ArticlePDF Available

Structural Role Complementarity in Entrepreneurial Teams


Abstract and Figures

To refine the understanding of the social network characteristics of entrepreneurial teams, we present a new construct: structural role complementarity. In particular, we examine the variation between team members’ respective abilities to act as network brokers. Based on the cofounding networks of 9,461 entrepreneurs and 2,446 large-scale industrial enterprises over 45 years in Russia’s emerging economy (1869–1913), our findings show that variation among team members’ brokering ability significantly predicts the starting capital raised by their firm. The effect is moderated by the team’s average brokering potential. When both the team’s average and variation in brokering potential is high, firms raise greater starting capital. By using multiple membership models, we demonstrate that greater starting capital is largely attributable to team factors rather than the attributes of the individual team members. We also take advantage of discriminatory laws that were passed in 1887 in an instrumental variable analysis to address potential endogeneity issues. The online appendix is available at . This paper was accepted by Olav Sorenson, organizations.
Content may be subject to copyright.
This article was downloaded by: [] On: 27 December 2017, At: 10:35
Publisher: Institute for Operations Research and the Management Sciences (INFORMS)
INFORMS is located in Maryland, USA
Management Science
Publication details, including instructions for authors and subscription information:
Structural Role Complementarity in Entrepreneurial
Brandy Aven, Henning Hillmann
To cite this article:
Brandy Aven, Henning Hillmann (2017) Structural Role Complementarity in Entrepreneurial Teams. Management Science
Published online in Articles in Advance 30 Nov 2017
Full terms and conditions of use:
This article may be used only for the purposes of research, teaching, and/or private study. Commercial use
or systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisher
approval, unless otherwise noted. For more information, contact
The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitness
for a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, or
inclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, or
support of claims made of that product, publication, or service.
Copyright © 2017, The Author(s)
Please scroll down for article—it is on subsequent pages
INFORMS is the largest professional society in the world for professionals in the fields of operations research, management
science, and analytics.
For more information on INFORMS, its publications, membership, or meetings visit
Articles in Advance,pp. 1–17 ISSN 0025-1909 (print), ISSN 1526-5501 (online)
Structural Role Complementarity in Entrepreneurial Teams
Brandy Aven,a,Henning Hillmannb
aTepper School of Business, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213; bDepartment of Sociology, School of Social
Sciences, University of Mannheim, 68131 Mannheim, Germany
Corresponding author
Contact:, (BA); (HH)
Received: November 22, 2016
Revised: June 11, 2017
Accepted: June 15, 2017
Published Online in Articles in Advance:
November 30, 2017
Copyright: ©2017 The Author(s)
Abstract. To refine the understanding of the social network characteristics of entrepre-
neurial teams, we present a new construct: structural role complementarity. In particular,
we examine the variation between team members’ respective abilities to act as network
brokers. Based on the cofounding networks of 9,461 entrepreneurs and 2,446 large-scale
industrial enterprises over 45 years in Russia’s emerging economy (1869–1913), our find-
ings show that variation among team members’ brokering ability significantly predicts the
starting capital raised by their firm. The effect is moderated by the team’s average brokering
potential. When both the team’s average and variation in brokering potential is high, firms
raise greater starting capital. By using multiple membership models, we demonstrate that
greater starting capital is largely attributable to team factors rather than the attributes of the
individual team members. We also take advantage of discriminatory laws that were passed
in 1887 in an instrumental variable analysis to address potential endogeneity issues.
Accepted by Olav Sorenson, organizations.
Open Access Statement:
This work is licensed under a Creative Commons Attribution-NonCommercial
4.0 International License. You are free to download this work and share with others for any purpose,
except commercially, and you must attribute this work as “Management Science. Copyright ©2017
The Author(s)., used under a Creative Commons Attribu-
tion License:”
Supplemental Material:
The online appendix is available at
social networks
structural roles
entrepreneurial teams
emerging economies
Entrepreneurship requires an array of knowledge and
skills, as well as social connections to various groups
that may not be solely accessible to a lone entrepre-
neur (Aldrich and Kim 2007, Ruef 2010). When teams
of entrepreneurs found new firms, it is often the diver-
sity of experience among those founders that under-
lies the new firm’s success (Beckman et al. 2007, Ruef
et al. 2003). The literature on the social capital of teams
suggests that for internally cohesive teams, the ben-
efit of compositional diversity partly stems from the
range of different ties that each member has external to
the team, which provide access to valuable resources
and information (Oh et al. 2004, Reagans et al. 2004).
Yet, at the same time, economic sociologists observe
that cohesion among a team’s external relations also
provides unique benefits important to these entrepre-
neurial teams, such as joint problem solving, increased
trust, and improved knowledge sharing (Ahuja 2000,
Aldrich and Zimmer 1986, Uzzi 1997, Vedres and Stark
2010). We argue that these seemingly inconsistent find-
ings arise in part from the attribution of social relations
to firms rather than to the individual entrepreneurs
who direct these firms (Sorenson and Rogan 2014).
Instead, we propose structural role complementarity as
a means unique to teams where individual members
“specialize” in structural roles, allowing entrepreneur-
ial teams to profit from brokering ties while simulta-
neously retaining the advantages of ties to a cohesive
For individual entrepreneurs, it is rarely possible to
fully retain the benefits of the social capital derived
from ties to a cohesive group while attempting to bro-
ker beyond that cohesive group (Hillmann 2008). The
mechanisms of group cohesion and norm enforcement
that provide social capital benefits to entrepreneurs
also act to discourage relations to non-group mem-
bers (Coleman 1988). When an entrepreneur connects
to non-group members, it tends to weaken the ability
of partners within the group to evaluate his business
dealings and to sanction him effectively (Portes and
Sensenbrenner 1993, Shipilov and Li 2008). Hence, as
the entrepreneur embedded in a cohesive group begins
to benefit from brokerage connections, it may also
threaten his “good standing” in the group. However,
entrepreneurial teams need not make this social cap-
ital trade-off between cohesive ties versus brokering
ties because entrepreneurial ties belong to founders,
not firms. Drawing on the sociological concept of struc-
tural roles (White 2001), we argue that the pattern of
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
2Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s)
relationships that each entrepreneur maintains is spe-
cific to the individual team member, and differentiates
them into a distinct role within the team. Thus, collec-
tively, entrepreneurial teams may simultaneously real-
ize the full benefits of both brokerage and cohesive ties
without incurring the trade-off faced by an individual
entrepreneur who attempts both roles.
We assess our argument using a data set uniquely
suited to investigate the networks of entrepreneurial
teams—namely, data on the chartering of all joint-stock
and share partnership firms in Russia between 1869
and 1913. Comprised of state-mandated founding char-
ters for 2,446 firms formed by entrepreneurial teams,
this data set provides not only detailed descriptives
for the firm, such as the firm’s industry and region,
but also key characteristics of the 9,461 associated
founders, such as founder ethnicity. Similar to modern
founding of firms, the majority of all new ventures in
our setting (65% of the 3,762 new firms) were estab-
lished by a team of two or more entrepreneurs (Aldrich
and Kim 2007, Ruef et al. 2003). To examine how struc-
tural role complementarity contributes to firm success,
we model the basic starting capital raised by the firm, a
critical event in the entrepreneurial process that is doc-
umented in the charters (Beckman et al. 2007, Hillmann
and Aven 2011, Stuart and Sorenson 2003).
These historical data permit us to examine the rela-
tional patterns of founding teams in the context of
an emerging economy with detailed longitudinal data
on both individuals and firms; comparable data are
rarely available for analogous contemporary settings.
Emerging economies currently account for 58% of the
world’s gross domestic product measured in purchas-
ing power parity, as reported by the International Mon-
etary Fund’s (2016)World Economic Outlook. In con-
trast to established economies, emerging economies
are characterized by high institutional uncertainty and
limited access to capital resources (Hillmann and Aven
2011, Khanna and Rivkin 2006, Nee and Opper 2012).
These conditions accentuate the value of reliable infor-
mation acquired via social networks (Eisenhardt and
Schoonhoven 1990), and our data on Russian entrepre-
neurial teams provide a unique opportunity to assess
the importance of these networks on the success of new
firms in emerging economies.
Although teams of founders initiate a large por-
tion of entrepreneurial ventures (Aldrich and Kim
2007, Ruef et al. 2003), the study of founding teams
poses methodological challenges. In particular, our
data include founders who join multiple entrepreneur-
ial teams over the 45-year study period, and the success
of teams is potentially attributable to the characteris-
tics or skills of these individual serial founders. How-
ever, our dependent variable of basic capital raised is
specific to entrepreneurial teams, not individuals, and
furthermore, only 1.5% of the entrepreneurial teams in
this sample are comprised of the same set of unique
founders. These considerations preclude the use of
conventional fixed effects to assess the extent to which
unobserved heterogeneity across team members, such
as their talent for entrepreneurship, drives capital
mobilization. We therefore use multiple membership
models with weighted random effects to account for
the nonindependence that stems from the repeated
observations of serial founders and the cross-sectional
outcome. These models provide more conservative
estimates than the corresponding GLM estimates while
also offering insight into the extent to which vari-
ance in entrepreneurial capital is attributable to indi-
vidual founders. Next, to address potential endogene-
ity concerns with cross-sectional data, such as reverse
causation or omitted variables, we take advantage of
discriminatory laws that were passed in 1887 in an
instrumental variable (IV) analysis. These laws ex-
cluded Jews and foreigners from the potential set of
founding team members for certain regions and indus-
tries. Therefore, these laws affect our independent vari-
able, a measure of the diversity of brokerage roles
among team members, but not our dependent variable
of capital raised by founding teams, making them well
suited for use as an instrument.
Brokerage and Cohesion in
Social relations, particularly to other founders, pro-
vide access to relevant market information, resources,
and social support, which in turn greatly enhance new
firm outcomes (Hillmann and Aven 2011, Khanna and
Rivkin 2006, Stuart et al. 1999, Vedres and Stark 2010).
In particular, social network ties have been found to
help new firms and entrepreneurs access financial cap-
ital (Aldrich and Zimmer 1986, Portes and Sensenbren-
ner 1993); permit admission to supporting institutions,
such as professional service organizations (Freeman
1999); and facilitate the identification of entrepreneur-
ial opportunities (Stuart and Sorenson 2005). In sum,
the individual founders’ relations to other founders
beyond the focal team provide crucial channels for
resource mobilization.
Mobilization of resources presents a critical twofold
challenge for entrepreneurs, and the lack of institu-
tional support common to emerging economies only
serves to accentuate these challenges (Marquis and
Raynard 2015). First, founders must search for poten-
tial resources such as investors with capital, employ-
ees with the appropriate skills, banks able to pro-
vide financing, and suppliers willing to extend credit.
Once the founders have identified potential resources,
they next must mobilize them by, for instance, con-
vincing investors to invest in the firm and prospective
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s) 3
employees to join the firm. Emerging economies’ char-
acteristic lack of economic institutions and infrastruc-
ture that promote and regulate markets hinders both
the search for available resources and the trust neces-
sary to mobilize resources (Hillmann and Aven 2011,
Khanna and Rivkin 2006, Nee and Opper 2012).
On the one hand, founders who act as brokers can
locate resources and opportunities in the market, such
as potential investors with capital, skilled labor, and
potential suppliers or distributors (Burt 2005, Ferriani
et al. 2009, Stuart et al. 1999, Stuart and Sorenson 2005).
Since their network ties connect them to an array of
disparate parties, brokers often have valuable infor-
mation about suppliers, customers, and competitors
(Vedres and Stark 2010). In essence, brokers benefit the
team largely by capturing novel market information,
which then enhances the identification of necessary
material resources (Ahuja 2000). Also, brokers have
the added capability of bridging competing parties,
which can produce financial advantages in negotia-
tions, such as obtaining favorable terms from investors
(Burt 1992, Pfeffer and Salancik 1978). While the bro-
ker’s diverse and disconnected relations facilitate the
search for available resources, the mobilization of these
resources is likely to be difficult and time consuming.
On the other hand, founders with relationships to
contacts who are densely connected benefit in that
their connections are characterized by high trust and a
greater motivation to cooperate (Ahuja 2000, Coleman
1988, Granovetter 1985, Oh et al. 2004, Shipilov and
Li 2008). In turn, this trust and cooperation enable
these founders to quickly marshal resources, especially
under uncertain conditions (Aldrich and Zimmer 1986,
Krackhardt 1999, Portes and Sensenbrenner 1993).
Direct and indirect relations to a founder enhance
the access to credible information about the founder,
which helps overcome the uncertainty of poten-
tial investors and prospective employees (Dahl and
Sorenson 2012, Ruef et al. 2003, Sorenson and Stuart
2001). Moreover, their nonbrokering relations facili-
tate both joint problem solving and the exchange of
private or tacit information, which are critical in the
early stages of an enterprise (Hansen 1999, Uzzi 1997).
Rather than simply being transactional, these early
investors and employees will attempt to find mutu-
ally beneficial solutions if the entrepreneur experiences
operational difficulties or financial setbacks (Uzzi and
Gillespie 2002). Finally, connections within a cohesive
group can guard individual entrepreneurs against eco-
nomic fraud by facilitating collective sanctioning of
opportunism (Coleman 1988, Granovetter 1985). These
forms of social safeguards often act as private-order
substitutes for absent or weak regulatory institutions
(Marquis and Raynard 2015). By way of an exam-
ple from economic history, the Maghribi trader coali-
tion illustrates how close-knit networks enabled the
sanctioning of opportunistic behavior by excluding
offenders from any future business with all members
of the coalition (Greif 2006). Hence, although founders
connected only to a cohesive group may have a limited
ability to search for resources throughout the market,
the resources they already have at their disposal may
be more readily mobilized via their connections.
Nonetheless, the social capital benefits brought about
by membership in a cohesive group may be under-
mined if the individual entrepreneur seeks to create
ties beyond the group in which he is heavily embed-
ded. Highlighting the double-edged sword of the social
capital generated from group cohesion, Portes and
Sensenbrenner (1993) point out that the solidarity
and norm enforcement that group cohesion facilitates
can also act to restrict ties to individuals outside of
group. Moreover, dense ties within a group coincide
with a strong social identity, which often leads these
group members to negatively stereotype and be sus-
picious of non-group members (Sherif et al. 1961, Rao
et al. 2000). Affiliations beyond the cohesive group
make an entrepreneur’s membership in that group less
certain, and ambivalence arises toward the entrepre-
neur because the group members question his com-
mitment to the group (Krackhardt 1999, Popielarz
and McPherson 1995, Portes and Sensenbrenner 1993).
When there is a strong in-group orientation, a single
member’s attempts to broker or create external connec-
tions beyond the group can consequently undermine
the group’s trust and motivation to provide support
to that member (Hillmann 2008, Shipilov and Li 2008,
Xiao and Tsui 2007). Thus, we argue that when indi-
vidual entrepreneurs tied to cohesive groups attempt
brokering successive connections beyond the group, it
can reduce their social capital within that group and
perhaps jeopardize it altogether.
This is not to suggest that the relationship of social
capital and the proportion of in-group ties to out-group
ties is linear and static for an individual founder. A one-
time founding beyond the cohesive group will prob-
ably not entirely compromise an individual founder’s
standing, but each additional external connection
potentially erodes it in a gradual manner (Portes and
Sensenbrenner 1993). Therefore, it would not behoove
these founders to attempt to become brokers them-
selves; instead, they should partner with one.
Clearly, both brokering ties and ties to a cohesive
group can foster the success of an entrepreneurial
endeavor, but through very different means. Consider
as an example an early-stage firm without capital back-
ing, which presents the greatest risk to investors. These
early investments are generally perceived to be riskier
largely because the firm has not yet been vetted by
other investors, and it is uncertain whether there will
be subsequent investors who support the new firm
(Eisenhardt and Schoonhoven 1990, Thornton 1999).
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
4Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s)
Faced with this liability of newness, the nonbroker
entrepreneur, embedded as he is in a tightly knit web
of connections, may rely on the reservoir of trust that
these ties provide to mobilize initial seed money. But in
this case, the small group of cohesive founders may not
have sufficient capital to fully fund the firm. This is a
situation where a broker can then leverage his diverse
ties to recruit the additional investors needed, a task
made easier given the early set of investors. Hence, the
complementarity of structural roles provides a potent
means for the members of an entrepreneurial team to
circumvent trade-offs and benefit from both brokerage
and cohesion.
Structural Role Complementarity
Previous research on networked teams has primar-
ily considered the external contacts that each indi-
vidual member contributes to the team collectively
(Ferriani et al. 2009, Hansen 1999, Oh et al. 2004,
Reagans et al. 2004, Uzzi and Spiro 2005). These
earlier studies frame a team’s relational characteris-
tics as either external (ties beyond the focal team)
or internal (ties within the team), and then exam-
ine the trade-off between the benefits of strong and
dense connections—cohesion—versus the advantages
of bridging connections—brokerage—for the team. For
example, prior research indicates that performance
benefits from coupling high internal cohesion among
team members with ties external to the team that
bridge disconnected others (Oh et al. 2004, Reagans
et al. 2004). Uzzi and Spiro (2005) focus exclusively
on the effect of teams’ external ties (past team affil-
iations) and find that triadic closure based on all of
the members’ external connections increases success,
but only up to a point. Vedres and Stark (2010) intro-
duce a somewhat different concept for team networks
(structural folds) to capture the degree to which team
members are embedded within multiple teams, and
they find that this overlap significantly improves the
team’s outcomes. Taken together, this earlier research
highlights that when examining the effects of networks
and teams, there are simultaneous internal and ex-
ternal dynamics that affect the performance of the
team. Our concept—structural role complementarity
extends this stream of literature by integrating external
connections with the team’s internal role composition.
Implicitly, these prior studies have assumed that
social relationships are transferrable among team
members and therefore can be aggregated to the team
level. This team-level aggregation not only leads to
the loss of detail but also obscures how individual
members vary in their brokerage ability (Vedres and
Stark 2010). Sorenson and Rogan (2014) argue, how-
ever, that the social connections of the individual mem-
bers should not necessarily be attributed directly to
the entrepreneurial firm, and that social ties are more
commonly “owned” by individuals rather than the
firm. In other words, entrepreneurs in a team can
employ their social ties in the service of the firm, but
the relationships are created, cultivated, and main-
tained by the individual entrepreneurs. Attributing
these external relationships specifically to the individ-
ual team members permits an enhanced understand-
ing of the uses of social capital within founding teams.
In particular, we can apply the concept of structural
roles to the entrepreneurial team members and catego-
rize each individual by the characteristics of their net-
works (White 2001, White et al. 1976). Structural roles
can be understood as reflecting equivalence in the gen-
eral pattern of relationships that entrepreneurs main-
tain, but this does not imply that two entrepreneurs
who share such roles (1) are directly connected to each
other or (2) that they are connected to exactly the same
set of other entrepreneurs. Our understanding of roles
is therefore closer to the concept of regular equiva-
lence than to the strict concept of structural equivalence
(Everett and Borgatti 1994, Faust 1988, Winship and
Mandel 1983). Thus, individual members on an entre-
preneurial team may be distinguishable from each
other by their different structural roles.
The literature suggests that team outcomes and pro-
cesses improve when members play different func-
tional roles based on their particular knowledge or
skills (Beckman 2006, Beckman and Burton 2008,
Bunderson 2003). In particular, these roles permit team
members to divide the required activities effectively
among the team as well as “play to each member’s
strength” (Ancona and Caldwell 1992). For example,
team members with strong social or language skills
may emerge as the team’s communication special-
ist, or inspirational members may become the infor-
mal team leader. This role differentiation can help the
team to develop greater skills, abilities, and knowledge,
and when the roles are complementary, it has been
linked to greater entrepreneurial performance (Beck-
man 2006, Beckman and Burton 2008, Eisenhardt and
Schoonhoven 1990, Ruef et al. 2003).
Given the complementary benefits of brokerage and
cohesion for entrepreneurship, we focus on the bro-
kerage and nonbrokerage roles within a team. Simi-
lar to functional roles, differentiation into structural
roles permits team members to specialize in comple-
mentary activities, such as maintaining brokerage rela-
tions or cohesive relations.1Moreover, the potential
trade-offs that an individual might experience between
brokerage and cohesive relations need not hinder a
team in which members specialize across these dif-
ferent but beneficial structural roles. Although it has
not been examined in the team context, we predict
that brokerage role diversity as a specific and particularly
important case of structural role complementarity will
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s) 5
enhance entrepreneurial teams’ functioning and out-
comes. While we specifically focus on brokerage role
diversity because of its significance for entrepreneurial
success, other contexts, such as project teams in orga-
nizations, might benefit from other forms of structural
role complementarity, such as member-level diversity
in the number of external connections or the average
path distance to non-team members.
A focus on the networks and roles of individual
founders reveals variation in brokerage that is not
necessarily evident when examining only firm-level
networks. To illustrate this point, consider the styl-
ized comparison of two firms and their respective
founders in Figure 1. The two panels on either side
of the figure show networks for two different firms
and their founders—one with high and the other with
Figure 1. Stylized Networks for Two Different Firms (Firm 1 and Firm 2) and Their Two Respective Founders (Founders A
and B and Founders C and D)
Bipartite networks of firms and founders
Firm network projections
Founder network projections
Founder A’s constraint = 0.617
Founder B’s constraint = 0.413
Founder C’s constraint = 0.413
Founder D’s constraint = 0.413
Firm 1’s constraint = 0.701 Firm 2’s constraint = 0.701
High brokerage role diversity
(founders’ constraint SD = 0.144)
Low brokerage role diversity
(founders’ constraint SD = 0.000)
Black indicates present cofoundings
Gray indicates prior cofoundings
Founder Firm
Notes. The contemporary founders and firms are highlighted in black, and each founder’s past cofoundings are presented in gray. The left-
hand panel presents Firm 1, which has high brokerage role diversity, whereas the right-hand panel presents Firm 2, which has low brokerage
role diversity. The first images in both panels represent the bipartite networks of both of the present and prior firms and their founders. The
second set of images in the left and right panels show the one-mode projection of the firm-to-firm networks. Both Firms 1 and 2 share identical
positions in the firm networks and share equal constraint values. The bottom images in the each of the panels show the one-mode projections
of the founder-to-founder networks. Firm 1’s founders’ constraint scores vary, but the constraint scores of Firm 2 founders do not.
low brokerage role diversity. The focal Firms 1 and 2
with their respective founders (A and B; C and D) are
highlighted in black at time t, and the surrounding
founders and firms depicted in gray represent previ-
ous cofoundings at time tn. Vertically, the right and
left panels depict the affiliations among these two firms
and their founders represented in three different man-
ners: the firm-founder bipartite network, the firm net-
work projections, and the founder network projections.
The first images of the bipartite networks of Firm 1 and
Firm 2 depict two founders who each have two prior
foundings. The second images for Firm 1 and Firm 2
in the panels are the conventional one-mode projec-
tions of the firm-to-firm networks, where firms are con-
nected via shared founders. As shown, both Firm 1
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
6Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s)
and Firm 2 share identical network positions in their
respective firm-level networks. From this perspective,
both firms have the same number of prior cofounding
ties (four) and, importantly, hold identical brokerage
positions. By contrast, the last images at the bottom
of the panels show the one-mode projections of the
founder-to-founder networks, where founders are con-
nected via firms that they cofounded together. As seen
in the founder-to-founder networks, Firm 1 exhibits
high brokerage role diversity among its founders, but
there is no variation in brokerage roles for members
of Firm 2. In other words, the members of Firm 1
(Founders A and B) vary in their pattern of connec-
tions and differ in their brokerage roles for the team,
whereas Firm 2’s founders (C and D) have identical
patterns of relations and therefore identical brokerage
roles. Hence, brokerage role diversity within a found-
ing team is a distinct and measurable construct that
captures variation in network positioning and relations
in ways that are not evident when focusing only on
the firm-to-firm network (details regarding our mea-
surement of brokerage role diversity follow below). We
therefore hypothesize that, in addition to the structural
advantages of brokerage that are evident in firm-level
network projections, measures of brokerage role diver-
sity in an entrepreneurial team will further predict
variation in the firm’s performance.
Data Sources and Measurement
An empirical assessment of our argument requires
both a measure of the team’s performance, such as
capital raised, and network data for a complete pop-
ulation of founders and their firms. Because tests of
our argument require accurate measures of team mem-
bers’ structural roles, complete data on founder-to-
founder networks are necessary to ensure that no con-
nections are artificially or systematically excluded from
the analysis. In addition, archival data that do not rely
on self-reports from entrepreneurs are less vulnerable
to systematic reporting biases of respondents (Bernard
et al. 1984). To complement these measures, the ideal
data set would also account for other attributes of
the entrepreneurial team members, such as their eth-
nic diversity and founding experience, which may
be confounded with network position. We employ
a data set that meets these challenging criteria: the
RUSCORP database collected by economic historian
Thomas Owen (1992) on all Russian incorporations
during the tsarist regime. During this period, across
the Russian empire, the establishment of the two domi-
nant forms of large corporations, the share partnership
and the joint-stock company, required the approval of
the central government. On such approval, a newly
established enterprise was granted a charter by the
tsarist administration. To the extent that no enterprise
was permitted to start its business until this adminis-
trative procedure was completed, our data set includes
all corporate foundings in Russia between 1869 and
1913 (Owen 1991).
Our data on the characteristics of corporate enter-
prises and the composition of their founding teams
come from the information recorded in these corpo-
rate charters. Approximately two-thirds of our initial
set of 3,762 newly founded corporate ventures were
established by teams of two or more entrepreneurs,
yielding a sample of 2,446 founding teams of corpora-
tions and their 9,461 individual founders, with com-
plete information on all of our variables of interest.
Although we investigate entrepreneurial teams specif-
ically, all firm foundings, including firms established
by a sole entrepreneur, are included in the founder-to-
founder networks. The firms are comprised of large-
scale industrial corporations and share partnerships
with capital-intensive operations, ranging from textile
mills and mining companies, to banks and railways.
The data set is left-censored at 1869 because, to the
best of our knowledge, a reliable price index to deflate
ruble values is not available for earlier years. The right-
censoring at the beginning of 1913 provides a natural
framing of our time window because it ends just before
the onset of World War I and the demise of Russia’s
political and economic systems.
Cofounding Network
While the original founders named on the corporate
charter retained their stock share issuance, the remain-
ing corporate shares were largely purchased by other
entrepreneurs, who sought board positions or voting
rights over managerial decisions (Owen 1991). Thus,
the cofounding networks are comprised of relations to
other founders through previous ventures, and these
connections would have been a crucial means to both
gain capital support and garner information on how
to successfully manage firms. Connections to other
founders provide not only opportunities for founders
to learn from each other but also valuable ties to those
most suited to offer relevant information, as in other
collaboration networks (Ahuja 2000, Ferriani et al. 2009,
Vedres and Stark 2010). Based on the charter infor-
mation on firm foundings, we code annual bipartite
networks of cofounding ties among the founders and
firms in our data set for the period 1869–1913. Each
annual affiliation network includes all of the founders
and their firms that had been established up until
the current year.2To facilitate measures of individual
brokerage roles within teams, we then generate the
founder-by-founder networks from the bipartite net-
works. Pairs of founders are linked to the extent that
they were partners in a prior entrepreneurial founding.
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s) 7
Basic capital. As our dependent variable, we use the
ruble amount of basic capital recorded in the firm’s cor-
porate charter. The basic capital recorded in the charter
included all shares and financial commitments of the
firm and is analogous to an initial public offering used
in the context of present-day firms (Stuart et al. 1999,
Stuart and Sorenson 2003, Thornton 1999). Tsarist cor-
porate law stipulated that only teams that were able
to accomplish these declared funding commitments
within a specified time were subsequently permitted
to begin operations. Any changes in the membership
of the founding teams during the process of realiz-
ing the capital commitments would have precipitated
a recharter document from the tsarist office, and such
recharters rarely occurred (fewer than 4% of the char-
ters in the data set). The basic capital is among the most
important responsibilities of these founding teams and
is commonly used as a performance measure for these
firms (Carstensen 1983, McKay 1970, Owen 2005). In
addition, systematic evidence from our Russian his-
torical setting indicates that greater basic capital sig-
nificantly increased the longevity of otherwise com-
parable firms (see the online supplementary materials
Appendix A).
As the kind of ruble—silver, copper, or paper assig-
nat—and the values of shares routinely varied from
charter to charter, even within the same year, all cap-
ital values are normalized according to the standard
ruble of account (Owen 1992). We then deflated all nor-
malized capital values using the standard Saint Peters-
burg Institute of Economic Research retail price index
(Gregory 1982). All capital values are denoted in thou-
sands of rubles with 1913 as the base year. Where basic
capital consisted of both stocks and bonds, the sum of
both amounts is used.3To account for the variable’s
positive skew, it is log transformed.
Brokerage role diversity. This variable is our instanti-
ation of structural role complementarity and our main
explanatory variable. We use Burt’s (1992) constraint
measure to assess each founder’s capacity to bridge
across the structural holes that separate other founders.
In our case, constraint measures the degree to which
a founder’s previous partners were also connected to
each other through prior cofoundings. A founder i’s
structural constraint is defined as
j1Pij +
Piq Pq i 2
where Pij represents the proportion of founder i’s be-
longing to founder j. The sum, Pn
q1Piq Pq i , captures the
indirect connections of ito qthat are also connected
to j. Brokerage opportunities for a focal founder arise
when triadic closure among his previous partners is
low, and consequently, constraint approximates a value
of zero. In contrast, a founder’s brokerage opportuni-
ties decrease as triadic closure increases, as reflected
by higher constraint. Notably, the lower bound for this
measure is zero, and the upper bound is 1.125 (Buskens
and van de Rijt 2008).
To investigate the role of variance of brokerage in
determining founding team performance, we followed
the recommendation of Sørensen (2002) and calculated
for each team both the mean and standard deviation for
the founding members’ constraint while also including
the interaction of these effects as a term in the model.
This approach demonstrates the effect of increasing
heterogeneity of the founding members’ constraint
scores (i.e., brokerage role diversity) while accounting for
variation in the founding team’s mean constraint scores
(Reagans et al. 2004).
Covariates. The RUSCORP data set, as coded from the
corporate charters, provides detailed information on
control variables that potentially predict basic capital.
Team size counts the number of partners on a team
at the time of founding and was log transformed to
address the variable’s positive skew. We also control
for the two legal forms of business organizations found
in these capital markets, the joint-stock company and
the share partnership, using a binary indicator for joint-
stock organization. Following Dahl and Sorenson (2012),
we use fixed effects to control for the year in which
firms were incorporated and the geographic regions in
which they were based. Finally, we use fixed effects to
denote the industry sector in which the new company
operated. Tables 1and 2provide descriptive informa-
tion for the firms.
We also include models in which we consider alter-
native explanations to brokerage role diversity. Because
ethnic diversity has been found to negatively affect
team performance (Reagans et al. 2004, Ruef et al.
2003), we use Blau’s (1977) heterogeneity index as a
measure of ethnic diversity within each team, where
a value of zero indicates ethnic homogeneity. Eth-
nic diversity is based on each founder’s ethnicity (see
Table 2), as the tsarist administration mandated that
it be recorded within the charters. Also, to account
for the possibility that team-level outcomes reflect the
members’ first-order network connections, we specify
models that include the teams’ mean and standard devia-
tion of degree centrality and the interaction of these main
effects (note that diversity in members’ degree cen-
trality may be regarded as another measure of struc-
tural role complementarity). Alternative explanations
also include the team’s overall experience and learning
from earlier entrepreneurial activities, which have been
found to improve success and innovation (Eisenhardt
and Schoonhoven 1990, Uzzi and Spiro 2005). For this
reason, and also because founders with little to no
experience in previous ventures necessarily will have
higher constraint scores, we control for the experience
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
8Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s)
Table 1. Firm and Founder Summary Statistics
Mean SD Min Max
Firm summary statistics (n2,446 teams
comprised of 9,461 unique founders)
(1) Basic capital in ’000 6.748 1.100 1.113 11.467
rubles (logged)
(2) Brokerage role diversity 0.057 0.119 0.000 0.679
(3) Constraint mean 0.841 0.255 0.057 1.132
(4) Team size (logged) 1.213 0.562 0.693 4.248
(5) Joint-stock organization 0.642 0.479 0 1
(6) Founding year 1896 13 1869 1913
(7) Ethnic diversity 0.317 0.341 0.000 0.833
(8) Degree centrality 1.174 3.779 0.000 69.296
standard deviation
(9) Degree centrality mean 3.930 4.831 1.000 69.000
(10) Total experience 0.622 1.501 0.000 16.000
(11) Percentage of serial founders 0.112 0.221 0.000 1.000
(12) Firm constraint 1.659 0.544 0.173 2.000
(13) Firm constraint dummy 0.705 0.456 0 1
Founder summary statistics
(n10,499 founder-firm observations)
(1) Founder constraint 0.68 0.34 0.03 1.125
(2) Founder degree centrality 7.96 11.01 0 121
(3) Founder experience 0.17 0.66 0 11
Notes. Capital amounts are standardized and deflated to the 1913
ruble. Founder variables are calculated for each founder at the time
of the founding. The values for serial founders therefore vary over
of the team members. For each team, we include total
experience, specifically the sum of previous foundings
by the respective team members, and an additional
variable for the percentage of serial founders on the team.
Finally, we have noted that brokerage role diversity is
a distinct construct from measures that capture the
brokerage positions seen in firm-level projections of
the founding networks. Whereas we have used Burt’s
(1992) constraint measure to estimate the brokerage
potential of individual founders, an analogous mea-
sure of firm-level brokerage has frequently been cal-
culated from firm-to-firm projections of the network
(Ferriani et al. 2009, Reagans et al. 2004, Shipilov and
Li 2008). We likewise calculated a measure of firm con-
straint, as derived from symmetric network matrices
where teams (firms) are the nodes and two teams are
linked if they share one or more founding members. In
some cases, however, new founding teams lack prior
ties to other firms in the network, which poses a chal-
lenge for this measure because supplying a value of
zero inappropriately implies perfect brokerage. There-
fore, following the suggestion of Buskens and van de
Rijt (2008), we assign a value of two to these firms,
reflecting the lack of brokerage opportunities beyond
the firm. Because this may be regarded as an extreme
value for firm constraint, however, we also generated
an additional binary variable that explains variance
unique to these isolated teams.
Table 2. Firm Region, Firm Industry, and Founder Ethnicity
Regional location of firm headquarter (n2,446 firms)
Caucasus 7.28
Center 19.30
Central Asia 1.55
Finland 0.45
North 20.81
Poland 12.76
Siberia 8.30
South 22.44
Volga-Ural 6.26
Foreign countries 0.61
In Russian empire, but exact location unknown 0.25
Industrial sector of firm (n2,446 firms)
Construction 1.84
Finance 12.31
Manufacturing 64.55
Mining 5.40
Public administration 0.20
Transportation 10.51
Wholesale 5.07
Unclassifiable 0.12
Founder ethnicity (n9,461 unique founders)
Armenian 2.09
French 1.98
German 17.46
Greek 1.25
Jewish 10.07
Norwegian 7.99
Russian 43.69
Tatar 1.27
Other or unknown 14.20
Models and Estimation
Two sets of analyses comprise our empirical strategy.
In the first set, we examine the relationship between
basic capital and brokerage role diversity using a vari-
ant of multilevel modeling that permits us to parti-
tion the variance associated with individual founders.
Employing the same modeling strategy, we also con-
sider alternative explanations to brokerage role diversity,
such as other forms of team heterogeneity. Next, as
a means to address potential endogeneity, we imple-
ment an instrumental variable (IV) analysis based on
governmental restrictions on the rights of firms with
Jewish and foreign founders to hold property in certain
regions. For this latter analysis, we examine both the
complete sample and samples of matched firms. The
use of an IV allows us to address two potential sources
of biases in the multiple membership model results.
The first is an omitted variable bias, where another
variable might simultaneously influence both brokerage
role diversity and basic capital and thus confound the
results. The second is a reverse causality bias, where
greater demands for basic capital might lead teams to
become increasingly diverse in brokerage ability.
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s) 9
Multiple Membership Models
Organizational research on team performance has con-
ventionally relied on fixed effects to account for unob-
served heterogeneity introduced by the participation
of individual members on multiple teams (e.g., Rea-
gans et al. 2004). In the RUSCORP sample of entrepre-
neurial firms, however, the number of serial founders
(n598) relative to the number of firms is high, and
models with fixed effects therefore risk overfitting on
the available data (Greene 2008). In addition, fixed
effects implicitly assume that the individuals’ impacts
on team-level outcomes are invariant across teams of
different sizes, and there is collinearity relating to the
cooccurrence of serial founders on successive found-
ing teams. For these reasons, we draw on advances
in multilevel modeling to account for individual-level
The data structure in this analysis differs from the
most common applications of multilevel modeling, in
which outcomes are measured at the lowest unit of
analysis (e.g., the performance of individuals, clus-
tered in firms). By contrast, our dependent variable is
a team-level outcome, and we therefore employ “mul-
tiple membership” models to partition the variance
of member-level contributions (Browne et al. 2001). In
these models, each team is characterized by a combi-
nation of random effects (varying intercepts) for each
team member, inversely weighted by the number of
team members. The random effects structure in these
multiple membership models is formally given by the
following notation, where for simplicity we present
the “intercept-only” version of the model with no
where yidenotes the (log-transformed) basic capital
raised by team i(i1, . . . , 2,446),β0denotes the inter-
cept, Pjmember(i)w(2)
jdenotes the weighted average
of the random effects for that team’s set of participat-
ing members, and eidenotes the residual. The sub-
script member(i)is a classification function that returns
the set of individuals who belong to team i. Thus,
the summation sums only over the set of individuals
who are in team i, as opposed to, say, all individu-
als in the data. Hence, although the multiple mem-
bership models can be interpreted similarly to ordi-
nary least-squares (OLS) regression models, they can
explain variance at the two hierarchical levels and pro-
vide more conservative estimates of the independent
variables.4To reduce bias in the variance estimates, we
follow the recommendation of Browne et al. (2001) by
fitting our models using Markov chain Monte Carlo
(MCMC) methods while specifying “flat” priors for all
Instrumenting for Brokerage Role Diversity and
Basic Capital
During our observation period, the tsar directed local
magistrates to enact legal restrictions on the rights
of Jewish and foreign entrepreneurs. In addition to
documenting ethnicity and citizenship of the found-
ing team, founding charters listed any legal restric-
tions regarding the corporations’ right to own prop-
erty. These restrictions banned firms with either Jewish
or foreign entrepreneurial members from owning, pur-
chasing, or leasing land in regions along the southern
and western borders of the empire. Accordingly, these
bans limited Jewish and foreign founders from partic-
ipating in companies that would require any property
ownership, such as manufacturing firms (Owen 1991).
In other words, these restrictions on ownership would
have influenced the selection set of potential entrepre-
neurial partners for founding teams and therefore our
variable of interest, brokerage role diversity, indicating
a viable instrument. In particular, we anticipate that
these property restrictions would increase brokerage role
diversity for two reasons. First, although the effect is
not statistically significant, for firms founded prior to
1887, before the restrictions were imposed, brokerage
role diversity is negatively correlated with the number
of Jewish and foreign team members. Second, the law
may have encouraged founders embedded in a cohe-
sive group that included Jews or foreigners to partner
beyond their focal group. In other words, because the
law would have limited partner choices within such
cohesive groups, some members might be forced to
select partners outside of their tightly interconnected
group, which would increase the opportunities for
founders with little brokerage potential to partner with
potential brokers outside of their preferred group.
For these restrictions to serve as an instrument in
this analysis, we also have to confirm that these restric-
tions did not affect basic capital directly. We therefore
replicated the full model (see Model 4 in Table 3) for
firms founded prior to 1887 and included two addi-
tional variables, the number of Jewish members and the
number of foreign members on the team. Neither the num-
ber of Jewish nor foreign members on a team signifi-
cantly predicted its basic capital, suggesting that subse-
quent restrictions on their involvement did not affect
the capital raised (see the online supplementary mate-
rials Table S.3). In addition, to assess the robustness of
the IV analysis in the case that our instrument, Jewish
and foreign property restrictions, does not perfectly satisfy
the exclusion restriction, we use the approach intro-
duced by Conley et al. (2012). This approach allows us
to estimate the extent to which the exclusion restriction
can be relaxed for the two-stage least squared (2SLS)
brokerage role diversity estimators to remain informative.
The results reported in online supplementary materials
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
10 Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s)
Appendix B indicate that even with substantial depar-
tures from the exclusion restriction, the Jewish and for-
eign property restrictions remains a viable instrument.
Another concern is that possible differences in the
types of firms founded by teams with Jewish or for-
eign members served to also reduce the basic capital
that the firms raised. Hence, we created a matched
sample of 1,035 firms using coarsened exact matching
(CEM) to define comparable sets of firms. A strength
of examining this matched sample is that the distri-
butions of the covariates for the control firms (i.e.,
firms founded before the law) and the treated firms
(i.e., firms founded after the law) are largely equivalent
owing to the mechanics of the CEM approach. Thus, by
construction, the CEM technique addresses potential
balance issues in the sample. We use region,firm indus-
try,team size, and joint-stock organization to balance the
We also investigate similar models to the main multi-
ple membership models reported with some minor dif-
ferences. First, we choose not to include the interaction
term of brokerage role diversity and constraint mean be-
cause interaction terms would have also included our
Table 3. Multiple Membership Estimates of Basic Capital Raised by Founding Teams (n2,446 teams)
(0) (1) (2) (3) (4)
Intercept 6.739∗∗∗ 6.688∗∗∗ 6.729∗∗∗ 6.543∗∗∗ 6.970∗∗∗
(0.023) (0.024) (0.092) (0.238) (0.345)
Brokerage role diversity 1.002∗∗∗ 3.938∗∗∗ 4.193∗∗∗ 4.029∗∗∗
(0.192) (1.066) (1.101) (1.067)
Constraint mean 0.051 0.141 0.032
(0.100) (0.176) (0.169)
Brokerage role diversity ×
Constraint mean 4.479∗∗ 5.111∗∗ 4.925∗∗∗
(1.538) (1.533) (1.477)
Team size (logged) 0.062 0.112
(0.071) (0.069)
Joint-stock organization 0.540∗∗∗ 0.524∗∗∗
(0.045) (0.051)
Region controls (indicator coding) No No No No Yes
Year controls (indicator coding) No No No No Yes
Firm industry controls No No No No Yes
(indicator coding)
Random effects
Founders 0.313 0.189 0.175 0.129 0.124
(0.129) (0.119) (0.116) (0.103) (0.096)
Residual 1.104 1.133 1.133 1.082 0.975
(0.052) (0.051) (0.051) (0.047) (0.042)
DIC 7,391.883 7,374.801 7,367.627 7,231.059 7,042.001
Notes. The dependent variable is the logged basic capital of firms, standardized and deflated to the 1913 ruble. Models
are estimated via MCMC estimation, with computationally efficient parameter expansion of the member-level random
effects to promote good mixing of the chains (Browne 2004). We run the models for 250,000 iterations, discarding the first
50,000 iterations as the “burn-in” period that allows the chains to converge to their stationary distributions. Estimates
in this table are the means and standard deviations of the 200,000 iterations, which are analogous to the parameter
estimates and standard errors obtained in a frequentist analysis. To compare models, we report the deviance information
criterion, or DIC, with lower values indicating preferred models that effectively balance overparameterization against
model fit.
p<0.05;∗∗p0.01;∗∗∗ p0.001.
endogenous variable. Although there are techniques
to address endogenous interaction terms, these results
are difficult to interpret within IV models (Angrist and
Pischke 2008), and the interaction term is not central
to our theory. Second, we specified clustered robust
standard errors on years rather than including year
fixed effects because our instrument occurs only after
1887. In addition to the two-stage least squared (2SLS)
regressions, we estimate Moreira’s conditional likeli-
hood ratio (CLR) models (Moreira 2003). 2SLS esti-
mates are very sensitive to finite-sample bias, whereas
Moreira’s CLR models condition the critical values to
obtain correct significance levels (Hahn and Hausman
2005). The strength of the CLR approach is that it
draws correct inferences regardless of the instrument’s
strength (Bascle 2008).
Variance Components
Our results begin with Model 0 in Table 3, which is
an “intercept-only” model that includes only the ran-
dom effects, not the fixed effect covariates. We use
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s) 11
this model to calculate variance partition coefficients
(VPCs). These VPCs allow us to quantify the rela-
tive importance of individual founders and team-level
variables as sources of variation in basic capital raised.
In Model 0, the estimated variance for the founder-
level random effects is 0.313, and the residual variance
is 1.104. We use expressions from Leckie and Owen
(2013) to calculate the VPC for the founder-level vari-
ance while accounting for team size. Assuming inverse
weights for the median team size of three, the numera-
tor to calculate the VPC is (((1
3)2) × 0.313
0.104), and the denominator is the sum of the calcu-
lated founder-level variance and the residual variance
3)2) × 0.313)+1.104 1.208). Therefore,
the intercept-only model shows that approximately 9%
(0.104/1.208 0.086)of the variance in the data set can
be attributed to heterogeneous founder-level effects.
In other words, the member-level VPC indicates that
heterogeneity among individual entrepreneurs is only
a nominal determinant of team performance. Instead,
most of the variation in this data set is at the level of
the outcome variable (i.e., the residual variance), which
suggests that team-level characteristics (such as broker-
age role diversity) are the primary source of variation in
the mobilization of capital (VPC 0.914). As predictor
variables are added to models in Table 3, the founder-
level variance diminishes further, which indicates that
the covariates explain much of the heterogeneous suc-
cess of individual founders’ teams. For example, the
founder-level variance is reduced from 0.313 in the
intercept-only model to 0.124 in Model 4. Thus, not
only is the majority of variance explained by the team-
level factors, but over half of the founder-level variance
is accounted for by the explanatory variables. Overall,
these estimates of founder-level variance suggest that
individual founders’ attributes, such as their ability,
learning from past ventures, or other endowments of
resources, only marginally account for the success of
the team.
Multiple Membership Estimates
To reiterate, we argue that structural role complemen-
tarity of an entrepreneurial team in the form of broker-
age provides a benefit for the mobilization of basic cap-
ital. In Table 3, we present a series of models beginning
with the effect of brokerage role diversity, subsequently
including covariates that are potentially confounded
with this variable.
Model 1 includes our main variables of interest, bro-
kerage role diversity, which is positive and significant
(β1.002). In other words, this model predicts that as
founding teams move from a complete lack of diversity
to one standard deviation above the mean in broker-
age role diversity, the expected capital increases from
802,715 rubles to 957,525 rubles, which represents a
notable increase of 19% for teams with greater varia-
tion in brokerage.
Figure 2. (Color online) Plot of Basic Capital in Rubles
for New Firms by Constraint Mean for Teams with
the Predictions for High (95th Percentile) and Low
(5th Percentile) Brokerage Role Diversity, as Based on
Parameters from Model 2 in Table 3
Constraint mean
Capital (rubles)
0.1 0.3 0.5 0.7 0.9 1.1
High brokerage role diversity
Low brokerage role diversity
Note. Points have been jittered to promote visualization and propor-
tionally sized to the value of brokerage role diversity.
While the main effect of constraint mean is not signif-
icant in Model 2, the negative and significant interac-
tion term in the model (β4.479) indicates that the
effects of brokerage role diversity are contingent on the
team’s constraint mean. To illustrate this interaction, Fig-
ure 2plots basic capital as a function of constraint mean
for teams with both high and low brokerage role diver-
sity. As the plot suggests, the benefits to high brokerage
role diversity are realized primarily when the constraint
mean score is below approximately 0.8. When the con-
straint mean score is high (>0.8), by contrast, greater
variation in team members’ constraint scores is not
beneficial to the team’s performance. In other words,
increased heterogeneity in the members’ network bro-
kerage potential improves performance for the firm
when the team is largely comprised of brokers rather
than nonbrokers, on average.
Our findings in columns (3) and (4) in Table 3fur-
ther show that the estimates for brokerage role diversity
remain robust and positive with the inclusion of addi-
tional covariates. In Model 3, although team size is not
significant for basic capital, this variable becomes more
informative with the inclusion of additional controls in
Model 4 (β0.112). Model 4 indicates that as teams
increase in size from two to six members, the expected
capital declines from 1,389,433 rubles to 887,716 rubles
while holding other predictors at their mean values.
The inclusion of joint-stock organizations in Model 3
shows that the organizational form is positively associ-
ated with basic capital (β0.540). Model 4 adds regional,
year, and firm industry controls, and with the inclusion
of these covariates, the main effect of brokerage role diver-
sity remains positive and informative (β4.029), and
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
12 Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s)
the interaction term remains comparable in magnitude
Alternative Explanatory Factors
In Table 4, we extend the models to consider poten-
tial alternative explanations for the effects of brokerage
role diversity on basic capital. One possible explanation
Table 4. Multiple Membership Estimates of Basic Capital Raised by Founding Teams (n2,446 teams)
Extended Models
(5) (6) (7) (8) (9)
Intercept 7.073∗∗∗ 7.072∗∗∗ 6.753∗∗∗ 7.136∗∗∗ 7.050∗∗∗
(0.337) (0.348) (0.356) (0.399) (0.435)
Brokerage role diversity 4.127∗∗∗ 5.049∗∗ 4.348∗∗ 4.303∗∗ 4.534∗∗
(1.070) (1.608) (1.582) (1.590) (1.655)
Constraint mean 0.051 0.070 0.204 0.254 0.240
(0.167) (0.177) (0.185) (0.188) (0.192)
Brokerage role diversity ×
Constraint mean 5.036∗∗ 6.229∗∗ 5.501∗∗ 5.725∗∗ 6.116∗∗
(1.483) (2.092) (2.071) (2.088) (2.220)
Team size (logged) 0.1180.087 0.018 0.019 0.023
(0.068) (0.078) (0.084) (0.086) (0.087)
Joint-stock organization 0.545∗∗∗ 0.545∗∗∗ 0.524∗∗∗ 0.525∗∗∗ 0.524∗∗∗
(0.050) (0.051) (0.051) (0.052) (0.051)
Ethnic diversity 0.240∗∗∗ 0.238∗∗∗ 0.228∗∗∗ 0.232∗∗∗ 0.231∗∗∗
(0.063) (0.063) (0.062) (0.062) (0.064)
Degree centrality standard deviation 0.003 0.0350.0420.042
(0.018) (0.019) (0.019) (0.019)
Degree centrality mean 0.006 0.013 0.010 0.010
(0.010) (0.010) (0.010) (0.010)
Degree centrality standard deviation ×
Degree centrality mean 0.000 0.000 0.001 0.001
(0.000) (0.000) (0.000) (0.000)
Total experience 0.110∗∗∗ 0.092∗∗∗ 0.099∗∗
(0.026) (0.027) (0.030)
Percentage of serial founders 0.256 0.037 0.017
(0.166) (0.195) (0.198)
Firm constraint 0.1960.076
(0.090) (0.255)
Firm constraint dummy 0.139
Region controls (indicator coding) Yes Yes Yes Yes Yes
Year controls (indicator coding) Yes Yes Yes Yes Yes
Firm industry controls (indicator coding) Yes Yes Yes Yes Yes
Random effects
Founders 0.119 0.113 0.058 0.065 0.063
(0.094) (0.093) (0.063) (0.069) (0.067)
Residual 0.970 0.973 0.977 0.972 0.973
(0.042) (0.042) (0.036) (0.036) (0.035)
DIC 7,026.832 7,031.341 7,000.552 6,996.847 7,000.239
Notes. The dependent variable is the logged basic capital of firms, standardized and deflated to the 1913 ruble. Models
are estimated via MCMC estimation, with computationally efficient parameter expansion of the member-level random
effects to promote good mixing of the chains (Browne 2004). We run the models for 250,000 iterations, discarding the first
50,000 iterations as the “burn-in” period that allows the chains to converge to their stationary distributions. Estimates
in this table are the means and standard deviations of the 200,000 iterations, which are analogous to the parameter
estimates and standard errors obtained in a frequentist analysis. To compare models, we report the deviance information
criterion, or DIC, with lower values indicating preferred models that effectively balance overparameterization against
model fit.
p<0.05;∗∗p0.01;∗∗∗ p0.001.
is that brokerage role diversity may simply reflect ethnic
differences among the groups, which potentially pro-
vide them with connections to different ethnic groups
(Ruef et al. 2003). As in previous studies, the addi-
tion of ethnic diversity in Model 5 significantly reduces
the basic capital raised by teams (β0.240), but the
effects of brokerage role diversity nevertheless remain
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s) 13
consistent. According to the results of Model 5, ethni-
cally homogeneous teams will outperform teams with
a diversity value of 0.5 by 189,832 rubles, while hold-
ing all other predictors in the model at their means.
Another alterative explanation would suggest that sim-
ply any variation in network characteristics, such as the
number of ties between founders (degree centrality),
might supplant brokerage role diversity in explaining the
basic capital. This is not the case. Model 6 shows that
the teams’ variation in degree centrality (degree cen-
trality standard deviation) (β0.003), degree centrality
mean (β0.006), and the interaction of these terms
(β0.000) are not strong predictors of basic capital.
Although this result suggests that diversity in degree
centrality is relatively unimportant for entrepreneurial
teams, we note that this form of structural role com-
plementarity could be an important predictor of team
performance in other settings.
Earlier research also suggests that the founders’
prior experience would be advantageous for the entre-
preneurial firms (Eisenhardt and Schoonhoven 1990,
Uzzi and Spiro 2005). Model 7 includes both total expe-
rience and percentage of serial founders. Although the
cumulative experience of all of the founders was signif-
icant and positive for predicting basic capital (β0.110),
the ratio of serial founders relative to rookies (β0.256)
is evidently not predictive. The interpretation of broker-
age role diversity, however, is unchanged.
In addition, we show that brokerage role diversity is
an independent construct from the aggregate of team
member ties or firm-level network constraint. There-
fore, we also include the brokerage position of the firm
within the firm-to-firm network. Model 8 indicates that
increases in firm constraint reflect a significant decline
in the basic capital (β0.196). These results support
insights from earlier research that high brokerage of
the team within the firm-to-firm network is beneficial
to its success (Ahuja 2000, Burt 2005, Shipilov and Li
2008, Stuart et al. 1999, Stuart and Sorenson 2005).
Finally, because we substitute the value of two for
firm constraint only in the case of the isolated firms in
the cofounding network, as recommended by Buskens
and van de Rijt (2008), we introduce a dummy variable,
firm constraint dummy, in Model 9. This variable distin-
guishes the isolated firms with substituted values for
firm constraint and therefore reveals the sensitivity of
model parameters to this substitution. Because the sig-
nificant effects of brokerage role diversity remain consis-
tent in both Models 8 and 9, our findings for brokerage
role diversity are robust to the inclusion of firm constraint
as a predictor, whether or not the isolated firms are
In summary, across these additional models, broker-
age role diversity consistently exhibits a significant and
positive effect on basic capital. While these latter mod-
els incorporate variables based on predictions of earlier
research, including the ethnic diversity of team mem-
bers, founder experience, and firm-level brokerage, the
capital advantage from greater brokerage role diversity
among team members nevertheless remains robust.
Instrumental Variable Estimates
The instrument Jewish and foreign property restrictions
has a Kleibergen and Paap (2006)F-statistic of 10.47 in
the (2SLS) regression for the full sample. This value is
above the Stock and Yogo (2005) critical value of 8.96
to have no more than 15% of the bias of OLS estimates
with one endogenous variable and one instrument. For
the matched sample (Models 3 and 4), the critical value
is reduced to 7.044 and therefore has no more than 20%
of the bias of OLS estimates. Nevertheless, the identi-
cal 2SLS and Moreira’s CLR coefficients presented in
Table 5indicate that our results are vulnerable nei-
ther to a finite-sample problem or to a weak-instrument
problem (Yogo 2004).
For both the full sample and matched sample, we
present the first-stage models followed by the 2SLS
and CLR regressions in Table 5. The coefficient for Jew-
ish and foreign property restrictions in the first stage is
positive for both samples, as we anticipated (Models 1
and 4). The legal restrictions on the potential set of
founders are positively correlated with brokerage role
diversity for teams. Both of the 2SLS and CLR results
in Models 2, 3, 5, and 6 lend support to our argument
that brokerage role diversity is associated with greater
mobilization of basic capital.
Finally, Appendix B in the online supplemental
material reports the Conley et al. (2012) IV sensitiv-
ity analyses for both the full and matched sample. The
results indicate for the full sample that conservatively
at least 41% of the reduced form effect of Jewish and for-
eign property restrictions on basic capital would have to be
through channels other than brokerage role diversity for
the IV results to be insignificant. In other words, this
sensitivity check suggests that even with potentially
large violations of the exclusion restriction, where the
Jewish and foreign property restrictions might have a rel-
atively large direct impact on basic capital, the positive
estimates of brokerage role diversity in Table 5remain
statistically significant.
We present a novel means of capturing the varia-
tion in the portfolios of structural roles of entrepre-
neurial team members, and we demonstrate how this
variation helps to explain firms’ performance in ways
that are not evident when examining only connections
among firms. In our empirical case, we have examined
returns to structural role complementarity in terms
of the basic capital that entrepreneurial teams attract
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
14 Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s)
Table 5. Instrumental Variable Estimates of Basic Capital Raised by Founding Teams
(1) (2) (3) (4) (5) (6)
Full sample (n2,446 teams) Matched sample (n1,035 teams)
2SLS model CLR model 2SLS model CLR model
First stage basic capital basic capital First stage basic capital basic capital
Jewish and foreign property 0.025∗∗ 0.021∗∗
restrictions (0.008) (0.008)
Brokerage role diversity 9.773∗∗ 9.773∗∗ 10.90010.900
(3.958) (4.046) (5.271) (5.056)
Constraint mean 0.430∗∗∗ 3.675∗∗ 3.675∗∗ 0.463∗∗∗ 4.454 4.454
(0.016) (1.714) (1.748) (0.029) (2.515) (2.388)
Team size (logged) 0.160∗∗∗ 1.350∗∗ 1.350∗∗ 0.092∗∗∗ 0.9650.965
(0.007) (0.637) (0.654) (0.011) (0.478) (0.479)
Joint-stock organization 0.007 0.410∗∗∗ 0.410∗∗∗ 0.0130.099 0.099
(0.004) (0.075) (0.069) (0.006) (0.153) (0.115)
Region controls Yes Yes Yes Yes Yes Yes
(indicator coding)
Firm industry controls Yes Yes Yes Yes Yes Yes
(indicator coding)
KP Wald F-statistic 10.471 7.044 Yes
R20.444 0.443
Year clusters (n) 44 44 44 43 43 43
Moreira’s CLR [3.374, 23.534] [3.173, 32.206]
p-value (0.002) (0.005)
Note. The dependent variable is the logged basic capital of firms, standardized and deflated to the 1913 ruble.
p<0.05;∗∗p0.01;∗∗∗ p0.001
from investors. Using systematic evidence from corpo-
rate foundings, our findings support our main argu-
ment that entrepreneurial teams with greater broker-
age role diversity among the founding partners realize
higher capital. The effects of brokerage role diversity
are further moderated by the average brokerage poten-
tial among the team members. That is, when founding
team members are comprised of members with high
brokerage ability on average, higher levels of broker-
age role diversity predict greater basic capital. Taken
together, these findings suggest that brokerage role
diversity, as a separate construct from firm-level bro-
kerage, enhances entrepreneurial success via increases
in starting capital. Our findings are robust in models
that consider the brokerage position of the firm, the
experience of the team members, and the competing
categorical characteristics of founding team members.
Our results do not contravene previous studies
showing the value of network brokering for firms
(Ahuja 2000, Burt 2005, Shipilov and Li 2008). While
we also find benefits to brokerage for entrepreneur-
ial firms, our research suggests that the benefits are
enhanced when brokerage ties are differentially dis-
tributed amongst the team members. Moreover, we
are not the first to suggest that brokerage and cohe-
sion are complementary constructs; however, prior re-
search tends to consider complementarity primarily at
the individual level rather than how it might operate
across individuals within a team, as we have done here
(Burt 2005, Vedres and Stark 2010). In addition, stud-
ies that have focused on the networks for teams typ-
ically define complementarity as cohesive ties inside
the team paired with brokerage ties outside of the team
(Oh et al. 2004, Reagans et al. 2004), whereas we focus
on the composition of structural roles among the team
members. In sum, brokerage role diversity extends our
understanding of external brokerage ties by suggest-
ing that firms benefit when the extent of brokering ties
beyond the team is heterogeneously distributed among
its members.
Drawing on previous research on individual-level
trade-offs between cohesive and brokering ties (e.g.,
Portes and Sensenbrenner 1993), we argue and find
evidence for how teams might circumvent this trade-
off by permitting different team members to assume
the roles of maintaining cohesive and brokering ties,
respectively. However, additional research is needed
to enrich our understanding of the underlying pro-
cesses of brokerage role diversity that afford benefits
to entrepreneurial teams. Our contention is that struc-
tural roles allow a network “division of labor” where
the ability to mobilize resources arises from differenti-
ating members based on the pattern of their relation-
ships. We anticipate opportunities for future studies
to elucidate the specific mechanisms by which these
teams benefit their firms’ success.
Our empirical emphasis is on the particular his-
torical case of entrepreneurship in Russia because it
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s) 15
is ideally suited for understanding founding teams
in emergent economies, which are rapidly increas-
ing in economic importance (International Monetary
Fund 2016). While the insights from our setting might
extend to founding teams in established economies,
we anticipate the findings to be the most robust
for contemporary emergent economies. Entrepreneurs
in emergent economies past or present often face
similar conditions of limited institutional infrastruc-
ture and volatile markets (Nee and Opper 2012).
A growing body of research indicates that in such
settings, entrepreneurs must rely more on alterna-
tive strategies, such as leveraging personal connec-
tions to aid in mobilizing resources, than those in
developed economies (Eisenhardt and Schoonhoven
1990, Khanna and Rivkin 2006, Vedres and Stark 2010).
Moreover, contemporary emerging economies are typ-
ically comprised of largely manufacturing and labor-
intensive industries, much like firms founded in our
empirical setting (Marquis and Raynard 2015). Given
the commonalities of emerging economies and the
challenges that they present for entrepreneurship, we
expect that the returns to brokerage role diversity in
founding teams would be particularly applicable to
firms in modern emergent economies.
Three main contributions arise from our study. First, by
integrating structural roles with team composition, this
new construct provides a means to uncover important
group processes from a network perspective. Our the-
orizing considers the relational patterns of individual-
level differences concurrently with team dynamics.
More importantly, the construct of structural role com-
plementarity may help to resolve issues of aggregat-
ing individual-level social capital to the team or firm
(Sorenson and Rogan 2014). To date, it is largely unclear
how network theorists should attribute the social bene-
fits of individuals’ connections to an organization. Pro-
vided that fostering and maintaining connections is the
province of individual founders and not firms them-
selves, our analysis of structural roles within teams
helps to resolve this issue by revealing how individual-
level variation in social resources might render to the
firm or group.
Second, heeding calls for greater consideration of the
multiple levels critical to a new entrepreneurial suc-
cess, such as individual-level abilities of the founders,
firm resources, and environmental conditions, our con-
struct allows us to account for both individual and firm
factors (Thornton 1999). Our findings suggest our find-
ings suggest that a more fine-grained understanding
of firm foundings should consider how the diversity of
structural roles facilitates entrepreneurship, above and
beyond the established influence of ascriptive diver-
sity (e.g., ethnicity) or past experiences (Beckman et al.
2007, Reagans et al. 2004, Ruef et al. 2003). In other
words, network measures of structural diversity repre-
sent a distinct approach for understanding the social
roles of founders in their teams.
Although we focus on the structural role comple-
mentarity of brokerage for capital mobilization, our
construct may have applications to a broader array
of settings, such as teams within organizations (Oh
et al. 2004, Reagans et al. 2004). While we illustrate
the value of brokerage role diversity to the success
of new ventures, the mechanisms of acquiring pri-
vate, diverse, and trustworthy information and mobi-
lizing resources can be equally significant in organiza-
tional settings (Hansen 1999); brokerage role diversity
may refine our understanding of the performance of
project teams within organizations. In addition, bro-
kerage role diversity is one important instantiation
of structural role complementarity, but other variants
also exist. For instance, a potential variant of struc-
tural role complementarity would include other forms
of centrality measures, such as degree centrality. We
included variation in degree centrality in our analysis
not because we anticipated strong effects for this vari-
able on basic capital but rather to test that our results
were not simply driven by variation in the number of
connections among members. However, in some set-
tings variation in degree centrality may be a critical
role distinction for team members. Future research on
structural roles might also consider the complementar-
ity of core versus periphery positions of team members
within a firm, drawing on March’s (1991) conceptual-
ization of organizational learning. In this case, team
members with core positions would be well suited to
exploit the firm’s stock of internal knowledge while
those with peripheral positions would have a greater
opportunity to explore beyond the firm’s boundaries
and potentially increase the firm’s stock of knowl-
edge. The potentially vast typology of structural role
complementarities and consequences for performance
highlight the need for an improved understanding
of the intersection of team dynamics and their social
The authors thank Linda Argote, Christine Beckman, Patrick
Doreian, Emily Erickson, Mark Fichman, David Krack-
hardt, Christina Gathmann, Giacomo Negro, Woody Powell,
department editor Olav Sorenson, an anonymous associate
editor, and the anonymous reviewers. For helpful comments
and critiques, the authors are also particularly grateful to
seminar participants at NETSCI: Economics in Networks,
SUNBELT, EGOS, Junior Organizational Theory Conference,
Annual Meeting of the American Sociological Association,
Carnegie Mellon University, UC Berkeley, Stanford Univer-
sity, University of Maryland, ESSEC, Lund University, and
University of Zurich. Sangyoon Shin contributed important
research assistance. For all remaining errors, the authors
alone are responsible.
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
16 Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s)
1Note that we do not mean to imply only two discrete roles of brokers
and nonbrokers, but instead roles along a continuum that ranges
between maximum cohesion and maximum brokerage. Empirically,
our measure of role differences captures even gradual variation in
brokerage roles among team members.
2Our data sources do not indicate how long such partnership ties
lasted, and ties between entrepreneurs might have lasted for the
45 years in our observation period. Nonetheless, we also analyzed
network measures where ties decayed after a certain period. For a
meaningful periodization, we rely on the average duration between
subsequent foundings by the same entrepreneurs. We decay ties
in the cofounding networks after 10 years, which corresponds to
about twice the average number of years that lapses between succes-
sive foundings (4.4 years, SD 6.4). The duration is therefore long
enough for prior ties to contribute to the portfolio of subsequent
founding partnerships. For example, if we consider a team of five
founding partners who established a company in 1890, then we take
into account all previous cofounding network ties that each of the
five partners maintained in the preceding 10 years (1880–1889). For
each founding year from 1869 through 1913, we coded correspond-
ing cofounding networks within these 10-year windows. We find
similar results to those presented here.
3Rarely did the firms’ charters list bonds; however, 2% of firms per-
mitted the issuance of bonds, which would be included in the basic
4The estimates from the corresponding GLM regression estimates
are provided in online supplementary materials Tables S.1 and S.2.
Ahuja G (2000) Collaboration networks, structural holes, and inno-
vation: A longitudinal study. Admin. Sci. Quart. 45(3):425–455.
Aldrich HE, Kim PH (2007) Small worlds, infinite possibilities?
How social networks affect entrepreneurial team formation and
search. Strategic Entrepreneurship J. 1(1–2):147–165.
Aldrich HE, Zimmer C (1986) Entrepreneurship through social net-
works. Sexton D, Smiler R, eds. The Art and Science of Entrepre-
neurship (Ballinger, New York), 3–23.
Ancona DG, Caldwell DF (1992) Bridging the boundary: External
activity and performance in organizational teams. Admin. Sci.
Quart. 37(4):634–665.
Angrist JD, Pischke JS (2008) Mostly Harmless Econometrics: An Em-
piricist’s Companion (Princeton University Press, Princeton, NJ).
Bascle G (2008) Controlling for endogeneity with instrumental vari-
ables in strategic management research. Strategic Organ. 6(3):
Beckman CM (2006) The influence of founding team company affili-
ations on firm behavior. Acad. Management J. 49(4):741–758.
Beckman CM, Burton MD (2008) Founding the future: Path depen-
dence in the evolution of top management teams from founding
to IPO. Organ. Sci. 19(1):3–24.
Beckman CM, Burton MD, O’Reilly C (2007) Early teams: The impact
of team demography on VC financing and going public. J. Bus.
Venturing 22(2):147–173.
Bernard H, Killworth P, Kronenfeld D, Sailer L (1984) The problem of
informant accuracy: The validity of retrospective data. Annual
Rev. Anthropol. 13(1):495–517.
Blau PM (1977) Inequality and Heterogeneity—A Primitive Theory of
Social Structure (Free Press, New York).
Browne W, Goldstein H, Rasbash J (2001) Multiple membership
multiple classification (MMMC) models. Statist. Modeling 1(2):
Browne WJ (2004) An illustration of the use of reparameterisation
methods for improving MCMC efficiency in crossed random
effect models. Multilevel Modeling Newsletter 16(1):13–25.
Bunderson S (2003) Recognizing and utilizing expertise in work
groups: A status characteristics perspective. Admin. Sci. Quart.
Burt R (1992) Structural Holes: The Social Structure of Competition (Har-
vard University Press, Cambridge, MA).
Burt R (2005) Brokerage and Closure: An Introduction to Social Capital
(Oxford University Press, New York).
Buskens V, van de Rijt A (2008) Dynamics of networks if everyone
strives for structural holes. Amer. J. Sociol. 114(2):371–407.
Carstensen FV (1983) Foreign participation in Russian economic life:
Notes on British enterprise, 1865–1914. Guroff G, Carstensen
FV, eds. Entrepreneurship in Imperial Russia and the Soviet Union
(Princeton University Press, Princeton, NJ), 140–158.
Coleman JS (1988) Social capital in the creation of human capital.
Amer. J. Sociol. 94:95–120.
Conley TG, Hansen CB, Rossi PE (2012) Plausibly exogenous. Rev.
Econom. Statist. 94(1):260–272.
Dahl MS, Sorenson O (2012) Home sweet home: Entrepreneurs’ loca-
tion choices and the performance of their ventures. Management
Sci. 58(6):1059–1071.
Eisenhardt KM, Schoonhoven CB (1990) Organizational growth:
Linking founding team, strategy, environment, and growth
among U.S. semiconductor ventures, 1978-1988. Admin. Sci.
Quart. 35(3):504–529.
Everett MG, Borgatti SP (1994) Regular equivalence: General theory.
J. Math. Sociol. 19(1):29–52.
Faust K (1988) Comparison of methods for positional analysis: Struc-
tural and general equivalences. Soc. Networks 10(4):313–341.
Ferriani S, Cattani G, Baden-Fuller C (2009) The relational anteced-
ents of project-entrepreneurship: Network centrality, team com-
position and project performance. Res. Policy 38(10):1545–1558.
Freeman J (1999) Venture capital as an economy of time. Leen-
ders RTAJ, Gabbay SM, eds. Corporate Social Capital and Liability
(Springer, Boston), 460–482.
Granovetter M (1985) Economic action and social structure: The
problem of embeddedness. Amer. J. Sociol. 91(3):481–510.
Greene W (2008) Econometric Analysis, 7th ed. (Prentice Hall, New
Gregory PR (1982) Russian National Income, 1885–1913 (Cambridge
University Press, Cambridge, UK).
Greif A (2006) Institutions and the Path to the Modern Economy:
Lessons from Medieval Trade (Cambridge University Press, Cam-
bridge, UK).
Hahn J, Hausman J (2005) Estimation with valid and invalid instru-
ments. Annales d’économie et de statistique 79/80:25–57.
Hansen MT (1999) The search-transfer problem: The role of weak ties
in sharing knowledge across organization subunits. Admin. Sci.
Quart. 44(1):82–111.
Hillmann H (2008) Localism and the limits of political broker-
age: Evidence from revolutionary Vermont. Amer. J. Sociol.
Hillmann H, Aven BL (2011) Fragmented networks and entre-
preneurship in late imperial Russia. Amer. J. Sociol. 117(2):
International Monetary Fund (2016) World economic outlook:
Too slow for too long. Last accessed November 15, 2017,
Khanna T, Rivkin J (2006) Interorganizational ties and business
group boundaries: Evidence from an emerging economy. Organ.
Sci. 17(3):333–352.
Kleibergen F, Paap R (2006) Generalized reduced rank tests using the
singular value decomposition. J. Econometrics 133(1):97–126.
Krackhardt D (1999) The ties that torture. Res. Sociol. Organ. 16(1):
Leckie G, Owen D (2013) Multiple membership multilevel models—
MLwiN practical. LEMMA VLE Module 13:1–48.
March JG (1991) Exploration and exploitation in organizational
learning. Organ. Sci. 2(1):71–87.
Marquis C, Raynard M (2015) Institutional strategies in emerging
markets. Acad. Management Ann. 9(1):291–335.
McKay JP (1970) Pioneers for Profit: Foreign Entrepreneurship and Rus-
sian Industrialization (University of Chicago Press, Chicago),
Moreira M (2003) A conditional likelihood ratio test for structural
models. Econometrica 71(4):1027–1048.
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
Aven and Hillmann: Structural Role Complementarity in Entrepreneurial Teams
Management Science, Articles in Advance, pp. 1–17, ©2017 The Author(s) 17
Nee V, Opper S (2012) Capitalism from Below—Markets and Institutional
Change in China (Harvard University Press, Cambridge, MA).
Oh H, Chung MH, Labianca G (2004) Group social capital and group
effectiveness: The role of informal socializing ties. Acad. Manage-
ment J. 47(6):860–875.
Owen TC (1991) The Corporation Under Russian Law, 1800–1917: A
Study in Tsarist Economic Policy (Cambridge University Press,
Cambridge, UK).
Owen TC (1992) RUSCORP: A Database of Corporations in the Russian
Empire, 1700–1914. Inter-University Consortium for Political and
Social Research, Ann Arbor, MI.
Owen TC (2005) Dilemmas of Russian Capitalism: Fedor Chizhov and
Corporate Enterprise in the Railroad Age (Harvard University Press,
Cambridge, MA).
Pfeffer J, Salancik GR (1978) The External Control of Organizations: A
Resource Dependence Perspective (Stanford University Press, Stan-
ford, CA).
Popielarz P, McPherson JM (1995) On the edge or in between: Niche
position, niche overlap, and the duration of voluntary associa-
tion memberships. Amer. J. Sociol. 101(3):698–720.
Portes A, Sensenbrenner J (1993) Embeddedness and immigration:
Notes on the social determinants of economic action. Amer. J.
Sociol. 98(6):1320–1350.
Rao H, Davis GF, Ward A (2000) Embeddedness, social identity and
mobility: Why firms leave the NASDAQ and join the New York
Stock Exchange. Admin. Sci. Quart. 45(2):268–292.
Reagans RE, Zuckerman EW, McEvily B (2004) How to make the
team: Social networks vs. demography as criteria for designing
effective teams. Admin. Sci. Quart. 49(1):101–133.
Ruef M (2010) The Entrepreneurial Group: Social Identities, Relations,
and Collective Action (Princeton University Press, Princeton, NJ).
Ruef M, Aldrich H, Carter N (2003) The structure of founding teams:
Homophily, strong ties, and isolation among US entrepreneurs.
Amer. Sociol. Rev. 68:195–222.
Sherif M, Harvey OJ, White BJ, Hood W, Sherif CW (1961) Intergroup
Conflict and Cooperation: The Robbers Cave Experiment (University
Book Exchange, Norman, OK).
Shipilov AV, Li SX (2008) Can you have your cake and eat it too?
Structural holes’ influence on status accumulation and market
performance in collaborative networks. Admin. Sci. Quart. 53(1):
Sørensen JB (2002) The use and misuse of the coefficient of varia-
tion in organizational demography research. Sociol. Methods Res.
Sorenson O, Rogan M (2014) (When) do organizations have social
capital? Annual Rev. Sociol. 40(1):261–280.
Sorenson O, Stuart TE (2001) Syndication networks and the spa-
tial distribution of venture capital investments. Amer. J. Sociol.
Stock J, Yogo M (2005) Testing for weak instruments in linear IV
regression. Andrews DWK, Stock JH, eds. Identification and
Inference for Econometric Models: Essays in Honor of Thomas
Rothenberg (Cambridge University Press, Cambridge, UK),
Stuart TE, Sorenson O (2003) Liquidity events and the geographic
distribution of entrepreneurial activity. Admin. Sci. Quart. 48(2):
Stuart TE, Sorenson O (2005) Social networks and entrepreneurship.
Alvarez SA, Agarwal R, Sorenson O, eds. Handbook of Entrepre-
neurship Research (Springer, New York), 233–252.
Stuart TE, Hoang H, Hybels R (1999) Interorganizational endorse-
ments and the performance of entrepreneurial ventures. Admin.
Sci. Quart. 44(2):315–349.
Thornton PH (1999) The sociology of entrepreneurship. Annual Rev.
Sociol. 25(1):19–46.
Uzzi B (1997) Social structure and competition in interfirm networks:
The paradox of embeddedness. Admin. Sci. Quart. 42(1):35–67.
Uzzi B, Gillespie JJ (2002) Knowledge spillover in corporate financ-
ing networks: Embeddedness and the firm’s debt performance.
Strategic Management J. 23(7):595–618.
Uzzi B, Spiro J (2005) Collaboration and creativity: The small world
problem. Amer. J. Sociol. 111(2):447–504.
Vedres B, Stark D (2010) Structural folds: Generative disruption in
overlapping groups. Amer. J. Sociol. 115(4):1150–1190.
White HC (2001) Markets from Networks: Socioeconomic Models of Pro-
duction (Princeton University Press, Princeton, NJ).
White HC, Boorman S, Breiger RL (1976) Social structure from mul-
tiple networks. I. Blockmodels of roles and positions. Amer. J.
Sociol. 81(4):730–780.
Winship C, Mandel M (1983) Roles and positions: A critique and
extension of the blockmodeling approach. Sociol. Methodol. 14:
Xiao Z, Tsui AS (2007) When brokers may not work: The cultural
contingency of social capital in Chinese high-tech firms. Admin.
Sci. Quart. 52(1):1–31.
Yogo M (2004) Estimating the elasticity of intertemporal substitu-
tion when instruments are weak. Rev. Econom. Statist. 86(3):
Downloaded from by [] on 27 December 2017, at 10:35 . For personal use only, all rights reserved.
... The fifth factor is team composition understood as the extent to which a team is heterogeneous regarding skills and knowledge. Working in a team composed of individuals with heterogeneous and complementary skills allows everyone to apply different structures and mental models that produce multifaceted dialog (Aven & Hillmann, 2017;Robbins, 2001) and hence to more effectively find creative solutions for problem-solving. ...
The question of “success” in collaborative creativity, from groups and teams to communities, is of enormous importance today. This is not only because we live in increasingly interdependent worlds, but also because the kinds of challenges we are confronted with, as individuals and societies, are becoming more and more complex. This chapter adopts a sociocultural approach and uses it to redefine the notions of “success,” “creativity,” and “teamwork.” We do this by outlining a sociocultural framework for collaborative creativity grounded in a vocabulary of difference, positions, perspectives, and dialog. Within this approach, creative success in collaboration cannot be judged exclusively based on its outcomes, but needs to take into account process and context. We illustrate this general principle with empirical research from two different settings: a field research on multidisciplinary groups working with innovation in a corporate environment, and another on coaches’ work with developing collaborative creativity in team sports. By focusing on process and context rather than outcomes and isolated individuals, this chapter will expand the notion of “creative success” to include the emergent and developmental quality of relations and interactions established within collaborative settings.
... Another key causal relationship explicating the probability of a particular tie forming is tie transitivity, or the formation of ties on the basis of common acquaintances or partners (Aven & Hillmann, 2017;Hite, 2005;Kirsch et al., 2009;Shane & Cable, 2002;Sorenson & Stuart, 2001;Zhang, Gupta, & Hallen, 2017). As established in the broader literature on inter-organizational networks, tie transitivity may be driven by indirect ties acting as effective information pipes, psychological biases around social proof and in-group preferences, and common partners creating "network closure" that encourages trustworthy behavior. ...
... In those cases, we replace individual hunter skill in the production equation with the weighted average of the skill of the group members. The statistical model follows the principles of a multiple membership model (29,30). When hunters are observed in different combinations of groups, it is possible to distinguish differences in skill between them. ...
Full-text available
Human adaptation depends on the integration of slow life history, complex production skills, and extensive sociality. Refining and testing models of the evolution of human life history and cultural learning benefit from increasingly accurate measurement of knowledge, skills, and rates of production with age. We pursue this goal by inferring hunters’ increases and declines of skill from approximately 23,000 hunting records generated by more than 1800 individuals at 40 locations. The data reveal an average age of peak productivity between 30 and 35 years of age, although high skill is maintained throughout much of adulthood. In addition, there is substantial variation both among individuals and sites. Within study sites, variation among individuals depends more on heterogeneity in rates of decline than in rates of increase. This analysis sharpens questions about the coevolution of human life history and cultural adaptation.
... This is presumed to be of particular value for knowledge creation and innovation (Berliant and Fujita, 2008;Niebuhr, 2010). More generally, ethnic diversity may contribute to the complementarity of roles within a team, which is important for access to information and, ultimately, firm performance (Aven and Hillmann, 2018). Migrants' mobility increases the spatial, organizational and social proximity among qualified human capital and facilitates face-to-face interaction, augmenting the chances that an unprecedented exchange of knowledge takes place. ...
Full-text available
Several studies have sought evidence as to whether ethnically diverse teams promote a diversity in knowledge and perspectives which is beneficial for innovation. In multicultural societies, however, there are multiple opportunities for exchange between people from different ethnic backgrounds, and the extent to which such encounters actually imply cognitive diversity seems debatable. We propose to regard as diverse those combinations of ethnic backgrounds that are relatively unlikely to occur under a hypothesis of random allocation to firms, based on the distribution of nationalities at NUTS3 level. We label this measure ``unusualness'' and apply it to the study of innovation in newly founded firms in Germany. Our results reveal that unusualness has a robust positive association with the probability of a start-up introducing an innovation within the first two years of business, while diversity as measured by a standard Blau index is insignificant. The results are robust to a large set of robustness checks. We interpret these results as an indication that not all combinations of national origins matter for innovation, but only those that are associated with differences in cognitive approaches and knowledge.
Full-text available
What determines performance among small businesses with five employees or less in Mexico? Based on a conceptual framework already used in Argentina and on previous research, a sample of 174 Mexican entrepreneurs from two different states (Jalisco and Nuevo León) was used to test a set of nine hypotheses. The dependent performance variables tested were an objective one, sales, and a subjective one, the personal assessment of performance (or success) of entrepreneurs. The independent variables considered included personal, sociological, and organizational characteristics. Results were obtained from two linear regression models on the two dependent variables. In terms of personal characteristics, variables that were positively related to sales included three Human Capital components (Education level, Business experience, and Weekly hours worked), having been pushed into self-employment by economic necessity, and belonging to the male gender. Regarding organizational variables, entrepreneurs with higher sales had obtained bank loans and had purchased their business (by opposition to starting it from scratch) and had economic necessity (extrinsic) reasons to be in business. Respondents who worked long hours and had obtained government support were more likely to be more satisfied of their own performance than others.
Research on entrepreneurial team diversity (ETD) has reached a critical juncture as inconsistent findings hinder further theory development. The main reasons for these inconsistencies are one-dimensional theoretical perspectives that make it difficult to comprehensively theorize the benefits and barriers of ETD from a cross-disciplinary perspective. To address this shortcoming, we systematically identify existing literature and classify 44 studies into an ‘inputs-mediators-outcomes’ (IMO) framework. We find that the field is considerably fragmented, especially with regard to disciplinary perspectives, study contexts, diversity dimensions, and outcome variables. To invigorate and advance the research stream, we highlight unresolved issues and knowledge gaps in the current literature and propose a multi-disciplinary research agenda presented in the form of an IMO framework. We conclude that, while existing research on ETD offers a solid foundation, it is far from having reached its full potential.
Interpersonal networks can be conceptualized not only as actual social structures surrounding individuals but also as cognitive social structures stemming from individuals’ perceptions of those relationships. Yet most research on social networks adopts either a structural or a perceptual perspective. In this article, I blend these two traditions to examine how actual and perceptual brokerage jointly determine innovation performance. I hypothesize that while actual brokerage benefits individuals by exposing them to nonredundant information, socially perceived brokerage—being perceived to bridge groups regardless of one’s actual network configuration—may trigger skepticism of brokers’ motives that could hinder their ability to innovate. Thus I argue that others’ perceptions of a focal actor’s brokerage opportunities constitute a critical contingency underlying network advantage. Using a multimethod approach, including a field study in a global consulting firm and a preregistered experiment, I find that individuals spanning structural holes achieve higher innovation performance when their colleagues perceive them to have closed rather than open networks, and that trust is the underlying mechanism driving this effect. Integrating insights from cognitive social structures into structural holes theory, this study illustrates the importance of considering both structural and perceptual mechanisms in modeling how individuals reap the benefits of brokerage.
We survey 398 Chinese entrepreneurs within 118 entrepreneurial teams to investigate external social capital's impact on: the decision to start a venture and the mediating role that risk perception plays in that decision. Our results confirm that external social capital has a significant and positive effect on the decision to start a venture, while risk perception partly mediates this relationship. Furthermore, we find that the three dimensions of external social capital drive the decision to start a venture differently. The most important domains of external social capital are commercial social capital, followed by technical social capital and institutional social capital.
Full-text available
Network studies in the entrepreneurship domain suffer from an incomplete theorization of how the content of social capital relates to network relationships and structures in which entrepreneurs are embedded or embed themselves. This study presents a systematic review of the various ways in which the interaction between content (e.g. cognition and resources) and social structure has been studied within entrepreneurship. Based on this review, we develop a more integrative account of the underlying action mechanisms that link the content and structure of social capital. These mechanisms cut across different research traditions and align areas of entrepreneurship research. In this way, we contribute an integrative review of prior work and a formative set of directions for further theorizing and research on social capital, networks and entrepreneurship.
Creativity increasingly becomes the determining factor of start-up success. Opportunities to found new ventures are based on the uniqueness of underlying business models/ideas and on the speed to transform these unique business models/ideas into business. Doing so requires to be capable of managing people to create and transform new ideas into the business, also defined as being creative. Since most start-ups foremost in the tech sector are based on teams the management of factors propelling creativity is vital. In this chapter, we discuss five drivers of creativity that can be managerially addressed in start-up teams: social networks, leadership, network structures, communication media, managing conflict, and creative culture. Based on this discussion, we draw conclusions and formulate recommendations on how creativity can be fostered in start-up teams.
This study analyses the legal framework imposed on corporations by the imperial Russian Government. It stresses the dual nature of the bureaucracy's policy toward modern capitalist enterprise: encouragement for the sake of economic development, and regimentation in the interest of maintaining autocratic control. By illuminating the political nature of the autocracy's economic agenda, Professor Owen seeks to explain why Russian corporate law became increasingly restrictive toward the end of the imperial period. Attention is also given to the practices of Russian capitalists, whose occasional abuses of corporate power justified restrictive laws in the eyes of officials. The emphasis of this study on the uneasy accommodation between tsarist autocracy and the modern corporation clarifies aspects of Russian political, economic, and cultural life that hindered the development of capitalism on the eastern periphery of Europe.
I find that a firm's innovation output increases with the number of collaborative linkages maintained by it, the number of structural holes it spans, and the number of partners of its partners. However, innovation is negatively related to the interaction between spanning many structural holes and having partners with many partners.