0.64*** �0.67***

(0.21) (0.22) (0.23) (0.23)

Big pharma experience 0.07 �0.11 �0.30 �0.32

(0.25) (0.25) (0.26) (0.26)

Entrepreneurial experience �0.66** �0.74*** �0.86*** �0.86***

(0.28) (0.28) (0.28) (0.29)

Hybrid experience [4 types] �0.58*** �0.78***

(0.21) (0.22)

Academic-pharma experience �4.23*

(2.32)

Academic-pharma-ent

experience

0.04

(0.65)


experience

�0.63**

(0.30)


experience

�0.83**

(0.36)

Log likelihood �227.3 �226.0 �227.2 �224.0 �222.0 �223.3 �215.3 �211.0

Wald Chi Squared 165.5 167.1 165.2 166.1 169.0 169.1 176.1 170.4

Observations 488 488 488 488 488 488 488 488



Dow




arolina at Chapel H


Deleted Text: Rothaermel, Agung, and Jiang, 2007;

in business is negative and statistically significant across all models, controlling for historical factors in the year the

firm starts (e.g., starting in post-recession 2004 vs. 2009 in the next recession; or a year with a major layoff of biotech

workers). The number of founders is positive and statistically significant across all models. The amount of private

funding is not statistically significant. The only experience variable with a statistically significant effect on the likeli-

hood of a merger or acquisition event is the variable representing big pharma experience: firms with higher propor-

tions of big pharma founders are more likely to experience a merger or acquisition event.

Table 12 presents regression results from our employment models. Unsurprisingly, significance of the total found-

ers control variable indicates that firms with more founders have consistently more employees. There are clear tem-

poral relationships between experience types and employment outcomes. Firms with higher proportions of founders

who have had previous entrepreneurial work experience have consistently higher employment counts across all years.

In contrast, prior academic employment is not significant in year one, but is positive and significant in years three,

six and nine. The same pattern holds for big pharma experience, with a larger coefficient. Hybrid results are sporadic.

The hybrid academic-big pharma experience is significant in only year nine and with a large positive coefficient, indi-

cating a strong, but later-in-life, relationship between firms with higher proportions of these entrepreneurs and em-

ployment outcomes, while the academic-big pharma- entrepreneurial hybrid is negative and significant in year six.

Academic-entrepreneurial is positive and significant, though only in years one and three.

Table 8. Probit regression analysis: association of experience with binary variable for any patents published


Age �0.05*** �0.05*** �0.04*** �0.04*** 0.00 �0.05*** �0.04*** �0.03***

(0.01) (0.01) (0.01) (0.01) (.) (0.01) (0.01) (0.01)

Number of founders �0.22*** �0.21*** �0.22*** �0.20*** �0.18*** �0.18*** �0.10* �0.11*

(0.05) (0.05) (0.05) (0.05) (0.05) (0.05) (0.06) (0.06)

Human therapeutics 0.02 0.03 0.04 0.01 0.05 0.09 0.15 0.19

(0.14) (0.14) (0.14) (0.14) (0.14) (0.14) (0.14) (0.15)

Public funding ($1000s) 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00*** 0.00***

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

Private funding ($millions) 0.02*** 0.02*** 0.02*** 0.02*** 0.02*** 0.02*** 0.02*** 0.02***

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

Academic experience �0.19 �0.34* �0.54*** �0.55***

(0.19) (0.20) (0.21) (0.21)

Big pharma experience �0.37 �0.54** �0.72*** �0.72***

(0.25) (0.26) (0.27) (0.27)

Entrepreneurial experience �0.64** �0.75*** �0.87*** �0.88***

(0.25) (0.26) (0.26) (0.26)

Hybrid experience [4 types] �0.64*** �0.87***

(0.21) (0.22)

Academic-pharma experience �1.04*

(0.60)

Academic-pharma-ent

experience

�0.36

(0.65)


experience

�1.26***

(0.36)


experience

�0.55

(0.34)

Log-likelihood �248.0 �247.5 �246.9 �244.4 �241.2 �242.9 �232.5 �231.0

Wald Chi-Squared 152.9 153.6 154.0 156.2 160.1 158.8 170.9 169.3

Number of Observations 488 488 488 488 488 488 488 488



Dow




arolina at Chapel H


6. Discussion

In this article we consider an underexplored aspect of ecosystem research—the effects of employment and career ex-

perience on subsequent entrepreneurial performance. Our results demonstrate that there is value to widening the

scholarly lens to capture and accommodate a broader array of entrepreneurial experience within an entrepreneurial

ecosystem. This means recognizing ecosystem processes that can emerge years, if not decades, before the launch of a

new entrepreneurial venture. In contrast to Wennberg et al. (2011), which finds differential performance of corporate

spawns in terms of sales growth, survival and employment when compared to university spawns, we move the ana-

lysis to consider a wider range of founders’ experience and the impact of team composition on firm performance.

Table 13 summarizes our empirical results: we find that no single pathway into entrepreneurship is consistently suc-

cessful across a broad range of economic outcome measures. Rather, different entrepreneurial pathways produce dif-

ferent economic results that, in aggregate, determine the perceived success of an ecosystem.

Table 9. Poisson regression analysis: association of experience with count of patents, for firms with any patents published


Age 0.15*** 0.15*** 0.15*** 0.14*** 0.14*** 0.15*** 0.14*** 0.13***

(0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01)

Number of founders 0.40*** 0.41*** 0.39*** 0.37*** 0.38*** 0.38*** 0.37*** 0.37***

(0.04) (0.04) (0.04) (0.04) (0.04) (0.04) (0.03) (0.03)

Human therapeutics �0.10 �0.10 �0.11 �0.10 �0.09 �0.13* �0.11 �0.12

(0.07) (0.07) (0.07) (0.07) (0.07) (0.07) (0.07) (0.08)

Public funding ($1000s) �0.00 �0.00 �0.00 �0.00 0.00 �0.00 �0.00 �0.00

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

Private funding ($millions) �0.00** �0.00 �0.00 �0.00 0.00 �0.00* �0.00 �0.00

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

Probability that patents>0a 1.72*** 1.40*** 1.52*** 1.49*** 0.85** 1.62*** 1.17*** 1.27***

(0.41) (0.41) (0.41) (0.43) (0.41) (0.41) (0.41) (0.41)

Academic experience �0.37*** �0.57*** �0.51*** �0.51***

(0.11) (0.12) (0.14) (0.14)

Big pharma experience 0.17* �0.11 �0.08 �0.08

(0.10) (0.11) (0.12) (0.12)

Entrepreneurial experience �1.07*** �1.41*** �1.30*** �1.18***

(0.25) (0.26) (0.26) (0.27)

Hybrid experience [4 types] 0.21 0.05

(0.13) (0.15)

Academic-pharma experience �0.05

(0.27)

Academic-pharma-ent

experience

�0.41

(0.61)


experience

0.59**

(0.24)


experience

�0.36

(0.25)

Constant �0.96*** �0.83*** �0.89*** �0.65*** �0.35 �0.93*** �0.46* �0.44*

(0.22) (0.22) (0.22) (0.23) (0.23) (0.23) (0.24) (0.24)

Log-likelihood �740.4 �731.7 �741.1 �723.3 �708.5 �741.4 �705.5 �701.5

LR Chi-Squared 629.1 646.4 627.7 663.3 692.9 627.1 698.8 706.8

Number of Observations 100 100 100 100 100 100 100 100

Note: Values are presented as coefficients (standard errors).aGenerated through probit model.


Dow




arolina at Chapel H


Deleted Text: paper

Our results indicate the differential performance of firms based on the founders’ career experience. Firms with

second-generation entrepreneurial founders have consistently high employment numbers, though their employment

growth rate slows in later years. Firms with higher proportions of academic experience start with lower employment

numbers, but then pick up in later years, with growth rates surpassing both second-generation entrepreneurial found-

ers and big pharma founders. Cirrus Pharmaceuticals typifies this pattern with two UNC Chapel Hill cofounders

(one faculty and one postdoctoral) and an increasingly higher rate of employment growth over time. The firm,

founded in 1996 and acquired in 2013, had grown to employ about 60 people in North Carolina within nine years of

founding, according to NETS data. Founders with big pharma experience tend to display employment numbers and

rates somewhere in the middle of second-generation and academic founded firms, except for a noticeable increase in

employment around year six. Icagen is one such example. Founded by Glaxo and Burroughs Wellcome veterans, the

company started to employ more people between years three and six of operation, eventually growing to over 70

employees. In 2011 the firm merged into Pfizer.

While the classic parable invoked in our title has a clear winner and loser, we do not find such a result in our

study. If the outcomes of survival or exit through merger or acquisition are considered the ultimate firm outcomes,

we see that the pace of employment growth associated with each type of founder experience does not align with any

specific survival or exit outcome. Rather, we find that firms whose growth patterns behave like the parable’s tortoise

are more likely to have a merger or acquisition when they have big pharma experienced founders, and are more likely

to survive as entrepreneurial firms when they have academic founders. Two successful firms in the PLACE: Research

Triangle database, Lineberry Research Associates and Agile Sciences, have followed these trajectories. Lineberry

Research Associates, a full-service CRO that specialized in clinical trial management, regulatory consulting, medical

Table 10. Cox proportional hazards regression analysis: association of experience with firm closure


Founding year �0.13*** �0.13*** �0.13*** �0.13*** �0.13*** �0.13*** �0.12*** �0.12***

(0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02)

Number of founders 0.17* 0.18** 0.17* 0.16* 0.18* 0.17* 0.18* 0.21**

(0.09) (0.09) (0.09) (0.09) (0.09) (0.09) (0.09) (0.10)

Private funding ($millions) �0.00 �0.00 �0.00 �0.00 �0.00 �0.00 �0.00 �0.00

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

Academic experience �0.46 �0.53 �0.56 �0.57*

(0.31) (0.33) (0.34) (0.34)

Big pharma experience 0.04 �0.12 �0.14 �0.17

(0.31) (0.33) (0.33) (0.34)

Entrepreneurial experience �0.22 �0.33 �0.36 �0.35

(0.33) (0.34) (0.35) (0.35)

Hybrid experience [4 types] 0.11 �0.09

(0.30) (0.32)


(0.55)

Academic-pharma-ent

experience

0.55

(0.70)


experience

�0.12

(0.45)


experience

�1.00

(0.78)

Log-likelihood �711.8 �710.7 �711.8 �711.6 �710.2 �711.8 �710.1 �708.6

LR Chi-Squared 58.25 60.55 58.27 58.69 61.59 58.37 61.67 64.61

Number of Observations 10,762 10,762 10,762 10,762 10,762 10,762 10,762 10,762



Dow




arolina at Chapel H


writing, and other development services was founded in 1995 by two individuals with almost 30 years combined ex-

perience at Burroughs Wellcome. One of the cofounders began their career as a professor in another state. The com-

pany was acquired by Constella Group in 2006 after growing to 75 employees, according to our data. Two North

Carolina State University faculty started Agile Sciences in 2007. Both founders have maintained positions at the firm,

while also holding faculty and other industry positions. The firm is still active and headquartered in Raleigh, though

according to NETS data it is not one of the region’s larger employers.

In contrast, the firms we liken to the parable’s hare—those with second-generation entrepreneurial founders—do

not show any significant results for their survival and exit outcomes. Still, employment is positive and statistically sig-

nificant for these firms, as well as for firms with hybrid academic-big pharma experience founders. Liposcience’s

founders held faculty positions at North Carolina State University, and executive positions at both multinationals

Glaxo and Eli Lilly, and at Research Triangle Startup Trimeris before co-founding the Liposcience firm in the mid-

1990s. Liposcience developed blood tests for the diagnosis of cardiovascular disease. Though the company grew to

employ as many as 165 people, it had difficulty after its 2014 initial public offering. The company was eventually

acquired by LabCorp, another Research Triangle regional anchor, for $85.3 million in 2014.

Firms associated with higher proportions of academic experience are the only firm type that shows a higher sur-

vival rate, and the only type of firm that has significantly more success in raising public financing; big pharma firms

are the only set of firms statistically more likely to experience a merger or acquisition. Yet, we also find unexpected

results among our hybrid types. For example, only those firms with higher proportions of founders with hybrid big

pharma-entrepreneurial experience are associated with higher patenting numbers among those firms that patent.

Table 11. Cox proportional hazards regression analysis: association of experience with merger or acquisition event


Age 0.11*** 0.11*** 0.11*** 0.11*** 0.11*** 0.10*** 0.11*** 0.10***

(0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02) (0.02)

Number of founders 0.21** 0.22** 0.23** 0.22** 0.24** 0.21** 0.24** 0.27**

(0.11) (0.11) (0.11) (0.11) (0.11) (0.11) (0.11) (0.11)

Human therapeutics 0.27 0.28 0.23 0.28 0.25 0.28 0.24 0.20

(0.24) (0.24) (0.24) (0.24) (0.25) (0.24) (0.25) (0.25)

Total public funding ($thousands) �0.00 �0.00 �0.00 �0.00 �0.00 �0.00 �0.00 �0.00

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

Total private funding ($millions) �0.00 �0.00 �0.00 �0.00 �0.00 �0.00 �0.00 �0.00

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

Academic experience �0.19 0.10 0.10 0.09

(0.36) (0.38) (0.40) (0.40)

Big pharma experience 0.92*** 1.01*** 1.01*** 0.98***

(0.33) (0.35) (0.37) (0.37)

Entrepreneurial experience 0.21 0.41 0.41 0.43

(0.38) (0.40) (0.41) (0.41)

Hybrid experience [4 types] �0.22 0.01

(0.40) (0.44)


(0.98)

Academic-pharma-ent experience 0.33

(1.08)

Entrepreneurial-pharma experience 0.41

(0.52)

Entrepreneurial-academic experience �1.58

(1.22)

Log-likelihood �479.0 �478.8 �475.6 �478.8 �475.1 �478.8 �475.1 �473.6

LR Chi-Squared 31.86 32.16 38.49 32.15 39.50 32.17 39.50 42.56

Number of Observations 10,768 10,768 10,768 10,768 10,768 10,768 10,768 10,768



Dow




arolina at Chapel H


Table 12. Negative binomial regression analysis: association of experience with employment count

Model 1 Model 2 Model 3 Model 4

Year 1 Employment Year 3 Employment Year 6 Employment Year 9 Employment

Number of founders 0.76*** 0.86*** 1.10*** 1.30***

(0.04) (0.05) (0.06) (0.08)

Public funding ($1000s) �0.00 �0.00 �0.00 �0.00*

(0.00) (0.00) (0.00) (0.00)

Private funding ($millions) 0.00 0.00 0.00* 0.00***

(0.00) (0.00) (0.00) (0.00)

Academic experience 0.17 0.39** 0.64*** 0.60**

(0.15) (0.17) (0.22) (0.28)

Big pharma experience 0.30 0.38* 1.29*** 0.90***

(0.19) (0.22) (0.28) (0.34)

Entrepreneurial experience 0.38** 0.82*** 0.78*** 0.82***

(0.18) (0.20) (0.24) (0.31)

Academic-pharma experience 0.49 0.67 0.93 1.66**

(0.38) (0.44) (0.61) (0.73)

Academic-pharma-ent experience �0.68 �0.99 �1.30* �1.38

(0.54) (0.60) (0.76) (0.89)

Entrepreneurial-pharma experience 0.02 �0.09 �0.10 0.24

(0.22) (0.26) (0.34) (0.45)

Entrepreneurial-academic experience �0.63** �0.67** �0.51 �0.30

(0.27) (0.31) (0.48) (0.73)

Log-likelihood �1441.5 �1539.9 �1153.4 �825.9

Wald Chi-Squared 861.7 861.0 886.8 689.2

Number of Observations 488 415 306 197


Table 13. Summary table of empirical results, different types of prior work experience

Public

funding

Private

funding

Patenting Patent

count

Survival M&A Employment

Academic experience Positive Negative Negative Negative More likely

to survive

Positive (years 3, 6, 9)

Big pharma experience Negative Negative Positive Positive Positive (years 3, 6, 9)

Entrepreneurial experience Negative Negative Negative Negative Positive (all years)

Academic-pharma experience Negative Negative Positive (last year)

Academic-pharma-ent

experience


experience

Negative Negative Negative Positive


experience

Negative


Dow




arolina at Chapel H


The fact that the region’s academic firms do not consistently outperform their peers is surprising, given the long-

standing and historical relationship between the development of the Research Triangle Park and the region’s three re-

search universities (Link, 1995; McCorkle, 2012). The result is also contrary to research that finds better

performance among firms with strong links to university technology programs and funding sources (Link and Ruhm,

2009; Siegel and Wessner, 2012). Our results could be unique to the region, reflecting the region’s universities’ late

entry into technology transfer (Bercovitz et al., 2001; Donegan, 2019); the early years of our study period occur

when few entrepreneurial companies emerged from the campuses. However, recent research suggests that univer-

sities’ strongest contributions to regional ecosystems remain in training a local labor force—not in patenting and aca-

demic entrepreneurship (Drucker, 2016). Our results underscore the need to consider universities as anchor

institutions that serve a variety of purposes, both direct and indirect. In addition to producing patents and entrepre-

neurial firms, the Research Triangle universities train the workers that go on to work in local big pharma and entre-

preneurial firms before founding their own companies. One serial cofounder received a PhD at Duke, then held a

postdoctoral position at Harvard. The founder returned to work at Glaxo for a few years before starting a series of

companies in the life sciences sector in North Carolina.

7. Conclusion and policy implications

From a policy-making perspective, there is clearly value to capturing a more complete picture of entrepreneurial ex-

perience within a regional ecosystem. For starters, regional economic development practices support transformative

processes of local economic and institutional change that involves a mix of concurrent interventions (Markusen and

Schrock, 2006; Lowe and Wolf-Powers, 2016). Our results suggest that a better understanding of the experience

composition of local entrepreneurs provides information to support and reinforce policy consequences. The Research

Triangle firms we showcased above would not have been recognized as consistent “winners” in their tenure in the re-

gion, yet all have contributed to the process of entrepreneurial ecosystem development (Spigel and Harrison, 2018).

Privileging one form of entrepreneurial entry can blind us to the way that entrepreneurship, as one of many eco-

nomic activities, intersects with other aspects of the ecosystem. Recognizing this, some regions have intentionally

adopted what some scholars have described as a “portfolio” approach to economic development policy (Markusen

and Schrock, 2006; Rodrik, 2014). In this policy environment, the goal is not simply to have parallel, yet independ-

ent, policy tracks, but to identify and seek out opportunities for integrating entrepreneurial supports in ways that bal-

ance and combine multiple developmental objectives (Stark, 2011). This policy logic is outlined in the call for smart

specialization in Europe, which puts strong emphasis on supporting entrepreneurship through combined goals of in-

novation, equity, and environment. It also reflects growing recommendations from researchers of state and local pol-

icy in the United States to embrace economic development planning as more than a set of competing public

investment choices, but instead as an integrated policy platform for connecting and layering multiple objectives

(Lowe and Feldman, 2018).

As the ecosystem concept continues to gain traction in the practitioner and academic communities, it is imperative

that the research begins to consider the many dynamics and interdependencies inherent in them. This article has

delineated several of the economic and innovation outcomes associated with prior work experience at varying estab-

lishments, which have varying influences on how entrepreneurs start and develop their firms. While this article offers

new insight into how multiple pathways for entrepreneurship may exist and intersect in one region, its generalizabil-

ity is limited by its focus on life sciences, an industry that lends itself to diverse founding teams in ways other indus-

tries may not. The focus on one region—particularly one that has historically had both large firms and multiple

research-intensive universities—also limits its generalizability, as few regions have such breadth in their established

institutions.

Still, we argue the findings have important insights that apply beyond similarly situated regions. The paper’s com-

plementary view of experience moves away from privileging one type of experience over another. In so doing it

presents clear evidence that the mix of current career options may eventually influence a region’s entrepreneurial tra-

jectory and ecosystem evolution—a finding with broad applicability for locales attempting to strengthen current, or

adopt new, industries. Regions are constantly compared against one another, with the expectation that their perform-

ance should be similar. Our results highlight that different types of entrepreneurs’ prior experiences may account for

noted differences. Moreover, recognizing the different profiles of entrepreneurs have different operational logics may

suggest specific interventions to improve long run performance.


Dow




arolina at Chapel H


Deleted Text: P

Deleted Text: I

Deleted Text: <xref ref-type=

Deleted Text: paper

Deleted Text: paper

Further research should consider the temporal development of the region and the possibility of different prior

work experiences dominating at different time periods. Ultimately, the societal goal is to establish vibrant entrepre-

neurial ecosystems. To this end, more work remains to be done.

Acknowledgment

This work was supported by the University of North Carolina at Chapel Hill Office of Economic and Business Development and the

Odum Institute for Social Science Research, as well as the National Science Foundation and the Ewing Marion Kauffman

Foundation. We also wish to acknowledge the creative contribution and forensic skills of our UNC undergraduate and graduate re-

search team, especially Keagan Sacripanti.

References

Aldrich, H. E. and P. H. Kim (2007), ‘Small worlds, infinite possibilities? How social networks affect entrepreneurial team formation

and search,’ Strategic Entrepreneurship Journal, 1(1–2), 147–165.

Astebro, T. and P. Thompson (2011), ‘Entrepreneurs, jacks of all trades or hobos?’ Research Policy, 40(5), 637–649.

Autio, E., S. Nambisan, L. D. Thomas and M. Wright (2018), ‘Digital affordances, spatial affordances, and the genesis of entrepre-

neurial ecosystems,’ Strategic Entrepreneurship Journal, 12(1), 72–95.

Baker, T., A. Miner and D. Eesley (2003), ‘Improvising firms: bricolage, account giving and improvisational competencies in the

founding process,’ Research Policy, 32(2), 255–276.

Baron, J. N., M. T. Hannan and M. D. Burton (2001), ‘Labor pains: change in organizational models and employee turnover in

young, high-tech firms,’ American Journal of Sociology, 106(4), 960–1012.

Bercovitz, J., M. Feldman, I. Feller and R. Burton (2001), ‘Organizational structure as a determinant of academic patent and licensing

behavior: an exploratory study of Duke, Johns Hopkins, and Pennsylvania State Universities,’ Journal of Technology Transfer,

26(1/2), 21–35.

Burton, M. D., J. B. Sorensen and C. Beckman (2002), ‘Coming from good stock: career histories and new venture formation,’ in M.

Lounsbury and M. Ventresca (eds), Research in the Sociology of Organizations: Vol. 19. Social Structure and Organizations

Revisited. JAI Press: Greenwich, CT, pp. 229–262.

Campbell, B., M. Ganco, A. Franco and R. Agarwal (2012), ‘Who leaves, where to, and why worry? Employee mobility, employee

entrepreneurship, and effects on source firm performance,’ Strategic Management Journal, 33(1), 65–87.

Chowdhury, S. (2005), ‘Demographic diversity for building an effective entrepreneurial team: is it important?’ Journal of Business

Venturing, 20(6), 727–746.

Clayton, P., M. Feldman and N. Lowe (2018), ‘Behind the scenes: intermediary organizations that facilitate science commercializa-

tion through entrepreneurship,’ Academy of Management Perspectives, 32(1), 104–124.

Dahl, M. S. and O. Sorenson (2012), ‘Home sweet home: entrepreneurs’ location choices and the performance of their ventures,’

Management Science, 58(6), 1059–1071.

Decker, R., J. Haltiwanger, R. Jarmin and J. Miranda (2014), ‘The role of entrepreneurship in US job creation and economic dyna-

mism,’ Journal of Economic Perspectives, 28(3), 3–24.

Degroof, J. J. and E. B. Roberts (2004), ‘Overcoming weak entrepreneurial infrastructures for academic spin-off ventures,’ Journal of

Technology Transfer, 29(3/4), 327–353.

Dencker, J. C., M. Gruber and S. K. Shah (2009), ‘Pre-entry knowledge, learning, and the survival of new firms,’ Organization

Science, 20(3), 516–537.

Dobrev, S. D. and W. P. Barnett (2005), ‘Organizational roles and transition to entrepreneurship,’ Academy of Management Journal,

48(3), 433–449.

Donegan, M. (2019), Bridging Commercialization and Redevelopment: Jurisdictions and University Policy Development. Urban

Studies, doi: 10.1177/0042098018804758.

Donegan, M. and N. Lowe (2017), ‘Going local: Placing entrepreneurial microgeographies in a larger regional context,’ Paper

Presented at American Association of Geographers Annual Meeting, Boston, MA.

Drucker, J. (2016), ‘Reconsidering the regional economic development impacts of higher education institutions in the United States,’

Regional Studies, 50(7), 1185–1202.

Elfenbein, D. W., B. H. Hamilton and T. R. Zenger (2010), ‘The small firm effect and the entrepreneurial spawning of scientists and

engineers,’ Management Science, 56(4), 659–681.

Ensley, M. D., J. W. Carland and J. C. Carland (1998), ‘The effect of entrepreneurial team skill heterogeneity and functional diversity

on new venture performance,’ Journal of Business and Entrepreneurship, 10(1), 1–9.

Feldman, M. P. and J. L. Francis (2004), ‘Homegrown solutions: fostering cluster formation,’ Economic Development Quarterly,

18(2), 127–137.


Dow




arolina at Chapel H


Feldman, M. P., J. Francis and J. Bercovitz (2005), ‘Creating a cluster while building a firm: entrepreneurs and the formation of in-

novative clusters,’ Regional Studies, 39(1), 129–142.

Feldman, M. P. and N. Lowe (2008), ‘Consensus from controversy: Cambridge’s biosafety ordinance and the anchoring of the biotech

industry,’ European Planning Studies, 16(3), 395–410.

Feldman, M. P. and N. Lowe (2015), ‘Triangulating regional economies: the promise of digital data,’ Research Policy, 44(9),

1785–1793.

Feldman, M. P. and N. Lowe (2018), ‘Policy and collective action in place,’ Cambridge Journal of Regions, Economy and Society,

11(2),335–351. Retrieved from https://doi.org/10.1093/cjres/rsy011.

Franklin, S. J., M. Wright and A. Lockett (2001), ‘Academic and surrogate entrepreneurs in university spin-out companies,’ The

Journal of Technology Transfer, 26(1/2), 127–141.

Kenny, M. and U. von Burg (1999), ‘Technology and path dependence: the divergence between Silicon Valley and Route 128,’

Industrial and Corporate Change, 8(1), 67–103.

Klepper, S. (2001), ‘Employee startups in high-tech industries,’ Industrial and Corporate Change, 10(3), 639–674.

Klepper, S. and S. Sleeper (2005), ‘Entry by spinoffs,’ Management Science, 51(8), 1291–1306.

Kolko, J. and D. Neumark (2007), Business Location Decisions and Employment Dynamics in California. Public Policy Institute of

California: San Francisco, CA.

Link, A. N. (1995), A Generosity of Spirit: The Early History of the Research Triangle Park. Research Triangle Foundation: North

Carolina, NC.

Link, A. N. and C. J. Ruhm (2009), ‘Bringing science to market: commercializing from NIH SBIR Awards,’ Economics of Innovation

and New Technology, 18(4), 381–402.

Lowe, N. J. and M. P. Feldman (2008), ‘Constructing entrepreneurship advantage: consensus building, technological uncertainty and

emerging industries,’ Cambridge Journal of Regions, Economy and Society, 1(2), 265–284.

Lowe, N. J. and M. P. Feldman (2017), ‘Institutional life within an entrepreneurial region,’ Geography Compass, 11(3), doi: 10.

1111/gec3.12306.

Lowe, N. J. and M. P. Feldman (2018), ‘Breaking the waves: innovative innovating at the intersections of local economic development

planning,’ Economic Development Quarterly, 32(3), 183–194.

Lowe, N. J. and L. Wolf-Powers (2016), ‘Who works in a working region?: Inclusive innovation in the new manufacturing economy,’

Paper Presented at the Annual Meeting of the Association of American Geographers, San Francisco, CA.

Luger, M. I. and H. A. Goldstein (1991), Technology in the Garden: Research Parks and Regional Economic Development. The

University of North Carolina Press: Chapel Hill, NC.

Markusen, A. (1996), ‘Sticky places in slippery space: a typology of industrial districts,’ Economic Geography, 72(3), 293–313.

Markusen, A. and G. Schrock (2006), ‘The distinctive city: divergent patterns in growth, hierarchy and specialization,’ Urban Studies,

43(8), 1301–1323.

Marquis, C. (2003), ‘The pressure of the past: network imprinting in intercorporate communities,’ Administrative Science Quarterly,

48(4), 655–689.

Mason, C. and R. Brown (2014), ‘Entrepreneurial ecosystems and growth oriented entrepreneurship,’ Background paper prepared

for the workshop organised by the OECD LEED Programme and the Dutch Ministry of Economic Affairs. The Hague,

Netherlands.

Mayer, H. (2011), Entrepreneurship and Innovation in Second Tier Regions. Edward Elgar Publishing: Northampton, MA.

McCorkle, M. (2012), ‘History and the “new economy” narrative: the case of Research Triangle Park and North Carolina’s economic

development,’ The Journal of the Historical Society, 12(4), 479–525.

Neumark, D., B. Wall and J. Zhang (2011), ‘Do small businesses create more jobs? New evidence for the United States from the

National Establishment Time Series,’ The Review of Economics and Statistics, 93(1), 16–29.

O’Gorman, C., O. Byrne and D. Pandya (2008), ‘How scientists commercialise new knowledge via entrepreneurship,’ The Journal of

Technology Transfer, 33(1), 23–43.

Roberts, P. W., S. Klepper and S. Hayward (2011), ‘Founder backgrounds and the evolution of firm size,’ Industrial and Corporate

Change, 20(6), 1515–1538.

Rodrik, D. (2014), ‘Green industrial policy,’ Oxford Review of Economic Policy, 30(3), 469–491.

Rothaermel, F. T., S. D. Agung and L. Jiang (2007), ‘University entrepreneurship: a taxonomy of the literature,’ Industrial and

Corporate Change, 16(4), 691–791.

Siegel, D. S. (2006), Technological Entrepreneurship: Institutions and Agents Involved in University Technology Transfer. Edward

Elgar Publishing: Cheltenham, UK.

Siegel, D. S., R. Veugelers and M. Wright (2007), ‘Technology transfer offices and commercialization of university intellectual prop-

erty: performance and policy implications,’ Oxford Review of Economic Policy, 23(4), 640–660.

Siegel, D. S. and C. Wessner (2012), ‘Universities and the success of entrepreneurial ventures: evidence from the Small Business

Innovation Research Program,’ Journal of Technology Transfer, 37(4), 404–415.


Dow




arolina at Chapel H


https://doi.org/10.1093/cjres/rsy011

http://10.1111/gec3.12306

http://10.1111/gec3.12306

Sørensen, J. B. and D. J. Phillips (2011), ‘Competence and commitment: employer size and entrepreneurial endurance,’ Industrial and

Corporate Change, 20, 1277–1304.

Spigel, B. and R. T. Harrison (2018), ‘Towards a process theory of entrepreneurial ecosystems,’ Strategic Entrepreneurship Journal,

12(1), 151–168.

Stam, E. and B. Spigel (2018), ’Entrepreneurial ecosystems,’ in R. Blackburn, D. De Clercq, J. Heinonen, and Z. Wang (eds), SAGE

Handbook of Entrepreneurship and Small Business. Sage: London, UK, pp. 407–422.

Stark, D. (2011), The Sense of Dissonance: Accounts of Worth in Economic Life. Princeton University Press: Princeton, NJ.

Ucbasaran, D., A. Lockett, M. Wright and P. Westhead (2003), ‘Entrepreneurial founder teams: factors associated with member entry

and exit,’ Entrepreneurship Theory and Practice, 28(2), 107–128.

Ucbasaran, D., P. Westhead, M. Wright and M. Flores (2010), ‘The nature of entrepreneurial experience, business failure and com-

parative optimism,’ Journal of Business Venturing, 25(6), 541–555.

Vanaelst, I., B. Clarysse, M. Wright, A. Lockett, N. Moray and R. S’Jegers (2006), ‘Entrepreneurial team development in academic

spinouts: an examination of team heterogeneity,’ Entrepreneurship Theory and Practice, 30(2), 249–271.

Walls and Associates (2013), National Establishment Time-Series (NETS) Database: 2013 Database Description.

Wennberg, K., J. Wiklund and M. Wright (2011), ‘The effectiveness of university knowledge spillovers: performance differences be-

tween university spinoffs and corporate spinoffs,’ Research Policy, 40(8), 1128–1143.

Westhead, P. and M. Wright (1998), ‘Novice, portfolio, and serial founders: are they different?’ Journal of Business Venturing, 13(3),

173–204.

Wright, M., S. Birley and S. Mosey (2004), ‘Entrepreneurship and university technology transfer,’ The Journal of Technology

Transfer, 29(3/4), 235–246.


Dow




arolina at Chapel H


The tortoise, the hare, and the hybrid: effects of prior ...maryannfeldman.web.unc.edu/files/2019/07/The-tortoise-the-hare-an… · The tortoise, the hare, and the hybrid: effects

Documents