Content uploaded by Bob Walrave
Author content
All content in this area was uploaded by Bob Walrave on Mar 06, 2024
Content may be subject to copyright.
Technovation 131 (2024) 102954
Available online 26 January 2024
0166-4972/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
Knowledge diversity and technological innovation: The moderating role of
top management teams
Bob Walrave
a
,
*
, Nino van de Wal
b
,
1
, Victor Gilsing
c
a
Eindhoven University of Technology, School of Industrial Engineering & Innovation Sciences, the Netherlands
b
University of Antwerp, Department of Management, Faculty of Applied Economics, Belgium
c
Vrije Universiteit Amsterdam, Department of Management & Organization, School of Business and Economics, The Netherlands & University of Antwerp, Department of
Management, Faculty of Applied Economics, Belgium
ARTICLE INFO
Keywords:
Knowledge diversity
Top management
Innovation
Structural attributes
Structures
TMT
Incommensurability
Transformational leadership
ABSTRACT
Knowledge diversity between a rm’s groups of inventors enables recombinatory search for innovation. Yet, such
diversity remains rather useless unless it is actively exchanged among inventor groups. Inventor groups, how-
ever, tend to specialize by engaging in so-called perspective-making activities, that is, in intra-group knowledge
exchange and specialization. This makes them increasingly unable to communicate and understand other in-
ventor groups and creates a risk of incommensurability, which attenuates a rm’s effectiveness in its recombi-
nation for innovation. Here, we draw on transformational leadership theory to understand how TMTs are enabled
to motivate and inspire their inventor groups to share information and knowledge, to mitigate incommensura-
bility risks. For a TMT to act as an effective transformational leader, information is key, and their ability to send,
receive, and process information is shaped, following classic organization theory, by their structural attributes.
Hence, we study three key TMT structural attributes that underlie its information-processing capacity: Hierar-
chical structure, functional structure, and administrative intensity. Based on a longitudinal dataset that includes
124 pharmaceutical rms, 2815 top managers, and 34,203 inventors, we show that the positive relation between
inventor group knowledge diversity and innovation performance strengthens with a functional structure yet
weakens with administrative intensity. We contribute to the literature with its emphasis on how TMT compo-
sitional characteristics inuence its cognitive processes and decision-making on innovation, by studying how
TMT structural characteristics shape its information-processing capacity to be effective as transformational
leaders in motivating and inspiring inventor groups to engage in perspective-taking and overcome
incommensurability.
1. Introduction
It is well known that knowledge diversity, which exists between
groups of inventors with different expertise, enables recombinatory
search processes for innovation (Caner et al., 2017; Carnabuci and
Operti, 2013; Fleming, 2001; Gittelman and Kogut, 2003; Maggitti et al.,
2013). Yet, an important and overlooked implication is that knowledge
diversity is only useful for recombinatory purposes if these various
pieces of knowledge are actively exchanged among groups of inventors
(Boland and Tenkasi, 1995; Carnabuci and Operti, 2013). In other
words, whereas knowledge diversity is a necessary condition for
recombinatory search, it is not sufcient (Kogut and Zander, 1992;
Polanyi, 1966; Savino et al., 2017).
Diverse groups of inventors are not necessarily motivated to engage
in knowledge exchange. As groups of inventors specialize, they tend to
develop expert vocabularies, references, and information connections,
which are difcult for others to recognize and understand (Anderson
and Lewis, 2014; Fraidin, 2004). That is, groups of inventors—as com-
munities of knowing—naturally engage in so-called perspective-making,
that is, in in-group knowledge development and specialization. Yet, such
specialization inherently limits their aptitude to engage in perspective--
taking, that is, understanding and incorporating the perspectives of other
groups (Boland and Tenkasi, 1995; Fraidin, 2004). Specically, while
perspective-making activities enable specialization within a group of
* Corresponding author.
E-mail address: b.walrave@tue.nl (B. Walrave).
1
This journal publication is based on the dissertation titled ’Leading Innovation: How Top Managers Inuence Technological Search and Innovation Performance
of Firms’ submitted by Nino van de Wal for the completion of the degree of Doctor of Philosophy at Antwerp University.
Contents lists available at ScienceDirect
Technovation
journal homepage: www.elsevier.com/locate/technovation
https://doi.org/10.1016/j.technovation.2024.102954
Received 15 July 2022; Received in revised form 5 December 2023; Accepted 7 January 2024
Technovation 131 (2024) 102954
2
inventors, and happens rather naturally, it also instigates, and increas-
ingly so, between-group misunderstandings, biases, and conict (Car-
lile, 2004; Milliken and Martins, 1996; Williams and O’Reilly, 1998).
This leads to incommensurability across inventor groups, which further
drives out perspective-taking (Boland and Tenkasi, 1995; Hoever et al.,
2012), thereby limiting the potential for recombinatory search.
Incommensurability among inventor groups, as such, jeopardizes a
rm’s innovation performance and, as a consequence, its future
competitiveness and viability (Ahuja and Lampert, 2001; Carnabuci and
Operti, 2013; Granstrand et al., 1997; Moreira et al., 2018). Such risk
makes it a key issue for TMTs to consider and act upon, through moti-
vating and inspiring their inventor groups to overcome incommensu-
rability and to engage in perspective-taking. In this respect, TMTs can
address incommensurability among inventor groups by acting as trans-
formational leaders. Transformational leadership ‘transforms’ followers
to transcend their self-interest to identify needed change, creating a
vision to guide the change through inuence and inspiration, and
through executing the change in tandem with committed members of a
group (Aryee et al., 2012; Bass, 1999; Bryman, 2011; Gumusluoglu and
Ilsev, 2009; Odumeru and Ogbonna, 2013). Following transformational
leadership theory (Burns, 1978; Siangchokyoo et al., 2020), we argue
that TMTs may seek to mitigate the risk of incommensurability by
raising inventor groups’ awareness of such risk, encouraging collabo-
rations and relationships, preventing and resolving conicts, and
fostering a sense of shared purpose. This resonates strongly with Law-
rence and Lorsch’s (1967) seminal idea that a rm’s TMT is tasked with
achieving ‘unity of effort’ among their organization’s diverse parts.
For TMTs to act as effective transformational leaders and address the
risk of incommensurability, receiving and sending information as well as
its processing are of key importance. Here, we draw on classic organi-
zation theory, which details that structural attributes determine
information-processing capacity (Galbraith, 1973; Lawrence and Lorsch,
1967). Here, we focus on three key TMT structural attributes—covering
both the richness of vertical (i.e., hierarchical structure) and horizontal
(i.e., functional structure) information processes
2
as well as cognitive
capacity to process this information and act on it towards the organi-
zation (i.e., administrative intensity)—and discuss how these inuence a
TMT’s ability to act as transformational leader to address incommen-
surability among its inventor groups.
Based on this, we specify and test how these TMT structural attri-
butes moderate the relationship between inventor groups’ knowledge
diversity and innovation performance. To test our hypotheses, we rely
on a unique dataset spanning 2000–2014, which includes 124 phar-
maceutical rms, 34,203 inventors, and 2815 top managers. We nd
that, among others, TMT functional structure strengthens, while TMT
administrative intensity attenuates, the positive relationship between
inventor groups’ knowledge diversity and innovation performance.
We contribute to the literature on knowledge diversity and recom-
bination. This body of literature discusses the merits of knowledge di-
versity as an enabler for recombinatory search for innovation
(Carnabuci and Operti, 2013; Moreira et al., 2018), and emphasizes the
value of different organizational measures to connect and create
linkages among (groups of) inventors, such as information support tools
and objects (Acharya et al., 2022; Boland and Tenkasi, 1995; Nicolini,
2011), informal exchanges (e.g., Hargadon, 2002; Chou et al., 2011;
Garud et al., 2011), formal team structures (e.g., Singh and Fleming,
2010; Toh and Polidoro, 2013), effective team leadership (e.g., Currie
and White, 2012) and organizational knowledge networks (Carnabucci
and Di´
oszegi, 2015; Moreira et al., 2018; Paruchuri and Awate, 2017).
Yet, this literature with its emphasis on organizational measures to
connect and create linkages implicitly assumes that diverse groups of
inventors are naturally inclined and motivated to engage in cross-group
collaboration and knowledge exchange. Whereas diverse inventor
groups naturally engage in perspective-making, they do not in
perspective-taking (Boland and Tenkasi, 1995; Hoever et al., 2012). This
implies that a TMT needs to ensure it not only provides these informa-
tion channels and structures for communication and knowledge ex-
change, but also inspires and motivates groups of inventors to overcome
incommensurability and engage in perspective-taking, through trans-
formational leadership. This motivational dimension to knowledge ex-
change across diverse inventor groups has remained relatively
unaddressed in the literature on recombination. Whereas there is
growing attention in the literature for the role of leadership in
addressing individuals’ motivation for creativity and knowledge ex-
change with others (Gardner et al., 2020; Hughes et al., 2018; Crosby
and Bryson, 2010), an understanding of motivation for knowledge ex-
change and perspective taking by diverse groups of inventors, across
disciplinary boundaries, is still missing.
We also contribute to the growing literature on how TMTs inuence
their rm’s innovation activities and outcomes. In most of this literature,
there has been a strong emphasis on how top managers’ backgrounds
and TMTs’ compositional characteristics shape their cognitions and
values, and how these inuence their processing of external information
and a rm’s strategic decision-making process on innovation (e.g.,
Kashmiri & Mahajan, 2017; Kiss et al., 2018; Kiss et al., 2020; Ruiz--
Jim´
enez et al., 2016; Talke et al., 2010; Zhang et al., 2021). This
dominant focus on information-processing and decision-making carries
a strong focus on the cognitive processes within a TMT, yet at the
expense of a predominantly motivational process that occurs largely
between a TMT and a rm’s inventor groups. To address this, we need to
look beyond cognitive processes within a TMT such as its processing of
information and decision-making on innovation strategy and focus
instead on its information-processing capacity to inuence the organi-
zation’s different inventor groups through transformational leadership,
to execute on this strategy. Here, our paper contributes by showing how
a TMT’s structural attributes shape its capacity for receiving, processing,
and sending information to be effective as transformational leaders in
motivating and inspiring inventor groups to engage in
perspective-taking, and overcome incommensurability, to enable
innovation.
2. Theoretical background and hypotheses
2.1. Knowledge diversity, recombinatory search, and innovation
As people are cognitively bounded, inventors can specialize only in a
limited number of technological domains (Gruber et al., 2013; Maggitti
et al., 2013). The implication is that to solve today’s complex problems,
organizations need to maintain so-called communities of knowing
(Boland and Tenkasi, 1995)—in this paper considered to be groups of
inventors specializing in various technical knowledge domains. Espe-
cially knowledge-intensive rms are likely to consist of different com-
munities, each holding highly specialized knowledge and associated
technology (Boland and Tenkasi, 1995; Purser et al., 1992). Such in-
ventor groups’ knowledge diversity, if shared, creates the potential for
innovation through recombinatory search processes (Acharya et al.,
2022; Boland and Tenkasi, 1995; Nicolini, 2011; Fleming, 2001; Git-
telman and Kogut, 2003)—through various mechanisms: First,
2
We refer to vertical information processes as the ow of information within
a hierarchical structure or along the chain of command in an organization. In a
vertically structured organization, information typically moves up and down
through different levels of management, from top to bottom or vice versa.
Horizontal information processes involve the exchange of information between
individuals or departments at the same hierarchical level within an organiza-
tion. This type of communication is often associated with collaboration and
coordination among peers rather than through a formal chain of command.
Information richness refers to the depth and complexity of information that can
be conveyed through a communication channel. Rich communication channels
allow for the transmission of a variety of cues, such as verbal and nonverbal
signals, and can convey a high level of detail and context.
B. Walrave et al.
Technovation 131 (2024) 102954
3
knowledge diversity and exchange bring about the possibility to
discover and execute new opportunities for technological innovation
(Brennecke and Rank, 2017; Taylor and Greve, 2006). That is, the
sharing of specialized, diverse knowledge exposes groups of inventors to
different areas of expertise, approaches, and viewpoints that, in turn,
result in novel knowledge associations and linkages (Brennecke and
Rank, 2017; Taylor and Greve, 2006). Second, knowledge diversity and
exchange, across groups of specialized inventors, also leads to novel
interpretations of existing knowledge, which helps discover hitherto
undiscovered opportunities or identify new ways to understand and
solve existing problems (Carnabuci and Operti, 2013). Third, inventor
groups’ knowledge diversity and exchange also aid in interpreting and
learning from the often-unexpected outcomes of experiments. That is,
diverse groups of specialists are simply better equipped to make sense of
such feedback (Henderson and Cockburn, 1994; Thomke et al., 1998).
As an example, in the pharmaceutical industry, the development of
new products demands the integration of disciplinary knowledge from a
wide array of different disciplines. Think about molecular biology,
physiology, synthetic chemistry, and pharmacology (Henderson, 1994).
Mark Esser, VP at AstraZeneca, once explained that their scientists
explored three potential sources for antibodies against COVID-19 in the
process of identifying a lead candidate, and that such exploration
induced the need for groups of specialists from different disciplines to
exchange their knowledge to develop a more profound understanding of
new disease domains and what could form effective new molecules to
combat the underlying causes of a new ‘target’ (Astrazeneca, 2021).
Summarizing, knowledge exchange among specialized groups of
inventors is key to enable recombinatory search processes that, in turn,
lead to innovation (Ahuja and Lampert, 2001; Boland and Tenkasi,
1995; Carnabuci and Operti, 2013; Fleming, 2001; Huang, 2009; Huang
and Chen, 2010). Hence, and in-line with standing literature (e.g.,
Brennecke and Rank, 2017; Taylor and Greve, 2006), we expect that.
Hypothesis 1. Knowledge diversity among a rm’s groups of in-
ventors is positively related to rm innovation performance.
2.2. Incommensurability among inventor groups: the need for information
exchange and the role of a TMT
As explained, the diverse specialized knowledge, skills, and expertise
are held by a rm’s various groups of inventors—and such knowledge
only becomes helpful for recombinatory search when these groups, or
communities of knowing, are motivated to actively exchange it (Bren-
necke and Rank, 2017; Kogut and Zander, 1992; Savino et al., 2017). In
this respect, Henderson and Cockburn (1994) found that maintaining an
extensive ow of information across different scientic disciplines and
technological domains benets innovation; but this does not mean that
between-group knowledge exchange comes about by itself.
The similarity between people operating in a particular community
of knowing yields cooperation, trust, and social cohesiveness (e.g.,
Harrison et al., 2002; Locke and Horowitz, 1990), enhancing an in-group
orientation. Such an in-group orientation also means a stronger focus on
the group’s unique knowledge and supports perspective-making, implying
the development of a common, yet specialized, language and shared
cognition that supports information exchange within a community of
knowing (Boland and Tenkasi, 1995; Lawrence and Lorsch, 1967).
3
However, the risk of an in-group orientation is that inventor groups
increasingly start to perceive other groups as less relevant, trustworthy,
or cooperative (Stephan and Stephan, 1985; Tsui et al., 1995), feeding
the development of stereotypes and prejudices (Boland and Tenkasi,
1995; Milliken and Martins, 1996). As Boland and Tenkasi (1995: p.
351) describe: “Thought worlds with different funds of knowledge and
systems of meaning cannot easily share ideas, and may view one an-
other’s central issues as esoteric, if not meaningless.”
Perspective-making therefore brings about the risk of incommensu-
rability, stiing communication and knowledge exchange across groups
of inventors, and increasingly inhibiting perspective-taking—which refers
to information exchange across communities of knowing, aimed at
bringing about an understanding and incorporating the perspectives of
other groups as part of a community’s way of knowing (Boland and
Tenkasi, 1995; Carnabuci and Operti, 2013; McGrath and Gruenfeld,
1993; Milliken and Martins, 1996). Meanwhile, there is a growing
literature detailing how inventors can access and connect to other
(groups of) inventors by emphasizing the use of information support
tools (e.g., Acharya et al., 2022; Boland and Tenkasi, 1995; Nicolini,
2011), informal exchanges (e.g., Hargadon, 2002; Chou et al., 2011;
Garud et al., 2011), formal team structures (e.g., Singh and Fleming,
2010; Toh and Polidoro, 2013), effective team leadership (e.g., Currie
and White, 2012), and organizational knowledge networks (e.g., Car-
nabucci and Di´
oszegi, 2015; Moreira et al., 2018; Paruchuri and Awate,
2017). This strong focus on different information channels and organi-
zation structures to connect and create linkages places an emphasis on
the ability for inventors to reach out to other (groups of) inventors,
under the tacit assumption that these groups are also motivated to do so.
Yet, groups of inventors are naturally inclined to engage in
perspective-making, but much less to also engage in perspective-taking
(Boland and Tenkasi, 1995; Hoever et al., 2012; Mathieu et al., 2017).
This brings about a lack of effective information exchange between in-
ventor groups, which deteriorates the potential of knowledge recombi-
nation for innovation—ultimately jeopardizing a rm’s future
competitiveness and viability (Ahuja and Lampert, 2001; Granstrand,
1998; Grigoriou and Rothaermel, 2017).
Such a high risk makes the management of incommensurability a key
concern for TMTs. At the same time, however, it has been well estab-
lished that a rich, decentral, and horizontal information exchange is
needed to overcome different frames of reference across groups (Daft
and Lengel, 1984; Maitlis and Christianson, 2014; Narayanan et al.,
2011). For a TMT, the implication is that it needs to enable and foster-
—but not necessary directly interfere in—such rich form of horizontal
communication and interaction among its inventor groups (Granstrand,
1998). This demands for a TMT to act as transformational leaders.
Transformational leadership theory postulates that leaders can inspire
and motivate followers to achieve their full potential and surpass their
own self-interests (and expectations) to the benet of the organization
(Aryee et al., 2012; Bass, 1999; Bryman, 2011; Gumusluoglu and Ilsev,
3
Perspective taking is a cognitive process in which inventors take stock of
and learn other inventors’ expertise and viewpoints to understand their pref-
erences, values, and needs (Grant and Berry, 2011). Yet, in organizations, the
motivation for sharing and jointly creating knowledge tends to be limited as
inventors, and people in general, often try to protect what they know and are
generally not motivated to engage in perspective-taking (Hoever et al., 2012),
which accentuates the need for leadership (Von Krogh et al., 2012). Whereas
there is growing attention in the literature for the role of leadership in
addressing individuals’ motivation for creativity and innovation (Gardner et al.,
2020; Hughes et al., 2018; Crosby and Bryson, 2010), how it affects the moti-
vation for knowledge exchange and perspective-taking across disciplinary
boundaries, by diverse groups of inventors, is still missing.
B. Walrave et al.
Technovation 131 (2024) 102954
4
2009; Odumeru and Ogbonna, 2013). We argue that transformational
leaders can play a crucial role in preventing incommensurability among
inventor groups by: First, clear and open communication on strategy,
vision, and goals, ensuring everyone understands the direction and
purpose of their work, to promote alignment. Second, by actively
encouraging collaboration and teamwork, between groups, by fostering
and stimulating a collaborative work environment. Third, by conict
resolution, addressing conicting viewpoints, incentives, and in-
terpretations by different inventor groups, to prevent incommensura-
bility from happening or escalating. And fourth, by building trust among
the various groups, to ensure inventors feel safe to express their
(potentially conicting) viewpoints, which also enhances horizontal
information exchange among inventor groups, further decreasing the
risk of incommensurability.
For TMTs to be effective as transformational leaders in addressing
the risk of incommensurability, its ability to obtain information from the
organization, to interpret this, and send information to the organization,
is of key importance (Daft and Weick, 1984; Maitlis and Christianson,
2014; Narayanan et al., 2011). Here, we draw on classic organization
theory and consider structure to determine information-processing ca-
pacity (Galbraith, 1973; Lawrence and Lorsch, 1967). Structures, in
general, serve to establish coherent connections between the various
agents and functions that make up teams, groups, departments, and,
ultimately, the entire organization. In this respect, Hambrick et al.
(2015) describe that a TMT’s structural attributes signicantly inuence
the extent to which units or individuals affect each other by setting the
basic contours of the team. In other words, how agents interact to
obtain, interpret, and share information is according to classic organi-
zation theory greatly determined by structures.
Here, we focus on three key TMT structural attributes that, we argue,
inuence TMTs ability to act as transformational leader to address
incommensurability among inventor groups. This idea is in line with
research which details the link between structures—such as organiza-
tional hierarchy and centralization—and transformational leadership
behavior and effectiveness (e.g., Walter and Bruch, 2010; Wright and
Pandey, 2010). We concentrate on those attributes that fully emerge
from structure, staying true to classical organization theory, namely:
Hierarchical structure, which inuences the richness of vertical infor-
mation processes and inuences, for instance, critical thinking and
constructive disagreement among TMT members; functional structure,
which captures diversity in functional expertise, inuencing the richness
of horizontal information processes, notably by bringing together a wider
range of ideas and experiences; and administrative intensity, a measure
of relative size and administrative demands, which affects the cognitive
capacity of a TMT to act as transformational leader by, for instance,
having the time to share compelling visions and goals, to foster a sense of
shared purpose and enable collaboration.
2.2.1. The moderating role of TMT hierarchical structure
A stronger hierarchy within a TMT brings about the opportunity for
TMTs, and especially the CEO, to inuence and inspire others. In this
respect, higher ranked ofcers are placed in a position of great leverage,
able to (quickly) act and direct the implementation of transformational
initiatives to combat, among others, incommensurability among in-
ventor groups. Moreover, a high hierarchical position also provides
greater visibility and exposure, allowing these top managers to serve as
role models, potentially allowing them to spur perspective-taking
among inventor groups.
At the same time, however, a strong hierarchy implies that rank
designations among the TMT are more distinct and that a sort of pecking
order may emerge. Consequently, TMT members will hold less salience
for each other, as they will view each other less as part of the same team
(Hambrick et al., 2015). In this respect, a strong hierarchy can limit a
TMT’s vertical information-processing capacity, thereby limiting their
ability to engage in transformational leadership in several ways.
First, a status quo bias might arise, where leaders in higher
hierarchical positions are more focused on preserving their authority
and the stability of the organization, rather than fostering an environ-
ment conducive of collaboration and innovation (Pizzolitto et al., 2023).
As a consequence, the concentration of power could discourage the
participation and involvement of other, lower-ranked TMT members
(Clark, 2022). That is, the top leaders (e.g., CEO and/or COO) may
dominate the decision-making processes, leaving little room for input
and inuence from other team members, or lower-level team members
may simply feel hesitant to voice their ideas or concerns to
higher-ranking executives due to power differentials and fear of re-
percussions (Mihalache et al., 2014). This can lower a TMT’s ability to
signal alignment and a sense of shared purpose to the organization to-
wards the organization, and to credibly encourage collaboration and
teamwork across different groups in the organization, including inven-
tor groups. This will limit the potential for transformational leadership
by a TMT to stimulate inventor groups to overcome incommensurability
and to engage in perspective-taking.
Second, TMT hierarchy can contribute to a culture where feedback
ows primarily from top to bottom, rather than being encouraged in
reverse. Leaders in higher hierarchical positions may be less receptive to
feedback and suggestions from lower-level team members, also because
there are less informal social relations among TMT members (Mihalache
et al., 2014). This can suppress minority dissent (Nijstad et al., 2014)
that hinders the open exchange of ideas that can impede a TMT’s efforts
to support trust building among various inventor groups, and to ensure a
perception of safety by inventor groups to express different viewpoints.
In this respect, TMT hierarchy could severely limit the upward infor-
mation process of ‘issue selling’ by inventor groups on incommensura-
bility issues (Dutton et al., 2001; Narayanan et al., 2011), which reduces
the likelihood that (looming) conicts between inventor groups reach a
TMT and can be effectively resolved.
These negative consequences limit the richness of the vertical in-
formation ow, to and from a TMT, hindering its ability to gather
diverse perspectives and ideas. This will negatively affect its ability to
prevent and resolve conicts, encourage collaborations across inventor
groups, and foster a shared sense of purpose—all which reduces its
effectiveness as transformational leader to address perspective-taking
across inventor groups. We therefore expect a negative moderation ef-
fect of TMT hierarchy on the relationship between inventor groups’
knowledge diversity and innovation.
Hypothesis 2. TMT hierarchical structure negatively moderates the
relationship between knowledge diversity among a rm’s groups of in-
ventors and rm innovation performance.
2.2.2. The moderating role of TMT functional structure
In functionally structured TMTs, each executive is responsible for a
specic functional part of the rm’s value-creation process in a way that
depends on the behavior and effectiveness of all other TMT members
(Hambrick et al., 2015; Menz, 2012). A functionally structured TMT
consists of different executive functions, like marketing, sales, R&D,
operations, nance, and engineering. This means that a highly func-
tionally structured TMT needs to process cross-functional information
and expertise to make effective team decisions (Menz, 2012). While such
a functional structure provides valuable expertise, it may also cause
functional leaders to prioritize the interests and goals of their respective
functions, potentially hindering the TMT’s ability to take a holistic and
organization-wide approach to transformational leadership.
On the other hand, the mutual dependency could also make them
more aware of the presence and value of functional differences in the
organization, and of diversity in general (Richard et al., 2019). In this
respect, functional diversity can be advantageous for TMT’s ability to
process information horizontally, to enable transformational leadership,
in various manners. First, each TMT member brings a unique set of skills
and capabilities associated with their functional expertise. Drawing on
their varied perspectives, insights, and skills enables them to synthesize
B. Walrave et al.
Technovation 131 (2024) 102954
5
from these diverse perspectives (Wang et al., 2019), which contributes
to fostering a shared sense of purpose and to stimulating collaboration
across inventor groups. In this respect, its functional structure enables a
TMT to be effective as transformational leader as it supports the align-
ment of diverse viewpoints and skill sets and thereby the alignment of
the organization at large, which enhances perspective-taking across in-
ventor groups for recombination and innovation.
Second, TMT functional diversity can be effective in addressing
conicting viewpoints by offering different interpretations and in this
way reduce risks of conicts, or resolve them, between different in-
ventor groups (Cao et al., 2010), which then also contributes to the
build-up of trust among these groups. Trust between inventor groups
breaks down silos, reduces risks of conicts, and facilitates
cross-functional collaboration, which enables them to engage in
perspective-taking and challenge existing assumptions, identify new
opportunities and develop innovative solutions.
In sum, we expect that a functionally structured TMT will provide it
with a rich horizontal information ow, equipping them to act as
effective transformation leaders to also address incommensurability
across inventor groups. Therefore, we expect a positive moderation ef-
fect of TMT functional structure on the relationship between inventor
groups’ knowledge diversity and a rm’s innovation performance.
Hypothesis 3. TMT functional structure positively moderates the
relationship between knowledge diversity among a rm’s groups of in-
ventors and rm innovation performance.
2.2.3. The moderating role of TMT administrative intensity
Administrative intensity is reected by the number of TMT members
in relation to the number of inventors—and is indicative of the required
level of TMT administrative tasks and oversight (see Sine et al., 2006). A
high administrative intensity is sometimes found to be benecial,
notably in the context of startups, by top management extensively
engaging in process and organization-building activities (Sine et al.,
2006).
However, a high administrative intensity can negatively inuence
TMTs ability to act as a transformational leader, in two ways. First, the
demands associated with high administrative intensity include tasks
such as budgeting, planning, and performance monitoring, which
consume a substantial portion of a TMT’s time and attention. As a
consequence, these administrative tasks may consume much of the
TMT’s cognitive resources, leaving little mental space and energy to be
effective as transformational leader to notice new opportunities for
innovation, and to address perspective-taking across inventor groups
(Shepherd et al., 2017). More specically, higher administrative in-
tensity reduces a TMT’s capacity for raising awareness among inventor
groups of the risk of incommensurability, for encouraging collaboration
among them, for preventing and resolving conicts, and for fostering a
sense of shared purpose across these groups.
Second, a high administrative intensity can nurture a culture where
alignment and a shared sense of purpose as well as fostering and stim-
ulating collaboration across inventor groups carries especially a focus on
day-to-day operational tasks, routine processes, and ensuring compli-
ance with established procedures (Cortes and Herrmann, 2021; Teece,
1999). Such focus prioritizes stability and risk mitigation, in favor of
maintaining stability and operational efciency, yet discourages
risk-taking, experimenting with new ideas, and knowledge exchange
across inventor groups, which augments the risk of incommensurability
across inventor groups.
In sum, a high administrative intensity increases the risk of incom-
mensurability across inventor groups while it also undermines a TMT’s
ability to address this risk by effective transformational leadership.
Therefore, we expect a negative moderation effect of TMT administra-
tive intensity on the relationship between inventor groups’ knowledge
diversity and a rm’s innovation performance.
Hypothesis 4. TMT administrative intensity negatively moderates the
relationship between knowledge diversity among a rm’s groups of in-
ventors and rm innovation performance.
3. Method
We drew our sample from the pharmaceutical industry for different
reasons. First, many pharmaceutical rms are relatively large and
known to be comprised of various communities of knowing—think of
domains such as molecular biology, physiology, biochemistry, synthetic
chemistry, and pharmacology (Henderson and Cockburn, 1994; Pisano,
2006), and the various specialisms that exist within these domains. This
ts well with our emphasis on the creation of innovations that come
from recombining unconnected elements of knowledge (e.g., Carnabuci
and Operti, 2013). Second, the interdisciplinary nature of drug discov-
ery makes the ability to exchange and combine specialized knowledge
within a rm key to innovation success and, ultimately, rm survival.
Third, practically speaking, pharma rms have a strong incentive to le
for patents, allowing us to study inventor groups’ knowledge diversity at
the rm level, based on the identication of technology classes and
associated inventor groups (Henderson and Cockburn, 1994), as well as
to establish the overall success of a rm’s innovation activities (Caner
et al., 2017).
3.1. Data collection
We compiled an initial list of 195 public US pharmaceutical rms
(SICs 2833–2836), for which data were available in the BoardEx data-
base, and that were, according to the Compustat database, among the
industry’s hundred largest employers at any time during the period
between 2000 and 2014. This sample ensured that we observed the vast
majority of innovation activity, employment, and assets in the phar-
maceutical industry. We shortened the study’s time panel to 2000–2011
to reduce truncation bias in patent citations (Hall et al., 2005). By doing
so, we removed two rms from the sample. Owing to the study’s focus
on rms that are actively engaged in pharmaceutical innovation, we
omitted 29 rms because they had fewer than ve active inventors over
at least one ve-year time window or because they were granted no
patents during the period of study (Carnabuci and Operti, 2013). List-
wise deletion to handle missing data removed 24 rms. Fourteen rms
were deleted because of gaps in their time panel or because they had
only one observation. The resulting sample consists of 124 rms, 34,203
inventors, 2815 top managers, and 917 rm-year observations.
We constructed a unique rm-level panel dataset that includes
detailed information on rms’ top management, R&D activities and
associated technological domains, and innovation outcomes resulting
from extensive data collection across different data sources. We rst
identied subsidiaries, joint ventures, and historical names using Se-
curities and Exchange Commission SEC 10-K lings and company
websites to construct detailed family trees of all rms (Caner et al.,
2017). Next, we collected data on all rms’ patenting activities through
extensive name matching of the entities in the family trees. All patent
documents of the US Patent and Trademark Ofce (USPTO) were
downloaded from the ofcial ReedTech website, which resulted in a
dataset covering patents granted between 1976 and 2015. Subsequently,
we matched rms to USPTO and SDC records based on company names,
USPTO assignee or SDC entity name, legal form, and country data to
company information contained in SEC lings. These data sources were
used as they match our sample of US public rms. Eventually, we
aggregated all data for our 124 focal rms and their wholly owned
subsidiaries at the ultimate parent level to capture each focal rm’s full
patenting and external R&D activity (Arora et al., 2014). We identied
rms’ executives and gathered data on their backgrounds using the
BoardEx, Execucomp, and Thomson Reuters Eikon databases; these data
were complemented with hand-collected data from a wide variety of
databases, such as company reports, SEC lings, Lexis Nexis, and
Bloomberg Executive Prole and Biography. All nancial data are from
B. Walrave et al.
Technovation 131 (2024) 102954
6
Compustat, and data on rms’ ownership structures are from the
Thomson Reuters Institutional Holdings 13F database.
3.2. Dependent variable: innovation performance
We examined patent data to assess each rm’s innovation perfor-
mance, measured as a citation-weighted patent count (Aghion et al.,
2013; Kaplan and Vakili, 2015). Our focus on patents as an indicator of
technological innovation follows the idea that “without inventions there
are no innovations” (Ahuja and Lampert, 2001: p. 524). These in-
novations can, as such, be viewed as successful when they serve as the
basis for many subsequent technical developments and innovation ini-
tiatives. Specically, we measured performance using Trajtenberg’s
(1990) citation weighted patent count: CWP
t
=∑nt
i=t(1+Ci),where each
patent i is weighed according to the subsequent citations C
i
it receives.
We took the patent application year as the observation year because this
is the closest to the actual innovation activity and summed all citations
received in the subsequent ve years (t to t +5). We only consider
patents granted before 2011, as this ensures a ve-year citation window
to prevent right censoring. Furthermore, drug patents tend to receive the
highest number of citations within three to four years after application
and receive the bulk of citations within ve years from the grant date
(Hall et al., 2005). Also, our patent dataset allowed us to correct for
patent families, as these might substantially affect patent counts.
4
3.3. Independent variable: inventor groups’ knowledge diversity
We measured inventor groups’ knowledge diversity as the extent to
which the knowledge held by a rm’s inventors is dispersed across
different technological domains (Carnabuci and Operti, 2013). We
matched our patent dataset with the disambiguated inventor names
provided by the FUNG database (see Li et al., 2014, for information on
the FUNG project). In the case of multiple patent assignees, we correctly
assigned unique inventors to our sampled rms using inventors’ his-
torical or future patent activity. In total, we identied 34,203 inventors.
We subsequently measured knowledge diversity at the rm level, rela-
tively, using Teachman’s entropy index (Teachman, 1980): knowledge
diversity
t-1
=∑N
j=1Pj×ln 1
Pj, where Pj is the share of the rm’s inventors
who led a patent in technology class j during the previous ve years,
summed over the total number of patent classes N in a rm’s patent
stock in this period (Carnabuci and Operti, 2013). We considered both
primary and secondary patent classications at a three-digit class level.
The index approaches ln(N) when the inventors are fully dispersed over
distinct technological domains (i.e., no knowledge groups exist). This is
a direct measure of inventor groups’ knowledge diversity that considers
the total number of technology classes and the distribution of inventors
over these classes, or knowledge groups, within the rm.
5
This is also
in-line with Harrison and Klein’s (2007) notion of diversity as ‘variety,’
meaning that members of different groups differ from each other qual-
itatively, that is, on a categorical variable like for example domain of
expertise, functional background or source of external information.
3.4. Moderator variables: administrative intensity, hierarchical structure,
and functional structure
We measured the three structural attributes of rms’ TMTs based on
data about each rm’s TMT. Our database includes 2815 different top
managers. We operationalized each rm’s TMT as consisting of execu-
tives who had an executive directorship or worked at the level of senior
vice president or higher (i.e., chairperson, vice-chairperson, CEO, CFO,
executive vice president, and senior vice president). Following Ham-
brick et al. (2015), when a team consisted of ve executives or less, we
also included executives with a vice president title.
6
This procedure
maintains consistency across rms to identify top management as the
CEO and the executives with whom (s)he regularly interacts to make and
implement important strategic decisions (Williams et al., 2017),
including matters such as stimulating knowledge exchange across
diverse groups of inventors for innovation.
Administrative intensity was measured as the number of top managers
divided by the number of inventors. This measure was adapted from
Blau and Schoenherr’s (1971) administrative ratio measure (i.e., the
ratio of administrators to employees). Classical sociological studies rst
introduced the concept of administrative intensity as a reection of the
intensity of coordination issues that rms have to manage (Sine et al.,
2006).
Hierarchical structure was determined by standardizing and averaging
the following two indicators: (1) number of distinct hierarchical levels as
indicated by the title gradations in the management team each year,
always including a CEO and possibly including COO, EVPs, SVPs, and
VPs; and (2) the presence of a COO; a value of 1 was given if a COO was
present and 0 if one was not (Hambrick et al., 2015). Notably, the
presence of a COO represents an important aspect of the hierarchical
structure of TMTs, as it indicates a structural distinction between
strategy formulation and implementation, adds an organizational layer
to management teams, and splits the reporting structure in and to the
team (see Hambrick and Cannella, 2004).
Functional structure: Each team’s functional roles were coded based
on Menz (2012). Functional structure was then calculated as the total
number of functional roles that exist within the TMT divided by the total
number of top managers.
7
3.5. Control variables
We controlled for variables that are common in research on search,
innovation, and TMTs (Aghion et al., 2013; Barney et al., 2018). These
include rm size (log. of the number of employees), rm age (years since
IPO or rst recording in Compustat), nancial performance (return on
assets, as net income divided by total assets), nancial slack (current
ratio, as current assets divided by current liabilities), R&D expenditure
4
Even after these extensive efforts and manual checks of patent data, some
citation-weighted patent counts of a few big-pharma rm observations remain
outliers in our dataset. The 99th percentile of the citation count variable con-
sists of eight observations that all relate to Johnson & Johnson. The 95th
percentile of this variable mainly concerns observations of other big pharma
rms. However, winsorizing the dependent variable at the 99th or 95th per-
centiles or removing these outliers from the sample did not substantially affect
our results. (See Appendix I. Note that all Appendices will be made available
online after acceptance of this manuscript.)
5
Patents typically have more than one assignee. This does not substantially
inuence our measure, however, as we consider the count of patent applica-
tions, in each domain, per inventor.
6
To study if the selective inclusion of VPs did not introduce any bias, we
conducted a sensitivity test (Hambrick et al., 2015). Notable, we included a
dummy variable that was coded 1 in case ‘vice presidents’ were included in our
measures on TMT structural attributes. The result obtained from this model is
highly similar to the results presented here (see Appendix II).
7
Notably, this procedure implies that TMTs that include EVPs, SVPs, or VPs
that are responsible for the same functional domains increase our measure of
functional structure. For instance, consider a team of 10, from which 1 SVP and
1 VP are responsible for innovation. Let us assume that the other 8 TMT
members all have ‘generic’ titles. In that case, our measure would be 2/10 (and
not 1/10). As an alternative measure, we adopted Hambrick et al.’s (2015) TMT
horizontal interdependence structure index, which was created by standard-
izing and averaging two indicators: (1) functional structure, which was coded 1
if the team was based entirely on functional roles or 0 if the team consisted of
multiple general managers; and (2) functional titles, which was the proportion
of functional titles within the senior management team. Although this index
measure resulted in a skewed distribution and troubled the interpretation of its
coefcient, it resulted in similar ndings (see Appendix III).
B. Walrave et al.
Technovation 131 (2024) 102954
7
(log. of R&D dollars invested by a rm), acquisitions (the absolute
number), diversication (Teachman’s entropy index to calculate the
proportional distribution of rm sales over business and geographical
segments), and board independence (number of independent directors
divided by board size). We also controlled for TMT size (number of TMT
members), TMT age (average TMT members’ age), functional heteroge-
neity (Herndahl-Hirschman index to calculate the concentration of
TMT members’ primary functional backgrounds), tenure heterogeneity
(standard deviation of each executive’s number of years in the TMT),
and proportion PhDs (proportion of executives holding a Ph.D. or M.D.
before TMT appointment). We included the number of inventors and
patent classes to control for size-related factors in each rm’s knowledge
diversity and patenting activity (Carnabuci and Operti, 2013). We also
controlled for the number of granted patents in year t, dated by appli-
cation year, and corrected by patent families (Caner et al., 2017)—
meaning that the coefcient estimates for other independent variables
capture marginal contributions to the mean impact of a rm’s innova-
tion performance. Finally, we account for variation across industry
segments and time by including a full set of four-digit SIC and year
dummies. All explanatory variables and controls were lagged by one
year to reduce possible simultaneity biases and to allow for the inuence
of the explanatory variables to become observable.
3.6. Analysis
We used generalized estimating equations (GEE) models to analyze
our longitudinal data because we had multiple observations for each
rm that may be correlated over repeated measures. We specied a
negative binomial distribution with a log-link function because the
mean for innovation performance, standard deviation, and the
likelihood-ratio test all indicate overdispersion of our count-based
dependent variable. To control for unobserved heterogeneity between
rms, we introduced xed-effects by including the pre-sample mean-
scaling estimator (Blundell et al., 1995), and we exploited our long
pre-sample history on patenting behavior (up to 25 years per rm) to
include a pre-sample average of citation-weighted patents. GEE makes it
possible to account for rm-specic factors reected in any remaining
correlation or heteroscedasticity between the residuals within the rm,
which the xed-effects estimator does not consider. We clustered robust
standard errors by rm and modeled rst-order serial autocorrelation
because the Wooldridge test for serial correlation in panel-data models
reported a signicant test statistic.
We exploited our rich panel dataset to control for endogeneity in two
ways. First, the panel structure of our data enables to address the risk of
simultaneity bias, as all our independent variables and controls are
lagged by one year. Second, to address concerns related to omitted
variable bias, our dataset enabled the use of a ve-year lag of inventor
groups’ knowledge diversity as an instrument (Bettis et al., 2014). This
instrument satises the exclusion criteria because it is unlikely that
depreciated knowledge diversity affects rms’ innovation performance
directly (Caner et al., 2017). As expected, our instrument is positively
and signicantly correlated with our knowledge diversity variable (r =
0.77; p =0.000). The Kleibergen-Paap rk Wald F statistic, which can be
used as an indicator of instrument strength in models with robust
standard errors, clearly exceeded the Stock and Yogo (2005) critical
value for a maximal instrumental variable IV bias of 10 percent (i.e.,
61.19 >16.38). This conrms that our instrument is relevant and that
our IV estimates are not severely biased due to weak instruments. Also,
the Davidson and MacKinnon test of endogeneity is insignicant, which
shows that the parameter estimates of knowledge diversity are not
biased by endogeneity. This test compares the estimated coefcient of
the assumed endogenous regressor from an OLS panel regression with
the estimate obtained using a two-stage instrumental variable panel
regression. The null hypothesis is that the OLS regression yields
consistent parameter estimates. The results of the IV regressions are
consistent with the reported ndings in our main analyses (see Appendix
V). Based on these measures, we carefully conclude that our results are
not signicantly affected by endogeneity related biases.
We estimated the interaction effects by hierarchically entering their
interaction terms into our models (see Table 2). Model 1 contains only
control variables. Model 2 includes the main predictor, inventor groups’
knowledge diversity, to assess the baseline model. Models 3, 4, and 5
each introduce one of the three interaction terms. Finally, Model 6 in-
cludes all variables and interaction terms. We report Wald chi-square
statistics to test overall model signicance and further include the
quasi-likelihood under the independence model criterion. Notably, the
QIC decreases by 16 (from 9630 to 9614) when the three TMT structure
variables are added to the model with control variables and the
knowledge diversity variable. The QIC further decreases by 67 (from
9632 to 9565) when all variables and all moderation terms are included
(compare models 2 and 6 in Table 2). Both decreases signal that the nal
model, that considers a TMT’s structural attributes, underlying its ability
to act as transformational leader to enable knowledge diversity for
innovation, is the most parsimonious one (Cui and Qian, 2007).
Table 1 provides descriptive statistics and correlations. The mean-
variance ination factor (VIF) of 4.49 indicated potential multi-
collinearity between the variables of patent classes inventors, rm size,
and pre-sample patent stock. We kept these variables in our models
because of the importance of controlling for size-related factors. We
investigated the potential impact of multicollinearity in two ways. First,
we estimated models with orthogonalized variables of the indicated
variables using a modied Gram-Schmidt procedure (Sine et al., 2006).
This technique’ partials out’ the common variance between collinear
variables. The resulting VIF in the models was 3.61, which is well below
the commonly maintained threshold of 10 (see Appendix IV) (Gururati,
2005). Second, we also specied models that include only one collinear
variable as a control, followed by a model without these variables
(Kalnins, 2018). The signs, signicance, and magnitudes of estimates
remained highly consistent in all models (see Appendix V), indicating
that multicollinearity did not substantially affect our results.
4. Results
Table 2 reports our results. With respect to our baseline Hypothesis 1,
models 2–5 show a signicant positive relationship between inventor
groups’ knowledge diversity and innovation performance. However,
this is not the case for model 6. And while we cannot conrm Hypothesis
1, this nding underscores our central thesis that the relationship be-
tween knowledge diversity and innovation performance can only be
comprehensively understood when including TMT structural attrib-
utes—as excluding these attributes results in omitted variable bias. We
now continue to discuss the results obtained from Model 6.
Model 6 reports that hierarchical structure as a moderator has a
positive yet insignicant effect on the relationship between knowledge
diversity and innovation performance (β =0.156; p =0.211). This
nding provides no support for Hypothesis 2. Model 6 also illustrates a
strong positive moderation effect of functional structure on the relation
between knowledge diversity and innovation performance (β =1.279; p
=0.008). This result provides support for Hypothesis 3. Furthermore,
we nd a negative moderation effect of administrative intensity on the
relationship between inventor groups’ knowledge diversity and inno-
vation performance (β = − 1.483; p =0.000). This nding validates
Hypothesis 4, which predicted such a negative moderation effect.
Finally, Model 6 include all three interaction terms, and the signs and
magnitude of the coefcients remain highly consistent to those observed
in models 3–5, in which we only included one interaction term at the
time. This outcome provides further support for hypotheses 3 and 4.
Since we estimated non-linear models, we tested the signicance of
the interaction terms by plotting marginal effects at the means (MEM)
with 95 percent condence intervals using the estimates of Model 6. The
predicted values of innovation performance are calculated over the
entire range of values for knowledge diversity, when the moderating
B. Walrave et al.
Technovation 131 (2024) 102954
8
Table 1
Descriptive statistics and correlation matrix.
Mean S.D. Min Max 1 2 3 4 5 6 7 8 9 10 11
1 Innovation performance 320.21 1423.39 0.00 17961.00
2 Knowledge diversity 2.14 0.62 0.00 4.08 0.43
3 Administrative intensity 0.22 0.23 0.00 1.60 −0.18 −0.61
4 Hierarchical structure 0.00 0.70 −1.42 1.30 −0.24 −0.19 0.16
5 Functional structure 0.87 0.16 0.14 1.00 −0.53 −0.45 0.13 0.13
6 Firm size
i
8535.15 21648.32 12.00 122200.00 0.41 0.64 −0.42 −0.22 −0.48
7 Firm age 25.84 35.36 0.00 161.00 0.43 0.50 −0.30 −0.19 −0.42 0.74
8 Financial performance −0.13 0.30 −1.33 0.76 0.16 0.27 −0.07 −0.02 −0.25 0.54 0.32
9 Financial slack 5.75 6.53 0.37 64.14 −0.11 −0.20 0.16 0.13 0.14 −0.36 −0.24 −0.05
10 R&D expenditure
i
0.56 1.43 0.00 12.18 0.37 0.55 −0.54 −0.18 −0.28 0.77 0.67 0.22 −0.21
11 Acquisitions 0.45 1.02 0.00 8.00 0.45 0.40 −0.19 −0.10 −0.37 0.51 0.47 0.22 −0.16 0.44
12 Diversication 0.68 0.85 0.00 3.66 0.31 0.54 −0.24 −0.11 −0.46 0.72 0.55 0.46 −0.32 0.41 0.45
13 Board independence 0.83 0.09 0.50 1.00 0.01 0.21 −0.15 −0.13 0.04 0.06 0.08 0.00 −0.07 0.09 0.00
14 TMT size 8.35 2.94 3.00 23.00 0.33 0.51 −0.28 0.02 −0.37 0.60 0.49 0.24 −0.17 0.63 0.40
15 TMT age 49.97 3.49 38.60 60.50 0.11 0.15 −0.20 0.01 −0.06 0.27 0.33 0.04 −0.26 0.28 0.14
16 Functional heterogeneity 0.81 0.07 0.48 0.95 0.22 0.39 −0.13 0.06 −0.29 0.54 0.39 0.26 −0.16 0.54 0.29
17 Tenure heterogeneity 3.77 1.85 0.00 11.36 0.01 0.05 −0.06 0.03 −0.07 0.13 0.11 0.13 0.01 0.05 0.03
18 Proportion PhDs 0.40 0.21 0.00 1.00 −0.20 −0.23 −0.06 0.15 0.25 −0.41 −0.31 −0.29 0.28 −0.13 −0.16
19 Inventors 238.93 536.00 5.00 3801.00 0.71 0.60 −0.35 −0.28 −0.51 0.70 0.77 0.26 −0.21 0.68 0.61
20 Classes 21.56 27.96 1.00 186.00 0.74 0.77 −0.42 −0.28 −0.60 0.73 0.71 0.30 −0.23 0.64 0.57
21 Granted patents 26.35 67.25 0.00 557.00 0.83 0.57 −0.30 −0.27 −0.54 0.62 0.65 0.24 −0.18 0.58 0.52
22 Presample patent stock 138.97 262.86 0.48 1582.67 0.72 0.62 −0.37 −0.24 −0.54 0.67 0.72 0.23 −0.20 0.64 0.54
12 13 14 15 16 17 18 19 20 21
13 Board independence 0.09
14 TMT size 0.42 0.12
15 TMT age 0.18 0.05 0.13
16 Functional heterogeneity 0.33 0.10 0.83 0.11
17 Tenure heterogeneity 0.03 −0.16 0.00 0.27 −0.01
18 Proportion PhDs −0.37 −0.14 −0.19 −0.07 −0.33 0.08
19 Inventors 0.51 0.11 0.53 0.24 0.38 0.03 −0.23
20 Classes 0.59 0.14 0.54 0.21 0.40 0.02 −0.30 0.92
21 Granted patents 0.45 0.09 0.46 0.19 0.32 0.01 −0.21 0.91 0.89
22 Presample patent stock 0.48 0.10 0.53 0.20 0.38 0.01 −0.22 0.92 0.90 0.88
Note: Correlations greater than 0.06 are signicant at p <0.05 and those greater than 0.08 are signicant at p <0.01.
i
Log-transformed variable but original values reported here (R&D expenditure in $m).
B. Walrave et al.
Technovation 131 (2024) 102954
9
variable is low or high (one SD below or above its mean), while all other
variables were held constant at their means. Fig. 1 shows that the pos-
itive relationship between knowledge diversity and innovation perfor-
mance decreases as administrative intensity increases. The MEM effect
of knowledge diversity on innovation performance decreases by 56.2
percent and 69.8 percent when administrative intensity increases from
low to mean and mean to high, respectively. Fig. 2 shows that when
functional structure increases from low to mean and from mean to high,
the MEM effect of knowledge diversity on innovation performance in-
creases by 42.2 percent and 26.9 percent, respectively. These ndings
provide further support for hypotheses 3 and 4—and are demonstrative
of how TMT structural attributes inuence TMT’s ability to act as
transformation leader, to address incommensurability among their in-
ventor groups, to enable recombinatory search for innovation.
5. Discussion and conclusions
Whereas the literature has meanwhile demonstrated that organiza-
tions have a wide range of instruments and measures available to sup-
port rich knowledge exchange among groups of inventors, perspective-
taking still does not come naturally in most rms. Whereas perspective-
taking activities are essential for effective recombinatory search, they
are mostly driven out by the emphasis on perspective-making by
specialized groups of inventors, that is, in-group knowledge develop-
ment and specialization activities (Boland and Tenkasi, 1995; Huang,
2009). When left unaddressed, this creates a risk of incommensurability
across inventor groups, which jeopardizes a rm’s innovation perfor-
mance and thereby its future competitiveness and viability (Ahuja and
Lampert, 2001; Granstrand, 1998; Grigoriou and Rothaermel, 2017).
Table 2
Results.
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
DV: Innovation Performance β se β se β se β se β se β se
Knowledge diversity (KD) 0.497 0.179 0.514 0.177 −0.872 0.486 1.033 0.255 −0.170 0.572
KD*Hierarchical structure 0.057 0.120 0.156 0.124
KD*Functional structure 1.414 0.494 1.279 0.485
KD*Administrative intensity −1.472 0.340 −1.483 0.385
Firm size
i
0.172 0.110 0.158 0.106 0.159 0.106 0.121 0.103 0.130 0.105 0.103 0.101
Firm age −0.005 0.003 −0.005 0.003 −0.005 0.003 −0.005 0.003 −0.004 0.003 −0.004 0.003
Financial performance 0.054 0.195 0.056 0.201 0.058 0.200 0.081 0.195 0.026 0.204 0.047 0.194
Financial slack 0.005 0.005 0.004 0.006 0.004 0.006 0.003 0.006 0.005 0.006 0.004 0.006
R&D expenditure
i
0.219 0.059 0.221 0.058 0.221 0.058 0.233 0.059 0.174 0.060 0.181 0.060
Acquisitions −0.022 0.030 −0.020 0.030 −0.021 0.030 −0.024 0.029 −0.012 0.031 −0.018 0.029
Diversication −0.059 0.095 −0.069 0.098 −0.072 0.097 −0.067 0.099 −0.047 0.100 −0.054 0.100
Board independence −0.714 0.566 −0.621 0.552 −0.614 0.548 −0.618 0.556 −0.811 0.608 −0.805 0.597
TMT size 0.050 0.028 0.046 0.027 0.044 0.027 0.051 0.027 0.053 0.026 0.050 0.026
TMT age −0.024 0.017 −0.025 0.017 −0.025 0.017 −0.025 0.016 −0.017 0.016 −0.017 0.016
Functional heterogeneity −1.714 1.369 −1.704 1.325 −1.648 1.347 −1.847 1.317 −1.263 1.324 −1.277 1.344
Tenure heterogeneity 0.012 0.032 0.006 0.032 0.005 0.032 0.012 0.032 0.011 0.033 0.017 0.033
Proportion PhDs −0.169 0.351 −0.125 0.344 −0.119 0.347 −0.282 0.351 −0.066 0.350 −0.194 0.353
Administrative intensity −0.894 0.386 −0.556 0.406 −0.522 0.409 −0.560 0.416 1.418 0.578 1.523 0.661
Hierarchical structure −0.108 0.087 −0.104 0.088 −0.228 0.268 −0.126 0.089 −0.115 0.090 −0.474 0.294
Functional structure −0.024 0.368 −0.028 0.366 −0.041 0.363 −3.254 1.101 −0.071 0.346 −3.022 1.088
Inventors −0.001 0.000 −0.001 0.000 −0.001 0.000 −0.001 0.000 −0.000 0.000 −0.000 0.000
Classes 0.014 0.008 0.001 0.009 0.001 0.010 0.008 0.009 −0.012 0.010 −0.006 0.010
Granted patents 0.012 0.002 0.013 0.002 0.013 0.002 0.013 0.002 0.013 0.002 0.014 0.002
Presample patent stock 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001
Constant 3.360 1.369 2.427 1.365 2.354 1.364 5.716 1.890 1.770 1.364 4.606 1.972
Wald chi-square 889.8 906.0 956.8 960.9 990.4 1119.5
QIC 9637.9 9614.3 9619.2 9610.3 9567.6 9563.7
Note: Table shows coefcients and robust standard errors in parentheses clustered by rms. All models include SIC and time dummies.
i
Log-transformed variable.
Fig. 1. Marginal effect of knowledge diversity given administrative intensity.
Fig. 2. Marginal effect of knowledge diversity given functional structure.
B. Walrave et al.
Technovation 131 (2024) 102954
10
This makes incommensurability among inventor groups a key concern
for TMTs.
The implication for a TMT is that it not only needs to equip groups of
inventors by providing them with different organizational measures to
connect and create linkages among them to engage in perspective-
taking, but it also needs to motivate them to execute on this. This de-
mands for a TMT to act as a transformational leader and motivate and
inspire inventor groups to transcend their self-interest to overcome
incommensurability and to engage in perspective-taking (Boland and
Tenkasi, 1995; Hoever et al., 2012). In our paper, we argued and showed
how a TMT’s structural attributes shape its information-processing ca-
pacity to be effective as transformational leaders in motivating and
inspiring inventor groups to overcome incommensurability and to
engage in perspective-taking. Following this, several results stand out.
We argued and found that TMT’s functional structure has a positive
moderating effect. A functionally structured TMT is well equipped to act
as transformation leader, by being able to draw on the varied perspec-
tives, insights, and skills of its members. This enables a TMT to allow for
the alignment of diverse beliefs and skill sets, and thereby strengthen the
alignment of the organization at large, and allow for the signaling of
clear and compelling goals and vision. This then enables functionally
structured TMTs to address complex challenges such as incommensu-
rability across inventor groups, through promotion of interdisciplinary
collaborations by breaking down silos and signaling of alignment, and to
address and manage conicts among these groups. This strengthens
perspective-taking, which will amplify the positive relationship between
knowledge diversity and innovation performance.
On the other hand, as predicted, we found that a TMT’s adminis-
trative intensity has a negative moderating effect on the relationship
between knowledge diversity and innovation performance. We argued
that a high TMT administrative intensity is associated by a high
administrative load that will consume a substantial portion of a TMT’s
time and attention, leaving less room for vision and strategy develop-
ment and sharing this across the organization to stimulate organiza-
tional alignment, and to stimulate collaboration and trust-building
among inventor groups. All of which is needed for inventor groups to
engage in perspective-taking activities.
Overall, these ndings suggest that the moderation of the relation-
ship between inventor groups’ knowledge diversity and innovation
performance needs to be understood by how two key TMT’s structural
attributes, that is, its administrative intensity and its functional struc-
ture, inuence its ability to act as transformational leader to address the
risk of incommensurability, for innovation. In contrast to our expecta-
tions, we did not nd that a TMT’s hierarchical structure moderates the
relationship between knowledge diversity and innovation performance.
An interpretation of this latter non-nding may be as follows. While
more hierarchy within a TMT might constrain the vertical information
ow, it also allows top managers to better serve as role-models, by
having great visibility and exposure. Moreover, a stronger hierarchy
might possibly elevate the potential for conict resolution and enhanced
decision-making speed, also with respect to resource decisions (Hen-
derson and Cockburn, 1994), which might have a positive effect on
addressing the risk of incommensurability. As a net result, the positive
and negative effects might cancel each other out, resulting in a
non-signicant nding for hierarchical structure in our analyses. We
leave this reasoning as an interesting direction for future work.
We contribute to the literature on knowledge diversity and recom-
binatory search (Ahuja and Lampert, 2001; Carnabuci and Operti, 2013;
Xiao et al., 2022). In this literature, there is a dominant focus on how
collaboration, networks, and teams of inventors inuence the relation-
ship between knowledge diversity and innovation (Carnabuci and
Operti, 2013; Moreira et al., 2018; Vakili and Kaplan, 2021). While in-
ventor groups’ knowledge diversity enables recombinatory search for
innovation, it does not mean that inventor groups are motivated to
collaborate and share their knowledge. On the contrary, inventors
typically engage in perspective-making, limiting their collaborative
processes to their direct peers. So, whereas perspective-taking is what
inventors should do normatively, from a more behavioral perspective,
this is different from what most inventors are inclined or willing to do
(Boland and Tenkasi, 1995; Hoever et al., 2012; Mathieu et al., 2017).
This means that the necessary congurational and structural adjust-
ments as endorsed by this literature are likely to be less effective, or even
ineffective, when this motivational side remains unaddressed. Here, we
inform this eld of incommensurability across diverse inventor groups,
and its attendant risks of diverging interests, potential for conict and
lack of motivation, and the important role that TMTs play in addressing
these risks through transformational leadership. More specically, we
argue and show how a TMT’s structural attributes shape its
information-processing capacity to be effective as transformational
leader, and in this way moderate the relationship between knowledge
diversity and a rm’s innovation performance.
Moreover, these ndings and conclusions also contribute to a better
understanding of the role of a rm’s TMT in strategy execution for
innovation. In the growing literature on how TMTs inuence their rm’s
innovation activities and outcomes, there is a major emphasis on the
cognitive process within a TMT by studying how its compositional
characteristics inuence this information-processing and decision-
making for innovation (e.g., Kashmiri & Mahajan, 2017; Kiss et al.,
2018; Kiss et al., 2020; Zhang et al., 2021). Yet this has been largely at
the expense of looking into the motivational process of strategy execu-
tion, a process that occurs largely between a TMT and lower-level em-
ployees, including inventor groups (Pryor et al., 2007). Here, we argued
and showed that TMTs play a key role in addressing this motivational
side to knowledge exchange across inventor groups by acting as trans-
formational leader. We show that effective transformational leadership
is enabled by their structural attributes as these shape their capacity to
receive, process and send information to these different inventor groups,
on awareness of incommensurability, encouraging collaborations and
relationships, preventing and resolving conicts, and fostering a sense of
shared purpose. By means of our focus on a TMT’s
information-processing capacity to inuence the organization’s
different inventor groups, through transformational leadership, our
paper complements the literature with a dominant emphasis how its
compositional characteristics inuence cognitive processes within a
TMT such as processing of external information and arriving at strategy
formulation, rather than enacting strategy execution.
5.1. Managerial implications
The ndings of this study have important implications for manage-
rial practice. Attempts to create value in the modern organization
through innovation have led to a recent surge in management concepts
such as holacracy (Robertson, 2015), podularity (Gray and Vander Wal,
2014), teal organizations (Laloux, 2014), delayering (Ostroff, 1999),
and agile management (Rigby et al., 2016). These approaches often
consider structure as a burden and management as a cost. While these
modern approaches can indeed enhance the swiftness, speed, and
adaptiveness of organizations, it also emphasizes the use of smaller,
autonomous groups. This brings along the risk of breeding an in-group
out-group attitude, which feeds perspective-making at the expense of
perspective-taking. Hence, an overemphasis on these novel organiza-
tional forms, without a TMT safeguarding perspective-taking, in concert
with perspective-making, might undermine recombinatory search ac-
tivities that are key to innovation and sustained organizational perfor-
mance. As our study shows, for a TMT to be effective in enabling
inventor groups’ knowledge diversity for innovation, it needs to act as
transformational leader. To be effective herein, we argued and showed
that by lowering its administrative intensity while also placing a key
emphasis on different functional roles, TMTs can strengthen their ca-
pacity for receiving, processing and sending information to and from
inventor groups, in order to motivate and inspire them for
perspective-taking.
B. Walrave et al.
Technovation 131 (2024) 102954
11
5.2. Limitations and future research
One limitation of our study arises from our empirical focus on the
pharmaceutical industry. This industry ts well with our interest and
emphasis on the creation of innovations that stem from recombination
and integration of knowledge from a broad array of different techno-
logical disciplines (Henderson and Cockburn, 1994); akin to the idea
that innovation originates from recombining unconnected elements of
knowledge, rather than linking these in new ways (Carnabuci and
Operti, 2013). Yet, it also forms a somewhat unique type of industry
given its highly science-based character (Gilsing et al., 2011; Pavitt,
1984). More work is needed to study if our ndings hold true for other
contexts, such as industries that are based on mechanical and electrical
engineering, computer science, and mathematics (Gilsing et al., 2011;
Marsili, 2001).
Another limitation is that we did not measure perspective-making,
perspective-taking, transformational leadership, and incommensura-
bility directly. Future work could be directed to open the black-box that
underlies these making versus taking activities and associated group
incommensurability. Moreover, despite the fact we took several reme-
dial measures to address omitted variable bias and simultaneity bias
(reverse causality), we cannot entirely rule out endogeneity in our
empirical analysis. In this respect, inventor groups’ knowledge diversity
and innovation performance may be jointly determined. That is, while
knowledge diversity serves innovation performance, innovation per-
formance may, in turn, affect company practices or strategy formulation
processes directed toward diminishing or enlarging knowledge di-
versity. Following performance feedback theory (Greve, 1998), for
instance, negative performance generally makes a rm more inclined to
adjust its strategy and become more risk-taking—which could indeed
include an increase in its knowledge diversity. We leave this reasoning
as an interesting suggestion for future work, but it also means that our
ndings should be treated with some care and best be interpreted as
associations rather than be seen as causations.
Moreover, we studied three TMT structural attributes—which we
argued are highly relevant to TMTs’ ability to act as transformational
leader. Nevertheless, various other TMT structural attributes exist, such
as TMT reward interdependence (Hambrick et al., 2015). We leave it for
future work to continue our work and study the effects of other
attributes.
Finally, we focused on the role of top management teams to combat
incommensurability by their ability to act as transformational leaders.
This also implies that we did not focus on the role that individual senior
executives can play in this process. Our decision to focus on at the team
level follows existing research that discusses ‘transformational leader-
ship climate’ — the degree to which leaders throughout an organization
engage in transformational leadership behaviors (Menges et al., 2011).
Future research, however, may consider the inuence of R&D directors,
which in concert with the TMT need to act as transformational leaders to
enable knowledge recombination for innovation.
A nal promising direction for future research is knowledge ex-
change and search across rm boundaries. This study adopted a within-
rm perspective, but external knowledge is another important source of
knowledge diversity that may enable innovation (Faems et al., 2005;
Moreira et al., 2018). Although external collaboration for innovation
will also contribute to knowledge diversity, we submit that the task of a
TMT will not change qualitatively from what we have studied here.
However, we leave it up to future research to ascertain whether this is
the case.
CRediT authorship contribution statement
Bob Walrave: Conceptualization, Data curation, Formal analysis,
Investigation, Methodology, Supervision, Writing – original draft,
Writing – review & editing. Nino van de Wal: Conceptualization, Data
curation, Investigation, Methodology, Resources, Supervision, Writing –
original draft, Writing – review & editing. Victor Gilsing: Conceptual-
ization, Data curation, Formal analysis, Investigation, Methodology,
Writing – original draft.
Declaration of competing interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Data availability
The authors do not have permission to share data.
Appendix I. Dependent variable Winsorized at 99th percentile and 95th percentile
Model 1 Model 2 Model 3
Dependent: CWP <99th β p se β p se β p se
Knowledge diversity (KD) 0.500 (0.005) 0.179 1.042 (0.000) 0.256
KD*Administrative intensity −1.485 (0.000) 0.341
KD*Hierarchical structure
KD*Functional structure
Firm size 0.173 (0.116) 0.110 0.160 (0.132) 0.106 0.131 (0.211) 0.105
Firm age −0.005 (0.150) 0.003 −0.004 (0.156) 0.003 −0.004 (0.175) 0.003
Financial performance 0.053 (0.786) 0.195 0.055 (0.783) 0.201 0.025 (0.901) 0.204
Financial slack 0.005 (0.353) 0.005 0.004 (0.499) 0.006 0.005 (0.436) 0.006
R&D expenditure 0.219 (0.000) 0.059 0.221 (0.000) 0.058 0.174 (0.004) 0.060
Acquisitions −0.021 (0.477) 0.030 −0.019 (0.520) 0.030 −0.011 (0.716) 0.030
Diversication −0.058 (0.538) 0.095 −0.068 (0.485) 0.098 −0.045 (0.649) 0.100
Board independence −0.713 (0.208) 0.566 −0.618 (0.263) 0.552 −0.808 (0.184) 0.609
TMT size 0.050 (0.075) 0.028 0.046 (0.084) 0.027 0.053 (0.043) 0.026
TMT age −0.024 (0.160) 0.017 −0.025 (0.133) 0.017 −0.017 (0.285) 0.016
Functional heterogeneity −1.713 (0.212) 1.371 −1.704 (0.199) 1.326 −1.263 (0.341) 1.326
Tenure heterogeneity 0.012 (0.719) 0.032 0.005 (0.868) 0.032 0.011 (0.752) 0.033
Proportion PhDs −0.167 (0.635) 0.351 −0.121 (0.725) 0.345 −0.061 (0.861) 0.351
Administrative intensity −0.895 (0.020) 0.386 −0.556 (0.171) 0.407 1.434 (0.013) 0.578
Hierarchical structure −0.108 (0.218) 0.087 −0.103 (0.241) 0.088 −0.115 (0.206) 0.091
Functional structure −0.018 (0.960) 0.369 −0.021 (0.954) 0.367 −0.061 (0.859) 0.347
Inventors −0.001 (0.009) 0.000 −0.001 (0.073) 0.000 −0.000 (0.348) 0.000
(continued on next page)
B. Walrave et al.
Technovation 131 (2024) 102954
12
(continued)
Model 1 Model 2 Model 3
Dependent: CWP <99th β p se β p se β p se
Classes 0.014 (0.076) 0.008 0.001 (0.924) 0.010 −0.013 (0.209) 0.010
Granted patents 0.012 (0.000) 0.002 0.012 (0.000) 0.002 0.012 (0.000) 0.002
Presample patent stock 0.001 (0.256) 0.001 0.001 (0.298) 0.001 0.001 (0.366) 0.001
Constant 3.352 (0.014) 1.370 2.413 (0.077) 1.366 1.750 (0.200) 1.366
Model 4 Model 5 Model 6
Dependent: CWP <99th β p se β p se β p se
Knowledge diversity (KD) 0.517 (0.004) 0.178 −0.896 (0.066) 0.487 −0.195 (0.733) 0.571
KD*Administrative intensity −1.502 (0.000) 0.386
KD*Hierarchical structure 0.058 (0.628) 0.120 0.160 (0.200) 0.125
KD*Functional structure 1.443 (0.004) 0.497 1.321 (0.007) 0.486
Firm size 0.160 (0.128) 0.106 0.122 (0.235) 0.103 0.104 (0.303) 0.101
Firm age −0.004 (0.156) 0.003 −0.005 (0.135) 0.003 −0.004 (0.136) 0.003
Financial performance 0.058 (0.773) 0.200 0.081 (0.680) 0.196 0.046 (0.812) 0.194
Financial slack 0.004 (0.477) 0.006 0.003 (0.671) 0.006 0.004 (0.509) 0.006
R&D expenditure 0.221 (0.000) 0.058 0.234 (0.000) 0.059 0.181 (0.002) 0.060
Acquisitions −0.020 (0.502) 0.030 −0.024 (0.414) 0.029 −0.017 (0.553) 0.029
Diversication −0.071 (0.469) 0.098 −0.065 (0.510) 0.099 −0.051 (0.607) 0.100
Board independence −0.611 (0.265) 0.548 −0.611 (0.272) 0.556 −0.797 (0.182) 0.597
TMT size 0.044 (0.107) 0.027 0.051 (0.059) 0.027 0.051 (0.054) 0.026
TMT age −0.025 (0.136) 0.017 −0.025 (0.130) 0.016 −0.017 (0.292) 0.016
Functional heterogeneity −1.647 (0.222) 1.349 −1.853 (0.160) 1.320 −1.283 (0.341) 1.347
Tenure heterogeneity 0.005 (0.875) 0.032 0.012 (0.705) 0.032 0.016 (0.624) 0.033
Proportion PhDs −0.116 (0.739) 0.347 −0.281 (0.424) 0.351 −0.192 (0.587) 0.354
Administrative intensity −0.521 (0.204) 0.410 −0.561 (0.179) 0.417 1.549 (0.019) 0.663
Hierarchical structure −0.231 (0.389) 0.268 −0.126 (0.159) 0.089 −0.485 (0.101) 0.295
Functional structure −0.034 (0.926) 0.364 −3.313 (0.003) 1.106 −3.110 (0.004) 1.090
Inventors −0.001 (0.071) 0.000 −0.001 (0.032) 0.000 −0.000 (0.242) 0.000
Classes 0.001 (0.920) 0.010 0.008 (0.386) 0.009 −0.006 (0.528) 0.010
Granted patents 0.012 (0.000) 0.002 0.013 (0.000) 0.002 0.013 (0.000) 0.002
Presample patent stock 0.001 (0.287) 0.001 0.001 (0.167) 0.001 0.001 (0.191) 0.001
Constant 2.337 (0.087) 1.365 5.772 (0.002) 1.894 4.684 (0.018) 1.974
Model 1 Model 2 Model 3
Dependent: CWP <95th β p se β p se β p se
Knowledge diversity (KD) 0.515 (0.004) 0.180 1.082 (0.000) 0.257
KD*Administrative intensity −1.549 (0.000) 0.342
KD*Hierarchical structure
KD*Functional structure
Firm size 0.188 (0.089) 0.110 0.176 (0.097) 0.106 0.148 (0.159) 0.105
Firm age −0.004 (0.196) 0.003 −0.004 (0.211) 0.003 −0.004 (0.246) 0.003
Financial performance 0.055 (0.778) 0.195 0.057 (0.778) 0.201 0.024 (0.907) 0.205
Financial slack 0.005 (0.337) 0.005 0.004 (0.491) 0.006 0.005 (0.427) 0.006
R&D expenditure 0.222 (0.000) 0.059 0.225 (0.000) 0.058 0.176 (0.003) 0.060
Acquisitions −0.019 (0.511) 0.029 −0.019 (0.526) 0.029 −0.011 (0.712) 0.030
Diversication −0.073 (0.434) 0.093 −0.085 (0.373) 0.096 −0.061 (0.533) 0.098
Board independence −0.693 (0.221) 0.566 −0.575 (0.297) 0.551 −0.768 (0.211) 0.613
TMT size 0.048 (0.098) 0.029 0.045 (0.104) 0.027 0.052 (0.053) 0.027
TMT age −0.024 (0.155) 0.017 −0.025 (0.125) 0.017 −0.017 (0.278) 0.016
Functional heterogeneity −1.716 (0.213) 1.379 −1.715 (0.198) 1.332 −1.257 (0.346) 1.333
Tenure heterogeneity 0.011 (0.740) 0.032 0.004 (0.898) 0.032 0.009 (0.782) 0.033
Proportion PhDs −0.119 (0.736) 0.353 −0.063 (0.856) 0.347 0.007 (0.984) 0.354
Administrative intensity −0.904 (0.020) 0.389 −0.556 (0.178) 0.413 1.518 (0.009) 0.578
Hierarchical structure −0.100 (0.262) 0.089 −0.094 (0.296) 0.090 −0.104 (0.260) 0.093
Functional structure 0.013 (0.973) 0.378 0.018 (0.961) 0.377 −0.007 (0.985) 0.356
Inventors −0.001 (0.016) 0.000 −0.001 (0.091) 0.000 −0.000 (0.382) 0.000
Classes 0.013 (0.099) 0.008 0.000 (0.993) 0.010 −0.014 (0.169) 0.010
Granted patents 0.011 (0.000) 0.002 0.011 (0.000) 0.002 0.011 (0.000) 0.002
Presample patent stock 0.001 (0.331) 0.001 0.001 (0.387) 0.001 0.001 (0.486) 0.001
Constant 3.234 (0.019) 1.384 2.241 (0.105) 1.381 1.511 (0.275) 1.385
Model 4 Model 5 Model 6
Dependent: CWP <95th β p se β p se β p se
Knowledge diversity (KD) 0.542 (0.002) 0.178 −1.020 (0.040) 0.497 −0.321 (0.574) 0.570
KD*Administrative intensity −1.595 (0.000) 0.391
KD*Hierarchical structure 0.081 (0.503) 0.121 0.202 (0.111) 0.127
KD*Functional structure 1.589 (0.002) 0.511 1.518 (0.002) 0.491
Firm size 0.178 (0.093) 0.106 0.136 (0.187) 0.103 0.119 (0.239) 0.101
Firm age −0.004 (0.216) 0.003 −0.004 (0.213) 0.003 −0.003 (0.228) 0.003
Financial performance 0.059 (0.766) 0.200 0.083 (0.672) 0.195 0.044 (0.820) 0.193
Financial slack 0.004 (0.463) 0.006 0.002 (0.681) 0.006 0.004 (0.506) 0.006
R&D expenditure 0.225 (0.000) 0.058 0.239 (0.000) 0.059 0.184 (0.002) 0.060
Acquisitions −0.019 (0.500) 0.029 −0.022 (0.414) 0.027 −0.017 (0.539) 0.028
(continued on next page)
B. Walrave et al.
Technovation 131 (2024) 102954
13
(continued)
Model 4 Model 5 Model 6
Dependent: CWP <95th β p se β p se β p se
Diversication −0.088 (0.354) 0.095 −0.078 (0.421) 0.097 −0.064 (0.516) 0.098
Board independence −0.565 (0.301) 0.546 −0.568 (0.307) 0.556 −0.756 (0.207) 0.599
TMT size 0.041 (0.142) 0.028 0.049 (0.079) 0.028 0.047 (0.080) 0.027
TMT age −0.026 (0.127) 0.017 −0.025 (0.124) 0.016 −0.017 (0.283) 0.016
Functional heterogeneity −1.631 (0.230) 1.359 −1.844 (0.165) 1.330 −1.219 (0.369) 1.357
Tenure heterogeneity 0.004 (0.908) 0.032 0.011 (0.735) 0.032 0.015 (0.649) 0.033
Proportion PhDs −0.053 (0.879) 0.349 −0.231 (0.516) 0.355 −0.130 (0.715) 0.356
Administrative intensity −0.506 (0.226) 0.418 −0.564 (0.186) 0.427 1.694 (0.012) 0.671
Hierarchical structure −0.271 (0.313) 0.269 −0.117 (0.198) 0.091 −0.566 (0.058) 0.298
Functional structure 0.003 (0.993) 0.374 −3.595 (0.001) 1.121 −3.498 (0.001) 1.091
Inventors −0.001 (0.089) 0.000 −0.001 (0.036) 0.000 −0.000 (0.228) 0.000
Classes 0.000 (0.995) 0.010 0.008 (0.391) 0.009 −0.007 (0.477) 0.010
Granted patents 0.011 (0.000) 0.002 0.012 (0.000) 0.002 0.012 (0.000) 0.002
Presample patent stock 0.001 (0.373) 0.001 0.001 (0.248) 0.001 0.001 (0.299) 0.001
Constant 2.126 (0.124) 1.382 5.877 (0.002) 1.913 4.792 (0.016) 1.985
Note: n =917. Table shows coefcients, p-values and robust standard errors clustered by rms. All models include SIC and time dummies.
Appendix II. Analyses that include a dummy in case ‘vice presidents’ were included in the measure on TMT structural attributes
Model 1 Model 2 Model 3
Dependent: CWP β p se β p se β p se
Knowledge diversity (KD) 0.496 (0.006) 0.179 1.038 (0.000) 0.253
KD*Administrative intensity −1.487 (0.000) 0.336
KD*Hierarchical structure
KD*Functional structure
VP included in TMT 0.108 (0.450) 0.143 0.108 (0.444) 0.141 0.122 (0.395) 0.144
Firm size 0.172 (0.113) 0.108 0.159 (0.128) 0.104 0.130 (0.204) 0.103
Firm age −0.005 (0.134) 0.003 −0.005 (0.137) 0.003 −0.004 (0.147) 0.003
Financial performance 0.062 (0.759) 0.201 0.063 (0.759) 0.207 0.033 (0.875) 0.212
Financial slack 0.005 (0.336) 0.005 0.004 (0.484) 0.006 0.005 (0.418) 0.006
R&D expenditure 0.230 (0.000) 0.056 0.232 (0.000) 0.055 0.186 (0.001) 0.058
Acquisitions −0.022 (0.474) 0.030 −0.020 (0.513) 0.030 −0.012 (0.696) 0.031
Diversication −0.059 (0.522) 0.092 −0.069 (0.468) 0.095 −0.047 (0.630) 0.097
Board independence −0.762 (0.170) 0.556 −0.669 (0.217) 0.542 −0.868 (0.148) 0.600
TMT size 0.051 (0.065) 0.028 0.047 (0.073) 0.026 0.054 (0.036) 0.026
TMT age −0.021 (0.220) 0.017 −0.022 (0.196) 0.017 −0.013 (0.416) 0.016
Functional heterogeneity −1.768 (0.190) 1.348 −1.761 (0.178) 1.308 −1.330 (0.309) 1.308
Tenure heterogeneity 0.011 (0.742) 0.033 0.004 (0.889) 0.032 0.010 (0.773) 0.034
Proportion PhDs −0.181 (0.603) 0.349 −0.139 (0.684) 0.342 −0.084 (0.808) 0.347
Administrative intensity −0.919 (0.017) 0.383 −0.580 (0.151) 0.404 1.412 (0.014) 0.575
Hierarchical structure −0.130 (0.181) 0.097 −0.125 (0.201) 0.098 −0.140 (0.164) 0.100
Functional structure 0.038 (0.921) 0.383 0.039 (0.917) 0.376 0.010 (0.977) 0.355
Inventors −0.001 (0.008) 0.000 −0.001 (0.069) 0.000 −0.000 (0.339) 0.000
Classes 0.014 (0.073) 0.008 0.001 (0.897) 0.010 −0.012 (0.225) 0.010
Granted patents 0.013 (0.000) 0.002 0.013 (0.000) 0.002 0.013 (0.000) 0.002
Presample patent stock 0.001 (0.275) 0.001 0.001 (0.316) 0.001 0.001 (0.388) 0.001
Constant 3.089 (0.030) 1.424 2.151 (0.132) 1.428 1.445 (0.320) 1.452
Model 4 Model 5 Model 6
Dependent: CWP β p se β p se β p se
Knowledge diversity (KD) 0.513 (0.004) 0.178 −0.864 (0.079) 0.492 −0.150 (0.794) 0.574
KD*Administrative intensity −1.504 (0.000) 0.379
KD*Hierarchical structure 0.061 (0.604) 0.118 0.160 (0.197) 0.124
KD*Functional structure 1.407 (0.004) 0.495 1.269 (0.009) 0.483
VP included in TMT 0.112 (0.427) 0.141 0.094 (0.512) 0.143 0.118 (0.408) 0.143
Firm size 0.159 (0.124) 0.104 0.121 (0.232) 0.102 0.104 (0.297) 0.099
Firm age −0.005 (0.136) 0.003 −0.005 (0.119) 0.003 −0.005 (0.115) 0.003
Financial performance 0.066 (0.746) 0.205 0.088 (0.663) 0.201 0.055 (0.786) 0.201
Financial slack 0.004 (0.462) 0.006 0.003 (0.657) 0.006 0.004 (0.488) 0.006
R&D expenditure 0.232 (0.000) 0.055 0.243 (0.000) 0.056 0.193 (0.001) 0.058
Acquisitions −0.021 (0.494) 0.030 −0.024 (0.408) 0.029 −0.018 (0.538) 0.029
Diversication −0.071 (0.451) 0.095 −0.067 (0.489) 0.097 −0.054 (0.581) 0.097
Board independence −0.663 (0.218) 0.538 −0.660 (0.224) 0.542 −0.858 (0.144) 0.586
TMT size 0.045 (0.094) 0.027 0.052 (0.052) 0.027 0.051 (0.047) 0.026
TMT age −0.021 (0.202) 0.017 −0.022 (0.188) 0.017 −0.013 (0.433) 0.016
Functional heterogeneity −1.703 (0.199) 1.326 −1.894 (0.147) 1.305 −1.332 (0.316) 1.328
Tenure heterogeneity 0.004 (0.897) 0.032 0.011 (0.731) 0.032 0.015 (0.651) 0.034
Proportion PhDs −0.134 (0.696) 0.343 −0.293 (0.402) 0.349 −0.210 (0.548) 0.350
Administrative intensity −0.544 (0.181) 0.407 −0.581 (0.163) 0.416 1.529 (0.019) 0.651
Hierarchical structure −0.261 (0.336) 0.272 −0.144 (0.138) 0.097 −0.507 (0.092) 0.301
Functional structure 0.027 (0.941) 0.373 −3.175 (0.004) 1.112 −2.919 (0.008) 1.092
(continued on next page)
B. Walrave et al.
Technovation 131 (2024) 102954
14
(continued)
Model 4 Model 5 Model 6
Dependent: CWP β p se β p se β p se
Inventors −0.001 (0.067) 0.000 −0.001 (0.029) 0.000 −0.000 (0.219) 0.000
Classes 0.001 (0.891) 0.010 0.008 (0.381) 0.009 −0.006 (0.554) 0.010
Granted patents 0.013 (0.000) 0.002 0.013 (0.000) 0.002 0.014 (0.000) 0.002
Presample patent stock 0.001 (0.304) 0.001 0.001 (0.174) 0.001 0.001 (0.197) 0.001
Constant 2.061 (0.149) 1.428 5.448 (0.006) 1.999 4.248 (0.043) 2.100
Note: n =917. Table shows coefcients, p-values and robust standard errors clustered by rms. All models include SIC and time dummies.
Appendix III. Analysis using TMT horizontal interdependence structure index variable
Fig. 1. Distribution of TMT horizontal interdependence structure index variable.
Model 1 Model 2 Model 3
Dependent: CWP β p se β p se β p se
Knowledge diversity (KD) 0.480 (0.006) 0.176 1.027 (0.000) 0.254
KD*Administrative intensity −1.505 (0.000) 0.344
KD*Hierarchical structure
KD*Functional structure
Firm size 0.155 (0.145) 0.106 0.144 (0.164) 0.103 0.114 (0.261) 0.101
Firm age −0.005 (0.117) 0.003 −0.005 (0.125) 0.003 −0.004 (0.141) 0.003
Financial performance 0.076 (0.688) 0.188 0.075 (0.701) 0.194 0.045 (0.818) 0.197
Financial slack 0.005 (0.351) 0.005 0.004 (0.487) 0.006 0.005 (0.436) 0.006
R&D expenditure 0.229 (0.000) 0.058 0.230 (0.000) 0.057 0.182 (0.002) 0.059
Acquisitions −0.019 (0.532) 0.030 −0.017 (0.577) 0.030 −0.009 (0.770) 0.030
Diversication −0.068 (0.464) 0.093 −0.077 (0.422) 0.096 −0.055 (0.577) 0.098
Board independence −0.669 (0.237) 0.566 −0.586 (0.292) 0.556 −0.787 (0.201) 0.615
TMT size 0.045 (0.103) 0.028 0.042 (0.116) 0.027 0.049 (0.063) 0.026
TMT age −0.026 (0.130) 0.017 −0.026 (0.114) 0.017 −0.018 (0.259) 0.016
Functional heterogeneity −1.758 (0.195) 1.357 −1.747 (0.184) 1.317 −1.300 (0.322) 1.312
Tenure heterogeneity 0.010 (0.741) 0.031 0.005 (0.882) 0.031 0.009 (0.771) 0.032
Proportion PhDs −0.194 (0.570) 0.341 −0.154 (0.647) 0.336 −0.092 (0.788) 0.342
Administrative intensity −0.915 (0.021) 0.397 −0.586 (0.157) 0.414 1.429 (0.015) 0.586
Hierarchical structure −0.131 (0.133) 0.087 −0.124 (0.159) 0.088 −0.136 (0.135) 0.091
Functional structure −0.106 (0.080) 0.060 −0.097 (0.108) 0.060 −0.105 (0.074) 0.059
Inventors −0.001 (0.011) 0.000 −0.001 (0.083) 0.000 −0.000 (0.386) 0.000
Classes 0.013 (0.096) 0.008 0.001 (0.927) 0.010 −0.013 (0.206) 0.010
Granted patents 0.012 (0.000) 0.002 0.012 (0.000) 0.002 0.013 (0.000) 0.002
Presample patent stock 0.001 (0.292) 0.001 0.001 (0.326) 0.001 0.001 (0.393) 0.001
Constant 3.501 (0.008) 1.330 2.583 (0.051) 1.326 1.874 (0.159) 1.330
Model 4 Model 5 Model 6
Dependent: CWP β p se β p se β p se
Knowledge diversity (KD) 0.500 (0.004) 0.174 0.376 (0.035) 0.179 0.983 (0.000) 0.270
KD*Administrative intensity −1.555 (0.000) 0.382
KD*Hierarchical structure 0.070 (0.557) 0.119 0.167 (0.182) 0.125
KD*Functional structure 0.165 (0.119) 0.106 0.159 (0.108) 0.099
Firm size 0.145 (0.159) 0.103 0.121 (0.235) 0.102 0.098 (0.324) 0.100
Firm age −0.005 (0.124) 0.003 −0.005 (0.141) 0.003 −0.004 (0.147) 0.003
Financial performance 0.077 (0.691) 0.193 0.093 (0.627) 0.191 0.061 (0.746) 0.189
(continued on next page)
B. Walrave et al.
Technovation 131 (2024) 102954
15
(continued)
Model 4 Model 5 Model 6
Dependent: CWP β p se β p se β p se
Financial slack 0.004 (0.465) 0.006 0.003 (0.615) 0.006 0.004 (0.484) 0.006
R&D expenditure 0.230 (0.000) 0.057 0.239 (0.000) 0.058 0.186 (0.002) 0.059
Acquisitions −0.018 (0.550) 0.030 −0.020 (0.496) 0.029 −0.015 (0.613) 0.029
Diversication −0.080 (0.405) 0.096 −0.072 (0.461) 0.098 −0.057 (0.563) 0.099
Board independence −0.577 (0.296) 0.552 −0.606 (0.276) 0.556 −0.804 (0.182) 0.602
TMT size 0.039 (0.150) 0.027 0.046 (0.088) 0.027 0.046 (0.081) 0.026
TMT age −0.026 (0.117) 0.017 −0.027 (0.108) 0.017 −0.018 (0.267) 0.016
Functional heterogeneity −1.678 (0.210) 1.338 −1.834 (0.166) 1.326 −1.245 (0.353) 1.342
Tenure heterogeneity 0.004 (0.887) 0.031 0.007 (0.814) 0.031 0.012 (0.710) 0.033
Proportion PhDs −0.147 (0.663) 0.337 −0.259 (0.455) 0.346 −0.176 (0.612) 0.348
Administrative intensity −0.545 (0.194) 0.420 −0.602 (0.161) 0.430 1.578 (0.017) 0.661
Hierarchical structure −0.277 (0.301) 0.268 −0.135 (0.125) 0.088 −0.509 (0.086) 0.297
Functional structure −0.098 (0.096) 0.059 −0.443 (0.038) 0.214 −0.442 (0.027) 0.200
Inventors −0.001 (0.081) 0.000 −0.001 (0.044) 0.000 −0.000 (0.276) 0.000
Classes 0.001 (0.920) 0.010 0.005 (0.601) 0.010 −0.009 (0.424) 0.011
Granted patents 0.013 (0.000) 0.002 0.013 (0.000) 0.002 0.013 (0.000) 0.002
Presample patent stock 0.001 (0.306) 0.001 0.001 (0.218) 0.001 0.001 (0.226) 0.001
Constant 2.479 (0.061) 1.322 2.941 (0.035) 1.398 2.010 (0.154) 1.409
Note: n =917. Table shows coefcients, p-values and robust standard errors clustered by rms. All models include SIC and time dummies.
Appendix IV. Analysis using orthogonalized TMT structure variables
Model 1 Model 2 Model 3
Dependent: CWP β p se β p se β p se
Knowledge diversity (KD) 0.497 (0.005) 0.179 1.033 (0.000) 0.255
KD*Administrative intensity −1.472 (0.000) 0.340
KD*Hierarchical structure
KD*Functional structure
Firm size 0.226 (0.125) 0.147 0.208 (0.143) 0.142 0.170 (0.225) 0.140
Firm age −0.005 (0.142) 0.003 −0.005 (0.147) 0.003 −0.004 (0.162) 0.003
Financial performance 0.054 (0.784) 0.195 0.056 (0.780) 0.201 0.026 (0.897) 0.204
Financial slack 0.005 (0.356) 0.005 0.004 (0.500) 0.006 0.005 (0.438) 0.006
R&D expenditure 0.219 (0.000) 0.059 0.221 (0.000) 0.058 0.174 (0.004) 0.060
Acquisitions −0.022 (0.470) 0.030 −0.020 (0.510) 0.030 −0.012 (0.695) 0.031
Diversication −0.059 (0.532) 0.095 −0.069 (0.478) 0.098 −0.047 (0.637) 0.100
Board independence −0.714 (0.208) 0.566 −0.621 (0.261) 0.552 −0.811 (0.182) 0.608
TMT size 0.050 (0.074) 0.028 0.046 (0.083) 0.027 0.053 (0.043) 0.026
TMT age −0.024 (0.161) 0.017 −0.025 (0.135) 0.017 −0.017 (0.292) 0.016
Functional heterogeneity −1.714 (0.211) 1.369 −1.704 (0.198) 1.325 −1.263 (0.340) 1.324
Tenure heterogeneity 0.012 (0.715) 0.032 0.006 (0.861) 0.032 0.011 (0.744) 0.033
Proportion PhDs −0.169 (0.631) 0.351 −0.125 (0.717) 0.344 −0.066 (0.850) 0.350
Administrative intensity −0.894 (0.021) 0.386 −0.556 (0.171) 0.406 1.418 (0.014) 0.578
Hierarchical structure −0.108 (0.216) 0.087 −0.104 (0.237) 0.088 −0.115 (0.201) 0.090
Functional structure −0.024 (0.949) 0.368 −0.028 (0.939) 0.366 −0.071 (0.838) 0.346
Inventors 0.281 (0.124) 0.182 0.103 (0.587) 0.190 −0.146 (0.436) 0.188
Classes 0.252 (0.001) 0.079 0.109 (0.262) 0.097 −0.051 (0.610) 0.101
Granted patents 0.012 (0.000) 0.002 0.013 (0.000) 0.002 0.013 (0.000) 0.002
Presample patent stock 0.093 (0.251) 0.081 0.090 (0.288) 0.085 0.075 (0.349) 0.080
Constant 4.728 (0.002) 1.492 3.505 (0.016) 1.458 2.416 (0.098) 1.462
Model 4 Model 5 Model 6
Dependent: CWP β p se β p se β p se
Knowledge diversity (KD) 0.514 (0.004) 0.177 −0.872 (0.073) 0.486 −0.170 (0.766) 0.572
KD*Administrative intensity −1.483 (0.000) 0.385
KD*Hierarchical structure 0.057 (0.636) 0.120 0.156 (0.211) 0.124
KD*Functional structure 1.414 (0.004) 0.494 1.279 (0.008) 0.485
Firm size 0.209 (0.139) 0.142 0.158 (0.251) 0.138 0.135 (0.321) 0.136
Firm age −0.005 (0.147) 0.003 −0.005 (0.125) 0.003 −0.004 (0.126) 0.003
Financial performance 0.058 (0.771) 0.200 0.081 (0.677) 0.195 0.047 (0.808) 0.194
Financial slack 0.004 (0.480) 0.006 0.003 (0.671) 0.006 0.004 (0.511) 0.006
R&D expenditure 0.221 (0.000) 0.058 0.233 (0.000) 0.059 0.181 (0.002) 0.060
Acquisitions −0.021 (0.492) 0.030 −0.024 (0.405) 0.029 −0.018 (0.533) 0.029
Diversication −0.072 (0.462) 0.097 −0.067 (0.499) 0.099 −0.054 (0.591) 0.100
Board independence −0.614 (0.263) 0.548 −0.618 (0.266) 0.556 −0.805 (0.177) 0.597
TMT size 0.044 (0.105) 0.027 0.051 (0.059) 0.027 0.050 (0.055) 0.026
TMT age −0.025 (0.138) 0.017 −0.025 (0.135) 0.016 −0.017 (0.305) 0.016
Functional heterogeneity −1.648 (0.221) 1.347 −1.847 (0.161) 1.317 −1.277 (0.342) 1.344
Tenure heterogeneity 0.005 (0.867) 0.032 0.012 (0.699) 0.032 0.017 (0.618) 0.033
Proportion PhDs −0.119 (0.731) 0.347 −0.282 (0.422) 0.351 −0.194 (0.583) 0.353
Administrative intensity −0.522 (0.203) 0.409 −0.560 (0.178) 0.416 1.523 (0.021) 0.661
Hierarchical structure −0.228 (0.394) 0.268 −0.126 (0.157) 0.089 −0.474 (0.107) 0.294
(continued on next page)
B. Walrave et al.
Technovation 131 (2024) 102954
16
(continued)
Model 4 Model 5 Model 6
Dependent: CWP β p se β p se β p se
Functional structure −0.041 (0.911) 0.363 −3.254 (0.003) 1.101 −3.022 (0.005) 1.088
Inventors 0.110 (0.570) 0.194 0.213 (0.268) 0.192 −0.010 (0.962) 0.210
Classes 0.110 (0.257) 0.097 0.172 (0.055) 0.090 0.013 (0.894) 0.100
Granted patents 0.013 (0.000) 0.002 0.013 (0.000) 0.002 0.014 (0.000) 0.002
Presample patent stock 0.092 (0.277) 0.085 0.105 (0.155) 0.074 0.096 (0.170) 0.070
Constant 3.440 (0.018) 1.460 6.684 (0.001) 1.995 5.220 (0.013) 2.093
Note: n =917. Table shows coefcients, p-values and robust standard errors clustered by rms. All models include SIC and time dummies.
Appendix V. GEE negative binomial analyses following Kalnins (2018) guidelines for mitigation of multicollinearity concerns
Model 1 Model 2 Model 3 Model 4
β p se β β p se β β p se β
Knowledge diversity (KD) 0.487 (0.007) 0.180 1.020 (0.000) 0.258 0.921 (0.004) 0.322
KD*Administrative intensity −1.464 (0.000) 0.343 −1.431 (0.000) 0.346
KD*Hierarchical structure −0.035 (0.375) 0.039
KD*Functional structure 0.086 (0.585) 0.158
Administrative intensity −0.890 (0.020) 0.384 −0.567 (0.158) 0.402 1.407 (0.014) 0.574 1.353 (0.020) 0.582
Firm size 0.188 (0.091) 0.111 0.173 (0.106) 0.107 0.146 (0.169) 0.106 0.137 (0.195) 0.106
Firm age −0.005 (0.143) 0.003 −0.004 (0.148) 0.003 −0.004 (0.159) 0.003 −0.004 (0.162) 0.003
Financial performance 0.032 (0.871) 0.198 0.035 (0.863) 0.204 0.004 (0.983) 0.208 0.019 (0.926) 0.208
Financial slack 0.005 (0.371) 0.005 0.004 (0.531) 0.006 0.004 (0.480) 0.006 0.004 (0.483) 0.006
R&D expenditure 0.222 (0.000) 0.059 0.224 (0.000) 0.058 0.178 (0.004) 0.061 0.174 (0.004) 0.061
Acquisitions −0.026 (0.408) 0.031 −0.023 (0.446) 0.031 −0.016 (0.612) 0.031 −0.015 (0.626) 0.030
Diversication −0.069 (0.466) 0.094 −0.078 (0.420) 0.097 −0.058 (0.557) 0.099 −0.049 (0.625) 0.101
Board independence −0.670 (0.255) 0.588 −0.566 (0.324) 0.574 −0.750 (0.234) 0.630 −0.798 (0.197) 0.619
TMT size 0.043 (0.121) 0.027 0.040 (0.130) 0.026 0.047 (0.066) 0.025 0.055 (0.037) 0.026
TMT age −0.027 (0.112) 0.017 −0.028 (0.093) 0.017 −0.020 (0.210) 0.016 −0.018 (0.272) 0.016
Functional heterogeneity −1.894 (0.151) 1.319 −1.884 (0.141) 1.279 −1.487 (0.244) 1.276 −1.402 (0.280) 1.297
Tenure heterogeneity 0.006 (0.851) 0.032 0.000 (0.996) 0.032 0.005 (0.872) 0.033 0.010 (0.764) 0.033
Proportion PhDs −0.167 (0.630) 0.347 −0.127 (0.707) 0.340 −0.074 (0.831) 0.346 −0.079 (0.822) 0.351
Inventors −0.001 (0.010) 0.000 −0.001 (0.077) 0.000 −0.000 (0.353) 0.000 −0.000 (0.305) 0.000
Classes 0.014 (0.057) 0.007 0.001 (0.876) 0.009 −0.012 (0.215) 0.010 −0.011 (0.306) 0.011
Granted patents 0.013 (0.000) 0.002 0.013 (0.000) 0.002 0.013 (0.000) 0.002 0.013 (0.000) 0.002
Presample patent stock 0.001 (0.263) 0.001 0.001 (0.297) 0.001 0.001 (0.350) 0.001 0.001 (0.327) 0.001
Constant 3.536 (0.007) 1.317 2.609 (0.048) 1.319 1.936 (0.145) 1.327 1.853 (0.159) 1.316
Model 5 Model 6 Model 7 Model 8
β p se β β p se β β p se β
Knowledge diversity (KD) 0.642 (0.001) 0.185 0.654 (0.000) 0.185 0.575 (0.022) 0.250
KD*Administrative intensity −0.631 (0.001) 0.183
KD*Hierarchical structure 0.097 (0.421) 0.120 0.056 (0.648) 0.123
KD*Functional structure 0.088 (0.571) 0.155
Hierarchical structure −0.110 (0.201) 0.086 −0.104 (0.228) 0.086 −0.316 (0.232) 0.264 −0.225 (0.428) 0.283
Firm size 0.201 (0.083) 0.116 0.172 (0.108) 0.107 0.172 (0.105) 0.106 0.140 (0.180) 0.105
Firm age −0.006 (0.118) 0.004 −0.005 (0.134) 0.003 −0.005 (0.137) 0.003 −0.004 (0.169) 0.003
Financial performance 0.032 (0.868) 0.191 0.041 (0.835) 0.198 0.047 (0.813) 0.197 0.051 (0.803) 0.204
Financial slack 0.005 (0.352) 0.005 0.004 (0.542) 0.006 0.004 (0.501) 0.006 0.004 (0.446) 0.006
R&D expenditure 0.267 (0.000) 0.054 0.248 (0.000) 0.053 0.245 (0.000) 0.054 0.181 (0.002) 0.059
Acquisitions −0.036 (0.240) 0.030 −0.027 (0.370) 0.030 −0.028 (0.349) 0.030 −0.015 (0.615) 0.030
Diversication −0.074 (0.437) 0.096 −0.080 (0.413) 0.098 −0.082 (0.401) 0.098 −0.052 (0.595) 0.098
Board independence −0.692 (0.218) 0.562 −0.573 (0.295) 0.546 −0.574 (0.287) 0.540 −0.749 (0.199) 0.583
TMT size 0.045 (0.099) 0.027 0.043 (0.100) 0.026 0.039 (0.142) 0.027 0.051 (0.062) 0.027
TMT age −0.021 (0.241) 0.018 −0.023 (0.177) 0.017 −0.022 (0.181) 0.017 −0.022 (0.191) 0.017
Functional heterogeneity −2.176 (0.086) 1.267 −1.959 (0.115) 1.241 −1.833 (0.150) 1.275 −1.332 (0.335) 1.381
Tenure heterogeneity 0.016 (0.599) 0.031 0.006 (0.832) 0.030 0.006 (0.849) 0.031 0.008 (0.803) 0.034
Proportion PhDs −0.089 (0.797) 0.348 −0.058 (0.866) 0.343 −0.054 (0.876) 0.345 −0.144 (0.686) 0.356
Inventors −0.001 (0.003) 0.000 −0.001 (0.077) 0.000 −0.001 (0.074) 0.000 −0.001 (0.133) 0.000
Classes 0.017 (0.031) 0.008 −0.001 (0.914) 0.009 −0.001 (0.941) 0.009 −0.003 (0.781) 0.010
Granted patents 0.013 (0.000) 0.002 0.013 (0.000) 0.002 0.013 (0.000) 0.002 0.013 (0.000) 0.002
Presample patent stock 0.001 (0.183) 0.001 0.001 (0.261) 0.001 0.001 (0.244) 0.001 0.001 (0.291) 0.001
Constant 2.618 (0.038) 1.260 1.699 (0.178) 1.260 1.627 (0.197) 1.261 2.433 (0.056) 1.272
Model 9 Model 10 Model 11 Model 12
β p se β β p se β β p se β
Knowledge diversity (KD) 0.634 (0.001) 0.190 −0.601 (0.185) 0.454 −0.594 (0.214) 0.479
KD*Administrative intensity −0.649 (0.000) 0.181
KD*Hierarchical structure −0.045 (0.248) 0.039
KD*Functional structure 1.275 (0.008) 0.479 1.286 (0.008) 0.482
Functional structure 0.120 (0.743) 0.366 0.089 (0.804) 0.357 −2.800 (0.008) 1.058 −2.980 (0.005) 1.062
Firm size 0.219 (0.067) 0.120 0.188 (0.086) 0.109 0.159 (0.139) 0.107 0.107 (0.298) 0.102
(continued on next page)
B. Walrave et al.
Technovation 131 (2024) 102954
17
(continued)
Model 9 Model 10 Model 11 Model 12
β p se β β p se β β p se β
Firm age −0.005 (0.123) 0.004 −0.005 (0.137) 0.003 −0.005 (0.118) 0.003 −0.004 (0.142) 0.003
Financial performance 0.010 (0.960) 0.196 0.020 (0.923) 0.203 0.034 (0.866) 0.199 0.066 (0.741) 0.199
Financial slack 0.005 (0.362) 0.005 0.003 (0.574) 0.006 0.002 (0.748) 0.006 0.003 (0.629) 0.006
R&D expenditure 0.267 (0.000) 0.055 0.250 (0.000) 0.054 0.262 (0.000) 0.054 0.198 (0.001) 0.060
Acquisitions −0.039 (0.209) 0.031 −0.031 (0.319) 0.031 −0.035 (0.239) 0.030 −0.018 (0.553) 0.030
Diversication −0.083 (0.388) 0.096 −0.089 (0.368) 0.099 −0.090 (0.372) 0.101 −0.051 (0.606) 0.099
Board independence −0.645 (0.270) 0.585 −0.510 (0.368) 0.567 −0.491 (0.391) 0.573 −0.739 (0.214) 0.595
TMT size 0.037 (0.161) 0.027 0.037 (0.147) 0.025 0.039 (0.119) 0.025 0.055 (0.041) 0.027
TMT age −0.024 (0.176) 0.017 −0.025 (0.126) 0.017 −0.025 (0.120) 0.016 −0.023 (0.154) 0.016
Functional heterogeneity −2.343 (0.053) 1.209 −2.143 (0.072) 1.193 −2.319 (0.051) 1.190 −1.625 (0.226) 1.342
Tenure heterogeneity 0.011 (0.725) 0.031 0.001 (0.966) 0.030 0.006 (0.839) 0.030 0.013 (0.697) 0.033
Proportion PhDs −0.082 (0.813) 0.345 −0.056 (0.868) 0.340 −0.198 (0.568) 0.347 −0.288 (0.417) 0.354
Inventors −0.001 (0.004) 0.000 −0.001 (0.083) 0.000 −0.001 (0.039) 0.000 −0.001 (0.086) 0.000
Classes 0.017 (0.031) 0.008 −0.001 (0.954) 0.010 0.006 (0.531) 0.009 0.002 (0.831) 0.009
Granted patents 0.014 (0.000) 0.002 0.013 (0.000) 0.002 0.014 (0.000) 0.002 0.013 (0.000) 0.002
Presample patent stock 0.001 (0.185) 0.001 0.001 (0.259) 0.001 0.001 (0.155) 0.001 0.001 (0.201) 0.001
Constant 2.694 (0.038) 1.298 1.801 (0.171) 1.316 4.755 (0.008) 1.797 5.655 (0.002) 1.824
Note: n =917. Table shows coefcients, p-values and robust standard errors clustered by rms. All models include SIC and time dummies.
References
Acharya, C., Ojha, D., Gokhale, R., Patel, P.C., 2022. Managing information for
innovation using knowledge integration capability: the role of boundary spanning
objects. Int. J. Inf. Manag. 62, 102438.
Aghion, P., Van Reenen, J., Zingales, L., 2013. Innovation and institutional ownership.
Am. Econ. Rev. 103 (1), 277–304.
Ahuja, G., Lampert, C.M., 2001. Entrepreneurship in the large corporation: a longitudinal
study of how established rms create breakthrough innovations. Strat. Manag. J. 22
(6–7), 521–543.
Anderson, E.G., Lewis, K., 2014. A dynamic model of individual and collective learning
amid disruption. Organ. Sci. 25 (2), 356–376.
Arora, A., Belenzon, S., Rios, L.A., 2014. Make, buy, organize: the interplay between
research, external knowledge, and rm structure. Strat. Manag. J. 35 (3), 317–337.
Aryee, S., Walumbwa, F.O., Zhou, Q., Hartnell, C.A., 2012. Transformational leadership,
innovative behavior, and task performance: test of mediation and moderation
processes. Hum. Perform. 25 (1), 1–25.
Astrazeneca “https://www.astrazeneca.com/media-centre/articles/2020/researching-a
ntibodies-to-target-covid-19.html.” Last visited: August 25th, 2021.
Barney, J.B., Foss, N.J., Lyngsie, J., 2018. The role of senior management in opportunity
formation: direct involvement or reactive selection? Strat. Manag. J. 39 (5),
1325–1349.
Bass, B.M., 1999. Two decades of research and development in transformational
leadership. Eur. J. Work. Organ. Psychol. 8 (1), 9–32.
Bettis, R., Gambardella, A., Helfat, C., Mitchell, W., 2014. Quantitative empirical analysis
in strategic management. Strat. Manag. J. 35 (7), 949–953.
Blau, P.M., Schoenherr, R.A., 1971. The Structure of Organizations. Basic Books, New
York.
Blundell, R., Grifth, R., Van Reenen, J., 1995. Dynamic count data models of
technological innovation. Econ. J. 105 (429), 333–344.
Boland, R.J., Tenkasi, R.V., 1995. Perspective making and perspective taking in
communities of knowing. Organ. Sci. 6 (4), 350–372.
Brennecke, J., Rank, O., 2017. The rm’s knowledge network and the transfer of advice
among corporate inventors—a multilevel network study. Res. Pol. 46 (4), 768–783.
Bryman, A., 2011. Mission accomplished? Research methods in the rst ve years of
Leadership. Leadership 7 (1), 73–83.
Burns, J.M., 1978. Leadership. Harper & Row, New York.
Caner, T., Cohen, S.K., Pil, F., 2017. Firm heterogeneity in complex problem solving: a
knowledge-based look at innovation. Strat. Manag. J. 38 (9), 1791–1811.
Cao, Q., Simsek, Z., Zhang, H., 2010. Modelling the joint impact of the CEO and the TMT
on organizational ambidexterity. J. Manag. Stud. 47 (7), 1272–1296.
Carlile, P.R., 2004. Transferring, translating, and transforming: an integrative framework
for managing knowledge across boundaries. Organ. Sci. 15 (5), 555–568.
Carnabuci, G., Di´
oszegi, B., 2015. Social networks, cognitive style, and innovative
performance: a contingency perspective. Acad. Manag. J. 58 (3), 881–905.
Carnabuci, G., Operti, E., 2013. Where do rms’ recombinant capabilities come from?
Intraorganizational networks, knowledge, and rms’ ability to innovate through
technological recombination. Strat. Manag. J. 34 (13), 1591–1613.
Chou, T.Y., Jiang, J.J., Klein, G., Chou, S.C.T., 2011. Organizational citizenship behavior
of information system personnel: the inuence of leader-member exchange. Inf.
Resour. Manag. J. 24 (4), 77–93.
Clark, T.R., 2022. Don’t let hierarchy stie innovation. Harv. Bus. Rev.
Cortes, A.F., Herrmann, P., 2021. Strategic leadership of innovation: a framework for
future research. Int. J. Manag. Rev. 23 (2), 224–243.
Crosby, B.C., Bryson, J.M., 2010. Integrative leadership and the creation and
maintenance of cross-sector collaborations. Leader. Q. 21 (2), 211–230.
Cui, J., Qian, G., 2007. Selection of working correlation structure and best model in GEE
analyses of longitudinal data. Commun. Stat. Simulat. Comput. 36 (5), 987–996.
Currie, G., White, L., 2012. Inter-professional barriers and knowledge brokering in an
organizational context: the case of healthcare. Organ. Stud. 33 (10), 1333–1361.
Daft, R.L., Lengel, R.H., 1984. Information richness: a new approach to managerial
behavior and organization design. Res. Organ. Behav. 6, 191–233.
Daft, R.L., Weick, K.E., 1984. Toward a model of organizations as interpretation systems.
Acad. Manag. Rev. 9 (2), 284–295.
Dutton, J.E., Ashford, S.J., O’Neill, R.M., Lawrence, K.A., 2001. Moves that matter: issue
selling and organizational change. Acad. Manag. J. 44 (4), 517–554.
Faems, D., Van Looy, B., Debackere, K., 2005. Interorganizational collaboration and
innovation: toward a portfolio approach. J. Prod. Innovat. Manag. 22 (3), 238–250.
Fleming, L., 2001. Recombinant uncertainty in technological search. Manag. Sci. 47 (1),
117–132.
Fraidin, S.N., 2004. When is one head better than two? Interdependent information in
group decision making. Organ. Behav. Hum. Decis. Process. 93 (2), 102–113.
Galbraith, J.R., 1973. Designing Complex Organizations. Addison-Wesley, Reading, MA.
Gardner, W.L., Lowe, K.B., Meuser, J.D., Noghani, F., Gullifor, D.P., Cogliser, C.C., 2020.
The leadership trilogy: a review of the third decade of the leadership quarterly.
Leader. Q. 31 (1), 101379.
Garud, R., Gehman, J., Kumaraswamy, A., 2011. Complexity arrangements for sustained
innovation: lessons from 3M Corporation. Organ. Stud. 32 (6), 737–767.
Gilsing, V.A., Bekkers, R., Bodas Freitas, I.M., van der Steen, M., 2011. Differences in
technology transfer between science-based and development-based industries:
transfer mechanisms and barriers. Technovation 31 (12), 638–647.
Gittelman, M., Kogut, B., 2003. Does good science lead to valuable knowledge?
Biotechnology rms and the evolutionary logic of citation patterns. Manag. Sci. 49
(4), 366–382.
Granstrand, O., 1998. Towards a theory of the technology-based rm. Res. Pol. 27 (5),
465–489.
Granstrand, O., Patel, P., Pavitt, K., 1997. Multi-technology corporations: why they have
‘distributed’ rather than ‘distinctive core’ competencies. Calif. Manag. Rev. 39 (4),
8–25.
Grant, A.M., Berry, J.W., 2011. The necessity of others is the mother of invention:
intrinsic and prosocial motivations, perspective taking, and creativity. Acad. Manag.
J. 54 (1), 73–96.
Gray, D., Vander Wal, T., 2014. The Connected Company. O’Reilly Media, Sebastopol,
CA.
Greve, H.R., 1998. Performance, aspirations, and risky organizational change. Adm. Sci.
Q. 43 (1), 58–86.
Grigoriou, K., Rothaermel, F.T., 2017. Organizing for knowledge generation: internal
knowledge networks and the contingent effect of external knowledge sourcing. Strat.
Manag. J. 38 (2), 395–414.
Gruber, M., Harhoff, D., Hoisl, K., 2013. Knowledge recombination across technological
boundaries: scientists vs. engineers. Manag. Sci. 59 (4), 837–851.
Gumusluoglu, L., Ilsev, A., 2009. Transformational leadership, creativity, and
organizational innovation. J. Bus. Res. 62 (4), 461–473.
Gururati, D., 2005. Basic Econometrics. McGraw Hill, New York, NY.
Hall, B.H., Jaffe, A., Trajtenberg, M., 2005. Market value and patent citations. Rand J.
Econ. 36 (1), 16–38.
Hambrick, D.C., Cannella, A.A., 2004. CEOs who have COOs: contingency analysis of an
unexplored structural form. Strat. Manag. J. 25 (10), 959–979.
Hambrick, D.C., Humphrey, S.E., Gupta, A., 2015. Structural interdependence within top
management teams: a key moderator of upper echelons predictions. Strat. Manag. J.
36 (3), 449–461.
Hargadon, A.B., 2002. Brokering knowledge: linking learning and innovation. Res.
Organ. Behav. 24, 41–85.
Harrison, D.A., Klein, K.J., 2007. What’s the difference? Diversity constructs as
separation, variety, or disparity in organizations. Acad. Manag. Rev. 32 (4),
1199–1228.
B. Walrave et al.
Technovation 131 (2024) 102954
18
Harrison, D.A., Price, K.H., Gavin, J.H., Florey, A.T., 2002. Time, teams, and task
performance: changing effects of surface-and deep-level diversity on group
functioning. Acad. Manag. J. 45 (5), 1029–1045.
Henderson, R.M., 1994. Managing innovation in the information age. Harv. Bus. Rev. 72
(1), 100–105.
Henderson, R.M., Cockburn, I., 1994. Measuring competence? Exploring rm effects in
pharmaceutical research. Strat. Manag. J. 15 (S1), 63–84.
Hoever, I.J., Van Knippenberg, D., Van Ginkel, W.P., Barkema, H.G., 2012. Fostering
team creativity: perspective taking as key to unlocking diversity’s potential. J. Appl.
Psychol. 97 (5), 982–996.
Huang, C., 2009. Knowledge sharing and group cohesiveness on performance: an
empirical study of technology R&D teams in Taiwan. Technovation 29, 786–797.
Huang, Y., Chen, C., 2010. The impact of technological diversity and organizational slack
on innovation. Technovation 30, 420–428.
Kalnins, A., 2018. Multicollinearity: how common factors cause type 1 errors in
multivariate regression. Strat. Manag. J. 39 (8), 2362–2385.
Kaplan, S., Vakili, K., 2015. The double-edge sword of recombination in breakthrough
innovation. Strat. Manag. J. 36 (10), 1435–1457.
Kashmiri, S., Mahajan, V., 2017. Values that shape marketing decisions: inuence of
chief executive ofcers’ political ideologies on innovation propensity, shareholder
value, and risk. J. Market. Res. 54 (2), 260–278.
Kiss, A.N., Libaers, D., Barr, P.S., Wang, T., Zachary, M.A., 2020. CEO cognitive
exibility, information search, and organizational ambidexterity. Strat. Manag. J. 41
(12), 2200–2233.
Kogut, B., Zander, U., 1992. Knowledge of the rm, combinative capabilities, and the
replication of technology. Organ. Sci. 3 (3), 383–397.
Laloux, F., 2014. Reinventing Organizations: A Guide to Creating Organizations Inspired
by the Next Stage of Human Consciousness. Nelson parker, Brussels.
Lawrence, P.R., Lorsch, J.W., 1967. Differentiation and integration in complex
organizations. Adm. Sci. Q. 12 (1), 1–47.
Li, G.-C., Lai, R., D’Amour, A., Doolin, D.M., Sun, Y., Torvik, V.I., Yu, A.Z., Fleming, L.,
2014. Disambiguation and co-authorship networks of the US patent inventor
database (1975–2010). Res. Pol. 43 (6), 941–955.
Locke, K.D., Horowitz, L.M., 1990. Satisfaction in interpersonal interactions as a function
of similarity in level of dysphoria. J. Pers. Soc. Psychol. 58 (5), 823–831.
Maggitti, P.G., Smith, K.G., Katila, R., 2013. The complex search process of innovation.
Res. Pol. 42 (1), 90–100.
Maitlis, S., Christianson, M., 2014. Sensemaking in organizations, taking stock and
moving forward. Acad. Manag. Ann. 8 (1), 57–125.
Marsili, O., 2001. The Anatomy and Evolution of Industries: Technical Change and
Industrial Dynamics. Edward Elgar Publishing, Cheltenham, UK.
Mathieu, J.E., Hollenbeck, J.R., van Knippenberg, D., Ilgen, D.R., 2017. A century of
work teams in the Journal of Applied Psychology. J. Appl. Psychol. 102 (3),
452–467.
McGrath, J.E., Gruenfeld, D.H., 1993. Toward a dynamic and systemic theory of groups:
an integration of six temporally enriched perspectives. In: Chemers, M.M., Ayman, R.
(Eds.), Leadership Theory and Research: Perspectives and Directions, vols. 217–243.
Academic Press, Claremont, CA.
Menges, J.I., Walter, F., Vogel, B., Bruch, H., 2011. Transformational leadership climate:
performance linkages, mechanisms, and boundary conditions at the organizational
level. Leader. Q. 22 (5), 893–909.
Menz, M., 2012. Functional top management team members: a review, synthesis, and
research agenda. J. Manag. 38 (1), 45–80.
Mihalache, O.R., Jansen, J.J., Van den Bosch, F.A., Volberda, H.W., 2014. Top
management team shared leadership and organizational ambidexterity: a moderated
mediation framework. Strateg. Entrep. J. 8 (2), 128–148.
Milliken, F.J., Martins, L.L., 1996. Searching for common threads: understanding the
multiple effects of diversity in organizational groups. Acad. Manag. Rev. 21 (2),
402–433.
Moreira, S., Markus, A., Laursen, K., 2018. Knowledge diversity and coordination: the
effect of intrarm inventor task networks on absorption speed. Strat. Manag. J. 39
(9), 2517–2546.
Narayanan, V.K., Zane, L.J., Kemmerer, B., 2011. The cognitive perspective in strategy:
an integrative review. J. Manag. 37 (1), 305–351.
Nicolini, D., 2011. Practice as the site of knowing: insights from the eld of telemedicine.
Organ. Sci. 22 (3), 602–620.
Nijstad, B.A., Berger-Selman, F., De Dreu, C.K., 2014. Innovation in top management
teams: minority dissent, transformational leadership, and radical innovations. Eur. J.
Work. Organ. Psychol. 23 (2), 310–322.
Odumeru, J.A., Ogbonna, I.G., 2013. Transformational vs. transactional leadership
theories: evidence in literature. Int. Rev. Manag. Bus. Res. 2 (2), 355–361.
Ostroff, F., 1999. The Horizontal Organization: what the Organization of the Future
Actually Looks like and How it Delivers Value to Customers. Oxford University Press,
Oxford.
Paruchuri, S., Awate, S., 2017. Organizational knowledge networks and local search: the
role of intra-organizational inventor networks. Strat. Manag. J. 38 (3), 657–675.
Pavitt, K., 1984. Sectoral patterns of innovation. Res. Pol. 13 (6), 343–373.
Pisano, G.P., 2006. Science Business: the Promise, the Reality, and the Future of Biotech.
Harvard Business School Press, Boston, MA.
Pizzolitto, E., Verna, I., Venditti, M., 2023. Authoritarian leadership styles and
performance: a systematic literature review and research agenda. Manag. Rev.
Quart. 73 (2), 841–871.
Polanyi, M., 1966. The Tacit Dimension. University of Chicago Press, Chicago, IL.
Pryor, M.G., Anderson, D., Toombs, L.A., Humphreys, J.H., 2007. Strategic
implementation as a core competency: the 5P’s model. J. Manag. Res. 7 (1), 3–17.
Purser, R.E., Pasmore, W.A., Tenkasi, R.V., 1992. The inuence of deliberations on
learning in new product development teams. J. Eng. Technol. Manag. 9 (1), 1–28.
Richard, O.C., Wu, J., Markoczy, L.A., Chung, Y., 2019. Top management team
demographic-faultline strength and strategic change: what role does environmental
dynamism play? Strat. Manag. J. 40 (6), 987–1009.
Rigby, D.K., Sutherland, J., Takeuchi, H., 2016. Embracing agile. Harv. Bus. Rev. 94 (5),
40–50.
Robertson, B. J. Holacracy, 2015. The New Management System for a Rapidly Changing
World. Henry Holt and Company, New York.
Ruiz-Jim´
enez, J.M., Fuentes-Fuentes, M.D.M., Ruiz-Arroyo, M., 2016. Knowledge
combination capability and innovation: the effects of gender diversity on top
management teams in technology-based rms. J. Bus. Ethics 135, 503–515.
Savino, T., Messeni Petruzzelli, A., Albino, V., 2017. Search and recombination process to
innovate: a review of the empirical evidence and a research agenda. Int. J. Manag.
Rev. 19 (1), 54–75.
Shepherd, D.A., Mcmullen, J.S., Ocasio, W., 2017. Is that an opportunity? An attention
model of top managers’ opportunity beliefs for strategic action. Strat. Manag. J. 38
(3), 626–644.
Siangchokyoo, N., Klinger, R.L., Campion, E.D., 2020. Follower transformation as the
linchpin of transformational leadership theory: a systematic review and future
research agenda. Leader. Q. 31 (1), 101341.
Sine, W.D., Mitsuhashi, H., Kirsch, D.A., 2006. Revisiting Burns and Stalker: formal
structure and new venture performance in emerging economic sectors. Acad. Manag.
J. 49 (1), 121–132.
Singh, J., Fleming, L., 2010. Lone inventors as sources of breakthroughs: myth or reality?
Manag. Sci. 56 (1), 41–56.
Stephan, W.G., Stephan, C.W., 1985. Intergroup anxiety. J. Soc. Issues 41 (3), 157–175.
Talke, K., Salomo, S., Rost, K., 2010. How top management team diversity affects
innovativeness and performance via the strategic choice to focus on innovation
elds. Res. Pol. 39 (7), 907–918.
Taylor, A., Greve, H.R., 2006. Superman or the Fantastic Four? Knowledge combination
and experience in innovative teams. Acad. Manag. J. 49 (4), 723–740.
Teachman, J.D., 1980. Analysis of population diversity: measures of qualitative
variation. Socio. Methods Res. 8 (3), 341–362.
Teece, D. J. 1999 “Design issues for innovative rms: bureaucracy, incentives and
industrial structure.” In A. D. Chandler, P. Hagstr¨
om, &
¨
O. S¨
olvell (Eds.), The
Dynamic Firm: The Role of Technology, Strategy, Organization and Regions,
134–165, Oxford University Press, Oxford, NY..
Thomke, S., Von Hippel, E., Franke, R., 1998. Modes of experimentation: an innovation
process—and competitive—variable. Res. Pol. 27 (3), 315–332.
Toh, P.K., Polidoro, F., 2013. A competition-based explanation of collaborative invention
within the rm. Strat. Manag. J. 34 (10), 1186–1208.
Trajtenberg, M., 1990. A penny for your quotes: patent citations and the value of
innovations. Rand J. Econ. 21 (1), 172–187.
Tsui, A.S., Ashford, S., Clair, L., Xin, K., 1995. Dealing with discrepant expectations:
response strategies and managerial effectiveness. Acad. Manag. J. 38 (6),
1515–1543.
Vakili, K., Kaplan, S., 2021. Organizing for innovation: a contingency view on innovative
team conguration. Strat. Manag. J. 42 (6), 1159–1183.
Von Krogh, G., Nonaka, I., Rechsteiner, L., 2012. Leadership in organizational knowledge
creation: a review and framework. J. Manag. Stud. 49 (1), 240–277.
Walter, F., Bruch, H., 2010. Structural impacts on the occurrence and effectiveness of
transformational leadership: an empirical study at the organizational level of
analysis. Leader. Q. 21 (5), 765–782.
Wang, D., Su, Z., Guo, H., 2019. Top management team conict and exploratory
innovation: the mediating impact of market orientation. Ind. Market. Manag. 82,
87–95.
Williams, C., Chen, P.L., Agarwal, R., 2017. Rookies and seasoned recruits: how
experience in different levels, rms, and industries shapes strategic renewal in top
management. Strat. Manag. J. 38 (7), 1391–1415.
Williams, K., O’Reilly, C., 1998. Demography and diversity in organizations. A review of
40 years research. Res. Organ. Behav. 20, 77–140.
Wright, B.E., Pandey, S.K., 2010. Transformational leadership in the public sector: does
structure matter? J. Publ. Adm. Res. Theor. 20 (1), 75–89.
Xiao, T., Makhija, M., Karim, S., 2022. A knowledge recombination perspective of
innovation: review and new research directions. J. Manag. 48 (6), 1724–1777.
Zhang, Y., Sharma, P., Xu, Y., Zhan, W., 2021. Challenges in internationalization of R&D
teams: impact of foreign technocrats in top management teams on rm innovations.
J. Bus. Res. 128, 728–741.
B. Walrave et al.
